d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

mlx5en: Check for pci_channel_offline() when draining sendqueue.

Browse Source 
This speeds up detach in hypervisor environments.

MFC after:	1 week
Reviewed by:	kib
Sponsored by:	Mellanox Technologies // NVIDIA Networking
This commit is contained in:
Hans Petter Selasky 2021-06-16 15:01:56 +02:00
parent 8abf5ac0e6
commit 4692d9808e
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
sys/dev/mlx5/mlx5_en/mlx5_en_main.c
View File

@ -1875,7 +1875,8 @@ mlx5e_drain_sq(struct mlx5e_sq *sq)
mtx_lock(&sq->lock);
while (sq->cc != sq->pc &&
(sq->priv->media_status_last & IFM_ACTIVE) != 0 &&
mdev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR) {
mdev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR &&
pci_channel_offline(mdev->pdev) == 0) {
mtx_unlock(&sq->lock);
msleep(1);
sq->cq.mcq.comp(&sq->cq.mcq, NULL);
@ -1893,7 +1894,8 @@ mlx5e_drain_sq(struct mlx5e_sq *sq)
/* wait till SQ is empty */
mtx_lock(&sq->lock);
while (sq->cc != sq->pc &&
mdev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR) {
mdev->state != MLX5_DEVICE_STATE_INTERNAL_ERROR &&
pci_channel_offline(mdev->pdev) == 0) {
mtx_unlock(&sq->lock);
msleep(1);
sq->cq.mcq.comp(&sq->cq.mcq, NULL);