From: Yevgeny Kliteynik Date: Sun, 17 Aug 2025 20:23:21 +0000 (+0300) Subject: net/mlx5: HWS, don't rehash on every kind of insertion failure X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=7c60952f83584bc4950057cfed2cc3c87343b5db;p=users%2Fjedix%2Flinux-maple.git net/mlx5: HWS, don't rehash on every kind of insertion failure If rule creation failed due to a full queue, due to timeout in polling for completion, or due to matcher being in resize, don't try to initiate rehash sequence - rehash would have failed anyway. Fixes: 2111bb970c78 ("net/mlx5: HWS, added backward-compatible API handling") Signed-off-by: Yevgeny Kliteynik Reviewed-by: Vlad Dogaru Signed-off-by: Mark Bloch Link: https://patch.msgid.link/20250817202323.308604-6-mbloch@nvidia.com Signed-off-by: Jakub Kicinski --- diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c index 2a59be11fe55..adeccc588e5d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c @@ -1063,6 +1063,21 @@ int mlx5hws_bwc_rule_create_simple(struct mlx5hws_bwc_rule *bwc_rule, return 0; /* rule inserted successfully */ } + /* Rule insertion could fail due to queue being full, timeout, or + * matcher in resize. In such cases, no point in trying to rehash. + */ + if (ret == -EBUSY || ret == -ETIMEDOUT || ret == -EAGAIN) { + mutex_unlock(queue_lock); + mlx5hws_err(ctx, + "BWC rule insertion failed - %s (%d)\n", + ret == -EBUSY ? "queue is full" : + ret == -ETIMEDOUT ? "timeout" : + ret == -EAGAIN ? "matcher in resize" : "N/A", + ret); + hws_bwc_rule_cnt_dec(bwc_rule); + return ret; + } + /* At this point the rule wasn't added. * It could be because there was collision, or some other problem. * Try rehash by size and insert rule again - last chance.