Hi, I have some notes regarding the patch:
- patch does not fix the root cause of SQ state setting FW call failure (https://bugs.dpdk.org/show_bug.cgi?id=1449 ) - MLX5_ASSERT is engaged only in debug mode and should not be used in release - removing "return " prevents the tx_burst from retrying other errors recovery What we should do - check if issue still exists and fix the root cause instead of crippling the handling code. With best regards, Slava > -----Original Message----- > From: Jun Wang <junwan...@cestc.cn> > Sent: Wednesday, June 12, 2024 10:06 AM > To: dev@dpdk.org > Subject: [PATCH v1] net/mlx5: error cqe handle Tx queue need assert > > When encountering errors in handling some queues, we should assert here; > otherwise, some queues will fail to send packets continuously, leading to > traffic forwarding issues. > > Bugzilla ID: 1449 > > Signed-off-by: Jun Wang <junwan...@cestc.cn> > --- > drivers/net/mlx5/mlx5_tx.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/net/mlx5/mlx5_tx.c b/drivers/net/mlx5/mlx5_tx.c index > 1fe9521..e149f38 100644 > --- a/drivers/net/mlx5/mlx5_tx.c > +++ b/drivers/net/mlx5/mlx5_tx.c > @@ -213,7 +213,7 @@ > * handling, we do not advance the index > * here, allowing to retry on next call. > */ > - return; > + MLX5_ASSERT(ret < 0); > } > /* > * We are going to fetch all entries with > -- > 1.8.3.1 > >
