From: Ziwei Xiao ziweixiao@google.com
[ Upstream commit f4c2e67c1773d2a2632381ee30e9139c1e744c16 ]
Clearing the PBA bit from the driver is race prone and it may lead to dropped interrupt events. This could potentially lead to the traffic being completely halted.
Fixes: 5e8c5adf95f8 ("gve: DQO: Add core netdev features") Signed-off-by: Ziwei Xiao ziweixiao@google.com Signed-off-by: Bailey Forrest bcf@google.com Reviewed-by: Simon Horman simon.horman@corigine.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/ethernet/google/gve/gve_main.c | 13 ------------- 1 file changed, 13 deletions(-)
diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index d3e3ac242bfc3..2e5e0a8872704 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -239,19 +239,6 @@ static int gve_napi_poll_dqo(struct napi_struct *napi, int budget) bool reschedule = false; int work_done = 0;
- /* Clear PCI MSI-X Pending Bit Array (PBA) - * - * This bit is set if an interrupt event occurs while the vector is - * masked. If this bit is set and we reenable the interrupt, it will - * fire again. Since we're just about to poll the queue state, we don't - * need it to fire again. - * - * Under high softirq load, it's possible that the interrupt condition - * is triggered twice before we got the chance to process it. - */ - gve_write_irq_doorbell_dqo(priv, block, - GVE_ITR_NO_UPDATE_DQO | GVE_ITR_CLEAR_PBA_BIT_DQO); - if (block->tx) reschedule |= gve_tx_poll_dqo(block, /*do_clean=*/true);