From 9e541466eed411cb5462fa9e6181c4d409e7e2ef Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:41 +0100 Subject: [PATCH 001/259] drm/i2c: tda998x: use HDMI constants This patch replaces hard coded values by hdmi constants. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index faa77f543a07..8571a66db0e8 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -526,12 +526,12 @@ tda998x_write_if(struct drm_encoder *encoder, uint8_t bit, uint16_t addr, static void tda998x_write_aif(struct drm_encoder *encoder, struct tda998x_encoder_params *p) { - uint8_t buf[PB(5) + 1]; + u8 buf[PB(HDMI_AUDIO_INFOFRAME_SIZE) + 1]; memset(buf, 0, sizeof(buf)); - buf[HB(0)] = 0x84; + buf[HB(0)] = HDMI_INFOFRAME_TYPE_AUDIO; buf[HB(1)] = 0x01; - buf[HB(2)] = 10; + buf[HB(2)] = HDMI_AUDIO_INFOFRAME_SIZE; buf[PB(1)] = p->audio_frame[1] & 0x07; /* CC */ buf[PB(2)] = p->audio_frame[2] & 0x1c; /* SF */ buf[PB(4)] = p->audio_frame[4]; @@ -544,12 +544,12 @@ tda998x_write_aif(struct drm_encoder *encoder, struct tda998x_encoder_params *p) static void tda998x_write_avi(struct drm_encoder *encoder, struct drm_display_mode *mode) { - uint8_t buf[PB(13) + 1]; + u8 buf[PB(HDMI_AVI_INFOFRAME_SIZE) + 1]; memset(buf, 0, sizeof(buf)); - buf[HB(0)] = 0x82; + buf[HB(0)] = HDMI_INFOFRAME_TYPE_AVI; buf[HB(1)] = 0x02; - buf[HB(2)] = 13; + buf[HB(2)] = HDMI_AVI_INFOFRAME_SIZE; buf[PB(1)] = HDMI_SCAN_MODE_UNDERSCAN; buf[PB(3)] = HDMI_QUANTIZATION_RANGE_FULL << 2; buf[PB(4)] = drm_match_cea_mode(mode); From bdf6345b3262d0ddbc6405fbc0fedd2941bec08e Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:40 +0100 Subject: [PATCH 002/259] drm/i2c: tda998x: add the active aspect in HDMI AVI frame The picture aspect setting was zero, which is reserved. A setting of Same As Picture makes more sense. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 8571a66db0e8..5be145ccbb08 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -551,6 +551,7 @@ tda998x_write_avi(struct drm_encoder *encoder, struct drm_display_mode *mode) buf[HB(1)] = 0x02; buf[HB(2)] = HDMI_AVI_INFOFRAME_SIZE; buf[PB(1)] = HDMI_SCAN_MODE_UNDERSCAN; + buf[PB(2)] = HDMI_ACTIVE_ASPECT_PICTURE; buf[PB(3)] = HDMI_QUANTIZATION_RANGE_FULL << 2; buf[PB(4)] = drm_match_cea_mode(mode); From f0b33b282c17337276504d6a700d0f558f1a6891 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:39 +0100 Subject: [PATCH 003/259] drm/i2c: tda998x: use ALSA IEC958 definitions and update audio frequency This patch sets the frequency as 'not indicated' instead of '48kHz' and uses the asound values in the channel status definition. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 5be145ccbb08..e5b764b95866 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -19,6 +19,7 @@ #include #include +#include #include #include @@ -649,10 +650,11 @@ tda998x_configure_audio(struct drm_encoder *encoder, reg_clear(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_CTS); /* Write the channel status */ - buf[0] = 0x04; + buf[0] = IEC958_AES0_CON_NOT_COPYRIGHT; buf[1] = 0x00; - buf[2] = 0x00; - buf[3] = 0xf1; + buf[2] = IEC958_AES3_CON_FS_NOTID; + buf[3] = IEC958_AES4_CON_ORIGFS_NOTID | + IEC958_AES4_CON_MAX_WORDLEN_24; reg_write_range(encoder, REG_CH_STAT_B(0), buf, 4); tda998x_audio_mute(encoder, true); From 2f7f730a4f0fd3376dda9266203f29ceccd0a67f Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:47 +0100 Subject: [PATCH 004/259] drm/i2c: tda998x: simplify the i2c read/write functions This patch simplifies the i2c read/write functions and permits them to be easily called in more contexts. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 320 +++++++++++++++--------------- 1 file changed, 161 insertions(+), 159 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index e5b764b95866..7df73baf3f9b 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -31,6 +31,7 @@ struct tda998x_priv { struct i2c_client *cec; + struct i2c_client *hdmi; uint16_t rev; uint8_t current_page; int dpms; @@ -329,9 +330,9 @@ struct tda998x_priv { #define TDA19988 0x0301 static void -cec_write(struct drm_encoder *encoder, uint16_t addr, uint8_t val) +cec_write(struct tda998x_priv *priv, uint16_t addr, uint8_t val) { - struct i2c_client *client = to_tda998x_priv(encoder)->cec; + struct i2c_client *client = priv->cec; uint8_t buf[] = {addr, val}; int ret; @@ -341,9 +342,9 @@ cec_write(struct drm_encoder *encoder, uint16_t addr, uint8_t val) } static uint8_t -cec_read(struct drm_encoder *encoder, uint8_t addr) +cec_read(struct tda998x_priv *priv, uint8_t addr) { - struct i2c_client *client = to_tda998x_priv(encoder)->cec; + struct i2c_client *client = priv->cec; uint8_t val; int ret; @@ -363,12 +364,10 @@ cec_read(struct drm_encoder *encoder, uint8_t addr) } static void -set_page(struct drm_encoder *encoder, uint16_t reg) +set_page(struct tda998x_priv *priv, uint16_t reg) { - struct tda998x_priv *priv = to_tda998x_priv(encoder); - if (REG2PAGE(reg) != priv->current_page) { - struct i2c_client *client = drm_i2c_encoder_get_client(encoder); + struct i2c_client *client = priv->hdmi; uint8_t buf[] = { REG_CURPAGE, REG2PAGE(reg) }; @@ -381,13 +380,13 @@ set_page(struct drm_encoder *encoder, uint16_t reg) } static int -reg_read_range(struct drm_encoder *encoder, uint16_t reg, char *buf, int cnt) +reg_read_range(struct tda998x_priv *priv, uint16_t reg, char *buf, int cnt) { - struct i2c_client *client = drm_i2c_encoder_get_client(encoder); + struct i2c_client *client = priv->hdmi; uint8_t addr = REG2ADDR(reg); int ret; - set_page(encoder, reg); + set_page(priv, reg); ret = i2c_master_send(client, &addr, sizeof(addr)); if (ret < 0) @@ -405,16 +404,16 @@ reg_read_range(struct drm_encoder *encoder, uint16_t reg, char *buf, int cnt) } static void -reg_write_range(struct drm_encoder *encoder, uint16_t reg, uint8_t *p, int cnt) +reg_write_range(struct tda998x_priv *priv, uint16_t reg, uint8_t *p, int cnt) { - struct i2c_client *client = drm_i2c_encoder_get_client(encoder); + struct i2c_client *client = priv->hdmi; uint8_t buf[cnt+1]; int ret; buf[0] = REG2ADDR(reg); memcpy(&buf[1], p, cnt); - set_page(encoder, reg); + set_page(priv, reg); ret = i2c_master_send(client, buf, cnt + 1); if (ret < 0) @@ -422,21 +421,21 @@ reg_write_range(struct drm_encoder *encoder, uint16_t reg, uint8_t *p, int cnt) } static uint8_t -reg_read(struct drm_encoder *encoder, uint16_t reg) +reg_read(struct tda998x_priv *priv, uint16_t reg) { uint8_t val = 0; - reg_read_range(encoder, reg, &val, sizeof(val)); + reg_read_range(priv, reg, &val, sizeof(val)); return val; } static void -reg_write(struct drm_encoder *encoder, uint16_t reg, uint8_t val) +reg_write(struct tda998x_priv *priv, uint16_t reg, uint8_t val) { - struct i2c_client *client = drm_i2c_encoder_get_client(encoder); + struct i2c_client *client = priv->hdmi; uint8_t buf[] = {REG2ADDR(reg), val}; int ret; - set_page(encoder, reg); + set_page(priv, reg); ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); if (ret < 0) @@ -444,13 +443,13 @@ reg_write(struct drm_encoder *encoder, uint16_t reg, uint8_t val) } static void -reg_write16(struct drm_encoder *encoder, uint16_t reg, uint16_t val) +reg_write16(struct tda998x_priv *priv, uint16_t reg, uint16_t val) { - struct i2c_client *client = drm_i2c_encoder_get_client(encoder); + struct i2c_client *client = priv->hdmi; uint8_t buf[] = {REG2ADDR(reg), val >> 8, val}; int ret; - set_page(encoder, reg); + set_page(priv, reg); ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); if (ret < 0) @@ -458,47 +457,47 @@ reg_write16(struct drm_encoder *encoder, uint16_t reg, uint16_t val) } static void -reg_set(struct drm_encoder *encoder, uint16_t reg, uint8_t val) +reg_set(struct tda998x_priv *priv, uint16_t reg, uint8_t val) { - reg_write(encoder, reg, reg_read(encoder, reg) | val); + reg_write(priv, reg, reg_read(priv, reg) | val); } static void -reg_clear(struct drm_encoder *encoder, uint16_t reg, uint8_t val) +reg_clear(struct tda998x_priv *priv, uint16_t reg, uint8_t val) { - reg_write(encoder, reg, reg_read(encoder, reg) & ~val); + reg_write(priv, reg, reg_read(priv, reg) & ~val); } static void -tda998x_reset(struct drm_encoder *encoder) +tda998x_reset(struct tda998x_priv *priv) { /* reset audio and i2c master: */ - reg_set(encoder, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); + reg_set(priv, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); msleep(50); - reg_clear(encoder, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); + reg_clear(priv, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); msleep(50); /* reset transmitter: */ - reg_set(encoder, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR); - reg_clear(encoder, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR); + reg_set(priv, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR); + reg_clear(priv, REG_MAIN_CNTRL0, MAIN_CNTRL0_SR); /* PLL registers common configuration */ - reg_write(encoder, REG_PLL_SERIAL_1, 0x00); - reg_write(encoder, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(1)); - reg_write(encoder, REG_PLL_SERIAL_3, 0x00); - reg_write(encoder, REG_SERIALIZER, 0x00); - reg_write(encoder, REG_BUFFER_OUT, 0x00); - reg_write(encoder, REG_PLL_SCG1, 0x00); - reg_write(encoder, REG_AUDIO_DIV, AUDIO_DIV_SERCLK_8); - reg_write(encoder, REG_SEL_CLK, SEL_CLK_SEL_CLK1 | SEL_CLK_ENA_SC_CLK); - reg_write(encoder, REG_PLL_SCGN1, 0xfa); - reg_write(encoder, REG_PLL_SCGN2, 0x00); - reg_write(encoder, REG_PLL_SCGR1, 0x5b); - reg_write(encoder, REG_PLL_SCGR2, 0x00); - reg_write(encoder, REG_PLL_SCG2, 0x10); + reg_write(priv, REG_PLL_SERIAL_1, 0x00); + reg_write(priv, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(1)); + reg_write(priv, REG_PLL_SERIAL_3, 0x00); + reg_write(priv, REG_SERIALIZER, 0x00); + reg_write(priv, REG_BUFFER_OUT, 0x00); + reg_write(priv, REG_PLL_SCG1, 0x00); + reg_write(priv, REG_AUDIO_DIV, AUDIO_DIV_SERCLK_8); + reg_write(priv, REG_SEL_CLK, SEL_CLK_SEL_CLK1 | SEL_CLK_ENA_SC_CLK); + reg_write(priv, REG_PLL_SCGN1, 0xfa); + reg_write(priv, REG_PLL_SCGN2, 0x00); + reg_write(priv, REG_PLL_SCGR1, 0x5b); + reg_write(priv, REG_PLL_SCGR2, 0x00); + reg_write(priv, REG_PLL_SCG2, 0x10); /* Write the default value MUX register */ - reg_write(encoder, REG_MUX_VP_VIP_OUT, 0x24); + reg_write(priv, REG_MUX_VP_VIP_OUT, 0x24); } static uint8_t tda998x_cksum(uint8_t *buf, size_t bytes) @@ -514,18 +513,18 @@ static uint8_t tda998x_cksum(uint8_t *buf, size_t bytes) #define PB(x) (HB(2) + 1 + (x)) static void -tda998x_write_if(struct drm_encoder *encoder, uint8_t bit, uint16_t addr, +tda998x_write_if(struct tda998x_priv *priv, uint8_t bit, uint16_t addr, uint8_t *buf, size_t size) { buf[PB(0)] = tda998x_cksum(buf, size); - reg_clear(encoder, REG_DIP_IF_FLAGS, bit); - reg_write_range(encoder, addr, buf, size); - reg_set(encoder, REG_DIP_IF_FLAGS, bit); + reg_clear(priv, REG_DIP_IF_FLAGS, bit); + reg_write_range(priv, addr, buf, size); + reg_set(priv, REG_DIP_IF_FLAGS, bit); } static void -tda998x_write_aif(struct drm_encoder *encoder, struct tda998x_encoder_params *p) +tda998x_write_aif(struct tda998x_priv *priv, struct tda998x_encoder_params *p) { u8 buf[PB(HDMI_AUDIO_INFOFRAME_SIZE) + 1]; @@ -538,12 +537,12 @@ tda998x_write_aif(struct drm_encoder *encoder, struct tda998x_encoder_params *p) buf[PB(4)] = p->audio_frame[4]; buf[PB(5)] = p->audio_frame[5] & 0xf8; /* DM_INH + LSV */ - tda998x_write_if(encoder, DIP_IF_FLAGS_IF4, REG_IF4_HB0, buf, + tda998x_write_if(priv, DIP_IF_FLAGS_IF4, REG_IF4_HB0, buf, sizeof(buf)); } static void -tda998x_write_avi(struct drm_encoder *encoder, struct drm_display_mode *mode) +tda998x_write_avi(struct tda998x_priv *priv, struct drm_display_mode *mode) { u8 buf[PB(HDMI_AVI_INFOFRAME_SIZE) + 1]; @@ -556,36 +555,36 @@ tda998x_write_avi(struct drm_encoder *encoder, struct drm_display_mode *mode) buf[PB(3)] = HDMI_QUANTIZATION_RANGE_FULL << 2; buf[PB(4)] = drm_match_cea_mode(mode); - tda998x_write_if(encoder, DIP_IF_FLAGS_IF2, REG_IF2_HB0, buf, + tda998x_write_if(priv, DIP_IF_FLAGS_IF2, REG_IF2_HB0, buf, sizeof(buf)); } -static void tda998x_audio_mute(struct drm_encoder *encoder, bool on) +static void tda998x_audio_mute(struct tda998x_priv *priv, bool on) { if (on) { - reg_set(encoder, REG_SOFTRESET, SOFTRESET_AUDIO); - reg_clear(encoder, REG_SOFTRESET, SOFTRESET_AUDIO); - reg_set(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); + reg_set(priv, REG_SOFTRESET, SOFTRESET_AUDIO); + reg_clear(priv, REG_SOFTRESET, SOFTRESET_AUDIO); + reg_set(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); } else { - reg_clear(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); + reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); } } static void -tda998x_configure_audio(struct drm_encoder *encoder, +tda998x_configure_audio(struct tda998x_priv *priv, struct drm_display_mode *mode, struct tda998x_encoder_params *p) { uint8_t buf[6], clksel_aip, clksel_fs, ca_i2s, cts_n, adiv; uint32_t n; /* Enable audio ports */ - reg_write(encoder, REG_ENA_AP, p->audio_cfg); - reg_write(encoder, REG_ENA_ACLK, p->audio_clk_cfg); + reg_write(priv, REG_ENA_AP, p->audio_cfg); + reg_write(priv, REG_ENA_ACLK, p->audio_clk_cfg); /* Set audio input source */ switch (p->audio_format) { case AFMT_SPDIF: - reg_write(encoder, REG_MUX_AP, 0x40); + reg_write(priv, REG_MUX_AP, 0x40); clksel_aip = AIP_CLKSEL_AIP(0); /* FS64SPDIF */ clksel_fs = AIP_CLKSEL_FS(2); @@ -594,7 +593,7 @@ tda998x_configure_audio(struct drm_encoder *encoder, break; case AFMT_I2S: - reg_write(encoder, REG_MUX_AP, 0x64); + reg_write(priv, REG_MUX_AP, 0x64); clksel_aip = AIP_CLKSEL_AIP(1); /* ACLK */ clksel_fs = AIP_CLKSEL_FS(0); @@ -607,12 +606,12 @@ tda998x_configure_audio(struct drm_encoder *encoder, return; } - reg_write(encoder, REG_AIP_CLKSEL, clksel_aip); - reg_clear(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_LAYOUT); + reg_write(priv, REG_AIP_CLKSEL, clksel_aip); + reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_LAYOUT); /* Enable automatic CTS generation */ - reg_clear(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_ACR_MAN); - reg_write(encoder, REG_CTS_N, cts_n); + reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_ACR_MAN); + reg_write(priv, REG_CTS_N, cts_n); /* * Audio input somehow depends on HDMI line rate which is @@ -625,7 +624,7 @@ tda998x_configure_audio(struct drm_encoder *encoder, adiv = AUDIO_DIV_SERCLK_16; else adiv = AUDIO_DIV_SERCLK_8; - reg_write(encoder, REG_AUDIO_DIV, adiv); + reg_write(priv, REG_AUDIO_DIV, adiv); /* * This is the approximate value of N, which happens to be @@ -640,14 +639,14 @@ tda998x_configure_audio(struct drm_encoder *encoder, buf[3] = n; buf[4] = n >> 8; buf[5] = n >> 16; - reg_write_range(encoder, REG_ACR_CTS_0, buf, 6); + reg_write_range(priv, REG_ACR_CTS_0, buf, 6); /* Set CTS clock reference */ - reg_write(encoder, REG_AIP_CLKSEL, clksel_aip | clksel_fs); + reg_write(priv, REG_AIP_CLKSEL, clksel_aip | clksel_fs); /* Reset CTS generator */ - reg_set(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_CTS); - reg_clear(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_CTS); + reg_set(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_CTS); + reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_CTS); /* Write the channel status */ buf[0] = IEC958_AES0_CON_NOT_COPYRIGHT; @@ -655,14 +654,14 @@ tda998x_configure_audio(struct drm_encoder *encoder, buf[2] = IEC958_AES3_CON_FS_NOTID; buf[3] = IEC958_AES4_CON_ORIGFS_NOTID | IEC958_AES4_CON_MAX_WORDLEN_24; - reg_write_range(encoder, REG_CH_STAT_B(0), buf, 4); + reg_write_range(priv, REG_CH_STAT_B(0), buf, 4); - tda998x_audio_mute(encoder, true); + tda998x_audio_mute(priv, true); mdelay(20); - tda998x_audio_mute(encoder, false); + tda998x_audio_mute(priv, false); /* Write the audio information packet */ - tda998x_write_aif(encoder, p); + tda998x_write_aif(priv, p); } /* DRM encoder functions */ @@ -704,19 +703,19 @@ tda998x_encoder_dpms(struct drm_encoder *encoder, int mode) switch (mode) { case DRM_MODE_DPMS_ON: /* enable video ports, audio will be enabled later */ - reg_write(encoder, REG_ENA_VP_0, 0xff); - reg_write(encoder, REG_ENA_VP_1, 0xff); - reg_write(encoder, REG_ENA_VP_2, 0xff); + reg_write(priv, REG_ENA_VP_0, 0xff); + reg_write(priv, REG_ENA_VP_1, 0xff); + reg_write(priv, REG_ENA_VP_2, 0xff); /* set muxing after enabling ports: */ - reg_write(encoder, REG_VIP_CNTRL_0, priv->vip_cntrl_0); - reg_write(encoder, REG_VIP_CNTRL_1, priv->vip_cntrl_1); - reg_write(encoder, REG_VIP_CNTRL_2, priv->vip_cntrl_2); + reg_write(priv, REG_VIP_CNTRL_0, priv->vip_cntrl_0); + reg_write(priv, REG_VIP_CNTRL_1, priv->vip_cntrl_1); + reg_write(priv, REG_VIP_CNTRL_2, priv->vip_cntrl_2); break; case DRM_MODE_DPMS_OFF: /* disable video ports */ - reg_write(encoder, REG_ENA_VP_0, 0x00); - reg_write(encoder, REG_ENA_VP_1, 0x00); - reg_write(encoder, REG_ENA_VP_2, 0x00); + reg_write(priv, REG_ENA_VP_0, 0x00); + reg_write(priv, REG_ENA_VP_1, 0x00); + reg_write(priv, REG_ENA_VP_2, 0x00); break; } @@ -834,57 +833,57 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, } /* mute the audio FIFO: */ - reg_set(encoder, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); + reg_set(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); /* set HDMI HDCP mode off: */ - reg_set(encoder, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); - reg_clear(encoder, REG_TX33, TX33_HDMI); + reg_set(priv, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); + reg_clear(priv, REG_TX33, TX33_HDMI); + reg_write(priv, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(0)); - reg_write(encoder, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(0)); /* no pre-filter or interpolator: */ - reg_write(encoder, REG_HVF_CNTRL_0, HVF_CNTRL_0_PREFIL(0) | + reg_write(priv, REG_HVF_CNTRL_0, HVF_CNTRL_0_PREFIL(0) | HVF_CNTRL_0_INTPOL(0)); - reg_write(encoder, REG_VIP_CNTRL_5, VIP_CNTRL_5_SP_CNT(0)); - reg_write(encoder, REG_VIP_CNTRL_4, VIP_CNTRL_4_BLANKIT(0) | + reg_write(priv, REG_VIP_CNTRL_5, VIP_CNTRL_5_SP_CNT(0)); + reg_write(priv, REG_VIP_CNTRL_4, VIP_CNTRL_4_BLANKIT(0) | VIP_CNTRL_4_BLC(0)); - reg_clear(encoder, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_CCIR); + reg_clear(priv, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_CCIR); - reg_clear(encoder, REG_PLL_SERIAL_1, PLL_SERIAL_1_SRL_MAN_IZ); - reg_clear(encoder, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_DE); - reg_write(encoder, REG_SERIALIZER, 0); - reg_write(encoder, REG_HVF_CNTRL_1, HVF_CNTRL_1_VQR(0)); + reg_clear(priv, REG_PLL_SERIAL_1, PLL_SERIAL_1_SRL_MAN_IZ); + reg_clear(priv, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_DE); + reg_write(priv, REG_SERIALIZER, 0); + reg_write(priv, REG_HVF_CNTRL_1, HVF_CNTRL_1_VQR(0)); /* TODO enable pixel repeat for pixel rates less than 25Msamp/s */ rep = 0; - reg_write(encoder, REG_RPT_CNTRL, 0); - reg_write(encoder, REG_SEL_CLK, SEL_CLK_SEL_VRF_CLK(0) | + reg_write(priv, REG_RPT_CNTRL, 0); + reg_write(priv, REG_SEL_CLK, SEL_CLK_SEL_VRF_CLK(0) | SEL_CLK_SEL_CLK1 | SEL_CLK_ENA_SC_CLK); - reg_write(encoder, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(div) | + reg_write(priv, REG_PLL_SERIAL_2, PLL_SERIAL_2_SRL_NOSC(div) | PLL_SERIAL_2_SRL_PR(rep)); /* set color matrix bypass flag: */ - reg_set(encoder, REG_MAT_CONTRL, MAT_CONTRL_MAT_BP); + reg_set(priv, REG_MAT_CONTRL, MAT_CONTRL_MAT_BP); /* set BIAS tmds value: */ - reg_write(encoder, REG_ANA_GENERAL, 0x09); + reg_write(priv, REG_ANA_GENERAL, 0x09); - reg_clear(encoder, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_MTHD); + reg_clear(priv, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_MTHD); /* * Sync on rising HSYNC/VSYNC */ - reg_write(encoder, REG_VIP_CNTRL_3, 0); - reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_SYNC_HS); + reg_write(priv, REG_VIP_CNTRL_3, 0); + reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_SYNC_HS); /* * TDA19988 requires high-active sync at input stage, * so invert low-active sync provided by master encoder here */ if (mode->flags & DRM_MODE_FLAG_NHSYNC) - reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_H_TGL); + reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_H_TGL); if (mode->flags & DRM_MODE_FLAG_NVSYNC) - reg_set(encoder, REG_VIP_CNTRL_3, VIP_CNTRL_3_V_TGL); + reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_V_TGL); /* * Always generate sync polarity relative to input sync and @@ -895,49 +894,49 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, reg |= TBG_CNTRL_1_H_TGL; if (mode->flags & DRM_MODE_FLAG_NVSYNC) reg |= TBG_CNTRL_1_V_TGL; - reg_write(encoder, REG_TBG_CNTRL_1, reg); + reg_write(priv, REG_TBG_CNTRL_1, reg); - reg_write(encoder, REG_VIDFORMAT, 0x00); - reg_write16(encoder, REG_REFPIX_MSB, ref_pix); - reg_write16(encoder, REG_REFLINE_MSB, ref_line); - reg_write16(encoder, REG_NPIX_MSB, n_pix); - reg_write16(encoder, REG_NLINE_MSB, n_line); - reg_write16(encoder, REG_VS_LINE_STRT_1_MSB, vs1_line_s); - reg_write16(encoder, REG_VS_PIX_STRT_1_MSB, vs1_pix_s); - reg_write16(encoder, REG_VS_LINE_END_1_MSB, vs1_line_e); - reg_write16(encoder, REG_VS_PIX_END_1_MSB, vs1_pix_e); - reg_write16(encoder, REG_VS_LINE_STRT_2_MSB, vs2_line_s); - reg_write16(encoder, REG_VS_PIX_STRT_2_MSB, vs2_pix_s); - reg_write16(encoder, REG_VS_LINE_END_2_MSB, vs2_line_e); - reg_write16(encoder, REG_VS_PIX_END_2_MSB, vs2_pix_e); - reg_write16(encoder, REG_HS_PIX_START_MSB, hs_pix_s); - reg_write16(encoder, REG_HS_PIX_STOP_MSB, hs_pix_e); - reg_write16(encoder, REG_VWIN_START_1_MSB, vwin1_line_s); - reg_write16(encoder, REG_VWIN_END_1_MSB, vwin1_line_e); - reg_write16(encoder, REG_VWIN_START_2_MSB, vwin2_line_s); - reg_write16(encoder, REG_VWIN_END_2_MSB, vwin2_line_e); - reg_write16(encoder, REG_DE_START_MSB, de_pix_s); - reg_write16(encoder, REG_DE_STOP_MSB, de_pix_e); + reg_write(priv, REG_VIDFORMAT, 0x00); + reg_write16(priv, REG_REFPIX_MSB, ref_pix); + reg_write16(priv, REG_REFLINE_MSB, ref_line); + reg_write16(priv, REG_NPIX_MSB, n_pix); + reg_write16(priv, REG_NLINE_MSB, n_line); + reg_write16(priv, REG_VS_LINE_STRT_1_MSB, vs1_line_s); + reg_write16(priv, REG_VS_PIX_STRT_1_MSB, vs1_pix_s); + reg_write16(priv, REG_VS_LINE_END_1_MSB, vs1_line_e); + reg_write16(priv, REG_VS_PIX_END_1_MSB, vs1_pix_e); + reg_write16(priv, REG_VS_LINE_STRT_2_MSB, vs2_line_s); + reg_write16(priv, REG_VS_PIX_STRT_2_MSB, vs2_pix_s); + reg_write16(priv, REG_VS_LINE_END_2_MSB, vs2_line_e); + reg_write16(priv, REG_VS_PIX_END_2_MSB, vs2_pix_e); + reg_write16(priv, REG_HS_PIX_START_MSB, hs_pix_s); + reg_write16(priv, REG_HS_PIX_STOP_MSB, hs_pix_e); + reg_write16(priv, REG_VWIN_START_1_MSB, vwin1_line_s); + reg_write16(priv, REG_VWIN_END_1_MSB, vwin1_line_e); + reg_write16(priv, REG_VWIN_START_2_MSB, vwin2_line_s); + reg_write16(priv, REG_VWIN_END_2_MSB, vwin2_line_e); + reg_write16(priv, REG_DE_START_MSB, de_pix_s); + reg_write16(priv, REG_DE_STOP_MSB, de_pix_e); if (priv->rev == TDA19988) { /* let incoming pixels fill the active space (if any) */ - reg_write(encoder, REG_ENABLE_SPACE, 0x00); + reg_write(priv, REG_ENABLE_SPACE, 0x00); } /* must be last register set: */ - reg_clear(encoder, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_ONCE); + reg_clear(priv, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_ONCE); /* Only setup the info frames if the sink is HDMI */ if (priv->is_hdmi_sink) { /* We need to turn HDMI HDCP stuff on to get audio through */ - reg_clear(encoder, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); - reg_write(encoder, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(1)); - reg_set(encoder, REG_TX33, TX33_HDMI); + reg_clear(priv, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); + reg_write(priv, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(1)); + reg_set(priv, REG_TX33, TX33_HDMI); - tda998x_write_avi(encoder, adjusted_mode); + tda998x_write_avi(priv, adjusted_mode); if (priv->params.audio_cfg) - tda998x_configure_audio(encoder, adjusted_mode, + tda998x_configure_audio(priv, adjusted_mode, &priv->params); } } @@ -946,7 +945,9 @@ static enum drm_connector_status tda998x_encoder_detect(struct drm_encoder *encoder, struct drm_connector *connector) { - uint8_t val = cec_read(encoder, REG_CEC_RXSHPDLEV); + struct tda998x_priv *priv = to_tda998x_priv(encoder); + uint8_t val = cec_read(priv, REG_CEC_RXSHPDLEV); + return (val & CEC_RXSHPDLEV_HPD) ? connector_status_connected : connector_status_disconnected; } @@ -954,29 +955,30 @@ tda998x_encoder_detect(struct drm_encoder *encoder, static int read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) { + struct tda998x_priv *priv = to_tda998x_priv(encoder); uint8_t offset, segptr; int ret, i; /* enable EDID read irq: */ - reg_set(encoder, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); + reg_set(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); offset = (blk & 1) ? 128 : 0; segptr = blk / 2; - reg_write(encoder, REG_DDC_ADDR, 0xa0); - reg_write(encoder, REG_DDC_OFFS, offset); - reg_write(encoder, REG_DDC_SEGM_ADDR, 0x60); - reg_write(encoder, REG_DDC_SEGM, segptr); + reg_write(priv, REG_DDC_ADDR, 0xa0); + reg_write(priv, REG_DDC_OFFS, offset); + reg_write(priv, REG_DDC_SEGM_ADDR, 0x60); + reg_write(priv, REG_DDC_SEGM, segptr); /* enable reading EDID: */ - reg_write(encoder, REG_EDID_CTRL, 0x1); + reg_write(priv, REG_EDID_CTRL, 0x1); /* flag must be cleared by sw: */ - reg_write(encoder, REG_EDID_CTRL, 0x0); + reg_write(priv, REG_EDID_CTRL, 0x0); /* wait for block read to complete: */ for (i = 100; i > 0; i--) { - uint8_t val = reg_read(encoder, REG_INT_FLAGS_2); + uint8_t val = reg_read(priv, REG_INT_FLAGS_2); if (val & INT_FLAGS_2_EDID_BLK_RD) break; msleep(1); @@ -985,14 +987,14 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) if (i == 0) return -ETIMEDOUT; - ret = reg_read_range(encoder, REG_EDID_DATA_0, buf, EDID_LENGTH); + ret = reg_read_range(priv, REG_EDID_DATA_0, buf, EDID_LENGTH); if (ret != EDID_LENGTH) { dev_err(encoder->dev->dev, "failed to read edid block %d: %d", blk, ret); return ret; } - reg_clear(encoder, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); + reg_clear(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); return 0; } @@ -1009,7 +1011,7 @@ do_get_edid(struct drm_encoder *encoder) return NULL; if (priv->rev == TDA19988) - reg_clear(encoder, REG_TX4, TX4_PD_RAM); + reg_clear(priv, REG_TX4, TX4_PD_RAM); /* base block fetch */ if (read_edid_block(encoder, block, 0)) @@ -1049,13 +1051,13 @@ do_get_edid(struct drm_encoder *encoder) done: if (priv->rev == TDA19988) - reg_set(encoder, REG_TX4, TX4_PD_RAM); + reg_set(priv, REG_TX4, TX4_PD_RAM); return block; fail: if (priv->rev == TDA19988) - reg_set(encoder, REG_TX4, TX4_PD_RAM); + reg_set(priv, REG_TX4, TX4_PD_RAM); dev_warn(encoder->dev->dev, "failed to read EDID\n"); kfree(block); return NULL; @@ -1141,7 +1143,6 @@ tda998x_encoder_init(struct i2c_client *client, struct drm_device *dev, struct drm_encoder_slave *encoder_slave) { - struct drm_encoder *encoder = &encoder_slave->base; struct tda998x_priv *priv; priv = kzalloc(sizeof(*priv), GFP_KERNEL); @@ -1153,6 +1154,7 @@ tda998x_encoder_init(struct i2c_client *client, priv->vip_cntrl_2 = VIP_CNTRL_2_SWAP_E(4) | VIP_CNTRL_2_SWAP_F(5); priv->current_page = 0xff; + priv->hdmi = client; priv->cec = i2c_new_dummy(client->adapter, 0x34); if (!priv->cec) { kfree(priv); @@ -1164,14 +1166,14 @@ tda998x_encoder_init(struct i2c_client *client, encoder_slave->slave_funcs = &tda998x_encoder_funcs; /* wake up the device: */ - cec_write(encoder, REG_CEC_ENAMODS, + cec_write(priv, REG_CEC_ENAMODS, CEC_ENAMODS_EN_RXSENS | CEC_ENAMODS_EN_HDMI); - tda998x_reset(encoder); + tda998x_reset(priv); /* read version: */ - priv->rev = reg_read(encoder, REG_VERSION_LSB) | - reg_read(encoder, REG_VERSION_MSB) << 8; + priv->rev = reg_read(priv, REG_VERSION_LSB) | + reg_read(priv, REG_VERSION_MSB) << 8; /* mask off feature bits: */ priv->rev &= ~0x30; /* not-hdcp and not-scalar bit */ @@ -1187,16 +1189,16 @@ tda998x_encoder_init(struct i2c_client *client, } /* after reset, enable DDC: */ - reg_write(encoder, REG_DDC_DISABLE, 0x00); + reg_write(priv, REG_DDC_DISABLE, 0x00); /* set clock on DDC channel: */ - reg_write(encoder, REG_TX3, 39); + reg_write(priv, REG_TX3, 39); /* if necessary, disable multi-master: */ if (priv->rev == TDA19989) - reg_set(encoder, REG_I2C_MASTER, I2C_MASTER_DIS_MM); + reg_set(priv, REG_I2C_MASTER, I2C_MASTER_DIS_MM); - cec_write(encoder, REG_CEC_FRO_IM_CLK_CTRL, + cec_write(priv, REG_CEC_FRO_IM_CLK_CTRL, CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL); return 0; From 7d2eadc9b9d4eacc6aa8cc0cb33e05b5a6d30256 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:45 +0100 Subject: [PATCH 005/259] drm/i2c: tda998x: check more I/O errors This patch adds more error checking inn I2C I/O functions. In case of I/O error, this permits to avoid writing in bad controller pages, a bad chipset detection or looping when getting the EDID. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 57 +++++++++++++++++++++++-------- 1 file changed, 43 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 7df73baf3f9b..9bd336cdb734 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -363,7 +363,7 @@ cec_read(struct tda998x_priv *priv, uint8_t addr) return 0; } -static void +static int set_page(struct tda998x_priv *priv, uint16_t reg) { if (REG2PAGE(reg) != priv->current_page) { @@ -372,11 +372,14 @@ set_page(struct tda998x_priv *priv, uint16_t reg) REG_CURPAGE, REG2PAGE(reg) }; int ret = i2c_master_send(client, buf, sizeof(buf)); - if (ret < 0) + if (ret < 0) { dev_err(&client->dev, "Error %d writing to REG_CURPAGE\n", ret); + return ret; + } priv->current_page = REG2PAGE(reg); } + return 0; } static int @@ -386,7 +389,9 @@ reg_read_range(struct tda998x_priv *priv, uint16_t reg, char *buf, int cnt) uint8_t addr = REG2ADDR(reg); int ret; - set_page(priv, reg); + ret = set_page(priv, reg); + if (ret < 0) + return ret; ret = i2c_master_send(client, &addr, sizeof(addr)); if (ret < 0) @@ -413,18 +418,24 @@ reg_write_range(struct tda998x_priv *priv, uint16_t reg, uint8_t *p, int cnt) buf[0] = REG2ADDR(reg); memcpy(&buf[1], p, cnt); - set_page(priv, reg); + ret = set_page(priv, reg); + if (ret < 0) + return; ret = i2c_master_send(client, buf, cnt + 1); if (ret < 0) dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg); } -static uint8_t +static int reg_read(struct tda998x_priv *priv, uint16_t reg) { uint8_t val = 0; - reg_read_range(priv, reg, &val, sizeof(val)); + int ret; + + ret = reg_read_range(priv, reg, &val, sizeof(val)); + if (ret < 0) + return ret; return val; } @@ -435,7 +446,9 @@ reg_write(struct tda998x_priv *priv, uint16_t reg, uint8_t val) uint8_t buf[] = {REG2ADDR(reg), val}; int ret; - set_page(priv, reg); + ret = set_page(priv, reg); + if (ret < 0) + return; ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); if (ret < 0) @@ -449,7 +462,9 @@ reg_write16(struct tda998x_priv *priv, uint16_t reg, uint16_t val) uint8_t buf[] = {REG2ADDR(reg), val >> 8, val}; int ret; - set_page(priv, reg); + ret = set_page(priv, reg); + if (ret < 0) + return; ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); if (ret < 0) @@ -459,13 +474,21 @@ reg_write16(struct tda998x_priv *priv, uint16_t reg, uint16_t val) static void reg_set(struct tda998x_priv *priv, uint16_t reg, uint8_t val) { - reg_write(priv, reg, reg_read(priv, reg) | val); + int old_val; + + old_val = reg_read(priv, reg); + if (old_val >= 0) + reg_write(priv, reg, old_val | val); } static void reg_clear(struct tda998x_priv *priv, uint16_t reg, uint8_t val) { - reg_write(priv, reg, reg_read(priv, reg) & ~val); + int old_val; + + old_val = reg_read(priv, reg); + if (old_val >= 0) + reg_write(priv, reg, old_val & ~val); } static void @@ -978,8 +1001,10 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) /* wait for block read to complete: */ for (i = 100; i > 0; i--) { - uint8_t val = reg_read(priv, REG_INT_FLAGS_2); - if (val & INT_FLAGS_2_EDID_BLK_RD) + ret = reg_read(priv, REG_INT_FLAGS_2); + if (ret < 0) + return ret; + if (ret & INT_FLAGS_2_EDID_BLK_RD) break; msleep(1); } @@ -1144,6 +1169,7 @@ tda998x_encoder_init(struct i2c_client *client, struct drm_encoder_slave *encoder_slave) { struct tda998x_priv *priv; + int ret; priv = kzalloc(sizeof(*priv), GFP_KERNEL); if (!priv) @@ -1172,8 +1198,11 @@ tda998x_encoder_init(struct i2c_client *client, tda998x_reset(priv); /* read version: */ - priv->rev = reg_read(priv, REG_VERSION_LSB) | - reg_read(priv, REG_VERSION_MSB) << 8; + ret = reg_read(priv, REG_VERSION_LSB) | + (reg_read(priv, REG_VERSION_MSB) << 8); + if (ret < 0) + goto fail; + priv->rev = ret; /* mask off feature bits: */ priv->rev &= ~0x30; /* not-hdcp and not-scalar bit */ From fb7544d7732f780df989fabf31c5852be953daad Mon Sep 17 00:00:00 2001 From: Russell King Date: Sun, 2 Feb 2014 16:18:24 +0000 Subject: [PATCH 006/259] drm/i2c: tda998x: clean up error chip version checking This is a nicer way, and results in proper return codes should the read of the MSB version register fail. Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 9bd336cdb734..19f418246e7b 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1169,7 +1169,7 @@ tda998x_encoder_init(struct i2c_client *client, struct drm_encoder_slave *encoder_slave) { struct tda998x_priv *priv; - int ret; + int rev_lo, rev_hi, ret; priv = kzalloc(sizeof(*priv), GFP_KERNEL); if (!priv) @@ -1198,11 +1198,14 @@ tda998x_encoder_init(struct i2c_client *client, tda998x_reset(priv); /* read version: */ - ret = reg_read(priv, REG_VERSION_LSB) | - (reg_read(priv, REG_VERSION_MSB) << 8); - if (ret < 0) + rev_lo = reg_read(priv, REG_VERSION_LSB); + rev_hi = reg_read(priv, REG_VERSION_MSB); + if (rev_lo < 0 || rev_hi < 0) { + ret = rev_lo < 0 ? rev_lo : rev_hi; goto fail; - priv->rev = ret; + } + + priv->rev = rev_lo | rev_hi << 8; /* mask off feature bits: */ priv->rev &= ~0x30; /* not-hdcp and not-scalar bit */ From 704d63f59900968b9b5ae92549c49db0686b87e3 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:46 +0100 Subject: [PATCH 007/259] drm/i2c: tda998x: code cleanup This patch: - replaces ARRAY_SIZE() by sizeof() when a number of bytes is needed, - adds a linefeed in an error message and - removes an useless variable setting. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 19f418246e7b..c5c9de536346 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -336,7 +336,7 @@ cec_write(struct tda998x_priv *priv, uint16_t addr, uint8_t val) uint8_t buf[] = {addr, val}; int ret; - ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); + ret = i2c_master_send(client, buf, sizeof(buf)); if (ret < 0) dev_err(&client->dev, "Error %d writing to cec:0x%x\n", ret, addr); } @@ -373,7 +373,8 @@ set_page(struct tda998x_priv *priv, uint16_t reg) }; int ret = i2c_master_send(client, buf, sizeof(buf)); if (ret < 0) { - dev_err(&client->dev, "Error %d writing to REG_CURPAGE\n", ret); + dev_err(&client->dev, "setpage %04x err %d\n", + reg, ret); return ret; } @@ -450,7 +451,7 @@ reg_write(struct tda998x_priv *priv, uint16_t reg, uint8_t val) if (ret < 0) return; - ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); + ret = i2c_master_send(client, buf, sizeof(buf)); if (ret < 0) dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg); } @@ -466,7 +467,7 @@ reg_write16(struct tda998x_priv *priv, uint16_t reg, uint16_t val) if (ret < 0) return; - ret = i2c_master_send(client, buf, ARRAY_SIZE(buf)); + ret = i2c_master_send(client, buf, sizeof(buf)); if (ret < 0) dev_err(&client->dev, "Error %d writing to 0x%x\n", ret, reg); } @@ -1014,7 +1015,7 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) ret = reg_read_range(priv, REG_EDID_DATA_0, buf, EDID_LENGTH); if (ret != EDID_LENGTH) { - dev_err(encoder->dev->dev, "failed to read edid block %d: %d", + dev_err(encoder->dev->dev, "failed to read edid block %d: %d\n", blk, ret); return ret; } @@ -1028,7 +1029,7 @@ static uint8_t * do_get_edid(struct drm_encoder *encoder) { struct tda998x_priv *priv = to_tda998x_priv(encoder); - int j = 0, valid_extensions = 0; + int j, valid_extensions = 0; uint8_t *block, *new; bool print_bad_edid = drm_debug & DRM_UT_KMS; From b728fab7026b9db5a9bb60c7638765cfa4ee50c1 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:46 +0100 Subject: [PATCH 008/259] drm/i2c: tda998x: change probe message origin On probe, a message giving the TDA chip version seems to come from the DRM driver: armada-drm armada-510-drm: found TDA19988 This patch changes the originator of the message to the TDA driver: tda998x 0-0070: found TDA19988 Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index c5c9de536346..c3eebd73a706 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1212,12 +1212,21 @@ tda998x_encoder_init(struct i2c_client *client, priv->rev &= ~0x30; /* not-hdcp and not-scalar bit */ switch (priv->rev) { - case TDA9989N2: dev_info(dev->dev, "found TDA9989 n2"); break; - case TDA19989: dev_info(dev->dev, "found TDA19989"); break; - case TDA19989N2: dev_info(dev->dev, "found TDA19989 n2"); break; - case TDA19988: dev_info(dev->dev, "found TDA19988"); break; + case TDA9989N2: + dev_info(&client->dev, "found TDA9989 n2"); + break; + case TDA19989: + dev_info(&client->dev, "found TDA19989"); + break; + case TDA19989N2: + dev_info(&client->dev, "found TDA19989 n2"); + break; + case TDA19988: + dev_info(&client->dev, "found TDA19988"); + break; default: - DBG("found unsupported device: %04x", priv->rev); + dev_err(&client->dev, "found unsupported device: %04x\n", + priv->rev); goto fail; } From 73d5e253ac641bf95f5836c064128be78f43cd0b Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:44 +0100 Subject: [PATCH 009/259] drm/i2c: tda998x: don't freeze the system at audio startup time This patch prevents the system to be freezed at audio startup time, replacing mdelay by msleep. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index c3eebd73a706..80b94b51e060 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -681,7 +681,7 @@ tda998x_configure_audio(struct tda998x_priv *priv, reg_write_range(priv, REG_CH_STAT_B(0), buf, 4); tda998x_audio_mute(priv, true); - mdelay(20); + msleep(20); tda998x_audio_mute(priv, false); /* Write the audio information packet */ From 81b53a166f5cdf4e5bec47fc8884c994de82dc6b Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:42 +0100 Subject: [PATCH 010/259] drm/i2c: tda998x: don't read write-only registers This patch takes care of the write-only registers of the tda998x. The registers SOFTRESET, TBG_CNTRL_0 and TBG_CNTRL_1 have all bits cleared after reset, so, they may be fully re-written. The register MAT_CONTRL is set to MAT_CONTRL_MAT_BP | MAT_CONTRL_MAT_SC(1) after reset, so, it may be fully set again to this value. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 46 ++++++++++++++++--------------- 1 file changed, 24 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 80b94b51e060..d31e1c170bb6 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -496,9 +496,9 @@ static void tda998x_reset(struct tda998x_priv *priv) { /* reset audio and i2c master: */ - reg_set(priv, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); + reg_write(priv, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); msleep(50); - reg_clear(priv, REG_SOFTRESET, SOFTRESET_AUDIO | SOFTRESET_I2C_MASTER); + reg_write(priv, REG_SOFTRESET, 0); msleep(50); /* reset transmitter: */ @@ -860,7 +860,7 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, reg_set(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_RST_FIFO); /* set HDMI HDCP mode off: */ - reg_set(priv, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); + reg_write(priv, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); reg_clear(priv, REG_TX33, TX33_HDMI); reg_write(priv, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(0)); @@ -887,38 +887,28 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, PLL_SERIAL_2_SRL_PR(rep)); /* set color matrix bypass flag: */ - reg_set(priv, REG_MAT_CONTRL, MAT_CONTRL_MAT_BP); + reg_write(priv, REG_MAT_CONTRL, MAT_CONTRL_MAT_BP | + MAT_CONTRL_MAT_SC(1)); /* set BIAS tmds value: */ reg_write(priv, REG_ANA_GENERAL, 0x09); - reg_clear(priv, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_MTHD); + reg_write(priv, REG_TBG_CNTRL_0, 0); /* * Sync on rising HSYNC/VSYNC */ - reg_write(priv, REG_VIP_CNTRL_3, 0); - reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_SYNC_HS); + reg = VIP_CNTRL_3_SYNC_HS; /* * TDA19988 requires high-active sync at input stage, * so invert low-active sync provided by master encoder here */ if (mode->flags & DRM_MODE_FLAG_NHSYNC) - reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_H_TGL); + reg |= VIP_CNTRL_3_H_TGL; if (mode->flags & DRM_MODE_FLAG_NVSYNC) - reg_set(priv, REG_VIP_CNTRL_3, VIP_CNTRL_3_V_TGL); - - /* - * Always generate sync polarity relative to input sync and - * revert input stage toggled sync at output stage - */ - reg = TBG_CNTRL_1_TGL_EN; - if (mode->flags & DRM_MODE_FLAG_NHSYNC) - reg |= TBG_CNTRL_1_H_TGL; - if (mode->flags & DRM_MODE_FLAG_NVSYNC) - reg |= TBG_CNTRL_1_V_TGL; - reg_write(priv, REG_TBG_CNTRL_1, reg); + reg |= VIP_CNTRL_3_V_TGL; + reg_write(priv, REG_VIP_CNTRL_3, reg); reg_write(priv, REG_VIDFORMAT, 0x00); reg_write16(priv, REG_REFPIX_MSB, ref_pix); @@ -947,13 +937,25 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, reg_write(priv, REG_ENABLE_SPACE, 0x00); } + /* + * Always generate sync polarity relative to input sync and + * revert input stage toggled sync at output stage + */ + reg = TBG_CNTRL_1_DWIN_DIS | TBG_CNTRL_1_TGL_EN; + if (mode->flags & DRM_MODE_FLAG_NHSYNC) + reg |= TBG_CNTRL_1_H_TGL; + if (mode->flags & DRM_MODE_FLAG_NVSYNC) + reg |= TBG_CNTRL_1_V_TGL; + reg_write(priv, REG_TBG_CNTRL_1, reg); + /* must be last register set: */ - reg_clear(priv, REG_TBG_CNTRL_0, TBG_CNTRL_0_SYNC_ONCE); + reg_write(priv, REG_TBG_CNTRL_0, 0); /* Only setup the info frames if the sink is HDMI */ if (priv->is_hdmi_sink) { /* We need to turn HDMI HDCP stuff on to get audio through */ - reg_clear(priv, REG_TBG_CNTRL_1, TBG_CNTRL_1_DWIN_DIS); + reg &= ~TBG_CNTRL_1_DWIN_DIS; + reg_write(priv, REG_TBG_CNTRL_1, reg); reg_write(priv, REG_ENC_CNTRL, ENC_CNTRL_CTL_CODE(1)); reg_set(priv, REG_TX33, TX33_HDMI); From 0d44ea190387e21a7e6f6d7c9dd44df2e85d007a Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:41 +0100 Subject: [PATCH 011/259] drm/i2c: tda998x: add DT support This patch adds DT support to the tda998x. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index d31e1c170bb6..1ea4d43781f2 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1172,6 +1172,8 @@ tda998x_encoder_init(struct i2c_client *client, struct drm_encoder_slave *encoder_slave) { struct tda998x_priv *priv; + struct device_node *np = client->dev.of_node; + u32 video; int rev_lo, rev_hi, ret; priv = kzalloc(sizeof(*priv), GFP_KERNEL); @@ -1245,6 +1247,17 @@ tda998x_encoder_init(struct i2c_client *client, cec_write(priv, REG_CEC_FRO_IM_CLK_CTRL, CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL); + if (!np) + return 0; /* non-DT */ + + /* get the optional video properties */ + ret = of_property_read_u32(np, "video-ports", &video); + if (ret == 0) { + priv->vip_cntrl_0 = video >> 16; + priv->vip_cntrl_1 = video >> 8; + priv->vip_cntrl_2 = video; + } + return 0; fail: @@ -1259,6 +1272,14 @@ tda998x_encoder_init(struct i2c_client *client, return -ENXIO; } +#ifdef CONFIG_OF +static const struct of_device_id tda998x_dt_ids[] = { + { .compatible = "nxp,tda998x", }, + { } +}; +MODULE_DEVICE_TABLE(of, tda998x_dt_ids); +#endif + static struct i2c_device_id tda998x_ids[] = { { "tda998x", 0 }, { } @@ -1271,6 +1292,7 @@ static struct drm_i2c_encoder_driver tda998x_driver = { .remove = tda998x_remove, .driver = { .name = "tda998x", + .of_match_table = of_match_ptr(tda998x_dt_ids), }, .id_table = tda998x_ids, }, From 7c82e064ea7b64b986104260c300b370feef403a Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:37 +0100 Subject: [PATCH 012/259] drm/i2c: tda998x: add DT documentation Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- .../devicetree/bindings/drm/i2c/tda998x.txt | 27 +++++++++++++++++++ 1 file changed, 27 insertions(+) create mode 100644 Documentation/devicetree/bindings/drm/i2c/tda998x.txt diff --git a/Documentation/devicetree/bindings/drm/i2c/tda998x.txt b/Documentation/devicetree/bindings/drm/i2c/tda998x.txt new file mode 100644 index 000000000000..d7df01c5bb3a --- /dev/null +++ b/Documentation/devicetree/bindings/drm/i2c/tda998x.txt @@ -0,0 +1,27 @@ +Device-Tree bindings for the NXP TDA998x HDMI transmitter + +Required properties; + - compatible: must be "nxp,tda998x" + +Optional properties: + - interrupts: interrupt number and trigger type + default: polling + + - pinctrl-0: pin control group to be used for + screen plug/unplug interrupt. + + - pinctrl-names: must contain a "default" entry. + + - video-ports: 24 bits value which defines how the video controller + output is wired to the TDA998x input - default: <0x230145> + +Example: + + tda998x: hdmi-encoder { + compatible = "nxp,tda998x"; + reg = <0x70>; + interrupt-parent = <&gpio0>; + interrupts = <27 2>; /* falling edge */ + pinctrl-0 = <&pmx_camera>; + pinctrl-names = "default"; + }; From e47826274e8871bc6b35f82d35aea53db0f4ae31 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:38 +0100 Subject: [PATCH 013/259] drm/i2c: tda998x: always enable EDID read IRQ There is no need to enable/disable EDID read IRQ at each EDID block read. This patch enables the IRQ at init time. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 1ea4d43781f2..fddac4c84c28 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -985,9 +985,6 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) uint8_t offset, segptr; int ret, i; - /* enable EDID read irq: */ - reg_set(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); - offset = (blk & 1) ? 128 : 0; segptr = blk / 2; @@ -1022,8 +1019,6 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) return ret; } - reg_clear(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); - return 0; } @@ -1247,6 +1242,9 @@ tda998x_encoder_init(struct i2c_client *client, cec_write(priv, REG_CEC_FRO_IM_CLK_CTRL, CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL); + /* enable EDID read irq: */ + reg_set(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); + if (!np) return 0; /* non-DT */ From 12473b7d8e6074c7d4c2816afa6027354ce9a502 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:38 +0100 Subject: [PATCH 014/259] drm/i2c: tda998x: use irq for connection status and EDID read This patch adds the optional treatment of the tda998x IRQ. The interrupt function is used to know the display connection status without polling and to speedup reading the EDID. The IRQ number and trigger type are defined in the i2c client either by platform data or in the DT. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 114 +++++++++++++++++++++++++++--- 1 file changed, 105 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index fddac4c84c28..c9882246d81f 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -19,6 +19,7 @@ #include #include +#include #include #include @@ -40,6 +41,10 @@ struct tda998x_priv { u8 vip_cntrl_1; u8 vip_cntrl_2; struct tda998x_encoder_params params; + + wait_queue_head_t wq_edid; + volatile int wq_edid_wait; + struct drm_encoder *encoder; }; #define to_tda998x_priv(x) ((struct tda998x_priv *)to_encoder_slave(x)->slave_priv) @@ -306,11 +311,16 @@ struct tda998x_priv { /* CEC registers: (not paged) */ +#define REG_CEC_INTSTATUS 0xee /* read */ +# define CEC_INTSTATUS_CEC (1 << 0) +# define CEC_INTSTATUS_HDMI (1 << 1) #define REG_CEC_FRO_IM_CLK_CTRL 0xfb /* read/write */ # define CEC_FRO_IM_CLK_CTRL_GHOST_DIS (1 << 7) # define CEC_FRO_IM_CLK_CTRL_ENA_OTP (1 << 6) # define CEC_FRO_IM_CLK_CTRL_IMCLK_SEL (1 << 1) # define CEC_FRO_IM_CLK_CTRL_FRO_DIV (1 << 0) +#define REG_CEC_RXSHPDINTENA 0xfc /* read/write */ +#define REG_CEC_RXSHPDINT 0xfd /* read */ #define REG_CEC_RXSHPDLEV 0xfe /* read */ # define CEC_RXSHPDLEV_RXSENS (1 << 0) # define CEC_RXSHPDLEV_HPD (1 << 1) @@ -524,6 +534,35 @@ tda998x_reset(struct tda998x_priv *priv) reg_write(priv, REG_MUX_VP_VIP_OUT, 0x24); } +/* + * only 2 interrupts may occur: screen plug/unplug and EDID read + */ +static irqreturn_t tda998x_irq_thread(int irq, void *data) +{ + struct tda998x_priv *priv = data; + u8 sta, cec, lvl, flag0, flag1, flag2; + + if (!priv) + return IRQ_HANDLED; + sta = cec_read(priv, REG_CEC_INTSTATUS); + cec = cec_read(priv, REG_CEC_RXSHPDINT); + lvl = cec_read(priv, REG_CEC_RXSHPDLEV); + flag0 = reg_read(priv, REG_INT_FLAGS_0); + flag1 = reg_read(priv, REG_INT_FLAGS_1); + flag2 = reg_read(priv, REG_INT_FLAGS_2); + DRM_DEBUG_DRIVER( + "tda irq sta %02x cec %02x lvl %02x f0 %02x f1 %02x f2 %02x\n", + sta, cec, lvl, flag0, flag1, flag2); + if ((flag2 & INT_FLAGS_2_EDID_BLK_RD) && priv->wq_edid_wait) { + priv->wq_edid_wait = 0; + wake_up(&priv->wq_edid); + } else if (cec != 0) { /* HPD change */ + if (priv->encoder && priv->encoder->dev) + drm_helper_hpd_irq_event(priv->encoder->dev); + } + return IRQ_HANDLED; +} + static uint8_t tda998x_cksum(uint8_t *buf, size_t bytes) { uint8_t sum = 0; @@ -994,23 +1033,36 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) reg_write(priv, REG_DDC_SEGM, segptr); /* enable reading EDID: */ + priv->wq_edid_wait = 1; reg_write(priv, REG_EDID_CTRL, 0x1); /* flag must be cleared by sw: */ reg_write(priv, REG_EDID_CTRL, 0x0); /* wait for block read to complete: */ - for (i = 100; i > 0; i--) { - ret = reg_read(priv, REG_INT_FLAGS_2); - if (ret < 0) - return ret; - if (ret & INT_FLAGS_2_EDID_BLK_RD) - break; - msleep(1); + if (priv->hdmi->irq) { + i = wait_event_timeout(priv->wq_edid, + !priv->wq_edid_wait, + msecs_to_jiffies(100)); + if (i < 0) { + dev_err(encoder->dev->dev, "read edid wait err %d\n", i); + return i; + } + } else { + for (i = 10; i > 0; i--) { + msleep(10); + ret = reg_read(priv, REG_INT_FLAGS_2); + if (ret < 0) + return ret; + if (ret & INT_FLAGS_2_EDID_BLK_RD) + break; + } } - if (i == 0) + if (i == 0) { + dev_err(encoder->dev->dev, "read edid timeout\n"); return -ETIMEDOUT; + } ret = reg_read_range(priv, REG_EDID_DATA_0, buf, EDID_LENGTH); if (ret != EDID_LENGTH) { @@ -1108,7 +1160,13 @@ static int tda998x_encoder_create_resources(struct drm_encoder *encoder, struct drm_connector *connector) { - DBG(""); + struct tda998x_priv *priv = to_tda998x_priv(encoder); + + if (priv->hdmi->irq) + connector->polled = DRM_CONNECTOR_POLL_HPD; + else + connector->polled = DRM_CONNECTOR_POLL_CONNECT | + DRM_CONNECTOR_POLL_DISCONNECT; return 0; } @@ -1127,6 +1185,13 @@ tda998x_encoder_destroy(struct drm_encoder *encoder) { struct tda998x_priv *priv = to_tda998x_priv(encoder); drm_i2c_encoder_destroy(encoder); + + /* disable all IRQs and free the IRQ handler */ + cec_write(priv, REG_CEC_RXSHPDINTENA, 0); + reg_clear(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); + if (priv->hdmi->irq) + free_irq(priv->hdmi->irq, priv); + if (priv->cec) i2c_unregister_device(priv->cec); kfree(priv); @@ -1186,6 +1251,8 @@ tda998x_encoder_init(struct i2c_client *client, kfree(priv); return -ENODEV; } + + priv->encoder = &encoder_slave->base; priv->dpms = DRM_MODE_DPMS_OFF; encoder_slave->slave_priv = priv; @@ -1242,6 +1309,35 @@ tda998x_encoder_init(struct i2c_client *client, cec_write(priv, REG_CEC_FRO_IM_CLK_CTRL, CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL); + /* initialize the optional IRQ */ + if (client->irq) { + int irqf_trigger; + + /* init read EDID waitqueue */ + init_waitqueue_head(&priv->wq_edid); + + /* clear pending interrupts */ + reg_read(priv, REG_INT_FLAGS_0); + reg_read(priv, REG_INT_FLAGS_1); + reg_read(priv, REG_INT_FLAGS_2); + + irqf_trigger = + irqd_get_trigger_type(irq_get_irq_data(client->irq)); + ret = request_threaded_irq(client->irq, NULL, + tda998x_irq_thread, + irqf_trigger | IRQF_ONESHOT, + "tda998x", priv); + if (ret) { + dev_err(&client->dev, + "failed to request IRQ#%u: %d\n", + client->irq, ret); + goto fail; + } + + /* enable HPD irq */ + cec_write(priv, REG_CEC_RXSHPDINTENA, CEC_RXSHPDLEV_HPD); + } + /* enable EDID read irq: */ reg_set(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); From 10df1a95d6457ead03ed804c9abece79023f3f77 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:40 +0100 Subject: [PATCH 015/259] drm/i2c: tda998x: make the audio code more readable This patch adds a definition of the values of the MUX_AP register and simplifies the macro's defining the fields of the AIP_CLKSEL register. This makes the format specific audio init sequence more readable. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 25 +++++++++++++------------ 1 file changed, 13 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index c9882246d81f..99d210bbf4e2 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -127,6 +127,8 @@ struct tda998x_priv { # define VIP_CNTRL_5_CKCASE (1 << 0) # define VIP_CNTRL_5_SP_CNT(x) (((x) & 3) << 1) #define REG_MUX_AP REG(0x00, 0x26) /* read/write */ +# define MUX_AP_SELECT_I2S 0x64 +# define MUX_AP_SELECT_SPDIF 0x40 #define REG_MUX_VP_VIP_OUT REG(0x00, 0x27) /* read/write */ #define REG_MAT_CONTRL REG(0x00, 0x80) /* write */ # define MAT_CONTRL_MAT_SC(x) (((x) & 3) << 0) @@ -204,10 +206,11 @@ struct tda998x_priv { #define REG_I2S_FORMAT REG(0x00, 0xfc) /* read/write */ # define I2S_FORMAT(x) (((x) & 3) << 0) #define REG_AIP_CLKSEL REG(0x00, 0xfd) /* write */ -# define AIP_CLKSEL_FS(x) (((x) & 3) << 0) -# define AIP_CLKSEL_CLK_POL(x) (((x) & 1) << 2) -# define AIP_CLKSEL_AIP(x) (((x) & 7) << 3) - +# define AIP_CLKSEL_AIP_SPDIF (0 << 3) +# define AIP_CLKSEL_AIP_I2S (1 << 3) +# define AIP_CLKSEL_FS_ACLK (0 << 0) +# define AIP_CLKSEL_FS_MCLK (1 << 0) +# define AIP_CLKSEL_FS_FS64SPDIF (2 << 0) /* Page 02h: PLL settings */ #define REG_PLL_SERIAL_1 REG(0x02, 0x00) /* read/write */ @@ -647,19 +650,17 @@ tda998x_configure_audio(struct tda998x_priv *priv, /* Set audio input source */ switch (p->audio_format) { case AFMT_SPDIF: - reg_write(priv, REG_MUX_AP, 0x40); - clksel_aip = AIP_CLKSEL_AIP(0); - /* FS64SPDIF */ - clksel_fs = AIP_CLKSEL_FS(2); + reg_write(priv, REG_MUX_AP, MUX_AP_SELECT_SPDIF); + clksel_aip = AIP_CLKSEL_AIP_SPDIF; + clksel_fs = AIP_CLKSEL_FS_FS64SPDIF; cts_n = CTS_N_M(3) | CTS_N_K(3); ca_i2s = 0; break; case AFMT_I2S: - reg_write(priv, REG_MUX_AP, 0x64); - clksel_aip = AIP_CLKSEL_AIP(1); - /* ACLK */ - clksel_fs = AIP_CLKSEL_FS(0); + reg_write(priv, REG_MUX_AP, MUX_AP_SELECT_I2S); + clksel_aip = AIP_CLKSEL_AIP_I2S; + clksel_fs = AIP_CLKSEL_FS_ACLK; cts_n = CTS_N_M(3) | CTS_N_K(3); ca_i2s = CA_I2S_CA_I2S(0); break; From 85c988bb26a3da46c04284bc43f93d732986547b Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:40 +0100 Subject: [PATCH 016/259] drm/i2c: tda998x: remove the unused variable ca_i2s ca_i2s is only ever written to, but never read, so let's get rid of it. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 99d210bbf4e2..4352294d75cd 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -640,7 +640,7 @@ static void tda998x_configure_audio(struct tda998x_priv *priv, struct drm_display_mode *mode, struct tda998x_encoder_params *p) { - uint8_t buf[6], clksel_aip, clksel_fs, ca_i2s, cts_n, adiv; + uint8_t buf[6], clksel_aip, clksel_fs, cts_n, adiv; uint32_t n; /* Enable audio ports */ @@ -654,7 +654,6 @@ tda998x_configure_audio(struct tda998x_priv *priv, clksel_aip = AIP_CLKSEL_AIP_SPDIF; clksel_fs = AIP_CLKSEL_FS_FS64SPDIF; cts_n = CTS_N_M(3) | CTS_N_K(3); - ca_i2s = 0; break; case AFMT_I2S: @@ -662,7 +661,6 @@ tda998x_configure_audio(struct tda998x_priv *priv, clksel_aip = AIP_CLKSEL_AIP_I2S; clksel_fs = AIP_CLKSEL_FS_ACLK; cts_n = CTS_N_M(3) | CTS_N_K(3); - ca_i2s = CA_I2S_CA_I2S(0); break; default: From a8b517e5312124e2dd7b6d6d9afac458aaecfbf3 Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:39 +0100 Subject: [PATCH 017/259] drm/i2c: tda998x: code optimization This patch reduces the number of I2C exchanges by setting many bits in one write and removing a useless write. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 4352294d75cd..666ee24329b9 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -669,10 +669,8 @@ tda998x_configure_audio(struct tda998x_priv *priv, } reg_write(priv, REG_AIP_CLKSEL, clksel_aip); - reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_LAYOUT); - - /* Enable automatic CTS generation */ - reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_ACR_MAN); + reg_clear(priv, REG_AIP_CNTRL_0, AIP_CNTRL_0_LAYOUT | + AIP_CNTRL_0_ACR_MAN); /* auto CTS */ reg_write(priv, REG_CTS_N, cts_n); /* @@ -908,10 +906,10 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, reg_write(priv, REG_VIP_CNTRL_5, VIP_CNTRL_5_SP_CNT(0)); reg_write(priv, REG_VIP_CNTRL_4, VIP_CNTRL_4_BLANKIT(0) | VIP_CNTRL_4_BLC(0)); - reg_clear(priv, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_CCIR); reg_clear(priv, REG_PLL_SERIAL_1, PLL_SERIAL_1_SRL_MAN_IZ); - reg_clear(priv, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_DE); + reg_clear(priv, REG_PLL_SERIAL_3, PLL_SERIAL_3_SRL_CCIR | + PLL_SERIAL_3_SRL_DE); reg_write(priv, REG_SERIALIZER, 0); reg_write(priv, REG_HVF_CNTRL_1, HVF_CNTRL_1_VQR(0)); @@ -931,8 +929,6 @@ tda998x_encoder_mode_set(struct drm_encoder *encoder, /* set BIAS tmds value: */ reg_write(priv, REG_ANA_GENERAL, 0x09); - reg_write(priv, REG_TBG_CNTRL_0, 0); - /* * Sync on rising HSYNC/VSYNC */ From 2470feccbf030652380c2d73304576137b0fb12e Mon Sep 17 00:00:00 2001 From: Jean-Francois Moine Date: Sat, 25 Jan 2014 18:14:36 +0100 Subject: [PATCH 018/259] drm/i2c: tda998x: adjust the audio clock divider for S/PDIF According to some tests on the Cubox (Marvell Armada 510 + TDA19988), the S/PDIF input asks for a greater audio clock divider. Tested-by: Russell King Acked-by: Russell King Signed-off-by: Jean-Francois Moine Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 666ee24329b9..d0f3a4cbb866 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -680,10 +680,14 @@ tda998x_configure_audio(struct tda998x_priv *priv, * There is no detailed info in the datasheet, so we just * assume 100MHz requires larger divider. */ + adiv = AUDIO_DIV_SERCLK_8; if (mode->clock > 100000) - adiv = AUDIO_DIV_SERCLK_16; - else - adiv = AUDIO_DIV_SERCLK_8; + adiv++; /* AUDIO_DIV_SERCLK_16 */ + + /* S/PDIF asks for a larger divider */ + if (p->audio_format == AFMT_SPDIF) + adiv++; /* AUDIO_DIV_SERCLK_16 or _32 */ + reg_write(priv, REG_AUDIO_DIV, adiv); /* From 5e7fe2fef4347d7a09bb15588d8bbe3cb83b6ed4 Mon Sep 17 00:00:00 2001 From: Russell King Date: Fri, 7 Feb 2014 19:13:23 +0000 Subject: [PATCH 019/259] drm/i2c: tda998x: always use the same device for all kernel messages Rather than using a mixture of the parent DRM device and the component device for messages from the driver, consistently use the component device for all messages. Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index d0f3a4cbb866..48af5cac1902 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1044,7 +1044,7 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) !priv->wq_edid_wait, msecs_to_jiffies(100)); if (i < 0) { - dev_err(encoder->dev->dev, "read edid wait err %d\n", i); + dev_err(&priv->hdmi->dev, "read edid wait err %d\n", i); return i; } } else { @@ -1059,14 +1059,14 @@ read_edid_block(struct drm_encoder *encoder, uint8_t *buf, int blk) } if (i == 0) { - dev_err(encoder->dev->dev, "read edid timeout\n"); + dev_err(&priv->hdmi->dev, "read edid timeout\n"); return -ETIMEDOUT; } ret = reg_read_range(priv, REG_EDID_DATA_0, buf, EDID_LENGTH); if (ret != EDID_LENGTH) { - dev_err(encoder->dev->dev, "failed to read edid block %d: %d\n", - blk, ret); + dev_err(&priv->hdmi->dev, "failed to read edid block %d: %d\n", + blk, ret); return ret; } @@ -1132,7 +1132,7 @@ do_get_edid(struct drm_encoder *encoder) fail: if (priv->rev == TDA19988) reg_set(priv, REG_TX4, TX4_PD_RAM); - dev_warn(encoder->dev->dev, "failed to read EDID\n"); + dev_warn(&priv->hdmi->dev, "failed to read EDID\n"); kfree(block); return NULL; } From 1c61eae469e0d1d2fb9d7b77f51ca50c1f8f3ce9 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 01:50:22 -0700 Subject: [PATCH 020/259] drm/radeon: fix CP semaphores on CIK MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The CP semaphore queue on CIK has a bug that triggers if uncompleted waits use the same address while a signal is still pending. Work around this by using different addresses for each sync. Signed-off-by: Christian König Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/radeon.h | 4 +++- drivers/gpu/drm/radeon/radeon_ring.c | 2 +- drivers/gpu/drm/radeon/radeon_semaphore.c | 19 ++++++++++++++++--- 3 files changed, 20 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 4a8ac1cd6b4c..024db37b1832 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -135,6 +135,9 @@ extern int radeon_hard_reset; /* R600+ */ #define R600_RING_TYPE_UVD_INDEX 5 +/* number of hw syncs before falling back on blocking */ +#define RADEON_NUM_SYNCS 4 + /* hardcode those limit for now */ #define RADEON_VA_IB_OFFSET (1 << 20) #define RADEON_VA_RESERVED_SIZE (8 << 20) @@ -554,7 +557,6 @@ int radeon_mode_dumb_mmap(struct drm_file *filp, /* * Semaphores. */ -/* everything here is constant */ struct radeon_semaphore { struct radeon_sa_bo *sa_bo; signed waiters; diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 1b783f0e6d3a..15e44a7281ab 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -139,7 +139,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, } /* 64 dwords should be enough for fence too */ - r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_RINGS * 8); + r = radeon_ring_lock(rdev, ring, 64 + RADEON_NUM_SYNCS * 8); if (r) { dev_err(rdev->dev, "scheduling IB failed (%d).\n", r); return r; diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c index 2b42aa1914f2..9006b32d5eed 100644 --- a/drivers/gpu/drm/radeon/radeon_semaphore.c +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -34,14 +34,15 @@ int radeon_semaphore_create(struct radeon_device *rdev, struct radeon_semaphore **semaphore) { + uint32_t *cpu_addr; int i, r; *semaphore = kmalloc(sizeof(struct radeon_semaphore), GFP_KERNEL); if (*semaphore == NULL) { return -ENOMEM; } - r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, - &(*semaphore)->sa_bo, 8, 8, true); + r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &(*semaphore)->sa_bo, + 8 * RADEON_NUM_SYNCS, 8, true); if (r) { kfree(*semaphore); *semaphore = NULL; @@ -49,7 +50,10 @@ int radeon_semaphore_create(struct radeon_device *rdev, } (*semaphore)->waiters = 0; (*semaphore)->gpu_addr = radeon_sa_bo_gpu_addr((*semaphore)->sa_bo); - *((uint64_t*)radeon_sa_bo_cpu_addr((*semaphore)->sa_bo)) = 0; + + cpu_addr = radeon_sa_bo_cpu_addr((*semaphore)->sa_bo); + for (i = 0; i < RADEON_NUM_SYNCS; ++i) + cpu_addr[i] = 0; for (i = 0; i < RADEON_NUM_RINGS; ++i) (*semaphore)->sync_to[i] = NULL; @@ -125,6 +129,7 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, struct radeon_semaphore *semaphore, int ring) { + unsigned count = 0; int i, r; for (i = 0; i < RADEON_NUM_RINGS; ++i) { @@ -140,6 +145,12 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, return -EINVAL; } + if (++count > RADEON_NUM_SYNCS) { + /* not enough room, wait manually */ + radeon_fence_wait_locked(fence); + continue; + } + /* allocate enough space for sync command */ r = radeon_ring_alloc(rdev, &rdev->ring[i], 16); if (r) { @@ -164,6 +175,8 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, radeon_ring_commit(rdev, &rdev->ring[i]); radeon_fence_note_sync(fence, ring); + + semaphore->gpu_addr += 8; } return 0; From d93f79376f210e0b19da57a3dc841ba332daa9d0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 23 May 2013 12:10:04 +0200 Subject: [PATCH 021/259] drm/radeon: initial VCE support v4 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Only VCE 2.0 support so far. v2: squashing multiple patches into this one v3: add IRQ support for CIK, major cleanups, basic code documentation v4: remove HAINAN from chipset list Signed-off-by: Christian König --- drivers/gpu/drm/radeon/Makefile | 6 + drivers/gpu/drm/radeon/cik.c | 60 +++ drivers/gpu/drm/radeon/cikd.h | 33 ++ drivers/gpu/drm/radeon/radeon.h | 56 ++- drivers/gpu/drm/radeon/radeon_asic.c | 17 + drivers/gpu/drm/radeon/radeon_asic.h | 13 + drivers/gpu/drm/radeon/radeon_cs.c | 4 + drivers/gpu/drm/radeon/radeon_kms.c | 1 + drivers/gpu/drm/radeon/radeon_ring.c | 4 + drivers/gpu/drm/radeon/radeon_test.c | 39 +- drivers/gpu/drm/radeon/radeon_vce.c | 588 +++++++++++++++++++++++++++ drivers/gpu/drm/radeon/sid.h | 47 +++ drivers/gpu/drm/radeon/vce_v1_0.c | 187 +++++++++ drivers/gpu/drm/radeon/vce_v2_0.c | 70 ++++ include/uapi/drm/radeon_drm.h | 1 + 15 files changed, 1117 insertions(+), 9 deletions(-) create mode 100644 drivers/gpu/drm/radeon/radeon_vce.c create mode 100644 drivers/gpu/drm/radeon/vce_v1_0.c create mode 100644 drivers/gpu/drm/radeon/vce_v2_0.c diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index 306364a1ecda..ed60caa32518 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -99,6 +99,12 @@ radeon-y += \ uvd_v3_1.o \ uvd_v4_2.o +# add VCE block +radeon-y += \ + radeon_vce.o \ + vce_v1_0.o \ + vce_v2_0.o \ + radeon-$(CONFIG_COMPAT) += radeon_ioc32.o radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o radeon-$(CONFIG_ACPI) += radeon_acpi.o diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index e6419ca7cd37..be6eb4d91284 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -7490,6 +7490,20 @@ int cik_irq_process(struct radeon_device *rdev) /* reset addr and status */ WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1); break; + case 167: /* VCE */ + DRM_DEBUG("IH: VCE int: 0x%08x\n", src_data); + switch (src_data) { + case 0: + radeon_fence_process(rdev, TN_RING_TYPE_VCE1_INDEX); + break; + case 1: + radeon_fence_process(rdev, TN_RING_TYPE_VCE2_INDEX); + break; + default: + DRM_ERROR("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; case 176: /* GFX RB CP_INT */ case 177: /* GFX IB CP_INT */ radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); @@ -7789,6 +7803,22 @@ static int cik_startup(struct radeon_device *rdev) if (r) rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0; + r = radeon_vce_resume(rdev); + if (!r) { + r = vce_v2_0_resume(rdev); + if (!r) + r = radeon_fence_driver_start_ring(rdev, + TN_RING_TYPE_VCE1_INDEX); + if (!r) + r = radeon_fence_driver_start_ring(rdev, + TN_RING_TYPE_VCE2_INDEX); + } + if (r) { + dev_err(rdev->dev, "VCE init error (%d).\n", r); + rdev->ring[TN_RING_TYPE_VCE1_INDEX].ring_size = 0; + rdev->ring[TN_RING_TYPE_VCE2_INDEX].ring_size = 0; + } + /* Enable IRQ */ if (!rdev->irq.installed) { r = radeon_irq_kms_init(rdev); @@ -7864,6 +7894,23 @@ static int cik_startup(struct radeon_device *rdev) DRM_ERROR("radeon: failed initializing UVD (%d).\n", r); } + r = -ENOENT; + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + if (ring->ring_size) + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + VCE_CMD_NO_OP); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + if (ring->ring_size) + r = radeon_ring_init(rdev, ring, ring->ring_size, 0, + VCE_CMD_NO_OP); + + if (!r) + r = vce_v1_0_init(rdev); + else if (r != -ENOENT) + DRM_ERROR("radeon: failed initializing VCE (%d).\n", r); + r = radeon_ib_pool_init(rdev); if (r) { dev_err(rdev->dev, "IB initialization failed (%d).\n", r); @@ -7934,6 +7981,7 @@ int cik_suspend(struct radeon_device *rdev) cik_sdma_enable(rdev, false); uvd_v1_0_fini(rdev); radeon_uvd_suspend(rdev); + radeon_vce_suspend(rdev); cik_fini_pg(rdev); cik_fini_cg(rdev); cik_irq_suspend(rdev); @@ -8066,6 +8114,17 @@ int cik_init(struct radeon_device *rdev) r600_ring_init(rdev, ring, 4096); } + r = radeon_vce_init(rdev); + if (!r) { + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 4096); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 4096); + } + rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -8127,6 +8186,7 @@ void cik_fini(struct radeon_device *rdev) radeon_irq_kms_fini(rdev); uvd_v1_0_fini(rdev); radeon_uvd_fini(rdev); + radeon_vce_fini(rdev); cik_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h index 98bae9d7b74d..459ae021d91c 100644 --- a/drivers/gpu/drm/radeon/cikd.h +++ b/drivers/gpu/drm/radeon/cikd.h @@ -2010,4 +2010,37 @@ /* UVD CTX indirect */ #define UVD_CGC_MEM_CTRL 0xC0 +/* VCE */ + +#define VCE_VCPU_CACHE_OFFSET0 0x20024 +#define VCE_VCPU_CACHE_SIZE0 0x20028 +#define VCE_VCPU_CACHE_OFFSET1 0x2002c +#define VCE_VCPU_CACHE_SIZE1 0x20030 +#define VCE_VCPU_CACHE_OFFSET2 0x20034 +#define VCE_VCPU_CACHE_SIZE2 0x20038 +#define VCE_RB_RPTR2 0x20178 +#define VCE_RB_WPTR2 0x2017c +#define VCE_RB_RPTR 0x2018c +#define VCE_RB_WPTR 0x20190 +#define VCE_CLOCK_GATING_A 0x202f8 +#define VCE_CLOCK_GATING_B 0x202fc +#define VCE_UENC_CLOCK_GATING 0x207bc +#define VCE_UENC_REG_CLOCK_GATING 0x207c0 +#define VCE_SYS_INT_EN 0x21300 +# define VCE_SYS_INT_TRAP_INTERRUPT_EN (1 << 3) +#define VCE_LMI_CTRL2 0x21474 +#define VCE_LMI_CTRL 0x21498 +#define VCE_LMI_VM_CTRL 0x214a0 +#define VCE_LMI_SWAP_CNTL 0x214b4 +#define VCE_LMI_SWAP_CNTL1 0x214b8 +#define VCE_LMI_CACHE_CTRL 0x214f4 + +#define VCE_CMD_NO_OP 0x00000000 +#define VCE_CMD_END 0x00000001 +#define VCE_CMD_IB 0x00000002 +#define VCE_CMD_FENCE 0x00000003 +#define VCE_CMD_TRAP 0x00000004 +#define VCE_CMD_IB_AUTO 0x00000005 +#define VCE_CMD_SEMAPHORE 0x00000006 + #endif diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 024db37b1832..a58a38942c73 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -113,19 +113,16 @@ extern int radeon_hard_reset; #define RADEONFB_CONN_LIMIT 4 #define RADEON_BIOS_NUM_SCRATCH 8 -/* max number of rings */ -#define RADEON_NUM_RINGS 6 - /* fence seq are set to this number when signaled */ #define RADEON_FENCE_SIGNALED_SEQ 0LL /* internal ring indices */ /* r1xx+ has gfx CP ring */ -#define RADEON_RING_TYPE_GFX_INDEX 0 +#define RADEON_RING_TYPE_GFX_INDEX 0 /* cayman has 2 compute CP rings */ -#define CAYMAN_RING_TYPE_CP1_INDEX 1 -#define CAYMAN_RING_TYPE_CP2_INDEX 2 +#define CAYMAN_RING_TYPE_CP1_INDEX 1 +#define CAYMAN_RING_TYPE_CP2_INDEX 2 /* R600+ has an async dma ring */ #define R600_RING_TYPE_DMA_INDEX 3 @@ -133,7 +130,14 @@ extern int radeon_hard_reset; #define CAYMAN_RING_TYPE_DMA1_INDEX 4 /* R600+ */ -#define R600_RING_TYPE_UVD_INDEX 5 +#define R600_RING_TYPE_UVD_INDEX 5 + +/* TN+ */ +#define TN_RING_TYPE_VCE1_INDEX 6 +#define TN_RING_TYPE_VCE2_INDEX 7 + +/* max number of rings */ +#define RADEON_NUM_RINGS 8 /* number of hw syncs before falling back on blocking */ #define RADEON_NUM_SYNCS 4 @@ -1591,6 +1595,42 @@ int radeon_uvd_calc_upll_dividers(struct radeon_device *rdev, int radeon_uvd_send_upll_ctlreq(struct radeon_device *rdev, unsigned cg_upll_func_cntl); +/* + * VCE + */ +#define RADEON_MAX_VCE_HANDLES 16 +#define RADEON_VCE_STACK_SIZE (1024*1024) +#define RADEON_VCE_HEAP_SIZE (4*1024*1024) + +struct radeon_vce { + struct radeon_bo *vcpu_bo; + void *cpu_addr; + uint64_t gpu_addr; + atomic_t handles[RADEON_MAX_VCE_HANDLES]; + struct drm_file *filp[RADEON_MAX_VCE_HANDLES]; +}; + +int radeon_vce_init(struct radeon_device *rdev); +void radeon_vce_fini(struct radeon_device *rdev); +int radeon_vce_suspend(struct radeon_device *rdev); +int radeon_vce_resume(struct radeon_device *rdev); +int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence); +int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence); +void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp); +int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi); +int radeon_vce_cs_parse(struct radeon_cs_parser *p); +bool radeon_vce_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait); +void radeon_vce_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); +void radeon_vce_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence); +int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring); +int radeon_vce_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); + struct r600_audio_pin { int channels; int rate; @@ -2186,6 +2226,7 @@ struct radeon_device { struct radeon_gem gem; struct radeon_pm pm; struct radeon_uvd uvd; + struct radeon_vce vce; uint32_t bios_scratch[RADEON_BIOS_NUM_SCRATCH]; struct radeon_wb wb; struct radeon_dummy_page dummy_page; @@ -2205,6 +2246,7 @@ struct radeon_device { const struct firmware *sdma_fw; /* CIK SDMA firmware */ const struct firmware *smc_fw; /* SMC firmware */ const struct firmware *uvd_fw; /* UVD firmware */ + const struct firmware *vce_fw; /* VCE firmware */ struct r600_vram_scratch vram_scratch; int msi_enabled; /* msi enabled */ struct r600_ih ih; /* r6/700 interrupt ring */ diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index dda02bfc10a4..4f059b2c05fb 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -1987,6 +1987,19 @@ static struct radeon_asic_ring ci_dma_ring = { .set_wptr = &cik_sdma_set_wptr, }; +static struct radeon_asic_ring ci_vce_ring = { + .ib_execute = &radeon_vce_ib_execute, + .emit_fence = &radeon_vce_fence_emit, + .emit_semaphore = &radeon_vce_semaphore_emit, + .cs_parse = &radeon_vce_cs_parse, + .ring_test = &radeon_vce_ring_test, + .ib_test = &radeon_vce_ib_test, + .is_lockup = &radeon_ring_test_lockup, + .get_rptr = &vce_v1_0_get_rptr, + .get_wptr = &vce_v1_0_get_wptr, + .set_wptr = &vce_v1_0_set_wptr, +}; + static struct radeon_asic ci_asic = { .init = &cik_init, .fini = &cik_fini, @@ -2015,6 +2028,8 @@ static struct radeon_asic ci_asic = { [R600_RING_TYPE_DMA_INDEX] = &ci_dma_ring, [CAYMAN_RING_TYPE_DMA1_INDEX] = &ci_dma_ring, [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, + [TN_RING_TYPE_VCE1_INDEX] = &ci_vce_ring, + [TN_RING_TYPE_VCE2_INDEX] = &ci_vce_ring, }, .irq = { .set = &cik_irq_set, @@ -2117,6 +2132,8 @@ static struct radeon_asic kv_asic = { [R600_RING_TYPE_DMA_INDEX] = &ci_dma_ring, [CAYMAN_RING_TYPE_DMA1_INDEX] = &ci_dma_ring, [R600_RING_TYPE_UVD_INDEX] = &cayman_uvd_ring, + [TN_RING_TYPE_VCE1_INDEX] = &ci_vce_ring, + [TN_RING_TYPE_VCE2_INDEX] = &ci_vce_ring, }, .irq = { .set = &cik_irq_set, diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index ae637cfda783..13f87bf5254b 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -863,4 +863,17 @@ bool uvd_v3_1_semaphore_emit(struct radeon_device *rdev, /* uvd v4.2 */ int uvd_v4_2_resume(struct radeon_device *rdev); +/* vce v1.0 */ +uint32_t vce_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring); +uint32_t vce_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +void vce_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring); +int vce_v1_0_init(struct radeon_device *rdev); +int vce_v1_0_start(struct radeon_device *rdev); + +/* vce v2.0 */ +int vce_v2_0_resume(struct radeon_device *rdev); + #endif diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index dfb5a1db87d4..701ee7981b5c 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -147,6 +147,10 @@ static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority case RADEON_CS_RING_UVD: p->ring = R600_RING_TYPE_UVD_INDEX; break; + case RADEON_CS_RING_VCE: + /* TODO: only use the low priority ring for now */ + p->ring = TN_RING_TYPE_VCE1_INDEX; + break; } return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 114d1672d616..0e078afc5db7 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -610,6 +610,7 @@ void radeon_driver_preclose_kms(struct drm_device *dev, if (rdev->cmask_filp == file_priv) rdev->cmask_filp = NULL; radeon_uvd_free_handles(rdev, file_priv); + radeon_vce_free_handles(rdev, file_priv); } /* diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 15e44a7281ab..d2980b03d1ad 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -814,6 +814,8 @@ static int cayman_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX; static int radeon_dma1_index = R600_RING_TYPE_DMA_INDEX; static int radeon_dma2_index = CAYMAN_RING_TYPE_DMA1_INDEX; static int r600_uvd_index = R600_RING_TYPE_UVD_INDEX; +static int si_vce1_index = TN_RING_TYPE_VCE1_INDEX; +static int si_vce2_index = TN_RING_TYPE_VCE2_INDEX; static struct drm_info_list radeon_debugfs_ring_info_list[] = { {"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_gfx_index}, @@ -822,6 +824,8 @@ static struct drm_info_list radeon_debugfs_ring_info_list[] = { {"radeon_ring_dma1", radeon_debugfs_ring_info, 0, &radeon_dma1_index}, {"radeon_ring_dma2", radeon_debugfs_ring_info, 0, &radeon_dma2_index}, {"radeon_ring_uvd", radeon_debugfs_ring_info, 0, &r600_uvd_index}, + {"radeon_ring_vce1", radeon_debugfs_ring_info, 0, &si_vce1_index}, + {"radeon_ring_vce2", radeon_debugfs_ring_info, 0, &si_vce2_index}, }; static int radeon_debugfs_sa_info(struct seq_file *m, void *data) diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c index 12e8099a0823..3a13e0d1055c 100644 --- a/drivers/gpu/drm/radeon/radeon_test.c +++ b/drivers/gpu/drm/radeon/radeon_test.c @@ -257,20 +257,36 @@ static int radeon_test_create_and_emit_fence(struct radeon_device *rdev, struct radeon_ring *ring, struct radeon_fence **fence) { + uint32_t handle = ring->idx ^ 0xdeafbeef; int r; if (ring->idx == R600_RING_TYPE_UVD_INDEX) { - r = radeon_uvd_get_create_msg(rdev, ring->idx, 1, NULL); + r = radeon_uvd_get_create_msg(rdev, ring->idx, handle, NULL); if (r) { DRM_ERROR("Failed to get dummy create msg\n"); return r; } - r = radeon_uvd_get_destroy_msg(rdev, ring->idx, 1, fence); + r = radeon_uvd_get_destroy_msg(rdev, ring->idx, handle, fence); if (r) { DRM_ERROR("Failed to get dummy destroy msg\n"); return r; } + + } else if (ring->idx == TN_RING_TYPE_VCE1_INDEX || + ring->idx == TN_RING_TYPE_VCE2_INDEX) { + r = radeon_vce_get_create_msg(rdev, ring->idx, handle, NULL); + if (r) { + DRM_ERROR("Failed to get dummy create msg\n"); + return r; + } + + r = radeon_vce_get_destroy_msg(rdev, ring->idx, handle, fence); + if (r) { + DRM_ERROR("Failed to get dummy destroy msg\n"); + return r; + } + } else { r = radeon_ring_lock(rdev, ring, 64); if (r) { @@ -486,6 +502,16 @@ static void radeon_test_ring_sync2(struct radeon_device *rdev, printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); } +static bool radeon_test_sync_possible(struct radeon_ring *ringA, + struct radeon_ring *ringB) +{ + if (ringA->idx == TN_RING_TYPE_VCE2_INDEX && + ringB->idx == TN_RING_TYPE_VCE1_INDEX) + return false; + + return true; +} + void radeon_test_syncing(struct radeon_device *rdev) { int i, j, k; @@ -500,6 +526,9 @@ void radeon_test_syncing(struct radeon_device *rdev) if (!ringB->ready) continue; + if (!radeon_test_sync_possible(ringA, ringB)) + continue; + DRM_INFO("Testing syncing between rings %d and %d...\n", i, j); radeon_test_ring_sync(rdev, ringA, ringB); @@ -511,6 +540,12 @@ void radeon_test_syncing(struct radeon_device *rdev) if (!ringC->ready) continue; + if (!radeon_test_sync_possible(ringA, ringC)) + continue; + + if (!radeon_test_sync_possible(ringB, ringC)) + continue; + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, j, k); radeon_test_ring_sync2(rdev, ringA, ringB, ringC); diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c new file mode 100644 index 000000000000..2547d8ea347a --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -0,0 +1,588 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König + */ + +#include +#include +#include +#include + +#include "radeon.h" +#include "radeon_asic.h" +#include "sid.h" + +/* Firmware Names */ +#define FIRMWARE_BONAIRE "radeon/BONAIRE_vce.bin" + +MODULE_FIRMWARE(FIRMWARE_BONAIRE); + +/** + * radeon_vce_init - allocate memory, load vce firmware + * + * @rdev: radeon_device pointer + * + * First step to get VCE online, allocate memory and load the firmware + */ +int radeon_vce_init(struct radeon_device *rdev) +{ + unsigned long bo_size; + const char *fw_name; + int i, r; + + switch (rdev->family) { + case CHIP_BONAIRE: + case CHIP_KAVERI: + case CHIP_KABINI: + fw_name = FIRMWARE_BONAIRE; + break; + + default: + return -EINVAL; + } + + r = request_firmware(&rdev->vce_fw, fw_name, rdev->dev); + if (r) { + dev_err(rdev->dev, "radeon_vce: Can't load firmware \"%s\"\n", + fw_name); + return r; + } + + bo_size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) + + RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE; + r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo); + if (r) { + dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r); + return r; + } + + r = radeon_vce_resume(rdev); + if (r) + return r; + + memset(rdev->vce.cpu_addr, 0, bo_size); + memcpy(rdev->vce.cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size); + + r = radeon_vce_suspend(rdev); + if (r) + return r; + + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + atomic_set(&rdev->vce.handles[i], 0); + rdev->vce.filp[i] = NULL; + } + + return 0; +} + +/** + * radeon_vce_fini - free memory + * + * @rdev: radeon_device pointer + * + * Last step on VCE teardown, free firmware memory + */ +void radeon_vce_fini(struct radeon_device *rdev) +{ + radeon_vce_suspend(rdev); + radeon_bo_unref(&rdev->vce.vcpu_bo); +} + +/** + * radeon_vce_suspend - unpin VCE fw memory + * + * @rdev: radeon_device pointer + * + * TODO: Test VCE suspend/resume + */ +int radeon_vce_suspend(struct radeon_device *rdev) +{ + int r; + + if (rdev->vce.vcpu_bo == NULL) + return 0; + + r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); + if (!r) { + radeon_bo_kunmap(rdev->vce.vcpu_bo); + radeon_bo_unpin(rdev->vce.vcpu_bo); + radeon_bo_unreserve(rdev->vce.vcpu_bo); + } + return r; +} + +/** + * radeon_vce_resume - pin VCE fw memory + * + * @rdev: radeon_device pointer + * + * TODO: Test VCE suspend/resume + */ +int radeon_vce_resume(struct radeon_device *rdev) +{ + int r; + + if (rdev->vce.vcpu_bo == NULL) + return -EINVAL; + + r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); + if (r) { + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) failed to reserve VCE bo\n", r); + return r; + } + + r = radeon_bo_pin(rdev->vce.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, + &rdev->vce.gpu_addr); + if (r) { + radeon_bo_unreserve(rdev->vce.vcpu_bo); + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) VCE bo pin failed\n", r); + return r; + } + + r = radeon_bo_kmap(rdev->vce.vcpu_bo, &rdev->vce.cpu_addr); + if (r) { + dev_err(rdev->dev, "(%d) VCE map failed\n", r); + return r; + } + + radeon_bo_unreserve(rdev->vce.vcpu_bo); + + return 0; +} + +/** + * radeon_vce_free_handles - free still open VCE handles + * + * @rdev: radeon_device pointer + * @filp: drm file pointer + * + * Close all VCE handles still open by this file pointer + */ +void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp) +{ + int i, r; + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + uint32_t handle = atomic_read(&rdev->vce.handles[i]); + if (!handle || rdev->vce.filp[i] != filp) + continue; + + r = radeon_vce_get_destroy_msg(rdev, TN_RING_TYPE_VCE1_INDEX, + handle, NULL); + if (r) + DRM_ERROR("Error destroying VCE handle (%d)!\n", r); + + rdev->vce.filp[i] = NULL; + atomic_set(&rdev->vce.handles[i], 0); + } +} + +/** + * radeon_vce_get_create_msg - generate a VCE create msg + * + * @rdev: radeon_device pointer + * @ring: ring we should submit the msg to + * @handle: VCE session handle to use + * @fence: optional fence to return + * + * Open up a stream for HW test + */ +int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence) +{ + const unsigned ib_size_dw = 1024; + struct radeon_ib ib; + uint64_t dummy; + int i, r; + + r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + dummy = ib.gpu_addr + 1024; + + /* stitch together an VCE create msg */ + ib.length_dw = 0; + ib.ptr[ib.length_dw++] = 0x0000000c; /* len */ + ib.ptr[ib.length_dw++] = 0x00000001; /* session cmd */ + ib.ptr[ib.length_dw++] = handle; + + ib.ptr[ib.length_dw++] = 0x00000030; /* len */ + ib.ptr[ib.length_dw++] = 0x01000001; /* create cmd */ + ib.ptr[ib.length_dw++] = 0x00000000; + ib.ptr[ib.length_dw++] = 0x00000042; + ib.ptr[ib.length_dw++] = 0x0000000a; + ib.ptr[ib.length_dw++] = 0x00000001; + ib.ptr[ib.length_dw++] = 0x00000080; + ib.ptr[ib.length_dw++] = 0x00000060; + ib.ptr[ib.length_dw++] = 0x00000100; + ib.ptr[ib.length_dw++] = 0x00000100; + ib.ptr[ib.length_dw++] = 0x0000000c; + ib.ptr[ib.length_dw++] = 0x00000000; + + ib.ptr[ib.length_dw++] = 0x00000014; /* len */ + ib.ptr[ib.length_dw++] = 0x05000005; /* feedback buffer */ + ib.ptr[ib.length_dw++] = upper_32_bits(dummy); + ib.ptr[ib.length_dw++] = dummy; + ib.ptr[ib.length_dw++] = 0x00000001; + + for (i = ib.length_dw; i < ib_size_dw; ++i) + ib.ptr[i] = 0x0; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + } + + if (fence) + *fence = radeon_fence_ref(ib.fence); + + radeon_ib_free(rdev, &ib); + + return r; +} + +/** + * radeon_vce_get_destroy_msg - generate a VCE destroy msg + * + * @rdev: radeon_device pointer + * @ring: ring we should submit the msg to + * @handle: VCE session handle to use + * @fence: optional fence to return + * + * Close up a stream for HW test or if userspace failed to do so + */ +int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, + uint32_t handle, struct radeon_fence **fence) +{ + const unsigned ib_size_dw = 1024; + struct radeon_ib ib; + uint64_t dummy; + int i, r; + + r = radeon_ib_get(rdev, ring, &ib, NULL, ib_size_dw * 4); + if (r) { + DRM_ERROR("radeon: failed to get ib (%d).\n", r); + return r; + } + + dummy = ib.gpu_addr + 1024; + + /* stitch together an VCE destroy msg */ + ib.length_dw = 0; + ib.ptr[ib.length_dw++] = 0x0000000c; /* len */ + ib.ptr[ib.length_dw++] = 0x00000001; /* session cmd */ + ib.ptr[ib.length_dw++] = handle; + + ib.ptr[ib.length_dw++] = 0x00000014; /* len */ + ib.ptr[ib.length_dw++] = 0x05000005; /* feedback buffer */ + ib.ptr[ib.length_dw++] = upper_32_bits(dummy); + ib.ptr[ib.length_dw++] = dummy; + ib.ptr[ib.length_dw++] = 0x00000001; + + ib.ptr[ib.length_dw++] = 0x00000008; /* len */ + ib.ptr[ib.length_dw++] = 0x02000001; /* destroy cmd */ + + for (i = ib.length_dw; i < ib_size_dw; ++i) + ib.ptr[i] = 0x0; + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); + } + + if (fence) + *fence = radeon_fence_ref(ib.fence); + + radeon_ib_free(rdev, &ib); + + return r; +} + +/** + * radeon_vce_cs_reloc - command submission relocation + * + * @p: parser context + * @lo: address of lower dword + * @hi: address of higher dword + * + * Patch relocation inside command stream with real buffer address + */ +int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi) +{ + struct radeon_cs_chunk *relocs_chunk; + uint64_t offset; + unsigned idx; + + relocs_chunk = &p->chunks[p->chunk_relocs_idx]; + offset = radeon_get_ib_value(p, lo); + idx = radeon_get_ib_value(p, hi); + + if (idx >= relocs_chunk->length_dw) { + DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", + idx, relocs_chunk->length_dw); + return -EINVAL; + } + + offset += p->relocs_ptr[(idx / 4)]->lobj.gpu_offset; + + p->ib.ptr[lo] = offset & 0xFFFFFFFF; + p->ib.ptr[hi] = offset >> 32; + + return 0; +} + +/** + * radeon_vce_cs_parse - parse and validate the command stream + * + * @p: parser context + * + */ +int radeon_vce_cs_parse(struct radeon_cs_parser *p) +{ + uint32_t handle = 0; + bool destroy = false; + int i, r; + + while (p->idx < p->chunks[p->chunk_ib_idx].length_dw) { + uint32_t len = radeon_get_ib_value(p, p->idx); + uint32_t cmd = radeon_get_ib_value(p, p->idx + 1); + + if ((len < 8) || (len & 3)) { + DRM_ERROR("invalid VCE command length (%d)!\n", len); + return -EINVAL; + } + + switch (cmd) { + case 0x00000001: // session + handle = radeon_get_ib_value(p, p->idx + 2); + break; + + case 0x00000002: // task info + case 0x01000001: // create + case 0x04000001: // config extension + case 0x04000002: // pic control + case 0x04000005: // rate control + case 0x04000007: // motion estimation + case 0x04000008: // rdo + break; + + case 0x03000001: // encode + r = radeon_vce_cs_reloc(p, p->idx + 10, p->idx + 9); + if (r) + return r; + + r = radeon_vce_cs_reloc(p, p->idx + 12, p->idx + 11); + if (r) + return r; + break; + + case 0x02000001: // destroy + destroy = true; + break; + + case 0x05000001: // context buffer + case 0x05000004: // video bitstream buffer + case 0x05000005: // feedback buffer + r = radeon_vce_cs_reloc(p, p->idx + 3, p->idx + 2); + if (r) + return r; + break; + + default: + DRM_ERROR("invalid VCE command (0x%x)!\n", cmd); + return -EINVAL; + } + + p->idx += len / 4; + } + + if (destroy) { + /* IB contains a destroy msg, free the handle */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) + atomic_cmpxchg(&p->rdev->vce.handles[i], handle, 0); + + return 0; + } + + /* create or encode, validate the handle */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + if (atomic_read(&p->rdev->vce.handles[i]) == handle) + return 0; + } + + /* handle not found try to alloc a new one */ + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { + if (!atomic_cmpxchg(&p->rdev->vce.handles[i], 0, handle)) { + p->rdev->vce.filp[i] = p->filp; + return 0; + } + } + + DRM_ERROR("No more free VCE handles!\n"); + return -EINVAL; +} + +/** + * radeon_vce_semaphore_emit - emit a semaphore command + * + * @rdev: radeon_device pointer + * @ring: engine to use + * @semaphore: address of semaphore + * @emit_wait: true=emit wait, false=emit signal + * + */ +bool radeon_vce_semaphore_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + uint64_t addr = semaphore->gpu_addr; + + radeon_ring_write(ring, VCE_CMD_SEMAPHORE); + radeon_ring_write(ring, (addr >> 3) & 0x000FFFFF); + radeon_ring_write(ring, (addr >> 23) & 0x000FFFFF); + radeon_ring_write(ring, 0x01003000 | (emit_wait ? 1 : 0)); + if (!emit_wait) + radeon_ring_write(ring, VCE_CMD_END); + + return true; +} + +/** + * radeon_vce_ib_execute - execute indirect buffer + * + * @rdev: radeon_device pointer + * @ib: the IB to execute + * + */ +void radeon_vce_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->ring]; + radeon_ring_write(ring, VCE_CMD_IB); + radeon_ring_write(ring, ib->gpu_addr); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr)); + radeon_ring_write(ring, ib->length_dw); +} + +/** + * radeon_vce_fence_emit - add a fence command to the ring + * + * @rdev: radeon_device pointer + * @fence: the fence + * + */ +void radeon_vce_fence_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + uint32_t addr = rdev->fence_drv[fence->ring].gpu_addr; + + radeon_ring_write(ring, VCE_CMD_FENCE); + radeon_ring_write(ring, addr); + radeon_ring_write(ring, upper_32_bits(addr)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, VCE_CMD_TRAP); + radeon_ring_write(ring, VCE_CMD_END); +} + +/** + * radeon_vce_ring_test - test if VCE ring is working + * + * @rdev: radeon_device pointer + * @ring: the engine to test on + * + */ +int radeon_vce_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + uint32_t rptr = vce_v1_0_get_rptr(rdev, ring); + unsigned i; + int r; + + r = radeon_ring_lock(rdev, ring, 16); + if (r) { + DRM_ERROR("radeon: vce failed to lock ring %d (%d).\n", + ring->idx, r); + return r; + } + radeon_ring_write(ring, VCE_CMD_END); + radeon_ring_unlock_commit(rdev, ring); + + for (i = 0; i < rdev->usec_timeout; i++) { + if (vce_v1_0_get_rptr(rdev, ring) != rptr) + break; + DRM_UDELAY(1); + } + + if (i < rdev->usec_timeout) { + DRM_INFO("ring test on %d succeeded in %d usecs\n", + ring->idx, i); + } else { + DRM_ERROR("radeon: ring %d test failed\n", + ring->idx); + r = -ETIMEDOUT; + } + + return r; +} + +/** + * radeon_vce_ib_test - test if VCE IBs are working + * + * @rdev: radeon_device pointer + * @ring: the engine to test on + * + */ +int radeon_vce_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) +{ + struct radeon_fence *fence = NULL; + int r; + + r = radeon_vce_get_create_msg(rdev, ring->idx, 1, NULL); + if (r) { + DRM_ERROR("radeon: failed to get create msg (%d).\n", r); + goto error; + } + + r = radeon_vce_get_destroy_msg(rdev, ring->idx, 1, &fence); + if (r) { + DRM_ERROR("radeon: failed to get destroy ib (%d).\n", r); + goto error; + } + + r = radeon_fence_wait(fence, false); + if (r) { + DRM_ERROR("radeon: fence wait failed (%d).\n", r); + } else { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + } +error: + radeon_fence_unref(&fence); + return r; +} diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h index 9239a6d29128..683532f84931 100644 --- a/drivers/gpu/drm/radeon/sid.h +++ b/drivers/gpu/drm/radeon/sid.h @@ -1798,4 +1798,51 @@ #define DMA_PACKET_CONSTANT_FILL 0xd #define DMA_PACKET_NOP 0xf +#define VCE_STATUS 0x20004 +#define VCE_VCPU_CNTL 0x20014 +#define VCE_CLK_EN (1 << 0) +#define VCE_VCPU_CACHE_OFFSET0 0x20024 +#define VCE_VCPU_CACHE_SIZE0 0x20028 +#define VCE_VCPU_CACHE_OFFSET1 0x2002c +#define VCE_VCPU_CACHE_SIZE1 0x20030 +#define VCE_VCPU_CACHE_OFFSET2 0x20034 +#define VCE_VCPU_CACHE_SIZE2 0x20038 +#define VCE_SOFT_RESET 0x20120 +#define VCE_ECPU_SOFT_RESET (1 << 0) +#define VCE_FME_SOFT_RESET (1 << 2) +#define VCE_RB_BASE_LO2 0x2016c +#define VCE_RB_BASE_HI2 0x20170 +#define VCE_RB_SIZE2 0x20174 +#define VCE_RB_RPTR2 0x20178 +#define VCE_RB_WPTR2 0x2017c +#define VCE_RB_BASE_LO 0x20180 +#define VCE_RB_BASE_HI 0x20184 +#define VCE_RB_SIZE 0x20188 +#define VCE_RB_RPTR 0x2018c +#define VCE_RB_WPTR 0x20190 +#define VCE_CLOCK_GATING_A 0x202f8 +#define VCE_CLOCK_GATING_B 0x202fc +#define VCE_UENC_CLOCK_GATING 0x205bc +#define VCE_UENC_REG_CLOCK_GATING 0x205c0 +#define VCE_FW_REG_STATUS 0x20e10 +# define VCE_FW_REG_STATUS_BUSY (1 << 0) +# define VCE_FW_REG_STATUS_PASS (1 << 3) +# define VCE_FW_REG_STATUS_DONE (1 << 11) +#define VCE_LMI_FW_START_KEYSEL 0x20e18 +#define VCE_LMI_FW_PERIODIC_CTRL 0x20e20 +#define VCE_LMI_CTRL2 0x20e74 +#define VCE_LMI_CTRL 0x20e98 +#define VCE_LMI_VM_CTRL 0x20ea0 +#define VCE_LMI_SWAP_CNTL 0x20eb4 +#define VCE_LMI_SWAP_CNTL1 0x20eb8 +#define VCE_LMI_CACHE_CTRL 0x20ef4 + +#define VCE_CMD_NO_OP 0x00000000 +#define VCE_CMD_END 0x00000001 +#define VCE_CMD_IB 0x00000002 +#define VCE_CMD_FENCE 0x00000003 +#define VCE_CMD_TRAP 0x00000004 +#define VCE_CMD_IB_AUTO 0x00000005 +#define VCE_CMD_SEMAPHORE 0x00000006 + #endif diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c new file mode 100644 index 000000000000..e0c3534356a1 --- /dev/null +++ b/drivers/gpu/drm/radeon/vce_v1_0.c @@ -0,0 +1,187 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König + */ + +#include +#include +#include "radeon.h" +#include "radeon_asic.h" +#include "sid.h" + +/** + * vce_v1_0_get_rptr - get read pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware read pointer + */ +uint32_t vce_v1_0_get_rptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + return RREG32(VCE_RB_RPTR); + else + return RREG32(VCE_RB_RPTR2); +} + +/** + * vce_v1_0_get_wptr - get write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Returns the current hardware write pointer + */ +uint32_t vce_v1_0_get_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + return RREG32(VCE_RB_WPTR); + else + return RREG32(VCE_RB_WPTR2); +} + +/** + * vce_v1_0_set_wptr - set write pointer + * + * @rdev: radeon_device pointer + * @ring: radeon_ring pointer + * + * Commits the write pointer to the hardware + */ +void vce_v1_0_set_wptr(struct radeon_device *rdev, + struct radeon_ring *ring) +{ + if (ring->idx == TN_RING_TYPE_VCE1_INDEX) + WREG32(VCE_RB_WPTR, ring->wptr); + else + WREG32(VCE_RB_WPTR2, ring->wptr); +} + +/** + * vce_v1_0_start - start VCE block + * + * @rdev: radeon_device pointer + * + * Setup and start the VCE block + */ +int vce_v1_0_start(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int i, j, r; + + /* set BUSY flag */ + WREG32_P(VCE_STATUS, 1, ~1); + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + WREG32(VCE_RB_RPTR, ring->rptr); + WREG32(VCE_RB_WPTR, ring->wptr); + WREG32(VCE_RB_BASE_LO, ring->gpu_addr); + WREG32(VCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32(VCE_RB_SIZE, ring->ring_size / 4); + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + WREG32(VCE_RB_RPTR2, ring->rptr); + WREG32(VCE_RB_WPTR2, ring->wptr); + WREG32(VCE_RB_BASE_LO2, ring->gpu_addr); + WREG32(VCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32(VCE_RB_SIZE2, ring->ring_size / 4); + + WREG32_P(VCE_VCPU_CNTL, VCE_CLK_EN, ~VCE_CLK_EN); + + WREG32_P(VCE_SOFT_RESET, + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET, ~( + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET)); + + mdelay(100); + + WREG32_P(VCE_SOFT_RESET, 0, ~( + VCE_ECPU_SOFT_RESET | + VCE_FME_SOFT_RESET)); + + for (i = 0; i < 10; ++i) { + uint32_t status; + for (j = 0; j < 100; ++j) { + status = RREG32(VCE_STATUS); + if (status & 2) + break; + mdelay(10); + } + r = 0; + if (status & 2) + break; + + DRM_ERROR("VCE not responding, trying to reset the ECPU!!!\n"); + WREG32_P(VCE_SOFT_RESET, VCE_ECPU_SOFT_RESET, ~VCE_ECPU_SOFT_RESET); + mdelay(10); + WREG32_P(VCE_SOFT_RESET, 0, ~VCE_ECPU_SOFT_RESET); + mdelay(10); + r = -1; + } + + /* clear BUSY flag */ + WREG32_P(VCE_STATUS, 0, ~1); + + if (r) { + DRM_ERROR("VCE not responding, giving up!!!\n"); + return r; + } + + return 0; +} + +int vce_v1_0_init(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int r; + + r = vce_v1_0_start(rdev); + if (r) + return r; + + ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; + ring->ready = true; + r = radeon_ring_test(rdev, TN_RING_TYPE_VCE1_INDEX, ring); + if (r) { + ring->ready = false; + return r; + } + + ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; + ring->ready = true; + r = radeon_ring_test(rdev, TN_RING_TYPE_VCE2_INDEX, ring); + if (r) { + ring->ready = false; + return r; + } + + DRM_INFO("VCE initialized successfully.\n"); + + return 0; +} diff --git a/drivers/gpu/drm/radeon/vce_v2_0.c b/drivers/gpu/drm/radeon/vce_v2_0.c new file mode 100644 index 000000000000..4911d1b00e3b --- /dev/null +++ b/drivers/gpu/drm/radeon/vce_v2_0.c @@ -0,0 +1,70 @@ +/* + * Copyright 2013 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * Authors: Christian König + */ + +#include +#include +#include "radeon.h" +#include "radeon_asic.h" +#include "cikd.h" + +int vce_v2_0_resume(struct radeon_device *rdev) +{ + uint64_t addr = rdev->vce.gpu_addr; + uint32_t size; + + WREG32_P(VCE_CLOCK_GATING_A, 0, ~(1 << 16)); + WREG32_P(VCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000); + WREG32_P(VCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F); + WREG32(VCE_CLOCK_GATING_B, 0xf7); + + WREG32(VCE_LMI_CTRL, 0x00398000); + WREG32_P(VCE_LMI_CACHE_CTRL, 0x0, ~0x1); + WREG32(VCE_LMI_SWAP_CNTL, 0); + WREG32(VCE_LMI_SWAP_CNTL1, 0); + WREG32(VCE_LMI_VM_CTRL, 0); + + size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size); + WREG32(VCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE0, size); + + addr += size; + size = RADEON_VCE_STACK_SIZE; + WREG32(VCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE1, size); + + addr += size; + size = RADEON_VCE_HEAP_SIZE; + WREG32(VCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff); + WREG32(VCE_VCPU_CACHE_SIZE2, size); + + WREG32_P(VCE_LMI_CTRL2, 0x0, ~0x100); + + WREG32_P(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, + ~VCE_SYS_INT_TRAP_INTERRUPT_EN); + + return 0; +} diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h index d9ea3a73afe2..6493ca55b463 100644 --- a/include/uapi/drm/radeon_drm.h +++ b/include/uapi/drm/radeon_drm.h @@ -919,6 +919,7 @@ struct drm_radeon_gem_va { #define RADEON_CS_RING_COMPUTE 1 #define RADEON_CS_RING_DMA 2 #define RADEON_CS_RING_UVD 3 +#define RADEON_CS_RING_VCE 4 /* The third dword of RADEON_CHUNK_ID_FLAGS is a sint32 that sets the priority */ /* 0 = normal, + = higher priority, - = lower priority */ From f7ba8b04b22d7c74898f53a0e118f31b9d40dcc2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 27 Jan 2014 10:16:06 -0700 Subject: [PATCH 022/259] drm/radeon: add VCE ring query MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon_kms.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 0e078afc5db7..ea018d53a85e 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -433,6 +433,9 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file case RADEON_CS_RING_UVD: *value = rdev->ring[R600_RING_TYPE_UVD_INDEX].ready; break; + case RADEON_CS_RING_VCE: + *value = rdev->ring[TN_RING_TYPE_VCE1_INDEX].ready; + break; default: return -EINVAL; } From 98ccc291ffdc34ccb9b13f0c29cc51d6eab24022 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 23 Jan 2014 09:50:49 -0700 Subject: [PATCH 023/259] drm/radeon: add VCE version parsing and checking MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Also make the result available to userspace. Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 2 ++ drivers/gpu/drm/radeon/radeon_kms.c | 6 ++++ drivers/gpu/drm/radeon/radeon_vce.c | 56 +++++++++++++++++++++++++---- include/uapi/drm/radeon_drm.h | 4 +++ 4 files changed, 62 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index a58a38942c73..d1491d41fc2d 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1606,6 +1606,8 @@ struct radeon_vce { struct radeon_bo *vcpu_bo; void *cpu_addr; uint64_t gpu_addr; + unsigned fw_version; + unsigned fb_version; atomic_t handles[RADEON_MAX_VCE_HANDLES]; struct drm_file *filp[RADEON_MAX_VCE_HANDLES]; }; diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index ea018d53a85e..baff98be65b1 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -480,6 +480,12 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file else *value = rdev->pm.default_sclk * 10; break; + case RADEON_INFO_VCE_FW_VERSION: + *value = rdev->vce.fw_version; + break; + case RADEON_INFO_VCE_FB_VERSION: + *value = rdev->vce.fb_version; + break; default: DRM_DEBUG_KMS("Invalid request %d\n", info->request); return -EINVAL; diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index 2547d8ea347a..f46563b60921 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -48,8 +48,11 @@ MODULE_FIRMWARE(FIRMWARE_BONAIRE); */ int radeon_vce_init(struct radeon_device *rdev) { - unsigned long bo_size; - const char *fw_name; + static const char *fw_version = "[ATI LIB=VCEFW,"; + static const char *fb_version = "[ATI LIB=VCEFWSTATS,"; + unsigned long size; + const char *fw_name, *c; + uint8_t start, mid, end; int i, r; switch (rdev->family) { @@ -70,9 +73,50 @@ int radeon_vce_init(struct radeon_device *rdev) return r; } - bo_size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) + - RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE; - r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true, + /* search for firmware version */ + + size = rdev->vce_fw->size - strlen(fw_version) - 9; + c = rdev->vce_fw->data; + for (;size > 0; --size, ++c) + if (strncmp(c, fw_version, strlen(fw_version)) == 0) + break; + + if (size == 0) + return -EINVAL; + + c += strlen(fw_version); + if (sscanf(c, "%2hhd.%2hhd.%2hhd]", &start, &mid, &end) != 3) + return -EINVAL; + + /* search for feedback version */ + + size = rdev->vce_fw->size - strlen(fb_version) - 3; + c = rdev->vce_fw->data; + for (;size > 0; --size, ++c) + if (strncmp(c, fb_version, strlen(fb_version)) == 0) + break; + + if (size == 0) + return -EINVAL; + + c += strlen(fb_version); + if (sscanf(c, "%2u]", &rdev->vce.fb_version) != 1) + return -EINVAL; + + DRM_INFO("Found VCE firmware/feedback version %hhd.%hhd.%hhd / %d!\n", + start, mid, end, rdev->vce.fb_version); + + rdev->vce.fw_version = (start << 24) | (mid << 16) | (end << 8); + + /* we can only work with this fw version for now */ + if (rdev->vce.fw_version != ((40 << 24) | (2 << 16) | (2 << 8))) + return -EINVAL; + + /* load firmware into VRAM */ + + size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) + + RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE; + r = radeon_bo_create(rdev, size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->vce.vcpu_bo); if (r) { dev_err(rdev->dev, "(%d) failed to allocate VCE bo\n", r); @@ -83,7 +127,7 @@ int radeon_vce_init(struct radeon_device *rdev) if (r) return r; - memset(rdev->vce.cpu_addr, 0, bo_size); + memset(rdev->vce.cpu_addr, 0, size); memcpy(rdev->vce.cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size); r = radeon_vce_suspend(rdev); diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h index 6493ca55b463..1cf18b4a39ec 100644 --- a/include/uapi/drm/radeon_drm.h +++ b/include/uapi/drm/radeon_drm.h @@ -988,6 +988,10 @@ struct drm_radeon_cs { #define RADEON_INFO_SI_BACKEND_ENABLED_MASK 0x19 /* max engine clock - needed for OpenCL */ #define RADEON_INFO_MAX_SCLK 0x1a +/* version of VCE firmware */ +#define RADEON_INFO_VCE_FW_VERSION 0x1b +/* version of VCE feedback */ +#define RADEON_INFO_VCE_FB_VERSION 0x1c struct drm_radeon_info { From b59b733397cac70be5b04c60e8810077ac6ca48d Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 20 Aug 2013 20:01:18 -0400 Subject: [PATCH 024/259] drm/radeon: add callback for setting vce clocks Similar to uvd clock setting. Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index d1491d41fc2d..2b26feb0a7c6 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1822,6 +1822,7 @@ struct radeon_asic { void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes); void (*set_clock_gating)(struct radeon_device *rdev, int enable); int (*set_uvd_clocks)(struct radeon_device *rdev, u32 vclk, u32 dclk); + int (*set_vce_clocks)(struct radeon_device *rdev, u32 evclk, u32 ecclk); int (*get_temperature)(struct radeon_device *rdev); } pm; /* dynamic power management */ @@ -2683,6 +2684,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l)) #define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e)) #define radeon_set_uvd_clocks(rdev, v, d) (rdev)->asic->pm.set_uvd_clocks((rdev), (v), (d)) +#define radeon_set_vce_clocks(rdev, ev, ec) (rdev)->asic->pm.set_vce_clocks((rdev), (ev), (ec)) #define radeon_get_temperature(rdev) (rdev)->asic->pm.get_temperature((rdev)) #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s))) #define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r))) From 82f79cc54b6a67c0b17aff4fb5ed43155ff3f0ea Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 21 Aug 2013 10:02:32 -0400 Subject: [PATCH 025/259] drm/radeon/dpm: move platform caps fetching to a separate function It's needed by by both the asic specific functions and the extended table parser. Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/btc_dpm.c | 4 ++++ drivers/gpu/drm/radeon/ci_dpm.c | 9 ++++++--- drivers/gpu/drm/radeon/cypress_dpm.c | 4 ++++ drivers/gpu/drm/radeon/kv_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/ni_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/r600_dpm.c | 20 ++++++++++++++++++++ drivers/gpu/drm/radeon/r600_dpm.h | 2 ++ drivers/gpu/drm/radeon/rs780_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/rv6xx_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/rv770_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/si_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/sumo_dpm.c | 7 ++++--- drivers/gpu/drm/radeon/trinity_dpm.c | 7 ++++--- 13 files changed, 68 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c index ea103ccdf4bd..f81d7ca134db 100644 --- a/drivers/gpu/drm/radeon/btc_dpm.c +++ b/drivers/gpu/drm/radeon/btc_dpm.c @@ -2601,6 +2601,10 @@ int btc_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 8d49104ca6c2..4a0c40186046 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -4959,9 +4959,6 @@ static int ci_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { u8 *idx; power_state = (union pplib_power_state *)power_state_offset; @@ -5077,6 +5074,12 @@ int ci_dpm_init(struct radeon_device *rdev) ci_dpm_fini(rdev); return ret; } + + ret = r600_get_platform_caps(rdev); + if (ret) { + ci_dpm_fini(rdev); + return ret; + } ret = ci_parse_power_table(rdev); if (ret) { ci_dpm_fini(rdev); diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c index cf783fc0ef21..5a9a5f4d7888 100644 --- a/drivers/gpu/drm/radeon/cypress_dpm.c +++ b/drivers/gpu/drm/radeon/cypress_dpm.c @@ -2036,6 +2036,10 @@ int cypress_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c index 351db361239d..b5bb3a5654ce 100644 --- a/drivers/gpu/drm/radeon/kv_dpm.c +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -2538,9 +2538,6 @@ static int kv_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { u8 *idx; power_state = (union pplib_power_state *)power_state_offset; @@ -2590,6 +2587,10 @@ int kv_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = r600_parse_extended_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c index 1217fbcbdcca..89fc5b976d9b 100644 --- a/drivers/gpu/drm/radeon/ni_dpm.c +++ b/drivers/gpu/drm/radeon/ni_dpm.c @@ -4025,9 +4025,6 @@ static int ni_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -4089,6 +4086,10 @@ int ni_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = ni_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c index e4cc9b314ce9..e8b6e4ab312b 100644 --- a/drivers/gpu/drm/radeon/r600_dpm.c +++ b/drivers/gpu/drm/radeon/r600_dpm.c @@ -834,6 +834,26 @@ static int r600_parse_clk_voltage_dep_table(struct radeon_clock_voltage_dependen return 0; } +int r600_get_platform_caps(struct radeon_device *rdev) +{ + struct radeon_mode_info *mode_info = &rdev->mode_info; + union power_info *power_info; + int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); + u16 data_offset; + u8 frev, crev; + + if (!atom_parse_data_header(mode_info->atom_context, index, NULL, + &frev, &crev, &data_offset)) + return -EINVAL; + power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); + + rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); + rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); + rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); + + return 0; +} + /* sizeof(ATOM_PPLIB_EXTENDEDHEADER) */ #define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V2 12 #define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3 14 diff --git a/drivers/gpu/drm/radeon/r600_dpm.h b/drivers/gpu/drm/radeon/r600_dpm.h index 07eab2b04e81..46b9d2a03018 100644 --- a/drivers/gpu/drm/radeon/r600_dpm.h +++ b/drivers/gpu/drm/radeon/r600_dpm.h @@ -215,6 +215,8 @@ void r600_stop_dpm(struct radeon_device *rdev); bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor); +int r600_get_platform_caps(struct radeon_device *rdev); + int r600_parse_extended_power_table(struct radeon_device *rdev); void r600_free_extended_power_table(struct radeon_device *rdev); diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c index 8512085b0aef..02f7710de470 100644 --- a/drivers/gpu/drm/radeon/rs780_dpm.c +++ b/drivers/gpu/drm/radeon/rs780_dpm.c @@ -807,9 +807,6 @@ static int rs780_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -859,6 +856,10 @@ int rs780_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rs780_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c index bebf31c4d841..e7045b085715 100644 --- a/drivers/gpu/drm/radeon/rv6xx_dpm.c +++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c @@ -1891,9 +1891,6 @@ static int rv6xx_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -1943,6 +1940,10 @@ int rv6xx_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv6xx_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c index 5b2ea8ac0731..9098c86ca006 100644 --- a/drivers/gpu/drm/radeon/rv770_dpm.c +++ b/drivers/gpu/drm/radeon/rv770_dpm.c @@ -2281,9 +2281,6 @@ int rv7xx_parse_power_table(struct radeon_device *rdev) power_info->pplib.ucNumStates, GFP_KERNEL); if (!rdev->pm.dpm.ps) return -ENOMEM; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < power_info->pplib.ucNumStates; i++) { power_state = (union pplib_power_state *) @@ -2361,6 +2358,10 @@ int rv770_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = rv7xx_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index eafb0e6bc67e..d5024778cabe 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -6271,9 +6271,6 @@ static int si_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { u8 *idx; power_state = (union pplib_power_state *)power_state_offset; @@ -6350,6 +6347,10 @@ int si_dpm_init(struct radeon_device *rdev) pi->min_vddc_in_table = 0; pi->max_vddc_in_table = 0; + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = si_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c index 8b47b3cd0357..3f0e8d7b8dbe 100644 --- a/drivers/gpu/drm/radeon/sumo_dpm.c +++ b/drivers/gpu/drm/radeon/sumo_dpm.c @@ -1484,9 +1484,6 @@ static int sumo_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { u8 *idx; power_state = (union pplib_power_state *)power_state_offset; @@ -1772,6 +1769,10 @@ int sumo_dpm_init(struct radeon_device *rdev) sumo_construct_boot_and_acpi_state(rdev); + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = sumo_parse_power_table(rdev); if (ret) return ret; diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c index 2da0e17eb960..2a2822c03329 100644 --- a/drivers/gpu/drm/radeon/trinity_dpm.c +++ b/drivers/gpu/drm/radeon/trinity_dpm.c @@ -1694,9 +1694,6 @@ static int trinity_parse_power_table(struct radeon_device *rdev) if (!rdev->pm.dpm.ps) return -ENOMEM; power_state_offset = (u8 *)state_array->states; - rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps); - rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime); - rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime); for (i = 0; i < state_array->ucNumEntries; i++) { u8 *idx; power_state = (union pplib_power_state *)power_state_offset; @@ -1895,6 +1892,10 @@ int trinity_dpm_init(struct radeon_device *rdev) trinity_construct_boot_state(rdev); + ret = r600_get_platform_caps(rdev); + if (ret) + return ret; + ret = trinity_parse_power_table(rdev); if (ret) return ret; From b62d628bd63f61e9aea3b8fab2ec638680bf4aa4 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 20 Aug 2013 20:29:05 -0400 Subject: [PATCH 026/259] drm/radeon/dpm: fill in some initial vce infrastructure Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 12 ++++++++++++ drivers/gpu/drm/radeon/radeon_pm.c | 8 ++++++++ 2 files changed, 20 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 2b26feb0a7c6..60c171c60a64 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1259,6 +1259,15 @@ enum radeon_dpm_event_src { RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL = 4 }; +enum radeon_vce_level { + RADEON_VCE_LEVEL_AC_ALL = 0, /* AC, All cases */ + RADEON_VCE_LEVEL_DC_EE = 1, /* DC, entropy encoding */ + RADEON_VCE_LEVEL_DC_LL_LOW = 2, /* DC, low latency queue, res <= 720 */ + RADEON_VCE_LEVEL_DC_LL_HIGH = 3, /* DC, low latency queue, 1080 >= res > 720 */ + RADEON_VCE_LEVEL_DC_GP_LOW = 4, /* DC, general purpose queue, res <= 720 */ + RADEON_VCE_LEVEL_DC_GP_HIGH = 5, /* DC, general purpose queue, 1080 >= res > 720 */ +}; + struct radeon_ps { u32 caps; /* vbios flags */ u32 class; /* vbios flags */ @@ -1269,6 +1278,8 @@ struct radeon_ps { /* VCE clocks */ u32 evclk; u32 ecclk; + bool vce_active; + enum radeon_vce_level vce_level; /* asic priv */ void *ps_priv; }; @@ -1480,6 +1491,7 @@ struct radeon_dpm { /* special states active */ bool thermal_active; bool uvd_active; + bool vce_active; /* thermal handling */ struct radeon_dpm_thermal thermal; /* forced levels */ diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index 8e8153e471c2..a4687e7b45f8 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -826,6 +826,9 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) /* no need to reprogram if nothing changed unless we are on BTC+ */ if (rdev->pm.dpm.current_ps == rdev->pm.dpm.requested_ps) { + /* vce just modifies an existing state so force a change */ + if (ps->vce_active != rdev->pm.dpm.vce_active) + goto force; if ((rdev->family < CHIP_BARTS) || (rdev->flags & RADEON_IS_IGP)) { /* for pre-BTC and APUs if the num crtcs changed but state is the same, * all we need to do is update the display configuration. @@ -862,16 +865,21 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) } } +force: if (radeon_dpm == 1) { printk("switching from power state:\n"); radeon_dpm_print_power_state(rdev, rdev->pm.dpm.current_ps); printk("switching to power state:\n"); radeon_dpm_print_power_state(rdev, rdev->pm.dpm.requested_ps); } + mutex_lock(&rdev->ddev->struct_mutex); down_write(&rdev->pm.mclk_lock); mutex_lock(&rdev->ring_lock); + /* update whether vce is active */ + ps->vce_active = rdev->pm.dpm.vce_active; + ret = radeon_dpm_pre_set_power_state(rdev); if (ret) goto done; From 58bd2a88facbdf3c39db0f834111cd4294400814 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 4 Sep 2013 16:13:56 -0400 Subject: [PATCH 027/259] drm/radeon/dpm: fetch vce states from the vbios Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/r600_dpm.c | 28 +++++++++++++++++++++++++++- drivers/gpu/drm/radeon/radeon.h | 16 ++++++++++++++++ 2 files changed, 43 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c index e8b6e4ab312b..cbf7e3269f84 100644 --- a/drivers/gpu/drm/radeon/r600_dpm.c +++ b/drivers/gpu/drm/radeon/r600_dpm.c @@ -1063,7 +1063,15 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) (mode_info->atom_context->bios + data_offset + le16_to_cpu(ext_hdr->usVCETableOffset) + 1 + 1 + array->ucNumEntries * sizeof(VCEClockInfo)); + ATOM_PPLIB_VCE_State_Table *states = + (ATOM_PPLIB_VCE_State_Table *) + (mode_info->atom_context->bios + data_offset + + le16_to_cpu(ext_hdr->usVCETableOffset) + 1 + + 1 + (array->ucNumEntries * sizeof (VCEClockInfo)) + + 1 + (limits->numEntries * sizeof(ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record))); ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record *entry; + ATOM_PPLIB_VCE_State_Record *state_entry; + VCEClockInfo *vce_clk; u32 size = limits->numEntries * sizeof(struct radeon_vce_clock_voltage_dependency_entry); rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries = @@ -1075,8 +1083,9 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count = limits->numEntries; entry = &limits->entries[0]; + state_entry = &states->entries[0]; for (i = 0; i < limits->numEntries; i++) { - VCEClockInfo *vce_clk = (VCEClockInfo *) + vce_clk = (VCEClockInfo *) ((u8 *)&array->entries[0] + (entry->ucVCEClockInfoIndex * sizeof(VCEClockInfo))); rdev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].evclk = @@ -1088,6 +1097,23 @@ int r600_parse_extended_power_table(struct radeon_device *rdev) entry = (ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record *) ((u8 *)entry + sizeof(ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record)); } + for (i = 0; i < states->numEntries; i++) { + if (i >= RADEON_MAX_VCE_LEVELS) + break; + vce_clk = (VCEClockInfo *) + ((u8 *)&array->entries[0] + + (state_entry->ucVCEClockInfoIndex * sizeof(VCEClockInfo))); + rdev->pm.dpm.vce_states[i].evclk = + le16_to_cpu(vce_clk->usEVClkLow) | (vce_clk->ucEVClkHigh << 16); + rdev->pm.dpm.vce_states[i].ecclk = + le16_to_cpu(vce_clk->usECClkLow) | (vce_clk->ucECClkHigh << 16); + rdev->pm.dpm.vce_states[i].clk_idx = + state_entry->ucClockInfoIndex & 0x3f; + rdev->pm.dpm.vce_states[i].pstate = + (state_entry->ucClockInfoIndex & 0xc0) >> 6; + state_entry = (ATOM_PPLIB_VCE_State_Record *) + ((u8 *)state_entry + sizeof(ATOM_PPLIB_VCE_State_Record)); + } } if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3) && ext_hdr->usUVDTableOffset) { diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 60c171c60a64..693a8fccd94d 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1259,6 +1259,8 @@ enum radeon_dpm_event_src { RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL = 4 }; +#define RADEON_MAX_VCE_LEVELS 6 + enum radeon_vce_level { RADEON_VCE_LEVEL_AC_ALL = 0, /* AC, All cases */ RADEON_VCE_LEVEL_DC_EE = 1, /* DC, entropy encoding */ @@ -1454,6 +1456,17 @@ enum radeon_dpm_forced_level { RADEON_DPM_FORCED_LEVEL_HIGH = 2, }; +struct radeon_vce_state { + /* vce clocks */ + u32 evclk; + u32 ecclk; + /* gpu clocks */ + u32 sclk; + u32 mclk; + u8 clk_idx; + u8 pstate; +}; + struct radeon_dpm { struct radeon_ps *ps; /* number of valid power states */ @@ -1466,6 +1479,9 @@ struct radeon_dpm { struct radeon_ps *boot_ps; /* default uvd power state */ struct radeon_ps *uvd_ps; + /* vce requirements */ + struct radeon_vce_state vce_states[RADEON_MAX_VCE_LEVELS]; + enum radeon_vce_level vce_level; enum radeon_pm_state_type state; enum radeon_pm_state_type user_state; u32 platform_caps; From 5ad6bf91ef8fd265aee252982a7d6fcf78436153 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 22 Aug 2013 17:09:06 -0400 Subject: [PATCH 028/259] drm/radeon: fill in set_vce_clocks for CIK asics Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/cik.c | 35 ++++++++++++++++++++++++++++ drivers/gpu/drm/radeon/cikd.h | 6 +++++ drivers/gpu/drm/radeon/radeon_asic.c | 2 ++ drivers/gpu/drm/radeon/radeon_asic.h | 1 + 4 files changed, 44 insertions(+) diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index be6eb4d91284..ecb16b14f049 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -8925,6 +8925,41 @@ int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk) return r; } +int cik_set_vce_clocks(struct radeon_device *rdev, u32 evclk, u32 ecclk) +{ + int r, i; + struct atom_clock_dividers dividers; + u32 tmp; + + r = radeon_atom_get_clock_dividers(rdev, COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, + ecclk, false, ÷rs); + if (r) + return r; + + for (i = 0; i < 100; i++) { + if (RREG32_SMC(CG_ECLK_STATUS) & ECLK_STATUS) + break; + mdelay(10); + } + if (i == 100) + return -ETIMEDOUT; + + tmp = RREG32_SMC(CG_ECLK_CNTL); + tmp &= ~(ECLK_DIR_CNTL_EN|ECLK_DIVIDER_MASK); + tmp |= dividers.post_divider; + WREG32_SMC(CG_ECLK_CNTL, tmp); + + for (i = 0; i < 100; i++) { + if (RREG32_SMC(CG_ECLK_STATUS) & ECLK_STATUS) + break; + mdelay(10); + } + if (i == 100) + return -ETIMEDOUT; + + return 0; +} + static void cik_pcie_gen3_enable(struct radeon_device *rdev) { struct pci_dev *root = rdev->pdev->bus->self; diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h index 459ae021d91c..ee16380ceba8 100644 --- a/drivers/gpu/drm/radeon/cikd.h +++ b/drivers/gpu/drm/radeon/cikd.h @@ -203,6 +203,12 @@ #define CTF_TEMP_MASK 0x0003fe00 #define CTF_TEMP_SHIFT 9 +#define CG_ECLK_CNTL 0xC05000AC +# define ECLK_DIVIDER_MASK 0x7f +# define ECLK_DIR_CNTL_EN (1 << 8) +#define CG_ECLK_STATUS 0xC05000B0 +# define ECLK_STATUS (1 << 0) + #define CG_SPLL_FUNC_CNTL 0xC0500140 #define SPLL_RESET (1 << 0) #define SPLL_PWRON (1 << 1) diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index 4f059b2c05fb..b8a24a75d4ff 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -2076,6 +2076,7 @@ static struct radeon_asic ci_asic = { .set_pcie_lanes = NULL, .set_clock_gating = NULL, .set_uvd_clocks = &cik_set_uvd_clocks, + .set_vce_clocks = &cik_set_vce_clocks, .get_temperature = &ci_get_temp, }, .dpm = { @@ -2180,6 +2181,7 @@ static struct radeon_asic kv_asic = { .set_pcie_lanes = NULL, .set_clock_gating = NULL, .set_uvd_clocks = &cik_set_uvd_clocks, + .set_vce_clocks = &cik_set_vce_clocks, .get_temperature = &kv_get_temp, }, .dpm = { diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index 13f87bf5254b..3d55a3a39e82 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -717,6 +717,7 @@ u32 cik_get_xclk(struct radeon_device *rdev); uint32_t cik_pciep_rreg(struct radeon_device *rdev, uint32_t reg); void cik_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk); +int cik_set_vce_clocks(struct radeon_device *rdev, u32 evclk, u32 ecclk); void cik_sdma_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); bool cik_sdma_semaphore_ring_emit(struct radeon_device *rdev, From 8cd366823e0045bfd450138204c7559ac06efcea Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 23 Aug 2013 11:05:24 -0400 Subject: [PATCH 029/259] drm/radeon: add vce dpm support for CI Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/ci_dpm.c | 50 ++++++++++++++++++++++++++------- 1 file changed, 40 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 4a0c40186046..c91d0ee9dfdc 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -746,6 +746,14 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev, u32 max_sclk_vddc, max_mclk_vddci, max_mclk_vddc; int i; + if (rps->vce_active) { + rps->evclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].evclk; + rps->ecclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].ecclk; + } else { + rps->evclk = 0; + rps->ecclk = 0; + } + if ((rdev->pm.dpm.new_active_crtc_count > 1) || ci_dpm_vblank_too_short(rdev)) disable_mclk_switching = true; @@ -804,6 +812,13 @@ static void ci_apply_state_adjust_rules(struct radeon_device *rdev, sclk = ps->performance_levels[0].sclk; } + if (rps->vce_active) { + if (sclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk) + sclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk; + if (mclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].mclk) + mclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].mclk; + } + ps->performance_levels[0].sclk = sclk; ps->performance_levels[0].mclk = mclk; @@ -3468,7 +3483,6 @@ static int ci_enable_uvd_dpm(struct radeon_device *rdev, bool enable) 0 : -EINVAL; } -#if 0 static int ci_enable_vce_dpm(struct radeon_device *rdev, bool enable) { struct ci_power_info *pi = ci_get_pi(rdev); @@ -3501,6 +3515,7 @@ static int ci_enable_vce_dpm(struct radeon_device *rdev, bool enable) 0 : -EINVAL; } +#if 0 static int ci_enable_samu_dpm(struct radeon_device *rdev, bool enable) { struct ci_power_info *pi = ci_get_pi(rdev); @@ -3587,7 +3602,6 @@ static int ci_update_uvd_dpm(struct radeon_device *rdev, bool gate) return ci_enable_uvd_dpm(rdev, !gate); } -#if 0 static u8 ci_get_vce_boot_level(struct radeon_device *rdev) { u8 i; @@ -3608,13 +3622,11 @@ static int ci_update_vce_dpm(struct radeon_device *rdev, struct radeon_ps *radeon_current_state) { struct ci_power_info *pi = ci_get_pi(rdev); - bool new_vce_clock_non_zero = (radeon_new_state->evclk != 0); - bool old_vce_clock_non_zero = (radeon_current_state->evclk != 0); int ret = 0; u32 tmp; - if (new_vce_clock_non_zero != old_vce_clock_non_zero) { - if (new_vce_clock_non_zero) { + if (radeon_current_state->evclk != radeon_new_state->evclk) { + if (radeon_new_state->evclk) { pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(rdev); tmp = RREG32_SMC(DPM_TABLE_475); @@ -3630,6 +3642,7 @@ static int ci_update_vce_dpm(struct radeon_device *rdev, return ret; } +#if 0 static int ci_update_samu_dpm(struct radeon_device *rdev, bool gate) { return ci_enable_samu_dpm(rdev, gate); @@ -4752,13 +4765,13 @@ int ci_dpm_set_power_state(struct radeon_device *rdev) DRM_ERROR("ci_generate_dpm_level_enable_mask failed\n"); return ret; } -#if 0 + ret = ci_update_vce_dpm(rdev, new_ps, old_ps); if (ret) { DRM_ERROR("ci_update_vce_dpm failed\n"); return ret; } -#endif + ret = ci_update_sclk_t(rdev); if (ret) { DRM_ERROR("ci_update_sclk_t failed\n"); @@ -4995,6 +5008,21 @@ static int ci_parse_power_table(struct radeon_device *rdev) power_state_offset += 2 + power_state->v2.ucNumDPMLevels; } rdev->pm.dpm.num_ps = state_array->ucNumEntries; + + /* fill in the vce power states */ + for (i = 0; i < RADEON_MAX_VCE_LEVELS; i++) { + u32 sclk, mclk; + clock_array_index = rdev->pm.dpm.vce_states[i].clk_idx; + clock_info = (union pplib_clock_info *) + &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + sclk = le16_to_cpu(clock_info->ci.usEngineClockLow); + sclk |= clock_info->ci.ucEngineClockHigh << 16; + mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow); + mclk |= clock_info->ci.ucMemoryClockHigh << 16; + rdev->pm.dpm.vce_states[i].sclk = sclk; + rdev->pm.dpm.vce_states[i].mclk = mclk; + } + return 0; } @@ -5080,12 +5108,14 @@ int ci_dpm_init(struct radeon_device *rdev) ci_dpm_fini(rdev); return ret; } - ret = ci_parse_power_table(rdev); + + ret = r600_parse_extended_power_table(rdev); if (ret) { ci_dpm_fini(rdev); return ret; } - ret = r600_parse_extended_power_table(rdev); + + ret = ci_parse_power_table(rdev); if (ret) { ci_dpm_fini(rdev); return ret; From ee35b0024a9d85f9c8745e0481c09d65f2507bd3 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 23 Aug 2013 11:09:21 -0400 Subject: [PATCH 030/259] drm/radeon: enable vce dpm on CI VCE dpm dynamically adjusts the uvd clocks on demand. Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/ci_dpm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index c91d0ee9dfdc..6669d3252f57 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -5153,6 +5153,7 @@ int ci_dpm_init(struct radeon_device *rdev) pi->caps_sclk_throttle_low_notification = false; pi->caps_uvd_dpm = true; + pi->caps_vce_dpm = true; ci_get_leakage_voltages(rdev); ci_patch_dependency_tables_with_leakage(rdev); From 4233290519c779e44a01816cf825f6df067a0886 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 4 Sep 2013 16:17:07 -0400 Subject: [PATCH 031/259] drm/radeon: add vce dpm support for KV/KB TODO: plug in cik_vce_suspend()/resume() so we can enable vce powergating. See XXX in code. Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/kv_dpm.c | 46 +++++++++++++++++++++++++-------- 1 file changed, 35 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c index b5bb3a5654ce..e972b885e297 100644 --- a/drivers/gpu/drm/radeon/kv_dpm.c +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -1338,13 +1338,11 @@ static int kv_enable_uvd_dpm(struct radeon_device *rdev, bool enable) PPSMC_MSG_UVDDPM_Enable : PPSMC_MSG_UVDDPM_Disable); } -#if 0 static int kv_enable_vce_dpm(struct radeon_device *rdev, bool enable) { return kv_notify_message_to_smu(rdev, enable ? PPSMC_MSG_VCEDPM_Enable : PPSMC_MSG_VCEDPM_Disable); } -#endif static int kv_enable_samu_dpm(struct radeon_device *rdev, bool enable) { @@ -1389,7 +1387,6 @@ static int kv_update_uvd_dpm(struct radeon_device *rdev, bool gate) return kv_enable_uvd_dpm(rdev, !gate); } -#if 0 static u8 kv_get_vce_boot_level(struct radeon_device *rdev) { u8 i; @@ -1414,6 +1411,8 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, int ret; if (radeon_new_state->evclk > 0 && radeon_current_state->evclk == 0) { + kv_dpm_powergate_vce(rdev, false); + /* XXX cik_vce_resume(); */ if (pi->caps_stable_p_state) pi->vce_boot_level = table->count - 1; else @@ -1436,11 +1435,12 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, kv_enable_vce_dpm(rdev, true); } else if (radeon_new_state->evclk == 0 && radeon_current_state->evclk > 0) { kv_enable_vce_dpm(rdev, false); + /* XXX cik_vce_suspend(); */ + kv_dpm_powergate_vce(rdev, true); } return 0; } -#endif static int kv_update_samu_dpm(struct radeon_device *rdev, bool gate) { @@ -1768,7 +1768,7 @@ int kv_dpm_set_power_state(struct radeon_device *rdev) { struct kv_power_info *pi = kv_get_pi(rdev); struct radeon_ps *new_ps = &pi->requested_rps; - /*struct radeon_ps *old_ps = &pi->current_rps;*/ + struct radeon_ps *old_ps = &pi->current_rps; int ret; if (pi->bapm_enable) { @@ -1798,13 +1798,12 @@ int kv_dpm_set_power_state(struct radeon_device *rdev) kv_set_enabled_levels(rdev); kv_force_lowest_valid(rdev); kv_unforce_levels(rdev); -#if 0 + ret = kv_update_vce_dpm(rdev, new_ps, old_ps); if (ret) { DRM_ERROR("kv_update_vce_dpm failed\n"); return ret; } -#endif kv_update_sclk_t(rdev); } } else { @@ -1823,13 +1822,11 @@ int kv_dpm_set_power_state(struct radeon_device *rdev) kv_program_nbps_index_settings(rdev, new_ps); kv_freeze_sclk_dpm(rdev, false); kv_set_enabled_levels(rdev); -#if 0 ret = kv_update_vce_dpm(rdev, new_ps, old_ps); if (ret) { DRM_ERROR("kv_update_vce_dpm failed\n"); return ret; } -#endif kv_update_acp_boot_level(rdev); kv_update_sclk_t(rdev); kv_enable_nb_dpm(rdev); @@ -2037,6 +2034,14 @@ static void kv_apply_state_adjust_rules(struct radeon_device *rdev, struct radeon_clock_and_voltage_limits *max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac; + if (new_rps->vce_active) { + new_rps->evclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].evclk; + new_rps->ecclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].ecclk; + } else { + new_rps->evclk = 0; + new_rps->ecclk = 0; + } + mclk = max_limits->mclk; sclk = min_sclk; @@ -2056,6 +2061,11 @@ static void kv_apply_state_adjust_rules(struct radeon_device *rdev, sclk = stable_p_state_sclk; } + if (new_rps->vce_active) { + if (sclk < rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk) + sclk = rdev->pm.dpm.vce_states[rdev->pm.dpm.vce_level].sclk; + } + ps->need_dfs_bypass = true; for (i = 0; i < ps->num_levels; i++) { @@ -2092,7 +2102,8 @@ static void kv_apply_state_adjust_rules(struct radeon_device *rdev, } } - pi->video_start = new_rps->dclk || new_rps->vclk; + pi->video_start = new_rps->dclk || new_rps->vclk || + new_rps->evclk || new_rps->ecclk; if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) @@ -2574,6 +2585,19 @@ static int kv_parse_power_table(struct radeon_device *rdev) power_state_offset += 2 + power_state->v2.ucNumDPMLevels; } rdev->pm.dpm.num_ps = state_array->ucNumEntries; + + /* fill in the vce power states */ + for (i = 0; i < RADEON_MAX_VCE_LEVELS; i++) { + u32 sclk; + clock_array_index = rdev->pm.dpm.vce_states[i].clk_idx; + clock_info = (union pplib_clock_info *) + &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; + sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow); + sclk |= clock_info->sumo.ucEngineClockHigh << 16; + rdev->pm.dpm.vce_states[i].sclk = sclk; + rdev->pm.dpm.vce_states[i].mclk = 0; + } + return 0; } @@ -2624,7 +2648,7 @@ int kv_dpm_init(struct radeon_device *rdev) pi->caps_fps = false; /* true? */ pi->caps_uvd_pg = true; pi->caps_uvd_dpm = true; - pi->caps_vce_pg = false; + pi->caps_vce_pg = false; /* XXX true */ pi->caps_samu_pg = false; pi->caps_acp_pg = false; pi->caps_stable_p_state = false; From 03afe6f6480f2544d6cd18866556f1f76bb05f14 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 23 Aug 2013 11:56:26 -0400 Subject: [PATCH 032/259] drm/radeon/dpm: enable dynamic vce state switching v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit enable vce states when vce is active. When vce is active, it adjusts the currently selected state (performance, battery, uvd, etc.) v2: add code comments Signed-off-by: Alex Deucher Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 3 ++ drivers/gpu/drm/radeon/radeon_cs.c | 3 ++ drivers/gpu/drm/radeon/radeon_pm.c | 17 ++++++++ drivers/gpu/drm/radeon/radeon_vce.c | 62 +++++++++++++++++++++++++++++ 4 files changed, 85 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 693a8fccd94d..540624e7491c 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1518,6 +1518,7 @@ struct radeon_dpm { }; void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable); +void radeon_dpm_enable_vce(struct radeon_device *rdev, bool enable); struct radeon_pm { struct mutex mutex; @@ -1638,6 +1639,7 @@ struct radeon_vce { unsigned fb_version; atomic_t handles[RADEON_MAX_VCE_HANDLES]; struct drm_file *filp[RADEON_MAX_VCE_HANDLES]; + struct delayed_work idle_work; }; int radeon_vce_init(struct radeon_device *rdev); @@ -1649,6 +1651,7 @@ int radeon_vce_get_create_msg(struct radeon_device *rdev, int ring, int radeon_vce_get_destroy_msg(struct radeon_device *rdev, int ring, uint32_t handle, struct radeon_fence **fence); void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp); +void radeon_vce_note_usage(struct radeon_device *rdev); int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi); int radeon_vce_cs_parse(struct radeon_cs_parser *p); bool radeon_vce_semaphore_emit(struct radeon_device *rdev, diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 701ee7981b5c..f28a8d82fa19 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -347,6 +347,9 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev, if (parser->ring == R600_RING_TYPE_UVD_INDEX) radeon_uvd_note_usage(rdev); + else if ((parser->ring == TN_RING_TYPE_VCE1_INDEX) || + (parser->ring == TN_RING_TYPE_VCE2_INDEX)) + radeon_vce_note_usage(rdev); radeon_cs_sync_rings(parser); r = radeon_ib_schedule(rdev, &parser->ib, NULL); diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index a4687e7b45f8..4ad9af9fc517 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -968,6 +968,23 @@ void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable) } } +void radeon_dpm_enable_vce(struct radeon_device *rdev, bool enable) +{ + if (enable) { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.vce_active = true; + /* XXX select vce level based on ring/task */ + rdev->pm.dpm.vce_level = RADEON_VCE_LEVEL_AC_ALL; + mutex_unlock(&rdev->pm.mutex); + } else { + mutex_lock(&rdev->pm.mutex); + rdev->pm.dpm.vce_active = false; + mutex_unlock(&rdev->pm.mutex); + } + + radeon_pm_compute_clocks(rdev); +} + static void radeon_pm_suspend_old(struct radeon_device *rdev) { mutex_lock(&rdev->pm.mutex); diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index f46563b60921..d130432e313a 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -34,11 +34,16 @@ #include "radeon_asic.h" #include "sid.h" +/* 1 second timeout */ +#define VCE_IDLE_TIMEOUT_MS 1000 + /* Firmware Names */ #define FIRMWARE_BONAIRE "radeon/BONAIRE_vce.bin" MODULE_FIRMWARE(FIRMWARE_BONAIRE); +static void radeon_vce_idle_work_handler(struct work_struct *work); + /** * radeon_vce_init - allocate memory, load vce firmware * @@ -55,6 +60,8 @@ int radeon_vce_init(struct radeon_device *rdev) uint8_t start, mid, end; int i, r; + INIT_DELAYED_WORK(&rdev->vce.idle_work, radeon_vce_idle_work_handler); + switch (rdev->family) { case CHIP_BONAIRE: case CHIP_KAVERI: @@ -219,6 +226,59 @@ int radeon_vce_resume(struct radeon_device *rdev) return 0; } +/** + * radeon_vce_idle_work_handler - power off VCE + * + * @work: pointer to work structure + * + * power of VCE when it's not used any more + */ +static void radeon_vce_idle_work_handler(struct work_struct *work) +{ + struct radeon_device *rdev = + container_of(work, struct radeon_device, vce.idle_work.work); + + if ((radeon_fence_count_emitted(rdev, TN_RING_TYPE_VCE1_INDEX) == 0) && + (radeon_fence_count_emitted(rdev, TN_RING_TYPE_VCE2_INDEX) == 0)) { + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + radeon_dpm_enable_vce(rdev, false); + } else { + radeon_set_vce_clocks(rdev, 0, 0); + } + } else { + schedule_delayed_work(&rdev->vce.idle_work, + msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); + } +} + +/** + * radeon_vce_note_usage - power up VCE + * + * @rdev: radeon_device pointer + * + * Make sure VCE is powerd up when we want to use it + */ +void radeon_vce_note_usage(struct radeon_device *rdev) +{ + bool streams_changed = false; + bool set_clocks = !cancel_delayed_work_sync(&rdev->vce.idle_work); + set_clocks &= schedule_delayed_work(&rdev->vce.idle_work, + msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); + + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + /* XXX figure out if the streams changed */ + streams_changed = false; + } + + if (set_clocks || streams_changed) { + if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) { + radeon_dpm_enable_vce(rdev, true); + } else { + radeon_set_vce_clocks(rdev, 53300, 40000); + } + } +} + /** * radeon_vce_free_handles - free still open VCE handles * @@ -235,6 +295,8 @@ void radeon_vce_free_handles(struct radeon_device *rdev, struct drm_file *filp) if (!handle || rdev->vce.filp[i] != filp) continue; + radeon_vce_note_usage(rdev); + r = radeon_vce_get_destroy_msg(rdev, TN_RING_TYPE_VCE1_INDEX, handle, NULL); if (r) From 44493ba959cfaa7506498441397f83d180e4a509 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 28 Aug 2013 18:53:50 -0400 Subject: [PATCH 033/259] drm/radeon/dpm: properly enable/disable vce when vce pg is enabled The adds the appropriate function calls to properly re-init vce before it's used after it has been power gated. Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/kv_dpm.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c index e972b885e297..9ee1f28bbd85 100644 --- a/drivers/gpu/drm/radeon/kv_dpm.c +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -1412,7 +1412,6 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, if (radeon_new_state->evclk > 0 && radeon_current_state->evclk == 0) { kv_dpm_powergate_vce(rdev, false); - /* XXX cik_vce_resume(); */ if (pi->caps_stable_p_state) pi->vce_boot_level = table->count - 1; else @@ -1435,7 +1434,6 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, kv_enable_vce_dpm(rdev, true); } else if (radeon_new_state->evclk == 0 && radeon_current_state->evclk > 0) { kv_enable_vce_dpm(rdev, false); - /* XXX cik_vce_suspend(); */ kv_dpm_powergate_vce(rdev, true); } @@ -1575,11 +1573,16 @@ static void kv_dpm_powergate_vce(struct radeon_device *rdev, bool gate) pi->vce_power_gated = gate; if (gate) { - if (pi->caps_vce_pg) + if (pi->caps_vce_pg) { + /* XXX do we need a vce_v1_0_stop() ? */ kv_notify_message_to_smu(rdev, PPSMC_MSG_VCEPowerOFF); + } } else { - if (pi->caps_vce_pg) + if (pi->caps_vce_pg) { kv_notify_message_to_smu(rdev, PPSMC_MSG_VCEPowerON); + vce_v2_0_resume(rdev); + vce_v1_0_start(rdev); + } } } From b9fa18837610483b09a07f1419e6b9f333c46023 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 5 Sep 2013 15:14:28 -0400 Subject: [PATCH 034/259] drm/radeon: add support for vce 2.0 clock gating Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/cikd.h | 10 +++ drivers/gpu/drm/radeon/vce_v2_0.c | 111 ++++++++++++++++++++++++++++++ 2 files changed, 121 insertions(+) diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h index ee16380ceba8..213873270d5f 100644 --- a/drivers/gpu/drm/radeon/cikd.h +++ b/drivers/gpu/drm/radeon/cikd.h @@ -2029,8 +2029,18 @@ #define VCE_RB_RPTR 0x2018c #define VCE_RB_WPTR 0x20190 #define VCE_CLOCK_GATING_A 0x202f8 +# define CGC_CLK_GATE_DLY_TIMER_MASK (0xf << 0) +# define CGC_CLK_GATE_DLY_TIMER(x) ((x) << 0) +# define CGC_CLK_GATER_OFF_DLY_TIMER_MASK (0xff << 4) +# define CGC_CLK_GATER_OFF_DLY_TIMER(x) ((x) << 4) +# define CGC_UENC_WAIT_AWAKE (1 << 18) #define VCE_CLOCK_GATING_B 0x202fc +#define VCE_CGTT_CLK_OVERRIDE 0x207a0 #define VCE_UENC_CLOCK_GATING 0x207bc +# define CLOCK_ON_DELAY_MASK (0xf << 0) +# define CLOCK_ON_DELAY(x) ((x) << 0) +# define CLOCK_OFF_DELAY_MASK (0xff << 4) +# define CLOCK_OFF_DELAY(x) ((x) << 4) #define VCE_UENC_REG_CLOCK_GATING 0x207c0 #define VCE_SYS_INT_EN 0x21300 # define VCE_SYS_INT_TRAP_INTERRUPT_EN (1 << 3) diff --git a/drivers/gpu/drm/radeon/vce_v2_0.c b/drivers/gpu/drm/radeon/vce_v2_0.c index 4911d1b00e3b..1ac7bb825a1b 100644 --- a/drivers/gpu/drm/radeon/vce_v2_0.c +++ b/drivers/gpu/drm/radeon/vce_v2_0.c @@ -31,6 +31,115 @@ #include "radeon_asic.h" #include "cikd.h" +static void vce_v2_0_set_sw_cg(struct radeon_device *rdev, bool gated) +{ + u32 tmp; + + if (gated) { + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0xe70000; + WREG32(VCE_CLOCK_GATING_B, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp |= 0xff000000; + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp &= ~0x3fc; + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + + WREG32(VCE_CGTT_CLK_OVERRIDE, 0); + } else { + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0xe7; + tmp &= ~0xe70000; + WREG32(VCE_CLOCK_GATING_B, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp |= 0x1fe000; + tmp &= ~0xff000000; + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp |= 0x3fc; + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + } +} + +static void vce_v2_0_set_dyn_cg(struct radeon_device *rdev, bool gated) +{ + u32 orig, tmp; + + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp &= ~0x00060006; + if (gated) { + tmp |= 0xe10000; + } else { + tmp |= 0xe1; + tmp &= ~0xe10000; + } + WREG32(VCE_CLOCK_GATING_B, tmp); + + orig = tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp &= ~0x1fe000; + tmp &= ~0xff000000; + if (tmp != orig) + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + orig = tmp = RREG32(VCE_UENC_REG_CLOCK_GATING); + tmp &= ~0x3fc; + if (tmp != orig) + WREG32(VCE_UENC_REG_CLOCK_GATING, tmp); + + if (gated) + WREG32(VCE_CGTT_CLK_OVERRIDE, 0); +} + +static void vce_v2_0_disable_cg(struct radeon_device *rdev) +{ + WREG32(VCE_CGTT_CLK_OVERRIDE, 7); +} + +void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable) +{ + bool sw_cg = false; + + if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_VCE_MGCG)) { + if (sw_cg) + vce_v2_0_set_sw_cg(rdev, true); + else + vce_v2_0_set_dyn_cg(rdev, true); + } else { + vce_v2_0_disable_cg(rdev); + + if (sw_cg) + vce_v2_0_set_sw_cg(rdev, false); + else + vce_v2_0_set_dyn_cg(rdev, false); + } +} + +static void vce_v2_0_init_cg(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG32(VCE_CLOCK_GATING_A); + tmp &= ~(CGC_CLK_GATE_DLY_TIMER_MASK | CGC_CLK_GATER_OFF_DLY_TIMER_MASK); + tmp |= (CGC_CLK_GATE_DLY_TIMER(0) | CGC_CLK_GATER_OFF_DLY_TIMER(4)); + tmp |= CGC_UENC_WAIT_AWAKE; + WREG32(VCE_CLOCK_GATING_A, tmp); + + tmp = RREG32(VCE_UENC_CLOCK_GATING); + tmp &= ~(CLOCK_ON_DELAY_MASK | CLOCK_OFF_DELAY_MASK); + tmp |= (CLOCK_ON_DELAY(0) | CLOCK_OFF_DELAY(4)); + WREG32(VCE_UENC_CLOCK_GATING, tmp); + + tmp = RREG32(VCE_CLOCK_GATING_B); + tmp |= 0x10; + tmp &= ~0x100000; + WREG32(VCE_CLOCK_GATING_B, tmp); +} + int vce_v2_0_resume(struct radeon_device *rdev) { uint64_t addr = rdev->vce.gpu_addr; @@ -66,5 +175,7 @@ int vce_v2_0_resume(struct radeon_device *rdev) WREG32_P(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN, ~VCE_SYS_INT_TRAP_INTERRUPT_EN); + vce_v2_0_init_cg(rdev); + return 0; } From a1d6f97c8cfa7c3554d0391c0b16505d1d97f380 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 6 Sep 2013 12:33:04 -0400 Subject: [PATCH 035/259] drm/radeon/cik: enable/disable vce cg when encoding v2 Some of the vce clocks are automatic, others need to be manually enabled. For ease, just disable cg when vce is active. v2: rebased Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/ci_dpm.c | 9 ++++++++- drivers/gpu/drm/radeon/cik.c | 5 +++++ drivers/gpu/drm/radeon/kv_dpm.c | 4 ++++ 3 files changed, 17 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 6669d3252f57..cad89a977527 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -172,6 +172,8 @@ extern void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev, extern void cik_enter_rlc_safe_mode(struct radeon_device *rdev); extern void cik_exit_rlc_safe_mode(struct radeon_device *rdev); extern int ci_mc_load_microcode(struct radeon_device *rdev); +extern void cik_update_cg(struct radeon_device *rdev, + u32 block, bool enable); static int ci_get_std_voltage_value_sidd(struct radeon_device *rdev, struct atom_voltage_table_entry *voltage_table, @@ -3627,8 +3629,10 @@ static int ci_update_vce_dpm(struct radeon_device *rdev, if (radeon_current_state->evclk != radeon_new_state->evclk) { if (radeon_new_state->evclk) { - pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(rdev); + /* turn the clocks on when encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, false); + pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(rdev); tmp = RREG32_SMC(DPM_TABLE_475); tmp &= ~VceBootLevel_MASK; tmp |= VceBootLevel(pi->smc_state_table.VceBootLevel); @@ -3636,6 +3640,9 @@ static int ci_update_vce_dpm(struct radeon_device *rdev, ret = ci_enable_vce_dpm(rdev, true); } else { + /* turn the clocks off when not encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, true); + ret = ci_enable_vce_dpm(rdev, false); } } diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index ecb16b14f049..2b31c3233a5e 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -75,6 +75,7 @@ extern void si_init_uvd_internal_cg(struct radeon_device *rdev); extern int cik_sdma_resume(struct radeon_device *rdev); extern void cik_sdma_enable(struct radeon_device *rdev, bool enable); extern void cik_sdma_fini(struct radeon_device *rdev); +extern void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable); static void cik_rlc_stop(struct radeon_device *rdev); static void cik_pcie_gen3_enable(struct radeon_device *rdev); static void cik_program_aspm(struct radeon_device *rdev); @@ -6141,6 +6142,10 @@ void cik_update_cg(struct radeon_device *rdev, cik_enable_hdp_mgcg(rdev, enable); cik_enable_hdp_ls(rdev, enable); } + + if (block & RADEON_CG_BLOCK_VCE) { + vce_v2_0_enable_mgcg(rdev, enable); + } } static void cik_init_cg(struct radeon_device *rdev) diff --git a/drivers/gpu/drm/radeon/kv_dpm.c b/drivers/gpu/drm/radeon/kv_dpm.c index 9ee1f28bbd85..16ec9d56a234 100644 --- a/drivers/gpu/drm/radeon/kv_dpm.c +++ b/drivers/gpu/drm/radeon/kv_dpm.c @@ -1412,6 +1412,8 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, if (radeon_new_state->evclk > 0 && radeon_current_state->evclk == 0) { kv_dpm_powergate_vce(rdev, false); + /* turn the clocks on when encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, false); if (pi->caps_stable_p_state) pi->vce_boot_level = table->count - 1; else @@ -1434,6 +1436,8 @@ static int kv_update_vce_dpm(struct radeon_device *rdev, kv_enable_vce_dpm(rdev, true); } else if (radeon_new_state->evclk == 0 && radeon_current_state->evclk > 0) { kv_enable_vce_dpm(rdev, false); + /* turn the clocks off when not encoding */ + cik_update_cg(rdev, RADEON_CG_BLOCK_VCE, true); kv_dpm_powergate_vce(rdev, true); } From ff212f25feb44a915ce9c0144faef7fae27a6e61 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 14:52:33 +0100 Subject: [PATCH 036/259] drm/radeon: drop drivers copy of the rptr MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In all cases where it really matters we are using the read functions anyway. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/cik.c | 7 ++----- drivers/gpu/drm/radeon/cik_sdma.c | 4 +--- drivers/gpu/drm/radeon/evergreen.c | 4 +--- drivers/gpu/drm/radeon/evergreen_dma.c | 2 +- drivers/gpu/drm/radeon/ni.c | 6 +++--- drivers/gpu/drm/radeon/ni_dma.c | 4 +--- drivers/gpu/drm/radeon/r100.c | 3 +-- drivers/gpu/drm/radeon/r600.c | 4 +--- drivers/gpu/drm/radeon/r600_dma.c | 4 +--- drivers/gpu/drm/radeon/radeon.h | 4 ++-- drivers/gpu/drm/radeon/radeon_ring.c | 27 +++++++++++++------------- drivers/gpu/drm/radeon/si.c | 8 +------- drivers/gpu/drm/radeon/si_dma.c | 2 +- drivers/gpu/drm/radeon/uvd_v1_0.c | 2 +- drivers/gpu/drm/radeon/vce_v1_0.c | 4 ++-- 15 files changed, 32 insertions(+), 53 deletions(-) diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 2b31c3233a5e..835dcfb78916 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -4031,8 +4031,6 @@ static int cik_cp_gfx_resume(struct radeon_device *rdev) WREG32(CP_RB0_BASE, rb_addr); WREG32(CP_RB0_BASE_HI, upper_32_bits(rb_addr)); - ring->rptr = RREG32(CP_RB0_RPTR); - /* start the ring */ cik_cp_gfx_start(rdev); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; @@ -4587,8 +4585,7 @@ static int cik_cp_compute_resume(struct radeon_device *rdev) rdev->ring[idx].wptr = 0; mqd->queue_state.cp_hqd_pq_wptr = rdev->ring[idx].wptr; WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr); - rdev->ring[idx].rptr = RREG32(CP_HQD_PQ_RPTR); - mqd->queue_state.cp_hqd_pq_rptr = rdev->ring[idx].rptr; + mqd->queue_state.cp_hqd_pq_rptr = RREG32(CP_HQD_PQ_RPTR); /* set the vmid for the queue */ mqd->queue_state.cp_hqd_vmid = 0; @@ -5118,7 +5115,7 @@ bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index 1ecb3f1070e3..e474760d714c 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -362,8 +362,6 @@ static int cik_sdma_gfx_resume(struct radeon_device *rdev) ring->wptr = 0; WREG32(SDMA0_GFX_RB_WPTR + reg_offset, ring->wptr << 2); - ring->rptr = RREG32(SDMA0_GFX_RB_RPTR + reg_offset) >> 2; - /* enable DMA RB */ WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl | SDMA_RB_ENABLE); @@ -713,7 +711,7 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) mask = RADEON_RESET_DMA1; if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force ring activities */ diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index f2b9e21ce4da..d9156be5b9a6 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -2990,8 +2990,6 @@ static int evergreen_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - ring->rptr = RREG32(CP_RB_RPTR); - evergreen_cp_start(rdev); ring->ready = true; r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); @@ -3952,7 +3950,7 @@ bool evergreen_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c index a37b54436382..d448961e9ab2 100644 --- a/drivers/gpu/drm/radeon/evergreen_dma.c +++ b/drivers/gpu/drm/radeon/evergreen_dma.c @@ -174,7 +174,7 @@ bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin u32 reset_mask = evergreen_gpu_check_soft_reset(rdev); if (!(reset_mask & RADEON_RESET_DMA)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force ring activities */ diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c index ea932ac66fc6..7601532b7372 100644 --- a/drivers/gpu/drm/radeon/ni.c +++ b/drivers/gpu/drm/radeon/ni.c @@ -1642,8 +1642,8 @@ static int cayman_cp_resume(struct radeon_device *rdev) ring = &rdev->ring[ridx[i]]; WREG32_P(cp_rb_cntl[i], RB_RPTR_WR_ENA, ~RB_RPTR_WR_ENA); - ring->rptr = ring->wptr = 0; - WREG32(cp_rb_rptr[i], ring->rptr); + ring->wptr = 0; + WREG32(cp_rb_rptr[i], 0); WREG32(cp_rb_wptr[i], ring->wptr); mdelay(1); @@ -1917,7 +1917,7 @@ bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c index 7cf96b15377f..95e533c61f83 100644 --- a/drivers/gpu/drm/radeon/ni_dma.c +++ b/drivers/gpu/drm/radeon/ni_dma.c @@ -248,8 +248,6 @@ int cayman_dma_resume(struct radeon_device *rdev) ring->wptr = 0; WREG32(DMA_RB_WPTR + reg_offset, ring->wptr << 2); - ring->rptr = RREG32(DMA_RB_RPTR + reg_offset) >> 2; - WREG32(DMA_RB_CNTL + reg_offset, rb_cntl | DMA_RB_ENABLE); ring->ready = true; @@ -302,7 +300,7 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) mask = RADEON_RESET_DMA1; if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force ring activities */ diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index ef024ce3f7cc..3a7438163d06 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -1193,7 +1193,6 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(RADEON_CP_RB_CNTL, tmp); udelay(10); - ring->rptr = RREG32(RADEON_CP_RB_RPTR); /* Set cp mode to bus mastering & enable cp*/ WREG32(RADEON_CP_CSQ_MODE, REG_SET(RADEON_INDIRECT2_START, indirect2_start) | @@ -2523,7 +2522,7 @@ bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) rbbm_status = RREG32(R_000E40_RBBM_STATUS); if (!G_000E40_GUI_ACTIVE(rbbm_status)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index cdbc4171fe73..085e02590dcf 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -1748,7 +1748,7 @@ bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ @@ -2604,8 +2604,6 @@ int r600_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - ring->rptr = RREG32(CP_RB_RPTR); - r600_cp_start(rdev); ring->ready = true; r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c index b2d4c91e6272..6944e1988426 100644 --- a/drivers/gpu/drm/radeon/r600_dma.c +++ b/drivers/gpu/drm/radeon/r600_dma.c @@ -176,8 +176,6 @@ int r600_dma_resume(struct radeon_device *rdev) ring->wptr = 0; WREG32(DMA_RB_WPTR, ring->wptr << 2); - ring->rptr = RREG32(DMA_RB_RPTR) >> 2; - WREG32(DMA_RB_CNTL, rb_cntl | DMA_RB_ENABLE); ring->ready = true; @@ -221,7 +219,7 @@ bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) u32 reset_mask = r600_gpu_check_soft_reset(rdev); if (!(reset_mask & RADEON_RESET_DMA)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force ring activities */ diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 540624e7491c..e1c4f9c6772b 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -793,7 +793,6 @@ struct radeon_ib { struct radeon_ring { struct radeon_bo *ring_obj; volatile uint32_t *ring; - unsigned rptr; unsigned rptr_offs; unsigned rptr_save_reg; u64 next_rptr_gpu_addr; @@ -958,7 +957,8 @@ void radeon_ring_undo(struct radeon_ring *ring); void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp); int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring *ring); -void radeon_ring_lockup_update(struct radeon_ring *ring); +void radeon_ring_lockup_update(struct radeon_device *rdev, + struct radeon_ring *ring); bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring); unsigned radeon_ring_backup(struct radeon_device *rdev, struct radeon_ring *ring, uint32_t **data); diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index d2980b03d1ad..0f78789d085a 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -342,9 +342,10 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev, */ void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring) { - ring->rptr = radeon_ring_get_rptr(rdev, ring); + uint32_t rptr = radeon_ring_get_rptr(rdev, ring); + /* This works because ring_size is a power of 2 */ - ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4)); + ring->ring_free_dw = rptr + (ring->ring_size / 4); ring->ring_free_dw -= ring->wptr; ring->ring_free_dw &= ring->ptr_mask; if (!ring->ring_free_dw) { @@ -376,7 +377,7 @@ int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsi /* This is an empty ring update lockup info to avoid * false positive. */ - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); } ndw = (ndw + ring->align_mask) & ~ring->align_mask; while (ndw > (ring->ring_free_dw - 1)) { @@ -490,8 +491,7 @@ void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring * { int r; - radeon_ring_free_size(rdev, ring); - if (ring->rptr == ring->wptr) { + if (radeon_ring_get_rptr(rdev, ring) == ring->wptr) { r = radeon_ring_alloc(rdev, ring, 1); if (!r) { radeon_ring_write(ring, ring->nop); @@ -507,9 +507,10 @@ void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring * * * Update the last rptr value and timestamp (all asics). */ -void radeon_ring_lockup_update(struct radeon_ring *ring) +void radeon_ring_lockup_update(struct radeon_device *rdev, + struct radeon_ring *ring) { - ring->last_rptr = ring->rptr; + ring->last_rptr = radeon_ring_get_rptr(rdev, ring); ring->last_activity = jiffies; } @@ -535,18 +536,18 @@ void radeon_ring_lockup_update(struct radeon_ring *ring) **/ bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { + uint32_t rptr = radeon_ring_get_rptr(rdev, ring); unsigned long cjiffies, elapsed; cjiffies = jiffies; if (!time_after(cjiffies, ring->last_activity)) { /* likely a wrap around */ - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } - ring->rptr = radeon_ring_get_rptr(rdev, ring); - if (ring->rptr != ring->last_rptr) { + if (rptr != ring->last_rptr) { /* CP is still working no lockup */ - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } elapsed = jiffies_to_msecs(cjiffies - ring->last_activity); @@ -709,7 +710,7 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsig if (radeon_debugfs_ring_init(rdev, ring)) { DRM_ERROR("Failed to register debugfs file for rings !\n"); } - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return 0; } @@ -780,8 +781,6 @@ static int radeon_debugfs_ring_info(struct seq_file *m, void *data) seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n", ring->wptr, ring->wptr); - seq_printf(m, "driver's copy of the rptr: 0x%08x [%5d]\n", - ring->rptr, ring->rptr); seq_printf(m, "last semaphore signal addr : 0x%016llx\n", ring->last_semaphore_signal_addr); seq_printf(m, "last semaphore wait addr : 0x%016llx\n", diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index 83578324e5d1..b406a48ef202 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -3434,8 +3434,6 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB0_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB0_RPTR); - /* ring1 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; @@ -3460,8 +3458,6 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB1_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB1_RPTR); - /* ring2 - compute only */ /* Set ring buffer size */ ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; @@ -3486,8 +3482,6 @@ static int si_cp_resume(struct radeon_device *rdev) WREG32(CP_RB2_BASE, ring->gpu_addr >> 8); - ring->rptr = RREG32(CP_RB2_RPTR); - /* start the rings */ si_cp_start(rdev); rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; @@ -3872,7 +3866,7 @@ bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) if (!(reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force CP activities */ diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c index 59be2cfcbb47..c75f5337f462 100644 --- a/drivers/gpu/drm/radeon/si_dma.c +++ b/drivers/gpu/drm/radeon/si_dma.c @@ -49,7 +49,7 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) mask = RADEON_RESET_DMA1; if (!(reset_mask & mask)) { - radeon_ring_lockup_update(ring); + radeon_ring_lockup_update(rdev, ring); return false; } /* force ring activities */ diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c index d4a68af1a279..0a243f0e5d68 100644 --- a/drivers/gpu/drm/radeon/uvd_v1_0.c +++ b/drivers/gpu/drm/radeon/uvd_v1_0.c @@ -262,7 +262,7 @@ int uvd_v1_0_start(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(UVD_RBC_RB_RPTR, 0x0); - ring->wptr = ring->rptr = RREG32(UVD_RBC_RB_RPTR); + ring->wptr = RREG32(UVD_RBC_RB_RPTR); WREG32(UVD_RBC_RB_WPTR, ring->wptr); /* set the ring address */ diff --git a/drivers/gpu/drm/radeon/vce_v1_0.c b/drivers/gpu/drm/radeon/vce_v1_0.c index e0c3534356a1..b44d9c842f7b 100644 --- a/drivers/gpu/drm/radeon/vce_v1_0.c +++ b/drivers/gpu/drm/radeon/vce_v1_0.c @@ -98,14 +98,14 @@ int vce_v1_0_start(struct radeon_device *rdev) WREG32_P(VCE_STATUS, 1, ~1); ring = &rdev->ring[TN_RING_TYPE_VCE1_INDEX]; - WREG32(VCE_RB_RPTR, ring->rptr); + WREG32(VCE_RB_RPTR, ring->wptr); WREG32(VCE_RB_WPTR, ring->wptr); WREG32(VCE_RB_BASE_LO, ring->gpu_addr); WREG32(VCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); WREG32(VCE_RB_SIZE, ring->ring_size / 4); ring = &rdev->ring[TN_RING_TYPE_VCE2_INDEX]; - WREG32(VCE_RB_RPTR2, ring->rptr); + WREG32(VCE_RB_RPTR2, ring->wptr); WREG32(VCE_RB_WPTR2, ring->wptr); WREG32(VCE_RB_BASE_LO2, ring->gpu_addr); WREG32(VCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); From 2d2fe3f9b60fd3cc9a19dcc3ae892a23825da07f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 12:37:50 +0100 Subject: [PATCH 037/259] drm/radeon: drop radeon_ring_force_activity MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The reason for the false positives was fixed quite some time ago and since most engines can still execute NOPs while being locked up it leads to false negatives. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/cik.c | 2 -- drivers/gpu/drm/radeon/cik_sdma.c | 2 -- drivers/gpu/drm/radeon/evergreen.c | 2 -- drivers/gpu/drm/radeon/evergreen_dma.c | 2 -- drivers/gpu/drm/radeon/ni.c | 2 -- drivers/gpu/drm/radeon/ni_dma.c | 2 -- drivers/gpu/drm/radeon/r100.c | 2 -- drivers/gpu/drm/radeon/r600.c | 2 -- drivers/gpu/drm/radeon/r600_dma.c | 2 -- drivers/gpu/drm/radeon/radeon.h | 1 - drivers/gpu/drm/radeon/radeon_ring.c | 38 +------------------------- drivers/gpu/drm/radeon/si.c | 2 -- drivers/gpu/drm/radeon/si_dma.c | 2 -- 13 files changed, 1 insertion(+), 60 deletions(-) diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 835dcfb78916..92e38b54efb9 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -5118,8 +5118,6 @@ bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index e474760d714c..00150ac49cd2 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -714,8 +714,6 @@ bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index d9156be5b9a6..c78d8ece9504 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -3953,8 +3953,6 @@ bool evergreen_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/evergreen_dma.c b/drivers/gpu/drm/radeon/evergreen_dma.c index d448961e9ab2..287fe966d7de 100644 --- a/drivers/gpu/drm/radeon/evergreen_dma.c +++ b/drivers/gpu/drm/radeon/evergreen_dma.c @@ -177,8 +177,6 @@ bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c index 7601532b7372..85168ecd216b 100644 --- a/drivers/gpu/drm/radeon/ni.c +++ b/drivers/gpu/drm/radeon/ni.c @@ -1920,8 +1920,6 @@ bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c index 95e533c61f83..6378e0276691 100644 --- a/drivers/gpu/drm/radeon/ni_dma.c +++ b/drivers/gpu/drm/radeon/ni_dma.c @@ -303,8 +303,6 @@ bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 3a7438163d06..1690a2dc0721 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -2525,8 +2525,6 @@ bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index 085e02590dcf..0f4ab928a15a 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -1751,8 +1751,6 @@ bool r600_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c index 6944e1988426..53fcb28f5578 100644 --- a/drivers/gpu/drm/radeon/r600_dma.c +++ b/drivers/gpu/drm/radeon/r600_dma.c @@ -222,8 +222,6 @@ bool r600_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index e1c4f9c6772b..a415f8e9d972 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -956,7 +956,6 @@ void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *c void radeon_ring_undo(struct radeon_ring *ring); void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp); int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); -void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring *ring); void radeon_ring_lockup_update(struct radeon_device *rdev, struct radeon_ring *ring); bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring); diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 0f78789d085a..668097abb424 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -478,28 +478,6 @@ void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *rin mutex_unlock(&rdev->ring_lock); } -/** - * radeon_ring_force_activity - add some nop packets to the ring - * - * @rdev: radeon_device pointer - * @ring: radeon_ring structure holding ring information - * - * Add some nop packets to the ring to force activity (all asics). - * Used for lockup detection to see if the rptr is advancing. - */ -void radeon_ring_force_activity(struct radeon_device *rdev, struct radeon_ring *ring) -{ - int r; - - if (radeon_ring_get_rptr(rdev, ring) == ring->wptr) { - r = radeon_ring_alloc(rdev, ring, 1); - if (!r) { - radeon_ring_write(ring, ring->nop); - radeon_ring_commit(rdev, ring); - } - } -} - /** * radeon_ring_lockup_update - update lockup variables * @@ -519,21 +497,7 @@ void radeon_ring_lockup_update(struct radeon_device *rdev, * @rdev: radeon device structure * @ring: radeon_ring structure holding ring information * - * We don't need to initialize the lockup tracking information as we will either - * have CP rptr to a different value of jiffies wrap around which will force - * initialization of the lockup tracking informations. - * - * A possible false positivie is if we get call after while and last_cp_rptr == - * the current CP rptr, even if it's unlikely it might happen. To avoid this - * if the elapsed time since last call is bigger than 2 second than we return - * false and update the tracking information. Due to this the caller must call - * radeon_ring_test_lockup several time in less than 2sec for lockup to be reported - * the fencing code should be cautious about that. - * - * Caller should write to the ring to force CP to do something so we don't get - * false positive when CP is just gived nothing to do. - * - **/ + */ bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t rptr = radeon_ring_get_rptr(rdev, ring); diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index b406a48ef202..8008cb8d5324 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -3869,8 +3869,6 @@ bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force CP activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } diff --git a/drivers/gpu/drm/radeon/si_dma.c b/drivers/gpu/drm/radeon/si_dma.c index c75f5337f462..cf0fdad8c278 100644 --- a/drivers/gpu/drm/radeon/si_dma.c +++ b/drivers/gpu/drm/radeon/si_dma.c @@ -52,8 +52,6 @@ bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) radeon_ring_lockup_update(rdev, ring); return false; } - /* force ring activities */ - radeon_ring_force_activity(rdev, ring); return radeon_ring_test_lockup(rdev, ring); } From 82dc62a31ce3ed7b4eeea9c65a3b69e81e2ea688 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 15:03:22 +0100 Subject: [PATCH 038/259] drm/radeon: cleanup false positive lockup handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Check always when we calculate the free dw, not just the first time. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon_ring.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 668097abb424..b14c86d57607 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -349,7 +349,10 @@ void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring) ring->ring_free_dw -= ring->wptr; ring->ring_free_dw &= ring->ptr_mask; if (!ring->ring_free_dw) { + /* this is an empty ring */ ring->ring_free_dw = ring->ring_size / 4; + /* update lockup info to avoid false positive */ + radeon_ring_lockup_update(rdev, ring); } } @@ -373,12 +376,6 @@ int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsi /* Align requested size with padding so unlock_commit can * pad safely */ radeon_ring_free_size(rdev, ring); - if (ring->ring_free_dw == (ring->ring_size / 4)) { - /* This is an empty ring update lockup info to avoid - * false positive. - */ - radeon_ring_lockup_update(rdev, ring); - } ndw = (ndw + ring->align_mask) & ~ring->align_mask; while (ndw > (ring->ring_free_dw - 1)) { radeon_ring_free_size(rdev, ring); From 6b1168e1617d9d4db73ef5276490627abf5adec4 Mon Sep 17 00:00:00 2001 From: Pavel Shilovsky Date: Fri, 14 Feb 2014 13:31:03 +0400 Subject: [PATCH 039/259] CIFS: Fix wrong pos argument of cifs_find_lock_conflict and use generic_file_aio_write rather than __generic_file_aio_write in cifs_writev. Signed-off-by: Pavel Shilovsky Reported-by: Al Viro Signed-off-by: Steve French --- fs/cifs/file.c | 24 ++++++------------------ 1 file changed, 6 insertions(+), 18 deletions(-) diff --git a/fs/cifs/file.c b/fs/cifs/file.c index 53c15074bb36..834fce759d80 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -2579,31 +2579,19 @@ cifs_writev(struct kiocb *iocb, const struct iovec *iov, struct cifsInodeInfo *cinode = CIFS_I(inode); struct TCP_Server_Info *server = tlink_tcon(cfile->tlink)->ses->server; ssize_t rc = -EACCES; + loff_t lock_pos = pos; - BUG_ON(iocb->ki_pos != pos); - + if (file->f_flags & O_APPEND) + lock_pos = i_size_read(inode); /* * We need to hold the sem to be sure nobody modifies lock list * with a brlock that prevents writing. */ down_read(&cinode->lock_sem); - if (!cifs_find_lock_conflict(cfile, pos, iov_length(iov, nr_segs), + if (!cifs_find_lock_conflict(cfile, lock_pos, iov_length(iov, nr_segs), server->vals->exclusive_lock_type, NULL, - CIFS_WRITE_OP)) { - mutex_lock(&inode->i_mutex); - rc = __generic_file_aio_write(iocb, iov, nr_segs, - &iocb->ki_pos); - mutex_unlock(&inode->i_mutex); - } - - if (rc > 0) { - ssize_t err; - - err = generic_write_sync(file, iocb->ki_pos - rc, rc); - if (err < 0) - rc = err; - } - + CIFS_WRITE_OP)) + rc = generic_file_aio_write(iocb, iov, nr_segs, pos); up_read(&cinode->lock_sem); return rc; } From a26054d184763969a411e3939fe243516715ff59 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 14 Feb 2014 07:21:00 -0500 Subject: [PATCH 040/259] cifs: sanity check length of data to send before sending We had a bug discovered recently where an upper layer function (cifs_iovec_write) could pass down a smb_rqst with an invalid amount of data in it. The length of the SMB frame would be correct, but the rqst struct would cause smb_send_rqst to send nearly 4GB of data. This should never be the case. Add some sanity checking to the beginning of smb_send_rqst that ensures that the amount of data we're going to send agrees with the length in the RFC1002 header. If it doesn't, WARN() and return -EIO to the upper layers. Signed-off-by: Jeff Layton Acked-by: Sachin Prabhu Reviewed-by: Pavel Shilovsky Signed-off-by: Steve French --- fs/cifs/transport.c | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index b37570952846..18cd5650a5fc 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -270,6 +270,26 @@ cifs_rqst_page_to_kvec(struct smb_rqst *rqst, unsigned int idx, iov->iov_len = rqst->rq_pagesz; } +static unsigned long +rqst_len(struct smb_rqst *rqst) +{ + unsigned int i; + struct kvec *iov = rqst->rq_iov; + unsigned long buflen = 0; + + /* total up iov array first */ + for (i = 0; i < rqst->rq_nvec; i++) + buflen += iov[i].iov_len; + + /* add in the page array if there is one */ + if (rqst->rq_npages) { + buflen += rqst->rq_pagesz * (rqst->rq_npages - 1); + buflen += rqst->rq_tailsz; + } + + return buflen; +} + static int smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) { @@ -277,6 +297,7 @@ smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) struct kvec *iov = rqst->rq_iov; int n_vec = rqst->rq_nvec; unsigned int smb_buf_length = get_rfc1002_length(iov[0].iov_base); + unsigned long send_length; unsigned int i; size_t total_len = 0, sent; struct socket *ssocket = server->ssocket; @@ -285,6 +306,14 @@ smb_send_rqst(struct TCP_Server_Info *server, struct smb_rqst *rqst) if (ssocket == NULL) return -ENOTSOCK; + /* sanity check send length */ + send_length = rqst_len(rqst); + if (send_length != smb_buf_length + 4) { + WARN(1, "Send length mismatch(send_length=%lu smb_buf_length=%u)\n", + send_length, smb_buf_length); + return -EIO; + } + cifs_dbg(FYI, "Sending smb: smb_len=%u\n", smb_buf_length); dump_smb(iov[0].iov_base, iov[0].iov_len); From c197db75ff5c1d4f015c7668a3715e230a5d7e27 Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Thu, 28 Nov 2013 11:31:00 +0100 Subject: [PATCH 041/259] drm/dp: Add AUX channel infrastructure This is a superset of the current i2c_dp_aux bus functionality and can be used to transfer native AUX in addition to I2C-over-AUX messages. Helpers are provided to read and write the DPCD, either blockwise or byte-wise. Many of the existing helpers for DisplayPort take a copy of a portion of the DPCD and operate on that, without a way to write data back to the DPCD (e.g. for configuration of the link). Subsequent patches will build upon this infrastructure to provide common functionality in a generic way. Reviewed-by: Alex Deucher Reviewed-by: Jani Nikula Signed-off-by: Thierry Reding --- Changes in v5: - move comments partially to struct drm_dp_aux_msg in header file - return -EPROTO on short reads in DPCD helpers Changes in v4: - fix a typo in a comment Changes in v3: - reorder drm_dp_dpcd_writeb() arguments to be more intuitive - return number of bytes transferred in drm_dp_dpcd_write() - factor out drm_dp_dpcd_access() - describe error codes --- drivers/gpu/drm/drm_dp_helper.c | 110 ++++++++++++++++++++++++++++++++ include/drm/drm_dp_helper.h | 79 +++++++++++++++++++++++ 2 files changed, 189 insertions(+) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 9e978aae8972..da6bcfe178a6 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -346,3 +346,113 @@ int drm_dp_bw_code_to_link_rate(u8 link_bw) } } EXPORT_SYMBOL(drm_dp_bw_code_to_link_rate); + +/** + * DOC: dp helpers + * + * The DisplayPort AUX channel is an abstraction to allow generic, driver- + * independent access to AUX functionality. Drivers can take advantage of + * this by filling in the fields of the drm_dp_aux structure. + * + * Transactions are described using a hardware-independent drm_dp_aux_msg + * structure, which is passed into a driver's .transfer() implementation. + * Both native and I2C-over-AUX transactions are supported. + * + * An AUX channel can also be used to transport I2C messages to a sink. A + * typical application of that is to access an EDID that's present in the + * sink device. The .transfer() function can also be used to execute such + * transactions. The drm_dp_aux_register_i2c_bus() function registers an + * I2C adapter that can be passed to drm_probe_ddc(). Upon removal, drivers + * should call drm_dp_aux_unregister_i2c_bus() to remove the I2C adapter. + */ + +static int drm_dp_dpcd_access(struct drm_dp_aux *aux, u8 request, + unsigned int offset, void *buffer, size_t size) +{ + struct drm_dp_aux_msg msg; + unsigned int retry; + int err; + + memset(&msg, 0, sizeof(msg)); + msg.address = offset; + msg.request = request; + msg.buffer = buffer; + msg.size = size; + + /* + * The specification doesn't give any recommendation on how often to + * retry native transactions, so retry 7 times like for I2C-over-AUX + * transactions. + */ + for (retry = 0; retry < 7; retry++) { + err = aux->transfer(aux, &msg); + if (err < 0) { + if (err == -EBUSY) + continue; + + return err; + } + + if (err < size) + return -EPROTO; + + switch (msg.reply & DP_AUX_NATIVE_REPLY_MASK) { + case DP_AUX_NATIVE_REPLY_ACK: + return err; + + case DP_AUX_NATIVE_REPLY_NACK: + return -EIO; + + case DP_AUX_NATIVE_REPLY_DEFER: + usleep_range(400, 500); + break; + } + } + + DRM_ERROR("too many retries, giving up\n"); + return -EIO; +} + +/** + * drm_dp_dpcd_read() - read a series of bytes from the DPCD + * @aux: DisplayPort AUX channel + * @offset: address of the (first) register to read + * @buffer: buffer to store the register values + * @size: number of bytes in @buffer + * + * Returns the number of bytes transferred on success, or a negative error + * code on failure. -EIO is returned if the request was NAKed by the sink or + * if the retry count was exceeded. If not all bytes were transferred, this + * function returns -EPROTO. Errors from the underlying AUX channel transfer + * function, with the exception of -EBUSY (which causes the transaction to + * be retried), are propagated to the caller. + */ +ssize_t drm_dp_dpcd_read(struct drm_dp_aux *aux, unsigned int offset, + void *buffer, size_t size) +{ + return drm_dp_dpcd_access(aux, DP_AUX_NATIVE_READ, offset, buffer, + size); +} +EXPORT_SYMBOL(drm_dp_dpcd_read); + +/** + * drm_dp_dpcd_write() - write a series of bytes to the DPCD + * @aux: DisplayPort AUX channel + * @offset: address of the (first) register to write + * @buffer: buffer containing the values to write + * @size: number of bytes in @buffer + * + * Returns the number of bytes transferred on success, or a negative error + * code on failure. -EIO is returned if the request was NAKed by the sink or + * if the retry count was exceeded. If not all bytes were transferred, this + * function returns -EPROTO. Errors from the underlying AUX channel transfer + * function, with the exception of -EBUSY (which causes the transaction to + * be retried), are propagated to the caller. + */ +ssize_t drm_dp_dpcd_write(struct drm_dp_aux *aux, unsigned int offset, + void *buffer, size_t size) +{ + return drm_dp_dpcd_access(aux, DP_AUX_NATIVE_WRITE, offset, buffer, + size); +} +EXPORT_SYMBOL(drm_dp_dpcd_write); diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 1d09050a8c00..70f18eb3ca1a 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -398,4 +398,83 @@ drm_dp_enhanced_frame_cap(const u8 dpcd[DP_RECEIVER_CAP_SIZE]) (dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP); } +/* + * DisplayPort AUX channel + */ + +/** + * struct drm_dp_aux_msg - DisplayPort AUX channel transaction + * @address: address of the (first) register to access + * @request: contains the type of transaction (see DP_AUX_* macros) + * @reply: upon completion, contains the reply type of the transaction + * @buffer: pointer to a transmission or reception buffer + * @size: size of @buffer + */ +struct drm_dp_aux_msg { + unsigned int address; + u8 request; + u8 reply; + void *buffer; + size_t size; +}; + +/** + * struct drm_dp_aux - DisplayPort AUX channel + * @dev: pointer to struct device that is the parent for this AUX channel + * @transfer: transfers a message representing a single AUX transaction + * + * The .dev field should be set to a pointer to the device that implements + * the AUX channel. + * + * Drivers provide a hardware-specific implementation of how transactions + * are executed via the .transfer() function. A pointer to a drm_dp_aux_msg + * structure describing the transaction is passed into this function. Upon + * success, the implementation should return the number of payload bytes + * that were transferred, or a negative error-code on failure. Helpers + * propagate errors from the .transfer() function, with the exception of + * the -EBUSY error, which causes a transaction to be retried. On a short, + * helpers will return -EPROTO to make it simpler to check for failure. + */ +struct drm_dp_aux { + struct device *dev; + + ssize_t (*transfer)(struct drm_dp_aux *aux, + struct drm_dp_aux_msg *msg); +}; + +ssize_t drm_dp_dpcd_read(struct drm_dp_aux *aux, unsigned int offset, + void *buffer, size_t size); +ssize_t drm_dp_dpcd_write(struct drm_dp_aux *aux, unsigned int offset, + void *buffer, size_t size); + +/** + * drm_dp_dpcd_readb() - read a single byte from the DPCD + * @aux: DisplayPort AUX channel + * @offset: address of the register to read + * @valuep: location where the value of the register will be stored + * + * Returns the number of bytes transferred (1) on success, or a negative + * error code on failure. + */ +static inline ssize_t drm_dp_dpcd_readb(struct drm_dp_aux *aux, + unsigned int offset, u8 *valuep) +{ + return drm_dp_dpcd_read(aux, offset, valuep, 1); +} + +/** + * drm_dp_dpcd_writeb() - write a single byte to the DPCD + * @aux: DisplayPort AUX channel + * @offset: address of the register to write + * @value: value to write to the register + * + * Returns the number of bytes transferred (1) on success, or a negative + * error code on failure. + */ +static inline ssize_t drm_dp_dpcd_writeb(struct drm_dp_aux *aux, + unsigned int offset, u8 value) +{ + return drm_dp_dpcd_write(aux, offset, &value, 1); +} + #endif /* _DRM_DP_HELPER_H_ */ From 8d4adc6a5807ca51999421b4d6d4f193c95775ba Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Fri, 22 Nov 2013 16:37:57 +0100 Subject: [PATCH 042/259] drm/dp: Add drm_dp_dpcd_read_link_status() The function reads the link status (6 bytes starting at offset 0x202) from the DPCD so that it can be conveniently passed to other DPCD helpers. Reviewed-by: Alex Deucher Reviewed-by: Jani Nikula Signed-off-by: Thierry Reding --- drivers/gpu/drm/drm_dp_helper.c | 16 ++++++++++++++++ include/drm/drm_dp_helper.h | 3 +++ 2 files changed, 19 insertions(+) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index da6bcfe178a6..84262ed64ce7 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -456,3 +456,19 @@ ssize_t drm_dp_dpcd_write(struct drm_dp_aux *aux, unsigned int offset, size); } EXPORT_SYMBOL(drm_dp_dpcd_write); + +/** + * drm_dp_dpcd_read_link_status() - read DPCD link status (bytes 0x202-0x207) + * @aux: DisplayPort AUX channel + * @status: buffer to store the link status in (must be at least 6 bytes) + * + * Returns the number of bytes transferred on success or a negative error + * code on failure. + */ +int drm_dp_dpcd_read_link_status(struct drm_dp_aux *aux, + u8 status[DP_LINK_STATUS_SIZE]) +{ + return drm_dp_dpcd_read(aux, DP_LANE0_1_STATUS, status, + DP_LINK_STATUS_SIZE); +} +EXPORT_SYMBOL(drm_dp_dpcd_read_link_status); diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 70f18eb3ca1a..a78711feb0d1 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -477,4 +477,7 @@ static inline ssize_t drm_dp_dpcd_writeb(struct drm_dp_aux *aux, return drm_dp_dpcd_write(aux, offset, &value, 1); } +int drm_dp_dpcd_read_link_status(struct drm_dp_aux *aux, + u8 status[DP_LINK_STATUS_SIZE]); + #endif /* _DRM_DP_HELPER_H_ */ From 516c0f7c0a608833cc01d3f5b2a357ee806b78a1 Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Mon, 9 Dec 2013 11:47:55 +0100 Subject: [PATCH 043/259] drm/dp: Add DisplayPort link helpers Add a helper to probe a DP link (read out the supported DPCD revision, maximum rate, link count and capabilities) as well as power up the DP link and configure it accordingly. Reviewed-by: Alex Deucher Reviewed-by: Jani Nikula Signed-off-by: Thierry Reding --- Changes in v5: - export helpers Changes in v4: - fix a couple of typos in comments as pointed out by Alex Deucher Changes in v3: - split into drm_dp_link_power_up() and drm_dp_link_configure() - do not change sink state for DPCD versions earlier than 1.1 - sleep for 1-2 ms after setting local sink to D0 state - read and write consecutive registers where possible - read DPCD revision when link is probed - remove duplicate kerneldoc --- drivers/gpu/drm/drm_dp_helper.c | 97 +++++++++++++++++++++++++++++++++ include/drm/drm_dp_helper.h | 17 ++++++ 2 files changed, 114 insertions(+) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 84262ed64ce7..177ac7bd1851 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -472,3 +472,100 @@ int drm_dp_dpcd_read_link_status(struct drm_dp_aux *aux, DP_LINK_STATUS_SIZE); } EXPORT_SYMBOL(drm_dp_dpcd_read_link_status); + +/** + * drm_dp_link_probe() - probe a DisplayPort link for capabilities + * @aux: DisplayPort AUX channel + * @link: pointer to structure in which to return link capabilities + * + * The structure filled in by this function can usually be passed directly + * into drm_dp_link_power_up() and drm_dp_link_configure() to power up and + * configure the link based on the link's capabilities. + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_link_probe(struct drm_dp_aux *aux, struct drm_dp_link *link) +{ + u8 values[3]; + int err; + + memset(link, 0, sizeof(*link)); + + err = drm_dp_dpcd_read(aux, DP_DPCD_REV, values, sizeof(values)); + if (err < 0) + return err; + + link->revision = values[0]; + link->rate = drm_dp_bw_code_to_link_rate(values[1]); + link->num_lanes = values[2] & DP_MAX_LANE_COUNT_MASK; + + if (values[2] & DP_ENHANCED_FRAME_CAP) + link->capabilities |= DP_LINK_CAP_ENHANCED_FRAMING; + + return 0; +} +EXPORT_SYMBOL(drm_dp_link_probe); + +/** + * drm_dp_link_power_up() - power up a DisplayPort link + * @aux: DisplayPort AUX channel + * @link: pointer to a structure containing the link configuration + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_link_power_up(struct drm_dp_aux *aux, struct drm_dp_link *link) +{ + u8 value; + int err; + + /* DP_SET_POWER register is only available on DPCD v1.1 and later */ + if (link->revision < 0x11) + return 0; + + err = drm_dp_dpcd_readb(aux, DP_SET_POWER, &value); + if (err < 0) + return err; + + value &= ~DP_SET_POWER_MASK; + value |= DP_SET_POWER_D0; + + err = drm_dp_dpcd_writeb(aux, DP_SET_POWER, value); + if (err < 0) + return err; + + /* + * According to the DP 1.1 specification, a "Sink Device must exit the + * power saving state within 1 ms" (Section 2.5.3.1, Table 5-52, "Sink + * Control Field" (register 0x600). + */ + usleep_range(1000, 2000); + + return 0; +} +EXPORT_SYMBOL(drm_dp_link_power_up); + +/** + * drm_dp_link_configure() - configure a DisplayPort link + * @aux: DisplayPort AUX channel + * @link: pointer to a structure containing the link configuration + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link) +{ + u8 values[2]; + int err; + + values[0] = drm_dp_link_rate_to_bw_code(link->rate); + values[1] = link->num_lanes; + + if (link->capabilities & DP_LINK_CAP_ENHANCED_FRAMING) + values[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN; + + err = drm_dp_dpcd_write(aux, DP_LINK_BW_SET, values, sizeof(values)); + if (err < 0) + return err; + + return 0; +} +EXPORT_SYMBOL(drm_dp_link_configure); diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index a78711feb0d1..28ab6f4e3b09 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -291,6 +291,7 @@ #define DP_SET_POWER 0x600 # define DP_SET_POWER_D0 0x1 # define DP_SET_POWER_D3 0x2 +# define DP_SET_POWER_MASK 0x3 #define DP_PSR_ERROR_STATUS 0x2006 /* XXX 1.2? */ # define DP_PSR_LINK_CRC_ERROR (1 << 0) @@ -480,4 +481,20 @@ static inline ssize_t drm_dp_dpcd_writeb(struct drm_dp_aux *aux, int drm_dp_dpcd_read_link_status(struct drm_dp_aux *aux, u8 status[DP_LINK_STATUS_SIZE]); +/* + * DisplayPort link + */ +#define DP_LINK_CAP_ENHANCED_FRAMING (1 << 0) + +struct drm_dp_link { + unsigned char revision; + unsigned int rate; + unsigned int num_lanes; + unsigned long capabilities; +}; + +int drm_dp_link_probe(struct drm_dp_aux *aux, struct drm_dp_link *link); +int drm_dp_link_power_up(struct drm_dp_aux *aux, struct drm_dp_link *link); +int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link); + #endif /* _DRM_DP_HELPER_H_ */ From 88759686c702f1fbbb8e737e6231b64a9880db73 Mon Sep 17 00:00:00 2001 From: Thierry Reding Date: Thu, 12 Dec 2013 09:57:53 +0100 Subject: [PATCH 044/259] drm/dp: Allow registering AUX channels as I2C busses Implements an I2C-over-AUX I2C adapter on top of the generic drm_dp_aux infrastructure. It extracts the retry logic from existing drivers, which should help in porting those drivers to this new helper. Reviewed-by: Alex Deucher Reviewed-by: Jani Nikula Signed-off-by: Thierry Reding --- Changes in v5: - move comments partially to to header file - keep MOT set between I2C messages - return -EPROTO on short reads Changes in v4: - fix typo "bitrate" -> "bit rate" Changes in v3: - add back DRM_DEBUG_KMS and DRM_ERROR messages - embed i2c_adapter within struct drm_dp_aux - fix typo in comment --- drivers/gpu/drm/drm_dp_helper.c | 186 ++++++++++++++++++++++++++++++-- include/drm/drm_dp_helper.h | 12 +++ 2 files changed, 191 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 177ac7bd1851..35251af3b14e 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -357,13 +357,6 @@ EXPORT_SYMBOL(drm_dp_bw_code_to_link_rate); * Transactions are described using a hardware-independent drm_dp_aux_msg * structure, which is passed into a driver's .transfer() implementation. * Both native and I2C-over-AUX transactions are supported. - * - * An AUX channel can also be used to transport I2C messages to a sink. A - * typical application of that is to access an EDID that's present in the - * sink device. The .transfer() function can also be used to execute such - * transactions. The drm_dp_aux_register_i2c_bus() function registers an - * I2C adapter that can be passed to drm_probe_ddc(). Upon removal, drivers - * should call drm_dp_aux_unregister_i2c_bus() to remove the I2C adapter. */ static int drm_dp_dpcd_access(struct drm_dp_aux *aux, u8 request, @@ -569,3 +562,182 @@ int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link) return 0; } EXPORT_SYMBOL(drm_dp_link_configure); + +/* + * I2C-over-AUX implementation + */ + +static u32 drm_dp_i2c_functionality(struct i2c_adapter *adapter) +{ + return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | + I2C_FUNC_SMBUS_READ_BLOCK_DATA | + I2C_FUNC_SMBUS_BLOCK_PROC_CALL | + I2C_FUNC_10BIT_ADDR; +} + +/* + * Transfer a single I2C-over-AUX message and handle various error conditions, + * retrying the transaction as appropriate. + */ +static int drm_dp_i2c_do_msg(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) +{ + unsigned int retry; + int err; + + /* + * DP1.2 sections 2.7.7.1.5.6.1 and 2.7.7.1.6.6.1: A DP Source device + * is required to retry at least seven times upon receiving AUX_DEFER + * before giving up the AUX transaction. + */ + for (retry = 0; retry < 7; retry++) { + err = aux->transfer(aux, msg); + if (err < 0) { + if (err == -EBUSY) + continue; + + DRM_DEBUG_KMS("transaction failed: %d\n", err); + return err; + } + + if (err < msg->size) + return -EPROTO; + + switch (msg->reply & DP_AUX_NATIVE_REPLY_MASK) { + case DP_AUX_NATIVE_REPLY_ACK: + /* + * For I2C-over-AUX transactions this isn't enough, we + * need to check for the I2C ACK reply. + */ + break; + + case DP_AUX_NATIVE_REPLY_NACK: + DRM_DEBUG_KMS("native nack\n"); + return -EREMOTEIO; + + case DP_AUX_NATIVE_REPLY_DEFER: + DRM_DEBUG_KMS("native defer"); + /* + * We could check for I2C bit rate capabilities and if + * available adjust this interval. We could also be + * more careful with DP-to-legacy adapters where a + * long legacy cable may force very low I2C bit rates. + * + * For now just defer for long enough to hopefully be + * safe for all use-cases. + */ + usleep_range(500, 600); + continue; + + default: + DRM_ERROR("invalid native reply %#04x\n", msg->reply); + return -EREMOTEIO; + } + + switch (msg->reply & DP_AUX_I2C_REPLY_MASK) { + case DP_AUX_I2C_REPLY_ACK: + /* + * Both native ACK and I2C ACK replies received. We + * can assume the transfer was successful. + */ + return 0; + + case DP_AUX_I2C_REPLY_NACK: + DRM_DEBUG_KMS("I2C nack\n"); + return -EREMOTEIO; + + case DP_AUX_I2C_REPLY_DEFER: + DRM_DEBUG_KMS("I2C defer\n"); + usleep_range(400, 500); + continue; + + default: + DRM_ERROR("invalid I2C reply %#04x\n", msg->reply); + return -EREMOTEIO; + } + } + + DRM_ERROR("too many retries, giving up\n"); + return -EREMOTEIO; +} + +static int drm_dp_i2c_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, + int num) +{ + struct drm_dp_aux *aux = adapter->algo_data; + unsigned int i, j; + + for (i = 0; i < num; i++) { + struct drm_dp_aux_msg msg; + int err; + + /* + * Many hardware implementations support FIFOs larger than a + * single byte, but it has been empirically determined that + * transferring data in larger chunks can actually lead to + * decreased performance. Therefore each message is simply + * transferred byte-by-byte. + */ + for (j = 0; j < msgs[i].len; j++) { + memset(&msg, 0, sizeof(msg)); + msg.address = msgs[i].addr; + + msg.request = (msgs[i].flags & I2C_M_RD) ? + DP_AUX_I2C_READ : + DP_AUX_I2C_WRITE; + + /* + * All messages except the last one are middle-of- + * transfer messages. + */ + if ((i < num - 1) || (j < msgs[i].len - 1)) + msg.request |= DP_AUX_I2C_MOT; + + msg.buffer = msgs[i].buf + j; + msg.size = 1; + + err = drm_dp_i2c_do_msg(aux, &msg); + if (err < 0) + return err; + } + } + + return num; +} + +static const struct i2c_algorithm drm_dp_i2c_algo = { + .functionality = drm_dp_i2c_functionality, + .master_xfer = drm_dp_i2c_xfer, +}; + +/** + * drm_dp_aux_register_i2c_bus() - register an I2C adapter for I2C-over-AUX + * @aux: DisplayPort AUX channel + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_aux_register_i2c_bus(struct drm_dp_aux *aux) +{ + aux->ddc.algo = &drm_dp_i2c_algo; + aux->ddc.algo_data = aux; + aux->ddc.retries = 3; + + aux->ddc.class = I2C_CLASS_DDC; + aux->ddc.owner = THIS_MODULE; + aux->ddc.dev.parent = aux->dev; + aux->ddc.dev.of_node = aux->dev->of_node; + + strncpy(aux->ddc.name, dev_name(aux->dev), sizeof(aux->ddc.name)); + + return i2c_add_adapter(&aux->ddc); +} +EXPORT_SYMBOL(drm_dp_aux_register_i2c_bus); + +/** + * drm_dp_aux_unregister_i2c_bus() - unregister an I2C-over-AUX adapter + * @aux: DisplayPort AUX channel + */ +void drm_dp_aux_unregister_i2c_bus(struct drm_dp_aux *aux) +{ + i2c_del_adapter(&aux->ddc); +} +EXPORT_SYMBOL(drm_dp_aux_unregister_i2c_bus); diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 28ab6f4e3b09..b7488c9849ad 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -421,6 +421,7 @@ struct drm_dp_aux_msg { /** * struct drm_dp_aux - DisplayPort AUX channel + * @ddc: I2C adapter that can be used for I2C-over-AUX communication * @dev: pointer to struct device that is the parent for this AUX channel * @transfer: transfers a message representing a single AUX transaction * @@ -435,8 +436,16 @@ struct drm_dp_aux_msg { * propagate errors from the .transfer() function, with the exception of * the -EBUSY error, which causes a transaction to be retried. On a short, * helpers will return -EPROTO to make it simpler to check for failure. + * + * An AUX channel can also be used to transport I2C messages to a sink. A + * typical application of that is to access an EDID that's present in the + * sink device. The .transfer() function can also be used to execute such + * transactions. The drm_dp_aux_register_i2c_bus() function registers an + * I2C adapter that can be passed to drm_probe_ddc(). Upon removal, drivers + * should call drm_dp_aux_unregister_i2c_bus() to remove the I2C adapter. */ struct drm_dp_aux { + struct i2c_adapter ddc; struct device *dev; ssize_t (*transfer)(struct drm_dp_aux *aux, @@ -497,4 +506,7 @@ int drm_dp_link_probe(struct drm_dp_aux *aux, struct drm_dp_link *link); int drm_dp_link_power_up(struct drm_dp_aux *aux, struct drm_dp_link *link); int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link); +int drm_dp_aux_register_i2c_bus(struct drm_dp_aux *aux); +void drm_dp_aux_unregister_i2c_bus(struct drm_dp_aux *aux); + #endif /* _DRM_DP_HELPER_H_ */ From aee4aa73a1af3176cc3eea5833cae596b4b7dd22 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 15:24:06 +0100 Subject: [PATCH 045/259] drm/radeon: improve ring lockup detection code v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use atomics and jiffies_64, so that we don't need to have the ring mutex locked any more and avoid wrap arounds. v2: fix some checkpatch warnings Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 4 ++-- drivers/gpu/drm/radeon/radeon_ring.c | 23 ++++++++++------------- 2 files changed, 12 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 4581df193932..e98fe5c7f77b 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -805,8 +805,8 @@ struct radeon_ring { unsigned ring_size; unsigned ring_free_dw; int count_dw; - unsigned long last_activity; - unsigned last_rptr; + atomic_t last_rptr; + atomic64_t last_activity; uint64_t gpu_addr; uint32_t align_mask; uint32_t ptr_mask; diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index b14c86d57607..4c4810272371 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -485,8 +485,8 @@ void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *rin void radeon_ring_lockup_update(struct radeon_device *rdev, struct radeon_ring *ring) { - ring->last_rptr = radeon_ring_get_rptr(rdev, ring); - ring->last_activity = jiffies; + atomic_set(&ring->last_rptr, radeon_ring_get_rptr(rdev, ring)); + atomic64_set(&ring->last_activity, jiffies_64); } /** @@ -498,22 +498,19 @@ void radeon_ring_lockup_update(struct radeon_device *rdev, bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t rptr = radeon_ring_get_rptr(rdev, ring); - unsigned long cjiffies, elapsed; + uint64_t last = atomic64_read(&ring->last_activity); + uint64_t elapsed; - cjiffies = jiffies; - if (!time_after(cjiffies, ring->last_activity)) { - /* likely a wrap around */ + if (rptr != atomic_read(&ring->last_rptr)) { + /* ring is still working, no lockup */ radeon_ring_lockup_update(rdev, ring); return false; } - if (rptr != ring->last_rptr) { - /* CP is still working no lockup */ - radeon_ring_lockup_update(rdev, ring); - return false; - } - elapsed = jiffies_to_msecs(cjiffies - ring->last_activity); + + elapsed = jiffies_to_msecs(jiffies_64 - last); if (radeon_lockup_timeout && elapsed >= radeon_lockup_timeout) { - dev_err(rdev->dev, "GPU lockup CP stall for more than %lumsec\n", elapsed); + dev_err(rdev->dev, "ring %d stalled for more than %llumsec\n", + ring->idx, elapsed); return true; } /* give a chance to the GPU ... */ From 37615527c5669f0c332534a797e5aaa175b6f3cb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 18 Feb 2014 15:58:31 +0100 Subject: [PATCH 046/259] drm/radeon: cleanup the fence ring locking code MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We no longer need to take the ring lock while checking for a gpu lockup, so just cleanup the code. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 5 +- drivers/gpu/drm/radeon/radeon_device.c | 4 +- drivers/gpu/drm/radeon/radeon_fence.c | 59 ++++------------------- drivers/gpu/drm/radeon/radeon_pm.c | 4 +- drivers/gpu/drm/radeon/radeon_ring.c | 2 +- drivers/gpu/drm/radeon/radeon_semaphore.c | 12 +++-- 6 files changed, 25 insertions(+), 61 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index e98fe5c7f77b..1ac3393bab66 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -363,9 +363,8 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, i void radeon_fence_process(struct radeon_device *rdev, int ring); bool radeon_fence_signaled(struct radeon_fence *fence); int radeon_fence_wait(struct radeon_fence *fence, bool interruptible); -int radeon_fence_wait_locked(struct radeon_fence *fence); -int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring); -int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring); +int radeon_fence_wait_next(struct radeon_device *rdev, int ring); +int radeon_fence_wait_empty(struct radeon_device *rdev, int ring); int radeon_fence_wait_any(struct radeon_device *rdev, struct radeon_fence **fences, bool intr); diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index b012cbbc3ed5..fa7841b9d7b6 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -1445,10 +1445,9 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon) /* evict vram memory */ radeon_bo_evict_vram(rdev); - mutex_lock(&rdev->ring_lock); /* wait for gpu to finish processing current batch */ for (i = 0; i < RADEON_NUM_RINGS; i++) { - r = radeon_fence_wait_empty_locked(rdev, i); + r = radeon_fence_wait_empty(rdev, i); if (r) { /* delay GPU reset to resume */ force_completion = true; @@ -1457,7 +1456,6 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon) if (force_completion) { radeon_fence_driver_force_completion(rdev); } - mutex_unlock(&rdev->ring_lock); radeon_save_bios_scratch_regs(rdev); diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index c37cb79a9489..a77b1c13ea43 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -288,7 +288,6 @@ static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq) * @rdev: radeon device pointer * @target_seq: sequence number(s) we want to wait for * @intr: use interruptable sleep - * @lock_ring: whether the ring should be locked or not * * Wait for the requested sequence number(s) to be written by any ring * (all asics). Sequnce number array is indexed by ring id. @@ -299,7 +298,7 @@ static bool radeon_fence_any_seq_signaled(struct radeon_device *rdev, u64 *seq) * -EDEADLK is returned when a GPU lockup has been detected. */ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq, - bool intr, bool lock_ring) + bool intr) { uint64_t last_seq[RADEON_NUM_RINGS]; bool signaled; @@ -358,9 +357,6 @@ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq, if (i != RADEON_NUM_RINGS) continue; - if (lock_ring) - mutex_lock(&rdev->ring_lock); - for (i = 0; i < RADEON_NUM_RINGS; ++i) { if (!target_seq[i]) continue; @@ -378,14 +374,9 @@ static int radeon_fence_wait_seq(struct radeon_device *rdev, u64 *target_seq, /* remember that we need an reset */ rdev->needs_reset = true; - if (lock_ring) - mutex_unlock(&rdev->ring_lock); wake_up_all(&rdev->fence_queue); return -EDEADLK; } - - if (lock_ring) - mutex_unlock(&rdev->ring_lock); } } return 0; @@ -416,7 +407,7 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) if (seq[fence->ring] == RADEON_FENCE_SIGNALED_SEQ) return 0; - r = radeon_fence_wait_seq(fence->rdev, seq, intr, true); + r = radeon_fence_wait_seq(fence->rdev, seq, intr); if (r) return r; @@ -464,7 +455,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev, if (num_rings == 0) return -ENOENT; - r = radeon_fence_wait_seq(rdev, seq, intr, true); + r = radeon_fence_wait_seq(rdev, seq, intr); if (r) { return r; } @@ -472,37 +463,7 @@ int radeon_fence_wait_any(struct radeon_device *rdev, } /** - * radeon_fence_wait_locked - wait for a fence to signal - * - * @fence: radeon fence object - * - * Wait for the requested fence to signal (all asics). - * Returns 0 if the fence has passed, error for all other cases. - */ -int radeon_fence_wait_locked(struct radeon_fence *fence) -{ - uint64_t seq[RADEON_NUM_RINGS] = {}; - int r; - - if (fence == NULL) { - WARN(1, "Querying an invalid fence : %p !\n", fence); - return -EINVAL; - } - - seq[fence->ring] = fence->seq; - if (seq[fence->ring] == RADEON_FENCE_SIGNALED_SEQ) - return 0; - - r = radeon_fence_wait_seq(fence->rdev, seq, false, false); - if (r) - return r; - - fence->seq = RADEON_FENCE_SIGNALED_SEQ; - return 0; -} - -/** - * radeon_fence_wait_next_locked - wait for the next fence to signal + * radeon_fence_wait_next - wait for the next fence to signal * * @rdev: radeon device pointer * @ring: ring index the fence is associated with @@ -511,7 +472,7 @@ int radeon_fence_wait_locked(struct radeon_fence *fence) * Returns 0 if the next fence has passed, error for all other cases. * Caller must hold ring lock. */ -int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring) +int radeon_fence_wait_next(struct radeon_device *rdev, int ring) { uint64_t seq[RADEON_NUM_RINGS] = {}; @@ -521,11 +482,11 @@ int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring) already the last emited fence */ return -ENOENT; } - return radeon_fence_wait_seq(rdev, seq, false, false); + return radeon_fence_wait_seq(rdev, seq, false); } /** - * radeon_fence_wait_empty_locked - wait for all fences to signal + * radeon_fence_wait_empty - wait for all fences to signal * * @rdev: radeon device pointer * @ring: ring index the fence is associated with @@ -534,7 +495,7 @@ int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring) * Returns 0 if the fences have passed, error for all other cases. * Caller must hold ring lock. */ -int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring) +int radeon_fence_wait_empty(struct radeon_device *rdev, int ring) { uint64_t seq[RADEON_NUM_RINGS] = {}; int r; @@ -543,7 +504,7 @@ int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring) if (!seq[ring]) return 0; - r = radeon_fence_wait_seq(rdev, seq, false, false); + r = radeon_fence_wait_seq(rdev, seq, false); if (r) { if (r == -EDEADLK) return -EDEADLK; @@ -794,7 +755,7 @@ void radeon_fence_driver_fini(struct radeon_device *rdev) for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { if (!rdev->fence_drv[ring].initialized) continue; - r = radeon_fence_wait_empty_locked(rdev, ring); + r = radeon_fence_wait_empty(rdev, ring); if (r) { /* no need to trigger GPU reset as we are unloading */ radeon_fence_driver_force_completion(rdev); diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index 4ad9af9fc517..0119af46ede3 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -260,7 +260,7 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) if (!ring->ready) { continue; } - r = radeon_fence_wait_empty_locked(rdev, i); + r = radeon_fence_wait_empty(rdev, i); if (r) { /* needs a GPU reset dont reset here */ mutex_unlock(&rdev->ring_lock); @@ -896,7 +896,7 @@ static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev) for (i = 0; i < RADEON_NUM_RINGS; i++) { struct radeon_ring *ring = &rdev->ring[i]; if (ring->ready) - radeon_fence_wait_empty_locked(rdev, i); + radeon_fence_wait_empty(rdev, i); } /* program the new power state */ diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 4c4810272371..fa140119cdb6 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -382,7 +382,7 @@ int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsi if (ndw < ring->ring_free_dw) { break; } - r = radeon_fence_wait_next_locked(rdev, ring->idx); + r = radeon_fence_wait_next(rdev, ring->idx); if (r) return r; } diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c index 9006b32d5eed..6140af6772c7 100644 --- a/drivers/gpu/drm/radeon/radeon_semaphore.c +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -147,7 +147,9 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, if (++count > RADEON_NUM_SYNCS) { /* not enough room, wait manually */ - radeon_fence_wait_locked(fence); + r = radeon_fence_wait(fence, false); + if (r) + return r; continue; } @@ -161,7 +163,9 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, if (!radeon_semaphore_emit_signal(rdev, i, semaphore)) { /* signaling wasn't successful wait manually */ radeon_ring_undo(&rdev->ring[i]); - radeon_fence_wait_locked(fence); + r = radeon_fence_wait(fence, false); + if (r) + return r; continue; } @@ -169,7 +173,9 @@ int radeon_semaphore_sync_rings(struct radeon_device *rdev, if (!radeon_semaphore_emit_wait(rdev, ring, semaphore)) { /* waiting wasn't successful wait manually */ radeon_ring_undo(&rdev->ring[i]); - radeon_fence_wait_locked(fence); + r = radeon_fence_wait(fence, false); + if (r) + return r; continue; } From 14a9579ddbf15dd1992a9481a4ec80b0b91656d5 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 21 Feb 2014 11:34:35 -0500 Subject: [PATCH 047/259] drm/radeon: use variable UVD clocks MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now that Christian fixed the performance problems with the feedback buffer in mesa, we can enable variable UVD clocks. There are multiple UVD power states associated with different types and numbers of streams. This uses the appropriate state based on that information rather than always using the fastest UVD clocks which saves some power. One possible downside is that this may adversely affect decode benchmarks since these power states target specific playback requirements rather than maximum performance. If that becomes an issue, we can add a sysfs attribute to force the max UVD state. Signed-off-by: Alex Deucher Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon_pm.c | 3 --- drivers/gpu/drm/radeon/radeon_uvd.c | 3 +-- 2 files changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index 0119af46ede3..ee738a524639 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -943,8 +943,6 @@ void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable) if (enable) { mutex_lock(&rdev->pm.mutex); rdev->pm.dpm.uvd_active = true; - /* disable this for now */ -#if 0 if ((rdev->pm.dpm.sd == 1) && (rdev->pm.dpm.hd == 0)) dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_SD; else if ((rdev->pm.dpm.sd == 2) && (rdev->pm.dpm.hd == 0)) @@ -954,7 +952,6 @@ void radeon_dpm_enable_uvd(struct radeon_device *rdev, bool enable) else if ((rdev->pm.dpm.sd == 0) && (rdev->pm.dpm.hd == 2)) dpm_state = POWER_STATE_TYPE_INTERNAL_UVD_HD2; else -#endif dpm_state = POWER_STATE_TYPE_INTERNAL_UVD; rdev->pm.dpm.state = dpm_state; mutex_unlock(&rdev->pm.mutex); diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c index 6781fee1eaad..ceb7b289b745 100644 --- a/drivers/gpu/drm/radeon/radeon_uvd.c +++ b/drivers/gpu/drm/radeon/radeon_uvd.c @@ -805,8 +805,7 @@ void radeon_uvd_note_usage(struct radeon_device *rdev) (rdev->pm.dpm.hd != hd)) { rdev->pm.dpm.sd = sd; rdev->pm.dpm.hd = hd; - /* disable this for now */ - /*streams_changed = true;*/ + streams_changed = true; } } From 48095d991d85687569ac025b18a6c7ae1632c9f7 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Mon, 3 Feb 2014 17:25:33 -0800 Subject: [PATCH 048/259] audit: Use struct net not pid_t to remember the network namespce to reply in In struct audit_netlink_list and audit_reply add a reference to the network namespace of the caller and remove the userspace pid of the caller. This cleanly remembers the callers network namespace, and removes a huge class of races and nasty failure modes that can occur when attempting to relook up the callers network namespace from a pid_t (including the caller's network namespace changing, pid wraparound, and the pid simply not being present). Signed-off-by: "Eric W. Biederman" --- kernel/audit.c | 10 ++++++---- kernel/audit.h | 2 +- kernel/auditfilter.c | 3 ++- 3 files changed, 9 insertions(+), 6 deletions(-) diff --git a/kernel/audit.c b/kernel/audit.c index 34c5a2310fbf..1e5756f16f6f 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -182,7 +182,7 @@ struct audit_buffer { struct audit_reply { __u32 portid; - pid_t pid; + struct net *net; struct sk_buff *skb; }; @@ -500,7 +500,7 @@ int audit_send_list(void *_dest) { struct audit_netlink_list *dest = _dest; struct sk_buff *skb; - struct net *net = get_net_ns_by_pid(dest->pid); + struct net *net = dest->net; struct audit_net *aunet = net_generic(net, audit_net_id); /* wait for parent to finish and send an ACK */ @@ -510,6 +510,7 @@ int audit_send_list(void *_dest) while ((skb = __skb_dequeue(&dest->q)) != NULL) netlink_unicast(aunet->nlsk, skb, dest->portid, 0); + put_net(net); kfree(dest); return 0; @@ -543,7 +544,7 @@ struct sk_buff *audit_make_reply(__u32 portid, int seq, int type, int done, static int audit_send_reply_thread(void *arg) { struct audit_reply *reply = (struct audit_reply *)arg; - struct net *net = get_net_ns_by_pid(reply->pid); + struct net *net = reply->net; struct audit_net *aunet = net_generic(net, audit_net_id); mutex_lock(&audit_cmd_mutex); @@ -552,6 +553,7 @@ static int audit_send_reply_thread(void *arg) /* Ignore failure. It'll only happen if the sender goes away, because our timeout is set to infinite. */ netlink_unicast(aunet->nlsk , reply->skb, reply->portid, 0); + put_net(net); kfree(reply); return 0; } @@ -583,8 +585,8 @@ static void audit_send_reply(__u32 portid, int seq, int type, int done, if (!skb) goto out; + reply->net = get_net(current->nsproxy->net_ns); reply->portid = portid; - reply->pid = task_pid_vnr(current); reply->skb = skb; tsk = kthread_run(audit_send_reply_thread, reply, "audit_send_reply"); diff --git a/kernel/audit.h b/kernel/audit.h index 57cc64d67718..8df132214606 100644 --- a/kernel/audit.h +++ b/kernel/audit.h @@ -247,7 +247,7 @@ extern void audit_panic(const char *message); struct audit_netlink_list { __u32 portid; - pid_t pid; + struct net *net; struct sk_buff_head q; }; diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index 14a78cca384e..a5e3d73d73e4 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -29,6 +29,7 @@ #include #include #include +#include #include "audit.h" /* @@ -1083,8 +1084,8 @@ int audit_list_rules_send(__u32 portid, int seq) dest = kmalloc(sizeof(struct audit_netlink_list), GFP_KERNEL); if (!dest) return -ENOMEM; + dest->net = get_net(current->nsproxy->net_ns); dest->portid = portid; - dest->pid = task_pid_vnr(current); skb_queue_head_init(&dest->q); mutex_lock(&audit_filter_mutex); From dca1c8d17a2feae056f9e334ea75a462ae4cb52a Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Sat, 22 Feb 2014 19:35:38 -0500 Subject: [PATCH 049/259] cifs: mask off top byte in get_rfc1002_length() The rfc1002 length actually includes a type byte, which we aren't masking off. In most cases, it's not a problem since the RFC1002_SESSION_MESSAGE type is 0, but when doing a RFC1002 session establishment, the type is non-zero and that throws off the returned length. Signed-off-by: Jeff Layton Tested-by: Sachin Prabhu Signed-off-by: Steve French --- fs/cifs/cifsglob.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index cf32f0393369..c0f3718b77a8 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -513,7 +513,7 @@ struct cifs_mnt_data { static inline unsigned int get_rfc1002_length(void *buf) { - return be32_to_cpu(*((__be32 *)buf)); + return be32_to_cpu(*((__be32 *)buf)) & 0xffffff; } static inline void From 6f285b19d09f72e801525f5eea1bdad22e559bf0 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Fri, 28 Feb 2014 19:44:55 -0800 Subject: [PATCH 050/259] audit: Send replies in the proper network namespace. In perverse cases of file descriptor passing the current network namespace of a process and the network namespace of a socket used by that socket may differ. Therefore use the network namespace of the appropiate socket to ensure replies always go to the appropiate socket. Signed-off-by: "Eric W. Biederman" --- include/linux/audit.h | 3 ++- kernel/audit.c | 21 ++++++++++----------- kernel/auditfilter.c | 7 +++++-- 3 files changed, 17 insertions(+), 14 deletions(-) diff --git a/include/linux/audit.h b/include/linux/audit.h index aa865a9a4c4f..ec1464df4c60 100644 --- a/include/linux/audit.h +++ b/include/linux/audit.h @@ -43,6 +43,7 @@ struct mq_attr; struct mqstat; struct audit_watch; struct audit_tree; +struct sk_buff; struct audit_krule { int vers_ops; @@ -463,7 +464,7 @@ extern int audit_filter_user(int type); extern int audit_filter_type(int type); extern int audit_rule_change(int type, __u32 portid, int seq, void *data, size_t datasz); -extern int audit_list_rules_send(__u32 portid, int seq); +extern int audit_list_rules_send(struct sk_buff *request_skb, int seq); extern u32 audit_enabled; #else /* CONFIG_AUDIT */ diff --git a/kernel/audit.c b/kernel/audit.c index 1e5756f16f6f..32086bff5564 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -570,9 +570,11 @@ static int audit_send_reply_thread(void *arg) * Allocates an skb, builds the netlink message, and sends it to the port id. * No failure notifications. */ -static void audit_send_reply(__u32 portid, int seq, int type, int done, +static void audit_send_reply(struct sk_buff *request_skb, int seq, int type, int done, int multi, const void *payload, int size) { + u32 portid = NETLINK_CB(request_skb).portid; + struct net *net = sock_net(NETLINK_CB(request_skb).sk); struct sk_buff *skb; struct task_struct *tsk; struct audit_reply *reply = kmalloc(sizeof(struct audit_reply), @@ -585,7 +587,7 @@ static void audit_send_reply(__u32 portid, int seq, int type, int done, if (!skb) goto out; - reply->net = get_net(current->nsproxy->net_ns); + reply->net = get_net(net); reply->portid = portid; reply->skb = skb; @@ -675,8 +677,7 @@ static int audit_get_feature(struct sk_buff *skb) seq = nlmsg_hdr(skb)->nlmsg_seq; - audit_send_reply(NETLINK_CB(skb).portid, seq, AUDIT_GET, 0, 0, - &af, sizeof(af)); + audit_send_reply(skb, seq, AUDIT_GET, 0, 0, &af, sizeof(af)); return 0; } @@ -796,8 +797,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) s.backlog = skb_queue_len(&audit_skb_queue); s.version = AUDIT_VERSION_LATEST; s.backlog_wait_time = audit_backlog_wait_time; - audit_send_reply(NETLINK_CB(skb).portid, seq, AUDIT_GET, 0, 0, - &s, sizeof(s)); + audit_send_reply(skb, seq, AUDIT_GET, 0, 0, &s, sizeof(s)); break; } case AUDIT_SET: { @@ -907,7 +907,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) seq, data, nlmsg_len(nlh)); break; case AUDIT_LIST_RULES: - err = audit_list_rules_send(NETLINK_CB(skb).portid, seq); + err = audit_list_rules_send(skb, seq); break; case AUDIT_TRIM: audit_trim_trees(); @@ -972,8 +972,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) memcpy(sig_data->ctx, ctx, len); security_release_secctx(ctx, len); } - audit_send_reply(NETLINK_CB(skb).portid, seq, AUDIT_SIGNAL_INFO, - 0, 0, sig_data, sizeof(*sig_data) + len); + audit_send_reply(skb, seq, AUDIT_SIGNAL_INFO, 0, 0, + sig_data, sizeof(*sig_data) + len); kfree(sig_data); break; case AUDIT_TTY_GET: { @@ -985,8 +985,7 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) s.log_passwd = tsk->signal->audit_tty_log_passwd; spin_unlock(&tsk->sighand->siglock); - audit_send_reply(NETLINK_CB(skb).portid, seq, - AUDIT_TTY_GET, 0, 0, &s, sizeof(s)); + audit_send_reply(skb, seq, AUDIT_TTY_GET, 0, 0, &s, sizeof(s)); break; } case AUDIT_TTY_SET: { diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index a5e3d73d73e4..e8d1c7c515d7 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -30,6 +30,7 @@ #include #include #include +#include #include "audit.h" /* @@ -1069,8 +1070,10 @@ int audit_rule_change(int type, __u32 portid, int seq, void *data, * @portid: target portid for netlink audit messages * @seq: netlink audit message sequence (serial) number */ -int audit_list_rules_send(__u32 portid, int seq) +int audit_list_rules_send(struct sk_buff *request_skb, int seq) { + u32 portid = NETLINK_CB(request_skb).portid; + struct net *net = sock_net(NETLINK_CB(request_skb).sk); struct task_struct *tsk; struct audit_netlink_list *dest; int err = 0; @@ -1084,7 +1087,7 @@ int audit_list_rules_send(__u32 portid, int seq) dest = kmalloc(sizeof(struct audit_netlink_list), GFP_KERNEL); if (!dest) return -ENOMEM; - dest->net = get_net(current->nsproxy->net_ns); + dest->net = get_net(net); dest->portid = portid; skb_queue_head_init(&dest->q); From bda72d58a20120aee1f78eb17d7eddb955d6696b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:17 +0100 Subject: [PATCH 051/259] drm/radeon: add a way to get and set initial buffer domains v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When passing buffers between processes, the receiving process needs to know the original buffer domain, so that it doesn't accidentally move the buffer. v2: reserve the buffer Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 3 +++ drivers/gpu/drm/radeon/radeon_drv.c | 3 ++- drivers/gpu/drm/radeon/radeon_gem.c | 36 ++++++++++++++++++++++++++ drivers/gpu/drm/radeon/radeon_kms.c | 1 + drivers/gpu/drm/radeon/radeon_object.c | 3 +++ include/uapi/drm/radeon_drm.h | 12 +++++++++ 6 files changed, 57 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 1ac3393bab66..c20d88c93940 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -456,6 +456,7 @@ struct radeon_bo { /* Protected by gem.mutex */ struct list_head list; /* Protected by tbo.reserved */ + u32 initial_domain; u32 placements[3]; struct ttm_placement placement; struct ttm_buffer_object tbo; @@ -2116,6 +2117,8 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_va_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); +int radeon_gem_op_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp); int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 84a1bbb75f91..4392b7c95ee6 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -79,9 +79,10 @@ * 2.35.0 - Add CIK macrotile mode array query * 2.36.0 - Fix CIK DCE tiling setup * 2.37.0 - allow GS ring setup on r6xx/r7xx + * 2.38.0 - RADEON_GEM_OP (GET_INITIAL_DOMAIN, SET_INITIAL_DOMAIN) */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 37 +#define KMS_DRIVER_MINOR 38 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index b96c819024b3..9863ca742494 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -533,6 +533,42 @@ int radeon_gem_va_ioctl(struct drm_device *dev, void *data, return r; } +int radeon_gem_op_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + struct drm_radeon_gem_op *args = data; + struct drm_gem_object *gobj; + struct radeon_bo *robj; + int r; + + gobj = drm_gem_object_lookup(dev, filp, args->handle); + if (gobj == NULL) { + return -ENOENT; + } + robj = gem_to_radeon_bo(gobj); + r = radeon_bo_reserve(robj, false); + if (unlikely(r)) + goto out; + + switch (args->op) { + case RADEON_GEM_OP_GET_INITIAL_DOMAIN: + args->value = robj->initial_domain; + break; + case RADEON_GEM_OP_SET_INITIAL_DOMAIN: + robj->initial_domain = args->value & (RADEON_GEM_DOMAIN_VRAM | + RADEON_GEM_DOMAIN_GTT | + RADEON_GEM_DOMAIN_CPU); + break; + default: + r = -EINVAL; + } + + radeon_bo_unreserve(robj); +out: + drm_gem_object_unreference_unlocked(gobj); + return r; +} + int radeon_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args) diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index baff98be65b1..0b631ebeeb18 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -814,5 +814,6 @@ const struct drm_ioctl_desc radeon_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(RADEON_GEM_OP, radeon_gem_op_ioctl, DRM_AUTH|DRM_UNLOCKED|DRM_RENDER_ALLOW), }; int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 08595cf90b01..dd12bb4025ac 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -145,6 +145,9 @@ int radeon_bo_create(struct radeon_device *rdev, bo->surface_reg = -1; INIT_LIST_HEAD(&bo->list); INIT_LIST_HEAD(&bo->va); + bo->initial_domain = domain & (RADEON_GEM_DOMAIN_VRAM | + RADEON_GEM_DOMAIN_GTT | + RADEON_GEM_DOMAIN_CPU); radeon_ttm_placement_from_domain(bo, domain); /* Kernel allocation are uninterruptible */ down_read(&rdev->pm.mclk_lock); diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h index 1cf18b4a39ec..cb5c93a4c266 100644 --- a/include/uapi/drm/radeon_drm.h +++ b/include/uapi/drm/radeon_drm.h @@ -510,6 +510,7 @@ typedef struct { #define DRM_RADEON_GEM_GET_TILING 0x29 #define DRM_RADEON_GEM_BUSY 0x2a #define DRM_RADEON_GEM_VA 0x2b +#define DRM_RADEON_GEM_OP 0x2c #define DRM_IOCTL_RADEON_CP_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_RADEON_CP_INIT, drm_radeon_init_t) #define DRM_IOCTL_RADEON_CP_START DRM_IO( DRM_COMMAND_BASE + DRM_RADEON_CP_START) @@ -552,6 +553,7 @@ typedef struct { #define DRM_IOCTL_RADEON_GEM_GET_TILING DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_GET_TILING, struct drm_radeon_gem_get_tiling) #define DRM_IOCTL_RADEON_GEM_BUSY DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_BUSY, struct drm_radeon_gem_busy) #define DRM_IOCTL_RADEON_GEM_VA DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_VA, struct drm_radeon_gem_va) +#define DRM_IOCTL_RADEON_GEM_OP DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_OP, struct drm_radeon_gem_op) typedef struct drm_radeon_init { enum { @@ -884,6 +886,16 @@ struct drm_radeon_gem_pwrite { uint64_t data_ptr; }; +/* Sets or returns a value associated with a buffer. */ +struct drm_radeon_gem_op { + uint32_t handle; /* buffer */ + uint32_t op; /* RADEON_GEM_OP_* */ + uint64_t value; /* input or return value */ +}; + +#define RADEON_GEM_OP_GET_INITIAL_DOMAIN 0 +#define RADEON_GEM_OP_SET_INITIAL_DOMAIN 1 + #define RADEON_VA_MAP 1 #define RADEON_VA_UNMAP 2 From 67e8e3f970ad747d3c854fb40f8ec0cecedd9089 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:18 +0100 Subject: [PATCH 052/259] drm/radeon: track memory statistics about VRAM and GTT usage and buffer moves v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The statistics are: - VRAM usage in bytes - GTT usage in bytes - number of bytes moved by TTM The last one is actually a counter, so you need to sample it before and after command submission and take the difference. This is useful for finding performance bottlenecks. Userspace queries are also added. v2: use atomic64_t Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 4 +++ drivers/gpu/drm/radeon/radeon_kms.c | 15 +++++++++++ drivers/gpu/drm/radeon/radeon_object.c | 36 +++++++++++++++++++++++++- drivers/gpu/drm/radeon/radeon_object.h | 2 +- drivers/gpu/drm/radeon/radeon_ttm.c | 8 +++++- include/uapi/drm/radeon_drm.h | 3 +++ 6 files changed, 65 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index c20d88c93940..7bb8fd96f3ce 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -2309,6 +2309,10 @@ struct radeon_device { /* virtual memory */ struct radeon_vm_manager vm_manager; struct mutex gpu_clock_mutex; + /* memory stats */ + atomic64_t vram_usage; + atomic64_t gtt_usage; + atomic64_t num_bytes_moved; /* ACPI interface */ struct radeon_atif atif; struct radeon_atcs atcs; diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 0b631ebeeb18..806506c03a79 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -486,6 +486,21 @@ static int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file case RADEON_INFO_VCE_FB_VERSION: *value = rdev->vce.fb_version; break; + case RADEON_INFO_NUM_BYTES_MOVED: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->num_bytes_moved); + break; + case RADEON_INFO_VRAM_USAGE: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->vram_usage); + break; + case RADEON_INFO_GTT_USAGE: + value = (uint32_t*)&value64; + value_size = sizeof(uint64_t); + value64 = atomic64_read(&rdev->gtt_usage); + break; default: DRM_DEBUG_KMS("Invalid request %d\n", info->request); return -EINVAL; diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index dd12bb4025ac..282d6a248396 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -56,11 +56,36 @@ static void radeon_bo_clear_va(struct radeon_bo *bo) } } +static void radeon_update_memory_usage(struct radeon_bo *bo, + unsigned mem_type, int sign) +{ + struct radeon_device *rdev = bo->rdev; + u64 size = (u64)bo->tbo.num_pages << PAGE_SHIFT; + + switch (mem_type) { + case TTM_PL_TT: + if (sign > 0) + atomic64_add(size, &rdev->gtt_usage); + else + atomic64_sub(size, &rdev->gtt_usage); + break; + case TTM_PL_VRAM: + if (sign > 0) + atomic64_add(size, &rdev->vram_usage); + else + atomic64_sub(size, &rdev->vram_usage); + break; + } +} + static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo) { struct radeon_bo *bo; bo = container_of(tbo, struct radeon_bo, tbo); + + radeon_update_memory_usage(bo, bo->tbo.mem.mem_type, -1); + mutex_lock(&bo->rdev->gem.mutex); list_del_init(&bo->list); mutex_unlock(&bo->rdev->gem.mutex); @@ -567,14 +592,23 @@ int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved, } void radeon_bo_move_notify(struct ttm_buffer_object *bo, - struct ttm_mem_reg *mem) + struct ttm_mem_reg *new_mem) { struct radeon_bo *rbo; + if (!radeon_ttm_bo_is_radeon_bo(bo)) return; + rbo = container_of(bo, struct radeon_bo, tbo); radeon_bo_check_tiling(rbo, 0, 1); radeon_vm_bo_invalidate(rbo->rdev, rbo); + + /* update statistics */ + if (!new_mem) + return; + + radeon_update_memory_usage(rbo, bo->mem.mem_type, -1); + radeon_update_memory_usage(rbo, new_mem->mem_type, 1); } int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo) diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index 209b11150263..a9a8c11bd80d 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -151,7 +151,7 @@ extern void radeon_bo_get_tiling_flags(struct radeon_bo *bo, extern int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved, bool force_drop); extern void radeon_bo_move_notify(struct ttm_buffer_object *bo, - struct ttm_mem_reg *mem); + struct ttm_mem_reg *new_mem); extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo); extern int radeon_bo_get_surface_reg(struct radeon_bo *bo); diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 77f5b0c3edb8..60dfce889ecf 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -406,8 +406,14 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, if (r) { memcpy: r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + if (r) { + return r; + } } - return r; + + /* update statistics */ + atomic64_add((u64)bo->num_pages << PAGE_SHIFT, &rdev->num_bytes_moved); + return 0; } static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) diff --git a/include/uapi/drm/radeon_drm.h b/include/uapi/drm/radeon_drm.h index cb5c93a4c266..aefa2f6afa3b 100644 --- a/include/uapi/drm/radeon_drm.h +++ b/include/uapi/drm/radeon_drm.h @@ -1004,6 +1004,9 @@ struct drm_radeon_cs { #define RADEON_INFO_VCE_FW_VERSION 0x1b /* version of VCE feedback */ #define RADEON_INFO_VCE_FB_VERSION 0x1c +#define RADEON_INFO_NUM_BYTES_MOVED 0x1d +#define RADEON_INFO_VRAM_USAGE 0x1e +#define RADEON_INFO_GTT_USAGE 0x1f struct drm_radeon_info { From 0bc490a8d9e0f2f54ec8f9d09a367db66605ff40 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:19 +0100 Subject: [PATCH 053/259] drm/radeon: deduplicate code in radeon_gem_busy_ioctl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon_gem.c | 13 +------------ 1 file changed, 1 insertion(+), 12 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index 9863ca742494..d09650c1d720 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -344,18 +344,7 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data, } robj = gem_to_radeon_bo(gobj); r = radeon_bo_wait(robj, &cur_placement, true); - switch (cur_placement) { - case TTM_PL_VRAM: - args->domain = RADEON_GEM_DOMAIN_VRAM; - break; - case TTM_PL_TT: - args->domain = RADEON_GEM_DOMAIN_GTT; - break; - case TTM_PL_SYSTEM: - args->domain = RADEON_GEM_DOMAIN_CPU; - default: - break; - } + args->domain = radeon_mem_type_to_domain(cur_placement); drm_gem_object_unreference_unlocked(gobj); r = radeon_gem_handle_lockup(rdev, r); return r; From 4330441a745ea0f1fd881438a0bbdfedda65f74a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:20 +0100 Subject: [PATCH 054/259] drm/radeon: add buffers to the LRU list from smallest to largest MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon_cs.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index f28a8d82fa19..d49a3f705e49 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -24,6 +24,7 @@ * Authors: * Jerome Glisse */ +#include #include #include #include "radeon_reg.h" @@ -290,6 +291,16 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) return 0; } +static int cmp_size_smaller_first(void *priv, struct list_head *a, + struct list_head *b) +{ + struct radeon_bo_list *la = list_entry(a, struct radeon_bo_list, tv.head); + struct radeon_bo_list *lb = list_entry(b, struct radeon_bo_list, tv.head); + + /* Sort A before B if A is smaller. */ + return (int)la->bo->tbo.num_pages - (int)lb->bo->tbo.num_pages; +} + /** * cs_parser_fini() - clean parser states * @parser: parser structure holding parsing context. @@ -303,6 +314,18 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo unsigned i; if (!error) { + /* Sort the buffer list from the smallest to largest buffer, + * which affects the order of buffers in the LRU list. + * This assures that the smallest buffers are added first + * to the LRU list, so they are likely to be later evicted + * first, instead of large buffers whose eviction is more + * expensive. + * + * This slightly lowers the number of bytes moved by TTM + * per frame under memory pressure. + */ + list_sort(NULL, &parser->validated, cmp_size_smaller_first); + ttm_eu_fence_buffer_objects(&parser->ticket, &parser->validated, parser->ib.fence); From c9b76548899cde2e729e3bca015d7e78ec5baad7 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:21 +0100 Subject: [PATCH 055/259] drm/radeon: validate relocations in the order determined by userspace v3 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Userspace should set the first 4 bits of drm_radeon_cs_reloc::flags to a number from 0 to 15. The higher the number, the higher the priority, which means a buffer with a higher number will be validated sooner. The old behavior is preserved: Buffers used for write are prioritized over read-only buffers if the userspace doesn't set the number. v2: add buffers to buckets directly, then concatenate them v3: use a stable sort Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 1 - drivers/gpu/drm/radeon/radeon_cs.c | 64 ++++++++++++++++++++++++-- drivers/gpu/drm/radeon/radeon_object.c | 10 ---- drivers/gpu/drm/radeon/radeon_object.h | 2 - 4 files changed, 61 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 7bb8fd96f3ce..efad56705e32 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -483,7 +483,6 @@ struct radeon_bo_list { struct ttm_validate_buffer tv; struct radeon_bo *bo; uint64_t gpu_offset; - bool written; unsigned domain; unsigned alt_domain; u32 tiling_flags; diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index d49a3f705e49..07e165128dbf 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -31,10 +31,52 @@ #include "radeon.h" #include "radeon_trace.h" +#define RADEON_CS_MAX_PRIORITY 32u +#define RADEON_CS_NUM_BUCKETS (RADEON_CS_MAX_PRIORITY + 1) + +/* This is based on the bucket sort with O(n) time complexity. + * An item with priority "i" is added to bucket[i]. The lists are then + * concatenated in descending order. + */ +struct radeon_cs_buckets { + struct list_head bucket[RADEON_CS_NUM_BUCKETS]; +}; + +static void radeon_cs_buckets_init(struct radeon_cs_buckets *b) +{ + unsigned i; + + for (i = 0; i < RADEON_CS_NUM_BUCKETS; i++) + INIT_LIST_HEAD(&b->bucket[i]); +} + +static void radeon_cs_buckets_add(struct radeon_cs_buckets *b, + struct list_head *item, unsigned priority) +{ + /* Since buffers which appear sooner in the relocation list are + * likely to be used more often than buffers which appear later + * in the list, the sort mustn't change the ordering of buffers + * with the same priority, i.e. it must be stable. + */ + list_add_tail(item, &b->bucket[min(priority, RADEON_CS_MAX_PRIORITY)]); +} + +static void radeon_cs_buckets_get_list(struct radeon_cs_buckets *b, + struct list_head *out_list) +{ + unsigned i; + + /* Connect the sorted buckets in the output list. */ + for (i = 0; i < RADEON_CS_NUM_BUCKETS; i++) { + list_splice(&b->bucket[i], out_list); + } +} + static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) { struct drm_device *ddev = p->rdev->ddev; struct radeon_cs_chunk *chunk; + struct radeon_cs_buckets buckets; unsigned i, j; bool duplicate; @@ -53,8 +95,12 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) if (p->relocs == NULL) { return -ENOMEM; } + + radeon_cs_buckets_init(&buckets); + for (i = 0; i < p->nrelocs; i++) { struct drm_radeon_cs_reloc *r; + unsigned priority; duplicate = false; r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4]; @@ -80,7 +126,14 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) p->relocs_ptr[i] = &p->relocs[i]; p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj); p->relocs[i].lobj.bo = p->relocs[i].robj; - p->relocs[i].lobj.written = !!r->write_domain; + + /* The userspace buffer priorities are from 0 to 15. A higher + * number means the buffer is more important. + * Also, the buffers used for write have a higher priority than + * the buffers used for read only, which doubles the range + * to 0 to 31. 32 is reserved for the kernel driver. + */ + priority = (r->flags & 0xf) * 2 + !!r->write_domain; /* the first reloc of an UVD job is the msg and that must be in VRAM, also but everything into VRAM on AGP cards to avoid @@ -94,6 +147,8 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) p->relocs[i].lobj.alt_domain = RADEON_GEM_DOMAIN_VRAM; + /* prioritize this over any other relocation */ + priority = RADEON_CS_MAX_PRIORITY; } else { uint32_t domain = r->write_domain ? r->write_domain : r->read_domains; @@ -107,9 +162,12 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo; p->relocs[i].handle = r->handle; - radeon_bo_list_add_object(&p->relocs[i].lobj, - &p->validated); + radeon_cs_buckets_add(&buckets, &p->relocs[i].lobj.tv.head, + priority); } + + radeon_cs_buckets_get_list(&buckets, &p->validated); + return radeon_bo_list_validate(&p->ticket, &p->validated, p->ring); } diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 282d6a248396..8399fe021769 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -366,16 +366,6 @@ void radeon_bo_fini(struct radeon_device *rdev) arch_phys_wc_del(rdev->mc.vram_mtrr); } -void radeon_bo_list_add_object(struct radeon_bo_list *lobj, - struct list_head *head) -{ - if (lobj->written) { - list_add(&lobj->tv.head, head); - } else { - list_add_tail(&lobj->tv.head, head); - } -} - int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, struct list_head *head, int ring) { diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index a9a8c11bd80d..6c3ca9edc2f4 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -138,8 +138,6 @@ extern int radeon_bo_evict_vram(struct radeon_device *rdev); extern void radeon_bo_force_delete(struct radeon_device *rdev); extern int radeon_bo_init(struct radeon_device *rdev); extern void radeon_bo_fini(struct radeon_device *rdev); -extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj, - struct list_head *head); extern int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, struct list_head *head, int ring); extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo, From 19dff56a5f4ba1f3a6e28282415a95a48c27bccf Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sun, 2 Mar 2014 00:56:22 +0100 Subject: [PATCH 056/259] drm/radeon: limit how much memory TTM can move per IB according to VRAM usage MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Marek Olšák Reviewed-by: Christian König --- drivers/gpu/drm/radeon/radeon_cs.c | 2 +- drivers/gpu/drm/radeon/radeon_object.c | 88 ++++++++++++++++++++++++-- drivers/gpu/drm/radeon/radeon_object.h | 3 +- 3 files changed, 85 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 07e165128dbf..5abae403ea4f 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -168,7 +168,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) radeon_cs_buckets_get_list(&buckets, &p->validated); - return radeon_bo_list_validate(&p->ticket, &p->validated, p->ring); + return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring); } static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority) diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 8399fe021769..ed03f2d15853 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -366,29 +366,105 @@ void radeon_bo_fini(struct radeon_device *rdev) arch_phys_wc_del(rdev->mc.vram_mtrr); } -int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, +/* Returns how many bytes TTM can move per IB. + */ +static u64 radeon_bo_get_threshold_for_moves(struct radeon_device *rdev) +{ + u64 real_vram_size = rdev->mc.real_vram_size; + u64 vram_usage = atomic64_read(&rdev->vram_usage); + + /* This function is based on the current VRAM usage. + * + * - If all of VRAM is free, allow relocating the number of bytes that + * is equal to 1/4 of the size of VRAM for this IB. + + * - If more than one half of VRAM is occupied, only allow relocating + * 1 MB of data for this IB. + * + * - From 0 to one half of used VRAM, the threshold decreases + * linearly. + * __________________ + * 1/4 of -|\ | + * VRAM | \ | + * | \ | + * | \ | + * | \ | + * | \ | + * | \ | + * | \________|1 MB + * |----------------| + * VRAM 0 % 100 % + * used used + * + * Note: It's a threshold, not a limit. The threshold must be crossed + * for buffer relocations to stop, so any buffer of an arbitrary size + * can be moved as long as the threshold isn't crossed before + * the relocation takes place. We don't want to disable buffer + * relocations completely. + * + * The idea is that buffers should be placed in VRAM at creation time + * and TTM should only do a minimum number of relocations during + * command submission. In practice, you need to submit at least + * a dozen IBs to move all buffers to VRAM if they are in GTT. + * + * Also, things can get pretty crazy under memory pressure and actual + * VRAM usage can change a lot, so playing safe even at 50% does + * consistently increase performance. + */ + + u64 half_vram = real_vram_size >> 1; + u64 half_free_vram = vram_usage >= half_vram ? 0 : half_vram - vram_usage; + u64 bytes_moved_threshold = half_free_vram >> 1; + return max(bytes_moved_threshold, 1024*1024ull); +} + +int radeon_bo_list_validate(struct radeon_device *rdev, + struct ww_acquire_ctx *ticket, struct list_head *head, int ring) { struct radeon_bo_list *lobj; struct radeon_bo *bo; - u32 domain; int r; + u64 bytes_moved = 0, initial_bytes_moved; + u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev); r = ttm_eu_reserve_buffers(ticket, head); if (unlikely(r != 0)) { return r; } + list_for_each_entry(lobj, head, tv.head) { bo = lobj->bo; if (!bo->pin_count) { - domain = lobj->domain; - + u32 domain = lobj->domain; + u32 current_domain = + radeon_mem_type_to_domain(bo->tbo.mem.mem_type); + + /* Check if this buffer will be moved and don't move it + * if we have moved too many buffers for this IB already. + * + * Note that this allows moving at least one buffer of + * any size, because it doesn't take the current "bo" + * into account. We don't want to disallow buffer moves + * completely. + */ + if (current_domain != RADEON_GEM_DOMAIN_CPU && + (domain & current_domain) == 0 && /* will be moved */ + bytes_moved > bytes_moved_threshold) { + /* don't move it */ + domain = current_domain; + } + retry: radeon_ttm_placement_from_domain(bo, domain); if (ring == R600_RING_TYPE_UVD_INDEX) radeon_uvd_force_into_uvd_segment(bo); - r = ttm_bo_validate(&bo->tbo, &bo->placement, - true, false); + + initial_bytes_moved = atomic64_read(&rdev->num_bytes_moved); + r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + bytes_moved += atomic64_read(&rdev->num_bytes_moved) - + initial_bytes_moved; + if (unlikely(r)) { if (r != -ERESTARTSYS && domain != lobj->alt_domain) { domain = lobj->alt_domain; diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index 6c3ca9edc2f4..7dff64d1f592 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -138,7 +138,8 @@ extern int radeon_bo_evict_vram(struct radeon_device *rdev); extern void radeon_bo_force_delete(struct radeon_device *rdev); extern int radeon_bo_init(struct radeon_device *rdev); extern void radeon_bo_fini(struct radeon_device *rdev); -extern int radeon_bo_list_validate(struct ww_acquire_ctx *ticket, +extern int radeon_bo_list_validate(struct radeon_device *rdev, + struct ww_acquire_ctx *ticket, struct list_head *head, int ring); extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo, struct vm_area_struct *vma); From f1e3dc708aaadb960b15ee40029f611475f14027 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 17:34:06 +0100 Subject: [PATCH 057/259] drm/radeon: fix missing bo reservation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon_kms.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 806506c03a79..7a810d0796a0 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -561,6 +561,10 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) radeon_vm_init(rdev, &fpriv->vm); + r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); + if (r) + return r; + /* map the ib pool buffer read only into * virtual address space */ bo_va = radeon_vm_bo_add(rdev, &fpriv->vm, @@ -568,6 +572,8 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) r = radeon_vm_bo_set_addr(rdev, bo_va, RADEON_VA_IB_OFFSET, RADEON_VM_PAGE_READABLE | RADEON_VM_PAGE_SNOOPED); + + radeon_bo_unreserve(rdev->ring_tmp_bo.bo); if (r) { radeon_vm_fini(rdev, &fpriv->vm); kfree(fpriv); From b03b4e4b6eb0563f2dc83c482b57b90b637ab81c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 28 Feb 2014 13:16:32 +0100 Subject: [PATCH 058/259] drm/radeon: fix VCE suspend/resume MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 1 - drivers/gpu/drm/radeon/radeon_vce.c | 63 ++++++++++++++++------------- 2 files changed, 34 insertions(+), 30 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index efad56705e32..40ab8a28c998 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -1634,7 +1634,6 @@ int radeon_uvd_send_upll_ctlreq(struct radeon_device *rdev, struct radeon_vce { struct radeon_bo *vcpu_bo; - void *cpu_addr; uint64_t gpu_addr; unsigned fw_version; unsigned fb_version; diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index d130432e313a..39ec7d8f33ab 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -119,7 +119,7 @@ int radeon_vce_init(struct radeon_device *rdev) if (rdev->vce.fw_version != ((40 << 24) | (2 << 16) | (2 << 8))) return -EINVAL; - /* load firmware into VRAM */ + /* allocate firmware, stack and heap BO */ size = RADEON_GPU_PAGE_ALIGN(rdev->vce_fw->size) + RADEON_VCE_STACK_SIZE + RADEON_VCE_HEAP_SIZE; @@ -130,16 +130,21 @@ int radeon_vce_init(struct radeon_device *rdev) return r; } - r = radeon_vce_resume(rdev); - if (r) + r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); + if (r) { + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) failed to reserve VCE bo\n", r); return r; + } - memset(rdev->vce.cpu_addr, 0, size); - memcpy(rdev->vce.cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size); - - r = radeon_vce_suspend(rdev); - if (r) + r = radeon_bo_pin(rdev->vce.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, + &rdev->vce.gpu_addr); + radeon_bo_unreserve(rdev->vce.vcpu_bo); + if (r) { + radeon_bo_unref(&rdev->vce.vcpu_bo); + dev_err(rdev->dev, "(%d) VCE bo pin failed\n", r); return r; + } for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) { atomic_set(&rdev->vce.handles[i], 0); @@ -158,8 +163,12 @@ int radeon_vce_init(struct radeon_device *rdev) */ void radeon_vce_fini(struct radeon_device *rdev) { - radeon_vce_suspend(rdev); + if (rdev->vce.vcpu_bo == NULL) + return; + radeon_bo_unref(&rdev->vce.vcpu_bo); + + release_firmware(rdev->vce_fw); } /** @@ -167,22 +176,23 @@ void radeon_vce_fini(struct radeon_device *rdev) * * @rdev: radeon_device pointer * - * TODO: Test VCE suspend/resume */ int radeon_vce_suspend(struct radeon_device *rdev) { - int r; + int i; if (rdev->vce.vcpu_bo == NULL) return 0; - r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); - if (!r) { - radeon_bo_kunmap(rdev->vce.vcpu_bo); - radeon_bo_unpin(rdev->vce.vcpu_bo); - radeon_bo_unreserve(rdev->vce.vcpu_bo); - } - return r; + for (i = 0; i < RADEON_MAX_VCE_HANDLES; ++i) + if (atomic_read(&rdev->vce.handles[i])) + break; + + if (i == RADEON_MAX_VCE_HANDLES) + return 0; + + /* TODO: suspending running encoding sessions isn't supported */ + return -EINVAL; } /** @@ -190,10 +200,10 @@ int radeon_vce_suspend(struct radeon_device *rdev) * * @rdev: radeon_device pointer * - * TODO: Test VCE suspend/resume */ int radeon_vce_resume(struct radeon_device *rdev) { + void *cpu_addr; int r; if (rdev->vce.vcpu_bo == NULL) @@ -201,26 +211,21 @@ int radeon_vce_resume(struct radeon_device *rdev) r = radeon_bo_reserve(rdev->vce.vcpu_bo, false); if (r) { - radeon_bo_unref(&rdev->vce.vcpu_bo); dev_err(rdev->dev, "(%d) failed to reserve VCE bo\n", r); return r; } - r = radeon_bo_pin(rdev->vce.vcpu_bo, RADEON_GEM_DOMAIN_VRAM, - &rdev->vce.gpu_addr); + r = radeon_bo_kmap(rdev->vce.vcpu_bo, &cpu_addr); if (r) { radeon_bo_unreserve(rdev->vce.vcpu_bo); - radeon_bo_unref(&rdev->vce.vcpu_bo); - dev_err(rdev->dev, "(%d) VCE bo pin failed\n", r); - return r; - } - - r = radeon_bo_kmap(rdev->vce.vcpu_bo, &rdev->vce.cpu_addr); - if (r) { dev_err(rdev->dev, "(%d) VCE map failed\n", r); return r; } + memcpy(cpu_addr, rdev->vce_fw->data, rdev->vce_fw->size); + + radeon_bo_kunmap(rdev->vce.vcpu_bo); + radeon_bo_unreserve(rdev->vce.vcpu_bo); return 0; From 2280ab57b6edc8581497d5e101c4694faf839c3e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 10:25:15 +0100 Subject: [PATCH 059/259] drm/radeon: separate gart and vm functions MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Both are complex enough on their own. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/Makefile | 2 +- drivers/gpu/drm/radeon/radeon_gart.c | 958 -------------------------- drivers/gpu/drm/radeon/radeon_vm.c | 981 +++++++++++++++++++++++++++ 3 files changed, 982 insertions(+), 959 deletions(-) create mode 100644 drivers/gpu/drm/radeon/radeon_vm.c diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index ed60caa32518..09433534dc47 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -80,7 +80,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \ r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \ rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \ trinity_smc.o ni_dpm.o si_smc.o si_dpm.o kv_smc.o kv_dpm.o ci_smc.o \ - ci_dpm.o dce6_afmt.o + ci_dpm.o dce6_afmt.o radeon_vm.o # add async DMA block radeon-y += \ diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c index a8f9b463bf2a..2e723651069b 100644 --- a/drivers/gpu/drm/radeon/radeon_gart.c +++ b/drivers/gpu/drm/radeon/radeon_gart.c @@ -28,8 +28,6 @@ #include #include #include "radeon.h" -#include "radeon_reg.h" -#include "radeon_trace.h" /* * GART @@ -394,959 +392,3 @@ void radeon_gart_fini(struct radeon_device *rdev) radeon_dummy_page_fini(rdev); } - -/* - * GPUVM - * GPUVM is similar to the legacy gart on older asics, however - * rather than there being a single global gart table - * for the entire GPU, there are multiple VM page tables active - * at any given time. The VM page tables can contain a mix - * vram pages and system memory pages and system memory pages - * can be mapped as snooped (cached system pages) or unsnooped - * (uncached system pages). - * Each VM has an ID associated with it and there is a page table - * associated with each VMID. When execting a command buffer, - * the kernel tells the the ring what VMID to use for that command - * buffer. VMIDs are allocated dynamically as commands are submitted. - * The userspace drivers maintain their own address space and the kernel - * sets up their pages tables accordingly when they submit their - * command buffers and a VMID is assigned. - * Cayman/Trinity support up to 8 active VMs at any given time; - * SI supports 16. - */ - -/* - * vm helpers - * - * TODO bind a default page at vm initialization for default address - */ - -/** - * radeon_vm_num_pde - return the number of page directory entries - * - * @rdev: radeon_device pointer - * - * Calculate the number of page directory entries (cayman+). - */ -static unsigned radeon_vm_num_pdes(struct radeon_device *rdev) -{ - return rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE; -} - -/** - * radeon_vm_directory_size - returns the size of the page directory in bytes - * - * @rdev: radeon_device pointer - * - * Calculate the size of the page directory in bytes (cayman+). - */ -static unsigned radeon_vm_directory_size(struct radeon_device *rdev) -{ - return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8); -} - -/** - * radeon_vm_manager_init - init the vm manager - * - * @rdev: radeon_device pointer - * - * Init the vm manager (cayman+). - * Returns 0 for success, error for failure. - */ -int radeon_vm_manager_init(struct radeon_device *rdev) -{ - struct radeon_vm *vm; - struct radeon_bo_va *bo_va; - int r; - unsigned size; - - if (!rdev->vm_manager.enabled) { - /* allocate enough for 2 full VM pts */ - size = radeon_vm_directory_size(rdev); - size += rdev->vm_manager.max_pfn * 8; - size *= 2; - r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager, - RADEON_GPU_PAGE_ALIGN(size), - RADEON_VM_PTB_ALIGN_SIZE, - RADEON_GEM_DOMAIN_VRAM); - if (r) { - dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n", - (rdev->vm_manager.max_pfn * 8) >> 10); - return r; - } - - r = radeon_asic_vm_init(rdev); - if (r) - return r; - - rdev->vm_manager.enabled = true; - - r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager); - if (r) - return r; - } - - /* restore page table */ - list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) { - if (vm->page_directory == NULL) - continue; - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } - } - return 0; -} - -/** - * radeon_vm_free_pt - free the page table for a specific vm - * - * @rdev: radeon_device pointer - * @vm: vm to unbind - * - * Free the page table of a specific vm (cayman+). - * - * Global and local mutex must be lock! - */ -static void radeon_vm_free_pt(struct radeon_device *rdev, - struct radeon_vm *vm) -{ - struct radeon_bo_va *bo_va; - int i; - - if (!vm->page_directory) - return; - - list_del_init(&vm->list); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } - - if (vm->page_tables == NULL) - return; - - for (i = 0; i < radeon_vm_num_pdes(rdev); i++) - radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence); - - kfree(vm->page_tables); -} - -/** - * radeon_vm_manager_fini - tear down the vm manager - * - * @rdev: radeon_device pointer - * - * Tear down the VM manager (cayman+). - */ -void radeon_vm_manager_fini(struct radeon_device *rdev) -{ - struct radeon_vm *vm, *tmp; - int i; - - if (!rdev->vm_manager.enabled) - return; - - mutex_lock(&rdev->vm_manager.lock); - /* free all allocated page tables */ - list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) { - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&vm->mutex); - } - for (i = 0; i < RADEON_NUM_VM; ++i) { - radeon_fence_unref(&rdev->vm_manager.active[i]); - } - radeon_asic_vm_fini(rdev); - mutex_unlock(&rdev->vm_manager.lock); - - radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager); - radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager); - rdev->vm_manager.enabled = false; -} - -/** - * radeon_vm_evict - evict page table to make room for new one - * - * @rdev: radeon_device pointer - * @vm: VM we want to allocate something for - * - * Evict a VM from the lru, making sure that it isn't @vm. (cayman+). - * Returns 0 for success, -ENOMEM for failure. - * - * Global and local mutex must be locked! - */ -static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm) -{ - struct radeon_vm *vm_evict; - - if (list_empty(&rdev->vm_manager.lru_vm)) - return -ENOMEM; - - vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, - struct radeon_vm, list); - if (vm_evict == vm) - return -ENOMEM; - - mutex_lock(&vm_evict->mutex); - radeon_vm_free_pt(rdev, vm_evict); - mutex_unlock(&vm_evict->mutex); - return 0; -} - -/** - * radeon_vm_alloc_pt - allocates a page table for a VM - * - * @rdev: radeon_device pointer - * @vm: vm to bind - * - * Allocate a page table for the requested vm (cayman+). - * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! - */ -int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm) -{ - unsigned pd_size, pd_entries, pts_size; - struct radeon_ib ib; - int r; - - if (vm == NULL) { - return -EINVAL; - } - - if (vm->page_directory != NULL) { - return 0; - } - - pd_size = radeon_vm_directory_size(rdev); - pd_entries = radeon_vm_num_pdes(rdev); - -retry: - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_directory, pd_size, - RADEON_VM_PTB_ALIGN_SIZE, false); - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - - } else if (r) { - return r; - } - - vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory); - - /* Initially clear the page directory */ - r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, - NULL, pd_entries * 2 + 64); - if (r) { - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return r; - } - - ib.length_dw = 0; - - radeon_asic_vm_set_page(rdev, &ib, vm->pd_gpu_addr, - 0, pd_entries, 0, 0); - - radeon_semaphore_sync_to(ib.semaphore, vm->fence); - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return r; - } - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(ib.fence); - radeon_ib_free(rdev, &ib); - radeon_fence_unref(&vm->last_flush); - - /* allocate page table array */ - pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *); - vm->page_tables = kzalloc(pts_size, GFP_KERNEL); - - if (vm->page_tables == NULL) { - DRM_ERROR("Cannot allocate memory for page table array\n"); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return -ENOMEM; - } - - return 0; -} - -/** - * radeon_vm_add_to_lru - add VMs page table to LRU list - * - * @rdev: radeon_device pointer - * @vm: vm to add to LRU - * - * Add the allocated page table to the LRU list (cayman+). - * - * Global mutex must be locked! - */ -void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm) -{ - list_del_init(&vm->list); - list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); -} - -/** - * radeon_vm_grab_id - allocate the next free VMID - * - * @rdev: radeon_device pointer - * @vm: vm to allocate id for - * @ring: ring we want to submit job to - * - * Allocate an id for the vm (cayman+). - * Returns the fence we need to sync to (if any). - * - * Global and local mutex must be locked! - */ -struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, - struct radeon_vm *vm, int ring) -{ - struct radeon_fence *best[RADEON_NUM_RINGS] = {}; - unsigned choices[2] = {}; - unsigned i; - - /* check if the id is still valid */ - if (vm->last_id_use && vm->last_id_use == rdev->vm_manager.active[vm->id]) - return NULL; - - /* we definately need to flush */ - radeon_fence_unref(&vm->last_flush); - - /* skip over VMID 0, since it is the system VM */ - for (i = 1; i < rdev->vm_manager.nvm; ++i) { - struct radeon_fence *fence = rdev->vm_manager.active[i]; - - if (fence == NULL) { - /* found a free one */ - vm->id = i; - trace_radeon_vm_grab_id(vm->id, ring); - return NULL; - } - - if (radeon_fence_is_earlier(fence, best[fence->ring])) { - best[fence->ring] = fence; - choices[fence->ring == ring ? 0 : 1] = i; - } - } - - for (i = 0; i < 2; ++i) { - if (choices[i]) { - vm->id = choices[i]; - trace_radeon_vm_grab_id(vm->id, ring); - return rdev->vm_manager.active[choices[i]]; - } - } - - /* should never happen */ - BUG(); - return NULL; -} - -/** - * radeon_vm_fence - remember fence for vm - * - * @rdev: radeon_device pointer - * @vm: vm we want to fence - * @fence: fence to remember - * - * Fence the vm (cayman+). - * Set the fence used to protect page table and id. - * - * Global and local mutex must be locked! - */ -void radeon_vm_fence(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_fence *fence) -{ - radeon_fence_unref(&rdev->vm_manager.active[vm->id]); - rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); - - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(fence); - - radeon_fence_unref(&vm->last_id_use); - vm->last_id_use = radeon_fence_ref(fence); -} - -/** - * radeon_vm_bo_find - find the bo_va for a specific vm & bo - * - * @vm: requested vm - * @bo: requested buffer object - * - * Find @bo inside the requested vm (cayman+). - * Search inside the @bos vm list for the requested vm - * Returns the found bo_va or NULL if none is found - * - * Object has to be reserved! - */ -struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - list_for_each_entry(bo_va, &bo->va, bo_list) { - if (bo_va->vm == vm) { - return bo_va; - } - } - return NULL; -} - -/** - * radeon_vm_bo_add - add a bo to a specific vm - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * - * Add @bo into the requested vm (cayman+). - * Add @bo to the list of bos associated with the vm - * Returns newly added bo_va or NULL for failure - * - * Object has to be reserved! - */ -struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); - if (bo_va == NULL) { - return NULL; - } - bo_va->vm = vm; - bo_va->bo = bo; - bo_va->soffset = 0; - bo_va->eoffset = 0; - bo_va->flags = 0; - bo_va->valid = false; - bo_va->ref_count = 1; - INIT_LIST_HEAD(&bo_va->bo_list); - INIT_LIST_HEAD(&bo_va->vm_list); - - mutex_lock(&vm->mutex); - list_add(&bo_va->vm_list, &vm->va); - list_add_tail(&bo_va->bo_list, &bo->va); - mutex_unlock(&vm->mutex); - - return bo_va; -} - -/** - * radeon_vm_bo_set_addr - set bos virtual address inside a vm - * - * @rdev: radeon_device pointer - * @bo_va: bo_va to store the address - * @soffset: requested offset of the buffer in the VM address space - * @flags: attributes of pages (read/write/valid/etc.) - * - * Set offset of @bo_va (cayman+). - * Validate and set the offset requested within the vm address space. - * Returns 0 for success, error for failure. - * - * Object has to be reserved! - */ -int radeon_vm_bo_set_addr(struct radeon_device *rdev, - struct radeon_bo_va *bo_va, - uint64_t soffset, - uint32_t flags) -{ - uint64_t size = radeon_bo_size(bo_va->bo); - uint64_t eoffset, last_offset = 0; - struct radeon_vm *vm = bo_va->vm; - struct radeon_bo_va *tmp; - struct list_head *head; - unsigned last_pfn; - - if (soffset) { - /* make sure object fit at this offset */ - eoffset = soffset + size; - if (soffset >= eoffset) { - return -EINVAL; - } - - last_pfn = eoffset / RADEON_GPU_PAGE_SIZE; - if (last_pfn > rdev->vm_manager.max_pfn) { - dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n", - last_pfn, rdev->vm_manager.max_pfn); - return -EINVAL; - } - - } else { - eoffset = last_pfn = 0; - } - - mutex_lock(&vm->mutex); - head = &vm->va; - last_offset = 0; - list_for_each_entry(tmp, &vm->va, vm_list) { - if (bo_va == tmp) { - /* skip over currently modified bo */ - continue; - } - - if (soffset >= last_offset && eoffset <= tmp->soffset) { - /* bo can be added before this one */ - break; - } - if (eoffset > tmp->soffset && soffset < tmp->eoffset) { - /* bo and tmp overlap, invalid offset */ - dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n", - bo_va->bo, (unsigned)bo_va->soffset, tmp->bo, - (unsigned)tmp->soffset, (unsigned)tmp->eoffset); - mutex_unlock(&vm->mutex); - return -EINVAL; - } - last_offset = tmp->eoffset; - head = &tmp->vm_list; - } - - bo_va->soffset = soffset; - bo_va->eoffset = eoffset; - bo_va->flags = flags; - bo_va->valid = false; - list_move(&bo_va->vm_list, head); - - mutex_unlock(&vm->mutex); - return 0; -} - -/** - * radeon_vm_map_gart - get the physical address of a gart page - * - * @rdev: radeon_device pointer - * @addr: the unmapped addr - * - * Look up the physical address of the page that the pte resolves - * to (cayman+). - * Returns the physical address of the page. - */ -uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr) -{ - uint64_t result; - - /* page table offset */ - result = rdev->gart.pages_addr[addr >> PAGE_SHIFT]; - - /* in case cpu page size != gpu page size*/ - result |= addr & (~PAGE_MASK); - - return result; -} - -/** - * radeon_vm_page_flags - translate page flags to what the hw uses - * - * @flags: flags comming from userspace - * - * Translate the flags the userspace ABI uses to hw flags. - */ -static uint32_t radeon_vm_page_flags(uint32_t flags) -{ - uint32_t hw_flags = 0; - hw_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0; - hw_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0; - hw_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0; - if (flags & RADEON_VM_PAGE_SYSTEM) { - hw_flags |= R600_PTE_SYSTEM; - hw_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0; - } - return hw_flags; -} - -/** - * radeon_vm_update_pdes - make sure that page directory is valid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @start: start of GPU address range - * @end: end of GPU address range - * - * Allocates new page tables if necessary - * and updates the page directory (cayman+). - * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! - */ -static int radeon_vm_update_pdes(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_ib *ib, - uint64_t start, uint64_t end) -{ - static const uint32_t incr = RADEON_VM_PTE_COUNT * 8; - - uint64_t last_pde = ~0, last_pt = ~0; - unsigned count = 0; - uint64_t pt_idx; - int r; - - start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; - end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; - - /* walk over the address space and update the page directory */ - for (pt_idx = start; pt_idx <= end; ++pt_idx) { - uint64_t pde, pt; - - if (vm->page_tables[pt_idx]) - continue; - -retry: - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_tables[pt_idx], - RADEON_VM_PTE_COUNT * 8, - RADEON_GPU_PAGE_SIZE, false); - - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - } else if (r) { - return r; - } - - pde = vm->pd_gpu_addr + pt_idx * 8; - - pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); - - if (((last_pde + 8 * count) != pde) || - ((last_pt + incr * count) != pt)) { - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, - last_pt, count, incr, - R600_PTE_VALID); - - count *= RADEON_VM_PTE_COUNT; - radeon_asic_vm_set_page(rdev, ib, last_pt, 0, - count, 0, 0); - } - - count = 1; - last_pde = pde; - last_pt = pt; - } else { - ++count; - } - } - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count, - incr, R600_PTE_VALID); - - count *= RADEON_VM_PTE_COUNT; - radeon_asic_vm_set_page(rdev, ib, last_pt, 0, - count, 0, 0); - } - - return 0; -} - -/** - * radeon_vm_update_ptes - make sure that page tables are valid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @start: start of GPU address range - * @end: end of GPU address range - * @dst: destination address to map to - * @flags: mapping flags - * - * Update the page tables in the range @start - @end (cayman+). - * - * Global and local mutex must be locked! - */ -static void radeon_vm_update_ptes(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_ib *ib, - uint64_t start, uint64_t end, - uint64_t dst, uint32_t flags) -{ - static const uint64_t mask = RADEON_VM_PTE_COUNT - 1; - - uint64_t last_pte = ~0, last_dst = ~0; - unsigned count = 0; - uint64_t addr; - - start = start / RADEON_GPU_PAGE_SIZE; - end = end / RADEON_GPU_PAGE_SIZE; - - /* walk over the address space and update the page tables */ - for (addr = start; addr < end; ) { - uint64_t pt_idx = addr >> RADEON_VM_BLOCK_SIZE; - unsigned nptes; - uint64_t pte; - - if ((addr & ~mask) == (end & ~mask)) - nptes = end - addr; - else - nptes = RADEON_VM_PTE_COUNT - (addr & mask); - - pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); - pte += (addr & mask) * 8; - - if ((last_pte + 8 * count) != pte) { - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pte, - last_dst, count, - RADEON_GPU_PAGE_SIZE, - flags); - } - - count = nptes; - last_pte = pte; - last_dst = dst; - } else { - count += nptes; - } - - addr += nptes; - dst += nptes * RADEON_GPU_PAGE_SIZE; - } - - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pte, - last_dst, count, - RADEON_GPU_PAGE_SIZE, flags); - } -} - -/** - * radeon_vm_bo_update - map a bo into the vm page table - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * @mem: ttm mem - * - * Fill in the page table entries for @bo (cayman+). - * Returns 0 for success, -EINVAL for failure. - * - * Object have to be reserved & global and local mutex must be locked! - */ -int radeon_vm_bo_update(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_bo *bo, - struct ttm_mem_reg *mem) -{ - struct radeon_ib ib; - struct radeon_bo_va *bo_va; - unsigned nptes, npdes, ndw; - uint64_t addr; - int r; - - /* nothing to do if vm isn't bound */ - if (vm->page_directory == NULL) - return 0; - - bo_va = radeon_vm_bo_find(vm, bo); - if (bo_va == NULL) { - dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); - return -EINVAL; - } - - if (!bo_va->soffset) { - dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n", - bo, vm); - return -EINVAL; - } - - if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL)) - return 0; - - bo_va->flags &= ~RADEON_VM_PAGE_VALID; - bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM; - if (mem) { - addr = mem->start << PAGE_SHIFT; - if (mem->mem_type != TTM_PL_SYSTEM) { - bo_va->flags |= RADEON_VM_PAGE_VALID; - bo_va->valid = true; - } - if (mem->mem_type == TTM_PL_TT) { - bo_va->flags |= RADEON_VM_PAGE_SYSTEM; - } else { - addr += rdev->vm_manager.vram_base_offset; - } - } else { - addr = 0; - bo_va->valid = false; - } - - trace_radeon_vm_bo_update(bo_va); - - nptes = radeon_bo_ngpu_pages(bo); - - /* assume two extra pdes in case the mapping overlaps the borders */ - npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2; - - /* padding, etc. */ - ndw = 64; - - if (RADEON_VM_BLOCK_SIZE > 11) - /* reserve space for one header for every 2k dwords */ - ndw += (nptes >> 11) * 4; - else - /* reserve space for one header for - every (1 << BLOCK_SIZE) entries */ - ndw += (nptes >> RADEON_VM_BLOCK_SIZE) * 4; - - /* reserve space for pte addresses */ - ndw += nptes * 2; - - /* reserve space for one header for every 2k dwords */ - ndw += (npdes >> 11) * 4; - - /* reserve space for pde addresses */ - ndw += npdes * 2; - - /* reserve space for clearing new page tables */ - ndw += npdes * 2 * RADEON_VM_PTE_COUNT; - - /* update too big for an IB */ - if (ndw > 0xfffff) - return -ENOMEM; - - r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4); - if (r) - return r; - ib.length_dw = 0; - - r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset); - if (r) { - radeon_ib_free(rdev, &ib); - return r; - } - - radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset, - addr, radeon_vm_page_flags(bo_va->flags)); - - radeon_semaphore_sync_to(ib.semaphore, vm->fence); - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - return r; - } - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(ib.fence); - radeon_ib_free(rdev, &ib); - radeon_fence_unref(&vm->last_flush); - - return 0; -} - -/** - * radeon_vm_bo_rmv - remove a bo to a specific vm - * - * @rdev: radeon_device pointer - * @bo_va: requested bo_va - * - * Remove @bo_va->bo from the requested vm (cayman+). - * Remove @bo_va->bo from the list of bos associated with the bo_va->vm and - * remove the ptes for @bo_va in the page table. - * Returns 0 for success. - * - * Object have to be reserved! - */ -int radeon_vm_bo_rmv(struct radeon_device *rdev, - struct radeon_bo_va *bo_va) -{ - int r = 0; - - mutex_lock(&rdev->vm_manager.lock); - mutex_lock(&bo_va->vm->mutex); - if (bo_va->soffset) { - r = radeon_vm_bo_update(rdev, bo_va->vm, bo_va->bo, NULL); - } - mutex_unlock(&rdev->vm_manager.lock); - list_del(&bo_va->vm_list); - mutex_unlock(&bo_va->vm->mutex); - list_del(&bo_va->bo_list); - - kfree(bo_va); - return r; -} - -/** - * radeon_vm_bo_invalidate - mark the bo as invalid - * - * @rdev: radeon_device pointer - * @vm: requested vm - * @bo: radeon buffer object - * - * Mark @bo as invalid (cayman+). - */ -void radeon_vm_bo_invalidate(struct radeon_device *rdev, - struct radeon_bo *bo) -{ - struct radeon_bo_va *bo_va; - - list_for_each_entry(bo_va, &bo->va, bo_list) { - bo_va->valid = false; - } -} - -/** - * radeon_vm_init - initialize a vm instance - * - * @rdev: radeon_device pointer - * @vm: requested vm - * - * Init @vm fields (cayman+). - */ -void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) -{ - vm->id = 0; - vm->fence = NULL; - vm->last_flush = NULL; - vm->last_id_use = NULL; - mutex_init(&vm->mutex); - INIT_LIST_HEAD(&vm->list); - INIT_LIST_HEAD(&vm->va); -} - -/** - * radeon_vm_fini - tear down a vm instance - * - * @rdev: radeon_device pointer - * @vm: requested vm - * - * Tear down @vm (cayman+). - * Unbind the VM and remove all bos from the vm bo list - */ -void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) -{ - struct radeon_bo_va *bo_va, *tmp; - int r; - - mutex_lock(&rdev->vm_manager.lock); - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&rdev->vm_manager.lock); - - if (!list_empty(&vm->va)) { - dev_err(rdev->dev, "still active bo inside vm\n"); - } - list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) { - list_del_init(&bo_va->vm_list); - r = radeon_bo_reserve(bo_va->bo, false); - if (!r) { - list_del_init(&bo_va->bo_list); - radeon_bo_unreserve(bo_va->bo); - kfree(bo_va); - } - } - radeon_fence_unref(&vm->fence); - radeon_fence_unref(&vm->last_flush); - radeon_fence_unref(&vm->last_id_use); - mutex_unlock(&vm->mutex); -} diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c new file mode 100644 index 000000000000..433b1ebd07ea --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -0,0 +1,981 @@ +/* + * Copyright 2008 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ +#include +#include +#include "radeon.h" +#include "radeon_trace.h" + +/* + * GPUVM + * GPUVM is similar to the legacy gart on older asics, however + * rather than there being a single global gart table + * for the entire GPU, there are multiple VM page tables active + * at any given time. The VM page tables can contain a mix + * vram pages and system memory pages and system memory pages + * can be mapped as snooped (cached system pages) or unsnooped + * (uncached system pages). + * Each VM has an ID associated with it and there is a page table + * associated with each VMID. When execting a command buffer, + * the kernel tells the the ring what VMID to use for that command + * buffer. VMIDs are allocated dynamically as commands are submitted. + * The userspace drivers maintain their own address space and the kernel + * sets up their pages tables accordingly when they submit their + * command buffers and a VMID is assigned. + * Cayman/Trinity support up to 8 active VMs at any given time; + * SI supports 16. + */ + +/** + * radeon_vm_num_pde - return the number of page directory entries + * + * @rdev: radeon_device pointer + * + * Calculate the number of page directory entries (cayman+). + */ +static unsigned radeon_vm_num_pdes(struct radeon_device *rdev) +{ + return rdev->vm_manager.max_pfn >> RADEON_VM_BLOCK_SIZE; +} + +/** + * radeon_vm_directory_size - returns the size of the page directory in bytes + * + * @rdev: radeon_device pointer + * + * Calculate the size of the page directory in bytes (cayman+). + */ +static unsigned radeon_vm_directory_size(struct radeon_device *rdev) +{ + return RADEON_GPU_PAGE_ALIGN(radeon_vm_num_pdes(rdev) * 8); +} + +/** + * radeon_vm_manager_init - init the vm manager + * + * @rdev: radeon_device pointer + * + * Init the vm manager (cayman+). + * Returns 0 for success, error for failure. + */ +int radeon_vm_manager_init(struct radeon_device *rdev) +{ + struct radeon_vm *vm; + struct radeon_bo_va *bo_va; + int r; + unsigned size; + + if (!rdev->vm_manager.enabled) { + /* allocate enough for 2 full VM pts */ + size = radeon_vm_directory_size(rdev); + size += rdev->vm_manager.max_pfn * 8; + size *= 2; + r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager, + RADEON_GPU_PAGE_ALIGN(size), + RADEON_VM_PTB_ALIGN_SIZE, + RADEON_GEM_DOMAIN_VRAM); + if (r) { + dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n", + (rdev->vm_manager.max_pfn * 8) >> 10); + return r; + } + + r = radeon_asic_vm_init(rdev); + if (r) + return r; + + rdev->vm_manager.enabled = true; + + r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager); + if (r) + return r; + } + + /* restore page table */ + list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) { + if (vm->page_directory == NULL) + continue; + + list_for_each_entry(bo_va, &vm->va, vm_list) { + bo_va->valid = false; + } + } + return 0; +} + +/** + * radeon_vm_free_pt - free the page table for a specific vm + * + * @rdev: radeon_device pointer + * @vm: vm to unbind + * + * Free the page table of a specific vm (cayman+). + * + * Global and local mutex must be lock! + */ +static void radeon_vm_free_pt(struct radeon_device *rdev, + struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va; + int i; + + if (!vm->page_directory) + return; + + list_del_init(&vm->list); + radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); + + list_for_each_entry(bo_va, &vm->va, vm_list) { + bo_va->valid = false; + } + + if (vm->page_tables == NULL) + return; + + for (i = 0; i < radeon_vm_num_pdes(rdev); i++) + radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence); + + kfree(vm->page_tables); +} + +/** + * radeon_vm_manager_fini - tear down the vm manager + * + * @rdev: radeon_device pointer + * + * Tear down the VM manager (cayman+). + */ +void radeon_vm_manager_fini(struct radeon_device *rdev) +{ + struct radeon_vm *vm, *tmp; + int i; + + if (!rdev->vm_manager.enabled) + return; + + mutex_lock(&rdev->vm_manager.lock); + /* free all allocated page tables */ + list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) { + mutex_lock(&vm->mutex); + radeon_vm_free_pt(rdev, vm); + mutex_unlock(&vm->mutex); + } + for (i = 0; i < RADEON_NUM_VM; ++i) { + radeon_fence_unref(&rdev->vm_manager.active[i]); + } + radeon_asic_vm_fini(rdev); + mutex_unlock(&rdev->vm_manager.lock); + + radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager); + radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager); + rdev->vm_manager.enabled = false; +} + +/** + * radeon_vm_evict - evict page table to make room for new one + * + * @rdev: radeon_device pointer + * @vm: VM we want to allocate something for + * + * Evict a VM from the lru, making sure that it isn't @vm. (cayman+). + * Returns 0 for success, -ENOMEM for failure. + * + * Global and local mutex must be locked! + */ +static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm) +{ + struct radeon_vm *vm_evict; + + if (list_empty(&rdev->vm_manager.lru_vm)) + return -ENOMEM; + + vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, + struct radeon_vm, list); + if (vm_evict == vm) + return -ENOMEM; + + mutex_lock(&vm_evict->mutex); + radeon_vm_free_pt(rdev, vm_evict); + mutex_unlock(&vm_evict->mutex); + return 0; +} + +/** + * radeon_vm_alloc_pt - allocates a page table for a VM + * + * @rdev: radeon_device pointer + * @vm: vm to bind + * + * Allocate a page table for the requested vm (cayman+). + * Returns 0 for success, error for failure. + * + * Global and local mutex must be locked! + */ +int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm) +{ + unsigned pd_size, pd_entries, pts_size; + struct radeon_ib ib; + int r; + + if (vm == NULL) { + return -EINVAL; + } + + if (vm->page_directory != NULL) { + return 0; + } + + pd_size = radeon_vm_directory_size(rdev); + pd_entries = radeon_vm_num_pdes(rdev); + +retry: + r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, + &vm->page_directory, pd_size, + RADEON_VM_PTB_ALIGN_SIZE, false); + if (r == -ENOMEM) { + r = radeon_vm_evict(rdev, vm); + if (r) + return r; + goto retry; + + } else if (r) { + return r; + } + + vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory); + + /* Initially clear the page directory */ + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, + NULL, pd_entries * 2 + 64); + if (r) { + radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); + return r; + } + + ib.length_dw = 0; + + radeon_asic_vm_set_page(rdev, &ib, vm->pd_gpu_addr, + 0, pd_entries, 0, 0); + + radeon_semaphore_sync_to(ib.semaphore, vm->fence); + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); + return r; + } + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(ib.fence); + radeon_ib_free(rdev, &ib); + radeon_fence_unref(&vm->last_flush); + + /* allocate page table array */ + pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *); + vm->page_tables = kzalloc(pts_size, GFP_KERNEL); + + if (vm->page_tables == NULL) { + DRM_ERROR("Cannot allocate memory for page table array\n"); + radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); + return -ENOMEM; + } + + return 0; +} + +/** + * radeon_vm_add_to_lru - add VMs page table to LRU list + * + * @rdev: radeon_device pointer + * @vm: vm to add to LRU + * + * Add the allocated page table to the LRU list (cayman+). + * + * Global mutex must be locked! + */ +void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm) +{ + list_del_init(&vm->list); + list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); +} + +/** + * radeon_vm_grab_id - allocate the next free VMID + * + * @rdev: radeon_device pointer + * @vm: vm to allocate id for + * @ring: ring we want to submit job to + * + * Allocate an id for the vm (cayman+). + * Returns the fence we need to sync to (if any). + * + * Global and local mutex must be locked! + */ +struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, + struct radeon_vm *vm, int ring) +{ + struct radeon_fence *best[RADEON_NUM_RINGS] = {}; + unsigned choices[2] = {}; + unsigned i; + + /* check if the id is still valid */ + if (vm->last_id_use && vm->last_id_use == rdev->vm_manager.active[vm->id]) + return NULL; + + /* we definately need to flush */ + radeon_fence_unref(&vm->last_flush); + + /* skip over VMID 0, since it is the system VM */ + for (i = 1; i < rdev->vm_manager.nvm; ++i) { + struct radeon_fence *fence = rdev->vm_manager.active[i]; + + if (fence == NULL) { + /* found a free one */ + vm->id = i; + trace_radeon_vm_grab_id(vm->id, ring); + return NULL; + } + + if (radeon_fence_is_earlier(fence, best[fence->ring])) { + best[fence->ring] = fence; + choices[fence->ring == ring ? 0 : 1] = i; + } + } + + for (i = 0; i < 2; ++i) { + if (choices[i]) { + vm->id = choices[i]; + trace_radeon_vm_grab_id(vm->id, ring); + return rdev->vm_manager.active[choices[i]]; + } + } + + /* should never happen */ + BUG(); + return NULL; +} + +/** + * radeon_vm_fence - remember fence for vm + * + * @rdev: radeon_device pointer + * @vm: vm we want to fence + * @fence: fence to remember + * + * Fence the vm (cayman+). + * Set the fence used to protect page table and id. + * + * Global and local mutex must be locked! + */ +void radeon_vm_fence(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_fence *fence) +{ + radeon_fence_unref(&rdev->vm_manager.active[vm->id]); + rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); + + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(fence); + + radeon_fence_unref(&vm->last_id_use); + vm->last_id_use = radeon_fence_ref(fence); +} + +/** + * radeon_vm_bo_find - find the bo_va for a specific vm & bo + * + * @vm: requested vm + * @bo: requested buffer object + * + * Find @bo inside the requested vm (cayman+). + * Search inside the @bos vm list for the requested vm + * Returns the found bo_va or NULL if none is found + * + * Object has to be reserved! + */ +struct radeon_bo_va *radeon_vm_bo_find(struct radeon_vm *vm, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + list_for_each_entry(bo_va, &bo->va, bo_list) { + if (bo_va->vm == vm) { + return bo_va; + } + } + return NULL; +} + +/** + * radeon_vm_bo_add - add a bo to a specific vm + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * + * Add @bo into the requested vm (cayman+). + * Add @bo to the list of bos associated with the vm + * Returns newly added bo_va or NULL for failure + * + * Object has to be reserved! + */ +struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); + if (bo_va == NULL) { + return NULL; + } + bo_va->vm = vm; + bo_va->bo = bo; + bo_va->soffset = 0; + bo_va->eoffset = 0; + bo_va->flags = 0; + bo_va->valid = false; + bo_va->ref_count = 1; + INIT_LIST_HEAD(&bo_va->bo_list); + INIT_LIST_HEAD(&bo_va->vm_list); + + mutex_lock(&vm->mutex); + list_add(&bo_va->vm_list, &vm->va); + list_add_tail(&bo_va->bo_list, &bo->va); + mutex_unlock(&vm->mutex); + + return bo_va; +} + +/** + * radeon_vm_bo_set_addr - set bos virtual address inside a vm + * + * @rdev: radeon_device pointer + * @bo_va: bo_va to store the address + * @soffset: requested offset of the buffer in the VM address space + * @flags: attributes of pages (read/write/valid/etc.) + * + * Set offset of @bo_va (cayman+). + * Validate and set the offset requested within the vm address space. + * Returns 0 for success, error for failure. + * + * Object has to be reserved! + */ +int radeon_vm_bo_set_addr(struct radeon_device *rdev, + struct radeon_bo_va *bo_va, + uint64_t soffset, + uint32_t flags) +{ + uint64_t size = radeon_bo_size(bo_va->bo); + uint64_t eoffset, last_offset = 0; + struct radeon_vm *vm = bo_va->vm; + struct radeon_bo_va *tmp; + struct list_head *head; + unsigned last_pfn; + + if (soffset) { + /* make sure object fit at this offset */ + eoffset = soffset + size; + if (soffset >= eoffset) { + return -EINVAL; + } + + last_pfn = eoffset / RADEON_GPU_PAGE_SIZE; + if (last_pfn > rdev->vm_manager.max_pfn) { + dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n", + last_pfn, rdev->vm_manager.max_pfn); + return -EINVAL; + } + + } else { + eoffset = last_pfn = 0; + } + + mutex_lock(&vm->mutex); + head = &vm->va; + last_offset = 0; + list_for_each_entry(tmp, &vm->va, vm_list) { + if (bo_va == tmp) { + /* skip over currently modified bo */ + continue; + } + + if (soffset >= last_offset && eoffset <= tmp->soffset) { + /* bo can be added before this one */ + break; + } + if (eoffset > tmp->soffset && soffset < tmp->eoffset) { + /* bo and tmp overlap, invalid offset */ + dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n", + bo_va->bo, (unsigned)bo_va->soffset, tmp->bo, + (unsigned)tmp->soffset, (unsigned)tmp->eoffset); + mutex_unlock(&vm->mutex); + return -EINVAL; + } + last_offset = tmp->eoffset; + head = &tmp->vm_list; + } + + bo_va->soffset = soffset; + bo_va->eoffset = eoffset; + bo_va->flags = flags; + bo_va->valid = false; + list_move(&bo_va->vm_list, head); + + mutex_unlock(&vm->mutex); + return 0; +} + +/** + * radeon_vm_map_gart - get the physical address of a gart page + * + * @rdev: radeon_device pointer + * @addr: the unmapped addr + * + * Look up the physical address of the page that the pte resolves + * to (cayman+). + * Returns the physical address of the page. + */ +uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr) +{ + uint64_t result; + + /* page table offset */ + result = rdev->gart.pages_addr[addr >> PAGE_SHIFT]; + + /* in case cpu page size != gpu page size*/ + result |= addr & (~PAGE_MASK); + + return result; +} + +/** + * radeon_vm_page_flags - translate page flags to what the hw uses + * + * @flags: flags comming from userspace + * + * Translate the flags the userspace ABI uses to hw flags. + */ +static uint32_t radeon_vm_page_flags(uint32_t flags) +{ + uint32_t hw_flags = 0; + hw_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0; + hw_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0; + hw_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0; + if (flags & RADEON_VM_PAGE_SYSTEM) { + hw_flags |= R600_PTE_SYSTEM; + hw_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0; + } + return hw_flags; +} + +/** + * radeon_vm_update_pdes - make sure that page directory is valid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @start: start of GPU address range + * @end: end of GPU address range + * + * Allocates new page tables if necessary + * and updates the page directory (cayman+). + * Returns 0 for success, error for failure. + * + * Global and local mutex must be locked! + */ +static int radeon_vm_update_pdes(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_ib *ib, + uint64_t start, uint64_t end) +{ + static const uint32_t incr = RADEON_VM_PTE_COUNT * 8; + + uint64_t last_pde = ~0, last_pt = ~0; + unsigned count = 0; + uint64_t pt_idx; + int r; + + start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; + end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; + + /* walk over the address space and update the page directory */ + for (pt_idx = start; pt_idx <= end; ++pt_idx) { + uint64_t pde, pt; + + if (vm->page_tables[pt_idx]) + continue; + +retry: + r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, + &vm->page_tables[pt_idx], + RADEON_VM_PTE_COUNT * 8, + RADEON_GPU_PAGE_SIZE, false); + + if (r == -ENOMEM) { + r = radeon_vm_evict(rdev, vm); + if (r) + return r; + goto retry; + } else if (r) { + return r; + } + + pde = vm->pd_gpu_addr + pt_idx * 8; + + pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); + + if (((last_pde + 8 * count) != pde) || + ((last_pt + incr * count) != pt)) { + + if (count) { + radeon_asic_vm_set_page(rdev, ib, last_pde, + last_pt, count, incr, + R600_PTE_VALID); + + count *= RADEON_VM_PTE_COUNT; + radeon_asic_vm_set_page(rdev, ib, last_pt, 0, + count, 0, 0); + } + + count = 1; + last_pde = pde; + last_pt = pt; + } else { + ++count; + } + } + + if (count) { + radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count, + incr, R600_PTE_VALID); + + count *= RADEON_VM_PTE_COUNT; + radeon_asic_vm_set_page(rdev, ib, last_pt, 0, + count, 0, 0); + } + + return 0; +} + +/** + * radeon_vm_update_ptes - make sure that page tables are valid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @start: start of GPU address range + * @end: end of GPU address range + * @dst: destination address to map to + * @flags: mapping flags + * + * Update the page tables in the range @start - @end (cayman+). + * + * Global and local mutex must be locked! + */ +static void radeon_vm_update_ptes(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_ib *ib, + uint64_t start, uint64_t end, + uint64_t dst, uint32_t flags) +{ + static const uint64_t mask = RADEON_VM_PTE_COUNT - 1; + + uint64_t last_pte = ~0, last_dst = ~0; + unsigned count = 0; + uint64_t addr; + + start = start / RADEON_GPU_PAGE_SIZE; + end = end / RADEON_GPU_PAGE_SIZE; + + /* walk over the address space and update the page tables */ + for (addr = start; addr < end; ) { + uint64_t pt_idx = addr >> RADEON_VM_BLOCK_SIZE; + unsigned nptes; + uint64_t pte; + + if ((addr & ~mask) == (end & ~mask)) + nptes = end - addr; + else + nptes = RADEON_VM_PTE_COUNT - (addr & mask); + + pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); + pte += (addr & mask) * 8; + + if ((last_pte + 8 * count) != pte) { + + if (count) { + radeon_asic_vm_set_page(rdev, ib, last_pte, + last_dst, count, + RADEON_GPU_PAGE_SIZE, + flags); + } + + count = nptes; + last_pte = pte; + last_dst = dst; + } else { + count += nptes; + } + + addr += nptes; + dst += nptes * RADEON_GPU_PAGE_SIZE; + } + + if (count) { + radeon_asic_vm_set_page(rdev, ib, last_pte, + last_dst, count, + RADEON_GPU_PAGE_SIZE, flags); + } +} + +/** + * radeon_vm_bo_update - map a bo into the vm page table + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * @mem: ttm mem + * + * Fill in the page table entries for @bo (cayman+). + * Returns 0 for success, -EINVAL for failure. + * + * Object have to be reserved & global and local mutex must be locked! + */ +int radeon_vm_bo_update(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo, + struct ttm_mem_reg *mem) +{ + struct radeon_ib ib; + struct radeon_bo_va *bo_va; + unsigned nptes, npdes, ndw; + uint64_t addr; + int r; + + /* nothing to do if vm isn't bound */ + if (vm->page_directory == NULL) + return 0; + + bo_va = radeon_vm_bo_find(vm, bo); + if (bo_va == NULL) { + dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); + return -EINVAL; + } + + if (!bo_va->soffset) { + dev_err(rdev->dev, "bo %p don't has a mapping in vm %p\n", + bo, vm); + return -EINVAL; + } + + if ((bo_va->valid && mem) || (!bo_va->valid && mem == NULL)) + return 0; + + bo_va->flags &= ~RADEON_VM_PAGE_VALID; + bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM; + if (mem) { + addr = mem->start << PAGE_SHIFT; + if (mem->mem_type != TTM_PL_SYSTEM) { + bo_va->flags |= RADEON_VM_PAGE_VALID; + bo_va->valid = true; + } + if (mem->mem_type == TTM_PL_TT) { + bo_va->flags |= RADEON_VM_PAGE_SYSTEM; + } else { + addr += rdev->vm_manager.vram_base_offset; + } + } else { + addr = 0; + bo_va->valid = false; + } + + trace_radeon_vm_bo_update(bo_va); + + nptes = radeon_bo_ngpu_pages(bo); + + /* assume two extra pdes in case the mapping overlaps the borders */ + npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2; + + /* padding, etc. */ + ndw = 64; + + if (RADEON_VM_BLOCK_SIZE > 11) + /* reserve space for one header for every 2k dwords */ + ndw += (nptes >> 11) * 4; + else + /* reserve space for one header for + every (1 << BLOCK_SIZE) entries */ + ndw += (nptes >> RADEON_VM_BLOCK_SIZE) * 4; + + /* reserve space for pte addresses */ + ndw += nptes * 2; + + /* reserve space for one header for every 2k dwords */ + ndw += (npdes >> 11) * 4; + + /* reserve space for pde addresses */ + ndw += npdes * 2; + + /* reserve space for clearing new page tables */ + ndw += npdes * 2 * RADEON_VM_PTE_COUNT; + + /* update too big for an IB */ + if (ndw > 0xfffff) + return -ENOMEM; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4); + if (r) + return r; + ib.length_dw = 0; + + r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset); + if (r) { + radeon_ib_free(rdev, &ib); + return r; + } + + radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset, + addr, radeon_vm_page_flags(bo_va->flags)); + + radeon_semaphore_sync_to(ib.semaphore, vm->fence); + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + return r; + } + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(ib.fence); + radeon_ib_free(rdev, &ib); + radeon_fence_unref(&vm->last_flush); + + return 0; +} + +/** + * radeon_vm_bo_rmv - remove a bo to a specific vm + * + * @rdev: radeon_device pointer + * @bo_va: requested bo_va + * + * Remove @bo_va->bo from the requested vm (cayman+). + * Remove @bo_va->bo from the list of bos associated with the bo_va->vm and + * remove the ptes for @bo_va in the page table. + * Returns 0 for success. + * + * Object have to be reserved! + */ +int radeon_vm_bo_rmv(struct radeon_device *rdev, + struct radeon_bo_va *bo_va) +{ + int r = 0; + + mutex_lock(&rdev->vm_manager.lock); + mutex_lock(&bo_va->vm->mutex); + if (bo_va->soffset) { + r = radeon_vm_bo_update(rdev, bo_va->vm, bo_va->bo, NULL); + } + mutex_unlock(&rdev->vm_manager.lock); + list_del(&bo_va->vm_list); + mutex_unlock(&bo_va->vm->mutex); + list_del(&bo_va->bo_list); + + kfree(bo_va); + return r; +} + +/** + * radeon_vm_bo_invalidate - mark the bo as invalid + * + * @rdev: radeon_device pointer + * @vm: requested vm + * @bo: radeon buffer object + * + * Mark @bo as invalid (cayman+). + */ +void radeon_vm_bo_invalidate(struct radeon_device *rdev, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + list_for_each_entry(bo_va, &bo->va, bo_list) { + bo_va->valid = false; + } +} + +/** + * radeon_vm_init - initialize a vm instance + * + * @rdev: radeon_device pointer + * @vm: requested vm + * + * Init @vm fields (cayman+). + */ +void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) +{ + vm->id = 0; + vm->fence = NULL; + vm->last_flush = NULL; + vm->last_id_use = NULL; + mutex_init(&vm->mutex); + INIT_LIST_HEAD(&vm->list); + INIT_LIST_HEAD(&vm->va); +} + +/** + * radeon_vm_fini - tear down a vm instance + * + * @rdev: radeon_device pointer + * @vm: requested vm + * + * Tear down @vm (cayman+). + * Unbind the VM and remove all bos from the vm bo list + */ +void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va, *tmp; + int r; + + mutex_lock(&rdev->vm_manager.lock); + mutex_lock(&vm->mutex); + radeon_vm_free_pt(rdev, vm); + mutex_unlock(&rdev->vm_manager.lock); + + if (!list_empty(&vm->va)) { + dev_err(rdev->dev, "still active bo inside vm\n"); + } + list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) { + list_del_init(&bo_va->vm_list); + r = radeon_bo_reserve(bo_va->bo, false); + if (!r) { + list_del_init(&bo_va->bo_list); + radeon_bo_unreserve(bo_va->bo); + kfree(bo_va); + } + } + radeon_fence_unref(&vm->fence); + radeon_fence_unref(&vm->last_flush); + radeon_fence_unref(&vm->last_id_use); + mutex_unlock(&vm->mutex); +} From fa68834342e992dcb58e6dd7d9dbe05c6b6b27d9 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 10:47:05 +0100 Subject: [PATCH 060/259] drm/radeon: further cleanup vm flushing & fencing MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 3 +++ drivers/gpu/drm/radeon/radeon_cs.c | 4 ---- drivers/gpu/drm/radeon/radeon_ring.c | 16 +++++++------- drivers/gpu/drm/radeon/radeon_vm.c | 31 +++++++++++++++++++++++++--- 4 files changed, 38 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 40ab8a28c998..644d922990ef 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -2804,6 +2804,9 @@ int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm); void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm); struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, struct radeon_vm *vm, int ring); +void radeon_vm_flush(struct radeon_device *rdev, + struct radeon_vm *vm, + int ring); void radeon_vm_fence(struct radeon_device *rdev, struct radeon_vm *vm, struct radeon_fence *fence); diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 5abae403ea4f..f92df2e8ebdd 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -511,10 +511,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, r = radeon_ib_schedule(rdev, &parser->ib, NULL); } - if (!r) { - radeon_vm_fence(rdev, vm, parser->ib.fence); - } - out: radeon_vm_add_to_lru(rdev, vm); mutex_unlock(&vm->mutex); diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index fa140119cdb6..665591a7faad 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -153,11 +153,9 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, return r; } - /* if we can't remember our last VM flush then flush now! */ - /* XXX figure out why we have to flush for every IB */ - if (ib->vm /*&& !ib->vm->last_flush*/) { - radeon_ring_vm_flush(rdev, ib->ring, ib->vm); - } + if (ib->vm) + radeon_vm_flush(rdev, ib->vm, ib->ring); + if (const_ib) { radeon_ring_ib_execute(rdev, const_ib->ring, const_ib); radeon_semaphore_free(rdev, &const_ib->semaphore, NULL); @@ -172,10 +170,10 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, if (const_ib) { const_ib->fence = radeon_fence_ref(ib->fence); } - /* we just flushed the VM, remember that */ - if (ib->vm && !ib->vm->last_flush) { - ib->vm->last_flush = radeon_fence_ref(ib->fence); - } + + if (ib->vm) + radeon_vm_fence(rdev, ib->vm, ib->fence); + radeon_ring_unlock_commit(rdev, ring); return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index 433b1ebd07ea..516017689793 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -378,6 +378,27 @@ struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, return NULL; } +/** + * radeon_vm_flush - hardware flush the vm + * + * @rdev: radeon_device pointer + * @vm: vm we want to flush + * @ring: ring to use for flush + * + * Flush the vm (cayman+). + * + * Global and local mutex must be locked! + */ +void radeon_vm_flush(struct radeon_device *rdev, + struct radeon_vm *vm, + int ring) +{ + /* if we can't remember our last VM flush then flush now! */ + /* XXX figure out why we have to flush all the time */ + if (!vm->last_flush || true) + radeon_ring_vm_flush(rdev, ring, vm); +} + /** * radeon_vm_fence - remember fence for vm * @@ -394,14 +415,18 @@ void radeon_vm_fence(struct radeon_device *rdev, struct radeon_vm *vm, struct radeon_fence *fence) { - radeon_fence_unref(&rdev->vm_manager.active[vm->id]); - rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); - radeon_fence_unref(&vm->fence); vm->fence = radeon_fence_ref(fence); + radeon_fence_unref(&rdev->vm_manager.active[vm->id]); + rdev->vm_manager.active[vm->id] = radeon_fence_ref(fence); + radeon_fence_unref(&vm->last_id_use); vm->last_id_use = radeon_fence_ref(fence); + + /* we just flushed the VM, remember that */ + if (!vm->last_flush) + vm->last_flush = radeon_fence_ref(fence); } /** From 6d2f2944e95e504a7d33385eeeb9bb7fcca72592 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 13:42:17 +0100 Subject: [PATCH 061/259] drm/radeon: use normal BOs for the page tables v4 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit No need to make it more complicated than necessary, just allocate the page tables as normal BO and flush whenever the address change. v2: update comments and function name v3: squash bug fixes, page directory and tables patch v4: rebased on Mareks changes Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon.h | 23 +- drivers/gpu/drm/radeon/radeon_cs.c | 44 ++- drivers/gpu/drm/radeon/radeon_device.c | 1 - drivers/gpu/drm/radeon/radeon_kms.c | 4 +- drivers/gpu/drm/radeon/radeon_vm.c | 492 +++++++++++-------------- 5 files changed, 269 insertions(+), 295 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 644d922990ef..c31e3c2c1b3a 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -857,17 +857,22 @@ struct radeon_mec { #define R600_PTE_READABLE (1 << 5) #define R600_PTE_WRITEABLE (1 << 6) +struct radeon_vm_pt { + struct radeon_bo *bo; + uint64_t addr; +}; + struct radeon_vm { - struct list_head list; struct list_head va; unsigned id; /* contains the page directory */ - struct radeon_sa_bo *page_directory; + struct radeon_bo *page_directory; uint64_t pd_gpu_addr; + unsigned max_pde_used; /* array of page tables, one for each page directory entry */ - struct radeon_sa_bo **page_tables; + struct radeon_vm_pt *page_tables; struct mutex mutex; /* last fence for cs using this vm */ @@ -880,9 +885,7 @@ struct radeon_vm { struct radeon_vm_manager { struct mutex lock; - struct list_head lru_vm; struct radeon_fence *active[RADEON_NUM_VM]; - struct radeon_sa_manager sa_manager; uint32_t max_pfn; /* number of VMIDs */ unsigned nvm; @@ -1011,6 +1014,7 @@ struct radeon_cs_parser { unsigned nrelocs; struct radeon_cs_reloc *relocs; struct radeon_cs_reloc **relocs_ptr; + struct radeon_bo_list *vm_bos; struct list_head validated; unsigned dma_reloc_idx; /* indices of various chunks */ @@ -2798,10 +2802,11 @@ extern void radeon_program_register_sequence(struct radeon_device *rdev, */ int radeon_vm_manager_init(struct radeon_device *rdev); void radeon_vm_manager_fini(struct radeon_device *rdev); -void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm); -int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm); -void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm); +struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head); struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, struct radeon_vm *vm, int ring); void radeon_vm_flush(struct radeon_device *rdev, @@ -2811,6 +2816,8 @@ void radeon_vm_fence(struct radeon_device *rdev, struct radeon_vm *vm, struct radeon_fence *fence); uint64_t radeon_vm_map_gart(struct radeon_device *rdev, uint64_t addr); +int radeon_vm_update_page_directory(struct radeon_device *rdev, + struct radeon_vm *vm); int radeon_vm_bo_update(struct radeon_device *rdev, struct radeon_vm *vm, struct radeon_bo *bo, diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index f92df2e8ebdd..420c28dd6a1c 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -168,6 +168,10 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) radeon_cs_buckets_get_list(&buckets, &p->validated); + if (p->cs_flags & RADEON_CS_USE_VM) + p->vm_bos = radeon_vm_get_bos(p->rdev, p->ib.vm, + &p->validated); + return radeon_bo_list_validate(p->rdev, &p->ticket, &p->validated, p->ring); } @@ -401,6 +405,7 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bo kfree(parser->track); kfree(parser->relocs); kfree(parser->relocs_ptr); + kfree(parser->vm_bos); for (i = 0; i < parser->nchunks; i++) drm_free_large(parser->chunks[i].kdata); kfree(parser->chunks); @@ -440,24 +445,32 @@ static int radeon_cs_ib_chunk(struct radeon_device *rdev, return r; } -static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser, +static int radeon_bo_vm_update_pte(struct radeon_cs_parser *p, struct radeon_vm *vm) { - struct radeon_device *rdev = parser->rdev; - struct radeon_bo_list *lobj; - struct radeon_bo *bo; - int r; + struct radeon_device *rdev = p->rdev; + int i, r; - r = radeon_vm_bo_update(rdev, vm, rdev->ring_tmp_bo.bo, &rdev->ring_tmp_bo.bo->tbo.mem); - if (r) { + r = radeon_vm_update_page_directory(rdev, vm); + if (r) return r; - } - list_for_each_entry(lobj, &parser->validated, tv.head) { - bo = lobj->bo; - r = radeon_vm_bo_update(parser->rdev, vm, bo, &bo->tbo.mem); - if (r) { + + r = radeon_vm_bo_update(rdev, vm, rdev->ring_tmp_bo.bo, + &rdev->ring_tmp_bo.bo->tbo.mem); + if (r) + return r; + + for (i = 0; i < p->nrelocs; i++) { + struct radeon_bo *bo; + + /* ignore duplicates */ + if (p->relocs_ptr[i] != &p->relocs[i]) + continue; + + bo = p->relocs[i].robj; + r = radeon_vm_bo_update(rdev, vm, bo, &bo->tbo.mem); + if (r) return r; - } } return 0; } @@ -491,10 +504,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, mutex_lock(&rdev->vm_manager.lock); mutex_lock(&vm->mutex); - r = radeon_vm_alloc_pt(rdev, vm); - if (r) { - goto out; - } r = radeon_bo_vm_update_pte(parser, vm); if (r) { goto out; @@ -512,7 +521,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, } out: - radeon_vm_add_to_lru(rdev, vm); mutex_unlock(&vm->mutex); mutex_unlock(&rdev->vm_manager.lock); return r; diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index fa7841b9d7b6..e58dbabd7d7d 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -1198,7 +1198,6 @@ int radeon_device_init(struct radeon_device *rdev, * Max GPUVM size for cayman and SI is 40 bits. */ rdev->vm_manager.max_pfn = 1 << 20; - INIT_LIST_HEAD(&rdev->vm_manager.lru_vm); /* Set asic functions */ r = radeon_asic_init(rdev); diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 7a810d0796a0..37b1deacb5b1 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -559,7 +559,9 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) return -ENOMEM; } - radeon_vm_init(rdev, &fpriv->vm); + r = radeon_vm_init(rdev, &fpriv->vm); + if (r) + return r; r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); if (r) diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index 516017689793..44b6918a87ba 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -84,84 +84,18 @@ static unsigned radeon_vm_directory_size(struct radeon_device *rdev) */ int radeon_vm_manager_init(struct radeon_device *rdev) { - struct radeon_vm *vm; - struct radeon_bo_va *bo_va; int r; - unsigned size; if (!rdev->vm_manager.enabled) { - /* allocate enough for 2 full VM pts */ - size = radeon_vm_directory_size(rdev); - size += rdev->vm_manager.max_pfn * 8; - size *= 2; - r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager, - RADEON_GPU_PAGE_ALIGN(size), - RADEON_VM_PTB_ALIGN_SIZE, - RADEON_GEM_DOMAIN_VRAM); - if (r) { - dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n", - (rdev->vm_manager.max_pfn * 8) >> 10); - return r; - } - r = radeon_asic_vm_init(rdev); if (r) return r; rdev->vm_manager.enabled = true; - - r = radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager); - if (r) - return r; - } - - /* restore page table */ - list_for_each_entry(vm, &rdev->vm_manager.lru_vm, list) { - if (vm->page_directory == NULL) - continue; - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } } return 0; } -/** - * radeon_vm_free_pt - free the page table for a specific vm - * - * @rdev: radeon_device pointer - * @vm: vm to unbind - * - * Free the page table of a specific vm (cayman+). - * - * Global and local mutex must be lock! - */ -static void radeon_vm_free_pt(struct radeon_device *rdev, - struct radeon_vm *vm) -{ - struct radeon_bo_va *bo_va; - int i; - - if (!vm->page_directory) - return; - - list_del_init(&vm->list); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - - list_for_each_entry(bo_va, &vm->va, vm_list) { - bo_va->valid = false; - } - - if (vm->page_tables == NULL) - return; - - for (i = 0; i < radeon_vm_num_pdes(rdev); i++) - radeon_sa_bo_free(rdev, &vm->page_tables[i], vm->fence); - - kfree(vm->page_tables); -} - /** * radeon_vm_manager_fini - tear down the vm manager * @@ -171,155 +105,59 @@ static void radeon_vm_free_pt(struct radeon_device *rdev, */ void radeon_vm_manager_fini(struct radeon_device *rdev) { - struct radeon_vm *vm, *tmp; int i; if (!rdev->vm_manager.enabled) return; mutex_lock(&rdev->vm_manager.lock); - /* free all allocated page tables */ - list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) { - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&vm->mutex); - } - for (i = 0; i < RADEON_NUM_VM; ++i) { + for (i = 0; i < RADEON_NUM_VM; ++i) radeon_fence_unref(&rdev->vm_manager.active[i]); - } radeon_asic_vm_fini(rdev); - mutex_unlock(&rdev->vm_manager.lock); - - radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager); - radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager); rdev->vm_manager.enabled = false; + mutex_unlock(&rdev->vm_manager.lock); } /** - * radeon_vm_evict - evict page table to make room for new one + * radeon_vm_get_bos - add the vm BOs to a validation list * - * @rdev: radeon_device pointer - * @vm: VM we want to allocate something for + * @vm: vm providing the BOs + * @head: head of validation list * - * Evict a VM from the lru, making sure that it isn't @vm. (cayman+). - * Returns 0 for success, -ENOMEM for failure. - * - * Global and local mutex must be locked! + * Add the page directory to the list of BOs to + * validate for command submission (cayman+). */ -static int radeon_vm_evict(struct radeon_device *rdev, struct radeon_vm *vm) +struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head) { - struct radeon_vm *vm_evict; + struct radeon_bo_list *list; + unsigned i, idx, size; - if (list_empty(&rdev->vm_manager.lru_vm)) - return -ENOMEM; + size = (radeon_vm_num_pdes(rdev) + 1) * sizeof(struct radeon_bo_list); + list = kmalloc(size, GFP_KERNEL); + if (!list) + return NULL; - vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, - struct radeon_vm, list); - if (vm_evict == vm) - return -ENOMEM; + /* add the vm page table to the list */ + list[0].bo = vm->page_directory; + list[0].domain = RADEON_GEM_DOMAIN_VRAM; + list[0].alt_domain = RADEON_GEM_DOMAIN_VRAM; + list[0].tv.bo = &vm->page_directory->tbo; + list_add(&list[0].tv.head, head); - mutex_lock(&vm_evict->mutex); - radeon_vm_free_pt(rdev, vm_evict); - mutex_unlock(&vm_evict->mutex); - return 0; -} + for (i = 0, idx = 1; i <= vm->max_pde_used; i++) { + if (!vm->page_tables[i].bo) + continue; -/** - * radeon_vm_alloc_pt - allocates a page table for a VM - * - * @rdev: radeon_device pointer - * @vm: vm to bind - * - * Allocate a page table for the requested vm (cayman+). - * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! - */ -int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm) -{ - unsigned pd_size, pd_entries, pts_size; - struct radeon_ib ib; - int r; - - if (vm == NULL) { - return -EINVAL; + list[idx].bo = vm->page_tables[i].bo; + list[idx].domain = RADEON_GEM_DOMAIN_VRAM; + list[idx].alt_domain = RADEON_GEM_DOMAIN_VRAM; + list[idx].tv.bo = &list[idx].bo->tbo; + list_add(&list[idx++].tv.head, head); } - if (vm->page_directory != NULL) { - return 0; - } - - pd_size = radeon_vm_directory_size(rdev); - pd_entries = radeon_vm_num_pdes(rdev); - -retry: - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_directory, pd_size, - RADEON_VM_PTB_ALIGN_SIZE, false); - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - - } else if (r) { - return r; - } - - vm->pd_gpu_addr = radeon_sa_bo_gpu_addr(vm->page_directory); - - /* Initially clear the page directory */ - r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, - NULL, pd_entries * 2 + 64); - if (r) { - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return r; - } - - ib.length_dw = 0; - - radeon_asic_vm_set_page(rdev, &ib, vm->pd_gpu_addr, - 0, pd_entries, 0, 0); - - radeon_semaphore_sync_to(ib.semaphore, vm->fence); - r = radeon_ib_schedule(rdev, &ib, NULL); - if (r) { - radeon_ib_free(rdev, &ib); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return r; - } - radeon_fence_unref(&vm->fence); - vm->fence = radeon_fence_ref(ib.fence); - radeon_ib_free(rdev, &ib); - radeon_fence_unref(&vm->last_flush); - - /* allocate page table array */ - pts_size = radeon_vm_num_pdes(rdev) * sizeof(struct radeon_sa_bo *); - vm->page_tables = kzalloc(pts_size, GFP_KERNEL); - - if (vm->page_tables == NULL) { - DRM_ERROR("Cannot allocate memory for page table array\n"); - radeon_sa_bo_free(rdev, &vm->page_directory, vm->fence); - return -ENOMEM; - } - - return 0; -} - -/** - * radeon_vm_add_to_lru - add VMs page table to LRU list - * - * @rdev: radeon_device pointer - * @vm: vm to add to LRU - * - * Add the allocated page table to the LRU list (cayman+). - * - * Global mutex must be locked! - */ -void radeon_vm_add_to_lru(struct radeon_device *rdev, struct radeon_vm *vm) -{ - list_del_init(&vm->list); - list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); + return list; } /** @@ -393,10 +231,14 @@ void radeon_vm_flush(struct radeon_device *rdev, struct radeon_vm *vm, int ring) { + uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory); + /* if we can't remember our last VM flush then flush now! */ /* XXX figure out why we have to flush all the time */ - if (!vm->last_flush || true) + if (!vm->last_flush || true || pd_addr != vm->pd_gpu_addr) { + vm->pd_gpu_addr = pd_addr; radeon_ring_vm_flush(rdev, ring, vm); + } } /** @@ -495,6 +337,63 @@ struct radeon_bo_va *radeon_vm_bo_add(struct radeon_device *rdev, return bo_va; } +/** + * radeon_vm_clear_bo - initially clear the page dir/table + * + * @rdev: radeon_device pointer + * @bo: bo to clear + */ +static int radeon_vm_clear_bo(struct radeon_device *rdev, + struct radeon_bo *bo) +{ + struct ttm_validate_buffer tv; + struct ww_acquire_ctx ticket; + struct list_head head; + struct radeon_ib ib; + unsigned entries; + uint64_t addr; + int r; + + memset(&tv, 0, sizeof(tv)); + tv.bo = &bo->tbo; + + INIT_LIST_HEAD(&head); + list_add(&tv.head, &head); + + r = ttm_eu_reserve_buffers(&ticket, &head); + if (r) + return r; + + r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + if (r) + goto error; + + addr = radeon_bo_gpu_offset(bo); + entries = radeon_bo_size(bo) / 8; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, + NULL, entries * 2 + 64); + if (r) + goto error; + + ib.length_dw = 0; + + radeon_asic_vm_set_page(rdev, &ib, addr, 0, entries, 0, 0); + + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) + goto error; + + ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence); + radeon_ib_free(rdev, &ib); + + return 0; + +error: + ttm_eu_backoff_reservation(&ticket, &head); + return r; +} + /** * radeon_vm_bo_set_addr - set bos virtual address inside a vm * @@ -519,7 +418,8 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev, struct radeon_vm *vm = bo_va->vm; struct radeon_bo_va *tmp; struct list_head *head; - unsigned last_pfn; + unsigned last_pfn, pt_idx; + int r; if (soffset) { /* make sure object fit at this offset */ @@ -570,8 +470,53 @@ int radeon_vm_bo_set_addr(struct radeon_device *rdev, bo_va->valid = false; list_move(&bo_va->vm_list, head); + soffset = (soffset / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; + eoffset = (eoffset / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; + + if (eoffset > vm->max_pde_used) + vm->max_pde_used = eoffset; + + radeon_bo_unreserve(bo_va->bo); + + /* walk over the address space and allocate the page tables */ + for (pt_idx = soffset; pt_idx <= eoffset; ++pt_idx) { + struct radeon_bo *pt; + + if (vm->page_tables[pt_idx].bo) + continue; + + /* drop mutex to allocate and clear page table */ + mutex_unlock(&vm->mutex); + + r = radeon_bo_create(rdev, RADEON_VM_PTE_COUNT * 8, + RADEON_GPU_PAGE_SIZE, false, + RADEON_GEM_DOMAIN_VRAM, NULL, &pt); + if (r) + return r; + + r = radeon_vm_clear_bo(rdev, pt); + if (r) { + radeon_bo_unref(&pt); + radeon_bo_reserve(bo_va->bo, false); + return r; + } + + /* aquire mutex again */ + mutex_lock(&vm->mutex); + if (vm->page_tables[pt_idx].bo) { + /* someone else allocated the pt in the meantime */ + mutex_unlock(&vm->mutex); + radeon_bo_unref(&pt); + mutex_lock(&vm->mutex); + continue; + } + + vm->page_tables[pt_idx].addr = 0; + vm->page_tables[pt_idx].bo = pt; + } + mutex_unlock(&vm->mutex); - return 0; + return radeon_bo_reserve(bo_va->bo, false); } /** @@ -631,58 +576,53 @@ static uint32_t radeon_vm_page_flags(uint32_t flags) * * Global and local mutex must be locked! */ -static int radeon_vm_update_pdes(struct radeon_device *rdev, - struct radeon_vm *vm, - struct radeon_ib *ib, - uint64_t start, uint64_t end) +int radeon_vm_update_page_directory(struct radeon_device *rdev, + struct radeon_vm *vm) { static const uint32_t incr = RADEON_VM_PTE_COUNT * 8; + uint64_t pd_addr = radeon_bo_gpu_offset(vm->page_directory); uint64_t last_pde = ~0, last_pt = ~0; - unsigned count = 0; - uint64_t pt_idx; + unsigned count = 0, pt_idx, ndw; + struct radeon_ib ib; int r; - start = (start / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; - end = (end / RADEON_GPU_PAGE_SIZE) >> RADEON_VM_BLOCK_SIZE; + /* padding, etc. */ + ndw = 64; + + /* assume the worst case */ + ndw += vm->max_pde_used * 12; + + /* update too big for an IB */ + if (ndw > 0xfffff) + return -ENOMEM; + + r = radeon_ib_get(rdev, R600_RING_TYPE_DMA_INDEX, &ib, NULL, ndw * 4); + if (r) + return r; + ib.length_dw = 0; /* walk over the address space and update the page directory */ - for (pt_idx = start; pt_idx <= end; ++pt_idx) { + for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) { + struct radeon_bo *bo = vm->page_tables[pt_idx].bo; uint64_t pde, pt; - if (vm->page_tables[pt_idx]) + if (bo == NULL) continue; -retry: - r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, - &vm->page_tables[pt_idx], - RADEON_VM_PTE_COUNT * 8, - RADEON_GPU_PAGE_SIZE, false); - - if (r == -ENOMEM) { - r = radeon_vm_evict(rdev, vm); - if (r) - return r; - goto retry; - } else if (r) { - return r; - } - - pde = vm->pd_gpu_addr + pt_idx * 8; - - pt = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); + pt = radeon_bo_gpu_offset(bo); + if (vm->page_tables[pt_idx].addr == pt) + continue; + vm->page_tables[pt_idx].addr = pt; + pde = pd_addr + pt_idx * 8; if (((last_pde + 8 * count) != pde) || ((last_pt + incr * count) != pt)) { if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, + radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count, incr, R600_PTE_VALID); - - count *= RADEON_VM_PTE_COUNT; - radeon_asic_vm_set_page(rdev, ib, last_pt, 0, - count, 0, 0); } count = 1; @@ -693,14 +633,22 @@ static int radeon_vm_update_pdes(struct radeon_device *rdev, } } - if (count) { - radeon_asic_vm_set_page(rdev, ib, last_pde, last_pt, count, + if (count) + radeon_asic_vm_set_page(rdev, &ib, last_pde, last_pt, count, incr, R600_PTE_VALID); - count *= RADEON_VM_PTE_COUNT; - radeon_asic_vm_set_page(rdev, ib, last_pt, 0, - count, 0, 0); + if (ib.length_dw != 0) { + radeon_semaphore_sync_to(ib.semaphore, vm->last_id_use); + r = radeon_ib_schedule(rdev, &ib, NULL); + if (r) { + radeon_ib_free(rdev, &ib); + return r; + } + radeon_fence_unref(&vm->fence); + vm->fence = radeon_fence_ref(ib.fence); + radeon_fence_unref(&vm->last_flush); } + radeon_ib_free(rdev, &ib); return 0; } @@ -745,7 +693,7 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev, else nptes = RADEON_VM_PTE_COUNT - (addr & mask); - pte = radeon_sa_bo_gpu_addr(vm->page_tables[pt_idx]); + pte = radeon_bo_gpu_offset(vm->page_tables[pt_idx].bo); pte += (addr & mask) * 8; if ((last_pte + 8 * count) != pte) { @@ -795,14 +743,10 @@ int radeon_vm_bo_update(struct radeon_device *rdev, { struct radeon_ib ib; struct radeon_bo_va *bo_va; - unsigned nptes, npdes, ndw; + unsigned nptes, ndw; uint64_t addr; int r; - /* nothing to do if vm isn't bound */ - if (vm->page_directory == NULL) - return 0; - bo_va = radeon_vm_bo_find(vm, bo); if (bo_va == NULL) { dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); @@ -840,9 +784,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev, nptes = radeon_bo_ngpu_pages(bo); - /* assume two extra pdes in case the mapping overlaps the borders */ - npdes = (nptes >> RADEON_VM_BLOCK_SIZE) + 2; - /* padding, etc. */ ndw = 64; @@ -857,15 +798,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev, /* reserve space for pte addresses */ ndw += nptes * 2; - /* reserve space for one header for every 2k dwords */ - ndw += (npdes >> 11) * 4; - - /* reserve space for pde addresses */ - ndw += npdes * 2; - - /* reserve space for clearing new page tables */ - ndw += npdes * 2 * RADEON_VM_PTE_COUNT; - /* update too big for an IB */ if (ndw > 0xfffff) return -ENOMEM; @@ -875,12 +807,6 @@ int radeon_vm_bo_update(struct radeon_device *rdev, return r; ib.length_dw = 0; - r = radeon_vm_update_pdes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset); - if (r) { - radeon_ib_free(rdev, &ib); - return r; - } - radeon_vm_update_ptes(rdev, vm, &ib, bo_va->soffset, bo_va->eoffset, addr, radeon_vm_page_flags(bo_va->flags)); @@ -957,15 +883,43 @@ void radeon_vm_bo_invalidate(struct radeon_device *rdev, * * Init @vm fields (cayman+). */ -void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) { + unsigned pd_size, pd_entries, pts_size; + int r; + vm->id = 0; vm->fence = NULL; vm->last_flush = NULL; vm->last_id_use = NULL; mutex_init(&vm->mutex); - INIT_LIST_HEAD(&vm->list); INIT_LIST_HEAD(&vm->va); + + pd_size = radeon_vm_directory_size(rdev); + pd_entries = radeon_vm_num_pdes(rdev); + + /* allocate page table array */ + pts_size = pd_entries * sizeof(struct radeon_vm_pt); + vm->page_tables = kzalloc(pts_size, GFP_KERNEL); + if (vm->page_tables == NULL) { + DRM_ERROR("Cannot allocate memory for page table array\n"); + return -ENOMEM; + } + + r = radeon_bo_create(rdev, pd_size, RADEON_VM_PTB_ALIGN_SIZE, false, + RADEON_GEM_DOMAIN_VRAM, NULL, + &vm->page_directory); + if (r) + return r; + + r = radeon_vm_clear_bo(rdev, vm->page_directory); + if (r) { + radeon_bo_unref(&vm->page_directory); + vm->page_directory = NULL; + return r; + } + + return 0; } /** @@ -980,12 +934,7 @@ void radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) { struct radeon_bo_va *bo_va, *tmp; - int r; - - mutex_lock(&rdev->vm_manager.lock); - mutex_lock(&vm->mutex); - radeon_vm_free_pt(rdev, vm); - mutex_unlock(&rdev->vm_manager.lock); + int i, r; if (!list_empty(&vm->va)) { dev_err(rdev->dev, "still active bo inside vm\n"); @@ -999,8 +948,17 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) kfree(bo_va); } } + + + for (i = 0; i < radeon_vm_num_pdes(rdev); i++) + radeon_bo_unref(&vm->page_tables[i].bo); + kfree(vm->page_tables); + + radeon_bo_unref(&vm->page_directory); + radeon_fence_unref(&vm->fence); radeon_fence_unref(&vm->last_flush); radeon_fence_unref(&vm->last_id_use); - mutex_unlock(&vm->mutex); + + mutex_destroy(&vm->mutex); } From 529364e05bc093dc41b0c9e67e94ac82442c1b35 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 19:33:15 +0100 Subject: [PATCH 062/259] drm/radeon: remove global vm lock MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Not needed any more. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon.h | 1 - drivers/gpu/drm/radeon/radeon_cs.c | 4 ---- drivers/gpu/drm/radeon/radeon_device.c | 3 +-- drivers/gpu/drm/radeon/radeon_ring.c | 7 +++++++ drivers/gpu/drm/radeon/radeon_vm.c | 10 +++------- 5 files changed, 11 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index c31e3c2c1b3a..cd6a48099547 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -884,7 +884,6 @@ struct radeon_vm { }; struct radeon_vm_manager { - struct mutex lock; struct radeon_fence *active[RADEON_NUM_VM]; uint32_t max_pfn; /* number of VMIDs */ diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 420c28dd6a1c..0570e7675fd3 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -502,7 +502,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, if (parser->ring == R600_RING_TYPE_UVD_INDEX) radeon_uvd_note_usage(rdev); - mutex_lock(&rdev->vm_manager.lock); mutex_lock(&vm->mutex); r = radeon_bo_vm_update_pte(parser, vm); if (r) { @@ -510,8 +509,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, } radeon_cs_sync_rings(parser); radeon_semaphore_sync_to(parser->ib.semaphore, vm->fence); - radeon_semaphore_sync_to(parser->ib.semaphore, - radeon_vm_grab_id(rdev, vm, parser->ring)); if ((rdev->family >= CHIP_TAHITI) && (parser->chunk_const_ib_idx != -1)) { @@ -522,7 +519,6 @@ static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, out: mutex_unlock(&vm->mutex); - mutex_unlock(&rdev->vm_manager.lock); return r; } diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index e58dbabd7d7d..7db44de90d6c 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -1191,8 +1191,7 @@ int radeon_device_init(struct radeon_device *rdev, r = radeon_gem_init(rdev); if (r) return r; - /* initialize vm here */ - mutex_init(&rdev->vm_manager.lock); + /* Adjust VM size here. * Currently set to 4GB ((1 << 20) 4k pages). * Max GPUVM size for cayman and SI is 40 bits. diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 665591a7faad..5321e249021f 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -145,6 +145,13 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib, return r; } + /* grab a vm id if necessary */ + if (ib->vm) { + struct radeon_fence *vm_id_fence; + vm_id_fence = radeon_vm_grab_id(rdev, ib->vm, ib->ring); + radeon_semaphore_sync_to(ib->semaphore, vm_id_fence); + } + /* sync with other rings */ r = radeon_semaphore_sync_rings(rdev, ib->semaphore, ib->ring); if (r) { diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index 44b6918a87ba..81d91b513ce1 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -110,12 +110,10 @@ void radeon_vm_manager_fini(struct radeon_device *rdev) if (!rdev->vm_manager.enabled) return; - mutex_lock(&rdev->vm_manager.lock); for (i = 0; i < RADEON_NUM_VM; ++i) radeon_fence_unref(&rdev->vm_manager.active[i]); radeon_asic_vm_fini(rdev); rdev->vm_manager.enabled = false; - mutex_unlock(&rdev->vm_manager.lock); } /** @@ -734,7 +732,7 @@ static void radeon_vm_update_ptes(struct radeon_device *rdev, * Fill in the page table entries for @bo (cayman+). * Returns 0 for success, -EINVAL for failure. * - * Object have to be reserved & global and local mutex must be locked! + * Object have to be reserved and mutex must be locked! */ int radeon_vm_bo_update(struct radeon_device *rdev, struct radeon_vm *vm, @@ -842,12 +840,10 @@ int radeon_vm_bo_rmv(struct radeon_device *rdev, { int r = 0; - mutex_lock(&rdev->vm_manager.lock); mutex_lock(&bo_va->vm->mutex); - if (bo_va->soffset) { + if (bo_va->soffset) r = radeon_vm_bo_update(rdev, bo_va->vm, bo_va->bo, NULL); - } - mutex_unlock(&rdev->vm_manager.lock); + list_del(&bo_va->vm_list); mutex_unlock(&bo_va->vm->mutex); list_del(&bo_va->bo_list); From 4d1526466296360f56f93c195848c1202b0cc10b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Thu, 20 Feb 2014 21:48:00 +0100 Subject: [PATCH 063/259] drm/radeon: drop non blocking allocations from sub allocator MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Not needed any more. Signed-off-by: Christian König --- drivers/gpu/drm/radeon/radeon_object.h | 2 +- drivers/gpu/drm/radeon/radeon_ring.c | 2 +- drivers/gpu/drm/radeon/radeon_sa.c | 7 ++----- drivers/gpu/drm/radeon/radeon_semaphore.c | 2 +- 4 files changed, 5 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index 7dff64d1f592..9e7b25a0629d 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -180,7 +180,7 @@ extern int radeon_sa_bo_manager_suspend(struct radeon_device *rdev, extern int radeon_sa_bo_new(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, struct radeon_sa_bo **sa_bo, - unsigned size, unsigned align, bool block); + unsigned size, unsigned align); extern void radeon_sa_bo_free(struct radeon_device *rdev, struct radeon_sa_bo **sa_bo, struct radeon_fence *fence); diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 5321e249021f..8b0dfdd23793 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -63,7 +63,7 @@ int radeon_ib_get(struct radeon_device *rdev, int ring, { int r; - r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256, true); + r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &ib->sa_bo, size, 256); if (r) { dev_err(rdev->dev, "failed to get a new IB (%d)\n", r); return r; diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c index c0625805cdd7..adcf3e2f07da 100644 --- a/drivers/gpu/drm/radeon/radeon_sa.c +++ b/drivers/gpu/drm/radeon/radeon_sa.c @@ -312,7 +312,7 @@ static bool radeon_sa_bo_next_hole(struct radeon_sa_manager *sa_manager, int radeon_sa_bo_new(struct radeon_device *rdev, struct radeon_sa_manager *sa_manager, struct radeon_sa_bo **sa_bo, - unsigned size, unsigned align, bool block) + unsigned size, unsigned align) { struct radeon_fence *fences[RADEON_NUM_RINGS]; unsigned tries[RADEON_NUM_RINGS]; @@ -353,14 +353,11 @@ int radeon_sa_bo_new(struct radeon_device *rdev, r = radeon_fence_wait_any(rdev, fences, false); spin_lock(&sa_manager->wq.lock); /* if we have nothing to wait for block */ - if (r == -ENOENT && block) { + if (r == -ENOENT) { r = wait_event_interruptible_locked( sa_manager->wq, radeon_sa_event(sa_manager, size, align) ); - - } else if (r == -ENOENT) { - r = -ENOMEM; } } while (!r); diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c index 6140af6772c7..dbd6bcde92de 100644 --- a/drivers/gpu/drm/radeon/radeon_semaphore.c +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -42,7 +42,7 @@ int radeon_semaphore_create(struct radeon_device *rdev, return -ENOMEM; } r = radeon_sa_bo_new(rdev, &rdev->ring_tmp_bo, &(*semaphore)->sa_bo, - 8 * RADEON_NUM_SYNCS, 8, true); + 8 * RADEON_NUM_SYNCS, 8); if (r) { kfree(*semaphore); *semaphore = NULL; From 1c37a72c1bd0b83be8b95cff7f1bc9b1f32bd433 Mon Sep 17 00:00:00 2001 From: Eliad Peller Date: Mon, 3 Mar 2014 13:37:14 +0200 Subject: [PATCH 064/259] mac80211: consider virtual mon when calculating min_def When calculating the current max bw required for a channel context, we didn't consider the virtual monitor interface, resulting in its channel context being narrower than configured. This broke monitor mode with iwlmvm, which uses the minimal width. Reported-by: Ido Yariv Signed-off-by: Eliad Peller Signed-off-by: Johannes Berg --- net/mac80211/chan.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/mac80211/chan.c b/net/mac80211/chan.c index f43613a97dd6..0c1ecfdf9a12 100644 --- a/net/mac80211/chan.c +++ b/net/mac80211/chan.c @@ -100,6 +100,12 @@ ieee80211_get_chanctx_max_required_bw(struct ieee80211_local *local, } max_bw = max(max_bw, width); } + + /* use the configured bandwidth in case of monitor interface */ + sdata = rcu_dereference(local->monitor_sdata); + if (sdata && rcu_access_pointer(sdata->vif.chanctx_conf) == conf) + max_bw = max(max_bw, conf->def.width); + rcu_read_unlock(); return max_bw; From bc00a91d627026f08abf69bf5d0015499dd30c2a Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 3 Mar 2014 13:55:54 +0100 Subject: [PATCH 065/259] cfg80211: remove racy beacon_interval assignment In case of AP mode, the beacon interval is already reset to zero inside cfg80211_stop_ap(), and in the other modes it isn't relevant. Remove the assignment to remove a potential race since the assignment isn't properly locked. Reported-by: Michal Kazior Signed-off-by: Johannes Berg --- net/wireless/core.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/net/wireless/core.c b/net/wireless/core.c index 010892b81a06..a3bf18d11609 100644 --- a/net/wireless/core.c +++ b/net/wireless/core.c @@ -788,8 +788,6 @@ void cfg80211_leave(struct cfg80211_registered_device *rdev, default: break; } - - wdev->beacon_interval = 0; } static int cfg80211_netdev_notifier_call(struct notifier_block *nb, From 24d01805ca652434e1ba7b83a1370cb42b618954 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Thu, 27 Feb 2014 09:19:30 -0600 Subject: [PATCH 066/259] drm/edid: request HDMI underscan by default MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Working with HDMI TVs is a real pain as they tend to overscan by default, meaning that the pixels around the edge of the framebuffer are not displayed. This is well explained here: http://mjg59.dreamwidth.org/8705.html There is a bit in the HDMI info frame that can request that the remote display shows the full pixel data ("underscan"). For the remote display, the HDMI spec states that this is optional - it doesn't have to listen. That means that most TVs will probably ignore this. But, maybe there are a handful of TVs for which this would help the situation. As we live in a digital world, ask the remote display not to overscan by default. Signed-off-by: Daniel Drake Reviewed-by: Ville Syrjälä Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_edid.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index b924306b8477..f8d8a1de9573 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -3599,6 +3599,7 @@ drm_hdmi_avi_infoframe_from_display_mode(struct hdmi_avi_infoframe *frame, frame->picture_aspect = HDMI_PICTURE_ASPECT_NONE; frame->active_aspect = HDMI_ACTIVE_ASPECT_PICTURE; + frame->scan_mode = HDMI_SCAN_MODE_UNDERSCAN; return 0; } From ea2787f350ea5f649fda6bb708c4182cf4ec06ca Mon Sep 17 00:00:00 2001 From: Aaro Koskinen Date: Tue, 4 Mar 2014 00:45:18 +0200 Subject: [PATCH 067/259] ASoC: n810: fix init with DT boot Since 3.14-rc1 only DT boot has been supported on N810, so this file fails to init. Make a minimal fix to retain functionality. This file should be properly converted to DT in longer term. There seems to be still other unresolved issues with N810 audio support, but this patch is needed at minimum as otherwise the machine driver probing would completely fail. Signed-off-by: Aaro Koskinen Acked-by: Jarkko Nikula Signed-off-by: Mark Brown --- sound/soc/omap/n810.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/sound/soc/omap/n810.c b/sound/soc/omap/n810.c index 3fde9e402710..d163e18d85d4 100644 --- a/sound/soc/omap/n810.c +++ b/sound/soc/omap/n810.c @@ -305,7 +305,9 @@ static int __init n810_soc_init(void) int err; struct device *dev; - if (!(machine_is_nokia_n810() || machine_is_nokia_n810_wimax())) + if (!of_have_populated_dt() || + (!of_machine_is_compatible("nokia,n810") && + !of_machine_is_compatible("nokia,n810-wimax"))) return -ENODEV; n810_snd_device = platform_device_alloc("soc-audio", -1); From df0af4403aa8df728a62ccb62a61b3244871068f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 3 Mar 2014 12:38:08 +0100 Subject: [PATCH 068/259] drm/radeon: remove struct radeon_bo_list MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Just move all fields into radeon_cs_reloc, removing unused/duplicated fields. Signed-off-by: Christian König Reviewed-by: Alex Deucher --- drivers/gpu/drm/radeon/evergreen_cs.c | 210 ++++++++++++------------- drivers/gpu/drm/radeon/r100.c | 40 ++--- drivers/gpu/drm/radeon/r200.c | 20 +-- drivers/gpu/drm/radeon/r300.c | 32 ++-- drivers/gpu/drm/radeon/r600_cs.c | 110 ++++++------- drivers/gpu/drm/radeon/radeon.h | 24 ++- drivers/gpu/drm/radeon/radeon_cs.c | 23 ++- drivers/gpu/drm/radeon/radeon_object.c | 4 +- drivers/gpu/drm/radeon/radeon_uvd.c | 2 +- drivers/gpu/drm/radeon/radeon_vce.c | 2 +- drivers/gpu/drm/radeon/radeon_vm.c | 22 ++- 11 files changed, 244 insertions(+), 245 deletions(-) diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c index c7cac07f139b..5c8b358f9fba 100644 --- a/drivers/gpu/drm/radeon/evergreen_cs.c +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -1165,7 +1165,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case DB_DEPTH_CONTROL: track->db_depth_control = radeon_get_ib_value(p, idx); @@ -1196,12 +1196,12 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } ib[idx] &= ~Z_ARRAY_MODE(0xf); track->db_z_info &= ~Z_ARRAY_MODE(0xf); - ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= DB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1237,7 +1237,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_z_read_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_z_read_bo = reloc->robj; track->db_dirty = true; break; @@ -1249,7 +1249,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_z_write_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_z_write_bo = reloc->robj; track->db_dirty = true; break; @@ -1261,7 +1261,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_s_read_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_s_read_bo = reloc->robj; track->db_dirty = true; break; @@ -1273,7 +1273,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_s_write_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_s_write_bo = reloc->robj; track->db_dirty = true; break; @@ -1297,7 +1297,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->vgt_strmout_bo[tmp] = reloc->robj; track->streamout_dirty = true; break; @@ -1317,7 +1317,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); case CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); track->cb_dirty = true; @@ -1381,8 +1381,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); + ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); } track->cb_dirty = true; break; @@ -1399,8 +1399,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); + ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); } track->cb_dirty = true; break; @@ -1461,10 +1461,10 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1489,10 +1489,10 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); @@ -1520,7 +1520,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_fmask_bo[tmp] = reloc->robj; break; case CB_COLOR0_CMASK: @@ -1537,7 +1537,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_cmask_bo[tmp] = reloc->robj; break; case CB_COLOR0_FMASK_SLICE: @@ -1578,7 +1578,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - CB_COLOR0_BASE) / 0x3c; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_bo[tmp] = reloc->robj; track->cb_dirty = true; break; @@ -1594,7 +1594,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_bo[tmp] = reloc->robj; track->cb_dirty = true; break; @@ -1606,7 +1606,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->htile_offset = radeon_get_ib_value(p, idx); - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->htile_bo = reloc->robj; track->db_dirty = true; break; @@ -1723,7 +1723,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MEMORY_EXPORT_BASE: if (p->rdev->family >= CHIP_CAYMAN) { @@ -1737,7 +1737,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case CAYMAN_SX_SCATTER_EXPORT_BASE: if (p->rdev->family < CHIP_CAYMAN) { @@ -1751,7 +1751,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MISC: track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; @@ -1836,7 +1836,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (idx_value & 0xfffffff0) + ((u64)(tmp & 0xff) << 32); @@ -1882,7 +1882,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1909,7 +1909,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1937,7 +1937,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + radeon_get_ib_value(p, idx+1) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2027,7 +2027,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DISPATCH_INDIRECT\n"); return -EINVAL; } - ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); + ib[idx+0] = idx_value + (u32)(reloc->gpu_offset & 0xffffffff); r = evergreen_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); @@ -2049,7 +2049,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2106,7 +2106,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx) + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA src buffer too small (%llu %lu)\n", @@ -2144,7 +2144,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx+2) + ((u64)(radeon_get_ib_value(p, idx+3) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA dst buffer too small (%llu %lu)\n", @@ -2174,7 +2174,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SURFACE_SYNC\n"); return -EINVAL; } - ib[idx+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+2] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_EVENT_WRITE: @@ -2190,7 +2190,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2212,7 +2212,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2234,7 +2234,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -2302,11 +2302,11 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { ib[idx+1+(i*8)+1] |= - TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->tiling_flags)); + if (reloc->tiling_flags & RADEON_TILING_MACRO) { unsigned bankw, bankh, mtaspect, tile_split; - evergreen_tiling_fields(reloc->lobj.tiling_flags, + evergreen_tiling_fields(reloc->tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); ib[idx+1+(i*8)+6] |= TEX_TILE_SPLIT(tile_split); @@ -2318,7 +2318,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } } texture = reloc->robj; - toffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + toffset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); /* tex mip base */ tex_dim = ib[idx+1+(i*8)+0] & 0x7; @@ -2337,7 +2337,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE (tex)\n"); return -EINVAL; } - moffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + moffset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); mipmap = reloc->robj; } @@ -2364,7 +2364,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset; } - offset64 = reloc->lobj.gpu_offset + offset; + offset64 = reloc->gpu_offset + offset; ib[idx+1+(i*8)+0] = offset64; ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | (upper_32_bits(offset64) & 0xff); @@ -2445,7 +2445,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } @@ -2464,7 +2464,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } @@ -2493,7 +2493,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 8, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+0] = offset; ib[idx+1] = upper_32_bits(offset) & 0xff; break; @@ -2518,7 +2518,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } else { @@ -2542,7 +2542,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } else { @@ -2717,7 +2717,7 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += count + 7; break; /* linear */ @@ -2725,8 +2725,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += count + 3; break; default: @@ -2768,10 +2768,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; break; /* Copy L2T/T2L */ @@ -2781,22 +2781,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx + 7); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+7); src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) { dev_warn(p->dev, "DMA L2T, src buffer too small (%llu %lu)\n", @@ -2827,10 +2827,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + count, radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xffffffff); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xffffffff); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; break; /* Copy L2L, partial */ @@ -2840,10 +2840,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) DRM_ERROR("L2L Partial is cayman only !\n"); return -EINVAL; } - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+5] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(src_reloc->gpu_offset & 0xffffffff); + ib[idx+2] += upper_32_bits(src_reloc->gpu_offset) & 0xff; + ib[idx+4] += (u32)(dst_reloc->gpu_offset & 0xffffffff); + ib[idx+5] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += 9; break; @@ -2876,12 +2876,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+4] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+5] += upper_32_bits(dst2_reloc->lobj.gpu_offset) & 0xff; - ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+4] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+5] += upper_32_bits(dst2_reloc->gpu_offset) & 0xff; + ib[idx+6] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 7; break; /* Copy L2T Frame to Field */ @@ -2916,10 +2916,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; /* Copy L2T/T2L, partial */ @@ -2932,16 +2932,16 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* detile bit */ if (radeon_get_ib_value(p, idx + 2) & (1 << 31)) { /* tiled src, linear dst */ - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } p->idx += 12; break; @@ -2978,10 +2978,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; /* Copy L2T/T2L (tile units) */ @@ -2992,22 +2992,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx+7); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+7); src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32; - ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+7] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+8] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) { dev_warn(p->dev, "DMA L2T, T2L src buffer too small (%llu %lu)\n", @@ -3028,8 +3028,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) DRM_ERROR("L2T, T2L Partial is cayman only !\n"); return -EINVAL; } - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); - ib[idx+4] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); + ib[idx+4] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += 13; break; /* Copy L2T broadcast (tile units) */ @@ -3065,10 +3065,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst2_offset + (count * 4), radeon_bo_size(dst2_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); - ib[idx+2] += (u32)(dst2_reloc->lobj.gpu_offset >> 8); - ib[idx+8] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+9] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); + ib[idx+2] += (u32)(dst2_reloc->gpu_offset >> 8); + ib[idx+8] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+9] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 10; break; default: @@ -3089,8 +3089,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p) dst_offset, radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) << 16) & 0x00ff0000; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) << 16) & 0x00ff0000; p->idx += 4; break; case DMA_PACKET_NOP: diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 1690a2dc0721..0a894aee7406 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -1274,12 +1274,12 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p, value = radeon_get_ib_value(p, idx); tmp = value & 0x003fffff; - tmp += (((u32)reloc->lobj.gpu_offset) >> 10); + tmp += (((u32)reloc->gpu_offset) >> 10); if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_DST_TILE_MACRO; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + if (reloc->tiling_flags & RADEON_TILING_MICRO) { if (reg == RADEON_SRC_PITCH_OFFSET) { DRM_ERROR("Cannot src blit from microtiled surface\n"); radeon_cs_dump_packet(p, pkt); @@ -1325,7 +1325,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, return r; } idx_value = radeon_get_ib_value(p, idx); - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); track->arrays[i + 0].esize = idx_value >> 8; track->arrays[i + 0].robj = reloc->robj; @@ -1337,7 +1337,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->lobj.gpu_offset); + ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->gpu_offset); track->arrays[i + 1].robj = reloc->robj; track->arrays[i + 1].esize = idx_value >> 24; track->arrays[i + 1].esize &= 0x7F; @@ -1351,7 +1351,7 @@ int r100_packet3_load_vbpntr(struct radeon_cs_parser *p, return r; } idx_value = radeon_get_ib_value(p, idx); - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); track->arrays[i + 0].robj = reloc->robj; track->arrays[i + 0].esize = idx_value >> 8; track->arrays[i + 0].esize &= 0x7F; @@ -1594,7 +1594,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_RB3D_COLOROFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -1607,7 +1607,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_TXOFFSET_0: case RADEON_PP_TXOFFSET_1: @@ -1621,16 +1621,16 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_TXO_MICRO_TILE_X2; tmp = idx_value & ~(0x7 << 2); tmp |= tile_flags; - ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + ib[idx] = tmp + ((u32)reloc->gpu_offset); } else - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1648,7 +1648,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[0].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[0].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1666,7 +1666,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[1].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[1].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1684,7 +1684,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[2].cube_info[i].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[2].cube_info[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1702,9 +1702,9 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -1772,7 +1772,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_CNTL: { @@ -1932,7 +1932,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->gpu_offset); r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); if (r) { return r; @@ -1946,7 +1946,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->lobj.gpu_offset); + ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->gpu_offset); track->num_arrays = 1; track->vtx_size = r100_get_vtx_size(radeon_get_ib_value(p, idx + 2)); diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index b3807edb1936..58f0473aa73f 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -185,7 +185,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_RB3D_COLOROFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -198,7 +198,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R200_PP_TXOFFSET_0: case R200_PP_TXOFFSET_1: @@ -215,16 +215,16 @@ int r200_packet0_check(struct radeon_cs_parser *p, return r; } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R200_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R200_TXO_MICRO_TILE; tmp = idx_value & ~(0x7 << 2); tmp |= tile_flags; - ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + ib[idx] = tmp + ((u32)reloc->gpu_offset); } else - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -268,7 +268,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, return r; } track->textures[i].cube_info[face - 1].offset = idx_value; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); track->textures[i].cube_info[face - 1].robj = reloc->robj; track->tex_dirty = true; break; @@ -287,9 +287,9 @@ int r200_packet0_check(struct radeon_cs_parser *p, } if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -362,7 +362,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case RADEON_PP_CNTL: { diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 7c63ef840e86..41cdf236ee9a 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -640,7 +640,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->cb[i].robj = reloc->robj; track->cb[i].offset = idx_value; track->cb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_ZB_DEPTHOFFSET: r = radeon_cs_packet_next_reloc(p, &reloc, 0); @@ -653,7 +653,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->zb.robj = reloc->robj; track->zb.offset = idx_value; track->zb_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_TX_OFFSET_0: case R300_TX_OFFSET_0+4: @@ -682,16 +682,16 @@ static int r300_packet0_check(struct radeon_cs_parser *p, if (p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) { ib[idx] = (idx_value & 31) | /* keep the 1st 5 bits */ - ((idx_value & ~31) + (u32)reloc->lobj.gpu_offset); + ((idx_value & ~31) + (u32)reloc->gpu_offset); } else { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_TXO_MACRO_TILE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_TXO_MICRO_TILE; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_TXO_MICRO_TILE_SQUARE; - tmp = idx_value + ((u32)reloc->lobj.gpu_offset); + tmp = idx_value + ((u32)reloc->gpu_offset); tmp |= tile_flags; ib[idx] = tmp; } @@ -753,11 +753,11 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_COLOR_MICROTILE_ENABLE; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE; tmp = idx_value & ~(0x7 << 16); @@ -838,11 +838,11 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_DEPTHMACROTILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + if (reloc->tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_DEPTHMICROTILE_TILED; - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + else if (reloc->tiling_flags & RADEON_TILING_MICRO_SQUARE) tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE; tmp = idx_value & ~(0x7 << 16); @@ -1052,7 +1052,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case 0x4e0c: /* RB3D_COLOR_CHANNEL_MASK */ @@ -1097,7 +1097,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->aa.robj = reloc->robj; track->aa.offset = idx_value; track->aa_dirty = true; - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->gpu_offset); break; case R300_RB3D_AARESOLVE_PITCH: track->aa.pitch = idx_value & 0x3FFE; @@ -1162,7 +1162,7 @@ static int r300_packet3_check(struct radeon_cs_parser *p, radeon_cs_dump_packet(p, pkt); return r; } - ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->gpu_offset); r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); if (r) { return r; diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c index 2812c7d1ae6f..12511bb5fd6f 100644 --- a/drivers/gpu/drm/radeon/r600_cs.c +++ b/drivers/gpu/drm/radeon/r600_cs.c @@ -1022,7 +1022,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SQ_CONFIG: track->sq_config = radeon_get_ib_value(p, idx); @@ -1043,7 +1043,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_depth_info = radeon_get_ib_value(p, idx); ib[idx] &= C_028010_ARRAY_MODE; track->db_depth_info &= C_028010_ARRAY_MODE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { ib[idx] |= S_028010_ARRAY_MODE(V_028010_ARRAY_2D_TILED_THIN1); track->db_depth_info |= S_028010_ARRAY_MODE(V_028010_ARRAY_2D_TILED_THIN1); } else { @@ -1084,9 +1084,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->vgt_strmout_bo[tmp] = reloc->robj; - track->vgt_strmout_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->vgt_strmout_bo_mc[tmp] = reloc->gpu_offset; track->streamout_dirty = true; break; case VGT_STRMOUT_BUFFER_SIZE_0: @@ -1105,7 +1105,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case R_028238_CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); @@ -1142,10 +1142,10 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4; track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + if (reloc->tiling_flags & RADEON_TILING_MACRO) { ib[idx] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_2D_TILED_THIN1); track->cb_color_info[tmp] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_2D_TILED_THIN1); - } else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + } else if (reloc->tiling_flags & RADEON_TILING_MICRO) { ib[idx] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_1D_TILED_THIN1); track->cb_color_info[tmp] |= S_0280A0_ARRAY_MODE(V_0280A0_ARRAY_1D_TILED_THIN1); } @@ -1214,7 +1214,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } track->cb_color_frag_bo[tmp] = reloc->robj; track->cb_color_frag_offset[tmp] = (u64)ib[idx] << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } if (G_0280A0_TILE_MODE(track->cb_color_info[tmp])) { track->cb_dirty = true; @@ -1245,7 +1245,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } track->cb_color_tile_bo[tmp] = reloc->robj; track->cb_color_tile_offset[tmp] = (u64)ib[idx] << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } if (G_0280A0_TILE_MODE(track->cb_color_info[tmp])) { track->cb_dirty = true; @@ -1281,10 +1281,10 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } tmp = (reg - CB_COLOR0_BASE) / 4; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->cb_color_base_last[tmp] = ib[idx]; track->cb_color_bo[tmp] = reloc->robj; - track->cb_color_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->cb_color_bo_mc[tmp] = reloc->gpu_offset; track->cb_dirty = true; break; case DB_DEPTH_BASE: @@ -1295,9 +1295,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->db_offset = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->db_bo = reloc->robj; - track->db_bo_mc = reloc->lobj.gpu_offset; + track->db_bo_mc = reloc->gpu_offset; track->db_dirty = true; break; case DB_HTILE_DATA_BASE: @@ -1308,7 +1308,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return -EINVAL; } track->htile_offset = radeon_get_ib_value(p, idx) << 8; - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); track->htile_bo = reloc->robj; track->db_dirty = true; break; @@ -1377,7 +1377,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MEMORY_EXPORT_BASE: r = radeon_cs_packet_next_reloc(p, &reloc, r600_nomm); @@ -1386,7 +1386,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); break; case SX_MISC: track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; @@ -1672,7 +1672,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (idx_value & 0xfffffff0) + ((u64)(tmp & 0xff) << 32); @@ -1713,7 +1713,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + idx_value + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); @@ -1765,7 +1765,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff0) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1805,7 +1805,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx) + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA src buffer too small (%llu %lu)\n", @@ -1835,7 +1835,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, tmp = radeon_get_ib_value(p, idx+2) + ((u64)(radeon_get_ib_value(p, idx+3) & 0xff) << 32); - offset = reloc->lobj.gpu_offset + tmp; + offset = reloc->gpu_offset + tmp; if ((tmp + size) > radeon_bo_size(reloc->robj)) { dev_warn(p->dev, "CP DMA dst buffer too small (%llu %lu)\n", @@ -1861,7 +1861,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SURFACE_SYNC\n"); return -EINVAL; } - ib[idx+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+2] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_EVENT_WRITE: @@ -1877,7 +1877,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1899,7 +1899,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - offset = reloc->lobj.gpu_offset + + offset = reloc->gpu_offset + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); @@ -1964,11 +1964,11 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE\n"); return -EINVAL; } - base_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + base_offset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + if (reloc->tiling_flags & RADEON_TILING_MACRO) ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1); - else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + else if (reloc->tiling_flags & RADEON_TILING_MICRO) ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1); } texture = reloc->robj; @@ -1978,13 +1978,13 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE\n"); return -EINVAL; } - mip_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + mip_offset = (u32)((reloc->gpu_offset >> 8) & 0xffffffff); mipmap = reloc->robj; r = r600_check_texture_resource(p, idx+(i*7)+1, texture, mipmap, base_offset + radeon_get_ib_value(p, idx+1+(i*7)+2), mip_offset + radeon_get_ib_value(p, idx+1+(i*7)+3), - reloc->lobj.tiling_flags); + reloc->tiling_flags); if (r) return r; ib[idx+1+(i*7)+2] += base_offset; @@ -2008,7 +2008,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj) - offset; } - offset64 = reloc->lobj.gpu_offset + offset; + offset64 = reloc->gpu_offset + offset; ib[idx+1+(i*8)+0] = offset64; ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | (upper_32_bits(offset64) & 0xff); @@ -2118,7 +2118,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + ib[idx+1] += (u32)((reloc->gpu_offset >> 8) & 0xffffffff); } break; case PACKET3_SURFACE_BASE_UPDATE: @@ -2151,7 +2151,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } @@ -2170,7 +2170,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } @@ -2199,7 +2199,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 8, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+0] = offset; ib[idx+1] = upper_32_bits(offset) & 0xff; break; @@ -2224,7 +2224,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+1] = offset; ib[idx+2] = upper_32_bits(offset) & 0xff; } else { @@ -2248,7 +2248,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, offset + 4, radeon_bo_size(reloc->robj)); return -EINVAL; } - offset += reloc->lobj.gpu_offset; + offset += reloc->gpu_offset; ib[idx+3] = offset; ib[idx+4] = upper_32_bits(offset) & 0xff; } else { @@ -2505,14 +2505,14 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); p->idx += count + 5; } else { dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; p->idx += count + 3; } if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) { @@ -2539,22 +2539,22 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) /* tiled src, linear dst */ src_offset = radeon_get_ib_value(p, idx+1); src_offset <<= 8; - ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(src_reloc->gpu_offset >> 8); dst_offset = radeon_get_ib_value(p, idx+5); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32; - ib[idx+5] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+6] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; + ib[idx+5] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+6] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; } else { /* linear src, tiled dst */ src_offset = radeon_get_ib_value(p, idx+5); src_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32; - ib[idx+5] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+5] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+6] += upper_32_bits(src_reloc->gpu_offset) & 0xff; dst_offset = radeon_get_ib_value(p, idx+1); dst_offset <<= 8; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8); + ib[idx+1] += (u32)(dst_reloc->gpu_offset >> 8); } p->idx += 7; } else { @@ -2564,10 +2564,10 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff; - ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(dst_reloc->gpu_offset) & 0xff; + ib[idx+4] += upper_32_bits(src_reloc->gpu_offset) & 0xff; p->idx += 5; } else { src_offset = radeon_get_ib_value(p, idx+2); @@ -2575,10 +2575,10 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset = radeon_get_ib_value(p, idx+1); dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff0000)) << 16; - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff; - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff) << 16; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+2] += (u32)(src_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += upper_32_bits(src_reloc->gpu_offset) & 0xff; + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) & 0xff) << 16; p->idx += 4; } } @@ -2610,8 +2610,8 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p) dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj)); return -EINVAL; } - ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc); - ib[idx+3] += (upper_32_bits(dst_reloc->lobj.gpu_offset) << 16) & 0x00ff0000; + ib[idx+1] += (u32)(dst_reloc->gpu_offset & 0xfffffffc); + ib[idx+3] += (upper_32_bits(dst_reloc->gpu_offset) << 16) & 0x00ff0000; p->idx += 4; break; case DMA_PACKET_NOP: diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index cd6a48099547..111deab2492a 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -479,15 +479,6 @@ struct radeon_bo { }; #define gem_to_radeon_bo(gobj) container_of((gobj), struct radeon_bo, gem_base) -struct radeon_bo_list { - struct ttm_validate_buffer tv; - struct radeon_bo *bo; - uint64_t gpu_offset; - unsigned domain; - unsigned alt_domain; - u32 tiling_flags; -}; - int radeon_gem_debugfs_init(struct radeon_device *rdev); /* sub-allocation manager, it has to be protected by another lock. @@ -987,9 +978,12 @@ void cayman_dma_fini(struct radeon_device *rdev); struct radeon_cs_reloc { struct drm_gem_object *gobj; struct radeon_bo *robj; - struct radeon_bo_list lobj; + struct ttm_validate_buffer tv; + uint64_t gpu_offset; + unsigned domain; + unsigned alt_domain; + uint32_t tiling_flags; uint32_t handle; - uint32_t flags; }; struct radeon_cs_chunk { @@ -1013,7 +1007,7 @@ struct radeon_cs_parser { unsigned nrelocs; struct radeon_cs_reloc *relocs; struct radeon_cs_reloc **relocs_ptr; - struct radeon_bo_list *vm_bos; + struct radeon_cs_reloc *vm_bos; struct list_head validated; unsigned dma_reloc_idx; /* indices of various chunks */ @@ -2803,9 +2797,9 @@ int radeon_vm_manager_init(struct radeon_device *rdev); void radeon_vm_manager_fini(struct radeon_device *rdev); int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm); -struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, - struct radeon_vm *vm, - struct list_head *head); +struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head); struct radeon_fence *radeon_vm_grab_id(struct radeon_device *rdev, struct radeon_vm *vm, int ring); void radeon_vm_flush(struct radeon_device *rdev, diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 0570e7675fd3..2b6e0ebcc13a 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -125,7 +125,6 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) } p->relocs_ptr[i] = &p->relocs[i]; p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj); - p->relocs[i].lobj.bo = p->relocs[i].robj; /* The userspace buffer priorities are from 0 to 15. A higher * number means the buffer is more important. @@ -141,10 +140,10 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) if (p->ring == R600_RING_TYPE_UVD_INDEX && (i == 0 || drm_pci_device_is_agp(p->rdev->ddev))) { /* TODO: is this still needed for NI+ ? */ - p->relocs[i].lobj.domain = + p->relocs[i].domain = RADEON_GEM_DOMAIN_VRAM; - p->relocs[i].lobj.alt_domain = + p->relocs[i].alt_domain = RADEON_GEM_DOMAIN_VRAM; /* prioritize this over any other relocation */ @@ -153,16 +152,16 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p) uint32_t domain = r->write_domain ? r->write_domain : r->read_domains; - p->relocs[i].lobj.domain = domain; + p->relocs[i].domain = domain; if (domain == RADEON_GEM_DOMAIN_VRAM) domain |= RADEON_GEM_DOMAIN_GTT; - p->relocs[i].lobj.alt_domain = domain; + p->relocs[i].alt_domain = domain; } - p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo; + p->relocs[i].tv.bo = &p->relocs[i].robj->tbo; p->relocs[i].handle = r->handle; - radeon_cs_buckets_add(&buckets, &p->relocs[i].lobj.tv.head, + radeon_cs_buckets_add(&buckets, &p->relocs[i].tv.head, priority); } @@ -356,11 +355,11 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) static int cmp_size_smaller_first(void *priv, struct list_head *a, struct list_head *b) { - struct radeon_bo_list *la = list_entry(a, struct radeon_bo_list, tv.head); - struct radeon_bo_list *lb = list_entry(b, struct radeon_bo_list, tv.head); + struct radeon_cs_reloc *la = list_entry(a, struct radeon_cs_reloc, tv.head); + struct radeon_cs_reloc *lb = list_entry(b, struct radeon_cs_reloc, tv.head); /* Sort A before B if A is smaller. */ - return (int)la->bo->tbo.num_pages - (int)lb->bo->tbo.num_pages; + return (int)la->robj->tbo.num_pages - (int)lb->robj->tbo.num_pages; } /** @@ -786,9 +785,9 @@ int radeon_cs_packet_next_reloc(struct radeon_cs_parser *p, /* FIXME: we assume reloc size is 4 dwords */ if (nomm) { *cs_reloc = p->relocs; - (*cs_reloc)->lobj.gpu_offset = + (*cs_reloc)->gpu_offset = (u64)relocs_chunk->kdata[idx + 3] << 32; - (*cs_reloc)->lobj.gpu_offset |= relocs_chunk->kdata[idx + 0]; + (*cs_reloc)->gpu_offset |= relocs_chunk->kdata[idx + 0]; } else *cs_reloc = p->relocs_ptr[(idx / 4)]; return 0; diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index ed03f2d15853..ca79431b2c1c 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -422,7 +422,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev, struct ww_acquire_ctx *ticket, struct list_head *head, int ring) { - struct radeon_bo_list *lobj; + struct radeon_cs_reloc *lobj; struct radeon_bo *bo; int r; u64 bytes_moved = 0, initial_bytes_moved; @@ -434,7 +434,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev, } list_for_each_entry(lobj, head, tv.head) { - bo = lobj->bo; + bo = lobj->robj; if (!bo->pin_count) { u32 domain = lobj->domain; u32 current_domain = diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c index ceb7b289b745..6a2e3ff68374 100644 --- a/drivers/gpu/drm/radeon/radeon_uvd.c +++ b/drivers/gpu/drm/radeon/radeon_uvd.c @@ -453,7 +453,7 @@ static int radeon_uvd_cs_reloc(struct radeon_cs_parser *p, } reloc = p->relocs_ptr[(idx / 4)]; - start = reloc->lobj.gpu_offset; + start = reloc->gpu_offset; end = start + radeon_bo_size(reloc->robj); start += offset; diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index 39ec7d8f33ab..76e9904bc537 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -461,7 +461,7 @@ int radeon_vce_cs_reloc(struct radeon_cs_parser *p, int lo, int hi) return -EINVAL; } - offset += p->relocs_ptr[(idx / 4)]->lobj.gpu_offset; + offset += p->relocs_ptr[(idx / 4)]->gpu_offset; p->ib.ptr[lo] = offset & 0xFFFFFFFF; p->ib.ptr[hi] = offset >> 32; diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c index 81d91b513ce1..2aae6ce49d32 100644 --- a/drivers/gpu/drm/radeon/radeon_vm.c +++ b/drivers/gpu/drm/radeon/radeon_vm.c @@ -125,33 +125,39 @@ void radeon_vm_manager_fini(struct radeon_device *rdev) * Add the page directory to the list of BOs to * validate for command submission (cayman+). */ -struct radeon_bo_list *radeon_vm_get_bos(struct radeon_device *rdev, - struct radeon_vm *vm, - struct list_head *head) +struct radeon_cs_reloc *radeon_vm_get_bos(struct radeon_device *rdev, + struct radeon_vm *vm, + struct list_head *head) { - struct radeon_bo_list *list; + struct radeon_cs_reloc *list; unsigned i, idx, size; - size = (radeon_vm_num_pdes(rdev) + 1) * sizeof(struct radeon_bo_list); + size = (radeon_vm_num_pdes(rdev) + 1) * sizeof(struct radeon_cs_reloc); list = kmalloc(size, GFP_KERNEL); if (!list) return NULL; /* add the vm page table to the list */ - list[0].bo = vm->page_directory; + list[0].gobj = NULL; + list[0].robj = vm->page_directory; list[0].domain = RADEON_GEM_DOMAIN_VRAM; list[0].alt_domain = RADEON_GEM_DOMAIN_VRAM; list[0].tv.bo = &vm->page_directory->tbo; + list[0].tiling_flags = 0; + list[0].handle = 0; list_add(&list[0].tv.head, head); for (i = 0, idx = 1; i <= vm->max_pde_used; i++) { if (!vm->page_tables[i].bo) continue; - list[idx].bo = vm->page_tables[i].bo; + list[idx].gobj = NULL; + list[idx].robj = vm->page_tables[i].bo; list[idx].domain = RADEON_GEM_DOMAIN_VRAM; list[idx].alt_domain = RADEON_GEM_DOMAIN_VRAM; - list[idx].tv.bo = &list[idx].bo->tbo; + list[idx].tv.bo = &list[idx].robj->tbo; + list[idx].tiling_flags = 0; + list[idx].handle = 0; list_add(&list[idx++].tv.head, head); } From 1e9291996c4eedf79883f47ec635235e39d3d6cd Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Tue, 4 Mar 2014 10:28:23 +0200 Subject: [PATCH 069/259] iwlwifi: mvm: don't WARN when statistics are handled late Since the statistics handler is asynchrous, it can very well be that we will handle the statistics (hence the RSSI fluctuation) when we already disassociated. Don't WARN on this case. This solves: https://bugzilla.redhat.com/show_bug.cgi?id=1071998 Cc: [3.10+] Fixes: 2b76ef13086f ("iwlwifi: mvm: implement reduced Tx power") Reviewed-by: Johannes Berg Signed-off-by: Emmanuel Grumbach --- drivers/net/wireless/iwlwifi/mvm/bt-coex.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/iwlwifi/mvm/bt-coex.c b/drivers/net/wireless/iwlwifi/mvm/bt-coex.c index 76cde6ce6551..18a895a949d4 100644 --- a/drivers/net/wireless/iwlwifi/mvm/bt-coex.c +++ b/drivers/net/wireless/iwlwifi/mvm/bt-coex.c @@ -872,8 +872,11 @@ void iwl_mvm_bt_rssi_event(struct iwl_mvm *mvm, struct ieee80211_vif *vif, lockdep_assert_held(&mvm->mutex); - /* Rssi update while not associated ?! */ - if (WARN_ON_ONCE(mvmvif->ap_sta_id == IWL_MVM_STATION_COUNT)) + /* + * Rssi update while not associated - can happen since the statistics + * are handled asynchronously + */ + if (mvmvif->ap_sta_id == IWL_MVM_STATION_COUNT) return; /* No BT - reports should be disabled */ From acfcd9ed588465e48efe5ceb67177096d51c95a9 Mon Sep 17 00:00:00 2001 From: Oren Givon Date: Tue, 4 Mar 2014 14:47:39 +0200 Subject: [PATCH 070/259] iwlwifi: fix and add 7265 series HW IDs Update of the HW IDs for the 7265 series. Signed-off-by: Oren Givon Signed-off-by: Emmanuel Grumbach --- drivers/net/wireless/iwlwifi/pcie/drv.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/net/wireless/iwlwifi/pcie/drv.c b/drivers/net/wireless/iwlwifi/pcie/drv.c index f47bcbe2945a..3872ead75488 100644 --- a/drivers/net/wireless/iwlwifi/pcie/drv.c +++ b/drivers/net/wireless/iwlwifi/pcie/drv.c @@ -359,13 +359,12 @@ static DEFINE_PCI_DEVICE_TABLE(iwl_hw_card_ids) = { /* 7265 Series */ {IWL_PCI_DEVICE(0x095A, 0x5010, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x5110, iwl7265_2ac_cfg)}, - {IWL_PCI_DEVICE(0x095A, 0x5112, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x5100, iwl7265_2ac_cfg)}, - {IWL_PCI_DEVICE(0x095A, 0x510A, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095B, 0x5310, iwl7265_2ac_cfg)}, - {IWL_PCI_DEVICE(0x095B, 0x5302, iwl7265_2ac_cfg)}, + {IWL_PCI_DEVICE(0x095B, 0x5302, iwl7265_n_cfg)}, {IWL_PCI_DEVICE(0x095B, 0x5210, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x5012, iwl7265_2ac_cfg)}, + {IWL_PCI_DEVICE(0x095A, 0x5412, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x5410, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x5400, iwl7265_2ac_cfg)}, {IWL_PCI_DEVICE(0x095A, 0x1010, iwl7265_2ac_cfg)}, From b6251243ac0a3cb63afe748133f6600f77526078 Mon Sep 17 00:00:00 2001 From: Ivaylo Dimitrov Date: Sat, 1 Mar 2014 14:52:06 +0200 Subject: [PATCH 071/259] wl1251: use skb_trim to make skb shorter the current code is directly setting skb->len, which is not correct and brings problems with HAVE_EFFICIENT_UNALIGNED_ACCESS enabled in config Signed-off-by: Ivaylo Dimitrov Signed-off-by: John W. Linville --- drivers/net/wireless/ti/wl1251/rx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/wireless/ti/wl1251/rx.c b/drivers/net/wireless/ti/wl1251/rx.c index 123c4bb50e0a..cde0eaf99714 100644 --- a/drivers/net/wireless/ti/wl1251/rx.c +++ b/drivers/net/wireless/ti/wl1251/rx.c @@ -180,7 +180,7 @@ static void wl1251_rx_body(struct wl1251 *wl, wl1251_mem_read(wl, rx_packet_ring_addr, rx_buffer, length); /* The actual length doesn't include the target's alignment */ - skb->len = desc->length - PLCP_HEADER_LENGTH; + skb_trim(skb, desc->length - PLCP_HEADER_LENGTH); fc = (u16 *)skb->data; From e4ad1accb28d0ed8cea6f12395d58686ad344ca7 Mon Sep 17 00:00:00 2001 From: Patrick Lai Date: Sun, 2 Mar 2014 11:52:57 -0800 Subject: [PATCH 072/259] ASoC: pcm: free path list before exiting from error conditions dpcm_path_get() allocates dynamic memory to hold path list. Corresponding dpcm_path_put() must be called to free the memory. dpcm_path_put() is not called under several error conditions. This leads to memory leak. Signed-off-by: Patrick Lai Signed-off-by: Mark Brown Cc: stable@vger.kernel.org --- sound/soc/soc-pcm.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/sound/soc/soc-pcm.c b/sound/soc/soc-pcm.c index 47e1ce771e65..28522bd03b8e 100644 --- a/sound/soc/soc-pcm.c +++ b/sound/soc/soc-pcm.c @@ -1989,6 +1989,7 @@ int soc_dpcm_runtime_update(struct snd_soc_card *card) paths = dpcm_path_get(fe, SNDRV_PCM_STREAM_PLAYBACK, &list); if (paths < 0) { + dpcm_path_put(&list); dev_warn(fe->dev, "ASoC: %s no valid %s path\n", fe->dai_link->name, "playback"); mutex_unlock(&card->mutex); @@ -2018,6 +2019,7 @@ int soc_dpcm_runtime_update(struct snd_soc_card *card) paths = dpcm_path_get(fe, SNDRV_PCM_STREAM_CAPTURE, &list); if (paths < 0) { + dpcm_path_put(&list); dev_warn(fe->dev, "ASoC: %s no valid %s path\n", fe->dai_link->name, "capture"); mutex_unlock(&card->mutex); @@ -2082,6 +2084,7 @@ static int dpcm_fe_dai_open(struct snd_pcm_substream *fe_substream) fe->dpcm[stream].runtime = fe_substream->runtime; if (dpcm_path_get(fe, stream, &list) <= 0) { + dpcm_path_put(&list); dev_dbg(fe->dev, "ASoC: %s no valid %s route\n", fe->dai_link->name, stream ? "capture" : "playback"); } From 864a6040f395464003af8dd0d8ca86fed19866d4 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Tue, 4 Mar 2014 13:46:53 +0100 Subject: [PATCH 073/259] mac80211: clear sequence/fragment number in QoS-null frames Avoid leaking data by sending uninitialized memory and setting an invalid (non-zero) fragment number (the sequence number is ignored anyway) by setting the seq_ctrl field to zero. Cc: stable@vger.kernel.org Fixes: 3f52b7e328c5 ("mac80211: mesh power save basics") Fixes: ce662b44ce22 ("mac80211: send (QoS) Null if no buffered frames") Reviewed-by: Emmanuel Grumbach Signed-off-by: Johannes Berg --- net/mac80211/mesh_ps.c | 1 + net/mac80211/sta_info.c | 1 + 2 files changed, 2 insertions(+) diff --git a/net/mac80211/mesh_ps.c b/net/mac80211/mesh_ps.c index 2802f9d9279d..ad8b377b4b9f 100644 --- a/net/mac80211/mesh_ps.c +++ b/net/mac80211/mesh_ps.c @@ -36,6 +36,7 @@ static struct sk_buff *mps_qos_null_get(struct sta_info *sta) sdata->vif.addr); nullfunc->frame_control = fc; nullfunc->duration_id = 0; + nullfunc->seq_ctrl = 0; /* no address resolution for this frame -> set addr 1 immediately */ memcpy(nullfunc->addr1, sta->sta.addr, ETH_ALEN); memset(skb_put(skb, 2), 0, 2); /* append QoS control field */ diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index a023b432143b..137a192e64bc 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -1206,6 +1206,7 @@ static void ieee80211_send_null_response(struct ieee80211_sub_if_data *sdata, memcpy(nullfunc->addr1, sta->sta.addr, ETH_ALEN); memcpy(nullfunc->addr2, sdata->vif.addr, ETH_ALEN); memcpy(nullfunc->addr3, sdata->vif.addr, ETH_ALEN); + nullfunc->seq_ctrl = 0; skb->priority = tid; skb_set_queue_mapping(skb, ieee802_1d_to_ac[tid]); From 24b9bf43e93e0edd89072da51cf1fab95fc69dec Mon Sep 17 00:00:00 2001 From: Nikolay Aleksandrov Date: Mon, 3 Mar 2014 23:19:18 +0100 Subject: [PATCH 074/259] net: fix for a race condition in the inet frag code I stumbled upon this very serious bug while hunting for another one, it's a very subtle race condition between inet_frag_evictor, inet_frag_intern and the IPv4/6 frag_queue and expire functions (basically the users of inet_frag_kill/inet_frag_put). What happens is that after a fragment has been added to the hash chain but before it's been added to the lru_list (inet_frag_lru_add) in inet_frag_intern, it may get deleted (either by an expired timer if the system load is high or the timer sufficiently low, or by the fraq_queue function for different reasons) before it's added to the lru_list, then after it gets added it's a matter of time for the evictor to get to a piece of memory which has been freed leading to a number of different bugs depending on what's left there. I've been able to trigger this on both IPv4 and IPv6 (which is normal as the frag code is the same), but it's been much more difficult to trigger on IPv4 due to the protocol differences about how fragments are treated. The setup I used to reproduce this is: 2 machines with 4 x 10G bonded in a RR bond, so the same flow can be seen on multiple cards at the same time. Then I used multiple instances of ping/ping6 to generate fragmented packets and flood the machines with them while running other processes to load the attacked machine. *It is very important to have the _same flow_ coming in on multiple CPUs concurrently. Usually the attacked machine would die in less than 30 minutes, if configured properly to have many evictor calls and timeouts it could happen in 10 minutes or so. An important point to make is that any caller (frag_queue or timer) of inet_frag_kill will remove both the timer refcount and the original/guarding refcount thus removing everything that's keeping the frag from being freed at the next inet_frag_put. All of this could happen before the frag was ever added to the LRU list, then it gets added and the evictor uses a freed fragment. An example for IPv6 would be if a fragment is being added and is at the stage of being inserted in the hash after the hash lock is released, but before inet_frag_lru_add executes (or is able to obtain the lru lock) another overlapping fragment for the same flow arrives at a different CPU which finds it in the hash, but since it's overlapping it drops it invoking inet_frag_kill and thus removing all guarding refcounts, and afterwards freeing it by invoking inet_frag_put which removes the last refcount added previously by inet_frag_find, then inet_frag_lru_add gets executed by inet_frag_intern and we have a freed fragment in the lru_list. The fix is simple, just move the lru_add under the hash chain locked region so when a removing function is called it'll have to wait for the fragment to be added to the lru_list, and then it'll remove it (it works because the hash chain removal is done before the lru_list one and there's no window between the two list adds when the frag can get dropped). With this fix applied I couldn't kill the same machine in 24 hours with the same setup. Fixes: 3ef0eb0db4bf ("net: frag, move LRU list maintenance outside of rwlock") CC: Florian Westphal CC: Jesper Dangaard Brouer CC: David S. Miller Signed-off-by: Nikolay Aleksandrov Acked-by: Jesper Dangaard Brouer Signed-off-by: David S. Miller --- net/ipv4/inet_fragment.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index bb075fc9a14f..322dcebfc588 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -278,9 +278,10 @@ static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf, atomic_inc(&qp->refcnt); hlist_add_head(&qp->list, &hb->chain); + inet_frag_lru_add(nf, qp); spin_unlock(&hb->chain_lock); read_unlock(&f->lock); - inet_frag_lru_add(nf, qp); + return qp; } From 6565b9eeef194afbb3beec80d6dd2447f4091f8c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Linus=20L=C3=BCssing?= Date: Tue, 4 Mar 2014 03:57:35 +0100 Subject: [PATCH 075/259] bridge: multicast: add sanity check for query source addresses MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit MLD queries are supposed to have an IPv6 link-local source address according to RFC2710, section 4 and RFC3810, section 5.1.14. This patch adds a sanity check to ignore such broken MLD queries. Without this check, such malformed MLD queries can result in a denial of service: The queries are ignored by any MLD listener therefore they will not respond with an MLD report. However, without this patch these malformed MLD queries would enable the snooping part in the bridge code, potentially shutting down the according ports towards these hosts for multicast traffic as the bridge did not learn about these listeners. Reported-by: Jan Stancek Signed-off-by: Linus Lüssing Reviewed-by: Hannes Frederic Sowa Signed-off-by: David S. Miller --- net/bridge/br_multicast.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index ef66365b7354..fb0e36fac668 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -1235,6 +1235,12 @@ static int br_ip6_multicast_query(struct net_bridge *br, (port && port->state == BR_STATE_DISABLED)) goto out; + /* RFC2710+RFC3810 (MLDv1+MLDv2) require link-local source addresses */ + if (!(ipv6_addr_type(&ip6h->saddr) & IPV6_ADDR_LINKLOCAL)) { + err = -EINVAL; + goto out; + } + if (skb->len == sizeof(*mld)) { if (!pskb_may_pull(skb, sizeof(*mld))) { err = -EINVAL; From 367d56f7b4d5ce61e883c64f81786c7a3ae88eea Mon Sep 17 00:00:00 2001 From: Gavin Shan Date: Tue, 4 Mar 2014 15:35:20 +0800 Subject: [PATCH 076/259] net/mlx4: Support shutdown() interface In kexec scenario, we failed to load the mlx4 driver in the second kernel because the ownership bit was hold by the first kernel without release correctly. The patch adds shutdown() interface so that the ownership can be released correctly in the first kernel. It also helps avoiding EEH error happened during boot stage of the second kernel because of undesired traffic, which can't be handled by hardware during that stage on Power platform. Signed-off-by: Gavin Shan Tested-by: Wei Yang Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/main.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c index d711158b0d4b..5a6105f1ba6d 100644 --- a/drivers/net/ethernet/mellanox/mlx4/main.c +++ b/drivers/net/ethernet/mellanox/mlx4/main.c @@ -2684,6 +2684,7 @@ static struct pci_driver mlx4_driver = { .name = DRV_NAME, .id_table = mlx4_pci_table, .probe = mlx4_init_one, + .shutdown = mlx4_remove_one, .remove = mlx4_remove_one, .err_handler = &mlx4_err_handler, }; From 10c3271712f58215f4d336a1e30aa25be09cd5d1 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Tue, 4 Mar 2014 20:47:48 +0800 Subject: [PATCH 077/259] r8152: disable the ECM mode There are known issues for switching the drivers between ECM mode and vendor mode. The interrup transfer may become abnormal. The hardware may have the opportunity to die if you change the configuration without unloading the current driver first, because all the control transfers of the current driver would fail after the command of switching the configuration. Although to use the ecm driver and vendor driver independently is fine, it may have problems to change the driver from one to the other by switching the configuration. Additionally, now the vendor mode driver is more powerful than the ECM driver. Thus, disable the ECM mode driver, and let r8152 to set the configuration to vendor mode and reset the device automatically. Signed-off-by: Hayes Wang Signed-off-by: David S. Miller --- drivers/net/usb/Makefile | 2 +- drivers/net/usb/cdc_ether.c | 7 + drivers/net/usb/r8152.c | 15 ++- drivers/net/usb/r815x.c | 248 ------------------------------------ 4 files changed, 17 insertions(+), 255 deletions(-) delete mode 100644 drivers/net/usb/r815x.c diff --git a/drivers/net/usb/Makefile b/drivers/net/usb/Makefile index 433f0a00c683..e2797f1e1b31 100644 --- a/drivers/net/usb/Makefile +++ b/drivers/net/usb/Makefile @@ -11,7 +11,7 @@ obj-$(CONFIG_USB_HSO) += hso.o obj-$(CONFIG_USB_NET_AX8817X) += asix.o asix-y := asix_devices.o asix_common.o ax88172a.o obj-$(CONFIG_USB_NET_AX88179_178A) += ax88179_178a.o -obj-$(CONFIG_USB_NET_CDCETHER) += cdc_ether.o r815x.o +obj-$(CONFIG_USB_NET_CDCETHER) += cdc_ether.o obj-$(CONFIG_USB_NET_CDC_EEM) += cdc_eem.o obj-$(CONFIG_USB_NET_DM9601) += dm9601.o obj-$(CONFIG_USB_NET_SR9700) += sr9700.o diff --git a/drivers/net/usb/cdc_ether.c b/drivers/net/usb/cdc_ether.c index 42e176912c8e..bd363b27e854 100644 --- a/drivers/net/usb/cdc_ether.c +++ b/drivers/net/usb/cdc_ether.c @@ -652,6 +652,13 @@ static const struct usb_device_id products[] = { .driver_info = 0, }, +/* Samsung USB Ethernet Adapters */ +{ + USB_DEVICE_AND_INTERFACE_INFO(SAMSUNG_VENDOR_ID, 0xa101, USB_CLASS_COMM, + USB_CDC_SUBCLASS_ETHERNET, USB_CDC_PROTO_NONE), + .driver_info = 0, +}, + /* WHITELIST!!! * * CDC Ether uses two interfaces, not necessarily consecutive. diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c index d89dbe395ad2..adb12f349a61 100644 --- a/drivers/net/usb/r8152.c +++ b/drivers/net/usb/r8152.c @@ -449,9 +449,6 @@ enum rtl8152_flags { #define MCU_TYPE_PLA 0x0100 #define MCU_TYPE_USB 0x0000 -#define REALTEK_USB_DEVICE(vend, prod) \ - USB_DEVICE_INTERFACE_CLASS(vend, prod, USB_CLASS_VENDOR_SPEC) - struct rx_desc { __le32 opts1; #define RX_LEN_MASK 0x7fff @@ -2739,6 +2736,12 @@ static int rtl8152_probe(struct usb_interface *intf, struct net_device *netdev; int ret; + if (udev->actconfig->desc.bConfigurationValue != 1) { + usb_driver_set_configuration(udev, 1); + return -ENODEV; + } + + usb_reset_device(udev); netdev = alloc_etherdev(sizeof(struct r8152)); if (!netdev) { dev_err(&intf->dev, "Out of memory\n"); @@ -2819,9 +2822,9 @@ static void rtl8152_disconnect(struct usb_interface *intf) /* table of devices that work with this driver */ static struct usb_device_id rtl8152_table[] = { - {REALTEK_USB_DEVICE(VENDOR_ID_REALTEK, PRODUCT_ID_RTL8152)}, - {REALTEK_USB_DEVICE(VENDOR_ID_REALTEK, PRODUCT_ID_RTL8153)}, - {REALTEK_USB_DEVICE(VENDOR_ID_SAMSUNG, PRODUCT_ID_SAMSUNG)}, + {USB_DEVICE(VENDOR_ID_REALTEK, PRODUCT_ID_RTL8152)}, + {USB_DEVICE(VENDOR_ID_REALTEK, PRODUCT_ID_RTL8153)}, + {USB_DEVICE(VENDOR_ID_SAMSUNG, PRODUCT_ID_SAMSUNG)}, {} }; diff --git a/drivers/net/usb/r815x.c b/drivers/net/usb/r815x.c deleted file mode 100644 index f0a8791b7636..000000000000 --- a/drivers/net/usb/r815x.c +++ /dev/null @@ -1,248 +0,0 @@ -#include -#include -#include -#include -#include -#include - -#define RTL815x_REQT_READ 0xc0 -#define RTL815x_REQT_WRITE 0x40 -#define RTL815x_REQ_GET_REGS 0x05 -#define RTL815x_REQ_SET_REGS 0x05 - -#define MCU_TYPE_PLA 0x0100 -#define OCP_BASE 0xe86c -#define BASE_MII 0xa400 - -#define BYTE_EN_DWORD 0xff -#define BYTE_EN_WORD 0x33 -#define BYTE_EN_BYTE 0x11 - -#define R815x_PHY_ID 32 -#define REALTEK_VENDOR_ID 0x0bda - - -static int pla_read_word(struct usb_device *udev, u16 index) -{ - int ret; - u8 shift = index & 2; - __le32 *tmp; - - tmp = kmalloc(sizeof(*tmp), GFP_KERNEL); - if (!tmp) - return -ENOMEM; - - index &= ~3; - - ret = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0), - RTL815x_REQ_GET_REGS, RTL815x_REQT_READ, - index, MCU_TYPE_PLA, tmp, sizeof(*tmp), 500); - if (ret < 0) - goto out2; - - ret = __le32_to_cpu(*tmp); - ret >>= (shift * 8); - ret &= 0xffff; - -out2: - kfree(tmp); - return ret; -} - -static int pla_write_word(struct usb_device *udev, u16 index, u32 data) -{ - __le32 *tmp; - u32 mask = 0xffff; - u16 byen = BYTE_EN_WORD; - u8 shift = index & 2; - int ret; - - tmp = kmalloc(sizeof(*tmp), GFP_KERNEL); - if (!tmp) - return -ENOMEM; - - data &= mask; - - if (shift) { - byen <<= shift; - mask <<= (shift * 8); - data <<= (shift * 8); - index &= ~3; - } - - ret = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0), - RTL815x_REQ_GET_REGS, RTL815x_REQT_READ, - index, MCU_TYPE_PLA, tmp, sizeof(*tmp), 500); - if (ret < 0) - goto out3; - - data |= __le32_to_cpu(*tmp) & ~mask; - *tmp = __cpu_to_le32(data); - - ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0), - RTL815x_REQ_SET_REGS, RTL815x_REQT_WRITE, - index, MCU_TYPE_PLA | byen, tmp, sizeof(*tmp), - 500); - -out3: - kfree(tmp); - return ret; -} - -static int ocp_reg_read(struct usbnet *dev, u16 addr) -{ - u16 ocp_base, ocp_index; - int ret; - - ocp_base = addr & 0xf000; - ret = pla_write_word(dev->udev, OCP_BASE, ocp_base); - if (ret < 0) - goto out; - - ocp_index = (addr & 0x0fff) | 0xb000; - ret = pla_read_word(dev->udev, ocp_index); - -out: - return ret; -} - -static int ocp_reg_write(struct usbnet *dev, u16 addr, u16 data) -{ - u16 ocp_base, ocp_index; - int ret; - - ocp_base = addr & 0xf000; - ret = pla_write_word(dev->udev, OCP_BASE, ocp_base); - if (ret < 0) - goto out1; - - ocp_index = (addr & 0x0fff) | 0xb000; - ret = pla_write_word(dev->udev, ocp_index, data); - -out1: - return ret; -} - -static int r815x_mdio_read(struct net_device *netdev, int phy_id, int reg) -{ - struct usbnet *dev = netdev_priv(netdev); - int ret; - - if (phy_id != R815x_PHY_ID) - return -EINVAL; - - if (usb_autopm_get_interface(dev->intf) < 0) - return -ENODEV; - - ret = ocp_reg_read(dev, BASE_MII + reg * 2); - - usb_autopm_put_interface(dev->intf); - return ret; -} - -static -void r815x_mdio_write(struct net_device *netdev, int phy_id, int reg, int val) -{ - struct usbnet *dev = netdev_priv(netdev); - - if (phy_id != R815x_PHY_ID) - return; - - if (usb_autopm_get_interface(dev->intf) < 0) - return; - - ocp_reg_write(dev, BASE_MII + reg * 2, val); - - usb_autopm_put_interface(dev->intf); -} - -static int r8153_bind(struct usbnet *dev, struct usb_interface *intf) -{ - int status; - - status = usbnet_cdc_bind(dev, intf); - if (status < 0) - return status; - - dev->mii.dev = dev->net; - dev->mii.mdio_read = r815x_mdio_read; - dev->mii.mdio_write = r815x_mdio_write; - dev->mii.phy_id_mask = 0x3f; - dev->mii.reg_num_mask = 0x1f; - dev->mii.phy_id = R815x_PHY_ID; - dev->mii.supports_gmii = 1; - - return status; -} - -static int r8152_bind(struct usbnet *dev, struct usb_interface *intf) -{ - int status; - - status = usbnet_cdc_bind(dev, intf); - if (status < 0) - return status; - - dev->mii.dev = dev->net; - dev->mii.mdio_read = r815x_mdio_read; - dev->mii.mdio_write = r815x_mdio_write; - dev->mii.phy_id_mask = 0x3f; - dev->mii.reg_num_mask = 0x1f; - dev->mii.phy_id = R815x_PHY_ID; - dev->mii.supports_gmii = 0; - - return status; -} - -static const struct driver_info r8152_info = { - .description = "RTL8152 ECM Device", - .flags = FLAG_ETHER | FLAG_POINTTOPOINT, - .bind = r8152_bind, - .unbind = usbnet_cdc_unbind, - .status = usbnet_cdc_status, - .manage_power = usbnet_manage_power, -}; - -static const struct driver_info r8153_info = { - .description = "RTL8153 ECM Device", - .flags = FLAG_ETHER | FLAG_POINTTOPOINT, - .bind = r8153_bind, - .unbind = usbnet_cdc_unbind, - .status = usbnet_cdc_status, - .manage_power = usbnet_manage_power, -}; - -static const struct usb_device_id products[] = { -{ - USB_DEVICE_AND_INTERFACE_INFO(REALTEK_VENDOR_ID, 0x8152, USB_CLASS_COMM, - USB_CDC_SUBCLASS_ETHERNET, USB_CDC_PROTO_NONE), - .driver_info = (unsigned long) &r8152_info, -}, - -{ - USB_DEVICE_AND_INTERFACE_INFO(REALTEK_VENDOR_ID, 0x8153, USB_CLASS_COMM, - USB_CDC_SUBCLASS_ETHERNET, USB_CDC_PROTO_NONE), - .driver_info = (unsigned long) &r8153_info, -}, - - { }, /* END */ -}; -MODULE_DEVICE_TABLE(usb, products); - -static struct usb_driver r815x_driver = { - .name = "r815x", - .id_table = products, - .probe = usbnet_probe, - .disconnect = usbnet_disconnect, - .suspend = usbnet_suspend, - .resume = usbnet_resume, - .reset_resume = usbnet_resume, - .supports_autosuspend = 1, - .disable_hub_initiated_lpm = 1, -}; - -module_usb_driver(r815x_driver); - -MODULE_AUTHOR("Hayes Wang"); -MODULE_DESCRIPTION("Realtek USB ECM device"); -MODULE_LICENSE("GPL"); From c485658bae87faccd7aed540fd2ca3ab37992310 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Tue, 4 Mar 2014 16:35:51 +0100 Subject: [PATCH 078/259] net: sctp: fix skb leakage in COOKIE ECHO path of chunk->auth_chunk While working on ec0223ec48a9 ("net: sctp: fix sctp_sf_do_5_1D_ce to verify if we/peer is AUTH capable"), we noticed that there's a skb memory leakage in the error path. Running the same reproducer as in ec0223ec48a9 and by unconditionally jumping to the error label (to simulate an error condition) in sctp_sf_do_5_1D_ce() receive path lets kmemleak detector bark about the unfreed chunk->auth_chunk skb clone: Unreferenced object 0xffff8800b8f3a000 (size 256): comm "softirq", pid 0, jiffies 4294769856 (age 110.757s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 89 ab 75 5e d4 01 58 13 00 00 00 00 00 00 00 00 ..u^..X......... backtrace: [] kmemleak_alloc+0x4e/0xb0 [] kmem_cache_alloc+0xc8/0x210 [] skb_clone+0x49/0xb0 [] sctp_endpoint_bh_rcv+0x1d9/0x230 [sctp] [] sctp_inq_push+0x4c/0x70 [sctp] [] sctp_rcv+0x82e/0x9a0 [sctp] [] ip_local_deliver_finish+0xa8/0x210 [] nf_reinject+0xbf/0x180 [] nfqnl_recv_verdict+0x1d2/0x2b0 [nfnetlink_queue] [] nfnetlink_rcv_msg+0x14b/0x250 [nfnetlink] [] netlink_rcv_skb+0xa9/0xc0 [] nfnetlink_rcv+0x23f/0x408 [nfnetlink] [] netlink_unicast+0x168/0x250 [] netlink_sendmsg+0x2e1/0x3f0 [] sock_sendmsg+0x8b/0xc0 [] ___sys_sendmsg+0x369/0x380 What happens is that commit bbd0d59809f9 clones the skb containing the AUTH chunk in sctp_endpoint_bh_rcv() when having the edge case that an endpoint requires COOKIE-ECHO chunks to be authenticated: ---------- INIT[RANDOM; CHUNKS; HMAC-ALGO] ----------> <------- INIT-ACK[RANDOM; CHUNKS; HMAC-ALGO] --------- ------------------ AUTH; COOKIE-ECHO ----------------> <-------------------- COOKIE-ACK --------------------- When we enter sctp_sf_do_5_1D_ce() and before we actually get to the point where we process (and subsequently free) a non-NULL chunk->auth_chunk, we could hit the "goto nomem_init" path from an error condition and thus leave the cloned skb around w/o freeing it. The fix is to centrally free such clones in sctp_chunk_destroy() handler that is invoked from sctp_chunk_free() after all refs have dropped; and also move both kfree_skb(chunk->auth_chunk) there, so that chunk->auth_chunk is either NULL (since sctp_chunkify() allocs new chunks through kmem_cache_zalloc()) or non-NULL with a valid skb pointer. chunk->skb and chunk->auth_chunk are the only skbs in the sctp_chunk structure that need to be handeled. While at it, we should use consume_skb() for both. It is the same as dev_kfree_skb() but more appropriately named as we are not a device but a protocol. Also, this effectively replaces the kfree_skb() from both invocations into consume_skb(). Functions are the same only that kfree_skb() assumes that the frame was being dropped after a failure (e.g. for tools like drop monitor), usage of consume_skb() seems more appropriate in function sctp_chunk_destroy() though. Fixes: bbd0d59809f9 ("[SCTP]: Implement the receive and verification of AUTH chunk") Signed-off-by: Daniel Borkmann Cc: Vlad Yasevich Cc: Neil Horman Acked-by: Vlad Yasevich Acked-by: Neil Horman Signed-off-by: David S. Miller --- net/sctp/sm_make_chunk.c | 4 ++-- net/sctp/sm_statefuns.c | 5 ----- 2 files changed, 2 insertions(+), 7 deletions(-) diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c index 632090b961c3..3a1767ef3201 100644 --- a/net/sctp/sm_make_chunk.c +++ b/net/sctp/sm_make_chunk.c @@ -1421,8 +1421,8 @@ static void sctp_chunk_destroy(struct sctp_chunk *chunk) BUG_ON(!list_empty(&chunk->list)); list_del_init(&chunk->transmitted_list); - /* Free the chunk skb data and the SCTP_chunk stub itself. */ - dev_kfree_skb(chunk->skb); + consume_skb(chunk->skb); + consume_skb(chunk->auth_chunk); SCTP_DBG_OBJCNT_DEC(chunk); kmem_cache_free(sctp_chunk_cachep, chunk); diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c index ae65b6b5973a..01e002430c85 100644 --- a/net/sctp/sm_statefuns.c +++ b/net/sctp/sm_statefuns.c @@ -760,7 +760,6 @@ sctp_disposition_t sctp_sf_do_5_1D_ce(struct net *net, /* Make sure that we and the peer are AUTH capable */ if (!net->sctp.auth_enable || !new_asoc->peer.auth_capable) { - kfree_skb(chunk->auth_chunk); sctp_association_free(new_asoc); return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands); } @@ -775,10 +774,6 @@ sctp_disposition_t sctp_sf_do_5_1D_ce(struct net *net, auth.transport = chunk->transport; ret = sctp_sf_authenticate(net, ep, new_asoc, type, &auth); - - /* We can now safely free the auth_chunk clone */ - kfree_skb(chunk->auth_chunk); - if (ret != SCTP_IERROR_NO_ERROR) { sctp_association_free(new_asoc); return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands); From 920309086652651c7d59791e5b83e2f10112b293 Mon Sep 17 00:00:00 2001 From: Soren Brinkmann Date: Tue, 4 Mar 2014 08:46:39 -0800 Subject: [PATCH 079/259] net: macb: Check DMA mappings for error With CONFIG_DMA_API_DEBUG enabled the following warning is printed: WARNING: CPU: 0 PID: 619 at lib/dma-debug.c:1101 check_unmap+0x758/0x894() macb e000b000.ethernet: DMA-API: device driver failed to check map error[device address=0x000000002d171c02] [size=322 bytes] [mapped as single] Modules linked in: CPU: 0 PID: 619 Comm: udhcpc Not tainted 3.14.0-rc3-xilinx-00219-gd158fc7f36a2 #63 [] (unwind_backtrace) from [] (show_stack+0x10/0x14) [] (show_stack) from [] (dump_stack+0x7c/0xc8) [] (dump_stack) from [] (warn_slowpath_common+0x60/0x84) [] (warn_slowpath_common) from [] (warn_slowpath_fmt+0x2c/0x3c) [] (warn_slowpath_fmt) from [] (check_unmap+0x758/0x894) [] (check_unmap) from [] (debug_dma_unmap_page+0x64/0x70) [] (debug_dma_unmap_page) from [] (macb_interrupt+0x1f8/0x2dc) [] (macb_interrupt) from [] (handle_irq_event_percpu+0x2c/0x178) [] (handle_irq_event_percpu) from [] (handle_irq_event+0x3c/0x5c) [] (handle_irq_event) from [] (handle_fasteoi_irq+0xb8/0x100) [] (handle_fasteoi_irq) from [] (generic_handle_irq+0x20/0x30) [] (generic_handle_irq) from [] (handle_IRQ+0x64/0x8c) [] (handle_IRQ) from [] (gic_handle_irq+0x3c/0x60) [] (gic_handle_irq) from [] (__irq_svc+0x44/0x78) Exception stack(0xed197f60 to 0xed197fa8) 7f60: 00000134 60000013 bd94362e bd94362e be96b37c 00000014 fffffd72 00000122 7f80: c000ebe4 ed196000 00000000 00000011 c032c0d8 ed197fa8 c0064008 c000ea20 7fa0: 60000013 ffffffff [] (__irq_svc) from [] (ret_fast_syscall+0x0/0x48) ---[ end trace 478f921d0d542d1e ]--- Mapped at: [] debug_dma_map_page+0x48/0x11c [] macb_start_xmit+0x184/0x2a8 [] dev_hard_start_xmit+0x334/0x470 [] sch_direct_xmit+0x78/0x2f8 [] __dev_queue_xmit+0x318/0x708 due to missing checks of the dma mapping. Add the appropriate checks to fix this. Signed-off-by: Soren Brinkmann Reviewed-by: Ben Hutchings Signed-off-by: David S. Miller --- drivers/net/ethernet/cadence/macb.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c index 3190d38e16fb..1f03e2041e86 100644 --- a/drivers/net/ethernet/cadence/macb.c +++ b/drivers/net/ethernet/cadence/macb.c @@ -632,11 +632,16 @@ static void gem_rx_refill(struct macb *bp) "Unable to allocate sk_buff\n"); break; } - bp->rx_skbuff[entry] = skb; /* now fill corresponding descriptor entry */ paddr = dma_map_single(&bp->pdev->dev, skb->data, bp->rx_buffer_size, DMA_FROM_DEVICE); + if (dma_mapping_error(&bp->pdev->dev, paddr)) { + dev_kfree_skb(skb); + break; + } + + bp->rx_skbuff[entry] = skb; if (entry == RX_RING_SIZE - 1) paddr |= MACB_BIT(RX_WRAP); @@ -1036,11 +1041,15 @@ static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev) } entry = macb_tx_ring_wrap(bp->tx_head); - bp->tx_head++; netdev_vdbg(bp->dev, "Allocated ring entry %u\n", entry); mapping = dma_map_single(&bp->pdev->dev, skb->data, len, DMA_TO_DEVICE); + if (dma_mapping_error(&bp->pdev->dev, mapping)) { + kfree_skb(skb); + goto unlock; + } + bp->tx_head++; tx_skb = &bp->tx_skb[entry]; tx_skb->skb = skb; tx_skb->mapping = mapping; @@ -1066,6 +1075,7 @@ static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev) if (CIRC_SPACE(bp->tx_head, bp->tx_tail, TX_RING_SIZE) < 1) netif_stop_queue(dev); +unlock: spin_unlock_irqrestore(&bp->lock, flags); return NETDEV_TX_OK; From 48330e08fa168395b9fd9f369f06cca1df204361 Mon Sep 17 00:00:00 2001 From: Soren Brinkmann Date: Tue, 4 Mar 2014 08:46:40 -0800 Subject: [PATCH 080/259] net: macb: DMA-unmap full rx-buffer When allocating RX buffers a fixed size is used, while freeing is based on actually received bytes, resulting in the following kernel warning when CONFIG_DMA_API_DEBUG is enabled: WARNING: CPU: 0 PID: 0 at lib/dma-debug.c:1051 check_unmap+0x258/0x894() macb e000b000.ethernet: DMA-API: device driver frees DMA memory with different size [device address=0x000000002d170040] [map size=1536 bytes] [unmap size=60 bytes] Modules linked in: CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.14.0-rc3-xilinx-00220-g49f84081ce4f #65 [] (unwind_backtrace) from [] (show_stack+0x10/0x14) [] (show_stack) from [] (dump_stack+0x7c/0xc8) [] (dump_stack) from [] (warn_slowpath_common+0x60/0x84) [] (warn_slowpath_common) from [] (warn_slowpath_fmt+0x2c/0x3c) [] (warn_slowpath_fmt) from [] (check_unmap+0x258/0x894) [] (check_unmap) from [] (debug_dma_unmap_page+0x64/0x70) [] (debug_dma_unmap_page) from [] (gem_rx+0x118/0x170) [] (gem_rx) from [] (macb_poll+0x24/0x94) [] (macb_poll) from [] (net_rx_action+0x6c/0x188) [] (net_rx_action) from [] (__do_softirq+0x108/0x280) [] (__do_softirq) from [] (irq_exit+0x84/0xf8) [] (irq_exit) from [] (handle_IRQ+0x68/0x8c) [] (handle_IRQ) from [] (gic_handle_irq+0x3c/0x60) [] (gic_handle_irq) from [] (__irq_svc+0x44/0x78) Exception stack(0xc056df20 to 0xc056df68) df20: 00000001 c0577430 00000000 c0577430 04ce8e0d 00000002 edfce238 00000000 df40: 04e20f78 00000002 c05981f4 00000000 00000008 c056df68 c0064008 c02d7658 df60: 20000013 ffffffff [] (__irq_svc) from [] (cpuidle_enter_state+0x54/0xf8) [] (cpuidle_enter_state) from [] (cpuidle_idle_call+0xe0/0x138) [] (cpuidle_idle_call) from [] (arch_cpu_idle+0x8/0x3c) [] (arch_cpu_idle) from [] (cpu_startup_entry+0xbc/0x124) [] (cpu_startup_entry) from [] (start_kernel+0x350/0x3b0) ---[ end trace d5fdc38641bd3a11 ]--- Mapped at: [] debug_dma_map_page+0x48/0x11c [] gem_rx_refill+0x154/0x1f8 [] macb_open+0x270/0x3e0 [] __dev_open+0x7c/0xfc [] __dev_change_flags+0x8c/0x140 Fixing this by passing the same size which is passed during mapping the memory to the unmap function as well. Signed-off-by: Soren Brinkmann Reviewed-by: Ben Hutchings Signed-off-by: David S. Miller --- drivers/net/ethernet/cadence/macb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c index 1f03e2041e86..d0c38e01e99f 100644 --- a/drivers/net/ethernet/cadence/macb.c +++ b/drivers/net/ethernet/cadence/macb.c @@ -730,7 +730,7 @@ static int gem_rx(struct macb *bp, int budget) skb_put(skb, len); addr = MACB_BF(RX_WADDR, MACB_BFEXT(RX_WADDR, addr)); dma_unmap_single(&bp->pdev->dev, addr, - len, DMA_FROM_DEVICE); + bp->rx_buffer_size, DMA_FROM_DEVICE); skb->protocol = eth_type_trans(skb, bp->dev); skb_checksum_none_assert(skb); From 1b07da516ee25250f458c76c012ebe4cd677a84f Mon Sep 17 00:00:00 2001 From: Haiyang Zhang Date: Tue, 4 Mar 2014 14:11:06 -0800 Subject: [PATCH 081/259] hyperv: Move state setting for link query It moves the state setting for query into rndis_filter_receive_response(). All callbacks including query-complete and status-callback are synchronized by channel->inbound_lock. This prevents pentential race between them. Signed-off-by: Haiyang Zhang Signed-off-by: David S. Miller --- drivers/net/hyperv/netvsc_drv.c | 4 ++++ drivers/net/hyperv/rndis_filter.c | 21 ++++++++++++++++++++- 2 files changed, 24 insertions(+), 1 deletion(-) diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c index 7141a1937360..d6fce9750b95 100644 --- a/drivers/net/hyperv/netvsc_drv.c +++ b/drivers/net/hyperv/netvsc_drv.c @@ -442,6 +442,8 @@ static int netvsc_probe(struct hv_device *dev, if (!net) return -ENOMEM; + netif_carrier_off(net); + net_device_ctx = netdev_priv(net); net_device_ctx->device_ctx = dev; hv_set_drvdata(dev, net); @@ -473,6 +475,8 @@ static int netvsc_probe(struct hv_device *dev, pr_err("Unable to register netdev.\n"); rndis_filter_device_remove(dev); free_netdev(net); + } else { + schedule_delayed_work(&net_device_ctx->dwork, 0); } return ret; diff --git a/drivers/net/hyperv/rndis_filter.c b/drivers/net/hyperv/rndis_filter.c index 1084e5de3ceb..b54fd257652b 100644 --- a/drivers/net/hyperv/rndis_filter.c +++ b/drivers/net/hyperv/rndis_filter.c @@ -243,6 +243,22 @@ static int rndis_filter_send_request(struct rndis_device *dev, return ret; } +static void rndis_set_link_state(struct rndis_device *rdev, + struct rndis_request *request) +{ + u32 link_status; + struct rndis_query_complete *query_complete; + + query_complete = &request->response_msg.msg.query_complete; + + if (query_complete->status == RNDIS_STATUS_SUCCESS && + query_complete->info_buflen == sizeof(u32)) { + memcpy(&link_status, (void *)((unsigned long)query_complete + + query_complete->info_buf_offset), sizeof(u32)); + rdev->link_state = link_status != 0; + } +} + static void rndis_filter_receive_response(struct rndis_device *dev, struct rndis_message *resp) { @@ -272,6 +288,10 @@ static void rndis_filter_receive_response(struct rndis_device *dev, sizeof(struct rndis_message) + RNDIS_EXT_LEN) { memcpy(&request->response_msg, resp, resp->msg_len); + if (request->request_msg.ndis_msg_type == + RNDIS_MSG_QUERY && request->request_msg.msg. + query_req.oid == RNDIS_OID_GEN_MEDIA_CONNECT_STATUS) + rndis_set_link_state(dev, request); } else { netdev_err(ndev, "rndis response buffer overflow " @@ -620,7 +640,6 @@ static int rndis_filter_query_device_link_status(struct rndis_device *dev) ret = rndis_filter_query_device(dev, RNDIS_OID_GEN_MEDIA_CONNECT_STATUS, &link_status, &size); - dev->link_state = (link_status != 0) ? true : false; return ret; } From 6cbde8253a8143ada18ec0d1711230747a7c1934 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Thu, 6 Mar 2014 03:30:46 +0000 Subject: [PATCH 082/259] ARM: KVM: fix non-VGIC compilation Add a stub for kvm_vgic_addr when compiling without CONFIG_KVM_ARM_VGIC. The usefulness of this configurarion is extremely doubtful, but let's fix it anyway (until we decide that we'll always support a VGIC). Reported-by: Michele Paolino Cc: Paolo Bonzini Cc: Christoffer Dall Signed-off-by: Marc Zyngier Signed-off-by: Paolo Bonzini --- include/kvm/arm_vgic.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index be85127bfed3..f27000f55a83 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -171,6 +171,11 @@ static inline int kvm_vgic_set_addr(struct kvm *kvm, unsigned long type, u64 add return 0; } +static inline int kvm_vgic_addr(struct kvm *kvm, unsigned long type, u64 *addr, bool write) +{ + return -ENXIO; +} + static inline int kvm_vgic_init(struct kvm *kvm) { return 0; From c99b1861c232e1f641f13b8645e0febb3712cc71 Mon Sep 17 00:00:00 2001 From: Amitkumar Karwar Date: Tue, 4 Mar 2014 18:43:13 -0800 Subject: [PATCH 083/259] mwifiex: copy AP's HT capability info correctly While preparing association request, intersection of device's HT capability information and corresponding fields advertised by AP is used. This patch fixes an error while copying this field from AP's beacon. Cc: Signed-off-by: Amitkumar Karwar Signed-off-by: Bing Zhao Signed-off-by: John W. Linville --- drivers/net/wireless/mwifiex/11n.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/net/wireless/mwifiex/11n.c b/drivers/net/wireless/mwifiex/11n.c index 6261f8c53d44..7db1a89fdd95 100644 --- a/drivers/net/wireless/mwifiex/11n.c +++ b/drivers/net/wireless/mwifiex/11n.c @@ -308,8 +308,7 @@ mwifiex_cmd_append_11n_tlv(struct mwifiex_private *priv, ht_cap->header.len = cpu_to_le16(sizeof(struct ieee80211_ht_cap)); memcpy((u8 *) ht_cap + sizeof(struct mwifiex_ie_types_header), - (u8 *) bss_desc->bcn_ht_cap + - sizeof(struct ieee_types_header), + (u8 *)bss_desc->bcn_ht_cap, le16_to_cpu(ht_cap->header.len)); mwifiex_fill_cap_info(priv, radio_type, ht_cap); From d51246481c7f28bbfa1f814ded2da65e531cd4b2 Mon Sep 17 00:00:00 2001 From: Amitkumar Karwar Date: Tue, 4 Mar 2014 18:43:14 -0800 Subject: [PATCH 084/259] mwifiex: save and copy AP's VHT capability info correctly While preparing association request, intersection of device's VHT capability information and corresponding field advertised by AP is used. This patch fixes a couple errors while saving and copying vht_cap and vht_oper fields from AP's beacon. Cc: # 3.9+ Signed-off-by: Amitkumar Karwar Signed-off-by: Bing Zhao Signed-off-by: John W. Linville --- drivers/net/wireless/mwifiex/11ac.c | 3 +-- drivers/net/wireless/mwifiex/scan.c | 8 ++++---- 2 files changed, 5 insertions(+), 6 deletions(-) diff --git a/drivers/net/wireless/mwifiex/11ac.c b/drivers/net/wireless/mwifiex/11ac.c index 5e0eec4d71c7..5d9a8084665d 100644 --- a/drivers/net/wireless/mwifiex/11ac.c +++ b/drivers/net/wireless/mwifiex/11ac.c @@ -189,8 +189,7 @@ int mwifiex_cmd_append_11ac_tlv(struct mwifiex_private *priv, vht_cap->header.len = cpu_to_le16(sizeof(struct ieee80211_vht_cap)); memcpy((u8 *)vht_cap + sizeof(struct mwifiex_ie_types_header), - (u8 *)bss_desc->bcn_vht_cap + - sizeof(struct ieee_types_header), + (u8 *)bss_desc->bcn_vht_cap, le16_to_cpu(vht_cap->header.len)); mwifiex_fill_vht_cap_tlv(priv, vht_cap, bss_desc->bss_band); diff --git a/drivers/net/wireless/mwifiex/scan.c b/drivers/net/wireless/mwifiex/scan.c index 0a8a26e10f01..668547c2de84 100644 --- a/drivers/net/wireless/mwifiex/scan.c +++ b/drivers/net/wireless/mwifiex/scan.c @@ -2101,12 +2101,12 @@ mwifiex_save_curr_bcn(struct mwifiex_private *priv) curr_bss->ht_info_offset); if (curr_bss->bcn_vht_cap) - curr_bss->bcn_ht_cap = (void *)(curr_bss->beacon_buf + - curr_bss->vht_cap_offset); + curr_bss->bcn_vht_cap = (void *)(curr_bss->beacon_buf + + curr_bss->vht_cap_offset); if (curr_bss->bcn_vht_oper) - curr_bss->bcn_ht_oper = (void *)(curr_bss->beacon_buf + - curr_bss->vht_info_offset); + curr_bss->bcn_vht_oper = (void *)(curr_bss->beacon_buf + + curr_bss->vht_info_offset); if (curr_bss->bcn_bss_co_2040) curr_bss->bcn_bss_co_2040 = From bb5016eac1656506df1a9d6057ce5bec342afbef Mon Sep 17 00:00:00 2001 From: Guillaume Nault Date: Thu, 6 Mar 2014 11:14:30 +0100 Subject: [PATCH 085/259] l2tp: fix manual sequencing (de)activation in L2TPv2 Commit e0d4435f "l2tp: Update PPP-over-L2TP driver to work over L2TPv3" broke the PPPOL2TP_SO_SENDSEQ setsockopt. The L2TP header length was previously computed by pppol2tp_l2t_header_len() before each call to l2tp_xmit_skb(). Now that header length is retrieved from the hdr_len session field, this field must be updated every time the L2TP header format is modified, or l2tp_xmit_skb() won't push the right amount of data for the L2TP header. This patch uses l2tp_session_set_header_len() to adjust hdr_len every time sequencing is (de)activated from userspace (either by the PPPOL2TP_SO_SENDSEQ setsockopt or the L2TP_ATTR_SEND_SEQ netlink attribute). Signed-off-by: Guillaume Nault Signed-off-by: David S. Miller --- net/l2tp/l2tp_core.c | 4 ++-- net/l2tp/l2tp_core.h | 1 + net/l2tp/l2tp_netlink.c | 4 +++- net/l2tp/l2tp_ppp.c | 1 + 4 files changed, 7 insertions(+), 3 deletions(-) diff --git a/net/l2tp/l2tp_core.c b/net/l2tp/l2tp_core.c index 735d0f60c83a..85d9d94c0a3c 100644 --- a/net/l2tp/l2tp_core.c +++ b/net/l2tp/l2tp_core.c @@ -112,7 +112,6 @@ struct l2tp_net { spinlock_t l2tp_session_hlist_lock; }; -static void l2tp_session_set_header_len(struct l2tp_session *session, int version); static void l2tp_tunnel_free(struct l2tp_tunnel *tunnel); static inline struct l2tp_tunnel *l2tp_tunnel(struct sock *sk) @@ -1863,7 +1862,7 @@ EXPORT_SYMBOL_GPL(l2tp_session_delete); /* We come here whenever a session's send_seq, cookie_len or * l2specific_len parameters are set. */ -static void l2tp_session_set_header_len(struct l2tp_session *session, int version) +void l2tp_session_set_header_len(struct l2tp_session *session, int version) { if (version == L2TP_HDR_VER_2) { session->hdr_len = 6; @@ -1876,6 +1875,7 @@ static void l2tp_session_set_header_len(struct l2tp_session *session, int versio } } +EXPORT_SYMBOL_GPL(l2tp_session_set_header_len); struct l2tp_session *l2tp_session_create(int priv_size, struct l2tp_tunnel *tunnel, u32 session_id, u32 peer_session_id, struct l2tp_session_cfg *cfg) { diff --git a/net/l2tp/l2tp_core.h b/net/l2tp/l2tp_core.h index 1f01ba3435bc..3f93ccd6ba97 100644 --- a/net/l2tp/l2tp_core.h +++ b/net/l2tp/l2tp_core.h @@ -263,6 +263,7 @@ void l2tp_recv_common(struct l2tp_session *session, struct sk_buff *skb, int length, int (*payload_hook)(struct sk_buff *skb)); int l2tp_session_queue_purge(struct l2tp_session *session); int l2tp_udp_encap_recv(struct sock *sk, struct sk_buff *skb); +void l2tp_session_set_header_len(struct l2tp_session *session, int version); int l2tp_xmit_skb(struct l2tp_session *session, struct sk_buff *skb, int hdr_len); diff --git a/net/l2tp/l2tp_netlink.c b/net/l2tp/l2tp_netlink.c index 4cfd722e9153..bd7387adea9e 100644 --- a/net/l2tp/l2tp_netlink.c +++ b/net/l2tp/l2tp_netlink.c @@ -578,8 +578,10 @@ static int l2tp_nl_cmd_session_modify(struct sk_buff *skb, struct genl_info *inf if (info->attrs[L2TP_ATTR_RECV_SEQ]) session->recv_seq = nla_get_u8(info->attrs[L2TP_ATTR_RECV_SEQ]); - if (info->attrs[L2TP_ATTR_SEND_SEQ]) + if (info->attrs[L2TP_ATTR_SEND_SEQ]) { session->send_seq = nla_get_u8(info->attrs[L2TP_ATTR_SEND_SEQ]); + l2tp_session_set_header_len(session, session->tunnel->version); + } if (info->attrs[L2TP_ATTR_LNS_MODE]) session->lns_mode = nla_get_u8(info->attrs[L2TP_ATTR_LNS_MODE]); diff --git a/net/l2tp/l2tp_ppp.c b/net/l2tp/l2tp_ppp.c index be5fadf34739..6bfeaa777135 100644 --- a/net/l2tp/l2tp_ppp.c +++ b/net/l2tp/l2tp_ppp.c @@ -1312,6 +1312,7 @@ static int pppol2tp_session_setsockopt(struct sock *sk, po->chan.hdrlen = val ? PPPOL2TP_L2TP_HDR_SIZE_SEQ : PPPOL2TP_L2TP_HDR_SIZE_NOSEQ; } + l2tp_session_set_header_len(session, session->tunnel->version); l2tp_info(session, PPPOL2TP_MSG_CONTROL, "%s: set send_seq=%d\n", session->name, session->send_seq); From 9e9cb6221aa7cb04765484fe87cc2d1b92edce64 Mon Sep 17 00:00:00 2001 From: Guillaume Nault Date: Thu, 6 Mar 2014 11:15:10 +0100 Subject: [PATCH 086/259] l2tp: fix userspace reception on plain L2TP sockets As pppol2tp_recv() never queues up packets to plain L2TP sockets, pppol2tp_recvmsg() never returns data to userspace, thus making the recv*() system calls unusable. Instead of dropping packets when the L2TP socket isn't bound to a PPP channel, this patch adds them to its reception queue. Signed-off-by: Guillaume Nault Signed-off-by: David S. Miller --- net/l2tp/l2tp_ppp.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/net/l2tp/l2tp_ppp.c b/net/l2tp/l2tp_ppp.c index 6bfeaa777135..5990919356a5 100644 --- a/net/l2tp/l2tp_ppp.c +++ b/net/l2tp/l2tp_ppp.c @@ -254,12 +254,14 @@ static void pppol2tp_recv(struct l2tp_session *session, struct sk_buff *skb, int po = pppox_sk(sk); ppp_input(&po->chan, skb); } else { - l2tp_info(session, PPPOL2TP_MSG_DATA, "%s: socket not bound\n", - session->name); + l2tp_dbg(session, PPPOL2TP_MSG_DATA, + "%s: recv %d byte data frame, passing to L2TP socket\n", + session->name, data_len); - /* Not bound. Nothing we can do, so discard. */ - atomic_long_inc(&session->stats.rx_errors); - kfree_skb(skb); + if (sock_queue_rcv_skb(sk, skb) < 0) { + atomic_long_inc(&session->stats.rx_errors); + kfree_skb(skb); + } } return; From 6d4ebeb4df0176b1973875840a9f7e91394c0685 Mon Sep 17 00:00:00 2001 From: Ying Xue Date: Thu, 6 Mar 2014 14:40:16 +0100 Subject: [PATCH 087/259] tipc: allow connection shutdown callback to be invoked in advance Currently connection shutdown callback function is called when connection instance is released in tipc_conn_kref_release(), and receiving packets and sending packets are running in different threads. Even if connection is closed by the thread of receiving packets, its shutdown callback may not be called immediately as the connection reference count is non-zero at that moment. So, although the connection is shut down by the thread of receiving packets, the thread of sending packets doesn't know it. Before its shutdown callback is invoked to tell the sending thread its connection has been closed, the sending thread may deliver messages by tipc_conn_sendmsg(), this is why the following error information appears: "Sending subscription event failed, no memory" To eliminate it, allow connection shutdown callback function to be called before connection id is removed in tipc_close_conn(), which makes the sending thread know the truth in time that its socket is closed so that it doesn't send message to it. We also remove the "Sending XXX failed..." error reporting for topology and config services. Signed-off-by: Ying Xue Signed-off-by: Erik Hugne Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/config.c | 9 ++------- net/tipc/server.c | 8 +++----- net/tipc/subscr.c | 8 ++------ 3 files changed, 7 insertions(+), 18 deletions(-) diff --git a/net/tipc/config.c b/net/tipc/config.c index e74eef2e7490..e6d721692ae0 100644 --- a/net/tipc/config.c +++ b/net/tipc/config.c @@ -376,7 +376,6 @@ static void cfg_conn_msg_event(int conid, struct sockaddr_tipc *addr, struct tipc_cfg_msg_hdr *req_hdr; struct tipc_cfg_msg_hdr *rep_hdr; struct sk_buff *rep_buf; - int ret; /* Validate configuration message header (ignore invalid message) */ req_hdr = (struct tipc_cfg_msg_hdr *)buf; @@ -398,12 +397,8 @@ static void cfg_conn_msg_event(int conid, struct sockaddr_tipc *addr, memcpy(rep_hdr, req_hdr, sizeof(*rep_hdr)); rep_hdr->tcm_len = htonl(rep_buf->len); rep_hdr->tcm_flags &= htons(~TCM_F_REQUEST); - - ret = tipc_conn_sendmsg(&cfgsrv, conid, addr, rep_buf->data, - rep_buf->len); - if (ret < 0) - pr_err("Sending cfg reply message failed, no memory\n"); - + tipc_conn_sendmsg(&cfgsrv, conid, addr, rep_buf->data, + rep_buf->len); kfree_skb(rep_buf); } } diff --git a/net/tipc/server.c b/net/tipc/server.c index 373979789a73..bbaa8f56024a 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -87,7 +87,6 @@ static void tipc_clean_outqueues(struct tipc_conn *con); static void tipc_conn_kref_release(struct kref *kref) { struct tipc_conn *con = container_of(kref, struct tipc_conn, kref); - struct tipc_server *s = con->server; if (con->sock) { tipc_sock_release_local(con->sock); @@ -95,10 +94,6 @@ static void tipc_conn_kref_release(struct kref *kref) } tipc_clean_outqueues(con); - - if (con->conid) - s->tipc_conn_shutdown(con->conid, con->usr_data); - kfree(con); } @@ -181,6 +176,9 @@ static void tipc_close_conn(struct tipc_conn *con) struct tipc_server *s = con->server; if (test_and_clear_bit(CF_CONNECTED, &con->flags)) { + if (con->conid) + s->tipc_conn_shutdown(con->conid, con->usr_data); + spin_lock_bh(&s->idr_lock); idr_remove(&s->conn_idr, con->conid); s->idr_in_use--; diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 7cb0bd5b1176..a6ce3bbf3eaf 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -96,20 +96,16 @@ static void subscr_send_event(struct tipc_subscription *sub, u32 found_lower, { struct tipc_subscriber *subscriber = sub->subscriber; struct kvec msg_sect; - int ret; msg_sect.iov_base = (void *)&sub->evt; msg_sect.iov_len = sizeof(struct tipc_event); - sub->evt.event = htohl(event, sub->swap); sub->evt.found_lower = htohl(found_lower, sub->swap); sub->evt.found_upper = htohl(found_upper, sub->swap); sub->evt.port.ref = htohl(port_ref, sub->swap); sub->evt.port.node = htohl(node, sub->swap); - ret = tipc_conn_sendmsg(&topsrv, subscriber->conid, NULL, - msg_sect.iov_base, msg_sect.iov_len); - if (ret < 0) - pr_err("Sending subscription event failed, no memory\n"); + tipc_conn_sendmsg(&topsrv, subscriber->conid, NULL, msg_sect.iov_base, + msg_sect.iov_len); } /** From 4652edb70e8a7eebbe47fa931940f65522c36e8f Mon Sep 17 00:00:00 2001 From: Ying Xue Date: Thu, 6 Mar 2014 14:40:17 +0100 Subject: [PATCH 088/259] tipc: fix connection refcount leak When tipc_conn_sendmsg() calls tipc_conn_lookup() to query a connection instance, its reference count value is increased if it's found. But subsequently if it's found that the connection is closed, the work of sending message is not queued into its server send workqueue, and the connection reference count is not decreased. This will cause a reference count leak. To reproduce this problem, an application would need to open and closes topology server connections with high intensity. We fix this by immediately decrementing the connection reference count if a send fails due to the connection being closed. Signed-off-by: Ying Xue Acked-by: Erik Hugne Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/server.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/net/tipc/server.c b/net/tipc/server.c index bbaa8f56024a..646a930eefbf 100644 --- a/net/tipc/server.c +++ b/net/tipc/server.c @@ -427,10 +427,12 @@ int tipc_conn_sendmsg(struct tipc_server *s, int conid, list_add_tail(&e->list, &con->outqueue); spin_unlock_bh(&con->outqueue_lock); - if (test_bit(CF_CONNECTED, &con->flags)) + if (test_bit(CF_CONNECTED, &con->flags)) { if (!queue_work(s->send_wq, &con->swork)) conn_put(con); - + } else { + conn_put(con); + } return 0; } From fe8e4649397915cf3b2ab0b695929a27e543967e Mon Sep 17 00:00:00 2001 From: Ying Xue Date: Thu, 6 Mar 2014 14:40:18 +0100 Subject: [PATCH 089/259] tipc: avoid to unnecessary process switch under non-block mode When messages are received via tipc socket under non-block mode, schedule_timeout() is called in tipc_wait_for_rcvmsg(), that is, the process of receiving messages will be scheduled once although timeout value passed to schedule_timeout() is 0. The same issue exists in accept()/wait_for_accept(). To avoid this unnecessary process switch, we only call schedule_timeout() if the timeout value is non-zero. Signed-off-by: Ying Xue Reviewed-by: Erik Hugne Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/socket.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/tipc/socket.c b/net/tipc/socket.c index a4cf274455aa..0ed0eaa62f29 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -997,7 +997,7 @@ static int tipc_wait_for_rcvmsg(struct socket *sock, long timeo) for (;;) { prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE); - if (skb_queue_empty(&sk->sk_receive_queue)) { + if (timeo && skb_queue_empty(&sk->sk_receive_queue)) { if (sock->state == SS_DISCONNECTING) { err = -ENOTCONN; break; @@ -1623,7 +1623,7 @@ static int tipc_wait_for_accept(struct socket *sock, long timeo) for (;;) { prepare_to_wait_exclusive(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE); - if (skb_queue_empty(&sk->sk_receive_queue)) { + if (timeo && skb_queue_empty(&sk->sk_receive_queue)) { release_sock(sk); timeo = schedule_timeout(timeo); lock_sock(sk); From edcc0511b5ee7235282a688cd604e3ae7f9e1fc9 Mon Sep 17 00:00:00 2001 From: Erik Hugne Date: Thu, 6 Mar 2014 14:40:19 +0100 Subject: [PATCH 090/259] tipc: drop subscriber connection id invalidation When a topology server subscriber is disconnected, the associated connection id is set to zero. A check vs zero is then done in the subscription timeout function to see if the subscriber have been shut down. This is unnecessary, because all subscription timers will be cancelled when a subscriber terminates. Setting the connection id to zero is actually harmful because id zero is the identity of the topology server listening socket, and can cause a race that leads to this socket being closed instead. Signed-off-by: Erik Hugne Acked-by: Ying Xue Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/subscr.c | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index a6ce3bbf3eaf..11c9ae00837d 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -149,14 +149,6 @@ static void subscr_timeout(struct tipc_subscription *sub) /* The spin lock per subscriber is used to protect its members */ spin_lock_bh(&subscriber->lock); - /* Validate if the connection related to the subscriber is - * closed (in case subscriber is terminating) - */ - if (subscriber->conid == 0) { - spin_unlock_bh(&subscriber->lock); - return; - } - /* Validate timeout (in case subscription is being cancelled) */ if (sub->timeout == TIPC_WAIT_FOREVER) { spin_unlock_bh(&subscriber->lock); @@ -211,9 +203,6 @@ static void subscr_release(struct tipc_subscriber *subscriber) spin_lock_bh(&subscriber->lock); - /* Invalidate subscriber reference */ - subscriber->conid = 0; - /* Destroy any existing subscriptions for subscriber */ list_for_each_entry_safe(sub, sub_temp, &subscriber->subscription_list, subscription_list) { From 1bb8dce57f4d15233688c68990852a10eb1cd79f Mon Sep 17 00:00:00 2001 From: Erik Hugne Date: Thu, 6 Mar 2014 14:40:20 +0100 Subject: [PATCH 091/259] tipc: fix memory leak during module removal When the TIPC module is removed, the tasklet handler is disabled before all other subsystems. This will cause lingering publications in the name table because the node_down tasklets responsible to clean up publications from an unreachable node will never run. When the name table is shut down, these publications are detected and an error message is logged: tipc: nametbl_stop(): orphaned hash chain detected This is actually a memory leak, introduced with commit 993b858e37b3120ee76d9957a901cca22312ffaa ("tipc: correct the order of stopping services at rmmod") Instead of just logging an error and leaking memory, we free the orphaned entries during nametable shutdown. Signed-off-by: Erik Hugne Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/name_table.c | 37 ++++++++++++++++++++++++++++++++++--- 1 file changed, 34 insertions(+), 3 deletions(-) diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 48302be175ce..042e8e3cabc0 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -941,17 +941,48 @@ int tipc_nametbl_init(void) return 0; } +/** + * tipc_purge_publications - remove all publications for a given type + * + * tipc_nametbl_lock must be held when calling this function + */ +static void tipc_purge_publications(struct name_seq *seq) +{ + struct publication *publ, *safe; + struct sub_seq *sseq; + struct name_info *info; + + if (!seq->sseqs) { + nameseq_delete_empty(seq); + return; + } + sseq = seq->sseqs; + info = sseq->info; + list_for_each_entry_safe(publ, safe, &info->zone_list, zone_list) { + tipc_nametbl_remove_publ(publ->type, publ->lower, publ->node, + publ->ref, publ->key); + } +} + void tipc_nametbl_stop(void) { u32 i; + struct name_seq *seq; + struct hlist_head *seq_head; + struct hlist_node *safe; - /* Verify name table is empty, then release it */ + /* Verify name table is empty and purge any lingering + * publications, then release the name table + */ write_lock_bh(&tipc_nametbl_lock); for (i = 0; i < TIPC_NAMETBL_SIZE; i++) { if (hlist_empty(&table.types[i])) continue; - pr_err("nametbl_stop(): orphaned hash chain detected\n"); - break; + seq_head = &table.types[i]; + hlist_for_each_entry_safe(seq, safe, seq_head, ns_list) { + tipc_purge_publications(seq); + } + continue; } kfree(table.types); table.types = NULL; From 2892505ea170094f982516bb38105eac45f274b1 Mon Sep 17 00:00:00 2001 From: Erik Hugne Date: Thu, 6 Mar 2014 14:40:21 +0100 Subject: [PATCH 092/259] tipc: don't log disabled tasklet handler errors Failure to schedule a TIPC tasklet with tipc_k_signal because the tasklet handler is disabled is not an error. It means TIPC is currently in the process of shutting down. We remove the error logging in this case. Signed-off-by: Erik Hugne Reviewed-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/handler.c | 1 - 1 file changed, 1 deletion(-) diff --git a/net/tipc/handler.c b/net/tipc/handler.c index e4bc8a296744..1fabf160501f 100644 --- a/net/tipc/handler.c +++ b/net/tipc/handler.c @@ -58,7 +58,6 @@ unsigned int tipc_k_signal(Handler routine, unsigned long argument) spin_lock_bh(&qitem_lock); if (!handler_enabled) { - pr_err("Signal request ignored by handler\n"); spin_unlock_bh(&qitem_lock); return -ENOPROTOOPT; } From e588e2f286ed7da011ed357c24c5b9a554e26595 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Thu, 6 Mar 2014 18:06:41 +0100 Subject: [PATCH 093/259] inet: frag: make sure forced eviction removes all frags Quoting Alexander Aring: While fragmentation and unloading of 6lowpan module I got this kernel Oops after few seconds: BUG: unable to handle kernel paging request at f88bbc30 [..] Modules linked in: ipv6 [last unloaded: 6lowpan] Call Trace: [] ? call_timer_fn+0x54/0xb3 [] ? process_timeout+0xa/0xa [] run_timer_softirq+0x140/0x15f Problem is that incomplete frags are still around after unload; when their frag expire timer fires, we get crash. When a netns is removed (also done when unloading module), inet_frag calls the evictor with 'force' argument to purge remaining frags. The evictor loop terminates when accounted memory ('work') drops to 0 or the lru-list becomes empty. However, the mem accounting is done via percpu counters and may not be accurate, i.e. loop may terminate prematurely. Alter evictor to only stop once the lru list is empty when force is requested. Reported-by: Phoebe Buckheister Reported-by: Alexander Aring Tested-by: Alexander Aring Signed-off-by: Florian Westphal Acked-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/inet_fragment.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index 322dcebfc588..3b01959bf4bb 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -208,7 +208,7 @@ int inet_frag_evictor(struct netns_frags *nf, struct inet_frags *f, bool force) } work = frag_mem_limit(nf) - nf->low_thresh; - while (work > 0) { + while (work > 0 || force) { spin_lock(&nf->lru_lock); if (list_empty(&nf->lru_list)) { From 4ae6e50c76def306d726a5d2678e88998ad5258e Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 4 Mar 2014 17:35:44 -0700 Subject: [PATCH 094/259] phy: fix compiler array bounds warning on settings[] With -Werror=array-bounds, gcc v4.7.x warns that in phy_find_valid(), the settings[] "array subscript is above array bounds", I think because idx is a signed integer and if the caller supplied idx < 0, we pass the guard but still reference out of bounds. Fix this by making idx unsigned here and elsewhere. Signed-off-by: Bjorn Helgaas Acked-by: Florian Fainelli Signed-off-by: David S. Miller --- drivers/net/phy/phy.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 19c9eca0ef26..76d96b9ebcdb 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -164,9 +164,9 @@ static const struct phy_setting settings[] = { * of that setting. Returns the index of the last setting if * none of the others match. */ -static inline int phy_find_setting(int speed, int duplex) +static inline unsigned int phy_find_setting(int speed, int duplex) { - int idx = 0; + unsigned int idx = 0; while (idx < ARRAY_SIZE(settings) && (settings[idx].speed != speed || settings[idx].duplex != duplex)) @@ -185,7 +185,7 @@ static inline int phy_find_setting(int speed, int duplex) * the mask in features. Returns the index of the last setting * if nothing else matches. */ -static inline int phy_find_valid(int idx, u32 features) +static inline unsigned int phy_find_valid(unsigned int idx, u32 features) { while (idx < MAX_NUM_SETTINGS && !(settings[idx].setting & features)) idx++; @@ -204,7 +204,7 @@ static inline int phy_find_valid(int idx, u32 features) static void phy_sanitize_settings(struct phy_device *phydev) { u32 features = phydev->supported; - int idx; + unsigned int idx; /* Sanitize settings based on PHY capabilities */ if ((features & SUPPORTED_Autoneg) == 0) @@ -954,7 +954,8 @@ int phy_init_eee(struct phy_device *phydev, bool clk_stop_enable) (phydev->interface == PHY_INTERFACE_MODE_RGMII))) { int eee_lp, eee_cap, eee_adv; u32 lp, cap, adv; - int idx, status; + int status; + unsigned int idx; /* Read phy status to properly get the right settings */ status = phy_read_status(phydev); From adca4767821e54c72d4a2f467af77923f2c87e07 Mon Sep 17 00:00:00 2001 From: Andrew Lutomirski Date: Tue, 4 Mar 2014 17:24:10 -0800 Subject: [PATCH 095/259] net: Improve SO_TIMESTAMPING documentation and fix a minor code bug The original documentation was very unclear. The code fix is presumably related to the formerly unclear documentation: SOCK_TIMESTAMPING_RX_SOFTWARE has no effect on __sock_recv_timestamp's behavior, so calling __sock_recv_ts_and_drops from sock_recv_ts_and_drops if only SOCK_TIMESTAMPING_RX_SOFTWARE is set is pointless. This should have no user-observable effect. Signed-off-by: Andy Lutomirski Acked-by: Richard Cochran Signed-off-by: David S. Miller --- Documentation/networking/timestamping.txt | 48 ++++++++++++++--------- include/net/sock.h | 1 - 2 files changed, 30 insertions(+), 19 deletions(-) diff --git a/Documentation/networking/timestamping.txt b/Documentation/networking/timestamping.txt index 661d3c316a17..048c92b487f6 100644 --- a/Documentation/networking/timestamping.txt +++ b/Documentation/networking/timestamping.txt @@ -21,26 +21,38 @@ has such a feature). SO_TIMESTAMPING: -Instructs the socket layer which kind of information is wanted. The -parameter is an integer with some of the following bits set. Setting -other bits is an error and doesn't change the current state. +Instructs the socket layer which kind of information should be collected +and/or reported. The parameter is an integer with some of the following +bits set. Setting other bits is an error and doesn't change the current +state. -SOF_TIMESTAMPING_TX_HARDWARE: try to obtain send time stamp in hardware -SOF_TIMESTAMPING_TX_SOFTWARE: if SOF_TIMESTAMPING_TX_HARDWARE is off or - fails, then do it in software -SOF_TIMESTAMPING_RX_HARDWARE: return the original, unmodified time stamp - as generated by the hardware -SOF_TIMESTAMPING_RX_SOFTWARE: if SOF_TIMESTAMPING_RX_HARDWARE is off or - fails, then do it in software -SOF_TIMESTAMPING_RAW_HARDWARE: return original raw hardware time stamp -SOF_TIMESTAMPING_SYS_HARDWARE: return hardware time stamp transformed to - the system time base -SOF_TIMESTAMPING_SOFTWARE: return system time stamp generated in - software +Four of the bits are requests to the stack to try to generate +timestamps. Any combination of them is valid. -SOF_TIMESTAMPING_TX/RX determine how time stamps are generated. -SOF_TIMESTAMPING_RAW/SYS determine how they are reported in the -following control message: +SOF_TIMESTAMPING_TX_HARDWARE: try to obtain send time stamps in hardware +SOF_TIMESTAMPING_TX_SOFTWARE: try to obtain send time stamps in software +SOF_TIMESTAMPING_RX_HARDWARE: try to obtain receive time stamps in hardware +SOF_TIMESTAMPING_RX_SOFTWARE: try to obtain receive time stamps in software + +The other three bits control which timestamps will be reported in a +generated control message. If none of these bits are set or if none of +the set bits correspond to data that is available, then the control +message will not be generated: + +SOF_TIMESTAMPING_SOFTWARE: report systime if available +SOF_TIMESTAMPING_SYS_HARDWARE: report hwtimetrans if available +SOF_TIMESTAMPING_RAW_HARDWARE: report hwtimeraw if available + +It is worth noting that timestamps may be collected for reasons other +than being requested by a particular socket with +SOF_TIMESTAMPING_[TR]X_(HARD|SOFT)WARE. For example, most drivers that +can generate hardware receive timestamps ignore +SOF_TIMESTAMPING_RX_HARDWARE. It is still a good idea to set that flag +in case future drivers pay attention. + +If timestamps are reported, they will appear in a control message with +cmsg_level==SOL_SOCKET, cmsg_type==SO_TIMESTAMPING, and a payload like +this: struct scm_timestamping { struct timespec systime; diff --git a/include/net/sock.h b/include/net/sock.h index 5c3f7c3624aa..7c4167bc8266 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -2186,7 +2186,6 @@ static inline void sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk, { #define FLAGS_TS_OR_DROPS ((1UL << SOCK_RXQ_OVFL) | \ (1UL << SOCK_RCVTSTAMP) | \ - (1UL << SOCK_TIMESTAMPING_RX_SOFTWARE) | \ (1UL << SOCK_TIMESTAMPING_SOFTWARE) | \ (1UL << SOCK_TIMESTAMPING_RAW_HARDWARE) | \ (1UL << SOCK_TIMESTAMPING_SYS_HARDWARE)) From 0a13404dd3bf4ea870e3d96270b5a382edca85c0 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 5 Mar 2014 14:29:58 +1100 Subject: [PATCH 096/259] net: unix socket code abuses csum_partial The unix socket code is using the result of csum_partial to hash into a lookup table: unix_hash_fold(csum_partial(sunaddr, len, 0)); csum_partial is only guaranteed to produce something that can be folded into a checksum, as its prototype explains: * returns a 32-bit number suitable for feeding into itself * or csum_tcpudp_magic The 32bit value should not be used directly. Depending on the alignment, the ppc64 csum_partial will return different 32bit partial checksums that will fold into the same 16bit checksum. This difference causes the following testcase (courtesy of Gustavo) to sometimes fail: #include #include int main() { int fd = socket(PF_LOCAL, SOCK_STREAM|SOCK_CLOEXEC, 0); int i = 1; setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &i, 4); struct sockaddr addr; addr.sa_family = AF_LOCAL; bind(fd, &addr, 2); listen(fd, 128); struct sockaddr_storage ss; socklen_t sslen = (socklen_t)sizeof(ss); getsockname(fd, (struct sockaddr*)&ss, &sslen); fd = socket(PF_LOCAL, SOCK_STREAM|SOCK_CLOEXEC, 0); if (connect(fd, (struct sockaddr*)&ss, sslen) == -1){ perror(NULL); return 1; } printf("OK\n"); return 0; } As suggested by davem, fix this by using csum_fold to fold the partial 32bit checksum into a 16bit checksum before using it. Signed-off-by: Anton Blanchard Cc: stable@vger.kernel.org Signed-off-by: David S. Miller --- net/unix/af_unix.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index 29fc8bee9702..ce6ec6c2f4de 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -163,9 +163,8 @@ static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb) static inline unsigned int unix_hash_fold(__wsum n) { - unsigned int hash = (__force unsigned int)n; + unsigned int hash = (__force unsigned int)csum_fold(n); - hash ^= hash>>16; hash ^= hash>>8; return hash&(UNIX_HASH_SIZE-1); } From d746ca9561440685edb62614d1bcbbc27ff50e66 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 5 Mar 2014 14:51:37 +1100 Subject: [PATCH 097/259] ibmveth: Fix endian issues with MAC addresses The code to load a MAC address into a u64 for passing to the hypervisor via a register is broken on little endian. Create a helper function called ibmveth_encode_mac_addr which does the right thing in both big and little endian. We were storing the MAC address in a long in struct ibmveth_adapter. It's never used so remove it - we don't need another place in the driver where we create endian issues with MAC addresses. Signed-off-by: Anton Blanchard Cc: stable@vger.kernel.org Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/ibmveth.c | 25 ++++++++++++++++--------- drivers/net/ethernet/ibm/ibmveth.h | 1 - 2 files changed, 16 insertions(+), 10 deletions(-) diff --git a/drivers/net/ethernet/ibm/ibmveth.c b/drivers/net/ethernet/ibm/ibmveth.c index 4be971590461..1fc8334fc181 100644 --- a/drivers/net/ethernet/ibm/ibmveth.c +++ b/drivers/net/ethernet/ibm/ibmveth.c @@ -522,10 +522,21 @@ static int ibmveth_register_logical_lan(struct ibmveth_adapter *adapter, return rc; } +static u64 ibmveth_encode_mac_addr(u8 *mac) +{ + int i; + u64 encoded = 0; + + for (i = 0; i < ETH_ALEN; i++) + encoded = (encoded << 8) | mac[i]; + + return encoded; +} + static int ibmveth_open(struct net_device *netdev) { struct ibmveth_adapter *adapter = netdev_priv(netdev); - u64 mac_address = 0; + u64 mac_address; int rxq_entries = 1; unsigned long lpar_rc; int rc; @@ -579,8 +590,7 @@ static int ibmveth_open(struct net_device *netdev) adapter->rx_queue.num_slots = rxq_entries; adapter->rx_queue.toggle = 1; - memcpy(&mac_address, netdev->dev_addr, netdev->addr_len); - mac_address = mac_address >> 16; + mac_address = ibmveth_encode_mac_addr(netdev->dev_addr); rxq_desc.fields.flags_len = IBMVETH_BUF_VALID | adapter->rx_queue.queue_len; @@ -1183,8 +1193,8 @@ static void ibmveth_set_multicast_list(struct net_device *netdev) /* add the addresses to the filter table */ netdev_for_each_mc_addr(ha, netdev) { /* add the multicast address to the filter table */ - unsigned long mcast_addr = 0; - memcpy(((char *)&mcast_addr)+2, ha->addr, ETH_ALEN); + u64 mcast_addr; + mcast_addr = ibmveth_encode_mac_addr(ha->addr); lpar_rc = h_multicast_ctrl(adapter->vdev->unit_address, IbmVethMcastAddFilter, mcast_addr); @@ -1372,9 +1382,6 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id) netif_napi_add(netdev, &adapter->napi, ibmveth_poll, 16); - adapter->mac_addr = 0; - memcpy(&adapter->mac_addr, mac_addr_p, ETH_ALEN); - netdev->irq = dev->irq; netdev->netdev_ops = &ibmveth_netdev_ops; netdev->ethtool_ops = &netdev_ethtool_ops; @@ -1383,7 +1390,7 @@ static int ibmveth_probe(struct vio_dev *dev, const struct vio_device_id *id) NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM; netdev->features |= netdev->hw_features; - memcpy(netdev->dev_addr, &adapter->mac_addr, netdev->addr_len); + memcpy(netdev->dev_addr, mac_addr_p, ETH_ALEN); for (i = 0; i < IBMVETH_NUM_BUFF_POOLS; i++) { struct kobject *kobj = &adapter->rx_buff_pool[i].kobj; diff --git a/drivers/net/ethernet/ibm/ibmveth.h b/drivers/net/ethernet/ibm/ibmveth.h index 451ba7949e15..1f37499d4398 100644 --- a/drivers/net/ethernet/ibm/ibmveth.h +++ b/drivers/net/ethernet/ibm/ibmveth.h @@ -138,7 +138,6 @@ struct ibmveth_adapter { struct napi_struct napi; struct net_device_stats stats; unsigned int mcastFilterSize; - unsigned long mac_addr; void * buffer_list_addr; void * filter_list_addr; dma_addr_t buffer_list_dma; From d2d273ffabd315eecefce21a4391d44b6e156b73 Mon Sep 17 00:00:00 2001 From: Anton Nayshtut Date: Wed, 5 Mar 2014 08:30:08 +0200 Subject: [PATCH 098/259] ipv6: Fix exthdrs offload registration. Without this fix, ipv6_exthdrs_offload_init doesn't register IPPROTO_DSTOPTS offload, but returns 0 (as the IPPROTO_ROUTING registration actually succeeds). This then causes the ipv6_gso_segment to drop IPv6 packets with IPPROTO_DSTOPTS header. The issue detected and the fix verified by running MS HCK Offload LSO test on top of QEMU Windows guests, as this test sends IPv6 packets with IPPROTO_DSTOPTS. Signed-off-by: Anton Nayshtut Signed-off-by: David S. Miller --- net/ipv6/exthdrs_offload.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/ipv6/exthdrs_offload.c b/net/ipv6/exthdrs_offload.c index cf77f3abfd06..447a7fbd1bb6 100644 --- a/net/ipv6/exthdrs_offload.c +++ b/net/ipv6/exthdrs_offload.c @@ -25,11 +25,11 @@ int __init ipv6_exthdrs_offload_init(void) int ret; ret = inet6_add_offload(&rthdr_offload, IPPROTO_ROUTING); - if (!ret) + if (ret) goto out; ret = inet6_add_offload(&dstopt_offload, IPPROTO_DSTOPTS); - if (!ret) + if (ret) goto out_rt; out: From 5bd4e4c158ceb4e76ce9ed005c876d59caad8af2 Mon Sep 17 00:00:00 2001 From: Sasha Levin Date: Thu, 6 Mar 2014 16:53:11 -0500 Subject: [PATCH 099/259] bonding: correctly handle out of range parameters for lp_interval We didn't correctly check cases where the value for lp_interval is not within the legal range due to a missing table terminator. This would let userspace trigger a kernel panic by specifying a value out of range: echo -1 > /sys/devices/virtual/net/bond0/bonding/lp_interval Introduced by commit 4325b374f84 ("bonding: convert lp_interval to use the new option API"). Acked-by: Nikolay Aleksandrov Signed-off-by: Sasha Levin Signed-off-by: David S. Miller --- drivers/net/bonding/bond_options.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/bonding/bond_options.c b/drivers/net/bonding/bond_options.c index c37878432717..298c26509095 100644 --- a/drivers/net/bonding/bond_options.c +++ b/drivers/net/bonding/bond_options.c @@ -121,6 +121,7 @@ static struct bond_opt_value bond_resend_igmp_tbl[] = { static struct bond_opt_value bond_lp_interval_tbl[] = { { "minval", 1, BOND_VALFLAG_MIN | BOND_VALFLAG_DEFAULT}, { "maxval", INT_MAX, BOND_VALFLAG_MAX}, + { NULL, -1, 0}, }; static struct bond_option bond_opts[] = { From 57352ef4f5f19969a50d42e84b274287993b576f Mon Sep 17 00:00:00 2001 From: Amir Vadai Date: Thu, 6 Mar 2014 18:28:16 +0200 Subject: [PATCH 100/259] net/mlx4_core: Fix memory access error in mlx4_QUERY_DEV_CAP_wrapper() Fix a regression introduced by [1]. outbox was accessed instead of outbox->buf. Typo was copy-pasted to [2] and [3]. [1] - cc1ade9 mlx4_core: Disable memory windows for virtual functions [2] - 4de6580 mlx4_core: Add support for steerable IB UD QPs [3] - 7ffdf72 net/mlx4_core: Add basic support for TCP/IP offloads under tunneling Signed-off-by: Or Gerlitz Signed-off-by: Amir Vadai Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/fw.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c index 91b69ff4b4a2..8726e34cee22 100644 --- a/drivers/net/ethernet/mellanox/mlx4/fw.c +++ b/drivers/net/ethernet/mellanox/mlx4/fw.c @@ -859,7 +859,7 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave, MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_CQ_TS_SUPPORT_OFFSET); /* For guests, disable vxlan tunneling */ - MLX4_GET(field, outbox, QUERY_DEV_CAP_VXLAN); + MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_VXLAN); field &= 0xf7; MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_VXLAN); @@ -869,7 +869,7 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave, MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_BF_OFFSET); /* For guests, disable mw type 2 */ - MLX4_GET(bmme_flags, outbox, QUERY_DEV_CAP_BMME_FLAGS_OFFSET); + MLX4_GET(bmme_flags, outbox->buf, QUERY_DEV_CAP_BMME_FLAGS_OFFSET); bmme_flags &= ~MLX4_BMME_FLAG_TYPE_2_WIN; MLX4_PUT(outbox->buf, bmme_flags, QUERY_DEV_CAP_BMME_FLAGS_OFFSET); @@ -883,7 +883,7 @@ int mlx4_QUERY_DEV_CAP_wrapper(struct mlx4_dev *dev, int slave, } /* turn off ipoib managed steering for guests */ - MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET); + MLX4_GET(field, outbox->buf, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET); field &= ~0x80; MLX4_PUT(outbox->buf, field, QUERY_DEV_CAP_FLOW_STEERING_IPOIB_OFFSET); From 97989356af0ec8b1b1658d804892abb354127330 Mon Sep 17 00:00:00 2001 From: Amir Vadai Date: Thu, 6 Mar 2014 18:28:17 +0200 Subject: [PATCH 101/259] net/mlx4_core: mlx4_init_slave() shouldn't access comm channel before PF is ready Currently, the PF call to pci_enable_sriov from the PF probe function stalls for 10 seconds times the number of VFs probed on the host. This happens because the way for such VFs to determine of the PF initialization finished, is by attempting to issue reset on the comm-channel and get timeout (after 10s). The PF probe function is called from a kenernel workqueue, and therefore during that time, rcu lock is being held and kernel's workqueue is stalled. This blocks other processes that try to use the workqueue or rcu lock. For example, interface renaming which is calling rcu_synchronize is blocked, and timedout by systemd. Changed mlx4_init_slave() to allow VF probed on the host to immediatly detect that the PF is not ready, and return EPROBE_DEFER instantly. Only when the PF finishes the initialization, allow such VFs to access the comm channel. This issue and fix are relevant only for probed VFs on the hypervisor, there is no way to pass this information to a VM until comm channel is ready, so in a VM, if PF is not ready, the first command will be timedout after 10 seconds and return EPROBE_DEFER. Signed-off-by: Amir Vadai Signed-off-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/main.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c index 5a6105f1ba6d..30a08a60f059 100644 --- a/drivers/net/ethernet/mellanox/mlx4/main.c +++ b/drivers/net/ethernet/mellanox/mlx4/main.c @@ -150,6 +150,8 @@ struct mlx4_port_config { struct pci_dev *pdev; }; +static atomic_t pf_loading = ATOMIC_INIT(0); + int mlx4_check_port_params(struct mlx4_dev *dev, enum mlx4_port_type *port_type) { @@ -1407,6 +1409,11 @@ static int mlx4_init_slave(struct mlx4_dev *dev) u32 slave_read; u32 cmd_channel_ver; + if (atomic_read(&pf_loading)) { + mlx4_warn(dev, "PF is not ready. Deferring probe\n"); + return -EPROBE_DEFER; + } + mutex_lock(&priv->cmd.slave_cmd_mutex); priv->cmd.max_cmds = 1; mlx4_warn(dev, "Sending reset\n"); @@ -2319,7 +2326,11 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data) if (num_vfs) { mlx4_warn(dev, "Enabling SR-IOV with %d VFs\n", num_vfs); + + atomic_inc(&pf_loading); err = pci_enable_sriov(pdev, num_vfs); + atomic_dec(&pf_loading); + if (err) { mlx4_err(dev, "Failed to enable SR-IOV, continuing without SR-IOV (err = %d).\n", err); From c88507fbad8055297c1d1e21e599f46960cbee39 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Thu, 6 Mar 2014 17:51:57 +0100 Subject: [PATCH 102/259] ipv6: don't set DST_NOCOUNT for remotely added routes DST_NOCOUNT should only be used if an authorized user adds routes locally. In case of routes which are added on behalf of router advertisments this flag must not get used as it allows an unlimited number of routes getting added remotely. Signed-off-by: Sabrina Dubroca Acked-by: Hannes Frederic Sowa Signed-off-by: David S. Miller --- net/ipv6/route.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 11dac21e6586..fba54a407bb2 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -1513,7 +1513,7 @@ int ip6_route_add(struct fib6_config *cfg) if (!table) goto out; - rt = ip6_dst_alloc(net, NULL, DST_NOCOUNT, table); + rt = ip6_dst_alloc(net, NULL, (cfg->fc_flags & RTF_ADDRCONF) ? 0 : DST_NOCOUNT, table); if (!rt) { err = -ENOMEM; From 9b745ab897199c2af9f21ca9681ef86d5b971002 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Fri, 7 Mar 2014 08:37:19 +0100 Subject: [PATCH 103/259] ALSA: hda - Fix loud click noise with IdeaPad 410Y Lenovo IdeaPad 410Y with ALC282 codec makes loud click noises at boot and shutdown. Also, it wrongly misdetects the acpi_thinkpad hook. This patch adds a device-specific fixup for disabling the shutup callback that is the cause of the click noise and also avoiding the thinpad_helper calls. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=71511 Reported-and-tested-by: Guilherme Amadio Cc: Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_realtek.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 850296a1e0ff..8d0a84436674 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -3616,6 +3616,19 @@ static void alc_fixup_auto_mute_via_amp(struct hda_codec *codec, } } +static void alc_no_shutup(struct hda_codec *codec) +{ +} + +static void alc_fixup_no_shutup(struct hda_codec *codec, + const struct hda_fixup *fix, int action) +{ + if (action == HDA_FIXUP_ACT_PRE_PROBE) { + struct alc_spec *spec = codec->spec; + spec->shutup = alc_no_shutup; + } +} + static void alc_fixup_headset_mode_alc668(struct hda_codec *codec, const struct hda_fixup *fix, int action) { @@ -3844,6 +3857,7 @@ enum { ALC269_FIXUP_HP_GPIO_LED, ALC269_FIXUP_INV_DMIC, ALC269_FIXUP_LENOVO_DOCK, + ALC269_FIXUP_NO_SHUTUP, ALC286_FIXUP_SONY_MIC_NO_PRESENCE, ALC269_FIXUP_PINCFG_NO_HP_TO_LINEOUT, ALC269_FIXUP_DELL1_MIC_NO_PRESENCE, @@ -4020,6 +4034,10 @@ static const struct hda_fixup alc269_fixups[] = { .type = HDA_FIXUP_FUNC, .v.func = alc_fixup_inv_dmic_0x12, }, + [ALC269_FIXUP_NO_SHUTUP] = { + .type = HDA_FIXUP_FUNC, + .v.func = alc_fixup_no_shutup, + }, [ALC269_FIXUP_LENOVO_DOCK] = { .type = HDA_FIXUP_PINS, .v.pins = (const struct hda_pintbl[]) { @@ -4405,6 +4423,7 @@ static const struct snd_pci_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x17aa, 0x2212, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST), SND_PCI_QUIRK(0x17aa, 0x2214, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST), SND_PCI_QUIRK(0x17aa, 0x2215, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST), + SND_PCI_QUIRK(0x17aa, 0x3978, "IdeaPad Y410P", ALC269_FIXUP_NO_SHUTUP), SND_PCI_QUIRK(0x17aa, 0x5013, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST), SND_PCI_QUIRK(0x17aa, 0x501a, "Thinkpad", ALC283_FIXUP_INT_MIC), SND_PCI_QUIRK(0x17aa, 0x5026, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST), From d211f177b28ec070c25b3d0b960aa55f352f731f Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sat, 8 Mar 2014 15:31:54 -0800 Subject: [PATCH 104/259] audit: Update kdoc for audit_send_reply and audit_list_rules_send The kbuild test robot reported: > tree: git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace.git for-next > head: 6f285b19d09f72e801525f5eea1bdad22e559bf0 > commit: 6f285b19d09f72e801525f5eea1bdad22e559bf0 [2/2] audit: Send replies in the proper network namespace. > reproduce: make htmldocs > > >> Warning(kernel/audit.c:575): No description found for parameter 'request_skb' > >> Warning(kernel/audit.c:575): Excess function parameter 'portid' description in 'audit_send_reply' > >> Warning(kernel/auditfilter.c:1074): No description found for parameter 'request_skb' > >> Warning(kernel/auditfilter.c:1074): Excess function parameter 'portid' description in 'audit_list_rules_s Which was caused by my failure to update the kdoc annotations when I updated the functions. Fix that small oversight now. Signed-off-by: "Eric W. Biederman" --- kernel/audit.c | 2 +- kernel/auditfilter.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/audit.c b/kernel/audit.c index 32086bff5564..3392d3e0254a 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -559,7 +559,7 @@ static int audit_send_reply_thread(void *arg) } /** * audit_send_reply - send an audit reply message via netlink - * @portid: netlink port to which to send reply + * @request_skb: skb of request we are replying to (used to target the reply) * @seq: sequence number * @type: audit message type * @done: done (last) flag diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index e8d1c7c515d7..92062fd6cc8c 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -1067,7 +1067,7 @@ int audit_rule_change(int type, __u32 portid, int seq, void *data, /** * audit_list_rules_send - list the audit rules - * @portid: target portid for netlink audit messages + * @request_skb: skb of request we are replying to (used to target the reply) * @seq: netlink audit message sequence (serial) number */ int audit_list_rules_send(struct sk_buff *request_skb, int seq) From a8d9bc2e9f5d1c5a25e33cec096d2a1652d3fd52 Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Sun, 9 Mar 2014 15:45:32 -0800 Subject: [PATCH 105/259] bnx2: Fix shutdown sequence The pci shutdown handler added in: bnx2: Add pci shutdown handler commit 25bfb1dd4ba3b2d9a49ce9d9b0cd7be1840e15ed created a shutdown down sequence without chip reset if the device was never brought up. This can cause the firmware to shutdown the PHY prematurely and cause MMIO read cycles to be unresponsive. On some systems, it may generate NMI in the bnx2's pci shutdown handler. The fix is to tell the firmware not to shutdown the PHY if there was no prior chip reset. Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2.c | 37 +++++++++++++++++++++++++--- drivers/net/ethernet/broadcom/bnx2.h | 5 ++++ 2 files changed, 38 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnx2.c b/drivers/net/ethernet/broadcom/bnx2.c index cda25ac45b47..6c9e1c9bdeb8 100644 --- a/drivers/net/ethernet/broadcom/bnx2.c +++ b/drivers/net/ethernet/broadcom/bnx2.c @@ -2507,6 +2507,7 @@ bnx2_fw_sync(struct bnx2 *bp, u32 msg_data, int ack, int silent) bp->fw_wr_seq++; msg_data |= bp->fw_wr_seq; + bp->fw_last_msg = msg_data; bnx2_shmem_wr(bp, BNX2_DRV_MB, msg_data); @@ -4000,8 +4001,23 @@ bnx2_setup_wol(struct bnx2 *bp) wol_msg = BNX2_DRV_MSG_CODE_SUSPEND_NO_WOL; } - if (!(bp->flags & BNX2_FLAG_NO_WOL)) - bnx2_fw_sync(bp, BNX2_DRV_MSG_DATA_WAIT3 | wol_msg, 1, 0); + if (!(bp->flags & BNX2_FLAG_NO_WOL)) { + u32 val; + + wol_msg |= BNX2_DRV_MSG_DATA_WAIT3; + if (bp->fw_last_msg || BNX2_CHIP(bp) != BNX2_CHIP_5709) { + bnx2_fw_sync(bp, wol_msg, 1, 0); + return; + } + /* Tell firmware not to power down the PHY yet, otherwise + * the chip will take a long time to respond to MMIO reads. + */ + val = bnx2_shmem_rd(bp, BNX2_PORT_FEATURE); + bnx2_shmem_wr(bp, BNX2_PORT_FEATURE, + val | BNX2_PORT_FEATURE_ASF_ENABLED); + bnx2_fw_sync(bp, wol_msg, 1, 0); + bnx2_shmem_wr(bp, BNX2_PORT_FEATURE, val); + } } @@ -4033,9 +4049,22 @@ bnx2_set_power_state(struct bnx2 *bp, pci_power_t state) if (bp->wol) pci_set_power_state(bp->pdev, PCI_D3hot); - } else { - pci_set_power_state(bp->pdev, PCI_D3hot); + break; + } + if (!bp->fw_last_msg && BNX2_CHIP(bp) == BNX2_CHIP_5709) { + u32 val; + + /* Tell firmware not to power down the PHY yet, + * otherwise the other port may not respond to + * MMIO reads. + */ + val = bnx2_shmem_rd(bp, BNX2_BC_STATE_CONDITION); + val &= ~BNX2_CONDITION_PM_STATE_MASK; + val |= BNX2_CONDITION_PM_STATE_UNPREP; + bnx2_shmem_wr(bp, BNX2_BC_STATE_CONDITION, val); + } + pci_set_power_state(bp->pdev, PCI_D3hot); /* No more memory access after this point until * device is brought back to D0. diff --git a/drivers/net/ethernet/broadcom/bnx2.h b/drivers/net/ethernet/broadcom/bnx2.h index f1cf2c44e7ed..e341bc366fa5 100644 --- a/drivers/net/ethernet/broadcom/bnx2.h +++ b/drivers/net/ethernet/broadcom/bnx2.h @@ -6900,6 +6900,7 @@ struct bnx2 { u16 fw_wr_seq; u16 fw_drv_pulse_wr_seq; + u32 fw_last_msg; int rx_max_ring; int rx_ring_size; @@ -7406,6 +7407,10 @@ struct bnx2_rv2p_fw_file { #define BNX2_CONDITION_MFW_RUN_NCSI 0x00006000 #define BNX2_CONDITION_MFW_RUN_NONE 0x0000e000 #define BNX2_CONDITION_MFW_RUN_MASK 0x0000e000 +#define BNX2_CONDITION_PM_STATE_MASK 0x00030000 +#define BNX2_CONDITION_PM_STATE_FULL 0x00030000 +#define BNX2_CONDITION_PM_STATE_PREP 0x00020000 +#define BNX2_CONDITION_PM_STATE_UNPREP 0x00010000 #define BNX2_BC_STATE_DEBUG_CMD 0x1dc #define BNX2_BC_STATE_BC_DBG_CMD_SIGNATURE 0x42440000 From f324777ea88bab2522602671e46fc0851d7d5e35 Mon Sep 17 00:00:00 2001 From: Chad Dupuis Date: Wed, 26 Feb 2014 04:15:14 -0500 Subject: [PATCH 106/259] [SCSI] qla2xxx: Fix multiqueue MSI-X registration. This fixes requesting of the MSI-X vectors for the base response queue. The iteration in the for loop in qla24xx_enable_msix() was incorrect. We should only iterate of the first two MSI-X vectors and not the total number of MSI-X vectors that have given to the driver for this device from pci_enable_msix() in this function. Cc: Signed-off-by: Chad Dupuis Signed-off-by: Saurav Kashyap Signed-off-by: James Bottomley --- drivers/scsi/qla2xxx/qla_isr.c | 46 ++++++++++++++++++++++------------ 1 file changed, 30 insertions(+), 16 deletions(-) diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c index 9bc86b9e86b1..0a1dcb43d18b 100644 --- a/drivers/scsi/qla2xxx/qla_isr.c +++ b/drivers/scsi/qla2xxx/qla_isr.c @@ -2880,6 +2880,7 @@ static int qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) { #define MIN_MSIX_COUNT 2 +#define ATIO_VECTOR 2 int i, ret; struct msix_entry *entries; struct qla_msix_entry *qentry; @@ -2936,34 +2937,47 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp) } /* Enable MSI-X vectors for the base queue */ - for (i = 0; i < ha->msix_count; i++) { + for (i = 0; i < 2; i++) { qentry = &ha->msix_entries[i]; - if (QLA_TGT_MODE_ENABLED() && IS_ATIO_MSIX_CAPABLE(ha)) { - ret = request_irq(qentry->vector, - qla83xx_msix_entries[i].handler, - 0, qla83xx_msix_entries[i].name, rsp); - } else if (IS_P3P_TYPE(ha)) { + if (IS_P3P_TYPE(ha)) ret = request_irq(qentry->vector, qla82xx_msix_entries[i].handler, 0, qla82xx_msix_entries[i].name, rsp); - } else { + else ret = request_irq(qentry->vector, msix_entries[i].handler, 0, msix_entries[i].name, rsp); - } - if (ret) { - ql_log(ql_log_fatal, vha, 0x00cb, - "MSI-X: unable to register handler -- %x/%d.\n", - qentry->vector, ret); - qla24xx_disable_msix(ha); - ha->mqenable = 0; - goto msix_out; - } + if (ret) + goto msix_register_fail; qentry->have_irq = 1; qentry->rsp = rsp; rsp->msix = qentry; } + /* + * If target mode is enable, also request the vector for the ATIO + * queue. + */ + if (QLA_TGT_MODE_ENABLED() && IS_ATIO_MSIX_CAPABLE(ha)) { + qentry = &ha->msix_entries[ATIO_VECTOR]; + ret = request_irq(qentry->vector, + qla83xx_msix_entries[ATIO_VECTOR].handler, + 0, qla83xx_msix_entries[ATIO_VECTOR].name, rsp); + qentry->have_irq = 1; + qentry->rsp = rsp; + rsp->msix = qentry; + } + +msix_register_fail: + if (ret) { + ql_log(ql_log_fatal, vha, 0x00cb, + "MSI-X: unable to register handler -- %x/%d.\n", + qentry->vector, ret); + qla24xx_disable_msix(ha); + ha->mqenable = 0; + goto msix_out; + } + /* Enable MSI-X vector for response queue update for queue 0 */ if (IS_QLA83XX(ha)) { if (ha->msixbase && ha->mqiobase && From 126e964a444f125bd428757fb88c24c730f6fcf9 Mon Sep 17 00:00:00 2001 From: Mike Christie Date: Thu, 19 Dec 2013 01:16:21 -0600 Subject: [PATCH 107/259] [SCSI] be2iscsi: fix bad if expression https://bugzilla.kernel.org/show_bug.cgi?id=67091 Cc: Jayamohan Kallickal Signed-off-by: James Bottomley --- drivers/scsi/be2iscsi/be_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 1f375051483a..5642a9b250c2 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -325,7 +325,7 @@ static int beiscsi_eh_device_reset(struct scsi_cmnd *sc) if (!abrt_task->sc || abrt_task->state == ISCSI_TASK_FREE) continue; - if (abrt_task->sc->device->lun != abrt_task->sc->device->lun) + if (sc->device->lun != abrt_task->sc->device->lun) continue; /* Invalidate WRB Posted for this Task */ From ddfadd7736b677de2d4ca2cd5b4b655368c85a7a Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Thu, 6 Feb 2014 12:23:01 -0800 Subject: [PATCH 108/259] [SCSI] isci: fix reset timeout handling Remove an erroneous BUG_ON() in the case of a hard reset timeout. The reset timeout handler puts the port into the "awaiting link-up" state. The timeout causes the device to be disconnected and we need to be in the awaiting link-up state to re-connect the port. The BUG_ON() made the incorrect assumption that resets never timeout and we always complete the reset in the "resetting" state. Testing this patch also uncovered that libata continues to attempt to reset the port long after the driver has torn down the context. Once the driver has committed to abandoning the link it must indicate to libata that recovery ends by returning -ENODEV from ->lldd_I_T_nexus_reset(). Cc: Acked-by: Lukasz Dorau Reported-by: David Milburn Reported-by: Xun Ni Tested-by: Xun Ni Signed-off-by: Dan Williams Signed-off-by: James Bottomley --- drivers/scsi/isci/port_config.c | 7 ------- drivers/scsi/isci/task.c | 2 +- 2 files changed, 1 insertion(+), 8 deletions(-) diff --git a/drivers/scsi/isci/port_config.c b/drivers/scsi/isci/port_config.c index 85c77f6b802b..ac879745ef80 100644 --- a/drivers/scsi/isci/port_config.c +++ b/drivers/scsi/isci/port_config.c @@ -615,13 +615,6 @@ static void sci_apc_agent_link_up(struct isci_host *ihost, SCIC_SDS_APC_WAIT_LINK_UP_NOTIFICATION); } else { /* the phy is already the part of the port */ - u32 port_state = iport->sm.current_state_id; - - /* if the PORT'S state is resetting then the link up is from - * port hard reset in this case, we need to tell the port - * that link up is recieved - */ - BUG_ON(port_state != SCI_PORT_RESETTING); port_agent->phy_ready_mask |= 1 << phy_index; sci_port_link_up(iport, iphy); } diff --git a/drivers/scsi/isci/task.c b/drivers/scsi/isci/task.c index 0d30ca849e8f..5d6fda72d659 100644 --- a/drivers/scsi/isci/task.c +++ b/drivers/scsi/isci/task.c @@ -801,7 +801,7 @@ int isci_task_I_T_nexus_reset(struct domain_device *dev) /* XXX: need to cleanup any ireqs targeting this * domain_device */ - ret = TMF_RESP_FUNC_COMPLETE; + ret = -ENODEV; goto out; } From c59053a23d586675c25d789a7494adfdc02fba57 Mon Sep 17 00:00:00 2001 From: Lukasz Dorau Date: Thu, 6 Feb 2014 12:23:20 -0800 Subject: [PATCH 109/259] [SCSI] isci: correct erroneous for_each_isci_host macro In the first place, the loop 'for' in the macro 'for_each_isci_host' (drivers/scsi/isci/host.h:314) is incorrect, because it accesses the 3rd element of 2 element array. After the 2nd iteration it executes the instruction: ihost = to_pci_info(pdev)->hosts[2] (while the size of the 'hosts' array equals 2) and reads an out of range element. In the second place, this loop is incorrectly optimized by GCC v4.8 (see http://marc.info/?l=linux-kernel&m=138998871911336&w=2). As a result, on platforms with two SCU controllers, the loop is executed more times than it can be (for i=0,1 and 2). It causes kernel panic during entering the S3 state and the following oops after 'rmmod isci': BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] __list_add+0x1b/0xc0 Oops: 0000 [#1] SMP RIP: 0010:[] [] __list_add+0x1b/0xc0 Call Trace: [] __mutex_lock_slowpath+0x114/0x1b0 [] mutex_lock+0x1f/0x30 [] sas_disable_events+0x1b/0x50 [libsas] [] sas_unregister_ha+0x18/0x60 [libsas] [] isci_unregister+0x1e/0x40 [isci] [] isci_pci_remove+0x5d/0x100 [isci] [] pci_device_remove+0x3b/0xb0 [] __device_release_driver+0x7f/0xf0 [] driver_detach+0xa8/0xb0 [] bus_remove_driver+0x9b/0x120 [] driver_unregister+0x2c/0x50 [] pci_unregister_driver+0x23/0x80 [] isci_exit+0x10/0x1e [isci] [] SyS_delete_module+0x16b/0x2d0 [] ? do_notify_resume+0x61/0xa0 [] system_call_fastpath+0x16/0x1b The loop has been corrected. This patch fixes kernel panic during entering the S3 state and the above oops. Signed-off-by: Lukasz Dorau Reviewed-by: Maciej Patelczyk Tested-by: Lukasz Dorau Cc: Signed-off-by: Dan Williams Signed-off-by: James Bottomley --- drivers/scsi/isci/host.h | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/isci/host.h b/drivers/scsi/isci/host.h index 4911310a38f5..22a9bb1abae1 100644 --- a/drivers/scsi/isci/host.h +++ b/drivers/scsi/isci/host.h @@ -311,9 +311,8 @@ static inline struct Scsi_Host *to_shost(struct isci_host *ihost) } #define for_each_isci_host(id, ihost, pdev) \ - for (id = 0, ihost = to_pci_info(pdev)->hosts[id]; \ - id < ARRAY_SIZE(to_pci_info(pdev)->hosts) && ihost; \ - ihost = to_pci_info(pdev)->hosts[++id]) + for (id = 0; id < SCI_MAX_CONTROLLERS && \ + (ihost = to_pci_info(pdev)->hosts[id]); id++) static inline void wait_for_start(struct isci_host *ihost) { From b77ed25c9f8402e8b3e49e220edb4ef09ecfbb53 Mon Sep 17 00:00:00 2001 From: Giridhar Malavali Date: Wed, 26 Feb 2014 04:15:12 -0500 Subject: [PATCH 110/259] [SCSI] qla2xxx: Poll during initialization for ISP25xx and ISP83xx Cc: stable@vger.kernel.org Signed-off-by: Giridhar Malavali Signed-off-by: Saurav Kashyap Signed-off-by: James Bottomley --- drivers/scsi/qla2xxx/qla_def.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h index e1fe95ef23e1..266724b6b899 100644 --- a/drivers/scsi/qla2xxx/qla_def.h +++ b/drivers/scsi/qla2xxx/qla_def.h @@ -2996,8 +2996,7 @@ struct qla_hw_data { IS_QLA82XX(ha) || IS_QLA83XX(ha) || \ IS_QLA8044(ha)) #define IS_MSIX_NACK_CAPABLE(ha) (IS_QLA81XX(ha) || IS_QLA83XX(ha)) -#define IS_NOPOLLING_TYPE(ha) ((IS_QLA25XX(ha) || IS_QLA81XX(ha) || \ - IS_QLA83XX(ha)) && (ha)->flags.msix_enabled) +#define IS_NOPOLLING_TYPE(ha) (IS_QLA81XX(ha) && (ha)->flags.msix_enabled) #define IS_FAC_REQUIRED(ha) (IS_QLA81XX(ha) || IS_QLA83XX(ha)) #define IS_NOCACHE_VPD_TYPE(ha) (IS_QLA81XX(ha) || IS_QLA83XX(ha)) #define IS_ALOGIO_CAPABLE(ha) (IS_QLA23XX(ha) || IS_FWI2_CAPABLE(ha)) From 58d4d3c976b33784a1443c446a3d7203bf2153f0 Mon Sep 17 00:00:00 2001 From: Lars-Peter Clausen Date: Tue, 4 Mar 2014 09:41:53 +0100 Subject: [PATCH 111/259] ASoC: si476x: Fix IO setup The si476x is a MFD device and the CODEC driver is using the regmap struct of the parent device, hence automatic IO setup will not work and we need to manually call snd_soc_codec_set_cache_io(). The issue was introduced commit d6173df35f ("ASoC: si476x: Remove custom register I/O implementation") Fixes: d6173df35f ("ASoC: si476x: Remove custom register I/O implementation") Signed-off-by: Lars-Peter Clausen Signed-off-by: Mark Brown Cc: stable@vger.kernel.org --- sound/soc/codecs/si476x.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/si476x.c b/sound/soc/codecs/si476x.c index 52e7cb08434b..fa2b8e07f420 100644 --- a/sound/soc/codecs/si476x.c +++ b/sound/soc/codecs/si476x.c @@ -210,7 +210,7 @@ static int si476x_codec_hw_params(struct snd_pcm_substream *substream, static int si476x_codec_probe(struct snd_soc_codec *codec) { codec->control_data = dev_get_regmap(codec->dev->parent, NULL); - return 0; + return snd_soc_codec_set_cache_io(codec, 0, 0, SND_SOC_REGMAP); } static struct snd_soc_dai_ops si476x_dai_ops = { From 8eeb5c15131d7b5061c10423eda3ae4c68db4eaf Mon Sep 17 00:00:00 2001 From: Lars-Peter Clausen Date: Tue, 4 Mar 2014 09:39:24 +0100 Subject: [PATCH 112/259] ASoC: 88pm860: Fix IO setup The 88pm860 is a MFD device and the CODEC driver is using the regmap struct of the parent device, hence automatic IO setup will not work and we need to manually call snd_soc_codec_set_cache_io(). The issue was introduced in commit f9ded3b2e7 ("ASoC: 88pm860x: Use regmap for I/O"). Fixes: f9ded3b2e7 ("ASoC: 88pm860x: Use regmap for I/O"). Signed-off-by: Lars-Peter Clausen Signed-off-by: Mark Brown Cc: stable@vger.kernel.org --- sound/soc/codecs/88pm860x-codec.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/sound/soc/codecs/88pm860x-codec.c b/sound/soc/codecs/88pm860x-codec.c index 75d0ad5d2dcb..647a72cda005 100644 --- a/sound/soc/codecs/88pm860x-codec.c +++ b/sound/soc/codecs/88pm860x-codec.c @@ -1328,6 +1328,9 @@ static int pm860x_probe(struct snd_soc_codec *codec) pm860x->codec = codec; codec->control_data = pm860x->regmap; + ret = snd_soc_codec_set_cache_io(codec, 0, 0, SND_SOC_REGMAP); + if (ret) + return ret; for (i = 0; i < 4; i++) { ret = request_threaded_irq(pm860x->irq[i], NULL, From 02c5bb4a352a4cca56e9b5d3a2a57d61062eb2e1 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 6 Feb 2014 01:00:41 +0000 Subject: [PATCH 113/259] perf trace: Decode architecture-specific signal numbers SIGSTKFLT is not defined on alpha, mips or sparc. SIGEMT and SIGSWI are defined on some architectures and should be decoded here if so. Signed-off-by: Ben Hutchings Fixes: 8bad5b0abfdb ('perf trace: Beautify signal number arg in several syscalls') Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: stable@vger.kernel.org Link: http://lkml.kernel.org/r/1391648441.3003.101.camel@deadeye.wl.decadent.org.uk Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-trace.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 6aa6fb6f7bd9..f954c26de231 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -825,7 +825,6 @@ static size_t syscall_arg__scnprintf_signum(char *bf, size_t size, struct syscal P_SIGNUM(PIPE); P_SIGNUM(ALRM); P_SIGNUM(TERM); - P_SIGNUM(STKFLT); P_SIGNUM(CHLD); P_SIGNUM(CONT); P_SIGNUM(STOP); @@ -841,6 +840,15 @@ static size_t syscall_arg__scnprintf_signum(char *bf, size_t size, struct syscal P_SIGNUM(IO); P_SIGNUM(PWR); P_SIGNUM(SYS); +#ifdef SIGEMT + P_SIGNUM(EMT); +#endif +#ifdef SIGSTKFLT + P_SIGNUM(STKFLT); +#endif +#ifdef SIGSWI + P_SIGNUM(SWI); +#endif default: break; } From 155b3a13a65d4217316dbe094843f2a7df0711fa Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Sun, 2 Mar 2014 14:32:07 +0100 Subject: [PATCH 114/259] perf symbols: Fix crash in elf_section_by_name Fixing crash in elf_section_by_name function caused by missing section name in elf binary. Reported-by: Albert Strasheim Signed-off-by: Jiri Olsa Cc: Albert Strasheim Cc: Corey Ashford Cc: David Ahern Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/1393767127-599-1-git-send-email-jolsa@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/symbol-elf.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c index 3e9f336740fa..516d19fb999b 100644 --- a/tools/perf/util/symbol-elf.c +++ b/tools/perf/util/symbol-elf.c @@ -151,15 +151,15 @@ Elf_Scn *elf_section_by_name(Elf *elf, GElf_Ehdr *ep, gelf_getshdr(sec, shp); str = elf_strptr(elf, ep->e_shstrndx, shp->sh_name); - if (!strcmp(name, str)) { + if (str && !strcmp(name, str)) { if (idx) *idx = cnt; - break; + return sec; } ++cnt; } - return sec; + return NULL; } #define elf_section__for_each_rel(reldata, pos, pos_mem, idx, nr_entries) \ From fdf57dd052d5cbd415533ae98f4d423286a85220 Mon Sep 17 00:00:00 2001 From: Don Zickus Date: Tue, 25 Feb 2014 22:43:45 -0500 Subject: [PATCH 115/259] perf machine: Use map as success in ip__resolve_ams When trying to map a bunch of instruction addresses to their respective threads, I kept getting a lot of bogus entries [I forget the exact reason as I patched my code months ago]. Looking through ip__resolve_ams, I noticed the check for if (al.sym) and realized, most times I have an al.map definition but sometimes an al.sym is undefined. In the cases where al.sym is undefined, the loop keeps going even though a valid al.map exists. Modify this check to use the more reliable al.map. This fixed my bogus entries. Signed-off-by: Don Zickus Cc: Jiri Olsa Cc: Stephane Eranian Link: http://lkml.kernel.org/r/1393386227-149412-2-git-send-email-dzickus@redhat.com Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/util/machine.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c index c872991e0f65..620a1983b76b 100644 --- a/tools/perf/util/machine.c +++ b/tools/perf/util/machine.c @@ -1213,7 +1213,7 @@ static void ip__resolve_ams(struct machine *machine, struct thread *thread, */ thread__find_addr_location(thread, machine, m, MAP__FUNCTION, ip, &al); - if (al.sym) + if (al.map) goto found; } found: From 83493d7e782d2630f1a55def14a79f0e7c4faac3 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 10 Mar 2014 11:13:43 -0400 Subject: [PATCH 116/259] libata: use wider match for blacklisting Crucial M500 We're now blacklisting "Crucial_CT???M500SSD1" and "Crucial_CT???M500SSD3". Also, "Micron_M500*" is blacklisted which is about the same devices as the crucial branded ones. Let's merge the two Crucial M500 entries and widen the match to "Crucial_CT???M500SSD*" so that we don't have to fiddle with new entries for similar devices. Signed-off-by: Tejun Heo Suggested-by: Linus Torvalds Cc: stable@vger.kernel.org --- drivers/ata/libata-core.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c index 65d3f1b5966c..8cb2522d592a 100644 --- a/drivers/ata/libata-core.c +++ b/drivers/ata/libata-core.c @@ -4225,8 +4225,7 @@ static const struct ata_blacklist_entry ata_device_blacklist [] = { /* devices that don't properly handle queued TRIM commands */ { "Micron_M500*", NULL, ATA_HORKAGE_NO_NCQ_TRIM, }, - { "Crucial_CT???M500SSD1", NULL, ATA_HORKAGE_NO_NCQ_TRIM, }, - { "Crucial_CT???M500SSD3", NULL, ATA_HORKAGE_NO_NCQ_TRIM, }, + { "Crucial_CT???M500SSD*", NULL, ATA_HORKAGE_NO_NCQ_TRIM, }, /* * Some WD SATA-I drives spin up and down erratically when the link From 1b56e98990bcdbb20b9fab163654b9315bf158e8 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 10 Feb 2014 15:18:55 -0500 Subject: [PATCH 117/259] ocfs2 syncs the wrong range... Cc: stable@vger.kernel.org Signed-off-by: Al Viro --- fs/ocfs2/file.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 8450262bcf2a..51632c40e896 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2393,8 +2393,8 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) || ((file->f_flags & O_DIRECT) && !direct_io)) { - ret = filemap_fdatawrite_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawrite_range(file->f_mapping, *ppos, + *ppos + count - 1); if (ret < 0) written = ret; @@ -2407,8 +2407,8 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, } if (!ret) - ret = filemap_fdatawait_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawait_range(file->f_mapping, *ppos, + *ppos + count - 1); } /* From 9c225f2655e36a470c4f58dbbc99244c5fc7f2d4 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 3 Mar 2014 09:36:58 -0800 Subject: [PATCH 118/259] vfs: atomic f_pos accesses as per POSIX Our write() system call has always been atomic in the sense that you get the expected thread-safe contiguous write, but we haven't actually guaranteed that concurrent writes are serialized wrt f_pos accesses, so threads (or processes) that share a file descriptor and use "write()" concurrently would quite likely overwrite each others data. This violates POSIX.1-2008/SUSv4 Section XSI 2.9.7 that says: "2.9.7 Thread Interactions with Regular File Operations All of the following functions shall be atomic with respect to each other in the effects specified in POSIX.1-2008 when they operate on regular files or symbolic links: [...]" and one of the effects is the file position update. This unprotected file position behavior is not new behavior, and nobody has ever cared. Until now. Yongzhi Pan reported unexpected behavior to Michael Kerrisk that was due to this. This resolves the issue with a f_pos-specific lock that is taken by read/write/lseek on file descriptors that may be shared across threads or processes. Reported-by: Yongzhi Pan Reported-by: Michael Kerrisk Cc: Al Viro Signed-off-by: Linus Torvalds Signed-off-by: Al Viro --- fs/file_table.c | 1 + fs/namei.c | 2 +- fs/open.c | 4 ++++ fs/read_write.c | 54 ++++++++++++++++++++++++++++++++------------ include/linux/file.h | 6 +++-- include/linux/fs.h | 6 ++++- 6 files changed, 55 insertions(+), 18 deletions(-) diff --git a/fs/file_table.c b/fs/file_table.c index 5fff9030be34..5b24008ea4f6 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -135,6 +135,7 @@ struct file *get_empty_filp(void) atomic_long_set(&f->f_count, 1); rwlock_init(&f->f_owner.lock); spin_lock_init(&f->f_lock); + mutex_init(&f->f_pos_lock); eventpoll_init_file(f); /* f->f_version: 0 */ return f; diff --git a/fs/namei.c b/fs/namei.c index 385f7817bfcc..2f730ef9b4b3 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1884,7 +1884,7 @@ static int path_init(int dfd, const char *name, unsigned int flags, nd->path = f.file->f_path; if (flags & LOOKUP_RCU) { - if (f.need_put) + if (f.flags & FDPUT_FPUT) *fp = f.file; nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq); rcu_read_lock(); diff --git a/fs/open.c b/fs/open.c index 4b3e1edf2fe4..b9ed8b25c108 100644 --- a/fs/open.c +++ b/fs/open.c @@ -705,6 +705,10 @@ static int do_dentry_open(struct file *f, return 0; } + /* POSIX.1-2008/SUSv4 Section XSI 2.9.7 */ + if (S_ISREG(inode->i_mode)) + f->f_mode |= FMODE_ATOMIC_POS; + f->f_op = fops_get(inode->i_fop); if (unlikely(WARN_ON(!f->f_op))) { error = -ENODEV; diff --git a/fs/read_write.c b/fs/read_write.c index edc5746a902a..932bb3414a96 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -264,10 +264,36 @@ loff_t vfs_llseek(struct file *file, loff_t offset, int whence) } EXPORT_SYMBOL(vfs_llseek); +/* + * We only lock f_pos if we have threads or if the file might be + * shared with another process. In both cases we'll have an elevated + * file count (done either by fdget() or by fork()). + */ +static inline struct fd fdget_pos(int fd) +{ + struct fd f = fdget(fd); + struct file *file = f.file; + + if (file && (file->f_mode & FMODE_ATOMIC_POS)) { + if (file_count(file) > 1) { + f.flags |= FDPUT_POS_UNLOCK; + mutex_lock(&file->f_pos_lock); + } + } + return f; +} + +static inline void fdput_pos(struct fd f) +{ + if (f.flags & FDPUT_POS_UNLOCK) + mutex_unlock(&f.file->f_pos_lock); + fdput(f); +} + SYSCALL_DEFINE3(lseek, unsigned int, fd, off_t, offset, unsigned int, whence) { off_t retval; - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); if (!f.file) return -EBADF; @@ -278,7 +304,7 @@ SYSCALL_DEFINE3(lseek, unsigned int, fd, off_t, offset, unsigned int, whence) if (res != (loff_t)retval) retval = -EOVERFLOW; /* LFS: should only happen on 32 bit platforms */ } - fdput(f); + fdput_pos(f); return retval; } @@ -498,7 +524,7 @@ static inline void file_pos_write(struct file *file, loff_t pos) SYSCALL_DEFINE3(read, unsigned int, fd, char __user *, buf, size_t, count) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret = -EBADF; if (f.file) { @@ -506,7 +532,7 @@ SYSCALL_DEFINE3(read, unsigned int, fd, char __user *, buf, size_t, count) ret = vfs_read(f.file, buf, count, &pos); if (ret >= 0) file_pos_write(f.file, pos); - fdput(f); + fdput_pos(f); } return ret; } @@ -514,7 +540,7 @@ SYSCALL_DEFINE3(read, unsigned int, fd, char __user *, buf, size_t, count) SYSCALL_DEFINE3(write, unsigned int, fd, const char __user *, buf, size_t, count) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret = -EBADF; if (f.file) { @@ -522,7 +548,7 @@ SYSCALL_DEFINE3(write, unsigned int, fd, const char __user *, buf, ret = vfs_write(f.file, buf, count, &pos); if (ret >= 0) file_pos_write(f.file, pos); - fdput(f); + fdput_pos(f); } return ret; @@ -797,7 +823,7 @@ EXPORT_SYMBOL(vfs_writev); SYSCALL_DEFINE3(readv, unsigned long, fd, const struct iovec __user *, vec, unsigned long, vlen) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret = -EBADF; if (f.file) { @@ -805,7 +831,7 @@ SYSCALL_DEFINE3(readv, unsigned long, fd, const struct iovec __user *, vec, ret = vfs_readv(f.file, vec, vlen, &pos); if (ret >= 0) file_pos_write(f.file, pos); - fdput(f); + fdput_pos(f); } if (ret > 0) @@ -817,7 +843,7 @@ SYSCALL_DEFINE3(readv, unsigned long, fd, const struct iovec __user *, vec, SYSCALL_DEFINE3(writev, unsigned long, fd, const struct iovec __user *, vec, unsigned long, vlen) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret = -EBADF; if (f.file) { @@ -825,7 +851,7 @@ SYSCALL_DEFINE3(writev, unsigned long, fd, const struct iovec __user *, vec, ret = vfs_writev(f.file, vec, vlen, &pos); if (ret >= 0) file_pos_write(f.file, pos); - fdput(f); + fdput_pos(f); } if (ret > 0) @@ -968,7 +994,7 @@ COMPAT_SYSCALL_DEFINE3(readv, compat_ulong_t, fd, const struct compat_iovec __user *,vec, compat_ulong_t, vlen) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret; loff_t pos; @@ -978,7 +1004,7 @@ COMPAT_SYSCALL_DEFINE3(readv, compat_ulong_t, fd, ret = compat_readv(f.file, vec, vlen, &pos); if (ret >= 0) f.file->f_pos = pos; - fdput(f); + fdput_pos(f); return ret; } @@ -1035,7 +1061,7 @@ COMPAT_SYSCALL_DEFINE3(writev, compat_ulong_t, fd, const struct compat_iovec __user *, vec, compat_ulong_t, vlen) { - struct fd f = fdget(fd); + struct fd f = fdget_pos(fd); ssize_t ret; loff_t pos; @@ -1045,7 +1071,7 @@ COMPAT_SYSCALL_DEFINE3(writev, compat_ulong_t, fd, ret = compat_writev(f.file, vec, vlen, &pos); if (ret >= 0) f.file->f_pos = pos; - fdput(f); + fdput_pos(f); return ret; } diff --git a/include/linux/file.h b/include/linux/file.h index cbacf4faf447..f2517fa2d610 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -28,12 +28,14 @@ static inline void fput_light(struct file *file, int fput_needed) struct fd { struct file *file; - int need_put; + unsigned int flags; }; +#define FDPUT_FPUT 1 +#define FDPUT_POS_UNLOCK 2 static inline void fdput(struct fd fd) { - if (fd.need_put) + if (fd.flags & FDPUT_FPUT) fput(fd.file); } diff --git a/include/linux/fs.h b/include/linux/fs.h index 60829565e552..ebfde04bca06 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -123,6 +123,9 @@ typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset, /* File is opened with O_PATH; almost nothing can be done with it */ #define FMODE_PATH ((__force fmode_t)0x4000) +/* File needs atomic accesses to f_pos */ +#define FMODE_ATOMIC_POS ((__force fmode_t)0x8000) + /* File was opened by fanotify and shouldn't generate fanotify events */ #define FMODE_NONOTIFY ((__force fmode_t)0x1000000) @@ -780,13 +783,14 @@ struct file { const struct file_operations *f_op; /* - * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR. + * Protects f_ep_links, f_flags. * Must not be taken from IRQ context. */ spinlock_t f_lock; atomic_long_t f_count; unsigned int f_flags; fmode_t f_mode; + struct mutex f_pos_lock; loff_t f_pos; struct fown_struct f_owner; const struct cred *f_cred; From 00e188ef6a7e7bf2883bcffc30d89e98a0bb76b3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 3 Mar 2014 23:48:18 -0500 Subject: [PATCH 119/259] sockfd_lookup_light(): switch to fdget^W^Waway from fget_light Signed-off-by: Al Viro --- net/socket.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/net/socket.c b/net/socket.c index 879933aaed4c..fd8d86e06f95 100644 --- a/net/socket.c +++ b/net/socket.c @@ -450,16 +450,17 @@ EXPORT_SYMBOL(sockfd_lookup); static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed) { - struct file *file; + struct fd f = fdget(fd); struct socket *sock; *err = -EBADF; - file = fget_light(fd, fput_needed); - if (file) { - sock = sock_from_file(file, err); - if (sock) + if (f.file) { + sock = sock_from_file(f.file, err); + if (likely(sock)) { + *fput_needed = f.flags; return sock; - fput_light(file, *fput_needed); + } + fdput(f); } return NULL; } From bd2a31d522344b3ac2fb680bd2366e77a9bd8209 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 4 Mar 2014 14:54:22 -0500 Subject: [PATCH 120/259] get rid of fget_light() instead of returning the flags by reference, we can just have the low-level primitive return those in lower bits of unsigned long, with struct file * derived from the rest. Signed-off-by: Al Viro --- fs/file.c | 56 ++++++++++++++++++++++++++++++++++---------- fs/read_write.c | 16 +------------ include/linux/file.h | 21 +++++++++-------- include/linux/fs.h | 2 +- 4 files changed, 56 insertions(+), 39 deletions(-) diff --git a/fs/file.c b/fs/file.c index db25c2bdfe46..60a45e9f5323 100644 --- a/fs/file.c +++ b/fs/file.c @@ -683,35 +683,65 @@ EXPORT_SYMBOL(fget_raw); * The fput_needed flag returned by fget_light should be passed to the * corresponding fput_light. */ -struct file *__fget_light(unsigned int fd, fmode_t mask, int *fput_needed) +static unsigned long __fget_light(unsigned int fd, fmode_t mask) { struct files_struct *files = current->files; struct file *file; - *fput_needed = 0; if (atomic_read(&files->count) == 1) { file = __fcheck_files(files, fd); - if (file && (file->f_mode & mask)) - file = NULL; + if (!file || unlikely(file->f_mode & mask)) + return 0; + return (unsigned long)file; } else { file = __fget(fd, mask); - if (file) - *fput_needed = 1; + if (!file) + return 0; + return FDPUT_FPUT | (unsigned long)file; } - - return file; } -struct file *fget_light(unsigned int fd, int *fput_needed) +unsigned long __fdget(unsigned int fd) { - return __fget_light(fd, FMODE_PATH, fput_needed); + return __fget_light(fd, FMODE_PATH); } -EXPORT_SYMBOL(fget_light); +EXPORT_SYMBOL(__fdget); -struct file *fget_raw_light(unsigned int fd, int *fput_needed) +unsigned long __fdget_raw(unsigned int fd) { - return __fget_light(fd, 0, fput_needed); + return __fget_light(fd, 0); } +unsigned long __fdget_pos(unsigned int fd) +{ + struct files_struct *files = current->files; + struct file *file; + unsigned long v; + + if (atomic_read(&files->count) == 1) { + file = __fcheck_files(files, fd); + v = 0; + } else { + file = __fget(fd, 0); + v = FDPUT_FPUT; + } + if (!file) + return 0; + + if (file->f_mode & FMODE_ATOMIC_POS) { + if (file_count(file) > 1) { + v |= FDPUT_POS_UNLOCK; + mutex_lock(&file->f_pos_lock); + } + } + return v | (unsigned long)file; +} + +/* + * We only lock f_pos if we have threads or if the file might be + * shared with another process. In both cases we'll have an elevated + * file count (done either by fdget() or by fork()). + */ + void set_close_on_exec(unsigned int fd, int flag) { struct files_struct *files = current->files; diff --git a/fs/read_write.c b/fs/read_write.c index 932bb3414a96..54e19b9392dc 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -264,23 +264,9 @@ loff_t vfs_llseek(struct file *file, loff_t offset, int whence) } EXPORT_SYMBOL(vfs_llseek); -/* - * We only lock f_pos if we have threads or if the file might be - * shared with another process. In both cases we'll have an elevated - * file count (done either by fdget() or by fork()). - */ static inline struct fd fdget_pos(int fd) { - struct fd f = fdget(fd); - struct file *file = f.file; - - if (file && (file->f_mode & FMODE_ATOMIC_POS)) { - if (file_count(file) > 1) { - f.flags |= FDPUT_POS_UNLOCK; - mutex_lock(&file->f_pos_lock); - } - } - return f; + return __to_fd(__fdget_pos(fd)); } static inline void fdput_pos(struct fd f) diff --git a/include/linux/file.h b/include/linux/file.h index f2517fa2d610..4d69123377a2 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -40,23 +40,24 @@ static inline void fdput(struct fd fd) } extern struct file *fget(unsigned int fd); -extern struct file *fget_light(unsigned int fd, int *fput_needed); +extern struct file *fget_raw(unsigned int fd); +extern unsigned long __fdget(unsigned int fd); +extern unsigned long __fdget_raw(unsigned int fd); +extern unsigned long __fdget_pos(unsigned int fd); + +static inline struct fd __to_fd(unsigned long v) +{ + return (struct fd){(struct file *)(v & ~3),v & 3}; +} static inline struct fd fdget(unsigned int fd) { - int b; - struct file *f = fget_light(fd, &b); - return (struct fd){f,b}; + return __to_fd(__fdget(fd)); } -extern struct file *fget_raw(unsigned int fd); -extern struct file *fget_raw_light(unsigned int fd, int *fput_needed); - static inline struct fd fdget_raw(unsigned int fd) { - int b; - struct file *f = fget_raw_light(fd, &b); - return (struct fd){f,b}; + return __to_fd(__fdget_raw(fd)); } extern int f_dupfd(unsigned int from, struct file *file, unsigned flags); diff --git a/include/linux/fs.h b/include/linux/fs.h index ebfde04bca06..23b2a35d712e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -812,7 +812,7 @@ struct file { #ifdef CONFIG_DEBUG_WRITECOUNT unsigned long f_mnt_write_state; #endif -}; +} __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */ struct file_handle { __u32 handle_bytes; From 37314363cd65d19c71bea5f222e5108c93dc3c78 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sat, 8 Mar 2014 08:01:19 -0800 Subject: [PATCH 121/259] pkt_sched: move the sanity test in qdisc_list_add() The WARN_ON(root == &noop_qdisc)) added in qdisc_list_add() can trigger in normal conditions when devices are not up. It should be done only right before the list_add_tail() call. Fixes: e57a784d8cae4 ("pkt_sched: set root qdisc before change() in attach_default_qdiscs()") Reported-by: Valdis Kletnieks Tested-by: Mirco Tischler Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/sched/sch_api.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 1313145e3b86..a07d55e75698 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -273,11 +273,12 @@ static struct Qdisc *qdisc_match_from_root(struct Qdisc *root, u32 handle) void qdisc_list_add(struct Qdisc *q) { - struct Qdisc *root = qdisc_dev(q)->qdisc; + if ((q->parent != TC_H_ROOT) && !(q->flags & TCQ_F_INGRESS)) { + struct Qdisc *root = qdisc_dev(q)->qdisc; - WARN_ON_ONCE(root == &noop_qdisc); - if ((q->parent != TC_H_ROOT) && !(q->flags & TCQ_F_INGRESS)) + WARN_ON_ONCE(root == &noop_qdisc); list_add_tail(&q->list, &root->list); + } } EXPORT_SYMBOL(qdisc_list_add); From bc48bc806442114ea44f61d6b18e02c2bd6236fd Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Sun, 9 Mar 2014 04:03:22 +0000 Subject: [PATCH 122/259] bna: Replace large udelay() with mdelay() udelay() does not work on some architectures for values above 2000, in particular on ARM: ERROR: "__bad_udelay" [drivers/net/ethernet/brocade/bna/bna.ko] undefined! Reported-by: Vagrant Cascadian Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- drivers/net/ethernet/brocade/bna/bfa_ioc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/brocade/bna/bfa_ioc.c b/drivers/net/ethernet/brocade/bna/bfa_ioc.c index 1803c3959044..354ae9792bad 100644 --- a/drivers/net/ethernet/brocade/bna/bfa_ioc.c +++ b/drivers/net/ethernet/brocade/bna/bfa_ioc.c @@ -1704,7 +1704,7 @@ bfa_flash_sem_get(void __iomem *bar) while (!bfa_raw_sem_get(bar)) { if (--n <= 0) return BFA_STATUS_BADFLASH; - udelay(10000); + mdelay(10); } return BFA_STATUS_OK; } From 2818fa0fa068bcbc87d6bd9064e3c1f72d6fcc2a Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 6 Mar 2014 22:57:52 -0800 Subject: [PATCH 123/259] pkt_sched: fq: do not hold qdisc lock while allocating memory Resizing fq hash table allocates memory while holding qdisc spinlock, with BH disabled. This is definitely not good, as allocation might sleep. We can drop the lock and get it when needed, we hold RTNL so no other changes can happen at the same time. Signed-off-by: Eric Dumazet Fixes: afe4fd062416 ("pkt_sched: fq: Fair Queue packet scheduler") Signed-off-by: David S. Miller --- net/sched/sch_fq.c | 21 +++++++++++++++------ 1 file changed, 15 insertions(+), 6 deletions(-) diff --git a/net/sched/sch_fq.c b/net/sched/sch_fq.c index 08ef7a42c0e4..21e251766eb1 100644 --- a/net/sched/sch_fq.c +++ b/net/sched/sch_fq.c @@ -601,6 +601,7 @@ static int fq_resize(struct Qdisc *sch, u32 log) { struct fq_sched_data *q = qdisc_priv(sch); struct rb_root *array; + void *old_fq_root; u32 idx; if (q->fq_root && log == q->fq_trees_log) @@ -615,13 +616,19 @@ static int fq_resize(struct Qdisc *sch, u32 log) for (idx = 0; idx < (1U << log); idx++) array[idx] = RB_ROOT; - if (q->fq_root) { - fq_rehash(q, q->fq_root, q->fq_trees_log, array, log); - fq_free(q->fq_root); - } + sch_tree_lock(sch); + + old_fq_root = q->fq_root; + if (old_fq_root) + fq_rehash(q, old_fq_root, q->fq_trees_log, array, log); + q->fq_root = array; q->fq_trees_log = log; + sch_tree_unlock(sch); + + fq_free(old_fq_root); + return 0; } @@ -697,9 +704,11 @@ static int fq_change(struct Qdisc *sch, struct nlattr *opt) q->flow_refill_delay = usecs_to_jiffies(usecs_delay); } - if (!err) + if (!err) { + sch_tree_unlock(sch); err = fq_resize(sch, fq_log); - + sch_tree_lock(sch); + } while (sch->q.qlen > sch->limit) { struct sk_buff *skb = fq_dequeue(sch); From e97ca8e5b864f88b028c1759ba8536fa827d6d96 Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Mon, 10 Mar 2014 15:49:43 -0700 Subject: [PATCH 124/259] mm: fix GFP_THISNODE callers and clarify GFP_THISNODE is for callers that implement their own clever fallback to remote nodes. It restricts the allocation to the specified node and does not invoke reclaim, assuming that the caller will take care of it when the fallback fails, e.g. through a subsequent allocation request without GFP_THISNODE set. However, many current GFP_THISNODE users only want the node exclusive aspect of the flag, without actually implementing their own fallback or triggering reclaim if necessary. This results in things like page migration failing prematurely even when there is easily reclaimable memory available, unless kswapd happens to be running already or a concurrent allocation attempt triggers the necessary reclaim. Convert all callsites that don't implement their own fallback strategy to __GFP_THISNODE. This restricts the allocation a single node too, but at the same time allows the allocator to enter the slowpath, wake kswapd, and invoke direct reclaim if necessary, to make the allocation happen when memory is full. Signed-off-by: Johannes Weiner Acked-by: Rik van Riel Cc: Jan Stancek Cc: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ia64/kernel/uncached.c | 2 +- arch/powerpc/platforms/cell/ras.c | 3 ++- drivers/misc/sgi-xp/xpc_uv.c | 2 +- include/linux/gfp.h | 4 ++++ include/linux/mmzone.h | 4 ++-- include/linux/slab.h | 2 +- kernel/profile.c | 4 ++-- mm/migrate.c | 11 ++++++----- 8 files changed, 19 insertions(+), 13 deletions(-) diff --git a/arch/ia64/kernel/uncached.c b/arch/ia64/kernel/uncached.c index a96bcf83a735..20e8a9b21d75 100644 --- a/arch/ia64/kernel/uncached.c +++ b/arch/ia64/kernel/uncached.c @@ -98,7 +98,7 @@ static int uncached_add_chunk(struct uncached_pool *uc_pool, int nid) /* attempt to allocate a granule's worth of cached memory pages */ page = alloc_pages_exact_node(nid, - GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, + GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE, IA64_GRANULE_SHIFT-PAGE_SHIFT); if (!page) { mutex_unlock(&uc_pool->add_chunk_mutex); diff --git a/arch/powerpc/platforms/cell/ras.c b/arch/powerpc/platforms/cell/ras.c index 5ec1e47a0d77..e865d748179b 100644 --- a/arch/powerpc/platforms/cell/ras.c +++ b/arch/powerpc/platforms/cell/ras.c @@ -123,7 +123,8 @@ static int __init cbe_ptcal_enable_on_node(int nid, int order) area->nid = nid; area->order = order; - area->pages = alloc_pages_exact_node(area->nid, GFP_KERNEL|GFP_THISNODE, + area->pages = alloc_pages_exact_node(area->nid, + GFP_KERNEL|__GFP_THISNODE, area->order); if (!area->pages) { diff --git a/drivers/misc/sgi-xp/xpc_uv.c b/drivers/misc/sgi-xp/xpc_uv.c index b9e2000969f0..95c894482fdd 100644 --- a/drivers/misc/sgi-xp/xpc_uv.c +++ b/drivers/misc/sgi-xp/xpc_uv.c @@ -240,7 +240,7 @@ xpc_create_gru_mq_uv(unsigned int mq_size, int cpu, char *irq_name, nid = cpu_to_node(cpu); page = alloc_pages_exact_node(nid, - GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, + GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE, pg_order); if (page == NULL) { dev_err(xpc_part, "xpc_create_gru_mq_uv() failed to alloc %d " diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 0437439bc047..39b81dc7d01a 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -123,6 +123,10 @@ struct vm_area_struct; __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN | \ __GFP_NO_KSWAPD) +/* + * GFP_THISNODE does not perform any reclaim, you most likely want to + * use __GFP_THISNODE to allocate from a given node without fallback! + */ #ifdef CONFIG_NUMA #define GFP_THISNODE (__GFP_THISNODE | __GFP_NOWARN | __GFP_NORETRY) #else diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 5f2052c83154..9b61b9bf81ac 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -590,10 +590,10 @@ static inline bool zone_is_empty(struct zone *zone) /* * The NUMA zonelists are doubled because we need zonelists that restrict the - * allocations to a single node for GFP_THISNODE. + * allocations to a single node for __GFP_THISNODE. * * [0] : Zonelist with fallback - * [1] : No fallback (GFP_THISNODE) + * [1] : No fallback (__GFP_THISNODE) */ #define MAX_ZONELISTS 2 diff --git a/include/linux/slab.h b/include/linux/slab.h index 9260abdd67df..b5b2df60299e 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -410,7 +410,7 @@ static __always_inline void *kmalloc_large(size_t size, gfp_t flags) * * %GFP_NOWAIT - Allocation will not sleep. * - * %GFP_THISNODE - Allocate node-local memory only. + * %__GFP_THISNODE - Allocate node-local memory only. * * %GFP_DMA - Allocation suitable for DMA. * Should only be used for kmalloc() caches. Otherwise, use a diff --git a/kernel/profile.c b/kernel/profile.c index 6631e1ef55ab..ebdd9c1a86b4 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -549,14 +549,14 @@ static int create_hash_tables(void) struct page *page; page = alloc_pages_exact_node(node, - GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, + GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE, 0); if (!page) goto out_cleanup; per_cpu(cpu_profile_hits, cpu)[1] = (struct profile_hit *)page_address(page); page = alloc_pages_exact_node(node, - GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, + GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE, 0); if (!page) goto out_cleanup; diff --git a/mm/migrate.c b/mm/migrate.c index 482a33d89134..b494fdb9a636 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1158,7 +1158,7 @@ static struct page *new_page_node(struct page *p, unsigned long private, pm->node); else return alloc_pages_exact_node(pm->node, - GFP_HIGHUSER_MOVABLE | GFP_THISNODE, 0); + GFP_HIGHUSER_MOVABLE | __GFP_THISNODE, 0); } /* @@ -1544,9 +1544,9 @@ static struct page *alloc_misplaced_dst_page(struct page *page, struct page *newpage; newpage = alloc_pages_exact_node(nid, - (GFP_HIGHUSER_MOVABLE | GFP_THISNODE | - __GFP_NOMEMALLOC | __GFP_NORETRY | - __GFP_NOWARN) & + (GFP_HIGHUSER_MOVABLE | + __GFP_THISNODE | __GFP_NOMEMALLOC | + __GFP_NORETRY | __GFP_NOWARN) & ~GFP_IOFS, 0); return newpage; @@ -1747,7 +1747,8 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, goto out_dropref; new_page = alloc_pages_node(node, - (GFP_TRANSHUGE | GFP_THISNODE) & ~__GFP_WAIT, HPAGE_PMD_ORDER); + (GFP_TRANSHUGE | __GFP_THISNODE) & ~__GFP_WAIT, + HPAGE_PMD_ORDER); if (!new_page) goto out_fail; From 2af120bc040c5ebcda156df6be6a66610ab6957f Mon Sep 17 00:00:00 2001 From: Laura Abbott Date: Mon, 10 Mar 2014 15:49:44 -0700 Subject: [PATCH 125/259] mm/compaction: break out of loop on !PageBuddy in isolate_freepages_block We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range: BUG: Bad page state in process Binder_A pfn:63202 page:d21130b0 count:0 mapcount:1 mapping: (null) index:0x7dfbf page flags: 0x40080068(uptodate|lru|active|swapbacked) Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator. There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range: if (strict && nr_strict_required > total_isolated) total_isolated = 0; After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does. Signed-off-by: Laura Abbott Acked-by: Minchan Kim Cc: Mel Gorman Acked-by: Vlastimil Babka Cc: Joonsoo Kim Acked-by: Bartlomiej Zolnierkiewicz Acked-by: Michal Nazarewicz Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/compaction.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) diff --git a/mm/compaction.c b/mm/compaction.c index b48c5259ea33..918577595ea8 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -251,7 +251,6 @@ static unsigned long isolate_freepages_block(struct compact_control *cc, { int nr_scanned = 0, total_isolated = 0; struct page *cursor, *valid_page = NULL; - unsigned long nr_strict_required = end_pfn - blockpfn; unsigned long flags; bool locked = false; @@ -264,11 +263,12 @@ static unsigned long isolate_freepages_block(struct compact_control *cc, nr_scanned++; if (!pfn_valid_within(blockpfn)) - continue; + goto isolate_fail; + if (!valid_page) valid_page = page; if (!PageBuddy(page)) - continue; + goto isolate_fail; /* * The zone lock must be held to isolate freepages. @@ -289,12 +289,10 @@ static unsigned long isolate_freepages_block(struct compact_control *cc, /* Recheck this is a buddy page under lock */ if (!PageBuddy(page)) - continue; + goto isolate_fail; /* Found a free page, break it into order-0 pages */ isolated = split_free_page(page); - if (!isolated && strict) - break; total_isolated += isolated; for (i = 0; i < isolated; i++) { list_add(&page->lru, freelist); @@ -305,7 +303,15 @@ static unsigned long isolate_freepages_block(struct compact_control *cc, if (isolated) { blockpfn += isolated - 1; cursor += isolated - 1; + continue; } + +isolate_fail: + if (strict) + break; + else + continue; + } trace_mm_compaction_isolate_freepages(nr_scanned, total_isolated); @@ -315,7 +321,7 @@ static unsigned long isolate_freepages_block(struct compact_control *cc, * pages requested were isolated. If there were any failures, 0 is * returned and CMA will fail. */ - if (strict && nr_strict_required > total_isolated) + if (strict && blockpfn < end_pfn) total_isolated = 0; if (locked) From 70335abb2689c8cd5df91bf2d95a65649addf50b Mon Sep 17 00:00:00 2001 From: Artem Fetishev Date: Mon, 10 Mar 2014 15:49:45 -0700 Subject: [PATCH 126/259] fs/proc/base.c: fix GPF in /proc/$PID/map_files The expected logic of proc_map_files_get_link() is either to return 0 and initialize 'path' or return an error and leave 'path' uninitialized. By the time dname_to_vma_addr() returns 0 the corresponding vma may have already be gone. In this case the path is not initialized but the return value is still 0. This results in 'general protection fault' inside d_path(). Steps to reproduce: CONFIG_CHECKPOINT_RESTORE=y fd = open(...); while (1) { mmap(fd, ...); munmap(fd, ...); } ls -la /proc/$PID/map_files Addresses https://bugzilla.kernel.org/show_bug.cgi?id=68991 Signed-off-by: Artem Fetishev Signed-off-by: Aleksandr Terekhov Reported-by: Acked-by: Pavel Emelyanov Acked-by: Cyrill Gorcunov Reviewed-by: "Eric W. Biederman" Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/proc/base.c b/fs/proc/base.c index 51507065263b..b9760628e1fd 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -1824,6 +1824,7 @@ static int proc_map_files_get_link(struct dentry *dentry, struct path *path) if (rc) goto out_mmput; + rc = -ENOENT; down_read(&mm->mmap_sem); vma = find_exact_vma(mm, vm_start, vm_end); if (vma && vma->vm_file) { From 2216ee853017f9c9371106c5c02d4fe42f61cbfa Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Mon, 10 Mar 2014 15:49:46 -0700 Subject: [PATCH 127/259] mm/Kconfig: fix URL for zsmalloc benchmark The help text for CONFIG_PGTABLE_MAPPING has an incorrect URL. While we're at it, remove the unnecessary footnote notation. Signed-off-by: Ben Hutchings Acked-by: Minchan Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/mm/Kconfig b/mm/Kconfig index 2d9f1504d75e..2888024e0b0a 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -575,5 +575,5 @@ config PGTABLE_MAPPING then you should select this. This causes zsmalloc to use page table mapping rather than copying for object mapping. - You can check speed with zsmalloc benchmark[1]. - [1] https://github.com/spartacus06/zsmalloc + You can check speed with zsmalloc benchmark: + https://github.com/spartacus06/zsmapbench From 2930ffc7593b64fe00fd7c5a0a7f543078d73ed9 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Mon, 10 Mar 2014 15:49:48 -0700 Subject: [PATCH 128/259] revert "kallsyms: fix absolute addresses for kASLR" Revert the recently applied 0f55159d091c ("kallsyms: fix absolute addresses for kASLR"). Kees said : This got NAKed, please don't apply -- this patch works for x86 and : ARM, but may cause problems for others: : : https://lkml.org/lkml/2014/2/24/718 It appears that Kees will be fixing all this up for 3.15. Cc: Andy Honig Cc: Kees Cook Cc: Michal Marek Cc: Rusty Russell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- scripts/kallsyms.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c index 276e84b8a8e5..10085de886fe 100644 --- a/scripts/kallsyms.c +++ b/scripts/kallsyms.c @@ -330,7 +330,8 @@ static void write_src(void) printf("\tPTR\t_text + %#llx\n", table[i].addr - _text); else - printf("\tPTR\t%#llx\n", table[i].addr); + printf("\tPTR\t_text - %#llx\n", + _text - table[i].addr); } else { printf("\tPTR\t%#llx\n", table[i].addr); } From 1443176fd6857744d6772e6a607d7d08f0f6afd9 Mon Sep 17 00:00:00 2001 From: Michael Opdenacker Date: Mon, 10 Mar 2014 15:49:49 -0700 Subject: [PATCH 129/259] MAINTAINERS: blackfin: add git repository Add the git repository currently in use for blackfin architecture development. This information was obtained from Steven Miao. Signed-off-by: Michael Opdenacker Cc: Steven Miao Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index b7befe758429..1ecfde109667 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1738,6 +1738,7 @@ F: include/uapi/linux/bfs_fs.h BLACKFIN ARCHITECTURE M: Steven Miao L: adi-buildroot-devel@lists.sourceforge.net +T: git git://git.code.sf.net/p/adi-linux/code W: http://blackfin.uclinux.org S: Supported F: arch/blackfin/ From 53942232369ec317f77dc8c2d5a0637bb85a6e84 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Mon, 10 Mar 2014 15:49:50 -0700 Subject: [PATCH 130/259] tools/testing/selftests/ipc/msgque.c: handle msgget failure return correctly A failed msgget causes the test to return an uninitialised value in ret. Assign ret to -errno on error exit. Signed-off-by: Colin Ian King Acked-by: Davidlohr Bueso Cc: Stanislav Kinsbursky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- tools/testing/selftests/ipc/msgque.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/ipc/msgque.c b/tools/testing/selftests/ipc/msgque.c index d66418237d21..aa290c0de6f5 100644 --- a/tools/testing/selftests/ipc/msgque.c +++ b/tools/testing/selftests/ipc/msgque.c @@ -201,6 +201,7 @@ int main(int argc, char **argv) msgque.msq_id = msgget(msgque.key, IPC_CREAT | IPC_EXCL | 0666); if (msgque.msq_id == -1) { + err = -errno; printf("Can't create queue\n"); goto err_out; } From d7d673a591701f131e53d4fd4e2b9352f1316642 Mon Sep 17 00:00:00 2001 From: Sergei Antonov Date: Mon, 10 Mar 2014 15:49:51 -0700 Subject: [PATCH 131/259] hfsplus: add HFSX subfolder count support Adds support for HFSX 'HasFolderCount' flag and a corresponding 'folderCount' field in folder records. (For reference see HFS_FOLDERCOUNT and kHFSHasFolderCountBit/kHFSHasFolderCountMask in Apple's source code.) Ignoring subfolder count leads to fs errors found by Mac: ... Checking catalog hierarchy. HasFolderCount flag needs to be set (id = 105) (It should be 0x10 instead of 0) Incorrect folder count in a directory (id = 2) (It should be 7 instead of 6) ... Steps to reproduce: Format with "newfs_hfs -s /dev/diskXXX". Mount in Linux. Create a new directory in root. Unmount. Run "fsck_hfs /dev/diskXXX". The patch handles directory creation, deletion, and rename. Signed-off-by: Sergei Antonov Reviewed-by: Vyacheslav Dubeyko Cc: Al Viro Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hfsplus/catalog.c | 41 ++++++++++++++++++++++++++++++++++++++++ fs/hfsplus/hfsplus_fs.h | 1 + fs/hfsplus/hfsplus_raw.h | 6 ++++-- fs/hfsplus/inode.c | 9 +++++++++ 4 files changed, 55 insertions(+), 2 deletions(-) diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index 968ce411db53..32602c667b4a 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -103,6 +103,8 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, folder = &entry->folder; memset(folder, 0, sizeof(*folder)); folder->type = cpu_to_be16(HFSPLUS_FOLDER); + if (test_bit(HFSPLUS_SB_HFSX, &sbi->flags)) + folder->flags |= cpu_to_be16(HFSPLUS_HAS_FOLDER_COUNT); folder->id = cpu_to_be32(inode->i_ino); HFSPLUS_I(inode)->create_date = folder->create_date = @@ -203,6 +205,36 @@ int hfsplus_find_cat(struct super_block *sb, u32 cnid, return hfs_brec_find(fd, hfs_find_rec_by_key); } +static void hfsplus_subfolders_inc(struct inode *dir) +{ + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); + + if (test_bit(HFSPLUS_SB_HFSX, &sbi->flags)) { + /* + * Increment subfolder count. Note, the value is only meaningful + * for folders with HFSPLUS_HAS_FOLDER_COUNT flag set. + */ + HFSPLUS_I(dir)->subfolders++; + } +} + +static void hfsplus_subfolders_dec(struct inode *dir) +{ + struct hfsplus_sb_info *sbi = HFSPLUS_SB(dir->i_sb); + + if (test_bit(HFSPLUS_SB_HFSX, &sbi->flags)) { + /* + * Decrement subfolder count. Note, the value is only meaningful + * for folders with HFSPLUS_HAS_FOLDER_COUNT flag set. + * + * Check for zero. Some subfolders may have been created + * by an implementation ignorant of this counter. + */ + if (HFSPLUS_I(dir)->subfolders) + HFSPLUS_I(dir)->subfolders--; + } +} + int hfsplus_create_cat(u32 cnid, struct inode *dir, struct qstr *str, struct inode *inode) { @@ -247,6 +279,8 @@ int hfsplus_create_cat(u32 cnid, struct inode *dir, goto err1; dir->i_size++; + if (S_ISDIR(inode->i_mode)) + hfsplus_subfolders_inc(dir); dir->i_mtime = dir->i_ctime = CURRENT_TIME_SEC; hfsplus_mark_inode_dirty(dir, HFSPLUS_I_CAT_DIRTY); @@ -336,6 +370,8 @@ int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) goto out; dir->i_size--; + if (type == HFSPLUS_FOLDER) + hfsplus_subfolders_dec(dir); dir->i_mtime = dir->i_ctime = CURRENT_TIME_SEC; hfsplus_mark_inode_dirty(dir, HFSPLUS_I_CAT_DIRTY); @@ -380,6 +416,7 @@ int hfsplus_rename_cat(u32 cnid, hfs_bnode_read(src_fd.bnode, &entry, src_fd.entryoffset, src_fd.entrylength); + type = be16_to_cpu(entry.type); /* create new dir entry with the data from the old entry */ hfsplus_cat_build_key(sb, dst_fd.search_key, dst_dir->i_ino, dst_name); @@ -394,6 +431,8 @@ int hfsplus_rename_cat(u32 cnid, if (err) goto out; dst_dir->i_size++; + if (type == HFSPLUS_FOLDER) + hfsplus_subfolders_inc(dst_dir); dst_dir->i_mtime = dst_dir->i_ctime = CURRENT_TIME_SEC; /* finally remove the old entry */ @@ -405,6 +444,8 @@ int hfsplus_rename_cat(u32 cnid, if (err) goto out; src_dir->i_size--; + if (type == HFSPLUS_FOLDER) + hfsplus_subfolders_dec(src_dir); src_dir->i_mtime = src_dir->i_ctime = CURRENT_TIME_SEC; /* remove old thread entry */ diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 08846425b67f..62d571eb69ba 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -242,6 +242,7 @@ struct hfsplus_inode_info { */ sector_t fs_blocks; u8 userflags; /* BSD user file flags */ + u32 subfolders; /* Subfolder count (HFSX only) */ struct list_head open_dir_list; loff_t phys_size; diff --git a/fs/hfsplus/hfsplus_raw.h b/fs/hfsplus/hfsplus_raw.h index 8ffb3a8ffe75..5a126828d85e 100644 --- a/fs/hfsplus/hfsplus_raw.h +++ b/fs/hfsplus/hfsplus_raw.h @@ -261,7 +261,7 @@ struct hfsplus_cat_folder { struct DInfo user_info; struct DXInfo finder_info; __be32 text_encoding; - u32 reserved; + __be32 subfolders; /* Subfolder count in HFSX. Reserved in HFS+. */ } __packed; /* HFS file info (stolen from hfs.h) */ @@ -301,11 +301,13 @@ struct hfsplus_cat_file { struct hfsplus_fork_raw rsrc_fork; } __packed; -/* File attribute bits */ +/* File and folder flag bits */ #define HFSPLUS_FILE_LOCKED 0x0001 #define HFSPLUS_FILE_THREAD_EXISTS 0x0002 #define HFSPLUS_XATTR_EXISTS 0x0004 #define HFSPLUS_ACL_EXISTS 0x0008 +#define HFSPLUS_HAS_FOLDER_COUNT 0x0010 /* Folder has subfolder count + * (HFSX only) */ /* HFS+ catalog thread (part of a cat_entry) */ struct hfsplus_cat_thread { diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index fa929f325f87..a4f45bd88a63 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -375,6 +375,7 @@ struct inode *hfsplus_new_inode(struct super_block *sb, umode_t mode) hip->extent_state = 0; hip->flags = 0; hip->userflags = 0; + hip->subfolders = 0; memset(hip->first_extents, 0, sizeof(hfsplus_extent_rec)); memset(hip->cached_extents, 0, sizeof(hfsplus_extent_rec)); hip->alloc_blocks = 0; @@ -494,6 +495,10 @@ int hfsplus_cat_read_inode(struct inode *inode, struct hfs_find_data *fd) inode->i_ctime = hfsp_mt2ut(folder->attribute_mod_date); HFSPLUS_I(inode)->create_date = folder->create_date; HFSPLUS_I(inode)->fs_blocks = 0; + if (folder->flags & cpu_to_be16(HFSPLUS_HAS_FOLDER_COUNT)) { + HFSPLUS_I(inode)->subfolders = + be32_to_cpu(folder->subfolders); + } inode->i_op = &hfsplus_dir_inode_operations; inode->i_fop = &hfsplus_dir_operations; } else if (type == HFSPLUS_FILE) { @@ -566,6 +571,10 @@ int hfsplus_cat_write_inode(struct inode *inode) folder->content_mod_date = hfsp_ut2mt(inode->i_mtime); folder->attribute_mod_date = hfsp_ut2mt(inode->i_ctime); folder->valence = cpu_to_be32(inode->i_size - 2); + if (folder->flags & cpu_to_be16(HFSPLUS_HAS_FOLDER_COUNT)) { + folder->subfolders = + cpu_to_be32(HFSPLUS_I(inode)->subfolders); + } hfs_bnode_write(fd.bnode, &entry, fd.entryoffset, sizeof(struct hfsplus_cat_folder)); } else if (HFSPLUS_IS_RSRC(inode)) { From 0eb808eb753cd44e740da8e64f5c97a2a8e07578 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Mon, 10 Mar 2014 15:49:52 -0700 Subject: [PATCH 132/259] cris: convert ffs from an object-like macro to a function-like macro This avoids bad interactions with code using identifiers called "ffs": drivers/usb/gadget/f_fs.c: In function 'ffsmod_init': drivers/usb/gadget/f_fs.c:2693:494: error: 'ffsusb_func' undeclared (first use in this function) drivers/usb/gadget/f_fs.c:2693:494: note: each undeclared identifier is reported only once for each function it appears in drivers/usb/gadget/f_fs.c: In function 'ffsmod_exit': drivers/usb/gadget/f_fs.c:2693:677: error: 'ffsusb_func' undeclared (first use in this function) drivers/usb/gadget/f_fs.c: At top level: drivers/usb/gadget/f_fs.c:2693:35: warning: 'kernel_ffsusb_func' defined but not used [-Wunused-variable] drivers/usb/gadget/f_fs.c: In function 'ffsmod_init': drivers/usb/gadget/f_fs.c:2693:15: warning: control reaches end of non-void function [-Wreturn-type] See http://kisskb.ellerman.id.au/kisskb/buildresult/10715817/ Signed-off-by: Geert Uytterhoeven Cc: Mikael Starvik Cc: Jesper Nilsson Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/cris/include/asm/bitops.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/cris/include/asm/bitops.h b/arch/cris/include/asm/bitops.h index 184066ceb1f6..053c17b36559 100644 --- a/arch/cris/include/asm/bitops.h +++ b/arch/cris/include/asm/bitops.h @@ -144,7 +144,7 @@ static inline int test_and_change_bit(int nr, volatile unsigned long *addr) * definition, which doesn't have the same semantics. We don't want to * use -fno-builtin, so just hide the name ffs. */ -#define ffs kernel_ffs +#define ffs(x) kernel_ffs(x) #include #include From 5bd076708664313f2bdbbc1cf71093313b7774a1 Mon Sep 17 00:00:00 2001 From: Annie Li Date: Mon, 10 Mar 2014 22:58:34 +0800 Subject: [PATCH 133/259] Xen-netback: Fix issue caused by using gso_type wrongly Current netback uses gso_type to check whether the skb contains gso offload, and this is wrong. Gso_size is the right one to check gso existence, and gso_type is only used to check gso type. Some skbs contains nonzero gso_type and zero gso_size, current netback would treat these skbs as gso and create wrong response for this. This also causes ssh failure to domu from other server. V2: use skb_is_gso function as Paul Durrant suggested Signed-off-by: Annie Li Acked-by: Wei Liu Reviewed-by: Paul Durrant Signed-off-by: David S. Miller --- drivers/net/xen-netback/netback.c | 39 ++++++++++++++----------------- 1 file changed, 18 insertions(+), 21 deletions(-) diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c index e5284bca2d90..438d0c09b7e6 100644 --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -240,7 +240,7 @@ static void xenvif_gop_frag_copy(struct xenvif *vif, struct sk_buff *skb, struct gnttab_copy *copy_gop; struct xenvif_rx_meta *meta; unsigned long bytes; - int gso_type; + int gso_type = XEN_NETIF_GSO_TYPE_NONE; /* Data must not cross a page boundary. */ BUG_ON(size + offset > PAGE_SIZE<gso_type & SKB_GSO_TCPV4) - gso_type = XEN_NETIF_GSO_TYPE_TCPV4; - else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) - gso_type = XEN_NETIF_GSO_TYPE_TCPV6; - else - gso_type = XEN_NETIF_GSO_TYPE_NONE; + if (skb_is_gso(skb)) { + if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4) + gso_type = XEN_NETIF_GSO_TYPE_TCPV4; + else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) + gso_type = XEN_NETIF_GSO_TYPE_TCPV6; + } if (*head && ((1 << gso_type) & vif->gso_mask)) vif->rx.req_cons++; @@ -338,19 +338,15 @@ static int xenvif_gop_skb(struct sk_buff *skb, int head = 1; int old_meta_prod; int gso_type; - int gso_size; old_meta_prod = npo->meta_prod; - if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4) { - gso_type = XEN_NETIF_GSO_TYPE_TCPV4; - gso_size = skb_shinfo(skb)->gso_size; - } else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) { - gso_type = XEN_NETIF_GSO_TYPE_TCPV6; - gso_size = skb_shinfo(skb)->gso_size; - } else { - gso_type = XEN_NETIF_GSO_TYPE_NONE; - gso_size = 0; + gso_type = XEN_NETIF_GSO_TYPE_NONE; + if (skb_is_gso(skb)) { + if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4) + gso_type = XEN_NETIF_GSO_TYPE_TCPV4; + else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) + gso_type = XEN_NETIF_GSO_TYPE_TCPV6; } /* Set up a GSO prefix descriptor, if necessary */ @@ -358,7 +354,7 @@ static int xenvif_gop_skb(struct sk_buff *skb, req = RING_GET_REQUEST(&vif->rx, vif->rx.req_cons++); meta = npo->meta + npo->meta_prod++; meta->gso_type = gso_type; - meta->gso_size = gso_size; + meta->gso_size = skb_shinfo(skb)->gso_size; meta->size = 0; meta->id = req->id; } @@ -368,7 +364,7 @@ static int xenvif_gop_skb(struct sk_buff *skb, if ((1 << gso_type) & vif->gso_mask) { meta->gso_type = gso_type; - meta->gso_size = gso_size; + meta->gso_size = skb_shinfo(skb)->gso_size; } else { meta->gso_type = XEN_NETIF_GSO_TYPE_NONE; meta->gso_size = 0; @@ -500,8 +496,9 @@ static void xenvif_rx_action(struct xenvif *vif) size = skb_frag_size(&skb_shinfo(skb)->frags[i]); max_slots_needed += DIV_ROUND_UP(size, PAGE_SIZE); } - if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4 || - skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) + if (skb_is_gso(skb) && + (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4 || + skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)) max_slots_needed++; /* If the skb may not fit then bail out now */ From dd38743b4cc2f86be250eaf156cf113ba3dd531a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Peter=20Bostr=C3=B6m?= Date: Mon, 10 Mar 2014 16:17:15 +0100 Subject: [PATCH 134/259] vlan: Set correct source MAC address with TX VLAN offload enabled MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With TX VLAN offload enabled the source MAC address for frames sent using the VLAN interface is currently set to the address of the real interface. This is wrong since the VLAN interface may be configured with a different address. The bug was introduced in commit 2205369a314e12fcec4781cc73ac9c08fc2b47de ("vlan: Fix header ops passthru when doing TX VLAN offload."). This patch sets the source address before calling the create function of the real interface. Signed-off-by: Peter Boström Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index de51c48c4393..4b65aa492fb6 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -538,6 +538,9 @@ static int vlan_passthru_hard_header(struct sk_buff *skb, struct net_device *dev struct vlan_dev_priv *vlan = vlan_dev_priv(dev); struct net_device *real_dev = vlan->real_dev; + if (saddr == NULL) + saddr = dev->dev_addr; + return dev_hard_header(skb, real_dev, type, daddr, saddr, len); } From d44753b843e093f9e1f2f14806fbe106fff74898 Mon Sep 17 00:00:00 2001 From: Juri Lelli Date: Mon, 3 Mar 2014 12:09:21 +0100 Subject: [PATCH 135/259] sched/deadline: Deny unprivileged users to set/change SCHED_DEADLINE policy Deny the use of SCHED_DEADLINE policy to unprivileged users. Even if root users can set the policy for normal users, we don't want the latter to be able to change their parameters (safest behavior). Signed-off-by: Juri Lelli Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1393844961-18097-1-git-send-email-juri.lelli@gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 6edbef296ece..f5c6635b806c 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -3338,6 +3338,15 @@ static int __sched_setscheduler(struct task_struct *p, return -EPERM; } + /* + * Can't set/change SCHED_DEADLINE policy at all for now + * (safest behavior); in the future we would like to allow + * unprivileged DL tasks to increase their relative deadline + * or reduce their runtime (both ways reducing utilization) + */ + if (dl_policy(policy)) + return -EPERM; + /* * Treat SCHED_IDLE as nice 20. Only allow a switch to * SCHED_NORMAL if the RLIMIT_NICE would normally permit it. From 177c53d943368fc97644ebc0a250dc8e2d124250 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 28 Feb 2014 13:39:05 +0100 Subject: [PATCH 136/259] stop_machine: Fix^2 race between stop_two_cpus() and stop_cpus() We must use smp_call_function_single(.wait=1) for the irq_cpu_stop_queue_work() to ensure the queueing is actually done under stop_cpus_lock. Without this we could have dropped the lock by the time we do the queueing and get the race we tried to fix. Fixes: 7053ea1a34fa ("stop_machine: Fix race between stop_two_cpus() and stop_cpus()") Signed-off-by: Peter Zijlstra Cc: Prarit Bhargava Cc: Rik van Riel Cc: Mel Gorman Cc: Christoph Hellwig Cc: Andrew Morton Link: http://lkml.kernel.org/r/20140228123905.GK3104@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- kernel/stop_machine.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c index 84571e09c907..01fbae5b97b7 100644 --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -293,7 +293,7 @@ int stop_two_cpus(unsigned int cpu1, unsigned int cpu2, cpu_stop_fn_t fn, void * */ smp_call_function_single(min(cpu1, cpu2), &irq_cpu_stop_queue_work, - &call_args, 0); + &call_args, 1); lg_local_unlock(&stop_cpus_lock); preempt_enable(); From 96b3d28bf4b00f62fc8386ff5d487d1830793a3d Mon Sep 17 00:00:00 2001 From: Fernando Luis Vazquez Cao Date: Thu, 6 Mar 2014 14:25:28 +0900 Subject: [PATCH 137/259] sched/clock: Prevent tracing recursion in sched_clock_cpu() Prevent tracing of preempt_disable/enable() in sched_clock_cpu(). When CONFIG_DEBUG_PREEMPT is enabled, preempt_disable/enable() are traced and this causes trace_clock() users (and probably others) to go into an infinite recursion. Systems with a stable sched_clock() are not affected. This problem is similar to that fixed by upstream commit 95ef1e52922 ("KVM guest: prevent tracing recursion with kvmclock"). Signed-off-by: Fernando Luis Vazquez Cao Signed-off-by: Peter Zijlstra Acked-by: Steven Rostedt Cc: Andrew Morton Cc: Linus Torvalds Link: http://lkml.kernel.org/r/1394083528.4524.3.camel@nexus Signed-off-by: Ingo Molnar --- kernel/sched/clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/sched/clock.c b/kernel/sched/clock.c index 43c2bcc35761..b30a2924ef14 100644 --- a/kernel/sched/clock.c +++ b/kernel/sched/clock.c @@ -301,14 +301,14 @@ u64 sched_clock_cpu(int cpu) if (unlikely(!sched_clock_running)) return 0ull; - preempt_disable(); + preempt_disable_notrace(); scd = cpu_sdc(cpu); if (cpu != smp_processor_id()) clock = sched_clock_remote(scd); else clock = sched_clock_local(scd); - preempt_enable(); + preempt_enable_notrace(); return clock; } From b7b4839d93e50adccef29eccb694807cdcb8bee3 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Thu, 6 Mar 2014 12:20:28 -0500 Subject: [PATCH 138/259] perf/x86: Fix leak in uncore_type_init failure paths The error path of uncore_type_init() frees up any allocations that were made along the way, but it relies upon type->pmus being set, which only happens if the function succeeds. As type->pmus remains null in this case, the call to uncore_type_exit will do nothing. Moving the assignment earlier will allow us to actually free those allocations should something go awry. Signed-off-by: Dave Jones Acked-by: Peter Zijlstra Link: http://lkml.kernel.org/r/20140306172028.GA552@redhat.com Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event_intel_uncore.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.c b/arch/x86/kernel/cpu/perf_event_intel_uncore.c index c88f7f4b03ee..047f540cf3f7 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c +++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.c @@ -3334,6 +3334,8 @@ static int __init uncore_type_init(struct intel_uncore_type *type) if (!pmus) return -ENOMEM; + type->pmus = pmus; + type->unconstrainted = (struct event_constraint) __EVENT_CONSTRAINT(0, (1ULL << type->num_counters) - 1, 0, type->num_counters, 0, 0); @@ -3369,7 +3371,6 @@ static int __init uncore_type_init(struct intel_uncore_type *type) } type->pmu_group = &uncore_pmu_attr_group; - type->pmus = pmus; return 0; fail: uncore_type_exit(type); From 09df7c4c8097ca4a11393b1edd4997d786daad52 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Mon, 10 Mar 2014 19:32:22 -0400 Subject: [PATCH 139/259] x86: Remove CONFIG_X86_OOSTORE This was an optimization that made memcpy type benchmarks a little faster on ancient (Circa 1998) IDT Winchip CPUs. In real-life workloads, it wasn't even noticable, and I doubt anyone is running benchmarks on 16 year old silicon any more. Given this code has likely seen very little use over the last decade, let's just remove it. Signed-off-by: Dave Jones Signed-off-by: Linus Torvalds --- arch/x86/Kconfig.cpu | 4 - arch/x86/include/asm/barrier.h | 8 +- arch/x86/include/asm/io.h | 2 +- arch/x86/include/asm/spinlock.h | 5 +- arch/x86/kernel/cpu/centaur.c | 272 -------------------------------- arch/x86/um/asm/barrier.h | 4 - 6 files changed, 5 insertions(+), 290 deletions(-) diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index c026cca5602c..f3aaf231b4e5 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -341,10 +341,6 @@ config X86_USE_3DNOW def_bool y depends on (MCYRIXIII || MK7 || MGEODE_LX) && !UML -config X86_OOSTORE - def_bool y - depends on (MWINCHIP3D || MWINCHIPC6) && MTRR - # # P6_NOPs are a relatively minor optimization that require a family >= # 6 processor, except that it is broken on certain VIA chips. diff --git a/arch/x86/include/asm/barrier.h b/arch/x86/include/asm/barrier.h index 04a48903b2eb..69bbb4845020 100644 --- a/arch/x86/include/asm/barrier.h +++ b/arch/x86/include/asm/barrier.h @@ -85,11 +85,7 @@ #else # define smp_rmb() barrier() #endif -#ifdef CONFIG_X86_OOSTORE -# define smp_wmb() wmb() -#else -# define smp_wmb() barrier() -#endif +#define smp_wmb() barrier() #define smp_read_barrier_depends() read_barrier_depends() #define set_mb(var, value) do { (void)xchg(&var, value); } while (0) #else /* !SMP */ @@ -100,7 +96,7 @@ #define set_mb(var, value) do { var = value; barrier(); } while (0) #endif /* SMP */ -#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) +#if defined(CONFIG_X86_PPRO_FENCE) /* * For either of these options x86 doesn't have a strong TSO memory diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h index 34f69cb9350a..91d9c69a629e 100644 --- a/arch/x86/include/asm/io.h +++ b/arch/x86/include/asm/io.h @@ -237,7 +237,7 @@ memcpy_toio(volatile void __iomem *dst, const void *src, size_t count) static inline void flush_write_buffers(void) { -#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE) +#if defined(CONFIG_X86_PPRO_FENCE) asm volatile("lock; addl $0,0(%%esp)": : :"memory"); #endif } diff --git a/arch/x86/include/asm/spinlock.h b/arch/x86/include/asm/spinlock.h index bf156ded74b5..0f62f5482d91 100644 --- a/arch/x86/include/asm/spinlock.h +++ b/arch/x86/include/asm/spinlock.h @@ -26,10 +26,9 @@ # define LOCK_PTR_REG "D" #endif -#if defined(CONFIG_X86_32) && \ - (defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE)) +#if defined(CONFIG_X86_32) && (defined(CONFIG_X86_PPRO_FENCE)) /* - * On PPro SMP or if we are using OOSTORE, we use a locked operation to unlock + * On PPro SMP, we use a locked operation to unlock * (PPro errata 66, 92) */ # define UNLOCK_LOCK_PREFIX LOCK_PREFIX diff --git a/arch/x86/kernel/cpu/centaur.c b/arch/x86/kernel/cpu/centaur.c index 8779edab684e..d8fba5c15fbd 100644 --- a/arch/x86/kernel/cpu/centaur.c +++ b/arch/x86/kernel/cpu/centaur.c @@ -8,236 +8,6 @@ #include "cpu.h" -#ifdef CONFIG_X86_OOSTORE - -static u32 power2(u32 x) -{ - u32 s = 1; - - while (s <= x) - s <<= 1; - - return s >>= 1; -} - - -/* - * Set up an actual MCR - */ -static void centaur_mcr_insert(int reg, u32 base, u32 size, int key) -{ - u32 lo, hi; - - hi = base & ~0xFFF; - lo = ~(size-1); /* Size is a power of 2 so this makes a mask */ - lo &= ~0xFFF; /* Remove the ctrl value bits */ - lo |= key; /* Attribute we wish to set */ - wrmsr(reg+MSR_IDT_MCR0, lo, hi); - mtrr_centaur_report_mcr(reg, lo, hi); /* Tell the mtrr driver */ -} - -/* - * Figure what we can cover with MCR's - * - * Shortcut: We know you can't put 4Gig of RAM on a winchip - */ -static u32 ramtop(void) -{ - u32 clip = 0xFFFFFFFFUL; - u32 top = 0; - int i; - - for (i = 0; i < e820.nr_map; i++) { - unsigned long start, end; - - if (e820.map[i].addr > 0xFFFFFFFFUL) - continue; - /* - * Don't MCR over reserved space. Ignore the ISA hole - * we frob around that catastrophe already - */ - if (e820.map[i].type == E820_RESERVED) { - if (e820.map[i].addr >= 0x100000UL && - e820.map[i].addr < clip) - clip = e820.map[i].addr; - continue; - } - start = e820.map[i].addr; - end = e820.map[i].addr + e820.map[i].size; - if (start >= end) - continue; - if (end > top) - top = end; - } - /* - * Everything below 'top' should be RAM except for the ISA hole. - * Because of the limited MCR's we want to map NV/ACPI into our - * MCR range for gunk in RAM - * - * Clip might cause us to MCR insufficient RAM but that is an - * acceptable failure mode and should only bite obscure boxes with - * a VESA hole at 15Mb - * - * The second case Clip sometimes kicks in is when the EBDA is marked - * as reserved. Again we fail safe with reasonable results - */ - if (top > clip) - top = clip; - - return top; -} - -/* - * Compute a set of MCR's to give maximum coverage - */ -static int centaur_mcr_compute(int nr, int key) -{ - u32 mem = ramtop(); - u32 root = power2(mem); - u32 base = root; - u32 top = root; - u32 floor = 0; - int ct = 0; - - while (ct < nr) { - u32 fspace = 0; - u32 high; - u32 low; - - /* - * Find the largest block we will fill going upwards - */ - high = power2(mem-top); - - /* - * Find the largest block we will fill going downwards - */ - low = base/2; - - /* - * Don't fill below 1Mb going downwards as there - * is an ISA hole in the way. - */ - if (base <= 1024*1024) - low = 0; - - /* - * See how much space we could cover by filling below - * the ISA hole - */ - - if (floor == 0) - fspace = 512*1024; - else if (floor == 512*1024) - fspace = 128*1024; - - /* And forget ROM space */ - - /* - * Now install the largest coverage we get - */ - if (fspace > high && fspace > low) { - centaur_mcr_insert(ct, floor, fspace, key); - floor += fspace; - } else if (high > low) { - centaur_mcr_insert(ct, top, high, key); - top += high; - } else if (low > 0) { - base -= low; - centaur_mcr_insert(ct, base, low, key); - } else - break; - ct++; - } - /* - * We loaded ct values. We now need to set the mask. The caller - * must do this bit. - */ - return ct; -} - -static void centaur_create_optimal_mcr(void) -{ - int used; - int i; - - /* - * Allocate up to 6 mcrs to mark as much of ram as possible - * as write combining and weak write ordered. - * - * To experiment with: Linux never uses stack operations for - * mmio spaces so we could globally enable stack operation wc - * - * Load the registers with type 31 - full write combining, all - * writes weakly ordered. - */ - used = centaur_mcr_compute(6, 31); - - /* - * Wipe unused MCRs - */ - for (i = used; i < 8; i++) - wrmsr(MSR_IDT_MCR0+i, 0, 0); -} - -static void winchip2_create_optimal_mcr(void) -{ - u32 lo, hi; - int used; - int i; - - /* - * Allocate up to 6 mcrs to mark as much of ram as possible - * as write combining, weak store ordered. - * - * Load the registers with type 25 - * 8 - weak write ordering - * 16 - weak read ordering - * 1 - write combining - */ - used = centaur_mcr_compute(6, 25); - - /* - * Mark the registers we are using. - */ - rdmsr(MSR_IDT_MCR_CTRL, lo, hi); - for (i = 0; i < used; i++) - lo |= 1<<(9+i); - wrmsr(MSR_IDT_MCR_CTRL, lo, hi); - - /* - * Wipe unused MCRs - */ - - for (i = used; i < 8; i++) - wrmsr(MSR_IDT_MCR0+i, 0, 0); -} - -/* - * Handle the MCR key on the Winchip 2. - */ -static void winchip2_unprotect_mcr(void) -{ - u32 lo, hi; - u32 key; - - rdmsr(MSR_IDT_MCR_CTRL, lo, hi); - lo &= ~0x1C0; /* blank bits 8-6 */ - key = (lo>>17) & 7; - lo |= key<<6; /* replace with unlock key */ - wrmsr(MSR_IDT_MCR_CTRL, lo, hi); -} - -static void winchip2_protect_mcr(void) -{ - u32 lo, hi; - - rdmsr(MSR_IDT_MCR_CTRL, lo, hi); - lo &= ~0x1C0; /* blank bits 8-6 */ - wrmsr(MSR_IDT_MCR_CTRL, lo, hi); -} -#endif /* CONFIG_X86_OOSTORE */ - #define ACE_PRESENT (1 << 6) #define ACE_ENABLED (1 << 7) #define ACE_FCR (1 << 28) /* MSR_VIA_FCR */ @@ -362,20 +132,6 @@ static void init_centaur(struct cpuinfo_x86 *c) fcr_clr = DPDC; printk(KERN_NOTICE "Disabling bugged TSC.\n"); clear_cpu_cap(c, X86_FEATURE_TSC); -#ifdef CONFIG_X86_OOSTORE - centaur_create_optimal_mcr(); - /* - * Enable: - * write combining on non-stack, non-string - * write combining on string, all types - * weak write ordering - * - * The C6 original lacks weak read order - * - * Note 0x120 is write only on Winchip 1 - */ - wrmsr(MSR_IDT_MCR_CTRL, 0x01F0001F, 0); -#endif break; case 8: switch (c->x86_mask) { @@ -392,40 +148,12 @@ static void init_centaur(struct cpuinfo_x86 *c) fcr_set = ECX8|DSMC|DTLOCK|EMMX|EBRPRED|ERETSTK| E2MMX|EAMD3D; fcr_clr = DPDC; -#ifdef CONFIG_X86_OOSTORE - winchip2_unprotect_mcr(); - winchip2_create_optimal_mcr(); - rdmsr(MSR_IDT_MCR_CTRL, lo, hi); - /* - * Enable: - * write combining on non-stack, non-string - * write combining on string, all types - * weak write ordering - */ - lo |= 31; - wrmsr(MSR_IDT_MCR_CTRL, lo, hi); - winchip2_protect_mcr(); -#endif break; case 9: name = "3"; fcr_set = ECX8|DSMC|DTLOCK|EMMX|EBRPRED|ERETSTK| E2MMX|EAMD3D; fcr_clr = DPDC; -#ifdef CONFIG_X86_OOSTORE - winchip2_unprotect_mcr(); - winchip2_create_optimal_mcr(); - rdmsr(MSR_IDT_MCR_CTRL, lo, hi); - /* - * Enable: - * write combining on non-stack, non-string - * write combining on string, all types - * weak write ordering - */ - lo |= 31; - wrmsr(MSR_IDT_MCR_CTRL, lo, hi); - winchip2_protect_mcr(); -#endif break; default: name = "??"; diff --git a/arch/x86/um/asm/barrier.h b/arch/x86/um/asm/barrier.h index 7d01b8c56c00..cc04e67bfd05 100644 --- a/arch/x86/um/asm/barrier.h +++ b/arch/x86/um/asm/barrier.h @@ -40,11 +40,7 @@ #define smp_rmb() barrier() #endif /* CONFIG_X86_PPRO_FENCE */ -#ifdef CONFIG_X86_OOSTORE -#define smp_wmb() wmb() -#else /* CONFIG_X86_OOSTORE */ #define smp_wmb() barrier() -#endif /* CONFIG_X86_OOSTORE */ #define smp_read_barrier_depends() read_barrier_depends() #define set_mb(var, value) do { (void)xchg(&var, value); } while (0) From 731bd6a93a6e9172094a2322bd0ee964bb1f4d63 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Sun, 2 Feb 2014 22:56:23 -0800 Subject: [PATCH 140/259] x86, fpu: Check tsk_used_math() in kernel_fpu_end() for eager FPU For non-eager fpu mode, thread's fpu state is allocated during the first fpu usage (in the context of device not available exception). This (math_state_restore()) can be a blocking call and hence we enable interrupts (which were originally disabled when the exception happened), allocate memory and disable interrupts etc. But the eager-fpu mode, call's the same math_state_restore() from kernel_fpu_end(). The assumption being that tsk_used_math() is always set for the eager-fpu mode and thus avoid the code path of enabling interrupts, allocating fpu state using blocking call and disable interrupts etc. But the below issue was noticed by Maarten Baert, Nate Eldredge and few others: If a user process dumps core on an ecrypt fs while aesni-intel is loaded, we get a BUG() in __find_get_block() complaining that it was called with interrupts disabled; then all further accesses to our ecrypt fs hang and we have to reboot. The aesni-intel code (encrypting the core file that we are writing) needs the FPU and quite properly wraps its code in kernel_fpu_{begin,end}(), the latter of which calls math_state_restore(). So after kernel_fpu_end(), interrupts may be disabled, which nobody seems to expect, and they stay that way until we eventually get to __find_get_block() which barfs. For eager fpu, most the time, tsk_used_math() is true. At few instances during thread exit, signal return handling etc, tsk_used_math() might be false. In kernel_fpu_end(), for eager-fpu, call math_state_restore() only if tsk_used_math() is set. Otherwise, don't bother. Kernel code path which cleared tsk_used_math() knows what needs to be done with the fpu state. Reported-by: Maarten Baert Reported-by: Nate Eldredge Suggested-by: Linus Torvalds Signed-off-by: Suresh Siddha Link: http://lkml.kernel.org/r/1391410583.3801.6.camel@europa Cc: George Spelvin Signed-off-by: H. Peter Anvin --- arch/x86/kernel/i387.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index e8368c6dd2a2..d5dd80814419 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -86,10 +86,19 @@ EXPORT_SYMBOL(__kernel_fpu_begin); void __kernel_fpu_end(void) { - if (use_eager_fpu()) - math_state_restore(); - else + if (use_eager_fpu()) { + /* + * For eager fpu, most the time, tsk_used_math() is true. + * Restore the user math as we are done with the kernel usage. + * At few instances during thread exit, signal handling etc, + * tsk_used_math() is false. Those few places will take proper + * actions, so we don't need to restore the math here. + */ + if (likely(tsk_used_math(current))) + math_state_restore(); + } else { stts(); + } } EXPORT_SYMBOL(__kernel_fpu_end); From d25f06ea466ea521b563b76661180b4e44714ae6 Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Mon, 10 Mar 2014 06:55:55 -0400 Subject: [PATCH 141/259] vmxnet3: fix netpoll race condition vmxnet3's netpoll driver is incorrectly coded. It directly calls vmxnet3_do_poll, which is the driver internal napi poll routine. As the netpoll controller method doesn't block real napi polls in any way, there is a potential for race conditions in which the netpoll controller method and the napi poll method run concurrently. The result is data corruption causing panics such as this one recently observed: PID: 1371 TASK: ffff88023762caa0 CPU: 1 COMMAND: "rs:main Q:Reg" #0 [ffff88023abd5780] machine_kexec at ffffffff81038f3b #1 [ffff88023abd57e0] crash_kexec at ffffffff810c5d92 #2 [ffff88023abd58b0] oops_end at ffffffff8152b570 #3 [ffff88023abd58e0] die at ffffffff81010e0b #4 [ffff88023abd5910] do_trap at ffffffff8152add4 #5 [ffff88023abd5970] do_invalid_op at ffffffff8100cf95 #6 [ffff88023abd5a10] invalid_op at ffffffff8100bf9b [exception RIP: vmxnet3_rq_rx_complete+1968] RIP: ffffffffa00f1e80 RSP: ffff88023abd5ac8 RFLAGS: 00010086 RAX: 0000000000000000 RBX: ffff88023b5dcee0 RCX: 00000000000000c0 RDX: 0000000000000000 RSI: 00000000000005f2 RDI: ffff88023b5dcee0 RBP: ffff88023abd5b48 R8: 0000000000000000 R9: ffff88023a3b6048 R10: 0000000000000000 R11: 0000000000000002 R12: ffff8802398d4cd8 R13: ffff88023af35140 R14: ffff88023b60c890 R15: 0000000000000000 ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018 #7 [ffff88023abd5b50] vmxnet3_do_poll at ffffffffa00f204a [vmxnet3] #8 [ffff88023abd5b80] vmxnet3_netpoll at ffffffffa00f209c [vmxnet3] #9 [ffff88023abd5ba0] netpoll_poll_dev at ffffffff81472bb7 The fix is to do as other drivers do, and have the poll controller call the top half interrupt handler, which schedules a napi poll properly to recieve frames Tested by myself, successfully. Signed-off-by: Neil Horman CC: Shreyas Bhatewara CC: "VMware, Inc." CC: "David S. Miller" CC: stable@vger.kernel.org Reviewed-by: Shreyas N Bhatewara Signed-off-by: David S. Miller --- drivers/net/vmxnet3/vmxnet3_drv.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c index 3be786faaaec..b7daa02ff026 100644 --- a/drivers/net/vmxnet3/vmxnet3_drv.c +++ b/drivers/net/vmxnet3/vmxnet3_drv.c @@ -1761,12 +1761,18 @@ static void vmxnet3_netpoll(struct net_device *netdev) { struct vmxnet3_adapter *adapter = netdev_priv(netdev); + int i; - if (adapter->intr.mask_mode == VMXNET3_IMM_ACTIVE) - vmxnet3_disable_all_intrs(adapter); - - vmxnet3_do_poll(adapter, adapter->rx_queue[0].rx_ring[0].size); - vmxnet3_enable_all_intrs(adapter); + switch (adapter->intr.type) { + case VMXNET3_IT_MSIX: + for (i = 0; i < adapter->num_rx_queues; i++) + vmxnet3_msix_rx(0, &adapter->rx_queue[i]); + break; + case VMXNET3_IT_MSI: + default: + vmxnet3_intr(0, adapter->netdev); + break; + } } #endif /* CONFIG_NET_POLL_CONTROLLER */ From 83bf79b6bb64e686ccf0b66b6828b7bfe9f70ae9 Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Mon, 10 Mar 2014 13:40:31 +0100 Subject: [PATCH 142/259] stmmac: disable at run-time the EEE if not supported This patch is to disable the EEE (so HW and timers) for example when the phy communicates that the EEE can be supported anymore. Signed-off-by: Giuseppe Cavallaro Signed-off-by: David S. Miller --- .../net/ethernet/stmicro/stmmac/stmmac_main.c | 25 +++++++++++++++---- 1 file changed, 20 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index 078ad0ec8593..b418f9437bf0 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -286,10 +286,25 @@ bool stmmac_eee_init(struct stmmac_priv *priv) /* MAC core supports the EEE feature. */ if (priv->dma_cap.eee) { - /* Check if the PHY supports EEE */ - if (phy_init_eee(priv->phydev, 1)) - goto out; + int tx_lpi_timer = priv->tx_lpi_timer; + /* Check if the PHY supports EEE */ + if (phy_init_eee(priv->phydev, 1)) { + /* To manage at run-time if the EEE cannot be supported + * anymore (for example because the lp caps have been + * changed). + * In that case the driver disable own timers. + */ + if (priv->eee_active) { + pr_debug("stmmac: disable EEE\n"); + del_timer_sync(&priv->eee_ctrl_timer); + priv->hw->mac->set_eee_timer(priv->ioaddr, 0, + tx_lpi_timer); + } + priv->eee_active = 0; + goto out; + } + /* Activate the EEE and start timers */ if (!priv->eee_active) { priv->eee_active = 1; init_timer(&priv->eee_ctrl_timer); @@ -300,13 +315,13 @@ bool stmmac_eee_init(struct stmmac_priv *priv) priv->hw->mac->set_eee_timer(priv->ioaddr, STMMAC_DEFAULT_LIT_LS, - priv->tx_lpi_timer); + tx_lpi_timer); } else /* Set HW EEE according to the speed */ priv->hw->mac->set_eee_pls(priv->ioaddr, priv->phydev->link); - pr_info("stmmac: Energy-Efficient Ethernet initialized\n"); + pr_debug("stmmac: Energy-Efficient Ethernet initialized\n"); ret = true; } From d916701c670701e1ab4d0ed3d4e64d6023bccec9 Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Mon, 10 Mar 2014 13:40:32 +0100 Subject: [PATCH 143/259] stmmac: fix and better tune the default buffer sizes This patch is to fix and tune the default buffer sizes. It reduces the default bufsize used by the driver from 4KiB to 1536 bytes. Patch has been tested on both ARM and SH4 platform based. Signed-off-by: Giuseppe Cavallaro Signed-off-by: David S. Miller --- drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index b418f9437bf0..536583655643 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -92,8 +92,8 @@ static int tc = TC_DEFAULT; module_param(tc, int, S_IRUGO | S_IWUSR); MODULE_PARM_DESC(tc, "DMA threshold control value"); -#define DMA_BUFFER_SIZE BUF_SIZE_4KiB -static int buf_sz = DMA_BUFFER_SIZE; +#define DEFAULT_BUFSIZE 1536 +static int buf_sz = DEFAULT_BUFSIZE; module_param(buf_sz, int, S_IRUGO | S_IWUSR); MODULE_PARM_DESC(buf_sz, "DMA buffer size"); @@ -136,8 +136,8 @@ static void stmmac_verify_args(void) dma_rxsize = DMA_RX_SIZE; if (unlikely(dma_txsize < 0)) dma_txsize = DMA_TX_SIZE; - if (unlikely((buf_sz < DMA_BUFFER_SIZE) || (buf_sz > BUF_SIZE_16KiB))) - buf_sz = DMA_BUFFER_SIZE; + if (unlikely((buf_sz < DEFAULT_BUFSIZE) || (buf_sz > BUF_SIZE_16KiB))) + buf_sz = DEFAULT_BUFSIZE; if (unlikely(flow_ctrl > 1)) flow_ctrl = FLOW_AUTO; else if (likely(flow_ctrl < 0)) @@ -901,10 +901,10 @@ static int stmmac_set_bfsize(int mtu, int bufsize) ret = BUF_SIZE_8KiB; else if (mtu >= BUF_SIZE_2KiB) ret = BUF_SIZE_4KiB; - else if (mtu >= DMA_BUFFER_SIZE) + else if (mtu > DEFAULT_BUFSIZE) ret = BUF_SIZE_2KiB; else - ret = DMA_BUFFER_SIZE; + ret = DEFAULT_BUFSIZE; return ret; } From 29896a674c8ef3d75134dacb7b9cbb3f5b894b6d Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Mon, 10 Mar 2014 13:40:33 +0100 Subject: [PATCH 144/259] stmmac: fix chained mode This patch is to fix the chain mode that was broken and generated a panic. This patch reviews the chain/ring modes now shaing the same structure and taking care about the pointers and callbacks. Signed-off-by: Giuseppe Cavallaro Signed-off-by: David S. Miller --- .../net/ethernet/stmicro/stmmac/chain_mode.c | 2 +- drivers/net/ethernet/stmicro/stmmac/common.h | 20 ++----- .../net/ethernet/stmicro/stmmac/ring_mode.c | 9 ++- .../net/ethernet/stmicro/stmmac/stmmac_main.c | 56 ++++++++----------- 4 files changed, 34 insertions(+), 53 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/chain_mode.c b/drivers/net/ethernet/stmicro/stmmac/chain_mode.c index 72d282bf33a5..c553f6b5a913 100644 --- a/drivers/net/ethernet/stmicro/stmmac/chain_mode.c +++ b/drivers/net/ethernet/stmicro/stmmac/chain_mode.c @@ -151,7 +151,7 @@ static void stmmac_clean_desc3(void *priv_ptr, struct dma_desc *p) sizeof(struct dma_desc))); } -const struct stmmac_chain_mode_ops chain_mode_ops = { +const struct stmmac_mode_ops chain_mode_ops = { .init = stmmac_init_dma_chain, .is_jumbo_frm = stmmac_is_jumbo_frm, .jumbo_frm = stmmac_jumbo_frm, diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h index 7834a3993946..74610f3aca9e 100644 --- a/drivers/net/ethernet/stmicro/stmmac/common.h +++ b/drivers/net/ethernet/stmicro/stmmac/common.h @@ -419,20 +419,13 @@ struct mii_regs { unsigned int data; /* MII Data */ }; -struct stmmac_ring_mode_ops { - unsigned int (*is_jumbo_frm) (int len, int ehn_desc); - unsigned int (*jumbo_frm) (void *priv, struct sk_buff *skb, int csum); - void (*refill_desc3) (void *priv, struct dma_desc *p); - void (*init_desc3) (struct dma_desc *p); - void (*clean_desc3) (void *priv, struct dma_desc *p); - int (*set_16kib_bfsize) (int mtu); -}; - -struct stmmac_chain_mode_ops { +struct stmmac_mode_ops { void (*init) (void *des, dma_addr_t phy_addr, unsigned int size, unsigned int extend_desc); unsigned int (*is_jumbo_frm) (int len, int ehn_desc); unsigned int (*jumbo_frm) (void *priv, struct sk_buff *skb, int csum); + int (*set_16kib_bfsize)(int mtu); + void (*init_desc3)(struct dma_desc *p); void (*refill_desc3) (void *priv, struct dma_desc *p); void (*clean_desc3) (void *priv, struct dma_desc *p); }; @@ -441,8 +434,7 @@ struct mac_device_info { const struct stmmac_ops *mac; const struct stmmac_desc_ops *desc; const struct stmmac_dma_ops *dma; - const struct stmmac_ring_mode_ops *ring; - const struct stmmac_chain_mode_ops *chain; + const struct stmmac_mode_ops *mode; const struct stmmac_hwtimestamp *ptp; struct mii_regs mii; /* MII register Addresses */ struct mac_link link; @@ -460,7 +452,7 @@ void stmmac_get_mac_addr(void __iomem *ioaddr, unsigned char *addr, void stmmac_set_mac(void __iomem *ioaddr, bool enable); void dwmac_dma_flush_tx_fifo(void __iomem *ioaddr); -extern const struct stmmac_ring_mode_ops ring_mode_ops; -extern const struct stmmac_chain_mode_ops chain_mode_ops; +extern const struct stmmac_mode_ops ring_mode_ops; +extern const struct stmmac_mode_ops chain_mode_ops; #endif /* __COMMON_H__ */ diff --git a/drivers/net/ethernet/stmicro/stmmac/ring_mode.c b/drivers/net/ethernet/stmicro/stmmac/ring_mode.c index a96c7c2f5f3f..650a4be6bce5 100644 --- a/drivers/net/ethernet/stmicro/stmmac/ring_mode.c +++ b/drivers/net/ethernet/stmicro/stmmac/ring_mode.c @@ -100,10 +100,9 @@ static void stmmac_refill_desc3(void *priv_ptr, struct dma_desc *p) { struct stmmac_priv *priv = (struct stmmac_priv *)priv_ptr; - if (unlikely(priv->plat->has_gmac)) - /* Fill DES3 in case of RING mode */ - if (priv->dma_buf_sz >= BUF_SIZE_8KiB) - p->des3 = p->des2 + BUF_SIZE_8KiB; + /* Fill DES3 in case of RING mode */ + if (priv->dma_buf_sz >= BUF_SIZE_8KiB) + p->des3 = p->des2 + BUF_SIZE_8KiB; } /* In ring mode we need to fill the desc3 because it is used as buffer */ @@ -126,7 +125,7 @@ static int stmmac_set_16kib_bfsize(int mtu) return ret; } -const struct stmmac_ring_mode_ops ring_mode_ops = { +const struct stmmac_mode_ops ring_mode_ops = { .is_jumbo_frm = stmmac_is_jumbo_frm, .jumbo_frm = stmmac_jumbo_frm, .refill_desc3 = stmmac_refill_desc3, diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index 536583655643..8543e1cfd55e 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -966,9 +966,9 @@ static int stmmac_init_rx_buffers(struct stmmac_priv *priv, struct dma_desc *p, p->des2 = priv->rx_skbuff_dma[i]; - if ((priv->mode == STMMAC_RING_MODE) && + if ((priv->hw->mode->init_desc3) && (priv->dma_buf_sz == BUF_SIZE_16KiB)) - priv->hw->ring->init_desc3(p); + priv->hw->mode->init_desc3(p); return 0; } @@ -999,11 +999,8 @@ static int init_dma_desc_rings(struct net_device *dev) unsigned int bfsize = 0; int ret = -ENOMEM; - /* Set the max buffer size according to the DESC mode - * and the MTU. Note that RING mode allows 16KiB bsize. - */ - if (priv->mode == STMMAC_RING_MODE) - bfsize = priv->hw->ring->set_16kib_bfsize(dev->mtu); + if (priv->hw->mode->set_16kib_bfsize) + bfsize = priv->hw->mode->set_16kib_bfsize(dev->mtu); if (bfsize < BUF_SIZE_16KiB) bfsize = stmmac_set_bfsize(dev->mtu, priv->dma_buf_sz); @@ -1044,15 +1041,15 @@ static int init_dma_desc_rings(struct net_device *dev) /* Setup the chained descriptor addresses */ if (priv->mode == STMMAC_CHAIN_MODE) { if (priv->extend_desc) { - priv->hw->chain->init(priv->dma_erx, priv->dma_rx_phy, - rxsize, 1); - priv->hw->chain->init(priv->dma_etx, priv->dma_tx_phy, - txsize, 1); + priv->hw->mode->init(priv->dma_erx, priv->dma_rx_phy, + rxsize, 1); + priv->hw->mode->init(priv->dma_etx, priv->dma_tx_phy, + txsize, 1); } else { - priv->hw->chain->init(priv->dma_rx, priv->dma_rx_phy, - rxsize, 0); - priv->hw->chain->init(priv->dma_tx, priv->dma_tx_phy, - txsize, 0); + priv->hw->mode->init(priv->dma_rx, priv->dma_rx_phy, + rxsize, 0); + priv->hw->mode->init(priv->dma_tx, priv->dma_tx_phy, + txsize, 0); } } @@ -1303,7 +1300,7 @@ static void stmmac_tx_clean(struct stmmac_priv *priv) DMA_TO_DEVICE); priv->tx_skbuff_dma[entry] = 0; } - priv->hw->ring->clean_desc3(priv, p); + priv->hw->mode->clean_desc3(priv, p); if (likely(skb != NULL)) { dev_kfree_skb(skb); @@ -1859,6 +1856,7 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev) int nfrags = skb_shinfo(skb)->nr_frags; struct dma_desc *desc, *first; unsigned int nopaged_len = skb_headlen(skb); + unsigned int enh_desc = priv->plat->enh_desc; if (unlikely(stmmac_tx_avail(priv) < nfrags + 1)) { if (!netif_queue_stopped(dev)) { @@ -1886,27 +1884,19 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev) first = desc; /* To program the descriptors according to the size of the frame */ - if (priv->mode == STMMAC_RING_MODE) { - is_jumbo = priv->hw->ring->is_jumbo_frm(skb->len, - priv->plat->enh_desc); - if (unlikely(is_jumbo)) - entry = priv->hw->ring->jumbo_frm(priv, skb, - csum_insertion); - } else { - is_jumbo = priv->hw->chain->is_jumbo_frm(skb->len, - priv->plat->enh_desc); - if (unlikely(is_jumbo)) - entry = priv->hw->chain->jumbo_frm(priv, skb, - csum_insertion); - } + if (enh_desc) + is_jumbo = priv->hw->mode->is_jumbo_frm(skb->len, enh_desc); + if (likely(!is_jumbo)) { desc->des2 = dma_map_single(priv->device, skb->data, nopaged_len, DMA_TO_DEVICE); priv->tx_skbuff_dma[entry] = desc->des2; priv->hw->desc->prepare_tx_desc(desc, 1, nopaged_len, csum_insertion, priv->mode); - } else + } else { desc = first; + entry = priv->hw->mode->jumbo_frm(priv, skb, csum_insertion); + } for (i = 0; i < nfrags; i++) { const skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; @@ -2044,7 +2034,7 @@ static inline void stmmac_rx_refill(struct stmmac_priv *priv) p->des2 = priv->rx_skbuff_dma[entry]; - priv->hw->ring->refill_desc3(priv, p); + priv->hw->mode->refill_desc3(priv, p); if (netif_msg_rx_status(priv)) pr_debug("\trefill entry #%d\n", entry); @@ -2648,11 +2638,11 @@ static int stmmac_hw_init(struct stmmac_priv *priv) /* To use the chained or ring mode */ if (chain_mode) { - priv->hw->chain = &chain_mode_ops; + priv->hw->mode = &chain_mode_ops; pr_info(" Chain mode enabled\n"); priv->mode = STMMAC_CHAIN_MODE; } else { - priv->hw->ring = &ring_mode_ops; + priv->hw->mode = &ring_mode_ops; pr_info(" Ring mode enabled\n"); priv->mode = STMMAC_RING_MODE; } From c5e9103dc397114523c65bd24ff0548bcdca54b4 Mon Sep 17 00:00:00 2001 From: Giuseppe CAVALLARO Date: Mon, 10 Mar 2014 13:40:34 +0100 Subject: [PATCH 145/259] stmmac: dwmac-sti: fix broken STiD127 compatibility This is to fix the compatibility to the STiD127 SoC. Signed-off-by: Giuseppe Cavallaro Cc: Srinivas Kandagatla Signed-off-by: David S. Miller --- drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c index c61bc72b8e90..8fb32a80f1c1 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c @@ -36,7 +36,7 @@ static const struct of_device_id stmmac_dt_ids[] = { #ifdef CONFIG_DWMAC_STI { .compatible = "st,stih415-dwmac", .data = &sti_gmac_data}, { .compatible = "st,stih416-dwmac", .data = &sti_gmac_data}, - { .compatible = "st,stih127-dwmac", .data = &sti_gmac_data}, + { .compatible = "st,stid127-dwmac", .data = &sti_gmac_data}, #endif /* SoC specific glue layers should come before generic bindings */ { .compatible = "st,spear600-gmac"}, From 89935315f192abf7068d0044cefc84f162c3c81f Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Tue, 11 Mar 2014 22:40:27 +0800 Subject: [PATCH 146/259] PNP / ACPI: proper handling of ACPI IO/Memory resource parsing failures Before commit b355cee88e3b (ACPI / resources: ignore invalid ACPI device resources), if acpi_dev_resource_memory()/acpi_dev_resource_io() returns false, it means the the resource is not a memeory/IO resource. But after commit b355cee88e3b, those functions return false if the given memory/IO resource entry is invalid (the length of the resource is zero). This breaks pnpacpi_allocated_resource(), because it now recognizes the invalid memory/io resources as resources of unknown type. Thus users see confusing warning messages on machines with zero length ACPI memory/IO resources. Fix the problem by rearranging pnpacpi_allocated_resource() so that it calls acpi_dev_resource_memory() for memory type and IO type resources only, respectively. Fixes: b355cee88e3b (ACPI / resources: ignore invalid ACPI device resources) Signed-off-by: Zhang Rui Reported-and-tested-by: Markus Trippelsdorf Reported-and-tested-by: Julian Wollrath Reported-and-tested-by: Paul Bolle [rjw: Changelog] Signed-off-by: Rafael J. Wysocki --- drivers/pnp/pnpacpi/rsparser.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/drivers/pnp/pnpacpi/rsparser.c b/drivers/pnp/pnpacpi/rsparser.c index 167f3d00c916..66977ebf13b3 100644 --- a/drivers/pnp/pnpacpi/rsparser.c +++ b/drivers/pnp/pnpacpi/rsparser.c @@ -183,9 +183,7 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res, struct resource r = {0}; int i, flags; - if (acpi_dev_resource_memory(res, &r) - || acpi_dev_resource_io(res, &r) - || acpi_dev_resource_address_space(res, &r) + if (acpi_dev_resource_address_space(res, &r) || acpi_dev_resource_ext_address_space(res, &r)) { pnp_add_resource(dev, &r); return AE_OK; @@ -217,6 +215,17 @@ static acpi_status pnpacpi_allocated_resource(struct acpi_resource *res, } switch (res->type) { + case ACPI_RESOURCE_TYPE_MEMORY24: + case ACPI_RESOURCE_TYPE_MEMORY32: + case ACPI_RESOURCE_TYPE_FIXED_MEMORY32: + if (acpi_dev_resource_memory(res, &r)) + pnp_add_resource(dev, &r); + break; + case ACPI_RESOURCE_TYPE_IO: + case ACPI_RESOURCE_TYPE_FIXED_IO: + if (acpi_dev_resource_io(res, &r)) + pnp_add_resource(dev, &r); + break; case ACPI_RESOURCE_TYPE_DMA: dma = &res->data.dma; if (dma->channel_count > 0 && dma->channels[0] != (u8) -1) From 8cb19905e9287a93ce7c2cbbdf742a060b00e219 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 10 Mar 2014 18:29:04 +0200 Subject: [PATCH 147/259] skbuff: skb_segment: s/frag/nskb_frag/ frag points at nskb, so name it appropriately Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- net/core/skbuff.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 5d6236d9fdce..60e8cd717cb3 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2876,7 +2876,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) do { struct sk_buff *nskb; - skb_frag_t *frag; + skb_frag_t *nskb_frag; int hsize; int size; @@ -2969,7 +2969,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) continue; } - frag = skb_shinfo(nskb)->frags; + nskb_frag = skb_shinfo(nskb)->frags; skb_copy_from_linear_data_offset(skb, offset, skb_put(nskb, hsize), hsize); @@ -2997,13 +2997,13 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) goto err; } - *frag = *skb_frag; - __skb_frag_ref(frag); - size = skb_frag_size(frag); + *nskb_frag = *skb_frag; + __skb_frag_ref(nskb_frag); + size = skb_frag_size(nskb_frag); if (pos < offset) { - frag->page_offset += offset - pos; - skb_frag_size_sub(frag, offset - pos); + nskb_frag->page_offset += offset - pos; + skb_frag_size_sub(nskb_frag, offset - pos); } skb_shinfo(nskb)->nr_frags++; @@ -3013,11 +3013,11 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) skb_frag++; pos += size; } else { - skb_frag_size_sub(frag, pos + size - (offset + len)); + skb_frag_size_sub(nskb_frag, pos + size - (offset + len)); goto skip_fraglist; } - frag++; + nskb_frag++; } skip_fraglist: From 4e1beba12d094c6c761ba5c49032b9b9e46380e8 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 10 Mar 2014 18:29:14 +0200 Subject: [PATCH 148/259] skbuff: skb_segment: s/skb_frag/frag/ skb_frag can in fact point at either skb or fskb so rename it generally "frag". Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- net/core/skbuff.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 60e8cd717cb3..d788a9845762 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2850,7 +2850,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) struct sk_buff *segs = NULL; struct sk_buff *tail = NULL; struct sk_buff *fskb = skb_shinfo(skb)->frag_list; - skb_frag_t *skb_frag = skb_shinfo(skb)->frags; + skb_frag_t *frag = skb_shinfo(skb)->frags; unsigned int mss = skb_shinfo(skb)->gso_size; unsigned int doffset = skb->data - skb_mac_header(skb); unsigned int offset = doffset; @@ -2896,19 +2896,19 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) i = 0; nfrags = skb_shinfo(fskb)->nr_frags; - skb_frag = skb_shinfo(fskb)->frags; + frag = skb_shinfo(fskb)->frags; pos += skb_headlen(fskb); while (pos < offset + len) { BUG_ON(i >= nfrags); - size = skb_frag_size(skb_frag); + size = skb_frag_size(frag); if (pos + size > offset + len) break; i++; pos += size; - skb_frag++; + frag++; } nskb = skb_clone(fskb, GFP_ATOMIC); @@ -2982,7 +2982,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) i = 0; nfrags = skb_shinfo(fskb)->nr_frags; - skb_frag = skb_shinfo(fskb)->frags; + frag = skb_shinfo(fskb)->frags; BUG_ON(!nfrags); @@ -2997,7 +2997,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) goto err; } - *nskb_frag = *skb_frag; + *nskb_frag = *frag; __skb_frag_ref(nskb_frag); size = skb_frag_size(nskb_frag); @@ -3010,7 +3010,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) if (pos + size <= offset + len) { i++; - skb_frag++; + frag++; pos += size; } else { skb_frag_size_sub(nskb_frag, pos + size - (offset + len)); From df5771ffefb13f8af5392bd54fd7e2b596a3a357 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 10 Mar 2014 18:29:19 +0200 Subject: [PATCH 149/259] skbuff: skb_segment: s/skb/head_skb/ rename local variable to make it easier to tell at a glance that we are dealing with a head skb. Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- net/core/skbuff.c | 46 ++++++++++++++++++++++++---------------------- 1 file changed, 24 insertions(+), 22 deletions(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index d788a9845762..fdc065dc869a 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2838,41 +2838,42 @@ EXPORT_SYMBOL_GPL(skb_pull_rcsum); /** * skb_segment - Perform protocol segmentation on skb. - * @skb: buffer to segment + * @head_skb: buffer to segment * @features: features for the output path (see dev->features) * * This function performs segmentation on the given skb. It returns * a pointer to the first in a list of new skbs for the segments. * In case of error it returns ERR_PTR(err). */ -struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) +struct sk_buff *skb_segment(struct sk_buff *head_skb, + netdev_features_t features) { struct sk_buff *segs = NULL; struct sk_buff *tail = NULL; - struct sk_buff *fskb = skb_shinfo(skb)->frag_list; - skb_frag_t *frag = skb_shinfo(skb)->frags; - unsigned int mss = skb_shinfo(skb)->gso_size; - unsigned int doffset = skb->data - skb_mac_header(skb); + struct sk_buff *fskb = skb_shinfo(head_skb)->frag_list; + skb_frag_t *frag = skb_shinfo(head_skb)->frags; + unsigned int mss = skb_shinfo(head_skb)->gso_size; + unsigned int doffset = head_skb->data - skb_mac_header(head_skb); unsigned int offset = doffset; - unsigned int tnl_hlen = skb_tnl_header_len(skb); + unsigned int tnl_hlen = skb_tnl_header_len(head_skb); unsigned int headroom; unsigned int len; __be16 proto; bool csum; int sg = !!(features & NETIF_F_SG); - int nfrags = skb_shinfo(skb)->nr_frags; + int nfrags = skb_shinfo(head_skb)->nr_frags; int err = -ENOMEM; int i = 0; int pos; - proto = skb_network_protocol(skb); + proto = skb_network_protocol(head_skb); if (unlikely(!proto)) return ERR_PTR(-EINVAL); csum = !!can_checksum_protocol(features, proto); - __skb_push(skb, doffset); - headroom = skb_headroom(skb); - pos = skb_headlen(skb); + __skb_push(head_skb, doffset); + headroom = skb_headroom(head_skb); + pos = skb_headlen(head_skb); do { struct sk_buff *nskb; @@ -2880,11 +2881,11 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) int hsize; int size; - len = skb->len - offset; + len = head_skb->len - offset; if (len > mss) len = mss; - hsize = skb_headlen(skb) - offset; + hsize = skb_headlen(head_skb) - offset; if (hsize < 0) hsize = 0; if (hsize > len || !sg) @@ -2933,7 +2934,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) __skb_push(nskb, doffset); } else { nskb = __alloc_skb(hsize + doffset + headroom, - GFP_ATOMIC, skb_alloc_rx_flag(skb), + GFP_ATOMIC, skb_alloc_rx_flag(head_skb), NUMA_NO_NODE); if (unlikely(!nskb)) @@ -2949,12 +2950,12 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) segs = nskb; tail = nskb; - __copy_skb_header(nskb, skb); - nskb->mac_len = skb->mac_len; + __copy_skb_header(nskb, head_skb); + nskb->mac_len = head_skb->mac_len; skb_headers_offset_update(nskb, skb_headroom(nskb) - headroom); - skb_copy_from_linear_data_offset(skb, -tnl_hlen, + skb_copy_from_linear_data_offset(head_skb, -tnl_hlen, nskb->data - tnl_hlen, doffset + tnl_hlen); @@ -2963,7 +2964,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) if (!sg) { nskb->ip_summed = CHECKSUM_NONE; - nskb->csum = skb_copy_and_csum_bits(skb, offset, + nskb->csum = skb_copy_and_csum_bits(head_skb, offset, skb_put(nskb, len), len, 0); continue; @@ -2971,10 +2972,11 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) nskb_frag = skb_shinfo(nskb)->frags; - skb_copy_from_linear_data_offset(skb, offset, + skb_copy_from_linear_data_offset(head_skb, offset, skb_put(nskb, hsize), hsize); - skb_shinfo(nskb)->tx_flags = skb_shinfo(skb)->tx_flags & SKBTX_SHARED_FRAG; + skb_shinfo(nskb)->tx_flags = skb_shinfo(head_skb)->tx_flags & + SKBTX_SHARED_FRAG; while (pos < offset + len) { if (i >= nfrags) { @@ -3031,7 +3033,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, netdev_features_t features) nskb->len - doffset, 0); nskb->ip_summed = CHECKSUM_NONE; } - } while ((offset += len) < skb->len); + } while ((offset += len) < head_skb->len); return segs; From 1a4cedaf65491e66e1e55b8428c89209da729209 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 10 Mar 2014 19:27:59 +0200 Subject: [PATCH 150/259] skbuff: skb_segment: s/fskb/list_skb/ fskb is unrelated to frag: it's coming from frag_list. Rename it list_skb to avoid confusion. Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- net/core/skbuff.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index fdc065dc869a..dc4f7683ff52 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2850,7 +2850,7 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, { struct sk_buff *segs = NULL; struct sk_buff *tail = NULL; - struct sk_buff *fskb = skb_shinfo(head_skb)->frag_list; + struct sk_buff *list_skb = skb_shinfo(head_skb)->frag_list; skb_frag_t *frag = skb_shinfo(head_skb)->frags; unsigned int mss = skb_shinfo(head_skb)->gso_size; unsigned int doffset = head_skb->data - skb_mac_header(head_skb); @@ -2891,14 +2891,14 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, if (hsize > len || !sg) hsize = len; - if (!hsize && i >= nfrags && skb_headlen(fskb) && - (skb_headlen(fskb) == len || sg)) { - BUG_ON(skb_headlen(fskb) > len); + if (!hsize && i >= nfrags && skb_headlen(list_skb) && + (skb_headlen(list_skb) == len || sg)) { + BUG_ON(skb_headlen(list_skb) > len); i = 0; - nfrags = skb_shinfo(fskb)->nr_frags; - frag = skb_shinfo(fskb)->frags; - pos += skb_headlen(fskb); + nfrags = skb_shinfo(list_skb)->nr_frags; + frag = skb_shinfo(list_skb)->frags; + pos += skb_headlen(list_skb); while (pos < offset + len) { BUG_ON(i >= nfrags); @@ -2912,8 +2912,8 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, frag++; } - nskb = skb_clone(fskb, GFP_ATOMIC); - fskb = fskb->next; + nskb = skb_clone(list_skb, GFP_ATOMIC); + list_skb = list_skb->next; if (unlikely(!nskb)) goto err; @@ -2980,15 +2980,15 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, while (pos < offset + len) { if (i >= nfrags) { - BUG_ON(skb_headlen(fskb)); + BUG_ON(skb_headlen(list_skb)); i = 0; - nfrags = skb_shinfo(fskb)->nr_frags; - frag = skb_shinfo(fskb)->frags; + nfrags = skb_shinfo(list_skb)->nr_frags; + frag = skb_shinfo(list_skb)->frags; BUG_ON(!nfrags); - fskb = fskb->next; + list_skb = list_skb->next; } if (unlikely(skb_shinfo(nskb)->nr_frags >= From 1fd819ecb90cc9b822cd84d3056ddba315d3340f Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Mon, 10 Mar 2014 19:28:08 +0200 Subject: [PATCH 151/259] skbuff: skb_segment: orphan frags before copying skb_segment copies frags around, so we need to copy them carefully to avoid accessing user memory after reporting completion to userspace through a callback. skb_segment doesn't normally happen on datapath: TSO needs to be disabled - so disabling zero copy in this case does not look like a big deal. Signed-off-by: Michael S. Tsirkin Acked-by: Herbert Xu Signed-off-by: David S. Miller --- net/core/skbuff.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/core/skbuff.c b/net/core/skbuff.c index dc4f7683ff52..869c7afe3b07 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2854,6 +2854,7 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, skb_frag_t *frag = skb_shinfo(head_skb)->frags; unsigned int mss = skb_shinfo(head_skb)->gso_size; unsigned int doffset = head_skb->data - skb_mac_header(head_skb); + struct sk_buff *frag_skb = head_skb; unsigned int offset = doffset; unsigned int tnl_hlen = skb_tnl_header_len(head_skb); unsigned int headroom; @@ -2898,6 +2899,7 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, i = 0; nfrags = skb_shinfo(list_skb)->nr_frags; frag = skb_shinfo(list_skb)->frags; + frag_skb = list_skb; pos += skb_headlen(list_skb); while (pos < offset + len) { @@ -2985,6 +2987,7 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, i = 0; nfrags = skb_shinfo(list_skb)->nr_frags; frag = skb_shinfo(list_skb)->frags; + frag_skb = list_skb; BUG_ON(!nfrags); @@ -2999,6 +3002,9 @@ struct sk_buff *skb_segment(struct sk_buff *head_skb, goto err; } + if (unlikely(skb_orphan_frags(frag_skb, GFP_ATOMIC))) + goto err; + *nskb_frag = *frag; __skb_frag_ref(nskb_frag); size = skb_frag_size(nskb_frag); From c3f9b01849ef3bc69024990092b9f42e20df7797 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 10 Mar 2014 09:50:11 -0700 Subject: [PATCH 152/259] tcp: tcp_release_cb() should release socket ownership Lars Persson reported following deadlock : -000 |M:0x0:0x802B6AF8(asm) <-- arch_spin_lock -001 |tcp_v4_rcv(skb = 0x8BD527A0) <-- sk = 0x8BE6B2A0 -002 |ip_local_deliver_finish(skb = 0x8BD527A0) -003 |__netif_receive_skb_core(skb = 0x8BD527A0, ?) -004 |netif_receive_skb(skb = 0x8BD527A0) -005 |elk_poll(napi = 0x8C770500, budget = 64) -006 |net_rx_action(?) -007 |__do_softirq() -008 |do_softirq() -009 |local_bh_enable() -010 |tcp_rcv_established(sk = 0x8BE6B2A0, skb = 0x87D3A9E0, th = 0x814EBE14, ?) -011 |tcp_v4_do_rcv(sk = 0x8BE6B2A0, skb = 0x87D3A9E0) -012 |tcp_delack_timer_handler(sk = 0x8BE6B2A0) -013 |tcp_release_cb(sk = 0x8BE6B2A0) -014 |release_sock(sk = 0x8BE6B2A0) -015 |tcp_sendmsg(?, sk = 0x8BE6B2A0, ?, ?) -016 |sock_sendmsg(sock = 0x8518C4C0, msg = 0x87D8DAA8, size = 4096) -017 |kernel_sendmsg(?, ?, ?, ?, size = 4096) -018 |smb_send_kvec() -019 |smb_send_rqst(server = 0x87C4D400, rqst = 0x87D8DBA0) -020 |cifs_call_async() -021 |cifs_async_writev(wdata = 0x87FD6580) -022 |cifs_writepages(mapping = 0x852096E4, wbc = 0x87D8DC88) -023 |__writeback_single_inode(inode = 0x852095D0, wbc = 0x87D8DC88) -024 |writeback_sb_inodes(sb = 0x87D6D800, wb = 0x87E4A9C0, work = 0x87D8DD88) -025 |__writeback_inodes_wb(wb = 0x87E4A9C0, work = 0x87D8DD88) -026 |wb_writeback(wb = 0x87E4A9C0, work = 0x87D8DD88) -027 |wb_do_writeback(wb = 0x87E4A9C0, force_wait = 0) -028 |bdi_writeback_workfn(work = 0x87E4A9CC) -029 |process_one_work(worker = 0x8B045880, work = 0x87E4A9CC) -030 |worker_thread(__worker = 0x8B045880) -031 |kthread(_create = 0x87CADD90) -032 |ret_from_kernel_thread(asm) Bug occurs because __tcp_checksum_complete_user() enables BH, assuming it is running from softirq context. Lars trace involved a NIC without RX checksum support but other points are problematic as well, like the prequeue stuff. Problem is triggered by a timer, that found socket being owned by user. tcp_release_cb() should call tcp_write_timer_handler() or tcp_delack_timer_handler() in the appropriate context : BH disabled and socket lock held, but 'owned' field cleared, as if they were running from timer handlers. Fixes: 6f458dfb4092 ("tcp: improve latencies of timer triggered events") Reported-by: Lars Persson Tested-by: Lars Persson Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/sock.h | 5 +++++ net/core/sock.c | 5 ++++- net/ipv4/tcp_output.c | 11 +++++++++++ 3 files changed, 20 insertions(+), 1 deletion(-) diff --git a/include/net/sock.h b/include/net/sock.h index 7c4167bc8266..b9586a137cad 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -1488,6 +1488,11 @@ static inline void sk_wmem_free_skb(struct sock *sk, struct sk_buff *skb) */ #define sock_owned_by_user(sk) ((sk)->sk_lock.owned) +static inline void sock_release_ownership(struct sock *sk) +{ + sk->sk_lock.owned = 0; +} + /* * Macro so as to not evaluate some arguments when * lockdep is not enabled. diff --git a/net/core/sock.c b/net/core/sock.c index 5b6a9431b017..c0fc6bdad1e3 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -2357,10 +2357,13 @@ void release_sock(struct sock *sk) if (sk->sk_backlog.tail) __release_sock(sk); + /* Warning : release_cb() might need to release sk ownership, + * ie call sock_release_ownership(sk) before us. + */ if (sk->sk_prot->release_cb) sk->sk_prot->release_cb(sk); - sk->sk_lock.owned = 0; + sock_release_ownership(sk); if (waitqueue_active(&sk->sk_lock.wq)) wake_up(&sk->sk_lock.wq); spin_unlock_bh(&sk->sk_lock.slock); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index f0eb4e337ec8..17a11e65e57f 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -767,6 +767,17 @@ void tcp_release_cb(struct sock *sk) if (flags & (1UL << TCP_TSQ_DEFERRED)) tcp_tsq_handler(sk); + /* Here begins the tricky part : + * We are called from release_sock() with : + * 1) BH disabled + * 2) sk_lock.slock spinlock held + * 3) socket owned by us (sk->sk_lock.owned == 1) + * + * But following code is meant to be called from BH handlers, + * so we should keep BH disabled, but early release socket ownership + */ + sock_release_ownership(sk); + if (flags & (1UL << TCP_WRITE_TIMER_DEFERRED)) { tcp_write_timer_handler(sk); __sock_put(sk); From 9ed973cc40c588abeaa58aea0683ea665132d11d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Linus=20L=C3=BCssing?= Date: Mon, 10 Mar 2014 22:25:24 +0100 Subject: [PATCH 153/259] bridge: multicast: add sanity check for general query destination MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit General IGMP and MLD queries are supposed to have the multicast link-local all-nodes address as their destination according to RFC2236 section 9, RFC3376 section 4.1.12/9.1, RFC2710 section 8 and RFC3810 section 5.1.15. Without this check, such malformed IGMP/MLD queries can result in a denial of service: The queries are ignored by most IGMP/MLD listeners therefore they will not respond with an IGMP/MLD report. However, without this patch these malformed MLD queries would enable the snooping part in the bridge code, potentially shutting down the according ports towards these hosts for multicast traffic as the bridge did not learn about these listeners. Reported-by: Jan Stancek Signed-off-by: Linus Lüssing Signed-off-by: David S. Miller --- net/bridge/br_multicast.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index fb0e36fac668..e56bae4f59ce 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -1181,6 +1181,14 @@ static int br_ip4_multicast_query(struct net_bridge *br, IGMPV3_MRC(ih3->code) * (HZ / IGMP_TIMER_SCALE) : 1; } + /* RFC2236+RFC3376 (IGMPv2+IGMPv3) require the multicast link layer + * all-systems destination addresses (224.0.0.1) for general queries + */ + if (!group && iph->daddr != htonl(INADDR_ALLHOSTS_GROUP)) { + err = -EINVAL; + goto out; + } + br_multicast_query_received(br, port, &br->ip4_querier, !!iph->saddr, max_delay); @@ -1228,6 +1236,7 @@ static int br_ip6_multicast_query(struct net_bridge *br, unsigned long max_delay; unsigned long now = jiffies; const struct in6_addr *group = NULL; + bool is_general_query; int err = 0; spin_lock(&br->multicast_lock); @@ -1262,6 +1271,16 @@ static int br_ip6_multicast_query(struct net_bridge *br, max_delay = max(msecs_to_jiffies(mldv2_mrc(mld2q)), 1UL); } + is_general_query = group && ipv6_addr_any(group); + + /* RFC2710+RFC3810 (MLDv1+MLDv2) require the multicast link layer + * all-nodes destination address (ff02::1) for general queries + */ + if (is_general_query && !ipv6_addr_is_ll_all_nodes(&ip6h->daddr)) { + err = -EINVAL; + goto out; + } + br_multicast_query_received(br, port, &br->ip6_querier, !ipv6_addr_any(&ip6h->saddr), max_delay); From 20a599bec95a52fa72432b2376a2ce47c5bb68fb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Linus=20L=C3=BCssing?= Date: Mon, 10 Mar 2014 22:25:25 +0100 Subject: [PATCH 154/259] bridge: multicast: enable snooping on general queries only MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Without this check someone could easily create a denial of service by injecting multicast-specific queries to enable the bridge snooping part if no real querier issuing periodic general queries is present on the link which would result in the bridge wrongly shutting down ports for multicast traffic as the bridge did not learn about these listeners. With this patch the snooping code is enabled upon receiving valid, general queries only. Signed-off-by: Linus Lüssing Signed-off-by: David S. Miller --- net/bridge/br_multicast.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index e56bae4f59ce..93067ecdb9a2 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -1127,9 +1127,10 @@ static void br_multicast_query_received(struct net_bridge *br, struct net_bridge_port *port, struct bridge_mcast_querier *querier, int saddr, + bool is_general_query, unsigned long max_delay) { - if (saddr) + if (saddr && is_general_query) br_multicast_update_querier_timer(br, querier, max_delay); else if (timer_pending(&querier->timer)) return; @@ -1190,7 +1191,7 @@ static int br_ip4_multicast_query(struct net_bridge *br, } br_multicast_query_received(br, port, &br->ip4_querier, !!iph->saddr, - max_delay); + !group, max_delay); if (!group) goto out; @@ -1282,7 +1283,8 @@ static int br_ip6_multicast_query(struct net_bridge *br, } br_multicast_query_received(br, port, &br->ip6_querier, - !ipv6_addr_any(&ip6h->saddr), max_delay); + !ipv6_addr_any(&ip6h->saddr), + is_general_query, max_delay); if (!group) goto out; From fdfaf64e75397567257e1051931f9a3377360665 Mon Sep 17 00:00:00 2001 From: Alexei Starovoitov Date: Mon, 10 Mar 2014 15:56:51 -0700 Subject: [PATCH 155/259] x86: bpf_jit: support negative offsets Commit a998d4342337 claimed to introduce negative offset support to x86 jit, but it couldn't be working, since at the time of the execution of LD+ABS or LD+IND instructions via call into bpf_internal_load_pointer_neg_helper() the %edx (3rd argument of this func) had junk value instead of access size in bytes (1 or 2 or 4). Store size into %edx instead of %ecx (what original commit intended to do) Fixes: a998d4342337 ("bpf jit: Let the x86 jit handle negative offsets") Signed-off-by: Alexei Starovoitov Cc: Jan Seiffert Cc: Eric Dumazet Acked-by: Eric Dumazet Signed-off-by: David S. Miller --- arch/x86/net/bpf_jit.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/net/bpf_jit.S b/arch/x86/net/bpf_jit.S index 877b9a1b2152..01495755701b 100644 --- a/arch/x86/net/bpf_jit.S +++ b/arch/x86/net/bpf_jit.S @@ -140,7 +140,7 @@ bpf_slow_path_byte_msh: push %r9; \ push SKBDATA; \ /* rsi already has offset */ \ - mov $SIZE,%ecx; /* size */ \ + mov $SIZE,%edx; /* size */ \ call bpf_internal_load_pointer_neg_helper; \ test %rax,%rax; \ pop SKBDATA; \ From 406764622fb71e5f4efb39ff111ef87713815f6d Mon Sep 17 00:00:00 2001 From: Markos Chandras Date: Tue, 11 Mar 2014 08:49:39 +0000 Subject: [PATCH 156/259] tools/net/Makefile: Define PACKAGE to fix build problems Fixes the following build problem with binutils-2.24 gcc -Wall -O2 -c -o bpf_jit_disasm.o bpf_jit_disasm.c In file included from bpf_jit_disasm.c:25:0: /usr/include/bfd.h:35:2: error: #error config.h must be included before this header #error config.h must be included before this header This is similar to commit 3ce711a6abc27abce1554e1d671a8762b7187690 "perf tools: bfd.h/libbfd detection fails with recent binutils" See: https://sourceware.org/bugzilla/show_bug.cgi?id=14243 CC: David S. Miller CC: Daniel Borkmann CC: netdev@vger.kernel.org Acked-by: Daniel Borkmann Signed-off-by: Markos Chandras Signed-off-by: David S. Miller --- tools/net/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/net/Makefile b/tools/net/Makefile index 004cd74734b6..ee577ea03ba5 100644 --- a/tools/net/Makefile +++ b/tools/net/Makefile @@ -12,7 +12,7 @@ YACC = bison all : bpf_jit_disasm bpf_dbg bpf_asm -bpf_jit_disasm : CFLAGS = -Wall -O2 +bpf_jit_disasm : CFLAGS = -Wall -O2 -DPACKAGE='bpf_jit_disasm' bpf_jit_disasm : LDLIBS = -lopcodes -lbfd -ldl bpf_jit_disasm : bpf_jit_disasm.o From f75761b6b5bf6277296505941d2dd8e11f9b5c35 Mon Sep 17 00:00:00 2001 From: hayeswang Date: Tue, 11 Mar 2014 15:11:59 +0800 Subject: [PATCH 157/259] r8169: fix the incorrect tx descriptor version The tx descriptor version of RTL8111B belong to RTL_TD_0. Signed-off-by: Hayes Wang Acked-by: Francois Romieu Signed-off-by: David S. Miller --- drivers/net/ethernet/realtek/r8169.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c index e9779653cd4c..3ff7bc3e7a23 100644 --- a/drivers/net/ethernet/realtek/r8169.c +++ b/drivers/net/ethernet/realtek/r8169.c @@ -209,7 +209,7 @@ static const struct { [RTL_GIGA_MAC_VER_16] = _R("RTL8101e", RTL_TD_0, NULL, JUMBO_1K, true), [RTL_GIGA_MAC_VER_17] = - _R("RTL8168b/8111b", RTL_TD_1, NULL, JUMBO_4K, false), + _R("RTL8168b/8111b", RTL_TD_0, NULL, JUMBO_4K, false), [RTL_GIGA_MAC_VER_18] = _R("RTL8168cp/8111cp", RTL_TD_1, NULL, JUMBO_6K, false), [RTL_GIGA_MAC_VER_19] = From b12bb60d6c350b348a4e1460cd68f97ccae9822e Mon Sep 17 00:00:00 2001 From: Ales Novak Date: Thu, 27 Feb 2014 11:03:30 +0100 Subject: [PATCH 158/259] [SCSI] storvsc: NULL pointer dereference fix If the initialization of storvsc fails, the storvsc_device_destroy() causes NULL pointer dereference. storvsc_bus_scan() scsi_scan_target() __scsi_scan_target() scsi_probe_and_add_lun(hostdata=NULL) scsi_alloc_sdev(hostdata=NULL) sdev->hostdata = hostdata now the host allocation fails __scsi_remove_device(sdev) calls sdev->host->hostt->slave_destroy() == storvsc_device_destroy(sdev) access of sdev->hostdata->request_mempool Signed-off-by: Ales Novak Signed-off-by: Thomas Abraham Reviewed-by: Jiri Kosina Acked-by: K. Y. Srinivasan Cc: stable@vger.kernel.org Signed-off-by: James Bottomley --- drivers/scsi/storvsc_drv.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c index 17d740427240..9969fa1ef7c4 100644 --- a/drivers/scsi/storvsc_drv.c +++ b/drivers/scsi/storvsc_drv.c @@ -1419,6 +1419,9 @@ static void storvsc_device_destroy(struct scsi_device *sdevice) { struct stor_mem_pools *memp = sdevice->hostdata; + if (!memp) + return; + mempool_destroy(memp->request_mempool); kmem_cache_destroy(memp->request_pool); kfree(memp); From 0e6d6ec02f867fe8d1f785312b7343b21052322f Mon Sep 17 00:00:00 2001 From: Thomas Hellstrom Date: Wed, 12 Mar 2014 10:41:32 +0100 Subject: [PATCH 159/259] drm/ttm: Work around performance regression with VM_PFNMAP A performance regression was introduced in TTM in linux 3.13 when we started using VM_PFNMAP for shared mappings. In theory this should've been faster due to less page book-keeping but it appears like VM_PFNMAP + x86 PAT + write-combine is a particularly cpu-hungry combination, as seen by largely increased cpu-usage on r200 GL video playback. Until we've sorted out why, revert to always use VM_MIXEDMAP. Reference: freedesktop.org bugzilla bug #75719 Reported-and-tested-by: Acked-by: Alex Deucher Signed-off-by: Thomas Hellstrom Cc: stable@vger.kernel.org --- drivers/gpu/drm/ttm/ttm_bo_vm.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index 801231c9ae48..0ce48e5a9cb4 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -339,11 +339,13 @@ int ttm_bo_mmap(struct file *filp, struct vm_area_struct *vma, vma->vm_private_data = bo; /* - * PFNMAP is faster than MIXEDMAP due to reduced page - * administration. So use MIXEDMAP only if private VMA, where - * we need to support COW. + * We'd like to use VM_PFNMAP on shared mappings, where + * (vma->vm_flags & VM_SHARED) != 0, for performance reasons, + * but for some reason VM_PFNMAP + x86 PAT + write-combine is very + * bad for performance. Until that has been sorted out, use + * VM_MIXEDMAP on all mappings. See freedesktop.org bug #75719 */ - vma->vm_flags |= (vma->vm_flags & VM_SHARED) ? VM_PFNMAP : VM_MIXEDMAP; + vma->vm_flags |= VM_MIXEDMAP; vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP; return 0; out_unref: @@ -359,7 +361,7 @@ int ttm_fbdev_mmap(struct vm_area_struct *vma, struct ttm_buffer_object *bo) vma->vm_ops = &ttm_bo_vm_ops; vma->vm_private_data = ttm_bo_reference(bo); - vma->vm_flags |= (vma->vm_flags & VM_SHARED) ? VM_PFNMAP : VM_MIXEDMAP; + vma->vm_flags |= VM_MIXEDMAP; vma->vm_flags |= VM_IO | VM_DONTEXPAND; return 0; } From 3cdeb713dc66057b50682048c151eae07b186c42 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 11 Mar 2014 14:22:19 -0600 Subject: [PATCH 160/259] PCI: Enable INTx in pci_reenable_device() only when MSI/MSI-X not enabled Andreas reported that after 1f42db786b14 ("PCI: Enable INTx if BIOS left them disabled"), pciehp surprise removal stopped working. This happens because pci_reenable_device() on the hotplug bridge (used in the pciehp_configure_device() path) clears the Interrupt Disable bit, which apparently breaks the bridge's MSI hotplug event reporting. Previously we cleared the Interrupt Disable bit in do_pci_enable_device(), which is used by both pci_enable_device() and pci_reenable_device(). But we use pci_reenable_device() after the driver may have enabled MSI or MSI-X, and we *set* Interrupt Disable as part of enabling MSI/MSI-X. This patch clears Interrupt Disable only when MSI/MSI-X has not been enabled. Fixes: 1f42db786b14 PCI: Enable INTx if BIOS left them disabled Link: https://bugzilla.kernel.org/show_bug.cgi?id=71691 Reported-and-tested-by: Andreas Noever Signed-off-by: Bjorn Helgaas CC: stable@vger.kernel.org CC: Sarah Sharp --- drivers/pci/pci.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 6b05f6134b68..fdbc294821e6 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1192,6 +1192,9 @@ static int do_pci_enable_device(struct pci_dev *dev, int bars) return err; pci_fixup_device(pci_fixup_enable, dev); + if (dev->msi_enabled || dev->msix_enabled) + return 0; + pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin); if (pin) { pci_read_config_word(dev, PCI_COMMAND, &cmd); From ac93ac7403493f8707b7734de9f40d5cb5db9045 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 11 Mar 2014 14:23:12 -0600 Subject: [PATCH 161/259] PCI: Don't check resource_size() in pci_bus_alloc_resource() Paul reported that after f75b99d5a77d ("PCI: Enforce bus address limits in resource allocation") on a 32-bit kernel (CONFIG_PHYS_ADDR_T_64BIT not set), intel-gtt complained "can't ioremap flush page - no chipset flushing". In addition, other PCI resource allocations, e.g., for bridge windows, failed. This happens because we incorrectly skip bus resources of [mem 0x00000000-0xffffffff] because we think they are of size zero. When resource_size_t is 32 bits wide, resource_size() on [mem 0x00000000-0xffffffff] returns 0 because (r->end - r->start + 1) overflows. Therefore, we can't use "resource_size() == 0" to decide that allocation from this resource will fail. allocate_resource() should fail anyway if it can't satisfy the address constraints, so we should just depend on that. A [mem 0x00000000-0xffffffff] bus resource is obviously not really valid, but we do fall back to it as a default when we don't have information about host bridge apertures. Link: https://bugzilla.kernel.org/show_bug.cgi?id=71611 Fixes: f75b99d5a77d PCI: Enforce bus address limits in resource allocation Reported-and-tested-by: Paul Bolle Signed-off-by: Bjorn Helgaas --- drivers/pci/bus.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/pci/bus.c b/drivers/pci/bus.c index 00660cc502c5..38901665c770 100644 --- a/drivers/pci/bus.c +++ b/drivers/pci/bus.c @@ -162,8 +162,6 @@ static int pci_bus_alloc_from_region(struct pci_bus *bus, struct resource *res, avail = *r; pci_clip_resource_to_region(bus, &avail, region); - if (!resource_size(&avail)) - continue; /* * "min" is typically PCIBIOS_MIN_IO or PCIBIOS_MIN_MEM to From 596f3142d2b7be307a1652d59e7b93adab918437 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Radim=20Kr=C4=8Dm=C3=A1=C5=99?= Date: Tue, 11 Mar 2014 19:11:18 +0100 Subject: [PATCH 162/259] KVM: SVM: fix cr8 intercept window MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We always disable cr8 intercept in its handler, but only re-enable it if handling KVM_REQ_EVENT, so there can be a window where we do not intercept cr8 writes, which allows an interrupt to disrupt a higher priority task. Fix this by disabling intercepts in the same function that re-enables them when needed. This fixes BSOD in Windows 2008. Cc: Signed-off-by: Radim Krčmář Reviewed-by: Marcelo Tosatti Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index e81df8fce027..2de1bc09a8d4 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -3002,10 +3002,8 @@ static int cr8_write_interception(struct vcpu_svm *svm) u8 cr8_prev = kvm_get_cr8(&svm->vcpu); /* instruction emulation calls kvm_set_cr8() */ r = cr_interception(svm); - if (irqchip_in_kernel(svm->vcpu.kvm)) { - clr_cr_intercept(svm, INTERCEPT_CR8_WRITE); + if (irqchip_in_kernel(svm->vcpu.kvm)) return r; - } if (cr8_prev <= kvm_get_cr8(&svm->vcpu)) return r; kvm_run->exit_reason = KVM_EXIT_SET_TPR; @@ -3567,6 +3565,8 @@ static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr) if (is_guest_mode(vcpu) && (vcpu->arch.hflags & HF_VINTR_MASK)) return; + clr_cr_intercept(svm, INTERCEPT_CR8_WRITE); + if (irr == -1) return; From 8b9d96666529a979acf4825391efcc7c8a3e9f12 Mon Sep 17 00:00:00 2001 From: Heinz Mauelshagen Date: Wed, 12 Mar 2014 00:40:05 +0100 Subject: [PATCH 163/259] dm cache: fix truncation bug when copying a block to/from >2TB fast device During demotion or promotion to a cache's >2TB fast device we must not truncate the cache block's associated sector to 32bits. The 32bit temporary result of from_cblock() caused a 32bit multiplication when calculating the sector of the fast device in issue_copy_real(). Use an intermediate 64bit type to store the 32bit from_cblock() to allow for proper 64bit multiplication. Here is an example of how this bug manifests on an ext4 filesystem: EXT4-fs error (device dm-0): ext4_mb_generate_buddy:756: group 17136, 32768 clusters in bitmap, 30688 in gd; block bitmap corrupt. JBD2: Spotted dirty metadata buffer (dev = dm-0, blocknr = 0). There's a risk of filesystem corruption in case of system crash. Signed-off-by: Heinz Mauelshagen Acked-by: Joe Thornber Signed-off-by: Mike Snitzer Cc: stable@vger.kernel.org --- drivers/md/dm-cache-target.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index 1af70145fab9..354bbc1b9a3c 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -979,12 +979,13 @@ static void issue_copy_real(struct dm_cache_migration *mg) int r; struct dm_io_region o_region, c_region; struct cache *cache = mg->cache; + sector_t cblock = from_cblock(mg->cblock); o_region.bdev = cache->origin_dev->bdev; o_region.count = cache->sectors_per_block; c_region.bdev = cache->cache_dev->bdev; - c_region.sector = from_cblock(mg->cblock) * cache->sectors_per_block; + c_region.sector = cblock * cache->sectors_per_block; c_region.count = cache->sectors_per_block; if (mg->writeback || mg->demote) { From e893fba90c09f9b57fb97daae204ea9cc2c52fa5 Mon Sep 17 00:00:00 2001 From: Heinz Mauelshagen Date: Wed, 12 Mar 2014 16:13:39 +0100 Subject: [PATCH 164/259] dm cache: fix access beyond end of origin device In order to avoid wasting cache space a partial block at the end of the origin device is not cached. Unfortunately, the check for such a partial block at the end of the origin device was flawed. Fix accesses beyond the end of the origin device that occured due to attempted promotion of an undetected partial block by: - initializing the per bio data struct to allow cache_end_io to work properly - recognizing access to the partial block at the end of the origin device - avoiding out of bounds access to the discard bitset Otherwise, users can experience errors like the following: attempt to access beyond end of device dm-5: rw=0, want=20971520, limit=20971456 ... device-mapper: cache: promotion failed; couldn't copy block Signed-off-by: Heinz Mauelshagen Acked-by: Joe Thornber Signed-off-by: Mike Snitzer Cc: stable@vger.kernel.org --- drivers/md/dm-cache-target.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index 354bbc1b9a3c..074b9c8e4cf0 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -2465,20 +2465,18 @@ static int cache_map(struct dm_target *ti, struct bio *bio) bool discarded_block; struct dm_bio_prison_cell *cell; struct policy_result lookup_result; - struct per_bio_data *pb; + struct per_bio_data *pb = init_per_bio_data(bio, pb_data_size); - if (from_oblock(block) > from_oblock(cache->origin_blocks)) { + if (unlikely(from_oblock(block) >= from_oblock(cache->origin_blocks))) { /* * This can only occur if the io goes to a partial block at * the end of the origin device. We don't cache these. * Just remap to the origin and carry on. */ - remap_to_origin_clear_discard(cache, bio, block); + remap_to_origin(cache, bio); return DM_MAPIO_REMAPPED; } - pb = init_per_bio_data(bio, pb_data_size); - if (bio->bi_rw & (REQ_FLUSH | REQ_FUA | REQ_DISCARD)) { defer_bio(cache, bio); return DM_MAPIO_SUBMITTED; From 9ef7506f7eff3fc42724269f62e30164c141661f Mon Sep 17 00:00:00 2001 From: Rob Clark Date: Wed, 12 Mar 2014 10:59:37 -0400 Subject: [PATCH 165/259] drm/ttm: don't oops if no invalidate_caches() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit A few of the simpler TTM drivers (cirrus, ast, mgag200) do not implement this function. Yet can end up somehow with an evicted bo: BUG: unable to handle kernel NULL pointer dereference at (null) IP: [< (null)>] (null) PGD 16e761067 PUD 16e6cf067 PMD 0 Oops: 0010 [#1] SMP Modules linked in: bnep bluetooth rfkill fuse ip6t_rpfilter ip6t_REJECT ipt_REJECT xt_conntrack ebtable_nat ebtable_broute bridge stp llc ebtable_filter ebtables ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw ip6table_filter ip6_tables iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw iptable_filter ip_tables sg btrfs zlib_deflate raid6_pq xor dm_queue_length iTCO_wdt iTCO_vendor_support coretemp kvm dcdbas dm_service_time microcode serio_raw pcspkr lpc_ich mfd_core i7core_edac edac_core ses enclosure ipmi_si ipmi_msghandler shpchp acpi_power_meter mperf nfsd auth_rpcgss nfs_acl lockd uinput sunrpc dm_multipath xfs libcrc32c ata_generic pata_acpi sr_mod cdrom sd_mod usb_storage mgag200 syscopyarea sysfillrect sysimgblt i2c_algo_bit lpfc drm_kms_helper ttm crc32c_intel ata_piix bfa drm ixgbe libata i2c_core mdio crc_t10dif ptp crct10dif_common pps_core scsi_transport_fc dca scsi_tgt megaraid_sas bnx2 dm_mirror dm_region_hash dm_log dm_mod CPU: 16 PID: 2572 Comm: X Not tainted 3.10.0-86.el7.x86_64 #1 Hardware name: Dell Inc. PowerEdge R810/0H235N, BIOS 0.3.0 11/14/2009 task: ffff8801799dabc0 ti: ffff88016c884000 task.ti: ffff88016c884000 RIP: 0010:[<0000000000000000>] [< (null)>] (null) RSP: 0018:ffff88016c885ad8 EFLAGS: 00010202 RAX: ffffffffa04e94c0 RBX: ffff880178937a20 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000240004 RDI: ffff880178937a00 RBP: ffff88016c885b60 R08: 00000000000171a0 R09: ffff88007cf171a0 R10: ffffea0005842540 R11: ffffffff810487b9 R12: ffff880178937b30 R13: ffff880178937a00 R14: ffff88016c885b78 R15: ffff880179929400 FS: 00007f81ba2ef980(0000) GS:ffff88007cf00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000016e763000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: ffffffffa0306fae ffff8801799295c0 0000000000260004 0000000000000001 ffff88016c885b60 ffffffffa0307669 00ff88007cf17738 ffff88017cf17700 ffff880178937a00 ffff880100000000 ffff880100000000 0000000079929400 Call Trace: [] ? ttm_bo_handle_move_mem+0x54e/0x5b0 [ttm] [] ? ttm_bo_mem_space+0x169/0x340 [ttm] [] ttm_bo_move_buffer+0x117/0x130 [ttm] [] ? perf_event_init_context+0x141/0x220 [] ttm_bo_validate+0xc1/0x130 [ttm] [] mgag200_bo_pin+0x87/0xc0 [mgag200] [] mga_crtc_cursor_set+0x474/0xbb0 [mgag200] [] ? __mem_cgroup_commit_charge+0x152/0x3b0 [] ? mutex_lock+0x12/0x2f [] drm_mode_cursor_common+0x123/0x170 [drm] [] drm_mode_cursor_ioctl+0x41/0x50 [drm] [] drm_ioctl+0x502/0x630 [drm] [] ? __do_page_fault+0x1f4/0x510 [] ? __restore_xstate_sig+0x218/0x4f0 [] do_vfs_ioctl+0x2e5/0x4d0 [] ? file_has_perm+0x8e/0xa0 [] SyS_ioctl+0x81/0xa0 [] system_call_fastpath+0x16/0x1b Code: Bad RIP value. RIP [< (null)>] (null) RSP CR2: 0000000000000000 Signed-off-by: Rob Clark Reviewed-by: Jérôme Glisse Reviewed-by: Thomas Hellstrom Cc: stable@vger.kernel.org --- drivers/gpu/drm/ttm/ttm_bo.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index a06651309388..214b7992a3aa 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -351,9 +351,11 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, moved: if (bo->evicted) { - ret = bdev->driver->invalidate_caches(bdev, bo->mem.placement); - if (ret) - pr_err("Can not flush read caches\n"); + if (bdev->driver->invalidate_caches) { + ret = bdev->driver->invalidate_caches(bdev, bo->mem.placement); + if (ret) + pr_err("Can not flush read caches\n"); + } bo->evicted = false; } From 836fbaf459f9c041826864021688e9bd131e722c Mon Sep 17 00:00:00 2001 From: Wei Liu Date: Tue, 11 Mar 2014 12:45:32 +0000 Subject: [PATCH 166/259] xen-netback: use skb_is_gso in xenvif_start_xmit In 5bd076708 ("Xen-netback: Fix issue caused by using gso_type wrongly") we use skb_is_gso to determine if we need an extra slot to accommodate the SKB. There's similar error in interface.c. Change that to use skb_is_gso as well. Signed-off-by: Wei Liu Cc: Annie Li Cc: Ian Campbell Cc: Paul Durrant Acked-by: Ian Campbell Signed-off-by: David S. Miller --- drivers/net/xen-netback/interface.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c index 7669d49a67e2..301cc037fda8 100644 --- a/drivers/net/xen-netback/interface.c +++ b/drivers/net/xen-netback/interface.c @@ -132,8 +132,7 @@ static int xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev) /* If the skb is GSO then we'll also need an extra slot for the * metadata. */ - if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4 || - skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) + if (skb_is_gso(skb)) min_slots_needed++; /* If the skb can't possibly fit in the remaining slots From 7848865914c6a63ead674f0f5604b77df7d3874f Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 11 Mar 2014 15:02:30 -0400 Subject: [PATCH 167/259] drm/radeon: fix runpm disabling on non-PX harder Make sure runtime pm is disabled on non-PX hardware. Should fix powerdown problems without displays attached. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/radeon_kms.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 2aecd6dc2610..66ed3ea71440 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -33,6 +33,13 @@ #include #include #include + +#if defined(CONFIG_VGA_SWITCHEROO) +bool radeon_is_px(void); +#else +static inline bool radeon_is_px(void) { return false; } +#endif + /** * radeon_driver_unload_kms - Main unload function for KMS. * @@ -130,7 +137,8 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) "Error during ACPI methods call\n"); } - if (radeon_runtime_pm != 0) { + if ((radeon_runtime_pm == 1) || + ((radeon_runtime_pm == -1) && radeon_is_px())) { pm_runtime_use_autosuspend(dev->dev); pm_runtime_set_autosuspend_delay(dev->dev, 5000); pm_runtime_set_active(dev->dev); From 7b1bbe883b3ed962ca2be4daf321f318f5091340 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 12 Mar 2014 15:15:58 -0400 Subject: [PATCH 168/259] drm/radeon/cik: properly set sdma ring status on disable When we disable the rings, set the status properly. If not other code pathes may try and use the rings which are not functional at this point. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/cik_sdma.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index 1ecb3f1070e3..5d7a5fa3741a 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -264,6 +264,8 @@ static void cik_sdma_gfx_stop(struct radeon_device *rdev) WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl); WREG32(SDMA0_GFX_IB_CNTL + reg_offset, 0); } + rdev->ring[R600_RING_TYPE_DMA_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX].ready = false; } /** From 07ae78c9798b79bad3d3adf983c94ba23fde54d4 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 12 Mar 2014 15:26:34 -0400 Subject: [PATCH 169/259] drm/radeon/cik: stop the sdma engines in the enable() function We always stop the rings when disabling the engines so just call the stop functions directly from the sdma enable function. This way the rings' status is set correctly on suspend so there are no problems on resume. Fixes resume failures that result in acceleration getting disabled. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/cik_sdma.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index 5d7a5fa3741a..94626ea90fa5 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -293,6 +293,11 @@ void cik_sdma_enable(struct radeon_device *rdev, bool enable) u32 me_cntl, reg_offset; int i; + if (enable == false) { + cik_sdma_gfx_stop(rdev); + cik_sdma_rlc_stop(rdev); + } + for (i = 0; i < 2; i++) { if (i == 0) reg_offset = SDMA0_REGISTER_OFFSET; @@ -422,10 +427,6 @@ static int cik_sdma_load_microcode(struct radeon_device *rdev) if (!rdev->sdma_fw) return -EINVAL; - /* stop the gfx rings and rlc compute queues */ - cik_sdma_gfx_stop(rdev); - cik_sdma_rlc_stop(rdev); - /* halt the MEs */ cik_sdma_enable(rdev, false); @@ -494,9 +495,6 @@ int cik_sdma_resume(struct radeon_device *rdev) */ void cik_sdma_fini(struct radeon_device *rdev) { - /* stop the gfx rings and rlc compute queues */ - cik_sdma_gfx_stop(rdev); - cik_sdma_rlc_stop(rdev); /* halt the MEs */ cik_sdma_enable(rdev, false); radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]); From 56cb456746a15c1025a178466492ca4c373b1a63 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 12 Mar 2014 17:16:30 +0200 Subject: [PATCH 170/259] net/mlx4_core: Fix wrong dump of the vxlan offloads device capability Fix the value used to dump the vxlan offloads device capability to align with the MLX4_DEV_CAP_FLAG2_yyy definition. While on that, add dump to the IPoIB flow-steering device capability and fix small typo. The vxlan cap value wasn't fully handled when a conflict was resolved between MLX4_DEV_CAP_FLAG2_DMFS_IPOIB coming from the IB tree to MLX4_DEV_CAP_FLAG2_VXLAN_OFFLOADS coming from net-next. Signed-off-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/fw.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c index 8726e34cee22..7e2995ecea6f 100644 --- a/drivers/net/ethernet/mellanox/mlx4/fw.c +++ b/drivers/net/ethernet/mellanox/mlx4/fw.c @@ -129,13 +129,14 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags) [0] = "RSS support", [1] = "RSS Toeplitz Hash Function support", [2] = "RSS XOR Hash Function support", - [3] = "Device manage flow steering support", + [3] = "Device managed flow steering support", [4] = "Automatic MAC reassignment support", [5] = "Time stamping support", [6] = "VST (control vlan insertion/stripping) support", [7] = "FSM (MAC anti-spoofing) support", [8] = "Dynamic QP updates support", - [9] = "TCP/IP offloads/flow-steering for VXLAN support" + [9] = "Device managed flow steering IPoIB support", + [10] = "TCP/IP offloads/flow-steering for VXLAN support" }; int i; From 2a2083f7f3568c0192daa6ac0e6fa35d953f47bd Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 12 Mar 2014 17:16:31 +0200 Subject: [PATCH 171/259] net/mlx4_en: Handle vxlan steering rules for mac address changes When the device mac address is changed, we must deregister the vxlan steering rule associated with the previous mac, and register a new steering rule using the new mac. Signed-off-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index fad45316200a..ba2126111573 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -742,6 +742,14 @@ static int mlx4_en_replace_mac(struct mlx4_en_priv *priv, int qpn, err = mlx4_en_uc_steer_add(priv, new_mac, &qpn, &entry->reg_id); + if (err) + return err; + if (priv->tunnel_reg_id) { + mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id); + priv->tunnel_reg_id = 0; + } + err = mlx4_en_tunnel_steer_add(priv, new_mac, qpn, + &priv->tunnel_reg_id); return err; } } From 7855bff42ea9938a0853321256f4c8ce3628aa73 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 12 Mar 2014 17:16:32 +0200 Subject: [PATCH 172/259] net/mlx4_core: Load the IB driver when the device supports IBoE When checking what protocol drivers to load, the IB driver should be requested also over Ethernet ports, if the device supports IBoE (RoCE). Fixes: b046ffe 'net/mlx4_core: Load higher level modules according to ports type' Signed-off-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c index 30a08a60f059..936c15364739 100644 --- a/drivers/net/ethernet/mellanox/mlx4/main.c +++ b/drivers/net/ethernet/mellanox/mlx4/main.c @@ -751,7 +751,7 @@ static void mlx4_request_modules(struct mlx4_dev *dev) has_eth_port = true; } - if (has_ib_port) + if (has_ib_port || (dev->caps.flags & MLX4_DEV_CAP_FLAG_IBOE)) request_module_nowait(IB_DRV_NAME); if (has_eth_port) request_module_nowait(EN_DRV_NAME); From a93c12560498066765c58b032530798a26bd8f70 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Tue, 11 Mar 2014 11:23:42 +0100 Subject: [PATCH 173/259] packet: doc: Spelling s/than/that/ Signed-off-by: Geert Uytterhoeven Cc: David S. Miller Cc: netdev@vger.kernel.org Signed-off-by: David S. Miller --- Documentation/networking/packet_mmap.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/networking/packet_mmap.txt b/Documentation/networking/packet_mmap.txt index 1404674c0a02..6fea79efb4cb 100644 --- a/Documentation/networking/packet_mmap.txt +++ b/Documentation/networking/packet_mmap.txt @@ -453,7 +453,7 @@ TP_STATUS_COPY : This flag indicates that the frame (and associated enabled previously with setsockopt() and the PACKET_COPY_THRESH option. - The number of frames than can be buffered to + The number of frames that can be buffered to be read with recvfrom is limited like a normal socket. See the SO_RCVBUF option in the socket (7) man page. From 7e814a6c50a1669118ffa4961568fea3aa955faf Mon Sep 17 00:00:00 2001 From: Tobias Klauser Date: Tue, 11 Mar 2014 12:06:24 +0100 Subject: [PATCH 174/259] MAINTAINERS: Add tools/net to NETWORKING [GENERAL] Make sure patches for these tools go to the netdev list as well. References: https://marc.info/?l=linux-kernel&m=139450284501328&w=2 Cc: David S. Miller Cc: Daniel Borkmann Signed-off-by: Tobias Klauser Signed-off-by: David S. Miller --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index e1297ff255e1..e1be77c78336 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -6002,6 +6002,7 @@ F: include/linux/netdevice.h F: include/uapi/linux/in.h F: include/uapi/linux/net.h F: include/uapi/linux/netdevice.h +F: tools/net/ NETWORKING [IPv4/IPv6] M: "David S. Miller" From b2b3d8d952e4f8d6ac2ce80be96b937f29f6e42e Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 12 Mar 2014 16:20:44 -0400 Subject: [PATCH 175/259] drm/radeon/cik: properly set compute ring status on disable When we disable the rings, set the status properly. If not other code pathes may try and use the rings which are not functional at this point. Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/cik.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index e22be8458d92..bbb17841a9e5 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -4134,8 +4134,11 @@ static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable) { if (enable) WREG32(CP_MEC_CNTL, 0); - else + else { WREG32(CP_MEC_CNTL, (MEC_ME1_HALT | MEC_ME2_HALT)); + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; + } udelay(50); } From dbb490b96584d4e958533fb637f08b557f505657 Mon Sep 17 00:00:00 2001 From: Matthew Leach Date: Tue, 11 Mar 2014 11:58:27 +0000 Subject: [PATCH 176/259] net: socket: error on a negative msg_namelen When copying in a struct msghdr from the user, if the user has set the msg_namelen parameter to a negative value it gets clamped to a valid size due to a comparison between signed and unsigned values. Ensure the syscall errors when the user passes in a negative value. Signed-off-by: Matthew Leach Signed-off-by: David S. Miller --- net/socket.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/socket.c b/net/socket.c index 879933aaed4c..32df5847efa3 100644 --- a/net/socket.c +++ b/net/socket.c @@ -1985,6 +1985,10 @@ static int copy_msghdr_from_user(struct msghdr *kmsg, { if (copy_from_user(kmsg, umsg, sizeof(struct msghdr))) return -EFAULT; + + if (kmsg->msg_namelen < 0) + return -EINVAL; + if (kmsg->msg_namelen > sizeof(struct sockaddr_storage)) kmsg->msg_namelen = sizeof(struct sockaddr_storage); return 0; From 2ed99e39cb9392312c100d9da591c20641c64d12 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Wed, 12 Mar 2014 21:49:33 +0100 Subject: [PATCH 177/259] cpufreq: Skip current frequency initialization for ->setpolicy drivers After commit da60ce9f2fac (cpufreq: call cpufreq_driver->get() after calling ->init()) __cpufreq_add_dev() sometimes fails for CPUs handled by intel_pstate, because that driver may return 0 from its ->get() callback if it has not run long enough to collect enough samples on the given CPU. That didn't happen before commit da60ce9f2fac which added policy->cur initialization to __cpufreq_add_dev() to help reduce code duplication in other cpufreq drivers. However, the code added by commit da60ce9f2fac need not be executed for cpufreq drivers having the ->setpolicy callback defined, because the subsequent invocation of cpufreq_set_policy() will use that callback to initialize the policy anyway and it doesn't need policy->cur to be initialized upfront. The analogous code in cpufreq_update_policy() is also unnecessary for cpufreq drivers having ->setpolicy set and may be skipped for them as well. Since intel_pstate provides ->setpolicy, skipping the upfront policy->cur initialization for cpufreq drivers with that callback set will cover intel_pstate and the problem it's been having after commit da60ce9f2fac will be addressed. Fixes: da60ce9f2fac (cpufreq: call cpufreq_driver->get() after calling ->init()) References: https://bugzilla.kernel.org/show_bug.cgi?id=71931 Reported-and-tested-by: Patrik Lundquist Acked-by: Dirk Brandewie Cc: 3.13+ # 3.13+ Signed-off-by: Rafael J. Wysocki --- drivers/cpufreq/cpufreq.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index cf485d928903..199b52b7c3e1 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -1129,7 +1129,7 @@ static int __cpufreq_add_dev(struct device *dev, struct subsys_interface *sif, per_cpu(cpufreq_cpu_data, j) = policy; write_unlock_irqrestore(&cpufreq_driver_lock, flags); - if (cpufreq_driver->get) { + if (cpufreq_driver->get && !cpufreq_driver->setpolicy) { policy->cur = cpufreq_driver->get(policy->cpu); if (!policy->cur) { pr_err("%s: ->get() failed\n", __func__); @@ -2143,7 +2143,7 @@ int cpufreq_update_policy(unsigned int cpu) * BIOS might change freq behind our back * -> ask driver for current freq and notify governors about a change */ - if (cpufreq_driver->get) { + if (cpufreq_driver->get && !cpufreq_driver->setpolicy) { new_policy.cur = cpufreq_driver->get(cpu); if (!policy->cur) { pr_debug("Driver did not initialize current freq"); From c4e1acbb35e4a3838cdfc0e7f8237e844aff00b6 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 13 Mar 2014 00:22:58 +0100 Subject: [PATCH 178/259] ACPI / init: Invoke early ACPI initialization later Commit 73f7d1ca3263 (ACPI / init: Run acpi_early_init() before timekeeping_init()) optimistically moved the early ACPI initialization before timekeeping_init(), but that didn't work, because it broke fast TSC calibration for Julian Wollrath on Thinkpad x121e (and most likely for others too). The reason is that acpi_early_init() enables the SCI and that interferes with the fast TSC calibration mechanism. Thus follow the original idea to execute acpi_early_init() before efi_enter_virtual_mode() to help the EFI people for now and we can revisit the other problem that commit 73f7d1ca3263 attempted to address in the future (if really necessary). Fixes: 73f7d1ca3263 (ACPI / init: Run acpi_early_init() before timekeeping_init()) Reported-by: Julian Wollrath Reviewed-by: Thomas Gleixner Signed-off-by: Rafael J. Wysocki --- init/main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/init/main.c b/init/main.c index eb03090cdced..9c7fd4c9249f 100644 --- a/init/main.c +++ b/init/main.c @@ -561,7 +561,6 @@ asmlinkage void __init start_kernel(void) init_timers(); hrtimers_init(); softirq_init(); - acpi_early_init(); timekeeping_init(); time_init(); sched_clock_postinit(); @@ -613,6 +612,7 @@ asmlinkage void __init start_kernel(void) calibrate_delay(); pidmap_init(); anon_vma_init(); + acpi_early_init(); #ifdef CONFIG_X86 if (efi_enabled(EFI_RUNTIME_SERVICES)) efi_enter_virtual_mode(); From 4b0c82529b92cda4723cf1d09e1e2ee9b9ae96f1 Mon Sep 17 00:00:00 2001 From: Thomas Hellstrom Date: Wed, 12 Mar 2014 20:42:31 +0100 Subject: [PATCH 179/259] drm/vmwgfx: Fix a surface reference corner-case in legacy emulation mode If running on a gb-object capable device with a non-gb capable surface exporter (X server) and a gb capable surface referencing client (GL driver), the referencing client expects to find a shareable backing buffer attached to the surface at reference time. This may not be the case if the surface has not yet been validated. This would cause the surface reference IOCTL to return an error. Signed-off-by: Thomas Hellstrom Reviewed-by: Jakob Bornecrantz --- drivers/gpu/drm/vmwgfx/vmwgfx_surface.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c index 82468d902915..e7af580ab977 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c @@ -830,6 +830,24 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, if (unlikely(ret != 0)) goto out_unlock; + /* + * A gb-aware client referencing a shared surface will + * expect a backup buffer to be present. + */ + if (dev_priv->has_mob && req->shareable) { + uint32_t backup_handle; + + ret = vmw_user_dmabuf_alloc(dev_priv, tfile, + res->backup_size, + true, + &backup_handle, + &res->backup); + if (unlikely(ret != 0)) { + vmw_resource_unreference(&res); + goto out_unlock; + } + } + tmp = vmw_resource_reference(&srf->res); ret = ttm_prime_object_init(tfile, res->backup_size, &user_srf->prime, req->shareable, VMW_RES_SURFACE, From 065a5027dca8e9383ac308de4310e8e850b0cafb Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 21 Jan 2014 12:01:41 +0100 Subject: [PATCH 180/259] drm/doc: Clarify the dumb object interfaces MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - This is _not_ a generic interface to create gem objects, but just an interface to make early boot services (like boot splash) with a generic KMS userspace driver possible. Hence it's better to move the documentation for this from the GEM section to the KMS section, next to the creation of framebuffer objects. - Make it really clear that the returned handle isn't necessarily a GEM object (it can also be e.g. a TTM handle when running on top of vmwgfx). - Add a paragraph to make it clear that this is just for unaccelarated userspace - gpu drivers need to have their own buffer object creation ioctl which is hardware specific. v2: Clarify that the documentation doesn't just apply to GEM-based drivers only but is now generally valid, as suggested by David. v3: Polish the intro sentence a bit and one s/objects/handles/ for clarification, both suggested by Laurent. v4: More text polish from Laurent's review. v5: More typo fixes from Dieter. Cc: Dieter Nützel Cc: David Herrmann Cc: Laurent Pinchart Acked-by: Laurent Pinchart Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 133 ++++++++++++++++++--------------- 1 file changed, 72 insertions(+), 61 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index ed1d6d289022..f2d0f5b89194 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -829,62 +829,6 @@ char *date; faults can implement their own mmap file operation handler. - - Dumb GEM Objects - - The GEM API doesn't standardize GEM objects creation and leaves it to - driver-specific ioctls. While not an issue for full-fledged graphics - stacks that include device-specific userspace components (in libdrm for - instance), this limit makes DRM-based early boot graphics unnecessarily - complex. - - - Dumb GEM objects partly alleviate the problem by providing a standard - API to create dumb buffers suitable for scanout, which can then be used - to create KMS frame buffers. - - - To support dumb GEM objects drivers must implement the - dumb_create, - dumb_destroy and - dumb_map_offset operations. - - - - int (*dumb_create)(struct drm_file *file_priv, struct drm_device *dev, - struct drm_mode_create_dumb *args); - - The dumb_create operation creates a GEM - object suitable for scanout based on the width, height and depth - from the struct drm_mode_create_dumb - argument. It fills the argument's handle, - pitch and size - fields with a handle for the newly created GEM object and its line - pitch and size in bytes. - - - - int (*dumb_destroy)(struct drm_file *file_priv, struct drm_device *dev, - uint32_t handle); - - The dumb_destroy operation destroys a dumb - GEM object created by dumb_create. - - - - int (*dumb_map_offset)(struct drm_file *file_priv, struct drm_device *dev, - uint32_t handle, uint64_t *offset); - - The dumb_map_offset operation associates an - mmap fake offset with the GEM object given by the handle and returns - it. Drivers must use the - drm_gem_create_mmap_offset function to - associate the fake offset as described in - . - - - - Memory Coherency @@ -968,9 +912,11 @@ int max_width, max_height; Frame buffers rely on the underneath memory manager for low-level memory operations. When creating a frame buffer applications pass a memory handle (or a list of memory handles for multi-planar formats) through - the drm_mode_fb_cmd2 argument. This document - assumes that the driver uses GEM, those handles thus reference GEM - objects. + the drm_mode_fb_cmd2 argument. For drivers using + GEM as their userspace buffer management interface this would be a GEM + handle. Drivers are however free to use their own backing storage object + handles, e.g. vmwgfx directly exposes special TTM handles to userspace + and so expects TTM handles in the create ioctl and not GEM handles. Drivers must first validate the requested frame buffer parameters passed @@ -992,7 +938,7 @@ int max_width, max_height; - The initailization of the new framebuffer instance is finalized with a + The initialization of the new framebuffer instance is finalized with a call to drm_framebuffer_init which takes a pointer to DRM frame buffer operations (struct drm_framebuffer_funcs). Note that this function @@ -1051,6 +997,71 @@ int max_width, max_height; unload time with drm_framebuffer_unregister_private. + + Dumb Buffer Objects + + The KMS API doesn't standardize backing storage object creation and + leaves it to driver-specific ioctls. Furthermore actually creating a + buffer object even for GEM-based drivers is done through a + driver-specific ioctl - GEM only has a common userspace interface for + sharing and destroying objects. While not an issue for full-fledged + graphics stacks that include device-specific userspace components (in + libdrm for instance), this limit makes DRM-based early boot graphics + unnecessarily complex. + + + Dumb objects partly alleviate the problem by providing a standard + API to create dumb buffers suitable for scanout, which can then be used + to create KMS frame buffers. + + + To support dumb objects drivers must implement the + dumb_create, + dumb_destroy and + dumb_map_offset operations. + + + + int (*dumb_create)(struct drm_file *file_priv, struct drm_device *dev, + struct drm_mode_create_dumb *args); + + The dumb_create operation creates a driver + object (GEM or TTM handle) suitable for scanout based on the + width, height and depth from the struct + drm_mode_create_dumb argument. It fills the + argument's handle, + pitch and size + fields with a handle for the newly created object and its line + pitch and size in bytes. + + + + int (*dumb_destroy)(struct drm_file *file_priv, struct drm_device *dev, + uint32_t handle); + + The dumb_destroy operation destroys a dumb + object created by dumb_create. + + + + int (*dumb_map_offset)(struct drm_file *file_priv, struct drm_device *dev, + uint32_t handle, uint64_t *offset); + + The dumb_map_offset operation associates an + mmap fake offset with the object given by the handle and returns + it. Drivers must use the + drm_gem_create_mmap_offset function to + associate the fake offset as described in + . + + + + + Note that dumb objects may not be used for gpu acceleration, as has been + attempted on some ARM embedded platforms. Such drivers really must have + a hardware-specific ioctl to allocate suitable buffer objects. + + Output Polling void (*output_poll_changed)(struct drm_device *dev); @@ -2134,7 +2145,7 @@ void intel_crt_init(struct drm_device *dev) set the display_info width_mm and height_mm fields if they haven't been set - already (for instance at initilization time when a fixed-size panel is + already (for instance at initialization time when a fixed-size panel is attached to the connector). The mode width_mm and height_mm fields are only used internally during EDID parsing and should not be set when creating modes manually. From fc66811ce1de3f79cbe3f83c2fdb1ef3e2cdac1c Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 21 Jan 2014 12:02:26 +0100 Subject: [PATCH 181/259] drm/doc: Fix up kerneldoc in drm_edid.c v2: Also do s/RETURNS/Returns/, less yelling in docs is always good. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_edid.c | 26 +++++++++++++++++++------- 1 file changed, 19 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index b924306b8477..2d54e460c95b 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -1098,10 +1098,14 @@ EXPORT_SYMBOL(drm_edid_is_valid); /** * Get EDID information via I2C. * - * \param adapter : i2c device adaptor - * \param buf : EDID data buffer to be filled - * \param len : EDID data buffer length - * \return 0 on success or -1 on failure. + * @adapter : i2c device adaptor + * @buf: EDID data buffer to be filled + * @block: 128 byte EDID block to start fetching from + * @len: EDID data buffer length to fetch + * + * Returns: + * + * 0 on success or -1 on failure. * * Try to fetch EDID information by calling i2c driver function. */ @@ -1243,9 +1247,11 @@ drm_do_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter) /** * Probe DDC presence. + * @adapter: i2c adapter to probe * - * \param adapter : i2c device adaptor - * \return 1 on success + * Returns: + * + * 1 on success */ bool drm_probe_ddc(struct i2c_adapter *adapter) @@ -1586,8 +1592,10 @@ bad_std_timing(u8 a, u8 b) /** * drm_mode_std - convert standard mode info (width, height, refresh) into mode + * @connector: connector of for the EDID block + * @edid: EDID block to scan * @t: standard timing params - * @timing_level: standard timing level + * @revision: standard timing level * * Take the standard timing params (in this case width, aspect, and refresh) * and convert them into a real mode using CVT/GTF/DMT. @@ -2132,6 +2140,7 @@ do_established_modes(struct detailed_timing *timing, void *c) /** * add_established_modes - get est. modes from EDID and add them + * @connector: connector of for the EDID block * @edid: EDID block to scan * * Each EDID block contains a bitmap of the supported "established modes" list @@ -2194,6 +2203,7 @@ do_standard_modes(struct detailed_timing *timing, void *c) /** * add_standard_modes - get std. modes from EDID and add them + * @connector: connector of for the EDID block * @edid: EDID block to scan * * Standard modes can be calculated using the appropriate standard (DMT, @@ -3300,6 +3310,7 @@ EXPORT_SYMBOL(drm_detect_hdmi_monitor); /** * drm_detect_monitor_audio - check monitor audio capability + * @edid: EDID block to scan * * Monitor should have CEA extension block. * If monitor has 'basic audio', but no CEA audio blocks, it's 'basic @@ -3345,6 +3356,7 @@ EXPORT_SYMBOL(drm_detect_monitor_audio); /** * drm_rgb_quant_range_selectable - is RGB quantization range selectable? + * @edid: EDID block to scan * * Check whether the monitor reports the RGB quantization range selection * as supported. The AVI infoframe can then be used to inform the monitor From 89d61fc0f5d384f07f3e93af2bb52009ce26283a Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 21 Jan 2014 12:39:00 +0100 Subject: [PATCH 182/259] drm/doc: Clean up and integrate kerneldoc for drm_gem.c Fairly incomplete, but at least a start. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 6 +++- drivers/gpu/drm/drm_gem.c | 63 +++++++++++++++++++++++++++++++--- 2 files changed, 64 insertions(+), 5 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index f2d0f5b89194..1cdca9ad8844 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -868,7 +868,11 @@ char *date; abstracted from the client in libdrm. - + + GEM Function Reference +!Edrivers/gpu/drm/drm_gem.c + + diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 5bbad873c798..2136052ccee1 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -85,9 +85,9 @@ #endif /** - * Initialize the GEM device fields + * drm_gem_init - Initialize the GEM device fields + * @dev: drm_devic structure to initialize */ - int drm_gem_init(struct drm_device *dev) { @@ -120,6 +120,11 @@ drm_gem_destroy(struct drm_device *dev) } /** + * drm_gem_object_init - initialize an allocated shmem-backed GEM object + * @dev: drm_device the object should be initialized for + * @obj: drm_gem_object to initialize + * @size: object size + * * Initialize an already allocated GEM object of the specified size with * shmfs backing store. */ @@ -141,6 +146,11 @@ int drm_gem_object_init(struct drm_device *dev, EXPORT_SYMBOL(drm_gem_object_init); /** + * drm_gem_object_init - initialize an allocated private GEM object + * @dev: drm_device the object should be initialized for + * @obj: drm_gem_object to initialize + * @size: object size + * * Initialize an already allocated GEM object of the specified size with * no GEM provided backing store. Instead the caller is responsible for * backing the object and handling it. @@ -176,6 +186,9 @@ drm_gem_remove_prime_handles(struct drm_gem_object *obj, struct drm_file *filp) } /** + * drm_gem_object_free - release resources bound to userspace handles + * @obj: GEM object to clean up. + * * Called after the last handle to the object has been closed * * Removes any name for the object. Note that this must be @@ -225,7 +238,12 @@ drm_gem_object_handle_unreference_unlocked(struct drm_gem_object *obj) } /** - * Removes the mapping from handle to filp for this object. + * drm_gem_handle_delete - deletes the given file-private handle + * @filp: drm file-private structure to use for the handle look up + * @handle: userspace handle to delete + * + * Removes the GEM handle from the @filp lookup table and if this is the last + * handle also cleans up linked resources like GEM names. */ int drm_gem_handle_delete(struct drm_file *filp, u32 handle) @@ -270,6 +288,9 @@ EXPORT_SYMBOL(drm_gem_handle_delete); /** * drm_gem_dumb_destroy - dumb fb callback helper for gem based drivers + * @file: drm file-private structure to remove the dumb handle from + * @dev: corresponding drm_device + * @handle: the dumb handle to remove * * This implements the ->dumb_destroy kms driver callback for drivers which use * gem to manage their backing storage. @@ -284,6 +305,9 @@ EXPORT_SYMBOL(drm_gem_dumb_destroy); /** * drm_gem_handle_create_tail - internal functions to create a handle + * @file_priv: drm file-private structure to register the handle for + * @obj: object to register + * @handlep: pionter to return the created handle to the caller * * This expects the dev->object_name_lock to be held already and will drop it * before returning. Used to avoid races in establishing new handles when @@ -336,6 +360,11 @@ drm_gem_handle_create_tail(struct drm_file *file_priv, } /** + * gem_handle_create - create a gem handle for an object + * @file_priv: drm file-private structure to register the handle for + * @obj: object to register + * @handlep: pionter to return the created handle to the caller + * * Create a handle for this object. This adds a handle reference * to the object, which includes a regular reference count. Callers * will likely want to dereference the object afterwards. @@ -536,6 +565,11 @@ drm_gem_object_lookup(struct drm_device *dev, struct drm_file *filp, EXPORT_SYMBOL(drm_gem_object_lookup); /** + * drm_gem_close_ioctl - implementation of the GEM_CLOSE ioctl + * @dev: drm_device + * @data: ioctl data + * @file_priv: drm file-private structure + * * Releases the handle to an mm object. */ int @@ -554,6 +588,11 @@ drm_gem_close_ioctl(struct drm_device *dev, void *data, } /** + * drm_gem_flink_ioctl - implementation of the GEM_FLINK ioctl + * @dev: drm_device + * @data: ioctl data + * @file_priv: drm file-private structure + * * Create a global name for an object, returning the name. * * Note that the name does not hold a reference; when the object @@ -601,6 +640,11 @@ drm_gem_flink_ioctl(struct drm_device *dev, void *data, } /** + * drm_gem_open - implementation of the GEM_OPEN ioctl + * @dev: drm_device + * @data: ioctl data + * @file_priv: drm file-private structure + * * Open an object using the global name, returning a handle and the size. * * This handle (of course) holds a reference to the object, so the object @@ -640,6 +684,10 @@ drm_gem_open_ioctl(struct drm_device *dev, void *data, } /** + * gem_gem_open - initalizes GEM file-private structures at devnode open time + * @dev: drm_device which is being opened by userspace + * @file_private: drm file-private structure to set up + * * Called at device open time, sets up the structure for handling refcounting * of mm objects. */ @@ -650,7 +698,7 @@ drm_gem_open(struct drm_device *dev, struct drm_file *file_private) spin_lock_init(&file_private->table_lock); } -/** +/* * Called at device close to release the file's * handle references on objects. */ @@ -674,6 +722,10 @@ drm_gem_object_release_handle(int id, void *ptr, void *data) } /** + * drm_gem_release - release file-private GEM resources + * @dev: drm_device which is being closed by userspace + * @file_private: drm file-private structure to clean up + * * Called at close time when the filp is going away. * * Releases any remaining references on objects by this filp. @@ -697,6 +749,9 @@ drm_gem_object_release(struct drm_gem_object *obj) EXPORT_SYMBOL(drm_gem_object_release); /** + * drm_gem_object_free - free a GEM object + * @kref: kref of the object to free + * * Called after the last reference to the object has been lost. * Must be called holding struct_ mutex * From 00153aebc22b8120ab18012e383c98e97fb509e4 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 21 Jan 2014 12:51:43 +0100 Subject: [PATCH 183/259] drm/doc: Remove from rendernode docs The stylesheet doesn't allow this in normal paragraphs. Cc: David Herrmann Acked-by: David Herrmann Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 1cdca9ad8844..0d2adf9825e9 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -2673,8 +2673,8 @@ int (*resume) (struct drm_device *); DRM core provides multiple character-devices for user-space to use. Depending on which device is opened, user-space can perform a different set of operations (mainly ioctls). The primary node is always created - and called card<num>. Additionally, a currently - unused control node, called controlD<num> is also + and called card<num>. Additionally, a currently + unused control node, called controlD<num> is also created. The primary node provides all legacy operations and historically was the only interface used by userspace. With KMS, the control node was introduced. However, the planned KMS control interface @@ -2689,21 +2689,21 @@ int (*resume) (struct drm_device *); nodes were introduced. Render nodes solely serve render clients, that is, no modesetting or privileged ioctls can be issued on render nodes. Only non-global rendering commands are allowed. If a driver supports - render nodes, it must advertise it via the DRIVER_RENDER + render nodes, it must advertise it via the DRIVER_RENDER DRM driver capability. If not supported, the primary node must be used for render clients together with the legacy drmAuth authentication procedure. If a driver advertises render node support, DRM core will create a - separate render node called renderD<num>. There will + separate render node called renderD<num>. There will be one render node per device. No ioctls except PRIME-related ioctls - will be allowed on this node. Especially GEM_OPEN will be + will be allowed on this node. Especially GEM_OPEN will be explicitly prohibited. Render nodes are designed to avoid the buffer-leaks, which occur if clients guess the flink names or mmap offsets on the legacy interface. Additionally to this basic interface, drivers must mark their driver-dependent render-only ioctls as - DRM_RENDER_ALLOW so render clients can use them. Driver + DRM_RENDER_ALLOW so render clients can use them. Driver authors must be careful not to allow any privileged ioctls on render nodes. From 3519f70ee7c1d786ef08a977c241128efc291227 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 12:21:16 +0100 Subject: [PATCH 184/259] drm/doc: Reorganize driver documentation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Split up the DocBook into the core drm part and a 2nd part for driver documentation. As an example add a very (very!) basic skeleton for i915. v1: Typo fixes from Dieter. Cc: Dieter Nützel Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 82 ++++++++++++++++++++++++++++++---- 1 file changed, 74 insertions(+), 8 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 0d2adf9825e9..e377b88304ef 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -60,7 +60,15 @@ - + + DRM Core + + + This first part of the DRM Developer's Guide documents core DRM code, + helper libraries for writting drivers and generic userspace interfaces + exposed by DRM drivers. + + Introduction @@ -2764,15 +2772,73 @@ int (*resume) (struct drm_device *); + + + DRM Drivers - - - - DRM Driver API + - Include auto-generated API reference here (need to reference it - from paragraphs above too). + This second part of the DRM Developer's Guide documents driver code, + implementation details and also all the driver-specific userspace + interfaces. Especially since all hardware-acceleration interfaces to + userspace are driver specific for efficiency and other reasons these + interfaces can be rather substantial. Hence every driver has its own + chapter. - + + + drm/i915 Intel GFX Driver + + The drm/i915 driver supports all (with the exception of some very early + models) integrated GFX chipsets with both Intel display and rendering + blocks. This excludes a set of SoC platforms with an SGX rendering unit, + those have basic support through the gma500 drm driver. + + + Display Hardware Handling + + This section covers everything related to the display hardware including + the mode setting infrastructure, plane, sprite and cursor handling and + display, output probing and related topics. + + + Mode Setting Infrastructure + + The i915 driver is thus far the only DRM driver which doesn't use the + common DRM helper code to implement mode setting sequences. Thus it + has its own tailor-made infrastructure for executing a display + configuration change. + + + + Plane Configuration + + This section covers plane configuration and composition with the + primary plane, sprites, cursors and overlays. This includes the + infrastructure to do atomic vsync'ed updates of all this state and + also tightly coupled topics like watermark setup and computation, + framebuffer compression and panel self refresh. + + + + Output Probing + + This section covers output probing and related infrastructure like the + hotplug interrupt storm detection and mitigation code. Note that the + i915 driver still uses most of the common DRM helper code for output + probing, so those sections fully apply. + + + + + + Memory Management and Command Submission + + This sections covers all things related to the GEM implementation in the + i915 driver. + + + + From 4c5acf3cc8423c90f620e578f14181a56fa7fb4e Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 12:28:42 +0100 Subject: [PATCH 185/259] drm/doc: Move the vma offset manager to the right spot Currently it's sitting in the mode setting helper section, which isn't quite right. Looks much better in the memory management section next to TTM and GEM. Cc: David Herrmann Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index e377b88304ef..2def6f3a9061 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -881,6 +881,12 @@ char *date; !Edrivers/gpu/drm/drm_gem.c + + VMA Offset Manager +!Pdrivers/gpu/drm/drm_vma_manager.c vma offset manager +!Edrivers/gpu/drm/drm_vma_manager.c +!Iinclude/drm/drm_vma_manager.h + @@ -2218,12 +2224,6 @@ void intel_crt_init(struct drm_device *dev) !Iinclude/drm/drm_flip_work.h !Edrivers/gpu/drm/drm_flip_work.c - - VMA Offset Manager -!Pdrivers/gpu/drm/drm_vma_manager.c vma offset manager -!Edrivers/gpu/drm/drm_vma_manager.c -!Iinclude/drm/drm_vma_manager.h - From 1aa12258d6056e47cfe49eb13cc7b652f7dab956 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 12:33:22 +0100 Subject: [PATCH 186/259] drm/doc: Remove the "command submissin and fencing" section This should be done in the driver chapter instead. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 2def6f3a9061..750ba8fb496a 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -2585,16 +2585,6 @@ int num_ioctls; - - Command submission & fencing - - This should cover a few device-specific command submission - implementations. - - - - - Suspend/Resume From e1f8ebdcc230a9ff9e9e17707c22a5f0a5a885ee Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 16:32:47 +0100 Subject: [PATCH 187/259] drm/doc: No more drm perf counters MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Those all died with commit 0111be42186fc5461b9e9d579014c70869ab3152 Author: Ville Syrjälä Date: Fri Oct 4 14:53:41 2013 +0300 drm: Kill drm perf counter leftovers Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 750ba8fb496a..26539ee3c63e 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -272,8 +272,8 @@ char *date; The load method is the driver and device initialization entry point. The method is responsible for allocating and - initializing driver private data, specifying supported performance - counters, performing resource allocation and mapping (e.g. acquiring + initializing driver private data, performing resource allocation and + mapping (e.g. acquiring clocks, mapping registers or allocating command buffers), initializing the memory manager (), installing the IRQ handler (), setting up @@ -303,7 +303,7 @@ char *date; their load method called with flags to 0. - Driver Private & Performance Counters + Driver Private Data The driver private hangs off the main drm_device structure and can be used for @@ -315,14 +315,6 @@ char *date; drm_device.dev_priv set to NULL when the driver is unloaded. - - DRM supports several counters which were used for rough performance - characterization. This stat counter system is deprecated and should not - be used. If performance monitoring is desired, the developer should - investigate and potentially enhance the kernel perf and tracing - infrastructure to export GPU related performance information for - consumption by performance monitoring tools and applications. - IRQ Registration From aa4cd9100e0709ea0bc6f85090188ace895e51fe Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 16:42:02 +0100 Subject: [PATCH 188/259] drm/doc: Document drm_helper_resume_force_mode Stumbled over while reviewing all occurences in the DRM doc talking about suspend/resume. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 7 +++++-- drivers/gpu/drm/drm_crtc_helper.c | 9 +++++++++ 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 26539ee3c63e..8e1052434bab 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -1151,8 +1151,11 @@ int max_width, max_height; This operation is called with the mode config lock held. - FIXME: How should set_config interact with DPMS? If the CRTC is - suspended, should it be resumed? + Note that the drm core has no notion of restoring the mode setting + state after resume, since all resume handling is in the full + responsibility of the driver. The common mode setting helper library + though provides a helper which can be used for this: + drm_helper_resume_force_mode. diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index ea92b827e787..85d476abea6c 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -943,6 +943,15 @@ int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, } EXPORT_SYMBOL(drm_helper_mode_fill_fb_struct); +/** + * drm_helper_resume_force_mode - force-restore mode setting configuration + * @dev: drm_device which should be restored + * + * Drivers which use the mode setting helpers can use this function to + * force-restore the mode setting configuration e.g. on resume or when something + * else might have trampled over the hw state (like some overzealous old BIOSen + * tended to do). + */ int drm_helper_resume_force_mode(struct drm_device *dev) { struct drm_crtc *crtc; From 4c6e2dfe08987b1e5d884939967037d68412d829 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 16:46:44 +0100 Subject: [PATCH 189/259] drm/doc: Hide legacy horrors better By consolidating them all into one section at the very end. And to make double-sure that no one gets confused start with a stern warning against any use of them. And prefix all subsections with "Legacy". Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 56 +++++++++++++++++++++------------- 1 file changed, 34 insertions(+), 22 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 8e1052434bab..0a9407ae4025 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -2579,32 +2579,44 @@ int num_ioctls; - - Suspend/Resume + Legacy Support Code - The DRM core provides some suspend/resume code, but drivers wanting full - suspend/resume support should provide save() and restore() functions. - These are called at suspend, hibernate, or resume time, and should perform - any state save or restore required by your device across suspend or - hibernate states. + The section very brievely covers some of the old legacy support code which + is only used by old DRM drivers which have done a so-called shadow-attach + to the underlying device instead of registering as a real driver. This + also includes some of the old generic buffer mangement and command + submission code. Do not use any of this in new and modern drivers. - int (*suspend) (struct drm_device *, pm_message_t state); -int (*resume) (struct drm_device *); - - Those are legacy suspend and resume methods. New driver should use the - power management interface provided by their bus type (usually through - the struct device_driver dev_pm_ops) and set - these methods to NULL. - - - - DMA services - - This should cover how DMA mapping etc. is supported by the core. - These functions are deprecated and should not be used. - + + Legacy Suspend/Resume + + The DRM core provides some suspend/resume code, but drivers wanting full + suspend/resume support should provide save() and restore() functions. + These are called at suspend, hibernate, or resume time, and should perform + any state save or restore required by your device across suspend or + hibernate states. + + int (*suspend) (struct drm_device *, pm_message_t state); + int (*resume) (struct drm_device *); + + Those are legacy suspend and resume methods which + only work with the legacy shadow-attach driver + registration functions. New driver should use the power management + interface provided by their bus type (usually through + the struct device_driver dev_pm_ops) and set + these methods to NULL. + + + + + Legacy DMA Services + + This should cover how DMA mapping etc. is supported by the core. + These functions are deprecated and should not be used. + + From 2d123f463669cb7b84b56aa00e073ce07fe7aff2 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 18:26:16 +0100 Subject: [PATCH 190/259] drm/docs: Include hdmi infoframe helper reference Thierry created such nice kerneldocs, it's a shame we've left them lingering! For the fun of it also add a bit of kerneldoc to the header so that we can also include that. Just in case someone adds kerneldoc in there. Cc: Thierry Reding Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 11 +++++++++++ include/linux/hdmi.h | 12 ++++++++++++ 2 files changed, 23 insertions(+) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 0a9407ae4025..0e0390ece122 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -2219,6 +2219,17 @@ void intel_crt_init(struct drm_device *dev) !Iinclude/drm/drm_flip_work.h !Edrivers/gpu/drm/drm_flip_work.c + + HDMI Infoframes Helper Reference + + Strictly speaking this is not a DRM helper library but generally useable + by any driver interfacing with HDMI outputs like v4l or alsa drivers. + But it nicely fits into the overall topic of mode setting helper + libraries and hence is also included here. + +!Iinclude/linux/hdmi.h +!Edrivers/video/hdmi.c + diff --git a/include/linux/hdmi.h b/include/linux/hdmi.h index 9231be9e90a2..11c0182a153b 100644 --- a/include/linux/hdmi.h +++ b/include/linux/hdmi.h @@ -262,6 +262,18 @@ union hdmi_vendor_any_infoframe { struct hdmi_vendor_infoframe hdmi; }; +/** + * union hdmi_infoframe - overall union of all abstract infoframe representations + * @any: generic infoframe + * @avi: avi infoframe + * @spd: spd infoframe + * @vendor: union of all vendor infoframes + * @audio: audio infoframe + * + * This is used by the generic pack function. This works since all infoframes + * have the same header which also indicates which type of infoframe should be + * packed. + */ union hdmi_infoframe { struct hdmi_any_infoframe any; struct hdmi_avi_infoframe avi; From 251261db7f71829968a8fe80ae3f296fc96851cd Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 18:46:33 +0100 Subject: [PATCH 191/259] drm/doc: Clarify PRIME documentation PRIME fds aren't actually GEM fds but are (like the modeset API) independent of the underlying buffer manager, as long as that one uses uint32_t as handles. So move that entire section out of the GEM section and reword it a bit to clarify which parts of PRIME are generic, and which are the mandatory pieces for GEM drivers to correctly implement the GEM lifetime rules. The rewording mostly consists of not mixing up GEM, PRIME and DRM. I've considered adding some blurbs to the GEM object lifetime section about interactions with dma-bufs, but then dropped that. As long as drivers use the right helpers they should have this all implemented correctly and hence can be regarded as an implementation detail of the PRIME/GEM helpers. So no need to confuse driver writers with those tricky interactions. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 125 +++++++++++++++++++-------------- 1 file changed, 74 insertions(+), 51 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 0e0390ece122..641db5cb656c 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -697,55 +697,16 @@ char *date; respectively. The conversion is handled by the DRM core without any driver-specific support. - - Similar to global names, GEM file descriptors are also used to share GEM - objects across processes. They offer additional security: as file - descriptors must be explicitly sent over UNIX domain sockets to be shared - between applications, they can't be guessed like the globally unique GEM - names. - - - Drivers that support GEM file descriptors, also known as the DRM PRIME - API, must set the DRIVER_PRIME bit in the struct - drm_driver - driver_features field, and implement the - prime_handle_to_fd and - prime_fd_to_handle operations. - - - int (*prime_handle_to_fd)(struct drm_device *dev, - struct drm_file *file_priv, uint32_t handle, - uint32_t flags, int *prime_fd); - int (*prime_fd_to_handle)(struct drm_device *dev, - struct drm_file *file_priv, int prime_fd, - uint32_t *handle); - Those two operations convert a handle to a PRIME file descriptor and - vice versa. Drivers must use the kernel dma-buf buffer sharing framework - to manage the PRIME file descriptors. - - - While non-GEM drivers must implement the operations themselves, GEM - drivers must use the drm_gem_prime_handle_to_fd - and drm_gem_prime_fd_to_handle helper functions. - Those helpers rely on the driver - gem_prime_export and - gem_prime_import operations to create a dma-buf - instance from a GEM object (dma-buf exporter role) and to create a GEM - object from a dma-buf instance (dma-buf importer role). - - - struct dma_buf * (*gem_prime_export)(struct drm_device *dev, - struct drm_gem_object *obj, - int flags); - struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, - struct dma_buf *dma_buf); - These two operations are mandatory for GEM drivers that support DRM - PRIME. - - - DRM PRIME Helper Functions Reference -!Pdrivers/gpu/drm/drm_prime.c PRIME Helpers - + + GEM also supports buffer sharing with dma-buf file descriptors through + PRIME. GEM-based drivers must use the provided helpers functions to + implement the exporting and importing correctly. See . + Since sharing file descriptors is inherently more secure than the + easily guessable and global GEM names it is the preferred buffer + sharing mechanism. Sharing buffers through GEM names is only supported + for legacy userspace. Furthermore PRIME also allows cross-device + buffer sharing since it is based on dma-bufs. + GEM Objects Mapping @@ -868,10 +829,10 @@ char *date; abstracted from the client in libdrm. - + GEM Function Reference !Edrivers/gpu/drm/drm_gem.c - + VMA Offset Manager @@ -879,6 +840,68 @@ char *date; !Edrivers/gpu/drm/drm_vma_manager.c !Iinclude/drm/drm_vma_manager.h + + PRIME Buffer Sharing + + PRIME is the cross device buffer sharing framework in drm, originally + created for the OPTIMUS range of multi-gpu platforms. To userspace + PRIME buffers are dma-buf based file descriptors. + + + Overview and Driver Interface + + Similar to GEM global names, PRIME file descriptors are + also used to share buffer objects across processes. They offer + additional security: as file descriptors must be explicitly sent over + UNIX domain sockets to be shared between applications, they can't be + guessed like the globally unique GEM names. + + + Drivers that support the PRIME + API must set the DRIVER_PRIME bit in the struct + drm_driver + driver_features field, and implement the + prime_handle_to_fd and + prime_fd_to_handle operations. + + + int (*prime_handle_to_fd)(struct drm_device *dev, + struct drm_file *file_priv, uint32_t handle, + uint32_t flags, int *prime_fd); +int (*prime_fd_to_handle)(struct drm_device *dev, + struct drm_file *file_priv, int prime_fd, + uint32_t *handle); + Those two operations convert a handle to a PRIME file descriptor and + vice versa. Drivers must use the kernel dma-buf buffer sharing framework + to manage the PRIME file descriptors. Similar to the mode setting + API PRIME is agnostic to the underlying buffer object manager, as + long as handles are 32bit unsinged integers. + + + While non-GEM drivers must implement the operations themselves, GEM + drivers must use the drm_gem_prime_handle_to_fd + and drm_gem_prime_fd_to_handle helper functions. + Those helpers rely on the driver + gem_prime_export and + gem_prime_import operations to create a dma-buf + instance from a GEM object (dma-buf exporter role) and to create a GEM + object from a dma-buf instance (dma-buf importer role). + + + struct dma_buf * (*gem_prime_export)(struct drm_device *dev, + struct drm_gem_object *obj, + int flags); +struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, + struct dma_buf *dma_buf); + These two operations are mandatory for GEM drivers that support + PRIME. + + + + PRIME Helper Functions Reference +!Pdrivers/gpu/drm/drm_prime.c PRIME Helpers + + From 39cc344acd414eda231f612325cf824b976025e5 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 19:16:30 +0100 Subject: [PATCH 192/259] drm/doc: Add PRIME function references For giant hilarity the DocBook reference overview is only generated when in a level 2 section, not in a level 3 section. So we need to move this up a bit as a side-by-side section to the main PRIME documentation. Whatever. To have a complete set of references add the missing kerneldoc for all functions exported to modules with the exception of the file private init/destroy functions - drivers have no business calling those, so let's just drop the EXPORT_SYMBOL instead. Also reflow the function parameters to align correctly and break at 80 chars - my OCD couldn't stand them while writing the kerneldoc ;-) Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 6 +- drivers/gpu/drm/drm_prime.c | 110 ++++++++++++++++++++++++++------- 2 files changed, 94 insertions(+), 22 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 641db5cb656c..f83622ebda9d 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -898,10 +898,14 @@ struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, - PRIME Helper Functions Reference + PRIME Helper Functions !Pdrivers/gpu/drm/drm_prime.c PRIME Helpers + + PRIME Function References +!Edrivers/gpu/drm/drm_prime.c + diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c index 56805c39c906..f1437b6c8dbf 100644 --- a/drivers/gpu/drm/drm_prime.c +++ b/drivers/gpu/drm/drm_prime.c @@ -68,7 +68,8 @@ struct drm_prime_attachment { enum dma_data_direction dir; }; -static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle) +static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, + struct dma_buf *dma_buf, uint32_t handle) { struct drm_prime_member *member; @@ -174,7 +175,7 @@ void drm_prime_remove_buf_handle_locked(struct drm_prime_file_private *prime_fpr } static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach, - enum dma_data_direction dir) + enum dma_data_direction dir) { struct drm_prime_attachment *prime_attach = attach->priv; struct drm_gem_object *obj = attach->dmabuf->priv; @@ -211,11 +212,19 @@ static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach, } static void drm_gem_unmap_dma_buf(struct dma_buf_attachment *attach, - struct sg_table *sgt, enum dma_data_direction dir) + struct sg_table *sgt, + enum dma_data_direction dir) { /* nothing to be done here */ } +/** + * drm_gem_dmabuf_release - dma_buf release implementation for GEM + * @dma_buf: buffer to be released + * + * Generic release function for dma_bufs exported as PRIME buffers. GEM drivers + * must use this in their dma_buf ops structure as the release callback. + */ void drm_gem_dmabuf_release(struct dma_buf *dma_buf) { struct drm_gem_object *obj = dma_buf->priv; @@ -242,30 +251,30 @@ static void drm_gem_dmabuf_vunmap(struct dma_buf *dma_buf, void *vaddr) } static void *drm_gem_dmabuf_kmap_atomic(struct dma_buf *dma_buf, - unsigned long page_num) + unsigned long page_num) { return NULL; } static void drm_gem_dmabuf_kunmap_atomic(struct dma_buf *dma_buf, - unsigned long page_num, void *addr) + unsigned long page_num, void *addr) { } static void *drm_gem_dmabuf_kmap(struct dma_buf *dma_buf, - unsigned long page_num) + unsigned long page_num) { return NULL; } static void drm_gem_dmabuf_kunmap(struct dma_buf *dma_buf, - unsigned long page_num, void *addr) + unsigned long page_num, void *addr) { } static int drm_gem_dmabuf_mmap(struct dma_buf *dma_buf, - struct vm_area_struct *vma) + struct vm_area_struct *vma) { struct drm_gem_object *obj = dma_buf->priv; struct drm_device *dev = obj->dev; @@ -315,6 +324,15 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops = { * driver's scatter/gather table */ +/** + * drm_gem_prime_export - helper library implemention of the export callback + * @dev: drm_device to export from + * @obj: GEM object to export + * @flags: flags like DRM_CLOEXEC + * + * This is the implementation of the gem_prime_export functions for GEM drivers + * using the PRIME helpers. + */ struct dma_buf *drm_gem_prime_export(struct drm_device *dev, struct drm_gem_object *obj, int flags) { @@ -355,9 +373,23 @@ static struct dma_buf *export_and_register_object(struct drm_device *dev, return dmabuf; } +/** + * drm_gem_prime_handle_to_fd - PRIME export function for GEM drivers + * @dev: dev to export the buffer from + * @file_priv: drm file-private structure + * @handle: buffer handle to export + * @flags: flags like DRM_CLOEXEC + * @prime_fd: pointer to storage for the fd id of the create dma-buf + * + * This is the PRIME export function which must be used mandatorily by GEM + * drivers to ensure correct lifetime management of the underlying GEM object. + * The actual exporting from GEM object to a dma-buf is done through the + * gem_prime_export driver callback. + */ int drm_gem_prime_handle_to_fd(struct drm_device *dev, - struct drm_file *file_priv, uint32_t handle, uint32_t flags, - int *prime_fd) + struct drm_file *file_priv, uint32_t handle, + uint32_t flags, + int *prime_fd) { struct drm_gem_object *obj; int ret = 0; @@ -441,6 +473,14 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev, } EXPORT_SYMBOL(drm_gem_prime_handle_to_fd); +/** + * drm_gem_prime_import - helper library implemention of the import callback + * @dev: drm_device to import into + * @dma_buf: dma-buf object to import + * + * This is the implementation of the gem_prime_import functions for GEM drivers + * using the PRIME helpers. + */ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev, struct dma_buf *dma_buf) { @@ -496,8 +536,21 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev, } EXPORT_SYMBOL(drm_gem_prime_import); +/** + * drm_gem_prime_fd_to_handle - PRIME import function for GEM drivers + * @dev: dev to export the buffer from + * @file_priv: drm file-private structure + * @prime_fd: fd id of the dma-buf which should be imported + * @handle: pointer to storage for the handle of the imported buffer object + * + * This is the PRIME import function which must be used mandatorily by GEM + * drivers to ensure correct lifetime management of the underlying GEM object. + * The actual importing of GEM object from the dma-buf is done through the + * gem_import_export driver callback. + */ int drm_gem_prime_fd_to_handle(struct drm_device *dev, - struct drm_file *file_priv, int prime_fd, uint32_t *handle) + struct drm_file *file_priv, int prime_fd, + uint32_t *handle) { struct dma_buf *dma_buf; struct drm_gem_object *obj; @@ -598,12 +651,14 @@ int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data, args->fd, &args->handle); } -/* - * drm_prime_pages_to_sg +/** + * drm_prime_pages_to_sg - converts a page array into an sg list + * @pages: pointer to the array of page pointers to convert + * @nr_pages: length of the page vector * - * this helper creates an sg table object from a set of pages + * This helper creates an sg table object from a set of pages * the driver is responsible for mapping the pages into the - * importers address space + * importers address space for use with dma_buf itself. */ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages) { @@ -628,9 +683,16 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages) } EXPORT_SYMBOL(drm_prime_pages_to_sg); -/* export an sg table into an array of pages and addresses - this is currently required by the TTM driver in order to do correct fault - handling */ +/** + * drm_prime_sg_to_page_addr_arrays - convert an sg table into a page array + * @sgt: scatter-gather table to convert + * @pages: array of page pointers to store the page array in + * @addrs: optional array to store the dma bus address of each page + * @max_pages: size of both the passed-in arrays + * + * Exports an sg table into an array of pages and addresses. This is currently + * required by the TTM driver in order to do correct fault handling. + */ int drm_prime_sg_to_page_addr_arrays(struct sg_table *sgt, struct page **pages, dma_addr_t *addrs, int max_pages) { @@ -663,7 +725,15 @@ int drm_prime_sg_to_page_addr_arrays(struct sg_table *sgt, struct page **pages, return 0; } EXPORT_SYMBOL(drm_prime_sg_to_page_addr_arrays); -/* helper function to cleanup a GEM/prime object */ + +/** + * drm_prime_gem_destroy - helper to clean up a PRIME-imported GEM object + * @obj: GEM object which was created from a dma-buf + * @sg: the sg-table which was pinned at import time + * + * This is the cleanup functions which GEM drivers need to call when they use + * @drm_gem_prime_import to import dma-bufs. + */ void drm_prime_gem_destroy(struct drm_gem_object *obj, struct sg_table *sg) { struct dma_buf_attachment *attach; @@ -683,11 +753,9 @@ void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv) INIT_LIST_HEAD(&prime_fpriv->head); mutex_init(&prime_fpriv->lock); } -EXPORT_SYMBOL(drm_prime_init_file_private); void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv) { /* by now drm_gem_release should've made sure the list is empty */ WARN_ON(!list_empty(&prime_fpriv->head)); } -EXPORT_SYMBOL(drm_prime_destroy_file_private); From 3a05700dc874f377b13e4368550e8c7228826ab4 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 22:38:57 +0100 Subject: [PATCH 193/259] drm/doc: Update copyright I've done quite a bit of cleanups, clarifications and mostly integrating kerneldoc. So I guess I should add myself. Also split up the copyright notices per holder to make it clear which year ranges are covered. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index f83622ebda9d..dd2a955031a7 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -29,12 +29,26 @@ + + Daniel + Vetter + Contributions all over the place + + Intel Corporation +
+ daniel.vetter@ffwll.ch +
+ + 2008-2009 - 2012 + 2013-2014 Intel Corporation + + + 2012 Laurent Pinchart From 79d1154478c057a7c894e34975ca4e5f16fe9803 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 22 Jan 2014 23:03:27 +0100 Subject: [PATCH 194/259] drm/mm: Remove MM_UNUSED_TARGET This was missed in commit c700c67bae6698fbc6bd20e2ae5dc62ddd367b3b Author: David Herrmann Date: Sat Jul 27 13:39:28 2013 +0200 drm/mm: remove unused API Cc: David Herrmann Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_mm.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c index af93cc55259f..d0a8e8482fe0 100644 --- a/drivers/gpu/drm/drm_mm.c +++ b/drivers/gpu/drm/drm_mm.c @@ -47,8 +47,6 @@ #include #include -#define MM_UNUSED_TARGET 4 - static struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm, unsigned long size, unsigned alignment, From 93110be69616df7dcd9cc3611e94400287fc26fb Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 00:31:48 +0100 Subject: [PATCH 195/259] drm/doc: Overview documentation for drm_mm.c kerneldoc polish will follow in the next patch. Hopefully documenting the lru scan support a bit better spurs someone to give this a shot in the ttm eviction code. At least in i915 it helped quite a lot with memory thrashing on platforms where eviction was (we've fixed that too meanwhile) fairly expensive. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 11 ++++++ drivers/gpu/drm/drm_mm.c | 67 ++++++++++++++++++++++++++++++++++ 2 files changed, 78 insertions(+) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index dd2a955031a7..2ac018bfbddf 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -920,6 +920,17 @@ struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, PRIME Function References !Edrivers/gpu/drm/drm_prime.c + + DRM MM Range Allocator + + Overview +!Pdrivers/gpu/drm/drm_mm.c Overview + + + LRU Scan/Eviction Support +!Pdrivers/gpu/drm/drm_mm.c lru scan roaster + + diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c index d0a8e8482fe0..276a7a27c166 100644 --- a/drivers/gpu/drm/drm_mm.c +++ b/drivers/gpu/drm/drm_mm.c @@ -47,6 +47,45 @@ #include #include +/** + * DOC: Overview + * + * drm_mm provides a simple range allocator. The drivers are free to use the + * resource allocator from the linux core if it suits them, the upside of drm_mm + * is that it's in the DRM core. Which means that it's easier to extend for + * some of the crazier special purpose needs of gpus. + * + * The main data struct is &drm_mm, allocations are tracked in &drm_mm_node. + * Drivers are free to embed either of them into their own suitable + * datastructures. drm_mm itself will not do any allocations of its own, so if + * drivers choose not to embed nodes they need to still allocate them + * themselves. + * + * The range allocator also supports reservation of preallocated blocks. This is + * useful for taking over initial mode setting configurations from the firmware, + * where an object needs to be created which exactly matches the firmware's + * scanout target. As long as the range is still free it can be inserted anytime + * after the allocator is initialized, which helps with avoiding looped + * depencies in the driver load sequence. + * + * drm_mm maintains a stack of most recently freed holes, which of all + * simplistic datastructures seems to be a fairly decent approach to clustering + * allocations and avoiding too much fragmentation. This means free space + * searches are O(num_holes). Given that all the fancy features drm_mm supports + * something better would be fairly complex and since gfx thrashing is a fairly + * steep cliff not a real concern. Removing a node again is O(1). + * + * drm_mm supports a few features: Alignment and range restrictions can be + * supplied. Further more every &drm_mm_node has a color value (which is just an + * opaqua unsigned long) which in conjunction with a driver callback can be used + * to implement sophisticated placement restrictions. The i915 DRM driver uses + * this to implement guard pages between incompatible caching domains in the + * graphics TT. + * + * Finally iteration helpers to walk all nodes and all holes are provided as are + * some basic allocator dumpers for debugging. + */ + static struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm, unsigned long size, unsigned alignment, @@ -399,6 +438,34 @@ void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new) } EXPORT_SYMBOL(drm_mm_replace_node); +/** + * DOC: lru scan roaster + * + * Very often GPUs need to have continuous allocations for a given object. When + * evicting objects to make space for a new one it is therefore not most + * efficient when we simply start to select all objects from the tail of an LRU + * until there's a suitable hole: Especially for big objects or nodes that + * otherwise have special allocation constraints there's a good chance we evict + * lots of (smaller) objects unecessarily. + * + * The DRM range allocator supports this use-case through the scanning + * interfaces. First a scan operation needs to be initialized with + * drm_mm_init_scan() or drm_mm_init_scan_with_range(). The the driver adds + * objects to the roaster (probably by walking an LRU list, but this can be + * freely implemented) until a suitable hole is found or there's no further + * evitable object. + * + * The the driver must walk through all objects again in exactly the reverse + * order to restore the allocator state. Note that while the allocator is used + * in the scan mode no other operation is allowed. + * + * Finally the driver evicts all objects selected in the scan. Adding and + * removing an object is O(1), and since freeing a node is also O(1) the overall + * complexity is O(scanned_objects). So like the free stack which needs to be + * walked before a scan operation even begins this is linear in the number of + * objects. It doesn't seem to hurt badly. + */ + /** * Initializa lru scanning. * From e18c04128faa2aa08547f8b73b9ecbf8fd6936af Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 00:39:13 +0100 Subject: [PATCH 196/259] drm/doc: Add function reference documentation for drm_mm.c While at it do a tiny bit of interface cleanup and convert boolean return values to bool. With this patch all exported functions and inline helpers which are part of the drm_mm public interface are documented. Also drop superflous extern function modifiers since most of drm_mm.h doesn't use them - more consistent that way. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 5 ++ drivers/gpu/drm/drm_mm.c | 144 ++++++++++++++++++++++++++----- include/drm/drm_mm.h | 152 +++++++++++++++++++++++++++------ 3 files changed, 250 insertions(+), 51 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 2ac018bfbddf..d68bb0a2dc06 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -931,6 +931,11 @@ struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, !Pdrivers/gpu/drm/drm_mm.c lru scan roaster + + DRM MM Range Allocator Function References +!Edrivers/gpu/drm/drm_mm.c +!Iinclude/drm/drm_mm.h + diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c index 276a7a27c166..a2d45b748f86 100644 --- a/drivers/gpu/drm/drm_mm.c +++ b/drivers/gpu/drm/drm_mm.c @@ -144,6 +144,20 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node, } } +/** + * drm_mm_reserve_node - insert an pre-initialized node + * @mm: drm_mm allocator to insert @node into + * @node: drm_mm_node to insert + * + * This functions inserts an already set-up drm_mm_node into the allocator, + * meaning that start, size and color must be set by the caller. This is useful + * to initialize the allocator with preallocated objects which must be set-up + * before the range allocator can be set-up, e.g. when taking over a firmware + * framebuffer. + * + * Returns: + * 0 on success, -ENOSPC if there's no hole where @node is. + */ int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node) { struct drm_mm_node *hole; @@ -185,9 +199,18 @@ int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node) EXPORT_SYMBOL(drm_mm_reserve_node); /** - * Search for free space and insert a preallocated memory node. Returns - * -ENOSPC if no suitable free area is available. The preallocated memory node - * must be cleared. + * drm_mm_insert_node_generic - search for space and insert @node + * @mm: drm_mm to allocate from + * @node: preallocate node to insert + * @size: size of the allocation + * @alignment: alignment of the allocation + * @color: opaque tag value to use for this node + * @flags: flags to fine-tune the allocation + * + * The preallocated node must be cleared to 0. + * + * Returns: + * 0 on success, -ENOSPC if there's no suitable hole. */ int drm_mm_insert_node_generic(struct drm_mm *mm, struct drm_mm_node *node, unsigned long size, unsigned alignment, @@ -259,9 +282,20 @@ static void drm_mm_insert_helper_range(struct drm_mm_node *hole_node, } /** - * Search for free space and insert a preallocated memory node. Returns - * -ENOSPC if no suitable free area is available. This is for range - * restricted allocations. The preallocated memory node must be cleared. + * drm_mm_insert_node_in_range_generic - ranged search for space and insert @node + * @mm: drm_mm to allocate from + * @node: preallocate node to insert + * @size: size of the allocation + * @alignment: alignment of the allocation + * @color: opaque tag value to use for this node + * @start: start of the allowed range for this node + * @end: end of the allowed range for this node + * @flags: flags to fine-tune the allocation + * + * The preallocated node must be cleared to 0. + * + * Returns: + * 0 on success, -ENOSPC if there's no suitable hole. */ int drm_mm_insert_node_in_range_generic(struct drm_mm *mm, struct drm_mm_node *node, unsigned long size, unsigned alignment, unsigned long color, @@ -284,7 +318,12 @@ int drm_mm_insert_node_in_range_generic(struct drm_mm *mm, struct drm_mm_node *n EXPORT_SYMBOL(drm_mm_insert_node_in_range_generic); /** - * Remove a memory node from the allocator. + * drm_mm_remove_node - Remove a memory node from the allocator. + * @node: drm_mm_node to remove + * + * This just removes a node from its drm_mm allocator. The node does not need to + * be cleared again before it can be re-inserted into this or any other drm_mm + * allocator. It is a bug to call this function on a un-allocated node. */ void drm_mm_remove_node(struct drm_mm_node *node) { @@ -421,7 +460,13 @@ static struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_ } /** - * Moves an allocation. To be used with embedded struct drm_mm_node. + * drm_mm_replace_node - move an allocation from @old to @new + * @old: drm_mm_node to remove from the allocator + * @new: drm_mm_node which should inherit @old's allocation + * + * This is useful for when drivers embed the drm_mm_node structure and hence + * can't move allocations by reassigning pointers. It's a combination of remove + * and insert with the guarantee that the allocation start will match. */ void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new) { @@ -467,12 +512,18 @@ EXPORT_SYMBOL(drm_mm_replace_node); */ /** - * Initializa lru scanning. + * drm_mm_init_scan - initialize lru scanning + * @mm: drm_mm to scan + * @size: size of the allocation + * @alignment: alignment of the allocation + * @color: opaque tag value to use for the allocation * * This simply sets up the scanning routines with the parameters for the desired - * hole. + * hole. Note that there's no need to specify allocation flags, since they only + * change the place a node is allocated from within a suitable hole. * - * Warning: As long as the scan list is non-empty, no other operations than + * Warning: + * As long as the scan list is non-empty, no other operations than * adding/removing nodes to/from the scan list are allowed. */ void drm_mm_init_scan(struct drm_mm *mm, @@ -492,12 +543,20 @@ void drm_mm_init_scan(struct drm_mm *mm, EXPORT_SYMBOL(drm_mm_init_scan); /** - * Initializa lru scanning. + * drm_mm_init_scan - initialize range-restricted lru scanning + * @mm: drm_mm to scan + * @size: size of the allocation + * @alignment: alignment of the allocation + * @color: opaque tag value to use for the allocation + * @start: start of the allowed range for the allocation + * @end: end of the allowed range for the allocation * * This simply sets up the scanning routines with the parameters for the desired - * hole. This version is for range-restricted scans. + * hole. Note that there's no need to specify allocation flags, since they only + * change the place a node is allocated from within a suitable hole. * - * Warning: As long as the scan list is non-empty, no other operations than + * Warning: + * As long as the scan list is non-empty, no other operations than * adding/removing nodes to/from the scan list are allowed. */ void drm_mm_init_scan_with_range(struct drm_mm *mm, @@ -521,12 +580,16 @@ void drm_mm_init_scan_with_range(struct drm_mm *mm, EXPORT_SYMBOL(drm_mm_init_scan_with_range); /** + * drm_mm_scan_add_block - add a node to the scan list + * @node: drm_mm_node to add + * * Add a node to the scan list that might be freed to make space for the desired * hole. * - * Returns non-zero, if a hole has been found, zero otherwise. + * Returns: + * True if a hole has been found, false otherwise. */ -int drm_mm_scan_add_block(struct drm_mm_node *node) +bool drm_mm_scan_add_block(struct drm_mm_node *node) { struct drm_mm *mm = node->mm; struct drm_mm_node *prev_node; @@ -566,15 +629,16 @@ int drm_mm_scan_add_block(struct drm_mm_node *node) mm->scan_size, mm->scan_alignment)) { mm->scan_hit_start = hole_start; mm->scan_hit_end = hole_end; - return 1; + return true; } - return 0; + return false; } EXPORT_SYMBOL(drm_mm_scan_add_block); /** - * Remove a node from the scan list. + * drm_mm_scan_remove_block - remove a node from the scan list + * @node: drm_mm_node to remove * * Nodes _must_ be removed in the exact same order from the scan list as they * have been added, otherwise the internal state of the memory manager will be @@ -584,10 +648,11 @@ EXPORT_SYMBOL(drm_mm_scan_add_block); * immediately following drm_mm_search_free with !DRM_MM_SEARCH_BEST will then * return the just freed block (because its at the top of the free_stack list). * - * Returns one if this block should be evicted, zero otherwise. Will always - * return zero when no hole has been found. + * Returns: + * True if this block should be evicted, false otherwise. Will always + * return false when no hole has been found. */ -int drm_mm_scan_remove_block(struct drm_mm_node *node) +bool drm_mm_scan_remove_block(struct drm_mm_node *node) { struct drm_mm *mm = node->mm; struct drm_mm_node *prev_node; @@ -608,7 +673,15 @@ int drm_mm_scan_remove_block(struct drm_mm_node *node) } EXPORT_SYMBOL(drm_mm_scan_remove_block); -int drm_mm_clean(struct drm_mm * mm) +/** + * drm_mm_clean - checks whether an allocator is clean + * @mm: drm_mm allocator to check + * + * Returns: + * True if the allocator is completely free, false if there's still a node + * allocated in it. + */ +bool drm_mm_clean(struct drm_mm * mm) { struct list_head *head = &mm->head_node.node_list; @@ -616,6 +689,14 @@ int drm_mm_clean(struct drm_mm * mm) } EXPORT_SYMBOL(drm_mm_clean); +/** + * drm_mm_init - initialize a drm-mm allocator + * @mm: the drm_mm structure to initialize + * @start: start of the range managed by @mm + * @size: end of the range managed by @mm + * + * Note that @mm must be cleared to 0 before calling this function. + */ void drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size) { INIT_LIST_HEAD(&mm->hole_stack); @@ -637,6 +718,13 @@ void drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size) } EXPORT_SYMBOL(drm_mm_init); +/** + * drm_mm_takedown - clean up a drm_mm allocator + * @mm: drm_mm allocator to clean up + * + * Note that it is a bug to call this function on an allocator which is not + * clean. + */ void drm_mm_takedown(struct drm_mm * mm) { WARN(!list_empty(&mm->head_node.node_list), @@ -662,6 +750,11 @@ static unsigned long drm_mm_debug_hole(struct drm_mm_node *entry, return 0; } +/** + * drm_mm_debug_table - dump allocator state to dmesg + * @mm: drm_mm allocator to dump + * @prefix: prefix to use for dumping to dmesg + */ void drm_mm_debug_table(struct drm_mm *mm, const char *prefix) { struct drm_mm_node *entry; @@ -700,6 +793,11 @@ static unsigned long drm_mm_dump_hole(struct seq_file *m, struct drm_mm_node *en return 0; } +/** + * drm_mm_dump_table - dump allocator state to a seq_file + * @m: seq_file to dump to + * @mm: drm_mm allocator to dump + */ int drm_mm_dump_table(struct seq_file *m, struct drm_mm *mm) { struct drm_mm_node *entry; diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h index cba67865d18f..8b6981ab3fcf 100644 --- a/include/drm/drm_mm.h +++ b/include/drm/drm_mm.h @@ -85,11 +85,31 @@ struct drm_mm { unsigned long *start, unsigned long *end); }; +/** + * drm_mm_node_allocated - checks whether a node is allocated + * @node: drm_mm_node to check + * + * Drivers should use this helpers for proper encapusulation of drm_mm + * internals. + * + * Returns: + * True if the @node is allocated. + */ static inline bool drm_mm_node_allocated(struct drm_mm_node *node) { return node->allocated; } +/** + * drm_mm_initialized - checks whether an allocator is initialized + * @mm: drm_mm to check + * + * Drivers should use this helpers for proper encapusulation of drm_mm + * internals. + * + * Returns: + * True if the @mm is initialized. + */ static inline bool drm_mm_initialized(struct drm_mm *mm) { return mm->hole_stack.next; @@ -100,6 +120,17 @@ static inline unsigned long __drm_mm_hole_node_start(struct drm_mm_node *hole_no return hole_node->start + hole_node->size; } +/** + * drm_mm_hole_node_start - computes the start of the hole following @node + * @hole_node: drm_mm_node which implicitly tracks the following hole + * + * This is useful for driver-sepific debug dumpers. Otherwise drivers should not + * inspect holes themselves. Drivers must check first whether a hole indeed + * follows by looking at node->hole_follows. + * + * Returns: + * Start of the subsequent hole. + */ static inline unsigned long drm_mm_hole_node_start(struct drm_mm_node *hole_node) { BUG_ON(!hole_node->hole_follows); @@ -112,18 +143,49 @@ static inline unsigned long __drm_mm_hole_node_end(struct drm_mm_node *hole_node struct drm_mm_node, node_list)->start; } +/** + * drm_mm_hole_node_end - computes the end of the hole following @node + * @hole_node: drm_mm_node which implicitly tracks the following hole + * + * This is useful for driver-sepific debug dumpers. Otherwise drivers should not + * inspect holes themselves. Drivers must check first whether a hole indeed + * follows by looking at node->hole_follows. + * + * Returns: + * End of the subsequent hole. + */ static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node) { return __drm_mm_hole_node_end(hole_node); } +/** + * drm_mm_for_each_node - iterator to walk over all allocated nodes + * @entry: drm_mm_node structure to assign to in each iteration step + * @mm: drm_mm allocator to walk + * + * This iterator walks over all nodes in the range allocator. It is implemented + * with list_for_each, so not save against removal of elements. + */ #define drm_mm_for_each_node(entry, mm) list_for_each_entry(entry, \ &(mm)->head_node.node_list, \ node_list) -/* Note that we need to unroll list_for_each_entry in order to inline - * setting hole_start and hole_end on each iteration and keep the - * macro sane. +/** + * drm_mm_for_each_hole - iterator to walk over all holes + * @entry: drm_mm_node used internally to track progress + * @mm: drm_mm allocator to walk + * @hole_start: ulong variable to assign the hole start to on each iteration + * @hole_end: ulong variable to assign the hole end to on each iteration + * + * This iterator walks over all holes in the range allocator. It is implemented + * with list_for_each, so not save against removal of elements. @entry is used + * internally and will not reflect a real drm_mm_node for the very first hole. + * Hence users of this iterator may not access it. + * + * Implementation Note: + * We need to inline list_for_each_entry in order to be able to set hole_start + * and hole_end on each iteration while keeping the macro sane. */ #define drm_mm_for_each_hole(entry, mm, hole_start, hole_end) \ for (entry = list_entry((mm)->hole_stack.next, struct drm_mm_node, hole_stack); \ @@ -136,14 +198,30 @@ static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node) /* * Basic range manager support (drm_mm.c) */ -extern int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node); +int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node); -extern int drm_mm_insert_node_generic(struct drm_mm *mm, - struct drm_mm_node *node, - unsigned long size, - unsigned alignment, - unsigned long color, - enum drm_mm_search_flags flags); +int drm_mm_insert_node_generic(struct drm_mm *mm, + struct drm_mm_node *node, + unsigned long size, + unsigned alignment, + unsigned long color, + enum drm_mm_search_flags flags); +/** + * drm_mm_insert_node - search for space and insert @node + * @mm: drm_mm to allocate from + * @node: preallocate node to insert + * @size: size of the allocation + * @alignment: alignment of the allocation + * @flags: flags to fine-tune the allocation + * + * This is a simplified version of drm_mm_insert_node_generic() with @color set + * to 0. + * + * The preallocated node must be cleared to 0. + * + * Returns: + * 0 on success, -ENOSPC if there's no suitable hole. + */ static inline int drm_mm_insert_node(struct drm_mm *mm, struct drm_mm_node *node, unsigned long size, @@ -153,14 +231,32 @@ static inline int drm_mm_insert_node(struct drm_mm *mm, return drm_mm_insert_node_generic(mm, node, size, alignment, 0, flags); } -extern int drm_mm_insert_node_in_range_generic(struct drm_mm *mm, - struct drm_mm_node *node, - unsigned long size, - unsigned alignment, - unsigned long color, - unsigned long start, - unsigned long end, - enum drm_mm_search_flags flags); +int drm_mm_insert_node_in_range_generic(struct drm_mm *mm, + struct drm_mm_node *node, + unsigned long size, + unsigned alignment, + unsigned long color, + unsigned long start, + unsigned long end, + enum drm_mm_search_flags flags); +/** + * drm_mm_insert_node_in_range - ranged search for space and insert @node + * @mm: drm_mm to allocate from + * @node: preallocate node to insert + * @size: size of the allocation + * @alignment: alignment of the allocation + * @start: start of the allowed range for this node + * @end: end of the allowed range for this node + * @flags: flags to fine-tune the allocation + * + * This is a simplified version of drm_mm_insert_node_in_range_generic() with + * @color set to 0. + * + * The preallocated node must be cleared to 0. + * + * Returns: + * 0 on success, -ENOSPC if there's no suitable hole. + */ static inline int drm_mm_insert_node_in_range(struct drm_mm *mm, struct drm_mm_node *node, unsigned long size, @@ -173,13 +269,13 @@ static inline int drm_mm_insert_node_in_range(struct drm_mm *mm, 0, start, end, flags); } -extern void drm_mm_remove_node(struct drm_mm_node *node); -extern void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new); -extern void drm_mm_init(struct drm_mm *mm, - unsigned long start, - unsigned long size); -extern void drm_mm_takedown(struct drm_mm *mm); -extern int drm_mm_clean(struct drm_mm *mm); +void drm_mm_remove_node(struct drm_mm_node *node); +void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new); +void drm_mm_init(struct drm_mm *mm, + unsigned long start, + unsigned long size); +void drm_mm_takedown(struct drm_mm *mm); +bool drm_mm_clean(struct drm_mm *mm); void drm_mm_init_scan(struct drm_mm *mm, unsigned long size, @@ -191,10 +287,10 @@ void drm_mm_init_scan_with_range(struct drm_mm *mm, unsigned long color, unsigned long start, unsigned long end); -int drm_mm_scan_add_block(struct drm_mm_node *node); -int drm_mm_scan_remove_block(struct drm_mm_node *node); +bool drm_mm_scan_add_block(struct drm_mm_node *node); +bool drm_mm_scan_remove_block(struct drm_mm_node *node); -extern void drm_mm_debug_table(struct drm_mm *mm, const char *prefix); +void drm_mm_debug_table(struct drm_mm *mm, const char *prefix); #ifdef CONFIG_DEBUG_FS int drm_mm_dump_table(struct seq_file *m, struct drm_mm *mm); #endif From 69fa5293bf8d0ade3fd726848c7af925227e9180 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 01:28:49 +0100 Subject: [PATCH 197/259] drm/kms: rip out drm_mode_connector_detach_encoder It's only used by imx, and that one gets it wrong - there's no need to deteach the encoder before removing it. And really, neither current drm modesetting code nor all the userspace we have can handle dynamic changes in the set of possible encoders for a given connector. So let's just remove this before someone starts doing something really nasty with it. As a plus, one less kerneldoc comment to write. Cc: Sascha Hauer Cc: Russell King Cc: Greg Kroah-Hartman Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc.c | 15 --------------- drivers/staging/imx-drm/imx-ldb.c | 2 -- drivers/staging/imx-drm/parallel-display.c | 2 -- include/drm/drm_crtc.h | 2 -- 4 files changed, 21 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 35ea15d5ffff..ea620f4cf6c7 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -3506,21 +3506,6 @@ int drm_mode_connector_attach_encoder(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_attach_encoder); -void drm_mode_connector_detach_encoder(struct drm_connector *connector, - struct drm_encoder *encoder) -{ - int i; - for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) { - if (connector->encoder_ids[i] == encoder->base.id) { - connector->encoder_ids[i] = 0; - if (connector->encoder == encoder) - connector->encoder = NULL; - break; - } - } -} -EXPORT_SYMBOL(drm_mode_connector_detach_encoder); - int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, int gamma_size) { diff --git a/drivers/staging/imx-drm/imx-ldb.c b/drivers/staging/imx-drm/imx-ldb.c index 7e593296ac47..c703e986b44c 100644 --- a/drivers/staging/imx-drm/imx-ldb.c +++ b/drivers/staging/imx-drm/imx-ldb.c @@ -595,8 +595,6 @@ static int imx_ldb_remove(struct platform_device *pdev) struct drm_connector *connector = &channel->connector; struct drm_encoder *encoder = &channel->encoder; - drm_mode_connector_detach_encoder(connector, encoder); - imx_drm_remove_connector(channel->imx_drm_connector); imx_drm_remove_encoder(channel->imx_drm_encoder); } diff --git a/drivers/staging/imx-drm/parallel-display.c b/drivers/staging/imx-drm/parallel-display.c index 351d61dede00..823d015d2140 100644 --- a/drivers/staging/imx-drm/parallel-display.c +++ b/drivers/staging/imx-drm/parallel-display.c @@ -244,8 +244,6 @@ static int imx_pd_remove(struct platform_device *pdev) struct drm_connector *connector = &imxpd->connector; struct drm_encoder *encoder = &imxpd->encoder; - drm_mode_connector_detach_encoder(connector, encoder); - imx_drm_remove_connector(imxpd->imx_drm_connector); imx_drm_remove_encoder(imxpd->imx_drm_encoder); diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h index f7646548660d..44c8576ddbe3 100644 --- a/include/drm/drm_crtc.h +++ b/include/drm/drm_crtc.h @@ -1082,8 +1082,6 @@ extern const char *drm_get_encoder_name(const struct drm_encoder *encoder); extern int drm_mode_connector_attach_encoder(struct drm_connector *connector, struct drm_encoder *encoder); -extern void drm_mode_connector_detach_encoder(struct drm_connector *connector, - struct drm_encoder *encoder); extern int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, int gamma_size); extern struct drm_mode_object *drm_mode_object_find(struct drm_device *dev, From 3ec0db819315c765b3c7bbf7e9dee2fe1f186f47 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 15:06:15 +0100 Subject: [PATCH 198/259] drm/doc: Integrate drm_modes.c kerneldoc And clean it up so that there's no kerneldoc warnings. There's still a lot to do with this one here. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 4 ++++ drivers/gpu/drm/drm_modes.c | 28 +++++++++++++++++----------- 2 files changed, 21 insertions(+), 11 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index d68bb0a2dc06..50af3298ac1f 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -963,6 +963,10 @@ int max_width, max_height; Mode setting functions. + + Display Modes Function Reference +!Edrivers/gpu/drm/drm_modes.c + Frame Buffer Creation struct drm_framebuffer *(*fb_create)(struct drm_device *dev, diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index b0733153dfd2..4892194937f3 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -39,12 +39,11 @@ #include Display Modes Function Reference +!Iinclude/drm/drm_modes.h !Edrivers/gpu/drm/drm_modes.c diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index cc352eed0191..8b410576fce4 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -63,9 +63,10 @@ EXPORT_SYMBOL(drm_mode_debug_printmodeline); * drm_mode_create - create a new display mode * @dev: DRM device * - * Create a new drm_display_mode, give it an ID, and return it. + * Create a new, cleared drm_display_mode with kzalloc, allocate an ID for it + * and return it. * - * RETURNS: + * Returns: * Pointer to new mode on success, NULL on error. */ struct drm_display_mode *drm_mode_create(struct drm_device *dev) @@ -90,7 +91,7 @@ EXPORT_SYMBOL(drm_mode_create); * @dev: DRM device * @mode: mode to remove * - * Free @mode's unique identifier, then free it. + * Release @mode's unique ID, then free it @mode structure itself using kfree. */ void drm_mode_destroy(struct drm_device *dev, struct drm_display_mode *mode) { @@ -104,11 +105,13 @@ void drm_mode_destroy(struct drm_device *dev, struct drm_display_mode *mode) EXPORT_SYMBOL(drm_mode_destroy); /** - * drm_mode_probed_add - add a mode to a connector's probed mode list + * drm_mode_probed_add - add a mode to a connector's probed_mode list * @connector: connector the new mode * @mode: mode data * - * Add @mode to @connector's mode list for later use. + * Add @mode to @connector's probed_mode list for later use. This list should + * then in a second step get filtered and all the modes actually supported by + * the hardware moved to the @connector's modes list. */ void drm_mode_probed_add(struct drm_connector *connector, struct drm_display_mode *mode) @@ -120,16 +123,14 @@ void drm_mode_probed_add(struct drm_connector *connector, EXPORT_SYMBOL(drm_mode_probed_add); /** - * drm_cvt_mode -create a modeline based on CVT algorithm - * @dev: DRM device + * drm_cvt_mode -create a modeline based on the CVT algorithm + * @dev: drm device * @hdisplay: hdisplay size * @vdisplay: vdisplay size - * @vrefresh : vrefresh rate - * @reduced : Whether the GTF calculation is simplified - * @interlaced:Whether the interlace is supported - * @margins: whether to add margins or not - * - * return the modeline based on CVT algorithm + * @vrefresh: vrefresh rate + * @reduced: whether to use reduced blanking + * @interlaced: whether to compute an interlaced mode + * @margins: whether to add margins (borders) * * This function is called to generate the modeline based on CVT algorithm * according to the hdisplay, vdisplay, vrefresh. @@ -139,6 +140,11 @@ EXPORT_SYMBOL(drm_mode_probed_add); * * And it is copied from xf86CVTmode in xserver/hw/xfree86/modes/xf86cvt.c. * What I have done is to translate it by using integer calculation. + * + * Returns: + * The modeline based on the CVT algorithm stored in a drm_display_mode object. + * The display mode object is allocated with drm_mode_create(). Returns NULL + * when no mode could be allocated. */ struct drm_display_mode *drm_cvt_mode(struct drm_device *dev, int hdisplay, int vdisplay, int vrefresh, @@ -338,23 +344,25 @@ struct drm_display_mode *drm_cvt_mode(struct drm_device *dev, int hdisplay, EXPORT_SYMBOL(drm_cvt_mode); /** - * drm_gtf_mode_complex - create the modeline based on full GTF algorithm - * - * @dev :drm device - * @hdisplay :hdisplay size - * @vdisplay :vdisplay size - * @vrefresh :vrefresh rate. - * @interlaced :whether the interlace is supported - * @margins :desired margin size + * drm_gtf_mode_complex - create the modeline based on the full GTF algorithm + * @dev: drm device + * @hdisplay: hdisplay size + * @vdisplay: vdisplay size + * @vrefresh: vrefresh rate. + * @interlaced: whether to compute an interlaced mode + * @margins: desired margin (borders) size * @GTF_M: extended GTF formula parameters * @GTF_2C: extended GTF formula parameters * @GTF_K: extended GTF formula parameters * @GTF_2J: extended GTF formula parameters * - * return the modeline based on full GTF algorithm. - * * GTF feature blocks specify C and J in multiples of 0.5, so we pass them * in here multiplied by two. For a C of 40, pass in 80. + * + * Returns: + * The modeline based on the full GTF algorithm stored in a drm_display_mode object. + * The display mode object is allocated with drm_mode_create(). Returns NULL + * when no mode could be allocated. */ struct drm_display_mode * drm_gtf_mode_complex(struct drm_device *dev, int hdisplay, int vdisplay, @@ -524,14 +532,13 @@ drm_gtf_mode_complex(struct drm_device *dev, int hdisplay, int vdisplay, EXPORT_SYMBOL(drm_gtf_mode_complex); /** - * drm_gtf_mode - create the modeline based on GTF algorithm - * - * @dev :drm device - * @hdisplay :hdisplay size - * @vdisplay :vdisplay size - * @vrefresh :vrefresh rate. - * @interlaced :whether the interlace is supported - * @margins :whether the margin is supported + * drm_gtf_mode - create the modeline based on the GTF algorithm + * @dev: drm device + * @hdisplay: hdisplay size + * @vdisplay: vdisplay size + * @vrefresh: vrefresh rate. + * @interlaced: whether to compute an interlaced mode + * @margins: desired margin (borders) size * * return the modeline based on GTF algorithm * @@ -550,6 +557,11 @@ EXPORT_SYMBOL(drm_gtf_mode_complex); * C = 40 * K = 128 * J = 20 + * + * Returns: + * The modeline based on the GTF algorithm stored in a drm_display_mode object. + * The display mode object is allocated with drm_mode_create(). Returns NULL + * when no mode could be allocated. */ struct drm_display_mode * drm_gtf_mode(struct drm_device *dev, int hdisplay, int vdisplay, int vrefresh, @@ -562,6 +574,13 @@ drm_gtf_mode(struct drm_device *dev, int hdisplay, int vdisplay, int vrefresh, EXPORT_SYMBOL(drm_gtf_mode); #ifdef CONFIG_VIDEOMODE_HELPERS +/** + * drm_display_mode_from_videomode - fill in @dmode using @vm, + * @vm: videomode structure to use as source + * @dmode: drm_display_mode structure to use as destination + * + * Fills out @dmode using the display mode specified in @vm. + */ void drm_display_mode_from_videomode(const struct videomode *vm, struct drm_display_mode *dmode) { @@ -606,6 +625,9 @@ EXPORT_SYMBOL_GPL(drm_display_mode_from_videomode); * This function is expensive and should only be used, if only one mode is to be * read from DT. To get multiple modes start with of_get_display_timings and * work with that instead. + * + * Returns: + * 0 on success, a negative errno code when no of videomode node was found. */ int of_get_drm_display_mode(struct device_node *np, struct drm_display_mode *dmode, int index) @@ -633,7 +655,8 @@ EXPORT_SYMBOL_GPL(of_get_drm_display_mode); * drm_mode_set_name - set the name on a mode * @mode: name will be set in this mode * - * Set the name of @mode to a standard format. + * Set the name of @mode to a standard format which is x + * with an optional 'i' suffix for interlaced modes. */ void drm_mode_set_name(struct drm_display_mode *mode) { @@ -648,7 +671,9 @@ EXPORT_SYMBOL(drm_mode_set_name); /** drm_mode_hsync - get the hsync of a mode * @mode: mode * - * Return @modes's hsync rate in kHz, rounded to the nearest int. + * Returns: + * @modes's hsync rate in kHz, rounded to the nearest integer. Calculates the + * value first if it is not yet set. */ int drm_mode_hsync(const struct drm_display_mode *mode) { @@ -672,14 +697,9 @@ EXPORT_SYMBOL(drm_mode_hsync); * drm_mode_vrefresh - get the vrefresh of a mode * @mode: mode * - * Return @mode's vrefresh rate in Hz or calculate it if necessary. - * - * FIXME: why is this needed? shouldn't vrefresh be set already? - * - * RETURNS: - * Vertical refresh rate. It will be the result of actual value plus 0.5. - * If it is 70.288, it will return 70Hz. - * If it is 59.6, it will return 60Hz. + * Returns: + * @modes's vrefresh rate in Hz, rounded to the nearest integer. Calculates the + * value first if it is not yet set. */ int drm_mode_vrefresh(const struct drm_display_mode *mode) { @@ -708,11 +728,11 @@ int drm_mode_vrefresh(const struct drm_display_mode *mode) EXPORT_SYMBOL(drm_mode_vrefresh); /** - * drm_mode_set_crtcinfo - set CRTC modesetting parameters + * drm_mode_set_crtcinfo - set CRTC modesetting timing parameters * @p: mode * @adjust_flags: a combination of adjustment flags * - * Setup the CRTC modesetting parameters for @p, adjusting if necessary. + * Setup the CRTC modesetting timing parameters for @p, adjusting if necessary. * * - The CRTC_INTERLACE_HALVE_V flag can be used to halve vertical timings of * interlaced modes. @@ -780,7 +800,6 @@ void drm_mode_set_crtcinfo(struct drm_display_mode *p, int adjust_flags) } EXPORT_SYMBOL(drm_mode_set_crtcinfo); - /** * drm_mode_copy - copy the mode * @dst: mode to overwrite @@ -807,6 +826,9 @@ EXPORT_SYMBOL(drm_mode_copy); * * Just allocate a new mode, copy the existing mode into it, and return * a pointer to it. Used to create new instances of established modes. + * + * Returns: + * Pointer to duplicated mode on success, NULL on error. */ struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev, const struct drm_display_mode *mode) @@ -830,7 +852,7 @@ EXPORT_SYMBOL(drm_mode_duplicate); * * Check to see if @mode1 and @mode2 are equivalent. * - * RETURNS: + * Returns: * True if the modes are equal, false otherwise. */ bool drm_mode_equal(const struct drm_display_mode *mode1, const struct drm_display_mode *mode2) @@ -859,7 +881,7 @@ EXPORT_SYMBOL(drm_mode_equal); * Check to see if @mode1 and @mode2 are equivalent, but * don't check the pixel clocks nor the stereo layout. * - * RETURNS: + * Returns: * True if the modes are equal, false otherwise. */ bool drm_mode_equal_no_clocks_no_stereo(const struct drm_display_mode *mode1, @@ -890,9 +912,10 @@ EXPORT_SYMBOL(drm_mode_equal_no_clocks_no_stereo); * @maxX: maximum width * @maxY: maximum height * - * The DRM device (@dev) has size and pitch limits. Here we validate the - * modes we probed for @dev against those limits and set their status as - * necessary. + * This function is a helper which can be used to validate modes against size + * limitations of the DRM device/connector. If a mode is too big its status + * memeber is updated with the appropriate validation failure code. The list + * itself is not changed. */ void drm_mode_validate_size(struct drm_device *dev, struct list_head *mode_list, @@ -916,9 +939,10 @@ EXPORT_SYMBOL(drm_mode_validate_size); * @mode_list: list of modes to check * @verbose: be verbose about it * - * Once mode list generation is complete, a caller can use this routine to - * remove invalid modes from a mode list. If any of the modes have a - * status other than %MODE_OK, they are removed from @mode_list and freed. + * This helper function can be used to prune a display mode list after + * validation has been completed. All modes who's status is not MODE_OK will be + * removed from the list, and if @verbose the status code and mode name is also + * printed to dmesg. */ void drm_mode_prune_invalid(struct drm_device *dev, struct list_head *mode_list, bool verbose) @@ -948,7 +972,7 @@ EXPORT_SYMBOL(drm_mode_prune_invalid); * Compare two modes, given by @lh_a and @lh_b, returning a value indicating * which is better. * - * RETURNS: + * Returns: * Negative if @lh_a is better than @lh_b, zero if they're equivalent, or * positive if @lh_b is better than @lh_a. */ @@ -976,9 +1000,9 @@ static int drm_mode_compare(void *priv, struct list_head *lh_a, struct list_head /** * drm_mode_sort - sort mode list - * @mode_list: list to sort + * @mode_list: list of drm_display_mode structures to sort * - * Sort @mode_list by favorability, putting good modes first. + * Sort @mode_list by favorability, moving good modes to the head of the list. */ void drm_mode_sort(struct list_head *mode_list) { @@ -992,8 +1016,10 @@ EXPORT_SYMBOL(drm_mode_sort); * * This moves the modes from the @connector probed_modes list * to the actual mode list. It compares the probed mode against the current - * list and only adds different modes. All modes unverified after this point - * will be removed by the prune invalid modes. + * list and only adds different/new modes. + * + * This is just a helper functions doesn't validate any modes itself and also + * doesn't prune any invalid modes. Callers need to do that themselves. */ void drm_mode_connector_list_update(struct drm_connector *connector) { @@ -1028,18 +1054,25 @@ void drm_mode_connector_list_update(struct drm_connector *connector) EXPORT_SYMBOL(drm_mode_connector_list_update); /** - * drm_mode_parse_command_line_for_connector - parse command line for connector - * @mode_option: per connector mode option - * @connector: connector to parse line for - * @mode: preallocated mode structure to fill out + * drm_mode_parse_command_line_for_connector - parse command line modeline for connector + * @mode_option: optional per connector mode option + * @connector: connector to parse modeline for + * @mode: preallocated drm_cmdline_mode structure to fill out * - * This parses the connector specific then generic command lines for - * modes and options to configure the connector. + * This parses @mode_option command line modeline for modes and options to + * configure the connector. If @mode_option is NULL the default command line + * modeline in fb_mode_option will be parsed instead. + * + * This uses the same parameters as the fb modedb.c, except for an extra + * force-enable, force-enable-digital and force-disable bit at the end: * - * This uses the same parameters as the fb modedb.c, except for extra * x[M][R][-][@][i][m][eDd] * - * enable/enable Digital/disable bit at the end + * The intermediate drm_cmdline_mode structure is required to store additional + * options from the command line modline like the force-enabel/disable flag. + * + * Returns: + * True if a valid modeline has been parsed, false otherwise. */ bool drm_mode_parse_command_line_for_connector(const char *mode_option, struct drm_connector *connector, @@ -1192,6 +1225,14 @@ bool drm_mode_parse_command_line_for_connector(const char *mode_option, } EXPORT_SYMBOL(drm_mode_parse_command_line_for_connector); +/** + * drm_mode_create_from_cmdline_mode - convert a command line modeline into a DRM display mode + * @dev: DRM device to create the new mode for + * @cmd: input command line modeline + * + * Returns: + * Pointer to converted mode on success, NULL on error. + */ struct drm_display_mode * drm_mode_create_from_cmdline_mode(struct drm_device *dev, struct drm_cmdline_mode *cmd) diff --git a/include/drm/drm_modes.h b/include/drm/drm_modes.h index b3507f15d010..995c34d91ef1 100644 --- a/include/drm/drm_modes.h +++ b/include/drm/drm_modes.h @@ -162,6 +162,14 @@ struct drm_cmdline_mode { enum drm_connector_force force; }; +/** + * drm_mode_is_stereo - check for stereo mode flags + * @mode: drm_display_mode to check + * + * Returns: + * True if the mode is one of the stereo modes (like side-by-side), false if + * not. + */ static inline bool drm_mode_is_stereo(const struct drm_display_mode *mode) { return mode->flags & DRM_MODE_FLAG_3D_MASK; From fa54143f924ac49ff1a40d4d30452ff33097c236 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 21:34:33 +0100 Subject: [PATCH 206/259] drm: remove drm_display_mode->private_size It' unused and there's also not really any way to make it work with the current code. So better rip it out. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- include/drm/drm_modes.h | 1 - 1 file changed, 1 deletion(-) diff --git a/include/drm/drm_modes.h b/include/drm/drm_modes.h index 995c34d91ef1..2dbbf9976669 100644 --- a/include/drm/drm_modes.h +++ b/include/drm/drm_modes.h @@ -138,7 +138,6 @@ struct drm_display_mode { int crtc_vtotal; /* Driver private mode info */ - int private_size; int *private; int private_flags; From 9ee984a5f735d6afc6f889e179b2e4b1f2ec335f Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 21:57:37 +0100 Subject: [PATCH 207/259] drm/doc: Fix misplaced Oops. This is a regression from commit 5d7a951537927555fa1286a338e1b91c3b8b7445 Author: Daniel Vetter Date: Fri Jan 4 22:31:20 2013 +0100 drm/doc: updates for new framebuffer lifetime rules Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- Documentation/DocBook/drm.tmpl | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 4268cbe6f95c..9f5457ac0373 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -1060,7 +1060,7 @@ int max_width, max_height; The lifetime of a drm framebuffer is controlled with a reference count, drivers can grab additional references with - drm_framebuffer_reference and drop them + drm_framebuffer_referenceand drop them again with drm_framebuffer_unreference. For driver-private framebuffers for which the last reference is never dropped (e.g. for the fbdev framebuffer when the struct @@ -1068,6 +1068,7 @@ int max_width, max_height; helper struct) drivers can manually clean up a framebuffer at module unload time with drm_framebuffer_unregister_private. + Dumb Buffer Objects From 9fd93784f1719532d796914935f87cc1c6afd687 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 22:16:24 +0100 Subject: [PATCH 208/259] drm: remove return value from drm_helper_mode_fill_fb_struct Rightfully no driver ever checked this - it can't fail. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc_helper.c | 6 ++---- include/drm/drm_crtc_helper.h | 4 ++-- 2 files changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 08b3359c655f..ae2fd5cd8dfa 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -923,8 +923,8 @@ void drm_helper_connector_dpms(struct drm_connector *connector, int mode) } EXPORT_SYMBOL(drm_helper_connector_dpms); -int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, - struct drm_mode_fb_cmd2 *mode_cmd) +void drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, + struct drm_mode_fb_cmd2 *mode_cmd) { int i; @@ -937,8 +937,6 @@ int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, drm_fb_get_bpp_depth(mode_cmd->pixel_format, &fb->depth, &fb->bits_per_pixel); fb->pixel_format = mode_cmd->pixel_format; - - return 0; } EXPORT_SYMBOL(drm_helper_mode_fill_fb_struct); diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h index b1388b5fe7ac..b6c17980cd00 100644 --- a/include/drm/drm_crtc_helper.h +++ b/include/drm/drm_crtc_helper.h @@ -139,8 +139,8 @@ extern void drm_helper_connector_dpms(struct drm_connector *connector, int mode) extern void drm_helper_move_panel_connectors_to_head(struct drm_device *); -extern int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, - struct drm_mode_fb_cmd2 *mode_cmd); +extern void drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, + struct drm_mode_fb_cmd2 *mode_cmd); static inline void drm_crtc_helper_add(struct drm_crtc *crtc, const struct drm_crtc_helper_funcs *funcs) From 62ff94a5492175759546f8bc61383189d6b49122 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 22:18:47 +0100 Subject: [PATCH 209/259] drm/crtc-helper: remove LOCKING from kerneldoc - It yells. - WARNing about incorrect locking is harder to ignore, so better than kerneldoc. - Since those have been written per-crtc locks were added ... So remove them and replace them by appropriate WARNs. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc_helper.c | 31 ++++++++++++------------------- 1 file changed, 12 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index ae2fd5cd8dfa..44d50f56afa3 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -105,9 +105,6 @@ static void drm_mode_validate_flag(struct drm_connector *connector, * @maxX: max width for modes * @maxY: max height for modes * - * LOCKING: - * Caller must hold mode config lock. - * * Based on the helper callbacks implemented by @connector try to detect all * valid modes. Modes will first be added to the connector's probed_modes list, * then culled (based on validity and the @maxX, @maxY parameters) and put into @@ -131,6 +128,8 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, int mode_flags = 0; bool verbose_prune = true; + WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); + DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, drm_get_connector_name(connector)); /* set all modes to the unverified state */ @@ -218,9 +217,6 @@ EXPORT_SYMBOL(drm_helper_probe_single_connector_modes); * drm_helper_encoder_in_use - check if a given encoder is in use * @encoder: encoder to check * - * LOCKING: - * Caller must hold mode config lock. - * * Walk @encoders's DRM device's mode_config and see if it's in use. * * RETURNS: @@ -230,6 +226,8 @@ bool drm_helper_encoder_in_use(struct drm_encoder *encoder) { struct drm_connector *connector; struct drm_device *dev = encoder->dev; + + WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); list_for_each_entry(connector, &dev->mode_config.connector_list, head) if (connector->encoder == encoder) return true; @@ -241,9 +239,6 @@ EXPORT_SYMBOL(drm_helper_encoder_in_use); * drm_helper_crtc_in_use - check if a given CRTC is in a mode_config * @crtc: CRTC to check * - * LOCKING: - * Caller must hold mode config lock. - * * Walk @crtc's DRM device's mode_config and see if it's in use. * * RETURNS: @@ -253,7 +248,8 @@ bool drm_helper_crtc_in_use(struct drm_crtc *crtc) { struct drm_encoder *encoder; struct drm_device *dev = crtc->dev; - /* FIXME: Locking around list access? */ + + WARN_ON(!mutex_is_locked(&dev->mode_config.mutex)); list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) if (encoder->crtc == crtc && drm_helper_encoder_in_use(encoder)) return true; @@ -282,9 +278,6 @@ drm_encoder_disable(struct drm_encoder *encoder) * drm_helper_disable_unused_functions - disable unused objects * @dev: DRM device * - * LOCKING: - * Caller must hold mode config lock. - * * If an connector or CRTC isn't part of @dev's mode_config, it can be disabled * by calling its dpms function, which should power it off. */ @@ -294,6 +287,8 @@ void drm_helper_disable_unused_functions(struct drm_device *dev) struct drm_connector *connector; struct drm_crtc *crtc; + drm_warn_on_modeset_not_all_locked(dev); + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { if (!connector->encoder) continue; @@ -354,9 +349,6 @@ drm_crtc_prepare_encoders(struct drm_device *dev) * @y: vertical offset into the surface * @old_fb: old framebuffer, for cleanup * - * LOCKING: - * Caller must hold mode config lock. - * * Try to set @mode on @crtc. Give @crtc and its associated connectors a chance * to fixup or reject the mode prior to trying to set it. This is an internal * helper that drivers could e.g. use to update properties that require the @@ -383,6 +375,8 @@ bool drm_crtc_helper_set_mode(struct drm_crtc *crtc, struct drm_encoder *encoder; bool ret = true; + drm_warn_on_modeset_not_all_locked(dev); + saved_enabled = crtc->enabled; crtc->enabled = drm_helper_crtc_in_use(crtc); if (!crtc->enabled) @@ -559,9 +553,6 @@ drm_crtc_helper_disable(struct drm_crtc *crtc) * drm_crtc_helper_set_config - set a new config from userspace * @set: mode set configuration * - * LOCKING: - * Caller must hold mode config lock. - * * Setup a new configuration, provided by the upper layers (either an ioctl call * from userspace or internally e.g. from the fbdev suppport code) in @set, and * enable it. This is the main helper functions for drivers that implement @@ -611,6 +602,8 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) dev = set->crtc->dev; + drm_warn_on_modeset_not_all_locked(dev); + /* * Allocate space for the backup of all (non-pointer) encoder and * connector data. From 00d762cbd1fb5df63bf005ffa1c8d0275f79890e Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 22:28:30 +0100 Subject: [PATCH 210/259] drm: drop error code for drm_helper_resume_force_mode No driver cares, and it should generally work. Add a big comment when drivers can't use this for recompense. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc_helper.c | 19 ++++++++++++++++--- include/drm/drm_crtc_helper.h | 2 +- 2 files changed, 17 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 44d50f56afa3..0f60150adfc3 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -941,13 +941,25 @@ EXPORT_SYMBOL(drm_helper_mode_fill_fb_struct); * force-restore the mode setting configuration e.g. on resume or when something * else might have trampled over the hw state (like some overzealous old BIOSen * tended to do). + * + * This helper doesn't provide a error return value since restoring the old + * config should never fail due to resource allocation issues since the driver + * has successfully set the restored configuration already. Hence this should + * boil down to the equivalent of a few dpms on calls, which also don't provide + * an error code. + * + * Drivers where simply restoring an old configuration again might fail (e.g. + * due to slight differences in allocating shared resources when the + * configuration is restored in a different order than when userspace set it up) + * need to use their own restore logic. */ -int drm_helper_resume_force_mode(struct drm_device *dev) +void drm_helper_resume_force_mode(struct drm_device *dev) { struct drm_crtc *crtc; struct drm_encoder *encoder; struct drm_crtc_helper_funcs *crtc_funcs; - int ret, encoder_dpms; + int encoder_dpms; + bool ret; list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { @@ -957,6 +969,7 @@ int drm_helper_resume_force_mode(struct drm_device *dev) ret = drm_crtc_helper_set_mode(crtc, &crtc->mode, crtc->x, crtc->y, crtc->fb); + /* Restoring the old config should never fail! */ if (ret == false) DRM_ERROR("failed to set mode on crtc %p\n", crtc); @@ -979,9 +992,9 @@ int drm_helper_resume_force_mode(struct drm_device *dev) drm_helper_choose_crtc_dpms(crtc)); } } + /* disable the unused connectors while restoring the modesetting */ drm_helper_disable_unused_functions(dev); - return 0; } EXPORT_SYMBOL(drm_helper_resume_force_mode); diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h index b6c17980cd00..0bb34ca2ad2b 100644 --- a/include/drm/drm_crtc_helper.h +++ b/include/drm/drm_crtc_helper.h @@ -160,7 +160,7 @@ static inline void drm_connector_helper_add(struct drm_connector *connector, connector->helper_private = (void *)funcs; } -extern int drm_helper_resume_force_mode(struct drm_device *dev); +extern void drm_helper_resume_force_mode(struct drm_device *dev); extern void drm_kms_helper_poll_init(struct drm_device *dev); extern void drm_kms_helper_poll_fini(struct drm_device *dev); extern bool drm_helper_hpd_irq_event(struct drm_device *dev); From 3fcc42e07c60cd1c459fa2f7e8b2b84e61116ac9 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 23 Jan 2014 22:58:26 +0100 Subject: [PATCH 211/259] drm: kerneldoc polish for drm_crtc_helper.c Most of this is newly added kerneldoc for the hotplug and output polling code. But I've also thrown in a bit lesser polish, most of it is tuning down the shouting RETURN: headers. Overview documentation for the output probing and mode setting support code will be added in later patches. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc_helper.c | 125 ++++++++++++++++++++++++++---- 1 file changed, 111 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 0f60150adfc3..a85517854073 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -114,8 +114,8 @@ static void drm_mode_validate_flag(struct drm_connector *connector, * @connector vfunc for drivers that use the crtc helpers for output mode * filtering and detection. * - * RETURNS: - * Number of modes found on @connector. + * Returns: + * The number of modes found on @connector. */ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, uint32_t maxX, uint32_t maxY) @@ -217,10 +217,12 @@ EXPORT_SYMBOL(drm_helper_probe_single_connector_modes); * drm_helper_encoder_in_use - check if a given encoder is in use * @encoder: encoder to check * - * Walk @encoders's DRM device's mode_config and see if it's in use. + * Checks whether @encoder is with the current mode setting output configuration + * in use by any connector. This doesn't mean that it is actually enabled since + * the DPMS state is tracked separately. * - * RETURNS: - * True if @encoder is part of the mode_config, false otherwise. + * Returns: + * True if @encoder is used, false otherwise. */ bool drm_helper_encoder_in_use(struct drm_encoder *encoder) { @@ -239,10 +241,12 @@ EXPORT_SYMBOL(drm_helper_encoder_in_use); * drm_helper_crtc_in_use - check if a given CRTC is in a mode_config * @crtc: CRTC to check * - * Walk @crtc's DRM device's mode_config and see if it's in use. + * Checks whether @crtc is with the current mode setting output configuration + * in use by any connector. This doesn't mean that it is actually enabled since + * the DPMS state is tracked separately. * - * RETURNS: - * True if @crtc is part of the mode_config, false otherwise. + * Returns: + * True if @crtc is used, false otherwise. */ bool drm_helper_crtc_in_use(struct drm_crtc *crtc) { @@ -278,8 +282,11 @@ drm_encoder_disable(struct drm_encoder *encoder) * drm_helper_disable_unused_functions - disable unused objects * @dev: DRM device * - * If an connector or CRTC isn't part of @dev's mode_config, it can be disabled - * by calling its dpms function, which should power it off. + * This function walks through the entire mode setting configuration of @dev. It + * will remove any crtc links of unused encoders and encoder links of + * disconnected connectors. Then it will disable all unused encoders and crtcs + * either by calling their disable callback if available or by calling their + * dpms callback with DRM_MODE_DPMS_OFF. */ void drm_helper_disable_unused_functions(struct drm_device *dev) { @@ -358,8 +365,8 @@ drm_crtc_prepare_encoders(struct drm_device *dev) * drm_crtc_helper_set_config() helper function to drive the mode setting * sequence. * - * RETURNS: - * True if the mode was set successfully, or false otherwise. + * Returns: + * True if the mode was set successfully, false otherwise. */ bool drm_crtc_helper_set_mode(struct drm_crtc *crtc, struct drm_display_mode *mode, @@ -559,8 +566,8 @@ drm_crtc_helper_disable(struct drm_crtc *crtc) * kernel mode setting with the crtc helper functions and the assorted * ->prepare(), ->modeset() and ->commit() helper callbacks. * - * RETURNS: - * Returns 0 on success, -ERRNO on failure. + * Returns: + * Returns 0 on success, negative errno numbers on failure. */ int drm_crtc_helper_set_config(struct drm_mode_set *set) { @@ -916,6 +923,14 @@ void drm_helper_connector_dpms(struct drm_connector *connector, int mode) } EXPORT_SYMBOL(drm_helper_connector_dpms); +/** + * drm_helper_mode_fill_fb_struct - fill out framebuffer metadata + * @fb: drm_framebuffer object to fill out + * @mode_cmd: metadata from the userspace fb creation request + * + * This helper can be used in a drivers fb_create callback to pre-fill the fb's + * metadata fields. + */ void drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, struct drm_mode_fb_cmd2 *mode_cmd) { @@ -998,6 +1013,22 @@ void drm_helper_resume_force_mode(struct drm_device *dev) } EXPORT_SYMBOL(drm_helper_resume_force_mode); +/** + * drm_kms_helper_hotplug_event - fire off KMS hotplug events + * @dev: drm_device whose connector state changed + * + * This function fires off the uevent for userspace and also calls the + * output_poll_changed function, which is most commonly used to inform the fbdev + * emulation code and allow it to update the fbcon output configuration. + * + * Drivers should call this from their hotplug handling code when a change is + * detected. Note that this function does not do any output detection of its + * own, like drm_helper_hpd_irq_event() does - this is assumed to be done by the + * driver already. + * + * This function must be called from process context with no mode + * setting locks held. + */ void drm_kms_helper_hotplug_event(struct drm_device *dev) { /* send a uevent + call fbdev */ @@ -1066,6 +1097,16 @@ static void output_poll_execute(struct work_struct *work) schedule_delayed_work(delayed_work, DRM_OUTPUT_POLL_PERIOD); } +/** + * drm_kms_helper_poll_disable - disable output polling + * @dev: drm_device + * + * This function disables the output polling work. + * + * Drivers can call this helper from their device suspend implementation. It is + * not an error to call this even when output polling isn't enabled or arlready + * disabled. + */ void drm_kms_helper_poll_disable(struct drm_device *dev) { if (!dev->mode_config.poll_enabled) @@ -1074,6 +1115,16 @@ void drm_kms_helper_poll_disable(struct drm_device *dev) } EXPORT_SYMBOL(drm_kms_helper_poll_disable); +/** + * drm_kms_helper_poll_enable - re-enable output polling. + * @dev: drm_device + * + * This function re-enables the output polling work. + * + * Drivers can call this helper from their device resume implementation. It is + * an error to call this when the output polling support has not yet been set + * up. + */ void drm_kms_helper_poll_enable(struct drm_device *dev) { bool poll = false; @@ -1093,6 +1144,25 @@ void drm_kms_helper_poll_enable(struct drm_device *dev) } EXPORT_SYMBOL(drm_kms_helper_poll_enable); +/** + * drm_kms_helper_poll_init - initialize and enable output polling + * @dev: drm_device + * + * This function intializes and then also enables output polling support for + * @dev. Drivers which do not have reliable hotplug support in hardware can use + * this helper infrastructure to regularly poll such connectors for changes in + * their connection state. + * + * Drivers can control which connectors are polled by setting the + * DRM_CONNECTOR_POLL_CONNECT and DRM_CONNECTOR_POLL_DISCONNECT flags. On + * connectors where probing live outputs can result in visual distortion drivers + * should not set the DRM_CONNECTOR_POLL_DISCONNECT flag to avoid this. + * Connectors which have no flag or only DRM_CONNECTOR_POLL_HPD set are + * completely ignored by the polling logic. + * + * Note that a connector can be both polled and probed from the hotplug handler, + * in case the hotplug interrupt is known to be unreliable. + */ void drm_kms_helper_poll_init(struct drm_device *dev) { INIT_DELAYED_WORK(&dev->mode_config.output_poll_work, output_poll_execute); @@ -1102,12 +1172,39 @@ void drm_kms_helper_poll_init(struct drm_device *dev) } EXPORT_SYMBOL(drm_kms_helper_poll_init); +/** + * drm_kms_helper_poll_fini - disable output polling and clean it up + * @dev: drm_device + */ void drm_kms_helper_poll_fini(struct drm_device *dev) { drm_kms_helper_poll_disable(dev); } EXPORT_SYMBOL(drm_kms_helper_poll_fini); +/** + * drm_helper_hpd_irq_event - hotplug processing + * @dev: drm_device + * + * Drivers can use this helper function to run a detect cycle on all connectors + * which have the DRM_CONNECTOR_POLL_HPD flag set in their &polled member. All + * other connectors are ignored, which is useful to avoid reprobing fixed + * panels. + * + * This helper function is useful for drivers which can't or don't track hotplug + * interrupts for each connector. + * + * Drivers which support hotplug interrupts for each connector individually and + * which have a more fine-grained detect logic should bypass this code and + * directly call drm_kms_helper_hotplug_event() in case the connector state + * changed. + * + * This function must be called from process context with no mode + * setting locks held. + * + * Note that a connector can be both polled and probed from the hotplug handler, + * in case the hotplug interrupt is known to be unreliable. + */ bool drm_helper_hpd_irq_event(struct drm_device *dev) { struct drm_connector *connector; From c8e32cc1219fc15135b696b726421571f68bd97e Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Mon, 10 Mar 2014 21:33:02 +0100 Subject: [PATCH 212/259] drm: kerneldoc polish for drm_crtc.c - Standardized on "Returns:" Block. - Sprinkle missing kerneldoc over all exported functions and all ioctls. - Add a stern warning that driver's really shouldn't use drm_mode_group_init_legacy_group. - Usual attempt at more consistency. - Add warnings that drm_mode_object_get/put don't do refcounting, despite what the names might lead to believe. - Try to clarify the framebuffer setup/cleanup functions wrt driver private framebuffers - I've fallen recently over this when reviewing i915 fbdev patches. - Align function parameters where the kerneldoc has been updated. - Most of the drm_get_*_name functions aren't thread safe. Add stern warnings where this is the case. Since a lot of the functions in drm_crtc.c are boilerplate to handle properties and create default sets of them it might be useful to extract all that code into a new file drm_property.c. Especially since properties will be used a lot more in the future. Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_crtc.c | 614 ++++++++++++++++++++++++++++++++++--- 1 file changed, 570 insertions(+), 44 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 4c2367181f3d..91d03e3acefb 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -45,7 +45,8 @@ * @dev: drm device * * This function takes all modeset locks, suitable where a more fine-grained - * scheme isn't (yet) implemented. + * scheme isn't (yet) implemented. Locks must be dropped with + * drm_modeset_unlock_all. */ void drm_modeset_lock_all(struct drm_device *dev) { @@ -61,6 +62,8 @@ EXPORT_SYMBOL(drm_modeset_lock_all); /** * drm_modeset_unlock_all - drop all modeset locks * @dev: device + * + * This function drop all modeset locks taken by drm_modeset_lock_all. */ void drm_modeset_unlock_all(struct drm_device *dev) { @@ -76,6 +79,8 @@ EXPORT_SYMBOL(drm_modeset_unlock_all); /** * drm_warn_on_modeset_not_all_locked - check that all modeset locks are locked * @dev: device + * + * Useful as a debug assert. */ void drm_warn_on_modeset_not_all_locked(struct drm_device *dev) { @@ -243,6 +248,15 @@ void drm_connector_ida_destroy(void) ida_destroy(&drm_connector_enum_list[i].ida); } +/** + * drm_get_encoder_name - return a string for encoder + * @encoder: encoder to compute name of + * + * Note that the buffer used by this function is globally shared and owned by + * the function itself. + * + * FIXME: This isn't really multithreading safe. + */ const char *drm_get_encoder_name(const struct drm_encoder *encoder) { static char buf[32]; @@ -254,6 +268,15 @@ const char *drm_get_encoder_name(const struct drm_encoder *encoder) } EXPORT_SYMBOL(drm_get_encoder_name); +/** + * drm_get_connector_name - return a string for connector + * @connector: connector to compute name of + * + * Note that the buffer used by this function is globally shared and owned by + * the function itself. + * + * FIXME: This isn't really multithreading safe. + */ const char *drm_get_connector_name(const struct drm_connector *connector) { static char buf[32]; @@ -265,6 +288,13 @@ const char *drm_get_connector_name(const struct drm_connector *connector) } EXPORT_SYMBOL(drm_get_connector_name); +/** + * drm_get_connector_status_name - return a string for connector status + * @status: connector status to compute name of + * + * In contrast to the other drm_get_*_name functions this one here returns a + * const pointer and hence is threadsafe. + */ const char *drm_get_connector_status_name(enum drm_connector_status status) { if (status == connector_status_connected) @@ -294,6 +324,15 @@ static char printable_char(int c) return isascii(c) && isprint(c) ? c : '?'; } +/** + * drm_get_format_name - return a string for drm fourcc format + * @format: format to compute name of + * + * Note that the buffer used by this function is globally shared and owned by + * the function itself. + * + * FIXME: This isn't really multithreading safe. + */ const char *drm_get_format_name(uint32_t format) { static char buf[32]; @@ -318,9 +357,11 @@ EXPORT_SYMBOL(drm_get_format_name); * @obj_type: object type * * Create a unique identifier based on @ptr in @dev's identifier space. Used - * for tracking modes, CRTCs and connectors. + * for tracking modes, CRTCs and connectors. Note that despite the _get postfix + * modeset identifiers are _not_ reference counted. Hence don't use this for + * reference counted modeset objects like framebuffers. * - * RETURNS: + * Returns: * New unique (relative to other objects in @dev) integer identifier for the * object. */ @@ -349,7 +390,9 @@ int drm_mode_object_get(struct drm_device *dev, * @dev: DRM device * @object: object to free * - * Free @id from @dev's unique identifier pool. + * Free @id from @dev's unique identifier pool. Note that despite the _get + * postfix modeset identifiers are _not_ reference counted. Hence don't use this + * for reference counted modeset objects like framebuffers. */ void drm_mode_object_put(struct drm_device *dev, struct drm_mode_object *object) @@ -402,7 +445,7 @@ EXPORT_SYMBOL(drm_mode_object_find); * since all the fb attributes are invariant over its lifetime, no further * locking but only correct reference counting is required. * - * RETURNS: + * Returns: * Zero on success, error code on failure. */ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb, @@ -463,7 +506,7 @@ static struct drm_framebuffer *__drm_framebuffer_lookup(struct drm_device *dev, * * If successful, this grabs an additional reference to the framebuffer - * callers need to make sure to eventually unreference the returned framebuffer - * again. + * again, using @drm_framebuffer_unreference. */ struct drm_framebuffer *drm_framebuffer_lookup(struct drm_device *dev, uint32_t id) @@ -496,6 +539,8 @@ EXPORT_SYMBOL(drm_framebuffer_unreference); /** * drm_framebuffer_reference - incr the fb refcnt * @fb: framebuffer + * + * This functions increments the fb's refcount. */ void drm_framebuffer_reference(struct drm_framebuffer *fb) { @@ -552,8 +597,9 @@ EXPORT_SYMBOL(drm_framebuffer_unregister_private); * drm_framebuffer_cleanup - remove a framebuffer object * @fb: framebuffer to remove * - * Cleanup references to a user-created framebuffer. This function is intended - * to be used from the drivers ->destroy callback. + * Cleanup framebuffer. This function is intended to be used from the drivers + * ->destroy callback. It can also be used to clean up driver private + * framebuffers embedded into a larger structure. * * Note that this function does not remove the fb from active usuage - if it is * still used anywhere, hilarity can ensue since userspace could call getfb on @@ -646,7 +692,7 @@ EXPORT_SYMBOL(drm_framebuffer_remove); * * Inits a new object created as base part of a driver crtc object. * - * RETURNS: + * Returns: * Zero on success, error code on failure. */ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc, @@ -746,7 +792,7 @@ static void drm_mode_remove(struct drm_connector *connector, * Initialises a preallocated connector. Connectors should be * subclassed as part of driver connector objects. * - * RETURNS: + * Returns: * Zero on success, error code on failure. */ int drm_connector_init(struct drm_device *dev, @@ -824,6 +870,14 @@ void drm_connector_cleanup(struct drm_connector *connector) } EXPORT_SYMBOL(drm_connector_cleanup); +/** + * drm_connector_unplug_all - unregister connector userspace interfaces + * @dev: drm device + * + * This function unregisters all connector userspace interfaces in sysfs. Should + * be call when the device is disconnected, e.g. from an usb driver's + * ->disconnect callback. + */ void drm_connector_unplug_all(struct drm_device *dev) { struct drm_connector *connector; @@ -835,6 +889,18 @@ void drm_connector_unplug_all(struct drm_device *dev) } EXPORT_SYMBOL(drm_connector_unplug_all); +/** + * drm_bridge_init - initialize a drm transcoder/bridge + * @dev: drm device + * @bridge: transcoder/bridge to set up + * @funcs: bridge function table + * + * Initialises a preallocated bridge. Bridges should be + * subclassed as part of driver connector objects. + * + * Returns: + * Zero on success, error code on failure. + */ int drm_bridge_init(struct drm_device *dev, struct drm_bridge *bridge, const struct drm_bridge_funcs *funcs) { @@ -858,6 +924,12 @@ int drm_bridge_init(struct drm_device *dev, struct drm_bridge *bridge, } EXPORT_SYMBOL(drm_bridge_init); +/** + * drm_bridge_cleanup - cleans up an initialised bridge + * @bridge: bridge to cleanup + * + * Cleans up the bridge but doesn't free the object. + */ void drm_bridge_cleanup(struct drm_bridge *bridge) { struct drm_device *dev = bridge->dev; @@ -870,6 +942,19 @@ void drm_bridge_cleanup(struct drm_bridge *bridge) } EXPORT_SYMBOL(drm_bridge_cleanup); +/** + * drm_encoder_init - Init a preallocated encoder + * @dev: drm device + * @encoder: the encoder to init + * @funcs: callbacks for this encoder + * @encoder_type: user visible type of the encoder + * + * Initialises a preallocated encoder. Encoder should be + * subclassed as part of driver encoder objects. + * + * Returns: + * Zero on success, error code on failure. + */ int drm_encoder_init(struct drm_device *dev, struct drm_encoder *encoder, const struct drm_encoder_funcs *funcs, @@ -897,6 +982,12 @@ int drm_encoder_init(struct drm_device *dev, } EXPORT_SYMBOL(drm_encoder_init); +/** + * drm_encoder_cleanup - cleans up an initialised encoder + * @encoder: encoder to cleanup + * + * Cleans up the encoder but doesn't free the object. + */ void drm_encoder_cleanup(struct drm_encoder *encoder) { struct drm_device *dev = encoder->dev; @@ -918,9 +1009,10 @@ EXPORT_SYMBOL(drm_encoder_cleanup); * @format_count: number of elements in @formats * @priv: plane is private (hidden from userspace)? * - * Inits a new object created as base part of a driver plane object. + * Inits a preallocate plane object created as base part of a driver plane + * object. * - * RETURNS: + * Returns: * Zero on success, error code on failure. */ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane, @@ -1224,6 +1316,10 @@ static int drm_mode_group_init(struct drm_device *dev, struct drm_mode_group *gr return 0; } +/* + * NOTE: Driver's shouldn't ever call drm_mode_group_init_legacy_group - it is + * the drm core's responsibility to set up mode control groups. + */ int drm_mode_group_init_legacy_group(struct drm_device *dev, struct drm_mode_group *group) { @@ -1300,7 +1396,7 @@ static void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out, * Convert a drm_mode_modeinfo into a drm_display_mode structure to return to * the caller. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ static int drm_crtc_convert_umode(struct drm_display_mode *out, @@ -1343,7 +1439,7 @@ static int drm_crtc_convert_umode(struct drm_display_mode *out, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_getresources(struct drm_device *dev, void *data, @@ -1528,7 +1624,7 @@ int drm_mode_getresources(struct drm_device *dev, void *data, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_getcrtc(struct drm_device *dev, @@ -1597,7 +1693,7 @@ static bool drm_mode_expose_to_userspace(const struct drm_display_mode *mode, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_getconnector(struct drm_device *dev, void *data, @@ -1732,6 +1828,19 @@ int drm_mode_getconnector(struct drm_device *dev, void *data, return ret; } +/** + * drm_mode_getencoder - get encoder configuration + * @dev: drm device for the ioctl + * @data: data pointer for the ioctl + * @file_priv: drm file for the ioctl call + * + * Construct a encoder configuration structure to return to the user. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_getencoder(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -1767,15 +1876,20 @@ int drm_mode_getencoder(struct drm_device *dev, void *data, } /** - * drm_mode_getplane_res - get plane info + * drm_mode_getplane_res - enumerate all plane resources * @dev: DRM device * @data: ioctl data * @file_priv: DRM file info * - * Return an plane count and set of IDs. + * Construct a list of plane ids to return to the user. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. */ int drm_mode_getplane_res(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_mode_get_plane_res *plane_resp = data; struct drm_mode_config *config; @@ -1813,16 +1927,20 @@ int drm_mode_getplane_res(struct drm_device *dev, void *data, } /** - * drm_mode_getplane - get plane info + * drm_mode_getplane - get plane configuration * @dev: DRM device * @data: ioctl data * @file_priv: DRM file info * - * Return plane info, including formats supported, gamma size, any - * current fb, etc. + * Construct a plane configuration structure to return to the user. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. */ int drm_mode_getplane(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_mode_get_plane *plane_resp = data; struct drm_mode_object *obj; @@ -1878,16 +1996,19 @@ int drm_mode_getplane(struct drm_device *dev, void *data, } /** - * drm_mode_setplane - set up or tear down an plane + * drm_mode_setplane - configure a plane's configuration * @dev: DRM device * @data: ioctl data* * @file_priv: DRM file info * - * Set plane info, including placement, fb, scaling, and other factors. + * Set plane configuration, including placement, fb, scaling, and other factors. * Or pass a NULL fb to disable. + * + * Returns: + * Zero on success, errno on failure. */ int drm_mode_setplane(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_mode_set_plane *plane_req = data; struct drm_mode_object *obj; @@ -2017,6 +2138,9 @@ int drm_mode_setplane(struct drm_device *dev, void *data, * * This is a little helper to wrap internal calls to the ->set_config driver * interface. The only thing it adds is correct refcounting dance. + * + * Returns: + * Zero on success, errno on failure. */ int drm_mode_set_config_internal(struct drm_mode_set *set) { @@ -2101,7 +2225,7 @@ static int drm_crtc_check_viewport(const struct drm_crtc *crtc, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_setcrtc(struct drm_device *dev, void *data, @@ -2303,8 +2427,23 @@ static int drm_mode_cursor_common(struct drm_device *dev, return ret; } + + +/** + * drm_mode_cursor_ioctl - set CRTC's cursor configuration + * @dev: drm device for the ioctl + * @data: data pointer for the ioctl + * @file_priv: drm file for the ioctl call + * + * Set the cursor configuration based on user request. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_cursor_ioctl(struct drm_device *dev, - void *data, struct drm_file *file_priv) + void *data, struct drm_file *file_priv) { struct drm_mode_cursor *req = data; struct drm_mode_cursor2 new_req; @@ -2315,6 +2454,21 @@ int drm_mode_cursor_ioctl(struct drm_device *dev, return drm_mode_cursor_common(dev, &new_req, file_priv); } +/** + * drm_mode_cursor2_ioctl - set CRTC's cursor configuration + * @dev: drm device for the ioctl + * @data: data pointer for the ioctl + * @file_priv: drm file for the ioctl call + * + * Set the cursor configuration based on user request. This implements the 2nd + * version of the cursor ioctl, which allows userspace to additionally specify + * the hotspot of the pointer. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_cursor2_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -2322,7 +2476,14 @@ int drm_mode_cursor2_ioctl(struct drm_device *dev, return drm_mode_cursor_common(dev, req, file_priv); } -/* Original addfb only supported RGB formats, so figure out which one */ +/** + * drm_mode_legacy_fb_format - compute drm fourcc code from legacy description + * @bpp: bits per pixels + * @depth: bit depth per pixel + * + * Computes a drm fourcc pixel format code for the given @bpp/@depth values. + * Useful in fbdev emulation code, since that deals in those values. + */ uint32_t drm_mode_legacy_fb_format(uint32_t bpp, uint32_t depth) { uint32_t fmt; @@ -2364,11 +2525,12 @@ EXPORT_SYMBOL(drm_mode_legacy_fb_format); * @data: data pointer for the ioctl * @file_priv: drm file for the ioctl call * - * Add a new FB to the specified CRTC, given a user request. + * Add a new FB to the specified CRTC, given a user request. This is the + * original addfb ioclt which only supported RGB formats. * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_addfb(struct drm_device *dev, @@ -2541,11 +2703,13 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r) * @data: data pointer for the ioctl * @file_priv: drm file for the ioctl call * - * Add a new FB to the specified CRTC, given a user request with format. + * Add a new FB to the specified CRTC, given a user request with format. This is + * the 2nd version of the addfb ioctl, which supports multi-planar framebuffers + * and uses fourcc codes as pixel format specifiers. * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_addfb2(struct drm_device *dev, @@ -2605,7 +2769,7 @@ int drm_mode_addfb2(struct drm_device *dev, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_rmfb(struct drm_device *dev, @@ -2659,7 +2823,7 @@ int drm_mode_rmfb(struct drm_device *dev, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ int drm_mode_getfb(struct drm_device *dev, @@ -2703,6 +2867,25 @@ int drm_mode_getfb(struct drm_device *dev, return ret; } +/** + * drm_mode_dirtyfb_ioctl - flush frontbuffer rendering on an FB + * @dev: drm device for the ioctl + * @data: data pointer for the ioctl + * @file_priv: drm file for the ioctl call + * + * Lookup the FB and flush out the damaged area supplied by userspace as a clip + * rectangle list. Generic userspace which does frontbuffer rendering must call + * this ioctl to flush out the changes on manual-update display outputs, e.g. + * usb display-link, mipi manual update panels or edp panel self refresh modes. + * + * Modesetting drivers which always update the frontbuffer do not need to + * implement the corresponding ->dirty framebuffer callback. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_dirtyfb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -2780,7 +2963,7 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev, * * Called by the user via ioctl. * - * RETURNS: + * Returns: * Zero on success, errno on failure. */ void drm_fb_release(struct drm_file *priv) @@ -2804,6 +2987,20 @@ void drm_fb_release(struct drm_file *priv) mutex_unlock(&priv->fbs_lock); } +/** + * drm_property_create - create a new property type + * @dev: drm device + * @flags: flags specifying the property type + * @name: name of the property + * @num_values: number of pre-defined values + * + * This creates a new generic drm property which can then be attached to a drm + * object with drm_object_attach_property. The returned property object must be + * freed with drm_property_destroy. + * + * Returns: + * A pointer to the newly created property on success, NULL on failure. + */ struct drm_property *drm_property_create(struct drm_device *dev, int flags, const char *name, int num_values) { @@ -2842,6 +3039,24 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags, } EXPORT_SYMBOL(drm_property_create); +/** + * drm_property_create - create a new enumeration property type + * @dev: drm device + * @flags: flags specifying the property type + * @name: name of the property + * @props: enumeration lists with property values + * @num_values: number of pre-defined values + * + * This creates a new generic drm property which can then be attached to a drm + * object with drm_object_attach_property. The returned property object must be + * freed with drm_property_destroy. + * + * Userspace is only allowed to set one of the predefined values for enumeration + * properties. + * + * Returns: + * A pointer to the newly created property on success, NULL on failure. + */ struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags, const char *name, const struct drm_prop_enum_list *props, @@ -2870,6 +3085,24 @@ struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags, } EXPORT_SYMBOL(drm_property_create_enum); +/** + * drm_property_create - create a new bitmask property type + * @dev: drm device + * @flags: flags specifying the property type + * @name: name of the property + * @props: enumeration lists with property bitflags + * @num_values: number of pre-defined values + * + * This creates a new generic drm property which can then be attached to a drm + * object with drm_object_attach_property. The returned property object must be + * freed with drm_property_destroy. + * + * Compared to plain enumeration properties userspace is allowed to set any + * or'ed together combination of the predefined property bitflag values + * + * Returns: + * A pointer to the newly created property on success, NULL on failure. + */ struct drm_property *drm_property_create_bitmask(struct drm_device *dev, int flags, const char *name, const struct drm_prop_enum_list *props, @@ -2898,6 +3131,24 @@ struct drm_property *drm_property_create_bitmask(struct drm_device *dev, } EXPORT_SYMBOL(drm_property_create_bitmask); +/** + * drm_property_create - create a new ranged property type + * @dev: drm device + * @flags: flags specifying the property type + * @name: name of the property + * @min: minimum value of the property + * @max: maximum value of the property + * + * This creates a new generic drm property which can then be attached to a drm + * object with drm_object_attach_property. The returned property object must be + * freed with drm_property_destroy. + * + * Userspace is allowed to set any interger value in the (min, max) range + * inclusive. + * + * Returns: + * A pointer to the newly created property on success, NULL on failure. + */ struct drm_property *drm_property_create_range(struct drm_device *dev, int flags, const char *name, uint64_t min, uint64_t max) @@ -2917,6 +3168,21 @@ struct drm_property *drm_property_create_range(struct drm_device *dev, int flags } EXPORT_SYMBOL(drm_property_create_range); +/** + * drm_property_add_enum - add a possible value to an enumeration property + * @property: enumeration property to change + * @index: index of the new enumeration + * @value: value of the new enumeration + * @name: symbolic name of the new enumeration + * + * This functions adds enumerations to a property. + * + * It's use is deprecated, drivers should use one of the more specific helpers + * to directly create the property with all enumerations already attached. + * + * Returns: + * Zero on success, error code on failure. + */ int drm_property_add_enum(struct drm_property *property, int index, uint64_t value, const char *name) { @@ -2956,6 +3222,14 @@ int drm_property_add_enum(struct drm_property *property, int index, } EXPORT_SYMBOL(drm_property_add_enum); +/** + * drm_property_destroy - destroy a drm property + * @dev: drm device + * @property: property to destry + * + * This function frees a property including any attached resources like + * enumeration values. + */ void drm_property_destroy(struct drm_device *dev, struct drm_property *property) { struct drm_property_enum *prop_enum, *pt; @@ -2973,6 +3247,16 @@ void drm_property_destroy(struct drm_device *dev, struct drm_property *property) } EXPORT_SYMBOL(drm_property_destroy); +/** + * drm_object_attach_property - attach a property to a modeset object + * @obj: drm modeset object + * @property: property to attach + * @init_val: initial value of the property + * + * This attaches the given property to the modeset object with the given initial + * value. Currently this function cannot fail since the properties are stored in + * a statically sized array. + */ void drm_object_attach_property(struct drm_mode_object *obj, struct drm_property *property, uint64_t init_val) @@ -2993,6 +3277,19 @@ void drm_object_attach_property(struct drm_mode_object *obj, } EXPORT_SYMBOL(drm_object_attach_property); +/** + * drm_object_property_set_value - set the value of a property + * @obj: drm mode object to set property value for + * @property: property to set + * @val: value the property should be set to + * + * This functions sets a given property on a given object. This function only + * changes the software state of the property, it does not call into the + * driver's ->set_property callback. + * + * Returns: + * Zero on success, error code on failure. + */ int drm_object_property_set_value(struct drm_mode_object *obj, struct drm_property *property, uint64_t val) { @@ -3009,6 +3306,20 @@ int drm_object_property_set_value(struct drm_mode_object *obj, } EXPORT_SYMBOL(drm_object_property_set_value); +/** + * drm_object_property_get_value - retrieve the value of a property + * @obj: drm mode object to get property value from + * @property: property to retrieve + * @val: storage for the property value + * + * This function retrieves the softare state of the given property for the given + * property. Since there is no driver callback to retrieve the current property + * value this might be out of sync with the hardware, depending upon the driver + * and property. + * + * Returns: + * Zero on success, error code on failure. + */ int drm_object_property_get_value(struct drm_mode_object *obj, struct drm_property *property, uint64_t *val) { @@ -3025,6 +3336,19 @@ int drm_object_property_get_value(struct drm_mode_object *obj, } EXPORT_SYMBOL(drm_object_property_get_value); +/** + * drm_mode_getproperty_ioctl - get the current value of a connector's property + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This function retrieves the current value for an connectors's property. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_getproperty_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3163,6 +3487,20 @@ static void drm_property_destroy_blob(struct drm_device *dev, kfree(blob); } +/** + * drm_mode_getblob_ioctl - get the contents of a blob property value + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This function retrieves the contents of a blob property. The value stored in + * an object's blob property is just a normal modeset object id. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_getblob_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3197,6 +3535,17 @@ int drm_mode_getblob_ioctl(struct drm_device *dev, return ret; } +/** + * drm_mode_connector_update_edid_property - update the edid property of a connector + * @connector: drm connector + * @edid: new value of the edid property + * + * This function creates a new blob modeset object and assigns its id to the + * connector's edid property. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_connector_update_edid_property(struct drm_connector *connector, struct edid *edid) { @@ -3254,6 +3603,20 @@ static bool drm_property_change_is_valid(struct drm_property *property, } } +/** + * drm_mode_connector_property_set_ioctl - set the current value of a connector property + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This function sets the current value for a connectors's property. It also + * calls into a driver's ->set_property callback to update the hardware state + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_connector_property_set_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3320,6 +3683,21 @@ static int drm_mode_plane_set_obj_prop(struct drm_mode_object *obj, return ret; } +/** + * drm_mode_getproperty_ioctl - get the current value of a object's property + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This function retrieves the current value for an object's property. Compared + * to the connector specific ioctl this one is extended to also work on crtc and + * plane objects. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3376,6 +3754,22 @@ int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data, return ret; } +/** + * drm_mode_obj_set_property_ioctl - set the current value of an object's property + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This function sets the current value for an object's property. It also calls + * into a driver's ->set_property callback to update the hardware state. + * Compared to the connector specific ioctl this one is extended to also work on + * crtc and plane objects. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3435,6 +3829,18 @@ int drm_mode_obj_set_property_ioctl(struct drm_device *dev, void *data, return ret; } +/** + * drm_mode_connector_attach_encoder - attach a connector to an encoder + * @connector: connector to attach + * @encoder: encoder to attach @connector to + * + * This function links up a connector to an encoder. Note that the routing + * restrictions between encoders and crtcs are exposed to userspace through the + * possible_clones and possible_crtcs bitmasks. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_connector_attach_encoder(struct drm_connector *connector, struct drm_encoder *encoder) { @@ -3450,8 +3856,20 @@ int drm_mode_connector_attach_encoder(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_attach_encoder); +/** + * drm_mode_crtc_set_gamma_size - set the gamma table size + * @crtc: CRTC to set the gamma table size for + * @gamma_size: size of the gamma table + * + * Drivers which support gamma tables should set this to the supported gamma + * table size when initializing the CRTC. Currently the drm core only supports a + * fixed gamma table size. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, - int gamma_size) + int gamma_size) { crtc->gamma_size = gamma_size; @@ -3465,6 +3883,20 @@ int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, } EXPORT_SYMBOL(drm_mode_crtc_set_gamma_size); +/** + * drm_mode_gamma_set_ioctl - set the gamma table + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * Set the gamma table of a CRTC to the one passed in by the user. Userspace can + * inquire the required gamma table size through drm_mode_gamma_get_ioctl. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_gamma_set_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3524,6 +3956,21 @@ int drm_mode_gamma_set_ioctl(struct drm_device *dev, } +/** + * drm_mode_gamma_get_ioctl - get the gamma table + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * Copy the current gamma table into the storage provided. This also provides + * the gamma table size the driver expects, which can be used to size the + * allocated storage. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_gamma_get_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3574,6 +4021,24 @@ int drm_mode_gamma_get_ioctl(struct drm_device *dev, return ret; } +/** + * drm_mode_page_flip_ioctl - schedule an asynchronous fb update + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This schedules an asynchronous update on a given CRTC, called page flip. + * Optionally a drm event is generated to signal the completion of the event. + * Generic drivers cannot assume that a pageflip with changed framebuffer + * properties (including driver specific metadata like tiling layout) will work, + * but some drivers support e.g. pixel format changes through the pageflip + * ioctl. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_page_flip_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3686,6 +4151,14 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, return ret; } +/** + * drm_mode_config_reset - call ->reset callbacks + * @dev: drm device + * + * This functions calls all the crtc's, encoder's and connector's ->reset + * callback. Drivers can use this in e.g. their driver load or resume code to + * reset hardware and software state. + */ void drm_mode_config_reset(struct drm_device *dev) { struct drm_crtc *crtc; @@ -3709,6 +4182,25 @@ void drm_mode_config_reset(struct drm_device *dev) } EXPORT_SYMBOL(drm_mode_config_reset); +/** + * drm_mode_create_dumb_ioctl - create a dumb backing storage buffer + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This creates a new dumb buffer in the driver's backing storage manager (GEM, + * TTM or something else entirely) and returns the resulting buffer handle. This + * handle can then be wrapped up into a framebuffer modeset object. + * + * Note that userspace is not allowed to use such objects for render + * acceleration - drivers must create their own private ioctls for such a use + * case. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_create_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3719,6 +4211,20 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev, return dev->driver->dumb_create(file_priv, dev, args); } +/** + * drm_mode_mmap_dumb_ioctl - create an mmap offset for a dumb backing storage buffer + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * Allocate an offset in the drm device node's address space to be able to + * memory map a dumb buffer. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_mmap_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3731,6 +4237,21 @@ int drm_mode_mmap_dumb_ioctl(struct drm_device *dev, return dev->driver->dumb_map_offset(file_priv, dev, args->handle, &args->offset); } +/** + * drm_mode_destroy_dumb_ioctl - destroy a dumb backing strage buffer + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * This destroys the userspace handle for the given dumb backing storage buffer. + * Since buffer objects must be reference counted in the kernel a buffer object + * won't be immediately freed if a framebuffer modeset object still uses it. + * + * Called by the user via ioctl. + * + * Returns: + * Zero on success, errno on failure. + */ int drm_mode_destroy_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -3742,9 +4263,14 @@ int drm_mode_destroy_dumb_ioctl(struct drm_device *dev, return dev->driver->dumb_destroy(file_priv, dev, args->handle); } -/* - * Just need to support RGB formats here for compat with code that doesn't - * use pixel formats directly yet. +/** + * drm_fb_get_bpp_depth - get the bpp/depth values for format + * @format: pixel format (DRM_FORMAT_*) + * @depth: storage for the depth value + * @bpp: storage for the bpp value + * + * This only supports RGB formats here for compat with code that doesn't use + * pixel formats directly yet. */ void drm_fb_get_bpp_depth(uint32_t format, unsigned int *depth, int *bpp) @@ -3816,7 +4342,7 @@ EXPORT_SYMBOL(drm_fb_get_bpp_depth); * drm_format_num_planes - get the number of planes for format * @format: pixel format (DRM_FORMAT_*) * - * RETURNS: + * Returns: * The number of planes used by the specified pixel format. */ int drm_format_num_planes(uint32_t format) @@ -3851,7 +4377,7 @@ EXPORT_SYMBOL(drm_format_num_planes); * @format: pixel format (DRM_FORMAT_*) * @plane: plane index * - * RETURNS: + * Returns: * The bytes per pixel value for the specified plane. */ int drm_format_plane_cpp(uint32_t format, int plane) @@ -3897,7 +4423,7 @@ EXPORT_SYMBOL(drm_format_plane_cpp); * drm_format_horz_chroma_subsampling - get the horizontal chroma subsampling factor * @format: pixel format (DRM_FORMAT_*) * - * RETURNS: + * Returns: * The horizontal chroma subsampling factor for the * specified pixel format. */ @@ -3932,7 +4458,7 @@ EXPORT_SYMBOL(drm_format_horz_chroma_subsampling); * drm_format_vert_chroma_subsampling - get the vertical chroma subsampling factor * @format: pixel format (DRM_FORMAT_*) * - * RETURNS: + * Returns: * The vertical chroma subsampling factor for the * specified pixel format. */ From f4e53f9a4f7d01cef9f096b1eb3705329a7ce9c2 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Thu, 13 Mar 2014 10:18:53 +0100 Subject: [PATCH 213/259] MAINTAINERS: add networking selftests to NETWORKING Add it to NETWORKING [GENERAL] to make sure patches for selftests go to the netdev list as well. Signed-off-by: Daniel Borkmann Signed-off-by: David S. Miller --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index e1be77c78336..b359eb056e5c 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -6003,6 +6003,7 @@ F: include/uapi/linux/in.h F: include/uapi/linux/net.h F: include/uapi/linux/netdevice.h F: tools/net/ +F: tools/testing/selftests/net/ NETWORKING [IPv4/IPv6] M: "David S. Miller" From 0a8d8c446b5429d15ff2d48f46e00d8a08552303 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 13 Mar 2014 10:44:34 +0100 Subject: [PATCH 214/259] vmxnet3: fix building without CONFIG_PCI_MSI Since commit d25f06ea466e "vmxnet3: fix netpoll race condition", the vmxnet3 driver fails to build when CONFIG_PCI_MSI is disabled, because it unconditionally references the vmxnet3_msix_rx() function. To fix this, use the same #ifdef in the caller that exists around the function definition. Signed-off-by: Arnd Bergmann Cc: Neil Horman Cc: Shreyas Bhatewara Cc: "VMware, Inc." Cc: "David S. Miller" Cc: stable@vger.kernel.org Acked-by: Neil Horman Signed-off-by: David S. Miller --- drivers/net/vmxnet3/vmxnet3_drv.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c index b7daa02ff026..0fa3b44f7342 100644 --- a/drivers/net/vmxnet3/vmxnet3_drv.c +++ b/drivers/net/vmxnet3/vmxnet3_drv.c @@ -1761,13 +1761,16 @@ static void vmxnet3_netpoll(struct net_device *netdev) { struct vmxnet3_adapter *adapter = netdev_priv(netdev); - int i; switch (adapter->intr.type) { - case VMXNET3_IT_MSIX: +#ifdef CONFIG_PCI_MSI + case VMXNET3_IT_MSIX: { + int i; for (i = 0; i < adapter->num_rx_queues; i++) vmxnet3_msix_rx(0, &adapter->rx_queue[i]); break; + } +#endif case VMXNET3_IT_MSI: default: vmxnet3_intr(0, adapter->netdev); From de123268300fd33b7f7668fda3264059daffa6ef Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Thu, 13 Mar 2014 14:52:15 +0200 Subject: [PATCH 215/259] net/mlx4_en: Deregister multicast vxlan steering rules when going down When mlx4_en_stop_port() is called, we need to deregister also the tunnel steering rules that relate to multicast. Signed-off-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index ba2126111573..84a96f70dfb5 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -1800,6 +1800,8 @@ void mlx4_en_stop_port(struct net_device *dev, int detach) mc_list[5] = priv->port; mlx4_multicast_detach(mdev->dev, &priv->rss_map.indir_qp, mc_list, MLX4_PROT_ETH, mclist->reg_id); + if (mclist->tunnel_reg_id) + mlx4_flow_detach(mdev->dev, mclist->tunnel_reg_id); } mlx4_en_clear_list(dev); list_for_each_entry_safe(mclist, tmp, &priv->curr_list, list) { From 6e07a1e0b53d77dbcb08f29d1cca07c6d33a926f Mon Sep 17 00:00:00 2001 From: Alexander Aring Date: Wed, 12 Mar 2014 08:21:24 +0100 Subject: [PATCH 216/259] at86rf230: fix lockdep splats This patch fix a lockdep in the at86rf230 driver, otherwise we get: [ 30.206517] ================================= [ 30.211078] [ INFO: inconsistent lock state ] [ 30.215647] 3.14.0-20140108-1-00994-g32e9426 #163 Not tainted [ 30.221660] --------------------------------- [ 30.226222] inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage. [ 30.232514] systemd-udevd/157 [HC1[1]:SC0[0]:HE0:SE1] takes: [ 30.238439] (&(&lp->lock)->rlock){?.+...}, at: [] at86rf230_isr+0x18/0x44 [ 30.246621] {HARDIRQ-ON-W} state was registered at: [ 30.251728] [] __lock_acquire+0x7a4/0x18d8 [ 30.257135] [] lock_acquire+0x68/0x7c [ 30.262071] [] _raw_spin_lock+0x28/0x38 [ 30.267203] [] at86rf230_xmit+0x1c/0x144 [ 30.272412] [] mac802154_xmit_worker+0x88/0x148 [ 30.278271] [] process_one_work+0x274/0x404 [ 30.283761] [] worker_thread+0x228/0x374 [ 30.288971] [] kthread+0xd0/0xe4 [ 30.293455] [] ret_from_fork+0x14/0x2c [ 30.298493] irq event stamp: 8948 [ 30.301963] hardirqs last enabled at (8947): [] __kmalloc+0xb4/0x110 [ 30.309636] hardirqs last disabled at (8948): [] __irq_svc+0x34/0x5c [ 30.317215] softirqs last enabled at (8452): [] __do_softirq+0x1dc/0x264 [ 30.325243] softirqs last disabled at (8439): [] irq_exit+0x80/0xf4 We use the lp->lock inside the isr of at86rf230, that's why we need the irqsave spinlock calls. Signed-off-by: Alexander Aring Signed-off-by: David S. Miller --- drivers/net/ieee802154/at86rf230.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/net/ieee802154/at86rf230.c b/drivers/net/ieee802154/at86rf230.c index ab31544bc254..a30258aad139 100644 --- a/drivers/net/ieee802154/at86rf230.c +++ b/drivers/net/ieee802154/at86rf230.c @@ -546,12 +546,12 @@ at86rf230_xmit(struct ieee802154_dev *dev, struct sk_buff *skb) int rc; unsigned long flags; - spin_lock(&lp->lock); + spin_lock_irqsave(&lp->lock, flags); if (lp->irq_busy) { - spin_unlock(&lp->lock); + spin_unlock_irqrestore(&lp->lock, flags); return -EBUSY; } - spin_unlock(&lp->lock); + spin_unlock_irqrestore(&lp->lock, flags); might_sleep(); @@ -725,10 +725,11 @@ static void at86rf230_irqwork_level(struct work_struct *work) static irqreturn_t at86rf230_isr(int irq, void *data) { struct at86rf230_local *lp = data; + unsigned long flags; - spin_lock(&lp->lock); + spin_lock_irqsave(&lp->lock, flags); lp->irq_busy = 1; - spin_unlock(&lp->lock); + spin_unlock_irqrestore(&lp->lock, flags); schedule_work(&lp->irqwork); From fb00bc2e6cd2046282ba4b03f4fe682aee70b2f8 Mon Sep 17 00:00:00 2001 From: dingtianhong Date: Wed, 12 Mar 2014 17:31:59 +0800 Subject: [PATCH 217/259] bonding: set correct vlan id for alb xmit path The commit d3ab3ffd1d728d7ee77340e7e7e2c7cfe6a4013e (bonding: use rlb_client_info->vlan_id instead of ->tag) remove the rlb_client_info->tag, but occur some issues, The vlan_get_tag() will return 0 for success and -EINVAL for error, so the client_info->vlan_id always be set to 0 if the vlan_get_tag return 0 for success, so the client_info would never get a correct vlan id. We should only set the vlan id to 0 when the vlan_get_tag return error. Fixes: d3ab3ffd1d7 (bonding: use rlb_client_info->vlan_id instead of ->tag) CC: Ding Tianhong CC: Jay Vosburgh CC: Andy Gospodarek Signed-off-by: Ding Tianhong Acked-by: Veaceslav Falico Signed-off-by: David S. Miller --- drivers/net/bonding/bond_alb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c index a2c47476804d..e8f133e926aa 100644 --- a/drivers/net/bonding/bond_alb.c +++ b/drivers/net/bonding/bond_alb.c @@ -730,7 +730,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon client_info->ntt = 0; } - if (!vlan_get_tag(skb, &client_info->vlan_id)) + if (vlan_get_tag(skb, &client_info->vlan_id)) client_info->vlan_id = 0; if (!client_info->assigned) { From 84fe61821e4ebab6322eeae3f3c27f77f0031978 Mon Sep 17 00:00:00 2001 From: Stefan Wahren Date: Wed, 12 Mar 2014 11:28:19 +0100 Subject: [PATCH 218/259] eth: fec: Fix lost promiscuous mode after reconnecting cable If the Freescale fec is in promiscuous mode and network cable is reconnected then the promiscuous mode get lost. The problem is caused by a too soon call of set_multicast_list to re-enable promisc mode. The FEC_R_CNTRL register changes are overwritten by fec_restart. This patch fixes this by moving the call behind the init of FEC_R_CNTRL register in fec_restart. Successful tested on a i.MX28 board. Signed-off-by: Stefan Wahren Signed-off-by: David S. Miller --- drivers/net/ethernet/freescale/fec_main.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c index 479a7cba45c0..03a351300013 100644 --- a/drivers/net/ethernet/freescale/fec_main.c +++ b/drivers/net/ethernet/freescale/fec_main.c @@ -528,13 +528,6 @@ fec_restart(struct net_device *ndev, int duplex) /* Clear any outstanding interrupt. */ writel(0xffc00000, fep->hwp + FEC_IEVENT); - /* Setup multicast filter. */ - set_multicast_list(ndev); -#ifndef CONFIG_M5272 - writel(0, fep->hwp + FEC_HASH_TABLE_HIGH); - writel(0, fep->hwp + FEC_HASH_TABLE_LOW); -#endif - /* Set maximum receive buffer size. */ writel(PKT_MAXBLR_SIZE, fep->hwp + FEC_R_BUFF_SIZE); @@ -655,6 +648,13 @@ fec_restart(struct net_device *ndev, int duplex) writel(rcntl, fep->hwp + FEC_R_CNTRL); + /* Setup multicast filter. */ + set_multicast_list(ndev); +#ifndef CONFIG_M5272 + writel(0, fep->hwp + FEC_HASH_TABLE_HIGH); + writel(0, fep->hwp + FEC_HASH_TABLE_LOW); +#endif + if (id_entry->driver_data & FEC_QUIRK_ENET_MAC) { /* enable ENET endian swap */ ecntl |= (1 << 8); From ecab67015ef6e3f3635551dcc9971cf363cc1cd5 Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Wed, 12 Mar 2014 22:13:19 +0100 Subject: [PATCH 219/259] ipv6: Avoid unnecessary temporary addresses being generated tmp_prefered_lft is an offset to ifp->tstamp, not now. Therefore age needs to be added to the condition. Age calculation in ipv6_create_tempaddr is different from the one in addrconf_verify and doesn't consider ADDRCONF_TIMER_FUZZ_MINUS. This can cause age in ipv6_create_tempaddr to be less than the one in addrconf_verify and therefore unnecessary temporary address to be generated. Use age calculation as in addrconf_modify to avoid this. Signed-off-by: Heiner Kallweit Signed-off-by: David S. Miller --- net/ipv6/addrconf.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index fdbfeca36d63..344e972426df 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -1103,8 +1103,11 @@ static int ipv6_create_tempaddr(struct inet6_ifaddr *ifp, struct inet6_ifaddr *i * Lifetime is greater than REGEN_ADVANCE time units. In particular, * an implementation must not create a temporary address with a zero * Preferred Lifetime. + * Use age calculation as in addrconf_verify to avoid unnecessary + * temporary addresses being generated. */ - if (tmp_prefered_lft <= regen_advance) { + age = (now - tmp_tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ; + if (tmp_prefered_lft <= regen_advance + age) { in6_ifa_put(ifp); in6_dev_put(idev); ret = -1; From a4e90bed511220ff601d064c9e5d583e91308f65 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 13 Mar 2014 22:11:39 +0100 Subject: [PATCH 220/259] ACPI / sleep: Add extra checks for HW Reduced ACPI mode sleep states If the HW Reduced ACPI mode bit is set in the FADT, ACPICA uses the optional sleep control and sleep status registers for making the system enter sleep states (including S5), so it is not possible to use system sleep states or power it off using ACPI if the HW Reduced ACPI mode bit is set and those registers are not available. For this reason, add a new function, acpi_sleep_state_supported(), checking if the HW Reduced ACPI mode bit is set and whether or not system sleep states are usable in that case in addition to checking the return value of acpi_get_sleep_type_data() and make the ACPI sleep setup routines use that function to check the availability of system sleep states. Among other things, this prevents the kernel from attempting to use ACPI for powering off HW Reduced ACPI systems without the sleep control and sleep status registers, because ACPI power off doesn't have a chance to work on them. That allows alternative power off mechanisms that may actually work to be used on those systems. The affected machines include Dell Venue 8 Pro, Asus T100TA, Haswell Desktop SDP and Ivy Bridge EP Demo depot. References: https://bugzilla.kernel.org/show_bug.cgi?id=70931 Reported-by: Adam Williamson Tested-by: Aubrey Li Cc: 3.4+ # 3.4+ Signed-off-by: Rafael J. Wysocki --- drivers/acpi/sleep.c | 32 +++++++++++++++----------------- 1 file changed, 15 insertions(+), 17 deletions(-) diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index b718806657cd..c40fb2e81bbc 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -71,6 +71,17 @@ static int acpi_sleep_prepare(u32 acpi_state) return 0; } +static bool acpi_sleep_state_supported(u8 sleep_state) +{ + acpi_status status; + u8 type_a, type_b; + + status = acpi_get_sleep_type_data(sleep_state, &type_a, &type_b); + return ACPI_SUCCESS(status) && (!acpi_gbl_reduced_hardware + || (acpi_gbl_FADT.sleep_control.address + && acpi_gbl_FADT.sleep_status.address)); +} + #ifdef CONFIG_ACPI_SLEEP static u32 acpi_target_sleep_state = ACPI_STATE_S0; @@ -604,15 +615,9 @@ static void acpi_sleep_suspend_setup(void) { int i; - for (i = ACPI_STATE_S1; i < ACPI_STATE_S4; i++) { - acpi_status status; - u8 type_a, type_b; - - status = acpi_get_sleep_type_data(i, &type_a, &type_b); - if (ACPI_SUCCESS(status)) { + for (i = ACPI_STATE_S1; i < ACPI_STATE_S4; i++) + if (acpi_sleep_state_supported(i)) sleep_states[i] = 1; - } - } suspend_set_ops(old_suspend_ordering ? &acpi_suspend_ops_old : &acpi_suspend_ops); @@ -740,11 +745,7 @@ static const struct platform_hibernation_ops acpi_hibernation_ops_old = { static void acpi_sleep_hibernate_setup(void) { - acpi_status status; - u8 type_a, type_b; - - status = acpi_get_sleep_type_data(ACPI_STATE_S4, &type_a, &type_b); - if (ACPI_FAILURE(status)) + if (!acpi_sleep_state_supported(ACPI_STATE_S4)) return; hibernation_set_ops(old_suspend_ordering ? @@ -793,8 +794,6 @@ static void acpi_power_off(void) int __init acpi_sleep_init(void) { - acpi_status status; - u8 type_a, type_b; char supported[ACPI_S_STATE_COUNT * 3 + 1]; char *pos = supported; int i; @@ -806,8 +805,7 @@ int __init acpi_sleep_init(void) acpi_sleep_suspend_setup(); acpi_sleep_hibernate_setup(); - status = acpi_get_sleep_type_data(ACPI_STATE_S5, &type_a, &type_b); - if (ACPI_SUCCESS(status)) { + if (acpi_sleep_state_supported(ACPI_STATE_S5)) { sleep_states[ACPI_STATE_S5] = 1; pm_power_off_prepare = acpi_power_off_prepare; pm_power_off = acpi_power_off; From 62c19c9d29e65086e5ae76df371ed2e6b23f00cd Mon Sep 17 00:00:00 2001 From: Richard Weinberger Date: Sun, 9 Feb 2014 19:47:40 +0100 Subject: [PATCH 221/259] i2c: Remove usage of orphaned symbol OF_I2C The symbol is an orphan, don't depend on it anymore. Signed-off-by: Richard Weinberger [wsa: enhanced commit message] Signed-off-by: Wolfram Sang Fixes: 687b81d083c0 (i2c: move OF helpers into the core) Cc: stable@kernel.org --- drivers/i2c/busses/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index f5ed03164d86..de17c5593d97 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -387,7 +387,7 @@ config I2C_CBUS_GPIO config I2C_CPM tristate "Freescale CPM1 or CPM2 (MPC8xx/826x)" - depends on (CPM1 || CPM2) && OF_I2C + depends on CPM1 || CPM2 help This supports the use of the I2C interface on Freescale processors with CPM1 or CPM2. From 847d7970defb45540735b3fb4e88471c27cacd85 Mon Sep 17 00:00:00 2001 From: Daniel J Blueman Date: Thu, 13 Mar 2014 19:43:01 +0800 Subject: [PATCH 222/259] x86/amd/numa: Fix northbridge quirk to assign correct NUMA node For systems with multiple servers and routed fabric, all northbridges get assigned to the first server. Fix this by also using the node reported from the PCI bus. For single-fabric systems, the northbriges are on PCI bus 0 by definition, which are on NUMA node 0 by definition, so this is invarient on most systems. Tested on fam10h and fam15h single and multi-fabric systems and candidate for stable. Signed-off-by: Daniel J Blueman Acked-by: Steffen Persvold Acked-by: Borislav Petkov Cc: Link: http://lkml.kernel.org/r/1394710981-3596-1-git-send-email-daniel@numascale.com Signed-off-by: Ingo Molnar --- arch/x86/kernel/quirks.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c index 7c6acd4b8995..ff898bbf579d 100644 --- a/arch/x86/kernel/quirks.c +++ b/arch/x86/kernel/quirks.c @@ -529,7 +529,7 @@ static void quirk_amd_nb_node(struct pci_dev *dev) return; pci_read_config_dword(nb_ht, 0x60, &val); - node = val & 7; + node = pcibus_to_node(dev->bus) | (val & 7); /* * Some hardware may return an invalid node ID, * so check it first: From fc1645ac826c82ebad4402aabbf65595b671ecca Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Sat, 15 Mar 2014 12:11:54 +0100 Subject: [PATCH 223/259] drm/imx: remove drm_mode_connector_detach_encoder harder Since the last time I've looked more of this stuff sprouted up. Stomp it down again. Repeating the original justification for ripping this all out: There's absolutely no need to deteach connectors before cleaning them up at driver unload time. And since drm doesn't support hotplugging kms objects at all it's positively dangerous to attempt this at runtime. Luckily imx only detachs at driver cleanup time and hence we can savely remove this. Reported-by: kbuild test robot Cc: Sascha Hauer Cc: Russell King Cc: Greg Kroah-Hartman Signed-off-by: Daniel Vetter --- drivers/staging/imx-drm/imx-hdmi.c | 1 - drivers/staging/imx-drm/imx-tve.c | 2 -- 2 files changed, 3 deletions(-) diff --git a/drivers/staging/imx-drm/imx-hdmi.c b/drivers/staging/imx-drm/imx-hdmi.c index 62ce0e86f14b..f996e082faf4 100644 --- a/drivers/staging/imx-drm/imx-hdmi.c +++ b/drivers/staging/imx-drm/imx-hdmi.c @@ -1883,7 +1883,6 @@ static int imx_hdmi_platform_remove(struct platform_device *pdev) struct drm_connector *connector = &hdmi->connector; struct drm_encoder *encoder = &hdmi->encoder; - drm_mode_connector_detach_encoder(connector, encoder); imx_drm_remove_connector(hdmi->imx_drm_connector); imx_drm_remove_encoder(hdmi->imx_drm_encoder); diff --git a/drivers/staging/imx-drm/imx-tve.c b/drivers/staging/imx-drm/imx-tve.c index 9abc7ca8b6cf..64729fa4a3d4 100644 --- a/drivers/staging/imx-drm/imx-tve.c +++ b/drivers/staging/imx-drm/imx-tve.c @@ -709,8 +709,6 @@ static int imx_tve_remove(struct platform_device *pdev) struct drm_connector *connector = &tve->connector; struct drm_encoder *encoder = &tve->encoder; - drm_mode_connector_detach_encoder(connector, encoder); - imx_drm_remove_connector(tve->imx_drm_connector); imx_drm_remove_encoder(tve->imx_drm_encoder); From 06c99161b66d36b0345c443bd0934cfc3f4d7f54 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 19:52:29 +0100 Subject: [PATCH 224/259] drm/udl: fix error-path when damage-req fails We need to call dma_buf_end_cpu_access() in case a damage-request. Unlikely, but might happen during device unplug. Reviewed-by: Daniel Vetter Signed-off-by: David Herrmann --- drivers/gpu/drm/udl/udl_fb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c index dbadd49e4c4a..377176372da8 100644 --- a/drivers/gpu/drm/udl/udl_fb.c +++ b/drivers/gpu/drm/udl/udl_fb.c @@ -421,7 +421,7 @@ static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb, clips[i].x2 - clips[i].x1, clips[i].y2 - clips[i].y1); if (ret) - goto unlock; + break; } if (ufb->obj->base.import_attach) { From 2b932d8ef009f37d397c211b1dc5d0b056f6ef64 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 19:54:18 +0100 Subject: [PATCH 225/259] drm/udl: fix Bpp calculation in dumb_create() Probably a typo.. we obviously need "(bpp + 7) / 8" instead of "(bpp + 1) / 8". Unlikely to be hit in any sane code, but lets be safe. Use DIV_ROUND_UP() to avoid the problem entirely and make the core more readable. Reviewed-by: Daniel Vetter Signed-off-by: David Herrmann --- drivers/gpu/drm/udl/udl_gem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c index 8d67b943ac05..be4fcd0f0e0f 100644 --- a/drivers/gpu/drm/udl/udl_gem.c +++ b/drivers/gpu/drm/udl/udl_gem.c @@ -60,7 +60,7 @@ int udl_dumb_create(struct drm_file *file, struct drm_device *dev, struct drm_mode_create_dumb *args) { - args->pitch = args->width * ((args->bpp + 1) / 8); + args->pitch = args->width * DIV_ROUND_UP(args->bpp, 8); args->size = args->pitch * args->height; return udl_gem_create(file, dev, args->size, &args->handle); From 16d2831d6f590681ef239562ac6d73c605e7d6dc Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 20:07:49 +0100 Subject: [PATCH 226/259] drm/gem: fix indentation Remove double-whitespace and wrong indentation. Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_gem.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 5bbad873c798..dd8e38a22e23 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -692,7 +692,7 @@ drm_gem_object_release(struct drm_gem_object *obj) WARN_ON(obj->dma_buf); if (obj->filp) - fput(obj->filp); + fput(obj->filp); } EXPORT_SYMBOL(drm_gem_object_release); @@ -782,7 +782,7 @@ int drm_gem_mmap_obj(struct drm_gem_object *obj, unsigned long obj_size, vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_ops = dev->driver->gem_vm_ops; vma->vm_private_data = obj; - vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); + vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); /* Take a ref for this mapping of the object, so that the fault * handler can dereference the mmap offset's pointer to the object. From 77472347972add74a3d89a0b9152b8eebc0ad2b0 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 20:05:43 +0100 Subject: [PATCH 227/259] drm/gem: free vma-node during object-cleanup All drivers currently need to clean up the vma-node manually. There is no fancy logic involved so lets just clean it up unconditionally. The vma-manager correctly catches multiple calls so we are fine. Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_gem.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index dd8e38a22e23..5ea622c54e76 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -693,6 +693,8 @@ drm_gem_object_release(struct drm_gem_object *obj) if (obj->filp) fput(obj->filp); + + drm_gem_free_mmap_offset(obj); } EXPORT_SYMBOL(drm_gem_object_release); From b28cd41f9e9bb8085f7362c80833fc129628d3d6 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 20:09:55 +0100 Subject: [PATCH 228/259] drm/crtc: add sanity checks to create_dumb() Lets make sure some basic expressions are always true: bpp != NULL width != NULL height != NULL stride = bpp * width < 2^32 size = stride * height < 2^32 PAGE_ALIGN(size) < 2^32 At least the udl driver doesn't check for multiplication-overflows, so lets just make sure it will never happen. These checks allow drivers to do any 32bit math without having to test for mult-overflows themselves. The two divisions might hurt performance a bit, but dumb_create() is only used for scanout-buffers, so that should be fine. We could use 64bit math to avoid the divisions, but that may be slow on 32bit machines.. Or maybe there should just be a "safe_mult32()" helper, which currently doesn't exist (I think?). Reviewed-by: Daniel Vetter Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_crtc.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 35ea15d5ffff..b1c2b278005c 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -3784,9 +3784,26 @@ int drm_mode_create_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct drm_mode_create_dumb *args = data; + u32 cpp, stride, size; if (!dev->driver->dumb_create) return -ENOSYS; + if (!args->width || !args->height || !args->bpp) + return -EINVAL; + + /* overflow checks for 32bit size calculations */ + cpp = DIV_ROUND_UP(args->bpp, 8); + if (cpp > 0xffffffffU / args->width) + return -EINVAL; + stride = cpp * args->width; + if (args->height > 0xffffffffU / stride) + return -EINVAL; + + /* test for wrap-around */ + size = args->height * stride; + if (PAGE_ALIGN(size) == 0) + return -EINVAL; + return dev->driver->dumb_create(file_priv, dev, args); } From a8469aa81de532180846b22e8ead3d8f4d2f96a2 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 20 Jan 2014 20:15:38 +0100 Subject: [PATCH 229/259] drm/gem: dont init "ret" in drm_gem_mmap() There is no need to initialize this variable, so drop it. Otherwise, the compiler won't warn if we use it unintialized. Reviewed-by: Daniel Vetter Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_gem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 5ea622c54e76..154d6c6955c1 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -820,7 +820,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma) struct drm_device *dev = priv->minor->dev; struct drm_gem_object *obj; struct drm_vma_offset_node *node; - int ret = 0; + int ret; if (drm_device_is_unplugged(dev)) return -ENODEV; From 31bbe16f6d88622d6731fa2cb4ab38d57d844ac1 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Fri, 3 Jan 2014 14:09:47 +0100 Subject: [PATCH 230/259] drm: add pseudo filesystem for shared inodes Our current DRM design uses a single address_space for all users of the same DRM device. However, there is no way to create an anonymous address_space without an underlying inode. Therefore, we wait for the first ->open() callback on a registered char-dev and take-over the inode of the char-dev. This worked well so far, but has several drawbacks: - We screw with FS internals and rely on some non-obvious invariants like inode->i_mapping being the same as inode->i_data for char-devs. - We don't have any address_space prior to the first ->open() from user-space. This leads to ugly fallback code and we cannot allocate global objects early. As pointed out by Al-Viro, fs/anon_inode.c is *not* supposed to be used by drivers for anonymous inode-allocation. Therefore, this patch follows the proposed alternative solution and adds a pseudo filesystem mount-point to DRM. We can then allocate private inodes including a private address_space for each DRM device at initialization time. Note that we could use: sysfs_get_inode(sysfs_mnt->mnt_sb, drm_device->dev->kobj.sd); to get access to the underlying sysfs-inode of a "struct device" object. However, most of this information is currently hidden and it's not clear whether this address_space is suitable for driver access. Thus, unless linux allows anonymous address_space objects or driver-core provides a public inode per device, we're left with our own private internal mount point. Cc: Al Viro Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_stub.c | 74 ++++++++++++++++++++++++++++++++++++++ fs/dcache.c | 1 + 2 files changed, 75 insertions(+) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 98a33c580ca1..f2903d7e3d8e 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -31,8 +31,10 @@ * DEALINGS IN THE SOFTWARE. */ +#include #include #include +#include #include #include #include @@ -416,6 +418,78 @@ void drm_unplug_dev(struct drm_device *dev) } EXPORT_SYMBOL(drm_unplug_dev); +/* + * DRM internal mount + * We want to be able to allocate our own "struct address_space" to control + * memory-mappings in VRAM (or stolen RAM, ...). However, core MM does not allow + * stand-alone address_space objects, so we need an underlying inode. As there + * is no way to allocate an independent inode easily, we need a fake internal + * VFS mount-point. + * + * The drm_fs_inode_new() function allocates a new inode, drm_fs_inode_free() + * frees it again. You are allowed to use iget() and iput() to get references to + * the inode. But each drm_fs_inode_new() call must be paired with exactly one + * drm_fs_inode_free() call (which does not have to be the last iput()). + * We use drm_fs_inode_*() to manage our internal VFS mount-point and share it + * between multiple inode-users. You could, technically, call + * iget() + drm_fs_inode_free() directly after alloc and sometime later do an + * iput(), but this way you'd end up with a new vfsmount for each inode. + */ + +static int drm_fs_cnt; +static struct vfsmount *drm_fs_mnt; + +static const struct dentry_operations drm_fs_dops = { + .d_dname = simple_dname, +}; + +static const struct super_operations drm_fs_sops = { + .statfs = simple_statfs, +}; + +static struct dentry *drm_fs_mount(struct file_system_type *fs_type, int flags, + const char *dev_name, void *data) +{ + return mount_pseudo(fs_type, + "drm:", + &drm_fs_sops, + &drm_fs_dops, + 0x010203ff); +} + +static struct file_system_type drm_fs_type = { + .name = "drm", + .owner = THIS_MODULE, + .mount = drm_fs_mount, + .kill_sb = kill_anon_super, +}; + +static struct inode *drm_fs_inode_new(void) +{ + struct inode *inode; + int r; + + r = simple_pin_fs(&drm_fs_type, &drm_fs_mnt, &drm_fs_cnt); + if (r < 0) { + DRM_ERROR("Cannot mount pseudo fs: %d\n", r); + return ERR_PTR(r); + } + + inode = alloc_anon_inode(drm_fs_mnt->mnt_sb); + if (IS_ERR(inode)) + simple_release_fs(&drm_fs_mnt, &drm_fs_cnt); + + return inode; +} + +static void drm_fs_inode_free(struct inode *inode) +{ + if (inode) { + iput(inode); + simple_release_fs(&drm_fs_mnt, &drm_fs_cnt); + } +} + /** * drm_dev_alloc - Allocate new drm device * @driver: DRM driver to allocate device for diff --git a/fs/dcache.c b/fs/dcache.c index 265e0ce9769c..66dc62cb766d 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -3112,6 +3112,7 @@ char *simple_dname(struct dentry *dentry, char *buffer, int buflen) end = ERR_PTR(-ENAMETOOLONG); return end; } +EXPORT_SYMBOL(simple_dname); /* * Write full pathname from the root of the filesystem into the buffer. From 6796cb16c088905bf3af40548fda68c09e6f6ee5 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Fri, 3 Jan 2014 14:24:19 +0100 Subject: [PATCH 231/259] drm: use anon-inode instead of relying on cdevs DRM drivers share a common address_space across all character-devices of a single DRM device. This allows simple buffer eviction and mapping-control. However, DRM core currently waits for the first ->open() on any char-dev to mark the underlying inode as backing inode of the device. This delayed initialization causes ugly conditions all over the place: if (dev->dev_mapping) do_sth(); To avoid delayed initialization and to stop reusing the inode of the char-dev, we allocate an anonymous inode for each DRM device and reset filp->f_mapping to it on ->open(). Signed-off-by: David Herrmann --- drivers/gpu/drm/ast/ast_ttm.c | 2 +- drivers/gpu/drm/bochs/bochs_mm.c | 2 +- drivers/gpu/drm/cirrus/cirrus_ttm.c | 2 +- drivers/gpu/drm/drm_fops.c | 25 +++---------------- drivers/gpu/drm/drm_stub.c | 12 ++++++++- drivers/gpu/drm/i915/i915_gem.c | 3 ++- drivers/gpu/drm/mgag200/mgag200_ttm.c | 2 +- drivers/gpu/drm/nouveau/nouveau_gem.c | 2 +- drivers/gpu/drm/omapdrm/omap_gem.c | 34 +++++++++++++------------- drivers/gpu/drm/qxl/qxl_object.c | 3 +-- drivers/gpu/drm/qxl/qxl_ttm.c | 3 +-- drivers/gpu/drm/radeon/radeon_object.c | 2 +- drivers/gpu/drm/radeon/radeon_ttm.c | 2 +- drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 2 +- include/drm/drmP.h | 2 +- 15 files changed, 44 insertions(+), 54 deletions(-) diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c index 4ea9b17ac17a..2b491539e60c 100644 --- a/drivers/gpu/drm/ast/ast_ttm.c +++ b/drivers/gpu/drm/ast/ast_ttm.c @@ -324,7 +324,7 @@ int ast_bo_create(struct drm_device *dev, int size, int align, } astbo->bo.bdev = &ast->ttm.bdev; - astbo->bo.bdev->dev_mapping = dev->dev_mapping; + astbo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; ast_ttm_placement(astbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c index ce6858765b37..2d8546d7b4af 100644 --- a/drivers/gpu/drm/bochs/bochs_mm.c +++ b/drivers/gpu/drm/bochs/bochs_mm.c @@ -359,7 +359,7 @@ static int bochs_bo_create(struct drm_device *dev, int size, int align, } bochsbo->bo.bdev = &bochs->ttm.bdev; - bochsbo->bo.bdev->dev_mapping = dev->dev_mapping; + bochsbo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; bochs_ttm_placement(bochsbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c index 8b37c25ff9bd..efcbd70a8774 100644 --- a/drivers/gpu/drm/cirrus/cirrus_ttm.c +++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c @@ -329,7 +329,7 @@ int cirrus_bo_create(struct drm_device *dev, int size, int align, } cirrusbo->bo.bdev = &cirrus->ttm.bdev; - cirrusbo->bo.bdev->dev_mapping = dev->dev_mapping; + cirrusbo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; cirrus_ttm_placement(cirrusbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 7f2af9aca038..147a84d9da9b 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -84,8 +84,6 @@ int drm_open(struct inode *inode, struct file *filp) struct drm_minor *minor; int retcode = 0; int need_setup = 0; - struct address_space *old_mapping; - struct address_space *old_imapping; minor = idr_find(&drm_minors_idr, minor_id); if (!minor) @@ -99,16 +97,9 @@ int drm_open(struct inode *inode, struct file *filp) if (!dev->open_count++) need_setup = 1; - mutex_lock(&dev->struct_mutex); - old_imapping = inode->i_mapping; - old_mapping = dev->dev_mapping; - if (old_mapping == NULL) - dev->dev_mapping = &inode->i_data; - /* ihold ensures nobody can remove inode with our i_data */ - ihold(container_of(dev->dev_mapping, struct inode, i_data)); - inode->i_mapping = dev->dev_mapping; - filp->f_mapping = dev->dev_mapping; - mutex_unlock(&dev->struct_mutex); + + /* share address_space across all char-devs of a single device */ + filp->f_mapping = dev->anon_inode->i_mapping; retcode = drm_open_helper(inode, filp, dev); if (retcode) @@ -121,12 +112,6 @@ int drm_open(struct inode *inode, struct file *filp) return 0; err_undo: - mutex_lock(&dev->struct_mutex); - filp->f_mapping = old_imapping; - inode->i_mapping = old_imapping; - iput(container_of(dev->dev_mapping, struct inode, i_data)); - dev->dev_mapping = old_mapping; - mutex_unlock(&dev->struct_mutex); dev->open_count--; return retcode; } @@ -434,7 +419,6 @@ int drm_lastclose(struct drm_device * dev) drm_legacy_dma_takedown(dev); - dev->dev_mapping = NULL; mutex_unlock(&dev->struct_mutex); drm_legacy_dev_reinit(dev); @@ -549,9 +533,6 @@ int drm_release(struct inode *inode, struct file *filp) } } - BUG_ON(dev->dev_mapping == NULL); - iput(container_of(dev->dev_mapping, struct inode, i_data)); - /* drop the reference held my the file priv */ if (file_priv->master) drm_master_put(&file_priv->master); diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index f2903d7e3d8e..04c25cedd4c1 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -526,8 +526,15 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, mutex_init(&dev->struct_mutex); mutex_init(&dev->ctxlist_mutex); - if (drm_ht_create(&dev->map_hash, 12)) + dev->anon_inode = drm_fs_inode_new(); + if (IS_ERR(dev->anon_inode)) { + ret = PTR_ERR(dev->anon_inode); + DRM_ERROR("Cannot allocate anonymous inode: %d\n", ret); goto err_free; + } + + if (drm_ht_create(&dev->map_hash, 12)) + goto err_inode; ret = drm_ctxbitmap_init(dev); if (ret) { @@ -549,6 +556,8 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, drm_ctxbitmap_cleanup(dev); err_ht: drm_ht_remove(&dev->map_hash); +err_inode: + drm_fs_inode_free(dev->anon_inode); err_free: kfree(dev); return NULL; @@ -576,6 +585,7 @@ void drm_dev_free(struct drm_device *dev) drm_ctxbitmap_cleanup(dev); drm_ht_remove(&dev->map_hash); + drm_fs_inode_free(dev->anon_inode); kfree(dev->devname); kfree(dev); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 3618bb0cda0a..928e15bb2cf7 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1508,7 +1508,8 @@ i915_gem_release_mmap(struct drm_i915_gem_object *obj) if (!obj->fault_mappable) return; - drm_vma_node_unmap(&obj->base.vma_node, obj->base.dev->dev_mapping); + drm_vma_node_unmap(&obj->base.vma_node, + obj->base.dev->anon_inode->i_mapping); obj->fault_mappable = false; } diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c index adb5166a5dfd..c1c2cb608ab3 100644 --- a/drivers/gpu/drm/mgag200/mgag200_ttm.c +++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c @@ -324,7 +324,7 @@ int mgag200_bo_create(struct drm_device *dev, int size, int align, } mgabo->bo.bdev = &mdev->ttm.bdev; - mgabo->bo.bdev->dev_mapping = dev->dev_mapping; + mgabo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; mgag200_ttm_placement(mgabo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 27c3fd89e8ce..5b193b898f5a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -228,7 +228,7 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data, struct nouveau_bo *nvbo = NULL; int ret = 0; - drm->ttm.bdev.dev_mapping = drm->dev->dev_mapping; + drm->ttm.bdev.dev_mapping = drm->dev->anon_inode->i_mapping; if (!pfb->memtype_valid(pfb, req->info.tile_flags)) { NV_ERROR(cli, "bad page flags: 0x%08x\n", req->info.tile_flags); diff --git a/drivers/gpu/drm/omapdrm/omap_gem.c b/drivers/gpu/drm/omapdrm/omap_gem.c index 5aec3e81fe24..c8d972763889 100644 --- a/drivers/gpu/drm/omapdrm/omap_gem.c +++ b/drivers/gpu/drm/omapdrm/omap_gem.c @@ -153,24 +153,24 @@ static struct { static void evict_entry(struct drm_gem_object *obj, enum tiler_fmt fmt, struct usergart_entry *entry) { - if (obj->dev->dev_mapping) { - struct omap_gem_object *omap_obj = to_omap_bo(obj); - int n = usergart[fmt].height; - size_t size = PAGE_SIZE * n; - loff_t off = mmap_offset(obj) + - (entry->obj_pgoff << PAGE_SHIFT); - const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE); - if (m > 1) { - int i; - /* if stride > than PAGE_SIZE then sparse mapping: */ - for (i = n; i > 0; i--) { - unmap_mapping_range(obj->dev->dev_mapping, - off, PAGE_SIZE, 1); - off += PAGE_SIZE * m; - } - } else { - unmap_mapping_range(obj->dev->dev_mapping, off, size, 1); + struct omap_gem_object *omap_obj = to_omap_bo(obj); + int n = usergart[fmt].height; + size_t size = PAGE_SIZE * n; + loff_t off = mmap_offset(obj) + + (entry->obj_pgoff << PAGE_SHIFT); + const int m = 1 + ((omap_obj->width << fmt) / PAGE_SIZE); + + if (m > 1) { + int i; + /* if stride > than PAGE_SIZE then sparse mapping: */ + for (i = n; i > 0; i--) { + unmap_mapping_range(obj->dev->anon_inode->i_mapping, + off, PAGE_SIZE, 1); + off += PAGE_SIZE * m; } + } else { + unmap_mapping_range(obj->dev->anon_inode->i_mapping, + off, size, 1); } entry->obj = NULL; diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c index 8691c76c5ef0..7e20c21b0879 100644 --- a/drivers/gpu/drm/qxl/qxl_object.c +++ b/drivers/gpu/drm/qxl/qxl_object.c @@ -82,8 +82,7 @@ int qxl_bo_create(struct qxl_device *qdev, enum ttm_bo_type type; int r; - if (unlikely(qdev->mman.bdev.dev_mapping == NULL)) - qdev->mman.bdev.dev_mapping = qdev->ddev->dev_mapping; + qdev->mman.bdev.dev_mapping = qdev->ddev->anon_inode->i_mapping; if (kernel) type = ttm_bo_type_kernel; else diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c index c7e7e6590c2b..78cbc40a8efb 100644 --- a/drivers/gpu/drm/qxl/qxl_ttm.c +++ b/drivers/gpu/drm/qxl/qxl_ttm.c @@ -518,8 +518,7 @@ int qxl_ttm_init(struct qxl_device *qdev) ((unsigned)num_io_pages * PAGE_SIZE) / (1024 * 1024)); DRM_INFO("qxl: %uM of Surface memory size\n", (unsigned)qdev->surfaceram_size / (1024 * 1024)); - if (unlikely(qdev->mman.bdev.dev_mapping == NULL)) - qdev->mman.bdev.dev_mapping = qdev->ddev->dev_mapping; + qdev->mman.bdev.dev_mapping = qdev->ddev->anon_inode->i_mapping; r = qxl_ttm_debugfs_init(qdev); if (r) { DRM_ERROR("Failed to init debugfs\n"); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index ca79431b2c1c..6a7f3c6ffbbe 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -145,7 +145,7 @@ int radeon_bo_create(struct radeon_device *rdev, size = ALIGN(size, PAGE_SIZE); - rdev->mman.bdev.dev_mapping = rdev->ddev->dev_mapping; + rdev->mman.bdev.dev_mapping = rdev->ddev->anon_inode->i_mapping; if (kernel) { type = ttm_bo_type_kernel; } else if (sg) { diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 60dfce889ecf..4663fbcfda28 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -745,7 +745,7 @@ int radeon_ttm_init(struct radeon_device *rdev) } DRM_INFO("radeon: %uM of GTT memory ready.\n", (unsigned)(rdev->mc.gtt_size / (1024 * 1024))); - rdev->mman.bdev.dev_mapping = rdev->ddev->dev_mapping; + rdev->mman.bdev.dev_mapping = rdev->ddev->anon_inode->i_mapping; r = radeon_ttm_debugfs_init(rdev); if (r) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 0083cbf99edf..df4b03e035bc 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -969,7 +969,7 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv) goto out_no_shman; file_priv->driver_priv = vmw_fp; - dev_priv->bdev.dev_mapping = dev->dev_mapping; + dev_priv->bdev.dev_mapping = dev->anon_inode->i_mapping; return 0; diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 04a7f31301f8..3227b716ffdf 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1183,7 +1183,7 @@ struct drm_device { struct drm_sg_mem *sg; /**< Scatter gather memory */ unsigned int num_crtcs; /**< Number of CRTCs on this device */ void *dev_private; /**< device private data */ - struct address_space *dev_mapping; + struct inode *anon_inode; struct drm_sigdata sigdata; /**< For block_all_signals */ sigset_t sigmask; From 44d847b7439bdea0b6c5640446427daa3ebcc7fa Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Tue, 13 Aug 2013 19:10:30 +0200 Subject: [PATCH 232/259] drm: init TTM dev_mapping in ttm_bo_device_init() With dev->anon_inode we have a global address_space ready for operation right from the beginning. Therefore, there is no need to do a delayed setup with TTM. Instead, set dev_mapping during initialization in ttm_bo_device_init() and remove any "if (dev_mapping)" conditions. Cc: Dave Airlie Cc: Ben Skeggs Cc: Maarten Lankhorst Cc: Alex Deucher Cc: Thomas Hellstrom Signed-off-by: David Herrmann --- drivers/gpu/drm/ast/ast_ttm.c | 5 +++-- drivers/gpu/drm/bochs/bochs_mm.c | 4 +++- drivers/gpu/drm/cirrus/cirrus_ttm.c | 5 +++-- drivers/gpu/drm/mgag200/mgag200_ttm.c | 5 +++-- drivers/gpu/drm/nouveau/nouveau_gem.c | 2 -- drivers/gpu/drm/nouveau/nouveau_ttm.c | 4 +++- drivers/gpu/drm/qxl/qxl_object.c | 1 - drivers/gpu/drm/qxl/qxl_ttm.c | 5 +++-- drivers/gpu/drm/radeon/radeon_object.c | 1 - drivers/gpu/drm/radeon/radeon_ttm.c | 5 +++-- drivers/gpu/drm/ttm/ttm_bo.c | 3 ++- drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 5 +++-- include/drm/drm_vma_manager.h | 6 +++--- include/drm/ttm/ttm_bo_driver.h | 2 ++ 14 files changed, 31 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c index 2b491539e60c..b8246227bab0 100644 --- a/drivers/gpu/drm/ast/ast_ttm.c +++ b/drivers/gpu/drm/ast/ast_ttm.c @@ -259,7 +259,9 @@ int ast_mm_init(struct ast_private *ast) ret = ttm_bo_device_init(&ast->ttm.bdev, ast->ttm.bo_global_ref.ref.object, - &ast_bo_driver, DRM_FILE_PAGE_OFFSET, + &ast_bo_driver, + dev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, true); if (ret) { DRM_ERROR("Error initialising bo driver; %d\n", ret); @@ -324,7 +326,6 @@ int ast_bo_create(struct drm_device *dev, int size, int align, } astbo->bo.bdev = &ast->ttm.bdev; - astbo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; ast_ttm_placement(astbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c index 2d8546d7b4af..f488be55d650 100644 --- a/drivers/gpu/drm/bochs/bochs_mm.c +++ b/drivers/gpu/drm/bochs/bochs_mm.c @@ -225,7 +225,9 @@ int bochs_mm_init(struct bochs_device *bochs) ret = ttm_bo_device_init(&bochs->ttm.bdev, bochs->ttm.bo_global_ref.ref.object, - &bochs_bo_driver, DRM_FILE_PAGE_OFFSET, + &bochs_bo_driver, + bochs->dev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, true); if (ret) { DRM_ERROR("Error initialising bo driver; %d\n", ret); diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c index efcbd70a8774..92e6b7786097 100644 --- a/drivers/gpu/drm/cirrus/cirrus_ttm.c +++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c @@ -259,7 +259,9 @@ int cirrus_mm_init(struct cirrus_device *cirrus) ret = ttm_bo_device_init(&cirrus->ttm.bdev, cirrus->ttm.bo_global_ref.ref.object, - &cirrus_bo_driver, DRM_FILE_PAGE_OFFSET, + &cirrus_bo_driver, + dev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, true); if (ret) { DRM_ERROR("Error initialising bo driver; %d\n", ret); @@ -329,7 +331,6 @@ int cirrus_bo_create(struct drm_device *dev, int size, int align, } cirrusbo->bo.bdev = &cirrus->ttm.bdev; - cirrusbo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; cirrus_ttm_placement(cirrusbo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c index c1c2cb608ab3..5a00e90696de 100644 --- a/drivers/gpu/drm/mgag200/mgag200_ttm.c +++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c @@ -259,7 +259,9 @@ int mgag200_mm_init(struct mga_device *mdev) ret = ttm_bo_device_init(&mdev->ttm.bdev, mdev->ttm.bo_global_ref.ref.object, - &mgag200_bo_driver, DRM_FILE_PAGE_OFFSET, + &mgag200_bo_driver, + dev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, true); if (ret) { DRM_ERROR("Error initialising bo driver; %d\n", ret); @@ -324,7 +326,6 @@ int mgag200_bo_create(struct drm_device *dev, int size, int align, } mgabo->bo.bdev = &mdev->ttm.bdev; - mgabo->bo.bdev->dev_mapping = dev->anon_inode->i_mapping; mgag200_ttm_placement(mgabo, TTM_PL_FLAG_VRAM | TTM_PL_FLAG_SYSTEM); diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 5b193b898f5a..c90c0dc0afe8 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -228,8 +228,6 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data, struct nouveau_bo *nvbo = NULL; int ret = 0; - drm->ttm.bdev.dev_mapping = drm->dev->anon_inode->i_mapping; - if (!pfb->memtype_valid(pfb, req->info.tile_flags)) { NV_ERROR(cli, "bad page flags: 0x%08x\n", req->info.tile_flags); return -EINVAL; diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c index d45d50da978f..be3a3c9feafa 100644 --- a/drivers/gpu/drm/nouveau/nouveau_ttm.c +++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c @@ -376,7 +376,9 @@ nouveau_ttm_init(struct nouveau_drm *drm) ret = ttm_bo_device_init(&drm->ttm.bdev, drm->ttm.bo_global_ref.ref.object, - &nouveau_bo_driver, DRM_FILE_PAGE_OFFSET, + &nouveau_bo_driver, + dev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, bits <= 32 ? true : false); if (ret) { NV_ERROR(drm, "error initialising bo driver, %d\n", ret); diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c index 7e20c21b0879..b95f144f0b49 100644 --- a/drivers/gpu/drm/qxl/qxl_object.c +++ b/drivers/gpu/drm/qxl/qxl_object.c @@ -82,7 +82,6 @@ int qxl_bo_create(struct qxl_device *qdev, enum ttm_bo_type type; int r; - qdev->mman.bdev.dev_mapping = qdev->ddev->anon_inode->i_mapping; if (kernel) type = ttm_bo_type_kernel; else diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c index 78cbc40a8efb..29c02e0e857f 100644 --- a/drivers/gpu/drm/qxl/qxl_ttm.c +++ b/drivers/gpu/drm/qxl/qxl_ttm.c @@ -493,7 +493,9 @@ int qxl_ttm_init(struct qxl_device *qdev) /* No others user of address space so set it to 0 */ r = ttm_bo_device_init(&qdev->mman.bdev, qdev->mman.bo_global_ref.ref.object, - &qxl_bo_driver, DRM_FILE_PAGE_OFFSET, 0); + &qxl_bo_driver, + qdev->ddev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, 0); if (r) { DRM_ERROR("failed initializing buffer object driver(%d).\n", r); return r; @@ -518,7 +520,6 @@ int qxl_ttm_init(struct qxl_device *qdev) ((unsigned)num_io_pages * PAGE_SIZE) / (1024 * 1024)); DRM_INFO("qxl: %uM of Surface memory size\n", (unsigned)qdev->surfaceram_size / (1024 * 1024)); - qdev->mman.bdev.dev_mapping = qdev->ddev->anon_inode->i_mapping; r = qxl_ttm_debugfs_init(qdev); if (r) { DRM_ERROR("Failed to init debugfs\n"); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 6a7f3c6ffbbe..1375ff85b08a 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -145,7 +145,6 @@ int radeon_bo_create(struct radeon_device *rdev, size = ALIGN(size, PAGE_SIZE); - rdev->mman.bdev.dev_mapping = rdev->ddev->anon_inode->i_mapping; if (kernel) { type = ttm_bo_type_kernel; } else if (sg) { diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 4663fbcfda28..2db486666d91 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -707,7 +707,9 @@ int radeon_ttm_init(struct radeon_device *rdev) /* No others user of address space so set it to 0 */ r = ttm_bo_device_init(&rdev->mman.bdev, rdev->mman.bo_global_ref.ref.object, - &radeon_bo_driver, DRM_FILE_PAGE_OFFSET, + &radeon_bo_driver, + rdev->ddev->anon_inode->i_mapping, + DRM_FILE_PAGE_OFFSET, rdev->need_dma32); if (r) { DRM_ERROR("failed initializing buffer object driver(%d).\n", r); @@ -745,7 +747,6 @@ int radeon_ttm_init(struct radeon_device *rdev) } DRM_INFO("radeon: %uM of GTT memory ready.\n", (unsigned)(rdev->mc.gtt_size / (1024 * 1024))); - rdev->mman.bdev.dev_mapping = rdev->ddev->anon_inode->i_mapping; r = radeon_ttm_debugfs_init(rdev); if (r) { diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index a06651309388..79238d2face9 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -1449,6 +1449,7 @@ EXPORT_SYMBOL(ttm_bo_device_release); int ttm_bo_device_init(struct ttm_bo_device *bdev, struct ttm_bo_global *glob, struct ttm_bo_driver *driver, + struct address_space *mapping, uint64_t file_page_offset, bool need_dma32) { @@ -1470,7 +1471,7 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev, 0x10000000); INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue); INIT_LIST_HEAD(&bdev->ddestroy); - bdev->dev_mapping = NULL; + bdev->dev_mapping = mapping; bdev->glob = glob; bdev->need_dma32 = need_dma32; bdev->val_seq = 0; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index df4b03e035bc..c35715f26f40 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -722,7 +722,9 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) ret = ttm_bo_device_init(&dev_priv->bdev, dev_priv->bo_global_ref.ref.object, - &vmw_bo_driver, VMWGFX_FILE_PAGE_OFFSET, + &vmw_bo_driver, + dev->anon_inode->i_mapping, + VMWGFX_FILE_PAGE_OFFSET, false); if (unlikely(ret != 0)) { DRM_ERROR("Failed initializing TTM buffer object driver.\n"); @@ -969,7 +971,6 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv) goto out_no_shman; file_priv->driver_priv = vmw_fp; - dev_priv->bdev.dev_mapping = dev->anon_inode->i_mapping; return 0; diff --git a/include/drm/drm_vma_manager.h b/include/drm/drm_vma_manager.h index c18a593d1744..8cd402c73a5f 100644 --- a/include/drm/drm_vma_manager.h +++ b/include/drm/drm_vma_manager.h @@ -221,8 +221,8 @@ static inline __u64 drm_vma_node_offset_addr(struct drm_vma_offset_node *node) * @file_mapping: Address space to unmap @node from * * Unmap all userspace mappings for a given offset node. The mappings must be - * associated with the @file_mapping address-space. If no offset exists or - * the address-space is invalid, nothing is done. + * associated with the @file_mapping address-space. If no offset exists + * nothing is done. * * This call is unlocked. The caller must guarantee that drm_vma_offset_remove() * is not called on this node concurrently. @@ -230,7 +230,7 @@ static inline __u64 drm_vma_node_offset_addr(struct drm_vma_offset_node *node) static inline void drm_vma_node_unmap(struct drm_vma_offset_node *node, struct address_space *file_mapping) { - if (file_mapping && drm_vma_node_has_offset(node)) + if (drm_vma_node_has_offset(node)) unmap_mapping_range(file_mapping, drm_vma_node_offset_addr(node), drm_vma_node_size(node) << PAGE_SHIFT, 1); diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h index 32d34ebf0706..5d8aabe68f6c 100644 --- a/include/drm/ttm/ttm_bo_driver.h +++ b/include/drm/ttm/ttm_bo_driver.h @@ -747,6 +747,7 @@ extern int ttm_bo_device_release(struct ttm_bo_device *bdev); * @bdev: A pointer to a struct ttm_bo_device to initialize. * @glob: A pointer to an initialized struct ttm_bo_global. * @driver: A pointer to a struct ttm_bo_driver set up by the caller. + * @mapping: The address space to use for this bo. * @file_page_offset: Offset into the device address space that is available * for buffer data. This ensures compatibility with other users of the * address space. @@ -758,6 +759,7 @@ extern int ttm_bo_device_release(struct ttm_bo_device *bdev); extern int ttm_bo_device_init(struct ttm_bo_device *bdev, struct ttm_bo_global *glob, struct ttm_bo_driver *driver, + struct address_space *mapping, uint64_t file_page_offset, bool need_dma32); /** From 45e212d20fdccaf958b194e95a23ad264188c59e Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Tue, 28 Jan 2014 16:00:35 +0100 Subject: [PATCH 233/259] drm: group dev-lifetime related members These members are all managed by DRM-core, lets group them together so they're not split across the whole device. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- include/drm/drmP.h | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 04a7f31301f8..d6cfca9042fe 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1098,6 +1098,18 @@ struct drm_device { char *devname; /**< For /proc/interrupts */ int if_version; /**< Highest interface version set */ + /** \name Lifetime Management */ + /*@{ */ + struct device *dev; /**< Device structure of bus-device */ + struct drm_driver *driver; /**< DRM driver managing the device */ + void *dev_private; /**< DRM driver private data */ + struct address_space *dev_mapping; /**< Private addr-space just for the device */ + struct drm_minor *control; /**< Control node */ + struct drm_minor *primary; /**< Primary node */ + struct drm_minor *render; /**< Render node */ + atomic_t unplugged; /**< Flag whether dev is dead */ + /*@} */ + /** \name Locks */ /*@{ */ spinlock_t count_lock; /**< For inuse, drm_device::open_count, drm_device::buf_use */ @@ -1171,7 +1183,6 @@ struct drm_device { struct drm_agp_head *agp; /**< AGP data */ - struct device *dev; /**< Device structure */ struct pci_dev *pdev; /**< PCI device structure */ #ifdef __alpha__ struct pci_controller *hose; @@ -1182,17 +1193,11 @@ struct drm_device { struct drm_sg_mem *sg; /**< Scatter gather memory */ unsigned int num_crtcs; /**< Number of CRTCs on this device */ - void *dev_private; /**< device private data */ - struct address_space *dev_mapping; struct drm_sigdata sigdata; /**< For block_all_signals */ sigset_t sigmask; - struct drm_driver *driver; struct drm_local_map *agp_buffer_map; unsigned int agp_buffer_token; - struct drm_minor *control; /**< Control node for card */ - struct drm_minor *primary; /**< render type primary screen head */ - struct drm_minor *render; /**< render node for card */ struct drm_mode_config mode_config; /**< Current mode config */ @@ -1203,8 +1208,6 @@ struct drm_device { struct drm_vma_offset_manager *vma_offset_manager; /*@} */ int switch_power_state; - - atomic_t unplugged; /* device has been unplugged or gone away */ }; #define DRM_SWITCH_POWER_ON 0 From f4aede2e3291896e7cb42755ecc5b6815b6cac97 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 10:18:02 +0100 Subject: [PATCH 234/259] drm: skip redundant minor-lookup in open path The drm_open_helper() function is only used internally for drm_open() so we can safely pass in the minor-object directly instead of the minor-id. This way, we avoid the additional minor IDR lookup, which we already do twice in drm_stub_open() and drm_open(). Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_fops.c | 19 +++++++------------ 1 file changed, 7 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 7f2af9aca038..6466cb5d8b1f 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -44,7 +44,7 @@ DEFINE_MUTEX(drm_global_mutex); EXPORT_SYMBOL(drm_global_mutex); static int drm_open_helper(struct inode *inode, struct file *filp, - struct drm_device * dev); + struct drm_minor *minor); static int drm_setup(struct drm_device * dev) { @@ -110,7 +110,7 @@ int drm_open(struct inode *inode, struct file *filp) filp->f_mapping = dev->dev_mapping; mutex_unlock(&dev->struct_mutex); - retcode = drm_open_helper(inode, filp, dev); + retcode = drm_open_helper(inode, filp, minor); if (retcode) goto err_undo; if (need_setup) { @@ -196,16 +196,16 @@ static int drm_cpu_valid(void) * * \param inode device inode. * \param filp file pointer. - * \param dev device. + * \param minor acquired minor-object. * \return zero on success or a negative number on failure. * * Creates and initializes a drm_file structure for the file private data in \p * filp and add it into the double linked list in \p dev. */ static int drm_open_helper(struct inode *inode, struct file *filp, - struct drm_device * dev) + struct drm_minor *minor) { - int minor_id = iminor(inode); + struct drm_device *dev = minor->dev; struct drm_file *priv; int ret; @@ -216,7 +216,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp, if (dev->switch_power_state != DRM_SWITCH_POWER_ON && dev->switch_power_state != DRM_SWITCH_POWER_DYNAMIC_OFF) return -EINVAL; - DRM_DEBUG("pid = %d, minor = %d\n", task_pid_nr(current), minor_id); + DRM_DEBUG("pid = %d, minor = %d\n", task_pid_nr(current), minor->index); priv = kzalloc(sizeof(*priv), GFP_KERNEL); if (!priv) @@ -226,11 +226,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp, priv->filp = filp; priv->uid = current_euid(); priv->pid = get_pid(task_pid(current)); - priv->minor = idr_find(&drm_minors_idr, minor_id); - if (!priv->minor) { - ret = -ENODEV; - goto out_put_pid; - } + priv->minor = minor; /* for compatibility root is always authenticated */ priv->always_authenticated = capable(CAP_SYS_ADMIN); @@ -336,7 +332,6 @@ static int drm_open_helper(struct inode *inode, struct file *filp, drm_prime_destroy_file_private(&priv->prime); if (dev->driver->driver_features & DRIVER_GEM) drm_gem_release(dev, priv); -out_put_pid: put_pid(priv->pid); kfree(priv); filp->private_data = NULL; From b9a0d15cc59e896dc6b6c07583157d78fcf72fbb Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 12:30:15 +0100 Subject: [PATCH 235/259] drm: remove unused DRM_MINOR_UNASSIGNED This constant is unused, remove it. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- include/drm/drmP.h | 1 - 1 file changed, 1 deletion(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index d6cfca9042fe..ff20b888d525 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1008,7 +1008,6 @@ struct drm_driver { struct list_head legacy_dev_list; }; -#define DRM_MINOR_UNASSIGNED 0 #define DRM_MINOR_LEGACY 1 #define DRM_MINOR_CONTROL 2 #define DRM_MINOR_RENDER 3 From cb8a239b03608079cbfb784e9ac2f522fe846c29 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 12:31:40 +0100 Subject: [PATCH 236/259] drm: turn DRM_MINOR_* into enum Use enum for DRM_MINOR_* constants to avoid hard-coding the IDs. Furthermore, add a DRM_MINOR_CNT so we can perform range-checks in follow-ups. This changes the IDs of the minor-types by -1, but they're not used as indices so this is fine. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- include/drm/drmP.h | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index ff20b888d525..2c322564ca7d 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1008,9 +1008,12 @@ struct drm_driver { struct list_head legacy_dev_list; }; -#define DRM_MINOR_LEGACY 1 -#define DRM_MINOR_CONTROL 2 -#define DRM_MINOR_RENDER 3 +enum drm_minor_type { + DRM_MINOR_LEGACY, + DRM_MINOR_CONTROL, + DRM_MINOR_RENDER, + DRM_MINOR_CNT, +}; /** * Info file list entry. This structure represents a debugfs or proc file to From 099d1c290e2ebc3b798961a6c177c3aef5f0b789 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 10:21:36 +0100 Subject: [PATCH 237/259] drm: provide device-refcount Lets not trick ourselves into thinking "drm_device" objects are not ref-counted. That's just utterly stupid. We manage "drm_minor" objects on each drm-device and each minor can have an unlimited number of open handles. Each of these handles has the drm_minor (and thus the drm_device) as private-data in the file-handle. Therefore, we may not destroy "drm_device" until all these handles are closed. It is *not* possible to reset all these pointers atomically and restrict access to them, and this is *not* how this is done! Instead, we use ref-counts to make sure the object is valid and not freed. Note that we currently use "dev->open_count" for that, which is *exactly* the same as a reference-count, just open coded. So this patch doesn't change any semantics on DRM devices (well, this patch just introduces the ref-count, anyway. Follow-up patches will replace open_count by it). Also note that generic VFS revoke support could allow us to drop this ref-count again. We could then just synchronously disable any fops->xy() calls. However, this is not the case, yet, and no such patches are in sight (and I seriously question the idea of dropping the ref-cnt again). Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_pci.c | 2 +- drivers/gpu/drm/drm_platform.c | 2 +- drivers/gpu/drm/drm_stub.c | 56 +++++++++++++++++++++++++--------- drivers/gpu/drm/drm_usb.c | 2 +- drivers/gpu/drm/tegra/bus.c | 2 +- include/drm/drmP.h | 5 ++- 6 files changed, 50 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c index 5736aaa7e86c..9ded847b05b4 100644 --- a/drivers/gpu/drm/drm_pci.c +++ b/drivers/gpu/drm/drm_pci.c @@ -351,7 +351,7 @@ int drm_get_pci_dev(struct pci_dev *pdev, const struct pci_device_id *ent, drm_pci_agp_destroy(dev); pci_disable_device(pdev); err_free: - drm_dev_free(dev); + drm_dev_unref(dev); return ret; } EXPORT_SYMBOL(drm_get_pci_dev); diff --git a/drivers/gpu/drm/drm_platform.c b/drivers/gpu/drm/drm_platform.c index 21fc82006b78..319ff5385601 100644 --- a/drivers/gpu/drm/drm_platform.c +++ b/drivers/gpu/drm/drm_platform.c @@ -64,7 +64,7 @@ static int drm_get_platform_dev(struct platform_device *platdev, return 0; err_free: - drm_dev_free(dev); + drm_dev_unref(dev); return ret; } diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 98a33c580ca1..f2f0249304b7 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -392,7 +392,7 @@ void drm_put_dev(struct drm_device *dev) } drm_dev_unregister(dev); - drm_dev_free(dev); + drm_dev_unref(dev); } EXPORT_SYMBOL(drm_put_dev); @@ -425,6 +425,9 @@ EXPORT_SYMBOL(drm_unplug_dev); * Call drm_dev_register() to advertice the device to user space and register it * with other core subsystems. * + * The initial ref-count of the object is 1. Use drm_dev_ref() and + * drm_dev_unref() to take and drop further ref-counts. + * * RETURNS: * Pointer to new DRM device, or NULL if out of memory. */ @@ -438,6 +441,7 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, if (!dev) return NULL; + kref_init(&dev->ref); dev->dev = parent; dev->driver = driver; @@ -481,18 +485,10 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, } EXPORT_SYMBOL(drm_dev_alloc); -/** - * drm_dev_free - Free DRM device - * @dev: DRM device to free - * - * Free a DRM device that has previously been allocated via drm_dev_alloc(). - * You must not use kfree() instead or you will leak memory. - * - * This must not be called once the device got registered. Use drm_put_dev() - * instead, which then calls drm_dev_free(). - */ -void drm_dev_free(struct drm_device *dev) +static void drm_dev_release(struct kref *ref) { + struct drm_device *dev = container_of(ref, struct drm_device, ref); + drm_put_minor(dev->control); drm_put_minor(dev->render); drm_put_minor(dev->primary); @@ -506,7 +502,39 @@ void drm_dev_free(struct drm_device *dev) kfree(dev->devname); kfree(dev); } -EXPORT_SYMBOL(drm_dev_free); + +/** + * drm_dev_ref - Take reference of a DRM device + * @dev: device to take reference of or NULL + * + * This increases the ref-count of @dev by one. You *must* already own a + * reference when calling this. Use drm_dev_unref() to drop this reference + * again. + * + * This function never fails. However, this function does not provide *any* + * guarantee whether the device is alive or running. It only provides a + * reference to the object and the memory associated with it. + */ +void drm_dev_ref(struct drm_device *dev) +{ + if (dev) + kref_get(&dev->ref); +} +EXPORT_SYMBOL(drm_dev_ref); + +/** + * drm_dev_unref - Drop reference of a DRM device + * @dev: device to drop reference of or NULL + * + * This decreases the ref-count of @dev by one. The device is destroyed if the + * ref-count drops to zero. + */ +void drm_dev_unref(struct drm_device *dev) +{ + if (dev) + kref_put(&dev->ref, drm_dev_release); +} +EXPORT_SYMBOL(drm_dev_unref); /** * drm_dev_register - Register DRM device @@ -581,7 +609,7 @@ EXPORT_SYMBOL(drm_dev_register); * * Unregister the DRM device from the system. This does the reverse of * drm_dev_register() but does not deallocate the device. The caller must call - * drm_dev_free() to free all resources. + * drm_dev_unref() to drop their final reference. */ void drm_dev_unregister(struct drm_device *dev) { diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c index 0f8cb1ae7607..c3406aad2944 100644 --- a/drivers/gpu/drm/drm_usb.c +++ b/drivers/gpu/drm/drm_usb.c @@ -30,7 +30,7 @@ int drm_get_usb_dev(struct usb_interface *interface, return 0; err_free: - drm_dev_free(dev); + drm_dev_unref(dev); return ret; } diff --git a/drivers/gpu/drm/tegra/bus.c b/drivers/gpu/drm/tegra/bus.c index e38e5967d77b..71cef5c13dc8 100644 --- a/drivers/gpu/drm/tegra/bus.c +++ b/drivers/gpu/drm/tegra/bus.c @@ -63,7 +63,7 @@ int drm_host1x_init(struct drm_driver *driver, struct host1x_device *device) return 0; err_free: - drm_dev_free(drm); + drm_dev_unref(drm); return ret; } diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 2c322564ca7d..4e53f1607355 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -43,6 +43,7 @@ #include #endif /* __alpha__ */ #include +#include #include #include #include @@ -1102,6 +1103,7 @@ struct drm_device { /** \name Lifetime Management */ /*@{ */ + struct kref ref; /**< Object ref-count */ struct device *dev; /**< Device structure of bus-device */ struct drm_driver *driver; /**< DRM driver managing the device */ void *dev_private; /**< DRM driver private data */ @@ -1666,7 +1668,8 @@ static __inline__ void drm_core_dropmap(struct drm_local_map *map) struct drm_device *drm_dev_alloc(struct drm_driver *driver, struct device *parent); -void drm_dev_free(struct drm_device *dev); +void drm_dev_ref(struct drm_device *dev); +void drm_dev_unref(struct drm_device *dev); int drm_dev_register(struct drm_device *dev, unsigned long flags); void drm_dev_unregister(struct drm_device *dev); /*@}*/ From 1616c525b98deb34b8f4b02eccf0ae3a1310fa27 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 10:49:19 +0100 Subject: [PATCH 238/259] drm: add minor-lookup/release helpers Instead of accessing drm_minors_idr directly, this adds a small helper to hide the internals. This will help us later to remove the drm_global_mutex requirement for minor-lookup. Furthermore, this also makes sure that minor->dev is always valid and takes a reference-count to the device as long as the minor is used in an open-file. This way, "struct file*"->private_data->dev is guaranteed to be valid (which it has to, as we cannot reset it). Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_fops.c | 50 ++++++++++++++++++++------------------ drivers/gpu/drm/drm_stub.c | 39 +++++++++++++++++++++++++++++ include/drm/drmP.h | 4 +++ 3 files changed, 70 insertions(+), 23 deletions(-) diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 6466cb5d8b1f..79478191404a 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -79,23 +79,22 @@ static int drm_setup(struct drm_device * dev) */ int drm_open(struct inode *inode, struct file *filp) { - struct drm_device *dev = NULL; - int minor_id = iminor(inode); + struct drm_device *dev; struct drm_minor *minor; - int retcode = 0; + int retcode; int need_setup = 0; struct address_space *old_mapping; struct address_space *old_imapping; - minor = idr_find(&drm_minors_idr, minor_id); - if (!minor) - return -ENODEV; + minor = drm_minor_acquire(iminor(inode)); + if (IS_ERR(minor)) + return PTR_ERR(minor); - if (!(dev = minor->dev)) - return -ENODEV; - - if (drm_device_is_unplugged(dev)) - return -ENODEV; + dev = minor->dev; + if (drm_device_is_unplugged(dev)) { + retcode = -ENODEV; + goto err_release; + } if (!dev->open_count++) need_setup = 1; @@ -128,6 +127,8 @@ int drm_open(struct inode *inode, struct file *filp) dev->dev_mapping = old_mapping; mutex_unlock(&dev->struct_mutex); dev->open_count--; +err_release: + drm_minor_release(minor); return retcode; } EXPORT_SYMBOL(drm_open); @@ -143,33 +144,33 @@ EXPORT_SYMBOL(drm_open); */ int drm_stub_open(struct inode *inode, struct file *filp) { - struct drm_device *dev = NULL; + struct drm_device *dev; struct drm_minor *minor; - int minor_id = iminor(inode); int err = -ENODEV; const struct file_operations *new_fops; DRM_DEBUG("\n"); mutex_lock(&drm_global_mutex); - minor = idr_find(&drm_minors_idr, minor_id); - if (!minor) - goto out; - - if (!(dev = minor->dev)) - goto out; + minor = drm_minor_acquire(iminor(inode)); + if (IS_ERR(minor)) + goto out_unlock; + dev = minor->dev; if (drm_device_is_unplugged(dev)) - goto out; + goto out_release; new_fops = fops_get(dev->driver->fops); if (!new_fops) - goto out; + goto out_release; replace_fops(filp, new_fops); if (filp->f_op->open) err = filp->f_op->open(inode, filp); -out: + +out_release: + drm_minor_release(minor); +out_unlock: mutex_unlock(&drm_global_mutex); return err; } @@ -453,7 +454,8 @@ int drm_lastclose(struct drm_device * dev) int drm_release(struct inode *inode, struct file *filp) { struct drm_file *file_priv = filp->private_data; - struct drm_device *dev = file_priv->minor->dev; + struct drm_minor *minor = file_priv->minor; + struct drm_device *dev = minor->dev; int retcode = 0; mutex_lock(&drm_global_mutex); @@ -575,6 +577,8 @@ int drm_release(struct inode *inode, struct file *filp) } mutex_unlock(&drm_global_mutex); + drm_minor_release(minor); + return retcode; } EXPORT_SYMBOL(drm_release); diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index f2f0249304b7..269048215e82 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -355,6 +355,45 @@ static void drm_unplug_minor(struct drm_minor *minor) idr_remove(&drm_minors_idr, minor->index); } +/** + * drm_minor_acquire - Acquire a DRM minor + * @minor_id: Minor ID of the DRM-minor + * + * Looks up the given minor-ID and returns the respective DRM-minor object. The + * refence-count of the underlying device is increased so you must release this + * object with drm_minor_release(). + * + * As long as you hold this minor, it is guaranteed that the object and the + * minor->dev pointer will stay valid! However, the device may get unplugged and + * unregistered while you hold the minor. + * + * Returns: + * Pointer to minor-object with increased device-refcount, or PTR_ERR on + * failure. + */ +struct drm_minor *drm_minor_acquire(unsigned int minor_id) +{ + struct drm_minor *minor; + + minor = idr_find(&drm_minors_idr, minor_id); + if (!minor) + return ERR_PTR(-ENODEV); + + drm_dev_ref(minor->dev); + return minor; +} + +/** + * drm_minor_release - Release DRM minor + * @minor: Pointer to DRM minor object + * + * Release a minor that was previously acquired via drm_minor_acquire(). + */ +void drm_minor_release(struct drm_minor *minor) +{ + drm_dev_unref(minor->dev); +} + /** * drm_put_minor - Destroy DRM minor * @minor: Minor to destroy diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 4e53f1607355..82963167f161 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1672,6 +1672,10 @@ void drm_dev_ref(struct drm_device *dev); void drm_dev_unref(struct drm_device *dev); int drm_dev_register(struct drm_device *dev, unsigned long flags); void drm_dev_unregister(struct drm_device *dev); + +struct drm_minor *drm_minor_acquire(unsigned int minor_id); +void drm_minor_release(struct drm_minor *minor); + /*@}*/ /* PCI section */ From 05b701f6f60201c9906167351cce50db2e9db7ae Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 12:43:56 +0100 Subject: [PATCH 239/259] drm: allocate minors early Instead of waiting for device-registration, we now allocate minor-objects during device allocation. The minors are not registered or assigned an ID. This is still postponed to device-registration. While at it, remove the superfluous output-parameter in drm_get_minor(). The reason for this early allocation is to make dev->primary/control/render available atomically. So once the device is alive, all of them are already set and we never have the situation where one of them is set after another (they're either NULL or set, but never changed). This will eventually allow us to reduce minor-ID allocation to one base-ID instead of a single ID for each. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_stub.c | 114 ++++++++++++++++++++++++------------- 1 file changed, 73 insertions(+), 41 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 269048215e82..b595b64291a2 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -260,21 +260,49 @@ int drm_dropmaster_ioctl(struct drm_device *dev, void *data, return 0; } +static struct drm_minor **drm_minor_get_slot(struct drm_device *dev, + unsigned int type) +{ + switch (type) { + case DRM_MINOR_LEGACY: + return &dev->primary; + case DRM_MINOR_RENDER: + return &dev->render; + case DRM_MINOR_CONTROL: + return &dev->control; + default: + return NULL; + } +} + +static int drm_minor_alloc(struct drm_device *dev, unsigned int type) +{ + struct drm_minor *minor; + + minor = kzalloc(sizeof(*minor), GFP_KERNEL); + if (!minor) + return -ENOMEM; + + minor->type = type; + minor->dev = dev; + INIT_LIST_HEAD(&minor->master_list); + + *drm_minor_get_slot(dev, type) = minor; + return 0; +} + /** - * drm_get_minor - Allocate and register new DRM minor + * drm_get_minor - Register DRM minor * @dev: DRM device - * @minor: Pointer to where new minor is stored * @type: Type of minor * - * Allocate a new minor of the given type and register it. A pointer to the new - * minor is returned in @minor. + * Register minor of given type. * Caller must hold the global DRM mutex. * * RETURNS: * 0 on success, negative error code on failure. */ -static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, - int type) +static int drm_get_minor(struct drm_device *dev, unsigned int type) { struct drm_minor *new_minor; int ret; @@ -282,21 +310,16 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, DRM_DEBUG("\n"); + new_minor = *drm_minor_get_slot(dev, type); + if (!new_minor) + return 0; + minor_id = drm_minor_get_id(dev, type); if (minor_id < 0) return minor_id; - new_minor = kzalloc(sizeof(struct drm_minor), GFP_KERNEL); - if (!new_minor) { - ret = -ENOMEM; - goto err_idr; - } - - new_minor->type = type; new_minor->device = MKDEV(DRM_MAJOR, minor_id); - new_minor->dev = dev; new_minor->index = minor_id; - INIT_LIST_HEAD(&new_minor->master_list); idr_replace(&drm_minors_idr, new_minor, minor_id); @@ -314,7 +337,6 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, "DRM: Error sysfs_device_add.\n"); goto err_debugfs; } - *minor = new_minor; DRM_DEBUG("new minor assigned %d\n", minor_id); return 0; @@ -325,10 +347,7 @@ static int drm_get_minor(struct drm_device *dev, struct drm_minor **minor, drm_debugfs_cleanup(new_minor); err_mem: #endif - kfree(new_minor); -err_idr: idr_remove(&drm_minors_idr, minor_id); - *minor = NULL; return ret; } @@ -495,8 +514,24 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, mutex_init(&dev->struct_mutex); mutex_init(&dev->ctxlist_mutex); + if (drm_core_check_feature(dev, DRIVER_MODESET)) { + ret = drm_minor_alloc(dev, DRM_MINOR_CONTROL); + if (ret) + goto err_minors; + } + + if (drm_core_check_feature(dev, DRIVER_RENDER) && drm_rnodes) { + ret = drm_minor_alloc(dev, DRM_MINOR_RENDER); + if (ret) + goto err_minors; + } + + ret = drm_minor_alloc(dev, DRM_MINOR_LEGACY); + if (ret) + goto err_minors; + if (drm_ht_create(&dev->map_hash, 12)) - goto err_free; + goto err_minors; ret = drm_ctxbitmap_init(dev); if (ret) { @@ -518,7 +553,10 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, drm_ctxbitmap_cleanup(dev); err_ht: drm_ht_remove(&dev->map_hash); -err_free: +err_minors: + drm_put_minor(dev->control); + drm_put_minor(dev->render); + drm_put_minor(dev->primary); kfree(dev); return NULL; } @@ -594,26 +632,22 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags) mutex_lock(&drm_global_mutex); - if (drm_core_check_feature(dev, DRIVER_MODESET)) { - ret = drm_get_minor(dev, &dev->control, DRM_MINOR_CONTROL); - if (ret) - goto out_unlock; - } - - if (drm_core_check_feature(dev, DRIVER_RENDER) && drm_rnodes) { - ret = drm_get_minor(dev, &dev->render, DRM_MINOR_RENDER); - if (ret) - goto err_control_node; - } - - ret = drm_get_minor(dev, &dev->primary, DRM_MINOR_LEGACY); + ret = drm_get_minor(dev, DRM_MINOR_CONTROL); if (ret) - goto err_render_node; + goto err_minors; + + ret = drm_get_minor(dev, DRM_MINOR_RENDER); + if (ret) + goto err_minors; + + ret = drm_get_minor(dev, DRM_MINOR_LEGACY); + if (ret) + goto err_minors; if (dev->driver->load) { ret = dev->driver->load(dev, flags); if (ret) - goto err_primary_node; + goto err_minors; } /* setup grouping for legacy outputs */ @@ -630,12 +664,10 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags) err_unload: if (dev->driver->unload) dev->driver->unload(dev); -err_primary_node: - drm_unplug_minor(dev->primary); -err_render_node: - drm_unplug_minor(dev->render); -err_control_node: +err_minors: drm_unplug_minor(dev->control); + drm_unplug_minor(dev->render); + drm_unplug_minor(dev->primary); out_unlock: mutex_unlock(&drm_global_mutex); return ret; From bd9dfa98187f6cb671e60d9df0801378e8a99ad9 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 12:55:48 +0100 Subject: [PATCH 240/259] drm: move drm_put_minor() to drm_minor_free() _put/get() are used for ref-counting, which we clearly don't do here. Rename it to _free() and also use the common drm_minor_* prefix. Furthermore, avoid passing the minor directly but instead use the type like the other functions do, this allows us to reset the slot. We also drop the redundant call to drm_unplug_minor() as drm_minor_free() is only used from paths were that has already be called. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_stub.c | 45 +++++++++++++++----------------------- 1 file changed, 18 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index b595b64291a2..e46c0904a706 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -291,6 +291,17 @@ static int drm_minor_alloc(struct drm_device *dev, unsigned int type) return 0; } +static void drm_minor_free(struct drm_device *dev, unsigned int type) +{ + struct drm_minor **slot; + + slot = drm_minor_get_slot(dev, type); + if (*slot) { + kfree(*slot); + *slot = NULL; + } +} + /** * drm_get_minor - Register DRM minor * @dev: DRM device @@ -413,26 +424,6 @@ void drm_minor_release(struct drm_minor *minor) drm_dev_unref(minor->dev); } -/** - * drm_put_minor - Destroy DRM minor - * @minor: Minor to destroy - * - * This calls drm_unplug_minor() on the given minor and then frees it. Nothing - * is done if @minor is NULL. It is fine to call this on already unplugged - * minors. - * The global DRM mutex must be held by the caller. - */ -static void drm_put_minor(struct drm_minor *minor) -{ - if (!minor) - return; - - DRM_DEBUG("release secondary minor %d\n", minor->index); - - drm_unplug_minor(minor); - kfree(minor); -} - /** * Called via drm_exit() at module unload time or when pci device is * unplugged. @@ -554,9 +545,9 @@ struct drm_device *drm_dev_alloc(struct drm_driver *driver, err_ht: drm_ht_remove(&dev->map_hash); err_minors: - drm_put_minor(dev->control); - drm_put_minor(dev->render); - drm_put_minor(dev->primary); + drm_minor_free(dev, DRM_MINOR_LEGACY); + drm_minor_free(dev, DRM_MINOR_RENDER); + drm_minor_free(dev, DRM_MINOR_CONTROL); kfree(dev); return NULL; } @@ -566,16 +557,16 @@ static void drm_dev_release(struct kref *ref) { struct drm_device *dev = container_of(ref, struct drm_device, ref); - drm_put_minor(dev->control); - drm_put_minor(dev->render); - drm_put_minor(dev->primary); - if (dev->driver->driver_features & DRIVER_GEM) drm_gem_destroy(dev); drm_ctxbitmap_cleanup(dev); drm_ht_remove(&dev->map_hash); + drm_minor_free(dev, DRM_MINOR_LEGACY); + drm_minor_free(dev, DRM_MINOR_RENDER); + drm_minor_free(dev, DRM_MINOR_CONTROL); + kfree(dev->devname); kfree(dev); } From afcdbc867460b7ee4119bf4904e60f0e171c6dfb Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 12:57:05 +0100 Subject: [PATCH 241/259] drm: rename drm_unplug/get_minor() to drm_minor_register/unregister() drm_get_minor() no longer allocates objects, and drm_unplug_minor() is now the exact reverse of it. Rename it to _register/unregister() so their name actually says what they do. Furthermore, remove the direct minor-ptr and instead pass the minor-type. This way we know the actual slot of the minor and can reset it if required. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_stub.c | 54 ++++++++++++-------------------------- 1 file changed, 17 insertions(+), 37 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index e46c0904a706..488600013e35 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -302,18 +302,7 @@ static void drm_minor_free(struct drm_device *dev, unsigned int type) } } -/** - * drm_get_minor - Register DRM minor - * @dev: DRM device - * @type: Type of minor - * - * Register minor of given type. - * Caller must hold the global DRM mutex. - * - * RETURNS: - * 0 on success, negative error code on failure. - */ -static int drm_get_minor(struct drm_device *dev, unsigned int type) +static int drm_minor_register(struct drm_device *dev, unsigned int type) { struct drm_minor *new_minor; int ret; @@ -362,18 +351,11 @@ static int drm_get_minor(struct drm_device *dev, unsigned int type) return ret; } -/** - * drm_unplug_minor - Unplug DRM minor - * @minor: Minor to unplug - * - * Unplugs the given DRM minor but keeps the object. So after this returns, - * minor->dev is still valid so existing open-files can still access it to get - * device information from their drm_file ojects. - * If the minor is already unplugged or if @minor is NULL, nothing is done. - * The global DRM mutex must be held by the caller. - */ -static void drm_unplug_minor(struct drm_minor *minor) +static void drm_minor_unregister(struct drm_device *dev, unsigned int type) { + struct drm_minor *minor; + + minor = *drm_minor_get_slot(dev, type); if (!minor || !minor->kdev) return; @@ -448,11 +430,9 @@ EXPORT_SYMBOL(drm_put_dev); void drm_unplug_dev(struct drm_device *dev) { /* for a USB device */ - if (drm_core_check_feature(dev, DRIVER_MODESET)) - drm_unplug_minor(dev->control); - if (dev->render) - drm_unplug_minor(dev->render); - drm_unplug_minor(dev->primary); + drm_minor_unregister(dev, DRM_MINOR_LEGACY); + drm_minor_unregister(dev, DRM_MINOR_RENDER); + drm_minor_unregister(dev, DRM_MINOR_CONTROL); mutex_lock(&drm_global_mutex); @@ -623,15 +603,15 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags) mutex_lock(&drm_global_mutex); - ret = drm_get_minor(dev, DRM_MINOR_CONTROL); + ret = drm_minor_register(dev, DRM_MINOR_CONTROL); if (ret) goto err_minors; - ret = drm_get_minor(dev, DRM_MINOR_RENDER); + ret = drm_minor_register(dev, DRM_MINOR_RENDER); if (ret) goto err_minors; - ret = drm_get_minor(dev, DRM_MINOR_LEGACY); + ret = drm_minor_register(dev, DRM_MINOR_LEGACY); if (ret) goto err_minors; @@ -656,9 +636,9 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags) if (dev->driver->unload) dev->driver->unload(dev); err_minors: - drm_unplug_minor(dev->control); - drm_unplug_minor(dev->render); - drm_unplug_minor(dev->primary); + drm_minor_unregister(dev, DRM_MINOR_LEGACY); + drm_minor_unregister(dev, DRM_MINOR_RENDER); + drm_minor_unregister(dev, DRM_MINOR_CONTROL); out_unlock: mutex_unlock(&drm_global_mutex); return ret; @@ -690,8 +670,8 @@ void drm_dev_unregister(struct drm_device *dev) list_for_each_entry_safe(r_list, list_temp, &dev->maplist, head) drm_rmmap(dev, r_list->map); - drm_unplug_minor(dev->control); - drm_unplug_minor(dev->render); - drm_unplug_minor(dev->primary); + drm_minor_unregister(dev, DRM_MINOR_LEGACY); + drm_minor_unregister(dev, DRM_MINOR_RENDER); + drm_minor_unregister(dev, DRM_MINOR_CONTROL); } EXPORT_SYMBOL(drm_dev_unregister); From cb0f93238b89c6178842ba89ecc1cd311f1a3e75 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 13:01:08 +0100 Subject: [PATCH 242/259] drm: remove unneeded #ifdef CONFIG_DEBUGFS No need to check for DEBUGFS, we already have dummy-fallbacks in our headers. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_stub.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 488600013e35..fe9595b750ea 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -323,13 +323,11 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) idr_replace(&drm_minors_idr, new_minor, minor_id); -#if defined(CONFIG_DEBUG_FS) ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root); if (ret) { DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n"); goto err_mem; } -#endif ret = drm_sysfs_device_add(new_minor); if (ret) { @@ -343,10 +341,8 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) err_debugfs: -#if defined(CONFIG_DEBUG_FS) drm_debugfs_cleanup(new_minor); err_mem: -#endif idr_remove(&drm_minors_idr, minor_id); return ret; } @@ -359,10 +355,7 @@ static void drm_minor_unregister(struct drm_device *dev, unsigned int type) if (!minor || !minor->kdev) return; -#if defined(CONFIG_DEBUG_FS) drm_debugfs_cleanup(minor); -#endif - drm_sysfs_device_remove(minor); idr_remove(&drm_minors_idr, minor->index); } From 5817878c6f4221c3ace4af63260080635063371e Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Wed, 29 Jan 2014 13:12:31 +0100 Subject: [PATCH 243/259] drm: remove redundant minor->device field Whenever we access minor->device, we are in a minor->kdev->...->fops callback so the minor->kdev pointer *must* be valid. Thus, simply use minor->kdev->devt instead of minor->device and remove the redundant field. Signed-off-by: David Herrmann Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_drv.c | 4 ++-- drivers/gpu/drm/drm_fops.c | 2 +- drivers/gpu/drm/drm_stub.c | 1 - include/drm/drmP.h | 1 - 4 files changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 345be03c23db..ec651be2f3cb 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -344,7 +344,7 @@ long drm_ioctl(struct file *filp, DRM_DEBUG("pid=%d, dev=0x%lx, auth=%d, %s\n", task_pid_nr(current), - (long)old_encode_dev(file_priv->minor->device), + (long)old_encode_dev(file_priv->minor->kdev->devt), file_priv->authenticated, ioctl->name); /* Do not trust userspace, use our own definition */ @@ -402,7 +402,7 @@ long drm_ioctl(struct file *filp, if (!ioctl) DRM_DEBUG("invalid ioctl: pid=%d, dev=0x%lx, auth=%d, cmd=0x%02x, nr=0x%02x\n", task_pid_nr(current), - (long)old_encode_dev(file_priv->minor->device), + (long)old_encode_dev(file_priv->minor->kdev->devt), file_priv->authenticated, cmd, nr); if (kdata != stack_kdata) diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 79478191404a..4ce5318d14bc 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -471,7 +471,7 @@ int drm_release(struct inode *inode, struct file *filp) DRM_DEBUG("pid = %d, device = 0x%lx, open_count = %d\n", task_pid_nr(current), - (long)old_encode_dev(file_priv->minor->device), + (long)old_encode_dev(file_priv->minor->kdev->devt), dev->open_count); /* Release any auth tokens that might point to this file_priv, diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index fe9595b750ea..96fe5dec3822 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -318,7 +318,6 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) if (minor_id < 0) return minor_id; - new_minor->device = MKDEV(DRM_MAJOR, minor_id); new_minor->index = minor_id; idr_replace(&drm_minors_idr, new_minor, minor_id); diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 82963167f161..538079030be0 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -1043,7 +1043,6 @@ struct drm_info_node { struct drm_minor { int index; /**< Minor device number */ int type; /**< Control or render */ - dev_t device; /**< Device number for mknod */ struct device *kdev; /**< Linux device */ struct drm_device *dev; From 1abbc43761793938fba9ae745e01d5b4730a9914 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 24 Feb 2014 15:32:00 +0100 Subject: [PATCH 244/259] drm: coding-style fixes in minor handling Properly name goto-labels, remove empty lines and use DRM_ERROR if possible. Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_stub.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 96fe5dec3822..5268ffc5d94e 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -325,23 +325,21 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root); if (ret) { DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n"); - goto err_mem; + goto err_id; } ret = drm_sysfs_device_add(new_minor); if (ret) { - printk(KERN_ERR - "DRM: Error sysfs_device_add.\n"); + DRM_ERROR("DRM: Error sysfs_device_add.\n"); goto err_debugfs; } DRM_DEBUG("new minor assigned %d\n", minor_id); return 0; - err_debugfs: drm_debugfs_cleanup(new_minor); -err_mem: +err_id: idr_remove(&drm_minors_idr, minor_id); return ret; } From 7d86cf1a4fc0c0bdb6947185c6fe71301dfea7b1 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 24 Feb 2014 15:35:09 +0100 Subject: [PATCH 245/259] drm: inline drm_minor_get_id() We can significantly simplify this helper by using plain multiplication. Note that we converted the minor-type to an enum earlier so this didn't work before. We also fix a minor range-bug here: the limit argument of idr_alloc() is *exclusive*, not inclusive, so we should use 64 instead of 63 as offset. Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_stub.c | 27 ++++++--------------------- 1 file changed, 6 insertions(+), 21 deletions(-) diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 5268ffc5d94e..83ef4a63358c 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -117,26 +117,6 @@ void drm_ut_debug_printk(unsigned int request_level, } EXPORT_SYMBOL(drm_ut_debug_printk); -static int drm_minor_get_id(struct drm_device *dev, int type) -{ - int ret; - int base = 0, limit = 63; - - if (type == DRM_MINOR_CONTROL) { - base += 64; - limit = base + 63; - } else if (type == DRM_MINOR_RENDER) { - base += 128; - limit = base + 63; - } - - mutex_lock(&dev->struct_mutex); - ret = idr_alloc(&drm_minors_idr, NULL, base, limit, GFP_KERNEL); - mutex_unlock(&dev->struct_mutex); - - return ret == -ENOSPC ? -EINVAL : ret; -} - struct drm_master *drm_master_create(struct drm_minor *minor) { struct drm_master *master; @@ -314,7 +294,12 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) if (!new_minor) return 0; - minor_id = drm_minor_get_id(dev, type); + minor_id = idr_alloc(&drm_minors_idr, + NULL, + 64 * type, + 64 * (type + 1), + GFP_KERNEL); + if (minor_id < 0) return minor_id; From 0d639883ee26359e1bf38195df1dbca0f879e239 Mon Sep 17 00:00:00 2001 From: David Herrmann Date: Mon, 24 Feb 2014 15:53:25 +0100 Subject: [PATCH 246/259] drm: make minors independent of global lock We used to protect minor-lookup and setup by the global drm lock. To continue our attempts of dropping drm_global_mutex, this patch makes the minor management independent of it. Furthermore, we make it all atomic and switch to spin-locks instead of a mutex. Now that minor-lookup is independent, we also move the "drm_is_unplugged()" test into the minor-lookup path. There is no reason to ever return a minor for unplugged objects, so keep that logic internal. Signed-off-by: David Herrmann --- drivers/gpu/drm/drm_fops.c | 11 +------- drivers/gpu/drm/drm_stub.c | 53 +++++++++++++++++++++++++++++++++----- 2 files changed, 47 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 4ce5318d14bc..8f46fe273ba3 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -39,7 +39,7 @@ #include #include -/* from BKL pushdown: note that nothing else serializes idr_find() */ +/* from BKL pushdown */ DEFINE_MUTEX(drm_global_mutex); EXPORT_SYMBOL(drm_global_mutex); @@ -91,11 +91,6 @@ int drm_open(struct inode *inode, struct file *filp) return PTR_ERR(minor); dev = minor->dev; - if (drm_device_is_unplugged(dev)) { - retcode = -ENODEV; - goto err_release; - } - if (!dev->open_count++) need_setup = 1; mutex_lock(&dev->struct_mutex); @@ -127,7 +122,6 @@ int drm_open(struct inode *inode, struct file *filp) dev->dev_mapping = old_mapping; mutex_unlock(&dev->struct_mutex); dev->open_count--; -err_release: drm_minor_release(minor); return retcode; } @@ -157,9 +151,6 @@ int drm_stub_open(struct inode *inode, struct file *filp) goto out_unlock; dev = minor->dev; - if (drm_device_is_unplugged(dev)) - goto out_release; - new_fops = fops_get(dev->driver->fops); if (!new_fops) goto out_release; diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 83ef4a63358c..c23eaf6442ff 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -70,6 +70,7 @@ module_param_named(vblankoffdelay, drm_vblank_offdelay, int, 0600); module_param_named(timestamp_precision_usec, drm_timestamp_precision, int, 0600); module_param_named(timestamp_monotonic, drm_timestamp_monotonic, int, 0600); +static DEFINE_SPINLOCK(drm_minor_lock); struct idr drm_minors_idr; struct class *drm_class; @@ -240,6 +241,19 @@ int drm_dropmaster_ioctl(struct drm_device *dev, void *data, return 0; } +/* + * DRM Minors + * A DRM device can provide several char-dev interfaces on the DRM-Major. Each + * of them is represented by a drm_minor object. Depending on the capabilities + * of the device-driver, different interfaces are registered. + * + * Minors can be accessed via dev->$minor_name. This pointer is either + * NULL or a valid drm_minor pointer and stays valid as long as the device is + * valid. This means, DRM minors have the same life-time as the underlying + * device. However, this doesn't mean that the minor is active. Minors are + * registered and unregistered dynamically according to device-state. + */ + static struct drm_minor **drm_minor_get_slot(struct drm_device *dev, unsigned int type) { @@ -285,6 +299,7 @@ static void drm_minor_free(struct drm_device *dev, unsigned int type) static int drm_minor_register(struct drm_device *dev, unsigned int type) { struct drm_minor *new_minor; + unsigned long flags; int ret; int minor_id; @@ -294,19 +309,21 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) if (!new_minor) return 0; + idr_preload(GFP_KERNEL); + spin_lock_irqsave(&drm_minor_lock, flags); minor_id = idr_alloc(&drm_minors_idr, NULL, 64 * type, 64 * (type + 1), - GFP_KERNEL); + GFP_NOWAIT); + spin_unlock_irqrestore(&drm_minor_lock, flags); + idr_preload_end(); if (minor_id < 0) return minor_id; new_minor->index = minor_id; - idr_replace(&drm_minors_idr, new_minor, minor_id); - ret = drm_debugfs_init(new_minor, minor_id, drm_debugfs_root); if (ret) { DRM_ERROR("DRM: Failed to initialize /sys/kernel/debug/dri.\n"); @@ -319,27 +336,40 @@ static int drm_minor_register(struct drm_device *dev, unsigned int type) goto err_debugfs; } + /* replace NULL with @minor so lookups will succeed from now on */ + spin_lock_irqsave(&drm_minor_lock, flags); + idr_replace(&drm_minors_idr, new_minor, new_minor->index); + spin_unlock_irqrestore(&drm_minor_lock, flags); + DRM_DEBUG("new minor assigned %d\n", minor_id); return 0; err_debugfs: drm_debugfs_cleanup(new_minor); err_id: + spin_lock_irqsave(&drm_minor_lock, flags); idr_remove(&drm_minors_idr, minor_id); + spin_unlock_irqrestore(&drm_minor_lock, flags); + new_minor->index = 0; return ret; } static void drm_minor_unregister(struct drm_device *dev, unsigned int type) { struct drm_minor *minor; + unsigned long flags; minor = *drm_minor_get_slot(dev, type); if (!minor || !minor->kdev) return; + spin_lock_irqsave(&drm_minor_lock, flags); + idr_remove(&drm_minors_idr, minor->index); + spin_unlock_irqrestore(&drm_minor_lock, flags); + minor->index = 0; + drm_debugfs_cleanup(minor); drm_sysfs_device_remove(minor); - idr_remove(&drm_minors_idr, minor->index); } /** @@ -361,12 +391,21 @@ static void drm_minor_unregister(struct drm_device *dev, unsigned int type) struct drm_minor *drm_minor_acquire(unsigned int minor_id) { struct drm_minor *minor; + unsigned long flags; + spin_lock_irqsave(&drm_minor_lock, flags); minor = idr_find(&drm_minors_idr, minor_id); - if (!minor) - return ERR_PTR(-ENODEV); + if (minor) + drm_dev_ref(minor->dev); + spin_unlock_irqrestore(&drm_minor_lock, flags); + + if (!minor) { + return ERR_PTR(-ENODEV); + } else if (drm_device_is_unplugged(minor->dev)) { + drm_dev_unref(minor->dev); + return ERR_PTR(-ENODEV); + } - drm_dev_ref(minor->dev); return minor; } From 4f87dac386cc43d5525da7a939d4b4e7edbea22c Mon Sep 17 00:00:00 2001 From: Michael Kerrisk Date: Mon, 10 Mar 2014 14:46:07 +0100 Subject: [PATCH 247/259] ipc: Fix 2 bugs in msgrcv() MSG_COPY implementation While testing and documenting the msgrcv() MSG_COPY flag that Stanislav Kinsbursky added in commit 4a674f34ba04 ("ipc: introduce message queue copy feature" => kernel 3.8), I discovered a couple of bugs in the implementation. The two bugs concern MSG_COPY interactions with other msgrcv() flags, namely: (A) MSG_COPY + MSG_EXCEPT (B) MSG_COPY + !IPC_NOWAIT The bugs are distinct (and the fix for the first one is obvious), however my fix for both is a single-line patch, which is why I'm combining them in a single mail, rather than writing two mails+patches. ===== (A) MSG_COPY + MSG_EXCEPT ===== With the addition of the MSG_COPY flag, there are now two msgrcv() flags--MSG_COPY and MSG_EXCEPT--that modify the meaning of the 'msgtyp' argument in unrelated ways. Specifying both in the same call is a logical error that is currently permitted, with the effect that MSG_COPY has priority and MSG_EXCEPT is ignored. The call should give an error if both flags are specified. The patch below implements that behavior. ===== (B) (B) MSG_COPY + !IPC_NOWAIT ===== The test code that was submitted in commit 3a665531a3b7 ("selftests: IPC message queue copy feature test") shows MSG_COPY being used in conjunction with IPC_NOWAIT. In other words, if there is no message at the position 'msgtyp'. return immediately with the error in ENOMSG. What was not (fully) tested is the behavior if MSG_COPY is specified *without* IPC_NOWAIT, and there is an odd behavior. If the queue contains less than 'msgtyp' messages, then the call blocks until the next message is written to the queue. At that point, the msgrcv() call returns a copy of the newly added message, regardless of whether that message is at the ordinal position 'msgtyp'. This is clearly bogus, and problematic for applications that might want to make use of the MSG_COPY flag. I considered the following possible solutions to this problem: (1) Force the call to block until a message *does* appear at the position 'msgtyp'. (2) If the MSG_COPY flag is specified, the kernel should implicitly add IPC_NOWAIT, so that the call fails with ENOMSG for this case. (3) If the MSG_COPY flag is specified, but IPC_NOWAIT is not, generate an error (probably, EINVAL is the right one). I do not know if any application would really want to have the functionality of solution (1), especially since an application can determine in advance the number of messages in the queue using msgctl() IPC_STAT. Obviously, this solution would be the most work to implement. Solution (2) would have the effect of silently fixing any applications that tried to employ broken behavior. However, it would mean that if we later decided to implement solution (1), then user-space could not easily detect what the kernel supports (but, since I'm somewhat doubtful that solution (1) is needed, I'm not sure that this is much of a problem). Solution (3) would have the effect of informing broken applications that they are doing something broken. The downside is that this would cause a ABI breakage for any applications that are currently employing the broken behavior. However: a) Those applications are almost certainly not getting the results they expect. b) Possibly, those applications don't even exist, because MSG_COPY is currently hidden behind CONFIG_CHECKPOINT_RESTORE. The upside of solution (3) is that if we later decided to implement solution (1), user-space could determine what the kernel supports, via the error return. In my view, solution (3) is mildly preferable to solution (2), and solution (1) could still be done later if anyone really cares. The patch below implements solution (3). PS. For anyone out there still listening, it's the usual story: documenting an API (and the thinking about, and the testing of the API, that documentation entails) is the one of the single best ways of finding bugs in the API, as I've learned from a lot of experience. Best to do that documentation before releasing the API. Signed-off-by: Michael Kerrisk Acked-by: Stanislav Kinsbursky Cc: Stanislav Kinsbursky Cc: stable@vger.kernel.org Cc: Serge Hallyn Cc: "Eric W. Biederman" Cc: Pavel Emelyanov Cc: Al Viro Cc: KOSAKI Motohiro Signed-off-by: Linus Torvalds --- ipc/msg.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/ipc/msg.c b/ipc/msg.c index 245db1140ad6..649853105a5d 100644 --- a/ipc/msg.c +++ b/ipc/msg.c @@ -901,6 +901,8 @@ long do_msgrcv(int msqid, void __user *buf, size_t bufsz, long msgtyp, int msgfl return -EINVAL; if (msgflg & MSG_COPY) { + if ((msgflg & MSG_EXCEPT) || !(msgflg & IPC_NOWAIT)) + return -EINVAL; copy = prepare_copy(buf, min_t(size_t, bufsz, ns->msg_ctlmax)); if (IS_ERR(copy)) return PTR_ERR(copy); From dcb99fd9b08cfe1afe426af4d8d3cbc429190f15 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 16 Mar 2014 18:51:24 -0700 Subject: [PATCH 248/259] Linux 3.14-rc7 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 1a2628ee5d91..ef779ec26f62 100644 --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ VERSION = 3 PATCHLEVEL = 14 SUBLEVEL = 0 -EXTRAVERSION = -rc6 +EXTRAVERSION = -rc7 NAME = Shuffling Zombie Juror # *DOCUMENTATION* From 366d48070008a0846a099b23efef297451b05640 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Wed, 26 Feb 2014 22:16:03 +0200 Subject: [PATCH 249/259] drm/fb-helper: Use drm_fb_helper_restore_fbdev_mode() in drm_fb_helper_set_par() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use drm_fb_helper_restore_fbdev_mode() in drm_fb_helper_set_par() to make sure extra planes get disabled whenever fbcon takes over. Otherwise the code in drm_fb_helper_set_par() was already doing the exact same thing as drm_fb_helper_restore_fbdev_mode(), so this doesn't change the behaviour in any other way. Signed-off-by: Ville Syrjälä Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_fb_helper.c | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index 98a03639b413..e5208e072ac7 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -809,8 +809,6 @@ int drm_fb_helper_set_par(struct fb_info *info) struct drm_fb_helper *fb_helper = info->par; struct drm_device *dev = fb_helper->dev; struct fb_var_screeninfo *var = &info->var; - int ret; - int i; if (var->pixclock != 0) { DRM_ERROR("PIXEL CLOCK SET\n"); @@ -818,13 +816,7 @@ int drm_fb_helper_set_par(struct fb_info *info) } drm_modeset_lock_all(dev); - for (i = 0; i < fb_helper->crtc_count; i++) { - ret = drm_mode_set_config_internal(&fb_helper->crtc_info[i].mode_set); - if (ret) { - drm_modeset_unlock_all(dev); - return ret; - } - } + drm_fb_helper_restore_fbdev_mode(fb_helper); drm_modeset_unlock_all(dev); if (fb_helper->delayed_hotplug) { From 409bbf1e3da29aaf57b520e29f904db9c7c2475e Mon Sep 17 00:00:00 2001 From: Damien Lespiau Date: Mon, 3 Mar 2014 23:59:07 +0000 Subject: [PATCH 250/259] drm: Check if the allocation has succeeded before dereferencing newmode We allocate memory in drm_display_mode_from_vic_index() and use it without checking the pointer is valid. Fix that. Signed-off-by: Damien Lespiau Reviewed-by: Alex Deucher Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_edid.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index f8d8a1de9573..f3cde90c1d98 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -2580,6 +2580,9 @@ drm_display_mode_from_vic_index(struct drm_connector *connector, return NULL; newmode = drm_mode_duplicate(dev, &edid_cea_modes[cea_mode]); + if (!newmode) + return NULL; + newmode->vrefresh = 0; return newmode; From 04cfe97eb11c28848d14ba8c88124da31a9f881c Mon Sep 17 00:00:00 2001 From: Xiubo Li Date: Mon, 10 Mar 2014 09:33:58 +0800 Subject: [PATCH 251/259] drm/fb-helper: Do the 'max_conn_count' zero check Since we cannot make sure the 'max_conn_count' will always be none zero from the users, and then if max_conn_count equals to zero, the kcalloc() will return ZERO_SIZE_PTR, which equals to ((void *)16). So this patch fix this with just doing the 'max_conn_count' zero check in the front of drm_fb_helper_init(). Signed-off-by: Xiubo Li CC: Jani Nikula Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_fb_helper.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index e5208e072ac7..89382dce1d59 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -516,6 +516,9 @@ int drm_fb_helper_init(struct drm_device *dev, struct drm_crtc *crtc; int i; + if (!max_conn_count) + return -EINVAL; + fb_helper->dev = dev; INIT_LIST_HEAD(&fb_helper->kernel_fb_list); From c94adc4a65c67a79f0d19285bf5c32fe4c00176f Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 30 Jan 2014 17:58:38 +0100 Subject: [PATCH 252/259] drm: Fix use-after-free in the shadow-attache exit code This regression has been introduced in commit b3f2333de8e81b089262b26d52272911523e605f Author: Daniel Vetter Date: Wed Dec 11 11:34:31 2013 +0100 drm: restrict the device list for shadow attached drivers Reported-by: Dave Jones Cc: Dave Jones Cc: Dave Airlie Cc: David Herrmann Signed-off-by: Daniel Vetter Reviewed-by: David Herrmann --- drivers/gpu/drm/drm_pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c index 5736aaa7e86c..f7af69bcf3f4 100644 --- a/drivers/gpu/drm/drm_pci.c +++ b/drivers/gpu/drm/drm_pci.c @@ -468,8 +468,8 @@ void drm_pci_exit(struct drm_driver *driver, struct pci_driver *pdriver) } else { list_for_each_entry_safe(dev, tmp, &driver->legacy_dev_list, legacy_dev_list) { - drm_put_dev(dev); list_del(&dev->legacy_dev_list); + drm_put_dev(dev); } } DRM_INFO("Module unloaded\n"); From 7550e3668ce1414f6c7edbe8b13c1169a1aa5960 Mon Sep 17 00:00:00 2001 From: Joonyoung Shim Date: Sat, 15 Mar 2014 16:30:28 +0900 Subject: [PATCH 253/259] drm/cma: remove to make sg_table when gem cma is created The sg_table made when gem cma is created isn't used anywhere. The sgt of struct drm_gem_cma_object will have only sg_tabel imported. Signed-off-by: Joonyoung Shim Acked-by: Laurent Pinchart Signed-off-by: Dave Airlie --- drivers/gpu/drm/drm_gem_cma_helper.c | 19 ------------------- 1 file changed, 19 deletions(-) diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c index 6b51bf90df0e..2c07cb9550ef 100644 --- a/drivers/gpu/drm/drm_gem_cma_helper.c +++ b/drivers/gpu/drm/drm_gem_cma_helper.c @@ -79,7 +79,6 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, unsigned int size) { struct drm_gem_cma_object *cma_obj; - struct sg_table *sgt = NULL; int ret; size = round_up(size, PAGE_SIZE); @@ -97,23 +96,9 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, goto error; } - sgt = kzalloc(sizeof(*cma_obj->sgt), GFP_KERNEL); - if (sgt == NULL) { - ret = -ENOMEM; - goto error; - } - - ret = dma_get_sgtable(drm->dev, sgt, cma_obj->vaddr, - cma_obj->paddr, size); - if (ret < 0) - goto error; - - cma_obj->sgt = sgt; - return cma_obj; error: - kfree(sgt); drm_gem_cma_free_object(&cma_obj->base); return ERR_PTR(ret); } @@ -175,10 +160,6 @@ void drm_gem_cma_free_object(struct drm_gem_object *gem_obj) if (cma_obj->vaddr) { dma_free_writecombine(gem_obj->dev->dev, cma_obj->base.size, cma_obj->vaddr, cma_obj->paddr); - if (cma_obj->sgt) { - sg_free_table(cma_obj->sgt); - kfree(cma_obj->sgt); - } } else if (gem_obj->import_attach) { drm_prime_gem_destroy(gem_obj, cma_obj->sgt); } From 9dc4056026e0df30f6b29109e1e7a6958e7bea62 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:12 +0200 Subject: [PATCH 254/259] drm/dp: let drivers specify the name of the I2C-over-AUX adapter Let the drivers specify the name of the I2C-over-AUX adapter to maintain backwards compatibility in the sysfs when converting to the new I2C-over-AUX helper infrastructure. The i915 driver currently uses DPDDC-A to DPDDC-D as names for the DP i2c adapters. These names show up in the i2c sysfs name attribute. We'd like to be able to maintain that when switching over to the new helpers. Due to i2c device and connector cleanup ordering issues we also recently made the drm device (instead of connector) the parent of the i2c adapters: commit 80f65de3c9b8101c1613fa82df500ba6a099a11c Author: Imre Deak Date: Tue Feb 11 17:12:49 2014 +0200 drm/i915: dp: fix order of dp aux i2c device cleanup With the name picked up from the adapter parent using dev_name(), it would be the same for all i2c adapters with the current I2C-over-AUX helpers. Signed-off-by: Jani Nikula Reviewed-by: Thierry Reding Acked-by: Dave Airlie Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_dp_helper.c | 3 ++- include/drm/drm_dp_helper.h | 4 ++++ 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 35251af3b14e..17832d048147 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -726,7 +726,8 @@ int drm_dp_aux_register_i2c_bus(struct drm_dp_aux *aux) aux->ddc.dev.parent = aux->dev; aux->ddc.dev.of_node = aux->dev->of_node; - strncpy(aux->ddc.name, dev_name(aux->dev), sizeof(aux->ddc.name)); + strlcpy(aux->ddc.name, aux->name ? aux->name : dev_name(aux->dev), + sizeof(aux->ddc.name)); return i2c_add_adapter(&aux->ddc); } diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 42947566e755..b4f58914bf7d 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -438,6 +438,9 @@ struct drm_dp_aux_msg { * The .dev field should be set to a pointer to the device that implements * the AUX channel. * + * The .name field may be used to specify the name of the I2C adapter. If set to + * NULL, dev_name() of .dev will be used. + * * Drivers provide a hardware-specific implementation of how transactions * are executed via the .transfer() function. A pointer to a drm_dp_aux_msg * structure describing the transaction is passed into this function. Upon @@ -455,6 +458,7 @@ struct drm_dp_aux_msg { * should call drm_dp_aux_unregister_i2c_bus() to remove the I2C adapter. */ struct drm_dp_aux { + const char *name; struct i2c_adapter ddc; struct device *dev; From adddaaf4885403a2f2180fb522b5b97e1469b328 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:13 +0200 Subject: [PATCH 255/259] drm/i915/dp: split edp_panel_vdd_on() for reuse Introduce _edp_panel_vdd_on() that returns true if the call enabled vdd, and a matching disable is needed. Keep edp_panel_vdd_on() as a helper for when it is expected the vdd is off. Signed-off-by: Jani Nikula Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_dp.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 49d12d341ab2..b463769b93e5 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -91,6 +91,7 @@ static struct intel_dp *intel_attached_dp(struct drm_connector *connector) } static void intel_dp_link_down(struct intel_dp *intel_dp); +static bool _edp_panel_vdd_on(struct intel_dp *intel_dp); static void edp_panel_vdd_on(struct intel_dp *intel_dp); static void edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync); @@ -1162,23 +1163,21 @@ static u32 ironlake_get_pp_control(struct intel_dp *intel_dp) return control; } -static void edp_panel_vdd_on(struct intel_dp *intel_dp) +static bool _edp_panel_vdd_on(struct intel_dp *intel_dp) { struct drm_device *dev = intel_dp_to_dev(intel_dp); struct drm_i915_private *dev_priv = dev->dev_private; u32 pp; u32 pp_stat_reg, pp_ctrl_reg; + bool need_to_disable = !intel_dp->want_panel_vdd; if (!is_edp(intel_dp)) - return; - - WARN(intel_dp->want_panel_vdd, - "eDP VDD already requested on\n"); + return false; intel_dp->want_panel_vdd = true; if (edp_have_panel_vdd(intel_dp)) - return; + return need_to_disable; intel_runtime_pm_get(dev_priv); @@ -1204,6 +1203,17 @@ static void edp_panel_vdd_on(struct intel_dp *intel_dp) DRM_DEBUG_KMS("eDP was not running\n"); msleep(intel_dp->panel_power_up_delay); } + + return need_to_disable; +} + +static void edp_panel_vdd_on(struct intel_dp *intel_dp) +{ + if (is_edp(intel_dp)) { + bool vdd = _edp_panel_vdd_on(intel_dp); + + WARN(!vdd, "eDP VDD already requested on\n"); + } } static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp) From 884f19e948894fc87b03b631fd03a0998c0ca1ef Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:14 +0200 Subject: [PATCH 256/259] drm/i915/dp: move edp vdd enable/disable at a lower level in i2c-over-aux This is prep work for conversion to generic drm i2c-over-aux helpers where we won't have the function to do this at. Signed-off-by: Jani Nikula Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_dp.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index b463769b93e5..17d73511e148 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -461,6 +461,9 @@ intel_dp_aux_ch(struct intel_dp *intel_dp, uint32_t status; int try, clock = 0; bool has_aux_irq = HAS_AUX_IRQ(dev); + bool vdd; + + vdd = _edp_panel_vdd_on(intel_dp); /* dp aux is extremely sensitive to irq latency, hence request the * lowest possible wakeup latency and so prevent the cpu from going into @@ -566,6 +569,9 @@ intel_dp_aux_ch(struct intel_dp *intel_dp, pm_qos_update_request(&dev_priv->pm_qos, PM_QOS_DEFAULT_VALUE); intel_aux_display_runtime_put(dev_priv); + if (vdd) + edp_panel_vdd_off(intel_dp, false); + return ret; } @@ -678,8 +684,6 @@ intel_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode, int reply_bytes; int ret; - edp_panel_vdd_on(intel_dp); - intel_dp_check_edp(intel_dp); /* Set up the command byte */ if (mode & MODE_I2C_READ) msg[0] = DP_AUX_I2C_READ << 4; @@ -781,7 +785,6 @@ intel_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode, ret = -EREMOTEIO; out: - edp_panel_vdd_off(intel_dp, false); return ret; } From 9d1a1031e84f30f3671f0a650fc38a7c588acc8a Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:15 +0200 Subject: [PATCH 257/259] drm/i915/dp: use the new drm helpers for dp aux Functionality remains largely the same as before. Note that the retry loops and native reply handling all moved into the core drm helper functions now. Signed-off-by: Jani Nikula [danvet: Fix up the stray ; Rodrigo spotted in his review and add a note to the commit message to answer Rodrigo's question in his review.] Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_dp.c | 263 ++++++++++++++----------------- drivers/gpu/drm/i915/intel_drv.h | 1 + 2 files changed, 119 insertions(+), 145 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 17d73511e148..b31f6db5d0c0 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -575,97 +575,77 @@ intel_dp_aux_ch(struct intel_dp *intel_dp, return ret; } -/* Write data to the aux channel in native mode */ -static int -intel_dp_aux_native_write(struct intel_dp *intel_dp, - uint16_t address, uint8_t *send, int send_bytes) +#define HEADER_SIZE 4 +static ssize_t +intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { + struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux); + uint8_t txbuf[20], rxbuf[20]; + size_t txsize, rxsize; int ret; - uint8_t msg[20]; - int msg_bytes; - uint8_t ack; - int retry; - if (WARN_ON(send_bytes > 16)) - return -E2BIG; + txbuf[0] = msg->request << 4; + txbuf[1] = msg->address >> 8; + txbuf[2] = msg->address & 0xff; + txbuf[3] = msg->size - 1; - intel_dp_check_edp(intel_dp); - msg[0] = DP_AUX_NATIVE_WRITE << 4; - msg[1] = address >> 8; - msg[2] = address & 0xff; - msg[3] = send_bytes - 1; - memcpy(&msg[4], send, send_bytes); - msg_bytes = send_bytes + 4; - for (retry = 0; retry < 7; retry++) { - ret = intel_dp_aux_ch(intel_dp, msg, msg_bytes, &ack, 1); - if (ret < 0) - return ret; - ack >>= 4; - if ((ack & DP_AUX_NATIVE_REPLY_MASK) == DP_AUX_NATIVE_REPLY_ACK) - return send_bytes; - else if ((ack & DP_AUX_NATIVE_REPLY_MASK) == DP_AUX_NATIVE_REPLY_DEFER) - usleep_range(400, 500); - else - return -EIO; - } + switch (msg->request & ~DP_AUX_I2C_MOT) { + case DP_AUX_NATIVE_WRITE: + case DP_AUX_I2C_WRITE: + txsize = HEADER_SIZE + msg->size; + rxsize = 1; - DRM_ERROR("too many retries, giving up\n"); - return -EIO; -} + if (WARN_ON(txsize > 20)) + return -E2BIG; -/* Write a single byte to the aux channel in native mode */ -static int -intel_dp_aux_native_write_1(struct intel_dp *intel_dp, - uint16_t address, uint8_t byte) -{ - return intel_dp_aux_native_write(intel_dp, address, &byte, 1); -} + memcpy(txbuf + HEADER_SIZE, msg->buffer, msg->size); -/* read bytes from a native aux channel */ -static int -intel_dp_aux_native_read(struct intel_dp *intel_dp, - uint16_t address, uint8_t *recv, int recv_bytes) -{ - uint8_t msg[4]; - int msg_bytes; - uint8_t reply[20]; - int reply_bytes; - uint8_t ack; - int ret; - int retry; + ret = intel_dp_aux_ch(intel_dp, txbuf, txsize, rxbuf, rxsize); + if (ret > 0) { + msg->reply = rxbuf[0] >> 4; - if (WARN_ON(recv_bytes > 19)) - return -E2BIG; - - intel_dp_check_edp(intel_dp); - msg[0] = DP_AUX_NATIVE_READ << 4; - msg[1] = address >> 8; - msg[2] = address & 0xff; - msg[3] = recv_bytes - 1; - - msg_bytes = 4; - reply_bytes = recv_bytes + 1; - - for (retry = 0; retry < 7; retry++) { - ret = intel_dp_aux_ch(intel_dp, msg, msg_bytes, - reply, reply_bytes); - if (ret == 0) - return -EPROTO; - if (ret < 0) - return ret; - ack = reply[0] >> 4; - if ((ack & DP_AUX_NATIVE_REPLY_MASK) == DP_AUX_NATIVE_REPLY_ACK) { - memcpy(recv, reply + 1, ret - 1); - return ret - 1; + /* Return payload size. */ + ret = msg->size; } - else if ((ack & DP_AUX_NATIVE_REPLY_MASK) == DP_AUX_NATIVE_REPLY_DEFER) - usleep_range(400, 500); - else - return -EIO; + break; + + case DP_AUX_NATIVE_READ: + case DP_AUX_I2C_READ: + txsize = HEADER_SIZE; + rxsize = msg->size + 1; + + if (WARN_ON(rxsize > 20)) + return -E2BIG; + + ret = intel_dp_aux_ch(intel_dp, txbuf, txsize, rxbuf, rxsize); + if (ret > 0) { + msg->reply = rxbuf[0] >> 4; + /* + * Assume happy day, and copy the data. The caller is + * expected to check msg->reply before touching it. + * + * Return payload size. + */ + ret--; + memcpy(msg->buffer, rxbuf + 1, ret); + } + break; + + default: + ret = -EINVAL; + break; } - DRM_ERROR("too many retries, giving up\n"); - return -EIO; + return ret; +} + +static void +intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) +{ + struct drm_device *dev = intel_dp_to_dev(intel_dp); + + intel_dp->aux.dev = dev->dev; + intel_dp->aux.transfer = intel_dp_aux_transfer; } static int @@ -1472,8 +1452,8 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode) return; if (mode != DRM_MODE_DPMS_ON) { - ret = intel_dp_aux_native_write_1(intel_dp, DP_SET_POWER, - DP_SET_POWER_D3); + ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, + DP_SET_POWER_D3); if (ret != 1) DRM_DEBUG_DRIVER("failed to write sink power state\n"); } else { @@ -1482,9 +1462,8 @@ void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode) * time to wake up. */ for (i = 0; i < 3; i++) { - ret = intel_dp_aux_native_write_1(intel_dp, - DP_SET_POWER, - DP_SET_POWER_D0); + ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, + DP_SET_POWER_D0); if (ret == 1) break; msleep(1); @@ -1708,13 +1687,11 @@ static void intel_edp_psr_enable_sink(struct intel_dp *intel_dp) /* Enable PSR in sink */ if (intel_dp->psr_dpcd[1] & DP_PSR_NO_TRAIN_ON_EXIT) - intel_dp_aux_native_write_1(intel_dp, DP_PSR_EN_CFG, - DP_PSR_ENABLE & - ~DP_PSR_MAIN_LINK_ACTIVE); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, + DP_PSR_ENABLE & ~DP_PSR_MAIN_LINK_ACTIVE); else - intel_dp_aux_native_write_1(intel_dp, DP_PSR_EN_CFG, - DP_PSR_ENABLE | - DP_PSR_MAIN_LINK_ACTIVE); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, + DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); /* Setup AUX registers */ I915_WRITE(EDP_PSR_AUX_DATA1(dev), EDP_PSR_DPCD_COMMAND); @@ -2026,26 +2003,25 @@ static void vlv_dp_pre_pll_enable(struct intel_encoder *encoder) /* * Native read with retry for link status and receiver capability reads for * cases where the sink may still be asleep. + * + * Sinks are *supposed* to come up within 1ms from an off state, but we're also + * supposed to retry 3 times per the spec. */ -static bool -intel_dp_aux_native_read_retry(struct intel_dp *intel_dp, uint16_t address, - uint8_t *recv, int recv_bytes) +static ssize_t +intel_dp_dpcd_read_wake(struct drm_dp_aux *aux, unsigned int offset, + void *buffer, size_t size) { - int ret, i; + ssize_t ret; + int i; - /* - * Sinks are *supposed* to come up within 1ms from an off state, - * but we're also supposed to retry 3 times per the spec. - */ for (i = 0; i < 3; i++) { - ret = intel_dp_aux_native_read(intel_dp, address, recv, - recv_bytes); - if (ret == recv_bytes) - return true; + ret = drm_dp_dpcd_read(aux, offset, buffer, size); + if (ret == size) + return ret; msleep(1); } - return false; + return ret; } /* @@ -2055,10 +2031,10 @@ intel_dp_aux_native_read_retry(struct intel_dp *intel_dp, uint16_t address, static bool intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE]) { - return intel_dp_aux_native_read_retry(intel_dp, - DP_LANE0_1_STATUS, - link_status, - DP_LINK_STATUS_SIZE); + return intel_dp_dpcd_read_wake(&intel_dp->aux, + DP_LANE0_1_STATUS, + link_status, + DP_LINK_STATUS_SIZE) == DP_LINK_STATUS_SIZE; } /* @@ -2572,8 +2548,8 @@ intel_dp_set_link_train(struct intel_dp *intel_dp, len = intel_dp->lane_count + 1; } - ret = intel_dp_aux_native_write(intel_dp, DP_TRAINING_PATTERN_SET, - buf, len); + ret = drm_dp_dpcd_write(&intel_dp->aux, DP_TRAINING_PATTERN_SET, + buf, len); return ret == len; } @@ -2602,9 +2578,8 @@ intel_dp_update_link_train(struct intel_dp *intel_dp, uint32_t *DP, I915_WRITE(intel_dp->output_reg, *DP); POSTING_READ(intel_dp->output_reg); - ret = intel_dp_aux_native_write(intel_dp, DP_TRAINING_LANE0_SET, - intel_dp->train_set, - intel_dp->lane_count); + ret = drm_dp_dpcd_write(&intel_dp->aux, DP_TRAINING_LANE0_SET, + intel_dp->train_set, intel_dp->lane_count); return ret == intel_dp->lane_count; } @@ -2660,11 +2635,11 @@ intel_dp_start_link_train(struct intel_dp *intel_dp) link_config[1] = intel_dp->lane_count; if (drm_dp_enhanced_frame_cap(intel_dp->dpcd)) link_config[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN; - intel_dp_aux_native_write(intel_dp, DP_LINK_BW_SET, link_config, 2); + drm_dp_dpcd_write(&intel_dp->aux, DP_LINK_BW_SET, link_config, 2); link_config[0] = 0; link_config[1] = DP_SET_ANSI_8B10B; - intel_dp_aux_native_write(intel_dp, DP_DOWNSPREAD_CTRL, link_config, 2); + drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2); DP |= DP_PORT_EN; @@ -2907,8 +2882,8 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp) char dpcd_hex_dump[sizeof(intel_dp->dpcd) * 3]; - if (intel_dp_aux_native_read_retry(intel_dp, 0x000, intel_dp->dpcd, - sizeof(intel_dp->dpcd)) == 0) + if (intel_dp_dpcd_read_wake(&intel_dp->aux, 0x000, intel_dp->dpcd, + sizeof(intel_dp->dpcd)) < 0) return false; /* aux transfer failed */ hex_dump_to_buffer(intel_dp->dpcd, sizeof(intel_dp->dpcd), @@ -2921,9 +2896,9 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp) /* Check if the panel supports PSR */ memset(intel_dp->psr_dpcd, 0, sizeof(intel_dp->psr_dpcd)); if (is_edp(intel_dp)) { - intel_dp_aux_native_read_retry(intel_dp, DP_PSR_SUPPORT, - intel_dp->psr_dpcd, - sizeof(intel_dp->psr_dpcd)); + intel_dp_dpcd_read_wake(&intel_dp->aux, DP_PSR_SUPPORT, + intel_dp->psr_dpcd, + sizeof(intel_dp->psr_dpcd)); if (intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED) { dev_priv->psr.sink_support = true; DRM_DEBUG_KMS("Detected EDP PSR Panel.\n"); @@ -2945,9 +2920,9 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp) if (intel_dp->dpcd[DP_DPCD_REV] == 0x10) return true; /* no per-port downstream info */ - if (intel_dp_aux_native_read_retry(intel_dp, DP_DOWNSTREAM_PORT_0, - intel_dp->downstream_ports, - DP_MAX_DOWNSTREAM_PORTS) == 0) + if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_DOWNSTREAM_PORT_0, + intel_dp->downstream_ports, + DP_MAX_DOWNSTREAM_PORTS) < 0) return false; /* downstream port status fetch failed */ return true; @@ -2963,11 +2938,11 @@ intel_dp_probe_oui(struct intel_dp *intel_dp) edp_panel_vdd_on(intel_dp); - if (intel_dp_aux_native_read_retry(intel_dp, DP_SINK_OUI, buf, 3)) + if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_SINK_OUI, buf, 3) == 3) DRM_DEBUG_KMS("Sink OUI: %02hx%02hx%02hx\n", buf[0], buf[1], buf[2]); - if (intel_dp_aux_native_read_retry(intel_dp, DP_BRANCH_OUI, buf, 3)) + if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_BRANCH_OUI, buf, 3) == 3) DRM_DEBUG_KMS("Branch OUI: %02hx%02hx%02hx\n", buf[0], buf[1], buf[2]); @@ -2982,46 +2957,40 @@ int intel_dp_sink_crc(struct intel_dp *intel_dp, u8 *crc) to_intel_crtc(intel_dig_port->base.base.crtc); u8 buf[1]; - if (!intel_dp_aux_native_read(intel_dp, DP_TEST_SINK_MISC, buf, 1)) + if (drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_SINK_MISC, buf) < 0) return -EAGAIN; if (!(buf[0] & DP_TEST_CRC_SUPPORTED)) return -ENOTTY; - if (!intel_dp_aux_native_write_1(intel_dp, DP_TEST_SINK, - DP_TEST_SINK_START)) + if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK, + DP_TEST_SINK_START) < 0) return -EAGAIN; /* Wait 2 vblanks to be sure we will have the correct CRC value */ intel_wait_for_vblank(dev, intel_crtc->pipe); intel_wait_for_vblank(dev, intel_crtc->pipe); - if (!intel_dp_aux_native_read(intel_dp, DP_TEST_CRC_R_CR, crc, 6)) + if (drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_CRC_R_CR, crc, 6) < 0) return -EAGAIN; - intel_dp_aux_native_write_1(intel_dp, DP_TEST_SINK, 0); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_SINK, 0); return 0; } static bool intel_dp_get_sink_irq(struct intel_dp *intel_dp, u8 *sink_irq_vector) { - int ret; - - ret = intel_dp_aux_native_read_retry(intel_dp, - DP_DEVICE_SERVICE_IRQ_VECTOR, - sink_irq_vector, 1); - if (!ret) - return false; - - return true; + return intel_dp_dpcd_read_wake(&intel_dp->aux, + DP_DEVICE_SERVICE_IRQ_VECTOR, + sink_irq_vector, 1) == 1; } static void intel_dp_handle_test_request(struct intel_dp *intel_dp) { /* NAK by default */ - intel_dp_aux_native_write_1(intel_dp, DP_TEST_RESPONSE, DP_TEST_NAK); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_RESPONSE, DP_TEST_NAK); } /* @@ -3060,9 +3029,9 @@ intel_dp_check_link_status(struct intel_dp *intel_dp) if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11 && intel_dp_get_sink_irq(intel_dp, &sink_irq_vector)) { /* Clear interrupt source */ - intel_dp_aux_native_write_1(intel_dp, - DP_DEVICE_SERVICE_IRQ_VECTOR, - sink_irq_vector); + drm_dp_dpcd_writeb(&intel_dp->aux, + DP_DEVICE_SERVICE_IRQ_VECTOR, + sink_irq_vector); if (sink_irq_vector & DP_AUTOMATED_TEST_REQUEST) intel_dp_handle_test_request(intel_dp); @@ -3097,9 +3066,11 @@ intel_dp_detect_dpcd(struct intel_dp *intel_dp) if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11 && intel_dp->downstream_ports[0] & DP_DS_PORT_HPD) { uint8_t reg; - if (!intel_dp_aux_native_read_retry(intel_dp, DP_SINK_COUNT, - ®, 1)) + + if (intel_dp_dpcd_read_wake(&intel_dp->aux, DP_SINK_COUNT, + ®, 1) < 0) return connector_status_unknown; + return DP_GET_SINK_COUNT(reg) ? connector_status_connected : connector_status_disconnected; } @@ -3925,6 +3896,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq); } + intel_dp_aux_init(intel_dp, intel_connector); + error = intel_dp_i2c_init(intel_dp, intel_connector, name); WARN(error, "intel_dp_i2c_init failed with error %d for port %c\n", error, port_name(port)); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 9c7090590776..578c18ed982c 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -487,6 +487,7 @@ struct intel_dp { uint8_t downstream_ports[DP_MAX_DOWNSTREAM_PORTS]; struct i2c_adapter adapter; struct i2c_algo_dp_aux_data algo; + struct drm_dp_aux aux; uint8_t train_set[4]; int panel_power_up_delay; int panel_power_down_delay; From 33ad6626a1a9155fcbb04869c7cdde0552976396 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:16 +0200 Subject: [PATCH 258/259] drm/i915/dp: move dp aux ch register init to aux init Do a slight rearrangement of the switch to prep for follow-up. Signed-off-by: Jani Nikula Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_dp.c | 42 +++++++++++++++++---------------- 1 file changed, 22 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index b31f6db5d0c0..623b50c46bbc 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -643,6 +643,28 @@ static void intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) { struct drm_device *dev = intel_dp_to_dev(intel_dp); + struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); + enum port port = intel_dig_port->port; + + switch (port) { + case PORT_A: + intel_dp->aux_ch_ctl_reg = DPA_AUX_CH_CTL; + break; + case PORT_B: + intel_dp->aux_ch_ctl_reg = PCH_DPB_AUX_CH_CTL; + break; + case PORT_C: + intel_dp->aux_ch_ctl_reg = PCH_DPC_AUX_CH_CTL; + break; + case PORT_D: + intel_dp->aux_ch_ctl_reg = PCH_DPD_AUX_CH_CTL; + break; + default: + BUG(); + } + + if (!HAS_DDI(dev)) + intel_dp->aux_ch_ctl_reg = intel_dp->output_reg + 0x10; intel_dp->aux.dev = dev->dev; intel_dp->aux.transfer = intel_dp_aux_transfer; @@ -3849,26 +3871,6 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, intel_connector->get_hw_state = intel_connector_get_hw_state; intel_connector->unregister = intel_dp_connector_unregister; - intel_dp->aux_ch_ctl_reg = intel_dp->output_reg + 0x10; - if (HAS_DDI(dev)) { - switch (intel_dig_port->port) { - case PORT_A: - intel_dp->aux_ch_ctl_reg = DPA_AUX_CH_CTL; - break; - case PORT_B: - intel_dp->aux_ch_ctl_reg = PCH_DPB_AUX_CH_CTL; - break; - case PORT_C: - intel_dp->aux_ch_ctl_reg = PCH_DPC_AUX_CH_CTL; - break; - case PORT_D: - intel_dp->aux_ch_ctl_reg = PCH_DPD_AUX_CH_CTL; - break; - default: - BUG(); - } - } - /* Set up the DDC bus. */ switch (port) { case PORT_A: From 0b99836f238f37a8632a3ab4f9a8cc2346a36d40 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Fri, 14 Mar 2014 16:51:17 +0200 Subject: [PATCH 259/259] drm/i915/dp: use the new drm helpers for dp i2c-over-aux The functionality remains largerly the same. The main difference is that i2c-over-aux defer timeouts are increased to be safe for all use cases instead of depending on DP device type and properties. Signed-off-by: Jani Nikula Reviewed-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_dp.c | 197 +++++-------------------------- drivers/gpu/drm/i915/intel_drv.h | 2 - 2 files changed, 30 insertions(+), 169 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 623b50c46bbc..160d5b30b3bc 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -645,19 +645,25 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) struct drm_device *dev = intel_dp_to_dev(intel_dp); struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); enum port port = intel_dig_port->port; + const char *name = NULL; + int ret; switch (port) { case PORT_A: intel_dp->aux_ch_ctl_reg = DPA_AUX_CH_CTL; + name = "DPDDC-A"; break; case PORT_B: intel_dp->aux_ch_ctl_reg = PCH_DPB_AUX_CH_CTL; + name = "DPDDC-B"; break; case PORT_C: intel_dp->aux_ch_ctl_reg = PCH_DPC_AUX_CH_CTL; + name = "DPDDC-C"; break; case PORT_D: intel_dp->aux_ch_ctl_reg = PCH_DPD_AUX_CH_CTL; + name = "DPDDC-D"; break; default: BUG(); @@ -666,128 +672,27 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) if (!HAS_DDI(dev)) intel_dp->aux_ch_ctl_reg = intel_dp->output_reg + 0x10; + intel_dp->aux.name = name; intel_dp->aux.dev = dev->dev; intel_dp->aux.transfer = intel_dp_aux_transfer; -} -static int -intel_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode, - uint8_t write_byte, uint8_t *read_byte) -{ - struct i2c_algo_dp_aux_data *algo_data = adapter->algo_data; - struct intel_dp *intel_dp = container_of(adapter, - struct intel_dp, - adapter); - uint16_t address = algo_data->address; - uint8_t msg[5]; - uint8_t reply[2]; - unsigned retry; - int msg_bytes; - int reply_bytes; - int ret; + DRM_DEBUG_KMS("registering %s bus for %s\n", name, + connector->base.kdev->kobj.name); - /* Set up the command byte */ - if (mode & MODE_I2C_READ) - msg[0] = DP_AUX_I2C_READ << 4; - else - msg[0] = DP_AUX_I2C_WRITE << 4; - - if (!(mode & MODE_I2C_STOP)) - msg[0] |= DP_AUX_I2C_MOT << 4; - - msg[1] = address >> 8; - msg[2] = address; - - switch (mode) { - case MODE_I2C_WRITE: - msg[3] = 0; - msg[4] = write_byte; - msg_bytes = 5; - reply_bytes = 1; - break; - case MODE_I2C_READ: - msg[3] = 0; - msg_bytes = 4; - reply_bytes = 2; - break; - default: - msg_bytes = 3; - reply_bytes = 1; - break; + ret = drm_dp_aux_register_i2c_bus(&intel_dp->aux); + if (ret < 0) { + DRM_ERROR("drm_dp_aux_register_i2c_bus() for %s failed (%d)\n", + name, ret); + return; } - /* - * DP1.2 sections 2.7.7.1.5.6.1 and 2.7.7.1.6.6.1: A DP Source device is - * required to retry at least seven times upon receiving AUX_DEFER - * before giving up the AUX transaction. - */ - for (retry = 0; retry < 7; retry++) { - ret = intel_dp_aux_ch(intel_dp, - msg, msg_bytes, - reply, reply_bytes); - if (ret < 0) { - DRM_DEBUG_KMS("aux_ch failed %d\n", ret); - goto out; - } - - switch ((reply[0] >> 4) & DP_AUX_NATIVE_REPLY_MASK) { - case DP_AUX_NATIVE_REPLY_ACK: - /* I2C-over-AUX Reply field is only valid - * when paired with AUX ACK. - */ - break; - case DP_AUX_NATIVE_REPLY_NACK: - DRM_DEBUG_KMS("aux_ch native nack\n"); - ret = -EREMOTEIO; - goto out; - case DP_AUX_NATIVE_REPLY_DEFER: - /* - * For now, just give more slack to branch devices. We - * could check the DPCD for I2C bit rate capabilities, - * and if available, adjust the interval. We could also - * be more careful with DP-to-Legacy adapters where a - * long legacy cable may force very low I2C bit rates. - */ - if (intel_dp->dpcd[DP_DOWNSTREAMPORT_PRESENT] & - DP_DWN_STRM_PORT_PRESENT) - usleep_range(500, 600); - else - usleep_range(300, 400); - continue; - default: - DRM_ERROR("aux_ch invalid native reply 0x%02x\n", - reply[0]); - ret = -EREMOTEIO; - goto out; - } - - switch ((reply[0] >> 4) & DP_AUX_I2C_REPLY_MASK) { - case DP_AUX_I2C_REPLY_ACK: - if (mode == MODE_I2C_READ) { - *read_byte = reply[1]; - } - ret = reply_bytes - 1; - goto out; - case DP_AUX_I2C_REPLY_NACK: - DRM_DEBUG_KMS("aux_i2c nack\n"); - ret = -EREMOTEIO; - goto out; - case DP_AUX_I2C_REPLY_DEFER: - DRM_DEBUG_KMS("aux_i2c defer\n"); - udelay(100); - break; - default: - DRM_ERROR("aux_i2c invalid reply 0x%02x\n", reply[0]); - ret = -EREMOTEIO; - goto out; - } + ret = sysfs_create_link(&connector->base.kdev->kobj, + &intel_dp->aux.ddc.dev.kobj, + intel_dp->aux.ddc.dev.kobj.name); + if (ret < 0) { + DRM_ERROR("sysfs_create_link() for %s failed (%d)\n", name, ret); + drm_dp_aux_unregister_i2c_bus(&intel_dp->aux); } - - DRM_ERROR("too many retries, giving up\n"); - ret = -EREMOTEIO; - -out: - return ret; } static void @@ -796,43 +701,10 @@ intel_dp_connector_unregister(struct intel_connector *intel_connector) struct intel_dp *intel_dp = intel_attached_dp(&intel_connector->base); sysfs_remove_link(&intel_connector->base.kdev->kobj, - intel_dp->adapter.dev.kobj.name); + intel_dp->aux.ddc.dev.kobj.name); intel_connector_unregister(intel_connector); } -static int -intel_dp_i2c_init(struct intel_dp *intel_dp, - struct intel_connector *intel_connector, const char *name) -{ - int ret; - - DRM_DEBUG_KMS("i2c_init %s\n", name); - intel_dp->algo.running = false; - intel_dp->algo.address = 0; - intel_dp->algo.aux_ch = intel_dp_i2c_aux_ch; - - memset(&intel_dp->adapter, '\0', sizeof(intel_dp->adapter)); - intel_dp->adapter.owner = THIS_MODULE; - intel_dp->adapter.class = I2C_CLASS_DDC; - strncpy(intel_dp->adapter.name, name, sizeof(intel_dp->adapter.name) - 1); - intel_dp->adapter.name[sizeof(intel_dp->adapter.name) - 1] = '\0'; - intel_dp->adapter.algo_data = &intel_dp->algo; - intel_dp->adapter.dev.parent = intel_connector->base.dev->dev; - - ret = i2c_dp_aux_add_bus(&intel_dp->adapter); - if (ret < 0) - return ret; - - ret = sysfs_create_link(&intel_connector->base.kdev->kobj, - &intel_dp->adapter.dev.kobj, - intel_dp->adapter.dev.kobj.name); - - if (ret < 0) - i2c_del_adapter(&intel_dp->adapter); - - return ret; -} - static void intel_dp_set_clock(struct intel_encoder *encoder, struct intel_crtc_config *pipe_config, int link_bw) @@ -3098,7 +2970,7 @@ intel_dp_detect_dpcd(struct intel_dp *intel_dp) } /* If no HPD, poke DDC gently */ - if (drm_probe_ddc(&intel_dp->adapter)) + if (drm_probe_ddc(&intel_dp->aux.ddc)) return connector_status_connected; /* Well we tried, say unknown for unreliable port types */ @@ -3266,7 +3138,7 @@ intel_dp_detect(struct drm_connector *connector, bool force) if (intel_dp->force_audio != HDMI_AUDIO_AUTO) { intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON); } else { - edid = intel_dp_get_edid(connector, &intel_dp->adapter); + edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc); if (edid) { intel_dp->has_audio = drm_detect_monitor_audio(edid); kfree(edid); @@ -3302,7 +3174,7 @@ static int intel_dp_get_modes(struct drm_connector *connector) power_domain = intel_display_port_power_domain(intel_encoder); intel_display_power_get(dev_priv, power_domain); - ret = intel_dp_get_edid_modes(connector, &intel_dp->adapter); + ret = intel_dp_get_edid_modes(connector, &intel_dp->aux.ddc); intel_display_power_put(dev_priv, power_domain); if (ret) return ret; @@ -3335,7 +3207,7 @@ intel_dp_detect_audio(struct drm_connector *connector) power_domain = intel_display_port_power_domain(intel_encoder); intel_display_power_get(dev_priv, power_domain); - edid = intel_dp_get_edid(connector, &intel_dp->adapter); + edid = intel_dp_get_edid(connector, &intel_dp->aux.ddc); if (edid) { has_audio = drm_detect_monitor_audio(edid); kfree(edid); @@ -3457,7 +3329,7 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder) struct intel_dp *intel_dp = &intel_dig_port->dp; struct drm_device *dev = intel_dp_to_dev(intel_dp); - i2c_del_adapter(&intel_dp->adapter); + drm_dp_aux_unregister_i2c_bus(&intel_dp->aux); drm_encoder_cleanup(encoder); if (is_edp(intel_dp)) { cancel_delayed_work_sync(&intel_dp->panel_vdd_work); @@ -3769,7 +3641,7 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, /* We now know it's not a ghost, init power sequence regs. */ intel_dp_init_panel_power_sequencer_registers(dev, intel_dp, power_seq); - edid = drm_get_edid(connector, &intel_dp->adapter); + edid = drm_get_edid(connector, &intel_dp->aux.ddc); if (edid) { if (drm_add_edid_modes(connector, edid)) { drm_mode_connector_update_edid_property(connector, @@ -3817,8 +3689,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, struct drm_i915_private *dev_priv = dev->dev_private; enum port port = intel_dig_port->port; struct edp_power_seq power_seq = { 0 }; - const char *name = NULL; - int type, error; + int type; /* intel_dp vfuncs */ if (IS_VALLEYVIEW(dev)) @@ -3871,23 +3742,19 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, intel_connector->get_hw_state = intel_connector_get_hw_state; intel_connector->unregister = intel_dp_connector_unregister; - /* Set up the DDC bus. */ + /* Set up the hotplug pin. */ switch (port) { case PORT_A: intel_encoder->hpd_pin = HPD_PORT_A; - name = "DPDDC-A"; break; case PORT_B: intel_encoder->hpd_pin = HPD_PORT_B; - name = "DPDDC-B"; break; case PORT_C: intel_encoder->hpd_pin = HPD_PORT_C; - name = "DPDDC-C"; break; case PORT_D: intel_encoder->hpd_pin = HPD_PORT_D; - name = "DPDDC-D"; break; default: BUG(); @@ -3900,14 +3767,10 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, intel_dp_aux_init(intel_dp, intel_connector); - error = intel_dp_i2c_init(intel_dp, intel_connector, name); - WARN(error, "intel_dp_i2c_init failed with error %d for port %c\n", - error, port_name(port)); - intel_dp->psr_setup_done = false; if (!intel_edp_init_connector(intel_dp, intel_connector, &power_seq)) { - i2c_del_adapter(&intel_dp->adapter); + drm_dp_aux_unregister_i2c_bus(&intel_dp->aux); if (is_edp(intel_dp)) { cancel_delayed_work_sync(&intel_dp->panel_vdd_work); mutex_lock(&dev->mode_config.mutex); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 578c18ed982c..5ca293b3ef5c 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -485,8 +485,6 @@ struct intel_dp { uint8_t dpcd[DP_RECEIVER_CAP_SIZE]; uint8_t psr_dpcd[EDP_PSR_RECEIVER_CAP_SIZE]; uint8_t downstream_ports[DP_MAX_DOWNSTREAM_PORTS]; - struct i2c_adapter adapter; - struct i2c_algo_dp_aux_data algo; struct drm_dp_aux aux; uint8_t train_set[4]; int panel_power_up_delay;