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 <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 16 insertions, 3 deletions

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));