From: Edward Cree Date: Wed, 16 Apr 2014 18:27:48 +0000 (+0100) Subject: sfc:On MCDI timeout, issue an FLR (and mark MCDI to fail-fast) X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=e283546c0465dd3026bc94f7b1a9de7f6b8969ec;p=GitHub%2FLineageOS%2Fandroid_kernel_motorola_exynos9610.git sfc:On MCDI timeout, issue an FLR (and mark MCDI to fail-fast) When an MCDI command times out (whether or not we find it completed when we poll), call efx_mcdi_abandon(), which tells all subsequent MCDI calls to fail-fast, and queues up an FLR. Because an FLR doesn't lead to receiving any reboot even from the MC (unlike most other types of reset), we have to call efx_ef10_reset_mc_allocations. In efx_start_all(), if a reset (of any kind) is pending, we bail out. Without this, attempts to reconfigure (e.g. change mtu) can cause driver/mc state inconsistency if the first MCDI call triggers an FLR. For similar reasons, on EF10, in efx_reset_down(method=RESET_TYPE_MCDI_TIMEOUT), set the number of active queues to zero before calling efx_stop_all(). And, on farch, in efx_reset_up(method=RESET_TYPE_MCDI_TIMEOUT), set active_queues and flushes pending & outstanding to zero. efx_mcdi_mode_{poll,event}() should not take us out of fail-fast mode. Instead, this is done by efx_mcdi_reset() after the FLR completes. The new FLR reset_type RESET_TYPE_MCDI_TIMEOUT doesn't really fit into the hierarchy of reset 'scopes' whereby efx_reset() decides some resets subsume others. Thus, it uses separate logic. Also, fixed up some inconsistency around RESET_TYPE_MC_BIST, which was in the wrong place in that hierarchy. Signed-off-by: Shradha Shah Signed-off-by: David S. Miller --- diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c index 21c20ea0dad0..b5ed30a39144 100644 --- a/drivers/net/ethernet/sfc/ef10.c +++ b/drivers/net/ethernet/sfc/ef10.c @@ -738,8 +738,11 @@ static int efx_ef10_reset(struct efx_nic *efx, enum reset_type reset_type) /* If it was a port reset, trigger reallocation of MC resources. * Note that on an MC reset nothing needs to be done now because we'll * detect the MC reset later and handle it then. + * For an FLR, we never get an MC reset event, but the MC has reset all + * resources assigned to us, so we have to trigger reallocation now. */ - if (reset_type == RESET_TYPE_ALL && !rc) + if ((reset_type == RESET_TYPE_ALL || + reset_type == RESET_TYPE_MCDI_TIMEOUT) && !rc) efx_ef10_reset_mc_allocations(efx); return rc; } @@ -2141,6 +2144,11 @@ static int efx_ef10_fini_dmaq(struct efx_nic *efx) return 0; } +static void efx_ef10_prepare_flr(struct efx_nic *efx) +{ + atomic_set(&efx->active_queues, 0); +} + static bool efx_ef10_filter_equal(const struct efx_filter_spec *left, const struct efx_filter_spec *right) { @@ -3603,6 +3611,8 @@ const struct efx_nic_type efx_hunt_a0_nic_type = { .probe_port = efx_mcdi_port_probe, .remove_port = efx_mcdi_port_remove, .fini_dmaq = efx_ef10_fini_dmaq, + .prepare_flr = efx_ef10_prepare_flr, + .finish_flr = efx_port_dummy_op_void, .describe_stats = efx_ef10_describe_stats, .update_stats = efx_ef10_update_stats, .start_stats = efx_mcdi_mac_start_stats, diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c index 57b971e5e6b2..63d595fd3cc5 100644 --- a/drivers/net/ethernet/sfc/efx.c +++ b/drivers/net/ethernet/sfc/efx.c @@ -76,6 +76,7 @@ const char *const efx_reset_type_names[] = { [RESET_TYPE_RECOVER_OR_ALL] = "RECOVER_OR_ALL", [RESET_TYPE_WORLD] = "WORLD", [RESET_TYPE_RECOVER_OR_DISABLE] = "RECOVER_OR_DISABLE", + [RESET_TYPE_MC_BIST] = "MC_BIST", [RESET_TYPE_DISABLE] = "DISABLE", [RESET_TYPE_TX_WATCHDOG] = "TX_WATCHDOG", [RESET_TYPE_INT_ERROR] = "INT_ERROR", @@ -83,7 +84,7 @@ const char *const efx_reset_type_names[] = { [RESET_TYPE_DMA_ERROR] = "DMA_ERROR", [RESET_TYPE_TX_SKIP] = "TX_SKIP", [RESET_TYPE_MC_FAILURE] = "MC_FAILURE", - [RESET_TYPE_MC_BIST] = "MC_BIST", + [RESET_TYPE_MCDI_TIMEOUT] = "MCDI_TIMEOUT (FLR)", }; /* Reset workqueue. If any NIC has a hardware failure then a reset will be @@ -1739,7 +1740,8 @@ static void efx_start_all(struct efx_nic *efx) /* Check that it is appropriate to restart the interface. All * of these flags are safe to read under just the rtnl lock */ - if (efx->port_enabled || !netif_running(efx->net_dev)) + if (efx->port_enabled || !netif_running(efx->net_dev) || + efx->reset_pending) return; efx_start_port(efx); @@ -2334,6 +2336,9 @@ void efx_reset_down(struct efx_nic *efx, enum reset_type method) { EFX_ASSERT_RESET_SERIALISED(efx); + if (method == RESET_TYPE_MCDI_TIMEOUT) + efx->type->prepare_flr(efx); + efx_stop_all(efx); efx_disable_interrupts(efx); @@ -2354,6 +2359,10 @@ int efx_reset_up(struct efx_nic *efx, enum reset_type method, bool ok) EFX_ASSERT_RESET_SERIALISED(efx); + if (method == RESET_TYPE_MCDI_TIMEOUT) + efx->type->finish_flr(efx); + + /* Ensure that SRAM is initialised even if we're disabling the device */ rc = efx->type->init(efx); if (rc) { netif_err(efx, drv, efx->net_dev, "failed to initialise NIC\n"); @@ -2417,7 +2426,10 @@ int efx_reset(struct efx_nic *efx, enum reset_type method) /* Clear flags for the scopes we covered. We assume the NIC and * driver are now quiescent so that there is no race here. */ - efx->reset_pending &= -(1 << (method + 1)); + if (method < RESET_TYPE_MAX_METHOD) + efx->reset_pending &= -(1 << (method + 1)); + else /* it doesn't fit into the well-ordered scope hierarchy */ + __clear_bit(method, &efx->reset_pending); /* Reinitialise bus-mastering, which may have been turned off before * the reset was scheduled. This is still appropriate, even in the @@ -2546,6 +2558,7 @@ void efx_schedule_reset(struct efx_nic *efx, enum reset_type type) case RESET_TYPE_DISABLE: case RESET_TYPE_RECOVER_OR_DISABLE: case RESET_TYPE_MC_BIST: + case RESET_TYPE_MCDI_TIMEOUT: method = type; netif_dbg(efx, drv, efx->net_dev, "scheduling %s reset\n", RESET_TYPE(method)); diff --git a/drivers/net/ethernet/sfc/enum.h b/drivers/net/ethernet/sfc/enum.h index 75ef7ef6450b..d1dbb5fb31bb 100644 --- a/drivers/net/ethernet/sfc/enum.h +++ b/drivers/net/ethernet/sfc/enum.h @@ -143,6 +143,7 @@ enum efx_loopback_mode { * @RESET_TYPE_WORLD: Reset as much as possible * @RESET_TYPE_RECOVER_OR_DISABLE: Try to recover. Apply RESET_TYPE_DISABLE if * unsuccessful. + * @RESET_TYPE_MC_BIST: MC entering BIST mode. * @RESET_TYPE_DISABLE: Reset datapath, MAC and PHY; leave NIC disabled * @RESET_TYPE_TX_WATCHDOG: reset due to TX watchdog * @RESET_TYPE_INT_ERROR: reset due to internal error @@ -150,14 +151,16 @@ enum efx_loopback_mode { * @RESET_TYPE_DMA_ERROR: DMA error * @RESET_TYPE_TX_SKIP: hardware completed empty tx descriptors * @RESET_TYPE_MC_FAILURE: MC reboot/assertion + * @RESET_TYPE_MCDI_TIMEOUT: MCDI timeout. */ enum reset_type { - RESET_TYPE_INVISIBLE = 0, - RESET_TYPE_RECOVER_OR_ALL = 1, - RESET_TYPE_ALL = 2, - RESET_TYPE_WORLD = 3, - RESET_TYPE_RECOVER_OR_DISABLE = 4, - RESET_TYPE_DISABLE = 5, + RESET_TYPE_INVISIBLE, + RESET_TYPE_RECOVER_OR_ALL, + RESET_TYPE_ALL, + RESET_TYPE_WORLD, + RESET_TYPE_RECOVER_OR_DISABLE, + RESET_TYPE_MC_BIST, + RESET_TYPE_DISABLE, RESET_TYPE_MAX_METHOD, RESET_TYPE_TX_WATCHDOG, RESET_TYPE_INT_ERROR, @@ -165,7 +168,13 @@ enum reset_type { RESET_TYPE_DMA_ERROR, RESET_TYPE_TX_SKIP, RESET_TYPE_MC_FAILURE, - RESET_TYPE_MC_BIST, + /* RESET_TYPE_MCDI_TIMEOUT is actually a method, not just a reason, but + * it doesn't fit the scope hierarchy (not well-ordered by inclusion). + * We encode this by having its enum value be greater than + * RESET_TYPE_MAX_METHOD. This also prevents issuing it with + * efx_ioctl_reset. + */ + RESET_TYPE_MCDI_TIMEOUT, RESET_TYPE_MAX, }; diff --git a/drivers/net/ethernet/sfc/falcon.c b/drivers/net/ethernet/sfc/falcon.c index 8ec20b713cc6..fae25a418647 100644 --- a/drivers/net/ethernet/sfc/falcon.c +++ b/drivers/net/ethernet/sfc/falcon.c @@ -2696,6 +2696,8 @@ const struct efx_nic_type falcon_a1_nic_type = { .fini_dmaq = efx_farch_fini_dmaq, .prepare_flush = falcon_prepare_flush, .finish_flush = efx_port_dummy_op_void, + .prepare_flr = efx_port_dummy_op_void, + .finish_flr = efx_farch_finish_flr, .describe_stats = falcon_describe_nic_stats, .update_stats = falcon_update_nic_stats, .start_stats = falcon_start_nic_stats, @@ -2790,6 +2792,8 @@ const struct efx_nic_type falcon_b0_nic_type = { .fini_dmaq = efx_farch_fini_dmaq, .prepare_flush = falcon_prepare_flush, .finish_flush = efx_port_dummy_op_void, + .prepare_flr = efx_port_dummy_op_void, + .finish_flr = efx_farch_finish_flr, .describe_stats = falcon_describe_nic_stats, .update_stats = falcon_update_nic_stats, .start_stats = falcon_start_nic_stats, diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c index a08761360cdf..0537381cd2f6 100644 --- a/drivers/net/ethernet/sfc/farch.c +++ b/drivers/net/ethernet/sfc/farch.c @@ -741,6 +741,28 @@ int efx_farch_fini_dmaq(struct efx_nic *efx) return rc; } +/* Reset queue and flush accounting after FLR + * + * One possible cause of FLR recovery is that DMA may be failing (eg. if bus + * mastering was disabled), in which case we don't receive (RXQ) flush + * completion events. This means that efx->rxq_flush_outstanding remained at 4 + * after the FLR; also, efx->active_queues was non-zero (as no flush completion + * events were received, and we didn't go through efx_check_tx_flush_complete()) + * If we don't fix this up, on the next call to efx_realloc_channels() we won't + * flush any RX queues because efx->rxq_flush_outstanding is at the limit of 4 + * for batched flush requests; and the efx->active_queues gets messed up because + * we keep incrementing for the newly initialised queues, but it never went to + * zero previously. Then we get a timeout every time we try to restart the + * queues, as it doesn't go back to zero when we should be flushing the queues. + */ +void efx_farch_finish_flr(struct efx_nic *efx) +{ + atomic_set(&efx->rxq_flush_pending, 0); + atomic_set(&efx->rxq_flush_outstanding, 0); + atomic_set(&efx->active_queues, 0); +} + + /************************************************************************** * * Event queue processing diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c index 7bd4b14bf3b3..5239cf9bdc56 100644 --- a/drivers/net/ethernet/sfc/mcdi.c +++ b/drivers/net/ethernet/sfc/mcdi.c @@ -52,12 +52,7 @@ static void efx_mcdi_timeout_async(unsigned long context); static int efx_mcdi_drv_attach(struct efx_nic *efx, bool driver_operating, bool *was_attached_out); static bool efx_mcdi_poll_once(struct efx_nic *efx); - -static inline struct efx_mcdi_iface *efx_mcdi(struct efx_nic *efx) -{ - EFX_BUG_ON_PARANOID(!efx->mcdi); - return &efx->mcdi->iface; -} +static void efx_mcdi_abandon(struct efx_nic *efx); int efx_mcdi_init(struct efx_nic *efx) { @@ -558,6 +553,8 @@ static int _efx_mcdi_rpc_finish(struct efx_nic *efx, unsigned cmd, size_t inlen, rc = 0; } + efx_mcdi_abandon(efx); + /* Close the race with efx_mcdi_ev_cpl() executing just too late * and completing a request we've just cancelled, by ensuring * that the seqno check therein fails. @@ -672,6 +669,9 @@ int efx_mcdi_rpc_start(struct efx_nic *efx, unsigned cmd, if (efx->mc_bist_for_other_fn) return -ENETDOWN; + if (mcdi->mode == MCDI_MODE_FAIL) + return -ENETDOWN; + efx_mcdi_acquire_sync(mcdi); efx_mcdi_send_request(efx, cmd, inbuf, inlen); return 0; @@ -812,7 +812,11 @@ void efx_mcdi_mode_poll(struct efx_nic *efx) return; mcdi = efx_mcdi(efx); - if (mcdi->mode == MCDI_MODE_POLL) + /* If already in polling mode, nothing to do. + * If in fail-fast state, don't switch to polled completion. + * FLR recovery will do that later. + */ + if (mcdi->mode == MCDI_MODE_POLL || mcdi->mode == MCDI_MODE_FAIL) return; /* We can switch from event completion to polled completion, because @@ -841,8 +845,8 @@ void efx_mcdi_flush_async(struct efx_nic *efx) mcdi = efx_mcdi(efx); - /* We must be in polling mode so no more requests can be queued */ - BUG_ON(mcdi->mode != MCDI_MODE_POLL); + /* We must be in poll or fail mode so no more requests can be queued */ + BUG_ON(mcdi->mode == MCDI_MODE_EVENTS); del_timer_sync(&mcdi->async_timer); @@ -875,8 +879,11 @@ void efx_mcdi_mode_event(struct efx_nic *efx) return; mcdi = efx_mcdi(efx); - - if (mcdi->mode == MCDI_MODE_EVENTS) + /* If already in event completion mode, nothing to do. + * If in fail-fast state, don't switch to event completion. FLR + * recovery will do that later. + */ + if (mcdi->mode == MCDI_MODE_EVENTS || mcdi->mode == MCDI_MODE_FAIL) return; /* We can't switch from polled to event completion in the middle of a @@ -966,6 +973,19 @@ static void efx_mcdi_ev_bist(struct efx_nic *efx) spin_unlock(&mcdi->iface_lock); } +/* MCDI timeouts seen, so make all MCDI calls fail-fast and issue an FLR to try + * to recover. + */ +static void efx_mcdi_abandon(struct efx_nic *efx) +{ + struct efx_mcdi_iface *mcdi = efx_mcdi(efx); + + if (xchg(&mcdi->mode, MCDI_MODE_FAIL) == MCDI_MODE_FAIL) + return; /* it had already been done */ + netif_dbg(efx, hw, efx->net_dev, "MCDI is timing out; trying to recover\n"); + efx_schedule_reset(efx, RESET_TYPE_MCDI_TIMEOUT); +} + /* Called from falcon_process_eventq for MCDI events */ void efx_mcdi_process_event(struct efx_channel *channel, efx_qword_t *event) @@ -1512,6 +1532,19 @@ int efx_mcdi_reset(struct efx_nic *efx, enum reset_type method) { int rc; + /* If MCDI is down, we can't handle_assertion */ + if (method == RESET_TYPE_MCDI_TIMEOUT) { + rc = pci_reset_function(efx->pci_dev); + if (rc) + return rc; + /* Re-enable polled MCDI completion */ + if (efx->mcdi) { + struct efx_mcdi_iface *mcdi = efx_mcdi(efx); + mcdi->mode = MCDI_MODE_POLL; + } + return 0; + } + /* Recover from a failed assertion pre-reset */ rc = efx_mcdi_handle_assertion(efx); if (rc) diff --git a/drivers/net/ethernet/sfc/mcdi.h b/drivers/net/ethernet/sfc/mcdi.h index 52931aebf3c3..56465f7465a2 100644 --- a/drivers/net/ethernet/sfc/mcdi.h +++ b/drivers/net/ethernet/sfc/mcdi.h @@ -28,9 +28,16 @@ enum efx_mcdi_state { MCDI_STATE_COMPLETED, }; +/** + * enum efx_mcdi_mode - MCDI transaction mode + * @MCDI_MODE_POLL: poll for MCDI completion, until timeout + * @MCDI_MODE_EVENTS: wait for an mcdi_event. On timeout, poll once + * @MCDI_MODE_FAIL: we think MCDI is dead, so fail-fast all calls + */ enum efx_mcdi_mode { MCDI_MODE_POLL, MCDI_MODE_EVENTS, + MCDI_MODE_FAIL, }; /** @@ -104,6 +111,12 @@ struct efx_mcdi_data { u32 fn_flags; }; +static inline struct efx_mcdi_iface *efx_mcdi(struct efx_nic *efx) +{ + EFX_BUG_ON_PARANOID(!efx->mcdi); + return &efx->mcdi->iface; +} + #ifdef CONFIG_SFC_MCDI_MON static inline struct efx_mcdi_mon *efx_mcdi_mon(struct efx_nic *efx) { diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h index 8a400a0595eb..5bdae8ed7c57 100644 --- a/drivers/net/ethernet/sfc/net_driver.h +++ b/drivers/net/ethernet/sfc/net_driver.h @@ -972,6 +972,8 @@ struct efx_mtd_partition { * (for Falcon architecture) * @finish_flush: Clean up after flushing the DMA queues (for Falcon * architecture) + * @prepare_flr: Prepare for an FLR + * @finish_flr: Clean up after an FLR * @describe_stats: Describe statistics for ethtool * @update_stats: Update statistics not provided by event handling. * Either argument may be %NULL. @@ -1100,6 +1102,8 @@ struct efx_nic_type { int (*fini_dmaq)(struct efx_nic *efx); void (*prepare_flush)(struct efx_nic *efx); void (*finish_flush)(struct efx_nic *efx); + void (*prepare_flr)(struct efx_nic *efx); + void (*finish_flr)(struct efx_nic *efx); size_t (*describe_stats)(struct efx_nic *efx, u8 *names); size_t (*update_stats)(struct efx_nic *efx, u64 *full_stats, struct rtnl_link_stats64 *core_stats); diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h index a001fae1a8d7..d3ad8ed8d901 100644 --- a/drivers/net/ethernet/sfc/nic.h +++ b/drivers/net/ethernet/sfc/nic.h @@ -757,6 +757,7 @@ static inline int efx_nic_irq_test_irq_cpu(struct efx_nic *efx) int efx_nic_flush_queues(struct efx_nic *efx); void siena_prepare_flush(struct efx_nic *efx); int efx_farch_fini_dmaq(struct efx_nic *efx); +void efx_farch_finish_flr(struct efx_nic *efx); void siena_finish_flush(struct efx_nic *efx); void falcon_start_nic_stats(struct efx_nic *efx); void falcon_stop_nic_stats(struct efx_nic *efx); diff --git a/drivers/net/ethernet/sfc/siena.c b/drivers/net/ethernet/sfc/siena.c index 23f3a6f7737a..50ffefed492c 100644 --- a/drivers/net/ethernet/sfc/siena.c +++ b/drivers/net/ethernet/sfc/siena.c @@ -921,6 +921,8 @@ const struct efx_nic_type siena_a0_nic_type = { .fini_dmaq = efx_farch_fini_dmaq, .prepare_flush = siena_prepare_flush, .finish_flush = siena_finish_flush, + .prepare_flr = efx_port_dummy_op_void, + .finish_flr = efx_farch_finish_flr, .describe_stats = siena_describe_nic_stats, .update_stats = siena_update_nic_stats, .start_stats = efx_mcdi_mac_start_stats,