sfc: Poll for MCDI completion once before timeout occurs
There is an as-yet unexplained bug that sometimes prevents (or delays)
the driver seeing the completion event for a completed MCDI request on
the SFC9120. The requested configuration change will have happened
but the driver assumes it to have failed, and this can result in
further failures. We can mitigate this by polling for completion
after unsuccessfully waiting for an event.
Fixes: 8127d661e77f ('sfc: Add support for Solarflare SFC9100 family')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c
index 9f26e46..4b0bd8a 100644
--- a/drivers/net/ethernet/sfc/mcdi.c
+++ b/drivers/net/ethernet/sfc/mcdi.c
@@ -630,6 +630,16 @@
rc = efx_mcdi_await_completion(efx);
if (rc != 0) {
+ netif_err(efx, hw, efx->net_dev,
+ "MC command 0x%x inlen %d mode %d timed out\n",
+ cmd, (int)inlen, mcdi->mode);
+
+ if (mcdi->mode == MCDI_MODE_EVENTS && efx_mcdi_poll_once(efx)) {
+ netif_err(efx, hw, efx->net_dev,
+ "MCDI request was completed without an event\n");
+ rc = 0;
+ }
+
/* 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.
@@ -638,11 +648,9 @@
++mcdi->seqno;
++mcdi->credits;
spin_unlock_bh(&mcdi->iface_lock);
+ }
- netif_err(efx, hw, efx->net_dev,
- "MC command 0x%x inlen %d mode %d timed out\n",
- cmd, (int)inlen, mcdi->mode);
- } else {
+ if (rc == 0) {
size_t hdr_len, data_len;
/* At the very least we need a memory barrier here to ensure