eth: fbnic: Add protection against pause storm

Add protection against TX pause storms. A pause storm occurs when a
device fails to send received packets up to the stack. When a pause
storm is detected (pause state persists beyond the configured timeout),
the device stops sending the pause frames and begins dropping packets
instead of back-pressuring.

The timeout is configurable via ethtool tunable (pfc-prevention-tout)
with a maximum value of 10485ms, and the default value of 500ms.

Once the device transitions to the storm-detected state, the service
task periodically attempts recovery, returning the device to normal
operation to handle any subsequent pause storm episodes.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Mohsin Bashir <mohsin.bashr@gmail.com>
Link: https://patch.msgid.link/20260302230149.1580195-4-mohsin.bashr@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This commit is contained in:
Mohsin Bashir 2026-03-02 15:01:47 -08:00 committed by Paolo Abeni
parent 817de93c34
commit 9b7c8728f5
7 changed files with 186 additions and 0 deletions

View File

@ -98,6 +98,9 @@ struct fbnic_dev {
/* MDIO bus for PHYs */
struct mii_bus *mdio_bus;
/* In units of ms since API supports values in ms */
u16 ps_timeout;
};
/* Reserve entry 0 in the MSI-X "others" array until we have filled all

View File

@ -230,6 +230,7 @@ enum {
#define FBNIC_INTR_MSIX_CTRL_VECTOR_MASK CSR_GENMASK(7, 0)
#define FBNIC_INTR_MSIX_CTRL_ENABLE CSR_BIT(31)
enum {
FBNIC_INTR_MSIX_CTRL_RXB_IDX = 7,
FBNIC_INTR_MSIX_CTRL_PCS_IDX = 34,
};
@ -560,6 +561,11 @@ enum {
#define FBNIC_RXB_DROP_THLD_CNT 8
#define FBNIC_RXB_DROP_THLD_ON CSR_GENMASK(12, 0)
#define FBNIC_RXB_DROP_THLD_OFF CSR_GENMASK(25, 13)
#define FBNIC_RXB_PAUSE_STORM(n) (0x08019 + (n)) /* 0x20064 + 4*n */
#define FBNIC_RXB_PAUSE_STORM_CNT 4
#define FBNIC_RXB_PAUSE_STORM_FORCE_NORMAL CSR_BIT(20)
#define FBNIC_RXB_PAUSE_STORM_THLD_TIME CSR_GENMASK(19, 0)
#define FBNIC_RXB_PAUSE_STORM_UNIT_WR 0x0801d /* 0x20074 */
#define FBNIC_RXB_ECN_THLD(n) (0x0801e + (n)) /* 0x20078 + 4*n */
#define FBNIC_RXB_ECN_THLD_CNT 8
#define FBNIC_RXB_ECN_THLD_ON CSR_GENMASK(12, 0)
@ -596,6 +602,9 @@ enum {
#define FBNIC_RXB_INTF_CREDIT_MASK2 CSR_GENMASK(11, 8)
#define FBNIC_RXB_INTF_CREDIT_MASK3 CSR_GENMASK(15, 12)
#define FBNIC_RXB_ERR_INTR_STS 0x08050 /* 0x20140 */
#define FBNIC_RXB_ERR_INTR_STS_PS CSR_GENMASK(15, 12)
#define FBNIC_RXB_ERR_INTR_MASK 0x08052 /* 0x20148 */
#define FBNIC_RXB_PAUSE_EVENT_CNT(n) (0x08053 + (n)) /* 0x2014c + 4*n */
#define FBNIC_RXB_DROP_FRMS_STS(n) (0x08057 + (n)) /* 0x2015c + 4*n */
#define FBNIC_RXB_DROP_BYTES_STS_L(n) \
@ -636,6 +645,7 @@ enum {
#define FBNIC_RXB_PBUF_FIFO_LEVEL(n) (0x0811d + (n)) /* 0x20474 + 4*n */
#define FBNIC_RXB_PAUSE_STORM_UNIT_RD 0x08125 /* 0x20494 */
#define FBNIC_RXB_INTEGRITY_ERR(n) (0x0812f + (n)) /* 0x204bc + 4*n */
#define FBNIC_RXB_MAC_ERR(n) (0x08133 + (n)) /* 0x204cc + 4*n */
#define FBNIC_RXB_PARSER_ERR(n) (0x08137 + (n)) /* 0x204dc + 4*n */

View File

@ -1641,6 +1641,47 @@ static void fbnic_get_ts_stats(struct net_device *netdev,
}
}
static int fbnic_get_tunable(struct net_device *netdev,
const struct ethtool_tunable *tun,
void *data)
{
struct fbnic_net *fbn = netdev_priv(netdev);
int err = 0;
switch (tun->id) {
case ETHTOOL_PFC_PREVENTION_TOUT:
*(u16 *)data = fbn->fbd->ps_timeout;
break;
default:
err = -EOPNOTSUPP;
break;
}
return err;
}
static int fbnic_set_tunable(struct net_device *netdev,
const struct ethtool_tunable *tun,
const void *data)
{
struct fbnic_net *fbn = netdev_priv(netdev);
int err;
switch (tun->id) {
case ETHTOOL_PFC_PREVENTION_TOUT: {
u16 ps_timeout = *(u16 *)data;
err = fbnic_mac_ps_protect_to_config(fbn->fbd, ps_timeout);
break;
}
default:
err = -EOPNOTSUPP;
break;
}
return err;
}
static int
fbnic_get_module_eeprom_by_page(struct net_device *netdev,
const struct ethtool_module_eeprom *page_data,
@ -1915,6 +1956,8 @@ static const struct ethtool_ops fbnic_ethtool_ops = {
.set_channels = fbnic_set_channels,
.get_ts_info = fbnic_get_ts_info,
.get_ts_stats = fbnic_get_ts_stats,
.get_tunable = fbnic_get_tunable,
.set_tunable = fbnic_set_tunable,
.get_link_ksettings = fbnic_phylink_ethtool_ksettings_get,
.get_fec_stats = fbnic_get_fec_stats,
.get_fecparam = fbnic_phylink_get_fecparam,

View File

@ -170,6 +170,8 @@ int fbnic_mac_request_irq(struct fbnic_dev *fbd)
fbnic_wr32(fbd, FBNIC_INTR_MSIX_CTRL(FBNIC_INTR_MSIX_CTRL_PCS_IDX),
FBNIC_PCS_MSIX_ENTRY | FBNIC_INTR_MSIX_CTRL_ENABLE);
fbnic_wr32(fbd, FBNIC_INTR_MSIX_CTRL(FBNIC_INTR_MSIX_CTRL_RXB_IDX), 0);
fbd->mac_msix_vector = vector;
return 0;

View File

@ -143,6 +143,7 @@ static void fbnic_mac_init_qm(struct fbnic_dev *fbd)
#define FBNIC_DROP_EN_MASK 0x7d
#define FBNIC_PAUSE_EN_MASK 0x14
#define FBNIC_ECN_EN_MASK 0x10
#define FBNIC_PS_EN_MASK 0x01
struct fbnic_fifo_config {
unsigned int addr;
@ -420,6 +421,14 @@ static void __fbnic_mac_stat_rd64(struct fbnic_dev *fbd, bool reset, u32 reg,
#define fbnic_mac_stat_rd64(fbd, reset, __stat, __CSR) \
__fbnic_mac_stat_rd64(fbd, reset, FBNIC_##__CSR##_L, &(__stat))
bool fbnic_mac_check_tx_pause(struct fbnic_dev *fbd)
{
u32 command_config;
command_config = rd32(fbd, FBNIC_MAC_COMMAND_CONFIG);
return !(command_config & FBNIC_MAC_COMMAND_CONFIG_TX_PAUSE_DIS);
}
static void fbnic_mac_tx_pause_config(struct fbnic_dev *fbd, bool tx_pause)
{
u32 rxb_pause_ctrl;
@ -434,6 +443,49 @@ static void fbnic_mac_tx_pause_config(struct fbnic_dev *fbd, bool tx_pause)
wr32(fbd, FBNIC_RXB_PAUSE_DROP_CTRL, rxb_pause_ctrl);
}
static void
fbnic_mac_ps_protect_to_reset(struct fbnic_dev *fbd, u16 timeout_ms)
{
wr32(fbd, FBNIC_RXB_PAUSE_STORM_UNIT_WR, FBNIC_RXB_PS_CLK_DIV);
wr32(fbd, FBNIC_RXB_PAUSE_STORM(FBNIC_RXB_INTF_NET),
FIELD_PREP(FBNIC_RXB_PAUSE_STORM_THLD_TIME,
FBNIC_MAC_RXB_PS_TO(timeout_ms)) |
FBNIC_RXB_PAUSE_STORM_FORCE_NORMAL);
wrfl(fbd);
wr32(fbd, FBNIC_RXB_PAUSE_STORM(FBNIC_RXB_INTF_NET),
FIELD_PREP(FBNIC_RXB_PAUSE_STORM_THLD_TIME,
FBNIC_MAC_RXB_PS_TO(timeout_ms)));
}
static void
fbnic_mac_ps_protect_config(struct fbnic_dev *fbd, bool ps_protect)
{
u16 timeout;
u32 reg;
ps_protect = ps_protect && fbd->ps_timeout;
timeout = ps_protect ? fbd->ps_timeout : FBNIC_MAC_PS_TO_DEFAULT_MS;
fbnic_mac_ps_protect_to_reset(fbd, timeout);
reg = rd32(fbd, FBNIC_RXB_PAUSE_DROP_CTRL);
reg &= ~FBNIC_RXB_PAUSE_DROP_CTRL_PS_ENABLE;
reg |= FIELD_PREP(FBNIC_RXB_PAUSE_DROP_CTRL_PS_ENABLE, ps_protect);
wr32(fbd, FBNIC_RXB_PAUSE_DROP_CTRL, reg);
/* Clear any pending interrupt status first */
wr32(fbd, FBNIC_RXB_ERR_INTR_STS,
FIELD_PREP(FBNIC_RXB_ERR_INTR_STS_PS, FBNIC_PS_EN_MASK));
/* Unmask the Network to Host PS interrupt if tx_pause is on */
reg = rd32(fbd, FBNIC_RXB_ERR_INTR_MASK);
reg |= FBNIC_RXB_ERR_INTR_STS_PS;
if (ps_protect)
reg &= ~FBNIC_RXB_ERR_INTR_STS_PS;
wr32(fbd, FBNIC_RXB_ERR_INTR_MASK, reg);
}
static int fbnic_mac_get_link_event(struct fbnic_dev *fbd)
{
u32 intr_mask = rd32(fbd, FBNIC_SIG_PCS_INTR_STS);
@ -658,6 +710,7 @@ static void fbnic_mac_link_up_asic(struct fbnic_dev *fbd,
u32 cmd_cfg, mac_ctrl;
fbnic_mac_tx_pause_config(fbd, tx_pause);
fbnic_mac_ps_protect_config(fbd, tx_pause);
cmd_cfg = __fbnic_mac_cmd_config_asic(fbd, tx_pause, rx_pause);
mac_ctrl = rd32(fbd, FBNIC_SIG_MAC_IN0);
@ -918,3 +971,46 @@ int fbnic_mac_init(struct fbnic_dev *fbd)
return 0;
}
int fbnic_mac_ps_protect_to_config(struct fbnic_dev *fbd, u16 timeout_ms)
{
u16 old_timeout_ms = fbd->ps_timeout;
if (timeout_ms == old_timeout_ms)
return 0;
if (timeout_ms == PFC_STORM_PREVENTION_AUTO)
timeout_ms = FBNIC_MAC_PS_TO_DEFAULT_MS;
if (timeout_ms > FBNIC_MAC_PS_TO_MAX_MS)
return -EINVAL;
fbd->ps_timeout = timeout_ms;
if (!fbnic_mac_check_tx_pause(fbd))
return 0;
if (timeout_ms == 0)
fbnic_mac_ps_protect_config(fbd, false);
else if (old_timeout_ms == 0)
fbnic_mac_ps_protect_config(fbd, true);
else
fbnic_mac_ps_protect_to_reset(fbd, fbd->ps_timeout);
return 0;
}
void fbnic_mac_ps_protect_handler(struct fbnic_dev *fbd)
{
u32 rxb_err_sts = rd32(fbd, FBNIC_RXB_ERR_INTR_STS);
/* Check if pause storm interrupt for network was triggered */
if (rxb_err_sts & FIELD_PREP(FBNIC_RXB_ERR_INTR_STS_PS,
FBNIC_PS_EN_MASK)) {
/* Write 1 to clear the interrupt status first */
wr32(fbd, FBNIC_RXB_ERR_INTR_STS,
FIELD_PREP(FBNIC_RXB_ERR_INTR_STS_PS, FBNIC_PS_EN_MASK));
fbnic_mac_ps_protect_to_reset(fbd, fbd->ps_timeout);
}
}

View File

@ -8,6 +8,30 @@
struct fbnic_dev;
/* The RXB clock runs at 600 MHZ in the ASIC and the PAUSE_STORM_UNIT_WR
* is 10us granularity, so set the clock to 6000 (0x1770)
*/
#define FBNIC_RXB_PS_CLK_DIV 0x1770
/* Convert milliseconds to pause storm timeout units (10us granularity) */
#define FBNIC_MAC_RXB_PS_TO(ms) ((ms) * 100)
/* Convert pause storm timeout units (10us granularity) to milliseconds */
#define FBNIC_MAC_RXB_PS_TO_MS(ps) ((ps) / 100)
/* Set the default timer to 500ms, which should be longer than any
* reasonable period of continuous pausing. The service task, which runs
* once per second, periodically resets the pause storm trigger.
*
* As a result, on a functioning system, if pause continues, we enforce
* a duty cycle determined by the configured pause storm timeout (50%
* default). A crashed system will not have the service task and therefore
* pause will remain disabled until reboot recovery.
*/
#define FBNIC_MAC_PS_TO_DEFAULT_MS 500
#define FBNIC_MAC_PS_TO_MAX_MS \
FBNIC_MAC_RXB_PS_TO_MS(FIELD_MAX(FBNIC_RXB_PAUSE_STORM_THLD_TIME))
#define FBNIC_MAX_JUMBO_FRAME_SIZE 9742
/* States loosely based on section 136.8.11.7.5 of IEEE 802.3-2022 Ethernet
@ -119,4 +143,7 @@ struct fbnic_mac {
int fbnic_mac_init(struct fbnic_dev *fbd);
void fbnic_mac_get_fw_settings(struct fbnic_dev *fbd, u8 *aui, u8 *fec);
int fbnic_mac_ps_protect_to_config(struct fbnic_dev *fbd, u16 timeout);
void fbnic_mac_ps_protect_handler(struct fbnic_dev *fbd);
bool fbnic_mac_check_tx_pause(struct fbnic_dev *fbd);
#endif /* _FBNIC_MAC_H_ */

View File

@ -220,6 +220,9 @@ static void fbnic_service_task(struct work_struct *work)
fbnic_get_hw_stats32(fbd);
if (fbd->ps_timeout && fbnic_mac_check_tx_pause(fbd))
fbnic_mac_ps_protect_handler(fbd);
fbnic_fw_check_heartbeat(fbd);
fbnic_health_check(fbd);
@ -296,6 +299,8 @@ static int fbnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
/* Populate driver with hardware-specific info and handlers */
fbd->max_num_queues = info->max_num_queues;
fbd->ps_timeout = FBNIC_MAC_PS_TO_DEFAULT_MS;
pci_set_master(pdev);
pci_save_state(pdev);