When driver has hit a parity event, HW can no longer write to host memory.
As a result, Tx completions cannot be written to the host SB memory, and
waiting for Tx completions eventually timeout.
As driver is willing to delay as much as 1-2 seconds per Tx queue for its
draining and this delay is sequential, the time to recover might greatly
lengthen needlessly in case the recovery is done under multi-connection
traffic.
Signed-off-by: Yuval Mintz <yuval.mi...@qlogic.com>
---
Hi Dave,
Wasn't sure whether to prepare this for net/net-next; But as it wasn't
a bug-fix by definition, I took the safer approach.
Please consider applying this to `net-next'.
Thanks,
Yuval
---
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index a949783..0a9108c 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -3042,8 +3042,12 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode,
bool keep_link)
bnx2x_save_statistics(bp);
}
- /* wait till consumers catch up with producers in all queues */
- bnx2x_drain_tx_queues(bp);
+ /* wait till consumers catch up with producers in all queues.
+ * If we're recovering, FW can't write to host so no reason
+ * to wait for the queues to complete all Tx.
+ */
+ if (unload_mode != UNLOAD_RECOVERY)
+ bnx2x_drain_tx_queues(bp);
/* if VF indicate to PF this function is going down (PF will delete sp
* elements and clear initializations
--
1.9.3
