From dec8b38be4b35cae5f7fa086daf2631e2cfa09c1 Mon Sep 17 00:00:00 2001 From: Victor Shih Date: Thu, 31 Jul 2025 14:57:50 +0800 Subject: [PATCH 1/5] mmc: sdhci-pci-gli: Add a new function to simplify the code In preparation to fix replay timer timeout, add sdhci_gli_mask_replay_timer_timeout() function to simplify some of the code, allowing it to be re-used. Signed-off-by: Victor Shih Fixes: 1ae1d2d6e555 ("mmc: sdhci-pci-gli: Add Genesys Logic GL9763E support") Cc: stable@vger.kernel.org Acked-by: Adrian Hunter Link: https://lore.kernel.org/r/20250731065752.450231-2-victorshihgli@gmail.com Signed-off-by: Ulf Hansson --- drivers/mmc/host/sdhci-pci-gli.c | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/drivers/mmc/host/sdhci-pci-gli.c b/drivers/mmc/host/sdhci-pci-gli.c index 4c2ae71770f7..f678c91f8d3e 100644 --- a/drivers/mmc/host/sdhci-pci-gli.c +++ b/drivers/mmc/host/sdhci-pci-gli.c @@ -287,6 +287,20 @@ #define GLI_MAX_TUNING_LOOP 40 /* Genesys Logic chipset */ +static void sdhci_gli_mask_replay_timer_timeout(struct pci_dev *pdev) +{ + int aer; + u32 value; + + /* mask the replay timer timeout of AER */ + aer = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_ERR); + if (aer) { + pci_read_config_dword(pdev, aer + PCI_ERR_COR_MASK, &value); + value |= PCI_ERR_COR_REP_TIMER; + pci_write_config_dword(pdev, aer + PCI_ERR_COR_MASK, value); + } +} + static inline void gl9750_wt_on(struct sdhci_host *host) { u32 wt_value; @@ -607,7 +621,6 @@ static void gl9750_hw_setting(struct sdhci_host *host) { struct sdhci_pci_slot *slot = sdhci_priv(host); struct pci_dev *pdev; - int aer; u32 value; pdev = slot->chip->pdev; @@ -626,12 +639,7 @@ static void gl9750_hw_setting(struct sdhci_host *host) pci_set_power_state(pdev, PCI_D0); /* mask the replay timer timeout of AER */ - aer = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_ERR); - if (aer) { - pci_read_config_dword(pdev, aer + PCI_ERR_COR_MASK, &value); - value |= PCI_ERR_COR_REP_TIMER; - pci_write_config_dword(pdev, aer + PCI_ERR_COR_MASK, value); - } + sdhci_gli_mask_replay_timer_timeout(pdev); gl9750_wt_off(host); } @@ -806,7 +814,6 @@ static void sdhci_gl9755_set_clock(struct sdhci_host *host, unsigned int clock) static void gl9755_hw_setting(struct sdhci_pci_slot *slot) { struct pci_dev *pdev = slot->chip->pdev; - int aer; u32 value; gl9755_wt_on(pdev); @@ -841,12 +848,7 @@ static void gl9755_hw_setting(struct sdhci_pci_slot *slot) pci_set_power_state(pdev, PCI_D0); /* mask the replay timer timeout of AER */ - aer = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_ERR); - if (aer) { - pci_read_config_dword(pdev, aer + PCI_ERR_COR_MASK, &value); - value |= PCI_ERR_COR_REP_TIMER; - pci_write_config_dword(pdev, aer + PCI_ERR_COR_MASK, value); - } + sdhci_gli_mask_replay_timer_timeout(pdev); gl9755_wt_off(pdev); } From 293ed0f5f34e1e9df888456af4b0a021f57b5f54 Mon Sep 17 00:00:00 2001 From: Victor Shih Date: Thu, 31 Jul 2025 14:57:51 +0800 Subject: [PATCH 2/5] mmc: sdhci-pci-gli: GL9763e: Rename the gli_set_gl9763e() for consistency In preparation to fix replay timer timeout, rename the gli_set_gl9763e() to gl9763e_hw_setting() for consistency. Signed-off-by: Victor Shih Fixes: 1ae1d2d6e555 ("mmc: sdhci-pci-gli: Add Genesys Logic GL9763E support") Cc: stable@vger.kernel.org Acked-by: Adrian Hunter Link: https://lore.kernel.org/r/20250731065752.450231-3-victorshihgli@gmail.com Signed-off-by: Ulf Hansson --- drivers/mmc/host/sdhci-pci-gli.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/mmc/host/sdhci-pci-gli.c b/drivers/mmc/host/sdhci-pci-gli.c index f678c91f8d3e..436f0460222f 100644 --- a/drivers/mmc/host/sdhci-pci-gli.c +++ b/drivers/mmc/host/sdhci-pci-gli.c @@ -1753,7 +1753,7 @@ static int gl9763e_add_host(struct sdhci_pci_slot *slot) return ret; } -static void gli_set_gl9763e(struct sdhci_pci_slot *slot) +static void gl9763e_hw_setting(struct sdhci_pci_slot *slot) { struct pci_dev *pdev = slot->chip->pdev; u32 value; @@ -1925,7 +1925,7 @@ static int gli_probe_slot_gl9763e(struct sdhci_pci_slot *slot) gli_pcie_enable_msi(slot); host->mmc_host_ops.hs400_enhanced_strobe = gl9763e_hs400_enhanced_strobe; - gli_set_gl9763e(slot); + gl9763e_hw_setting(slot); sdhci_enable_v4_mode(host); return 0; From 340be332e420ed37d15d4169a1b4174e912ad6cb Mon Sep 17 00:00:00 2001 From: Victor Shih Date: Thu, 31 Jul 2025 14:57:52 +0800 Subject: [PATCH 3/5] mmc: sdhci-pci-gli: GL9763e: Mask the replay timer timeout of AER Due to a flaw in the hardware design, the GL9763e replay timer frequently times out when ASPM is enabled. As a result, the warning messages will often appear in the system log when the system accesses the GL9763e PCI config. Therefore, the replay timer timeout must be masked. Signed-off-by: Victor Shih Fixes: 1ae1d2d6e555 ("mmc: sdhci-pci-gli: Add Genesys Logic GL9763E support") Cc: stable@vger.kernel.org Acked-by: Adrian Hunter Link: https://lore.kernel.org/r/20250731065752.450231-4-victorshihgli@gmail.com Signed-off-by: Ulf Hansson --- drivers/mmc/host/sdhci-pci-gli.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/mmc/host/sdhci-pci-gli.c b/drivers/mmc/host/sdhci-pci-gli.c index 436f0460222f..3a1de477e9af 100644 --- a/drivers/mmc/host/sdhci-pci-gli.c +++ b/drivers/mmc/host/sdhci-pci-gli.c @@ -1782,6 +1782,9 @@ static void gl9763e_hw_setting(struct sdhci_pci_slot *slot) value |= FIELD_PREP(GLI_9763E_HS400_RXDLY, GLI_9763E_HS400_RXDLY_5); pci_write_config_dword(pdev, PCIE_GLI_9763E_CLKRXDLY, value); + /* mask the replay timer timeout of AER */ + sdhci_gli_mask_replay_timer_timeout(pdev); + pci_read_config_dword(pdev, PCIE_GLI_9763E_VHS, &value); value &= ~GLI_9763E_VHS_REV; value |= FIELD_PREP(GLI_9763E_VHS_REV, GLI_9763E_VHS_REV_R); From e251709aaddb3ee1e8ac1ed5e361a608a1cc92de Mon Sep 17 00:00:00 2001 From: Sai Krishna Potthuri Date: Wed, 30 Jul 2025 11:35:43 +0530 Subject: [PATCH 4/5] mmc: sdhci-of-arasan: Ensure CD logic stabilization before power-up During SD suspend/resume without a full card rescan (when using non-removable SD cards for rootfs), the SD card initialization may fail after resume. This occurs because, after a host controller reset, the card detect logic may take time to stabilize due to debounce logic. Without waiting for stabilization, the host may attempt powering up the card prematurely, leading to command timeouts during resume flow. Add sdhci_arasan_set_power_and_bus_voltage() to wait for the card detect stable bit before power up the card. Since the stabilization time is not fixed, a maximum timeout of one second is used to ensure sufficient wait time for the card detect signal to stabilize. Signed-off-by: Sai Krishna Potthuri Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20250730060543.1735971-1-sai.krishna.potthuri@amd.com Signed-off-by: Ulf Hansson --- drivers/mmc/host/sdhci-of-arasan.c | 33 ++++++++++++++++++++++++++++-- 1 file changed, 31 insertions(+), 2 deletions(-) diff --git a/drivers/mmc/host/sdhci-of-arasan.c b/drivers/mmc/host/sdhci-of-arasan.c index 42878474e56e..60dbc815e501 100644 --- a/drivers/mmc/host/sdhci-of-arasan.c +++ b/drivers/mmc/host/sdhci-of-arasan.c @@ -99,6 +99,9 @@ #define HIWORD_UPDATE(val, mask, shift) \ ((val) << (shift) | (mask) << ((shift) + 16)) +#define CD_STABLE_TIMEOUT_US 1000000 +#define CD_STABLE_MAX_SLEEP_US 10 + /** * struct sdhci_arasan_soc_ctl_field - Field used in sdhci_arasan_soc_ctl_map * @@ -206,12 +209,15 @@ struct sdhci_arasan_data { * 19MHz instead */ #define SDHCI_ARASAN_QUIRK_CLOCK_25_BROKEN BIT(2) +/* Enable CD stable check before power-up */ +#define SDHCI_ARASAN_QUIRK_ENSURE_CD_STABLE BIT(3) }; struct sdhci_arasan_of_data { const struct sdhci_arasan_soc_ctl_map *soc_ctl_map; const struct sdhci_pltfm_data *pdata; const struct sdhci_arasan_clk_ops *clk_ops; + u32 quirks; }; static const struct sdhci_arasan_soc_ctl_map rk3399_soc_ctl_map = { @@ -514,6 +520,24 @@ static int sdhci_arasan_voltage_switch(struct mmc_host *mmc, return -EINVAL; } +static void sdhci_arasan_set_power_and_bus_voltage(struct sdhci_host *host, unsigned char mode, + unsigned short vdd) +{ + struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host); + struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host); + u32 reg; + + /* + * Ensure that the card detect logic has stabilized before powering up, this is + * necessary after a host controller reset. + */ + if (mode == MMC_POWER_UP && sdhci_arasan->quirks & SDHCI_ARASAN_QUIRK_ENSURE_CD_STABLE) + read_poll_timeout(sdhci_readl, reg, reg & SDHCI_CD_STABLE, CD_STABLE_MAX_SLEEP_US, + CD_STABLE_TIMEOUT_US, false, host, SDHCI_PRESENT_STATE); + + sdhci_set_power_and_bus_voltage(host, mode, vdd); +} + static const struct sdhci_ops sdhci_arasan_ops = { .set_clock = sdhci_arasan_set_clock, .get_max_clock = sdhci_pltfm_clk_get_max_clock, @@ -521,7 +545,7 @@ static const struct sdhci_ops sdhci_arasan_ops = { .set_bus_width = sdhci_set_bus_width, .reset = sdhci_arasan_reset, .set_uhs_signaling = sdhci_set_uhs_signaling, - .set_power = sdhci_set_power_and_bus_voltage, + .set_power = sdhci_arasan_set_power_and_bus_voltage, .hw_reset = sdhci_arasan_hw_reset, }; @@ -570,7 +594,7 @@ static const struct sdhci_ops sdhci_arasan_cqe_ops = { .set_bus_width = sdhci_set_bus_width, .reset = sdhci_arasan_reset, .set_uhs_signaling = sdhci_set_uhs_signaling, - .set_power = sdhci_set_power_and_bus_voltage, + .set_power = sdhci_arasan_set_power_and_bus_voltage, .irq = sdhci_arasan_cqhci_irq, }; @@ -1447,6 +1471,7 @@ static const struct sdhci_arasan_clk_ops zynqmp_clk_ops = { static struct sdhci_arasan_of_data sdhci_arasan_zynqmp_data = { .pdata = &sdhci_arasan_zynqmp_pdata, .clk_ops = &zynqmp_clk_ops, + .quirks = SDHCI_ARASAN_QUIRK_ENSURE_CD_STABLE, }; static const struct sdhci_arasan_clk_ops versal_clk_ops = { @@ -1457,6 +1482,7 @@ static const struct sdhci_arasan_clk_ops versal_clk_ops = { static struct sdhci_arasan_of_data sdhci_arasan_versal_data = { .pdata = &sdhci_arasan_zynqmp_pdata, .clk_ops = &versal_clk_ops, + .quirks = SDHCI_ARASAN_QUIRK_ENSURE_CD_STABLE, }; static const struct sdhci_arasan_clk_ops versal_net_clk_ops = { @@ -1467,6 +1493,7 @@ static const struct sdhci_arasan_clk_ops versal_net_clk_ops = { static struct sdhci_arasan_of_data sdhci_arasan_versal_net_data = { .pdata = &sdhci_arasan_versal_net_pdata, .clk_ops = &versal_net_clk_ops, + .quirks = SDHCI_ARASAN_QUIRK_ENSURE_CD_STABLE, }; static struct sdhci_arasan_of_data intel_keembay_emmc_data = { @@ -1937,6 +1964,8 @@ static int sdhci_arasan_probe(struct platform_device *pdev) if (of_device_is_compatible(np, "rockchip,rk3399-sdhci-5.1")) sdhci_arasan_update_clockmultiplier(host, 0x0); + sdhci_arasan->quirks |= data->quirks; + if (of_device_is_compatible(np, "intel,keembay-sdhci-5.1-emmc") || of_device_is_compatible(np, "intel,keembay-sdhci-5.1-sd") || of_device_is_compatible(np, "intel,keembay-sdhci-5.1-sdio")) { From 99d7ab8db9d8230b243f5ed20ba0229e54cc0dfa Mon Sep 17 00:00:00 2001 From: Jiayi Li Date: Mon, 4 Aug 2025 09:36:04 +0800 Subject: [PATCH 5/5] memstick: Fix deadlock by moving removing flag earlier The existing memstick core patch: commit 62c59a8786e6 ("memstick: Skip allocating card when removing host") sets host->removing in memstick_remove_host(),but still exists a critical time window where memstick_check can run after host->eject is set but before removing is set. In the rtsx_usb_ms driver, the problematic sequence is: rtsx_usb_ms_drv_remove: memstick_check: host->eject = true cancel_work_sync(handle_req) if(!host->removing) ... memstick_alloc_card() memstick_set_rw_addr() memstick_new_req() rtsx_usb_ms_request() if(!host->eject) skip schedule_work wait_for_completion() memstick_remove_host: [blocks indefinitely] host->removing = true flush_workqueue() [block] 1. rtsx_usb_ms_drv_remove sets host->eject = true 2. cancel_work_sync(&host->handle_req) runs 3. memstick_check work may be executed here <-- danger window 4. memstick_remove_host sets removing = 1 During this window (step 3), memstick_check calls memstick_alloc_card, which may indefinitely waiting for mrq_complete completion that will never occur because rtsx_usb_ms_request sees eject=true and skips scheduling work, memstick_set_rw_addr waits forever for completion. This causes a deadlock when memstick_remove_host tries to flush_workqueue, waiting for memstick_check to complete, while memstick_check is blocked waiting for mrq_complete completion. Fix this by setting removing=true at the start of rtsx_usb_ms_drv_remove, before any work cancellation. This ensures memstick_check will see the removing flag immediately and exit early, avoiding the deadlock. Fixes: 62c59a8786e6 ("memstick: Skip allocating card when removing host") Signed-off-by: Jiayi Li Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20250804013604.1311218-1-lijiayi@kylinos.cn Signed-off-by: Ulf Hansson --- drivers/memstick/core/memstick.c | 1 - drivers/memstick/host/rtsx_usb_ms.c | 1 + 2 files changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/memstick/core/memstick.c b/drivers/memstick/core/memstick.c index 7f3f47db4c98..e4275f8ee5db 100644 --- a/drivers/memstick/core/memstick.c +++ b/drivers/memstick/core/memstick.c @@ -555,7 +555,6 @@ EXPORT_SYMBOL(memstick_add_host); */ void memstick_remove_host(struct memstick_host *host) { - host->removing = 1; flush_workqueue(workqueue); mutex_lock(&host->lock); if (host->card) diff --git a/drivers/memstick/host/rtsx_usb_ms.c b/drivers/memstick/host/rtsx_usb_ms.c index 3878136227e4..5b5e9354fb2e 100644 --- a/drivers/memstick/host/rtsx_usb_ms.c +++ b/drivers/memstick/host/rtsx_usb_ms.c @@ -812,6 +812,7 @@ static void rtsx_usb_ms_drv_remove(struct platform_device *pdev) int err; host->eject = true; + msh->removing = true; cancel_work_sync(&host->handle_req); cancel_delayed_work_sync(&host->poll_card);