From: Saeed Mahameed <saeed@kernel.org>
To: "David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Eric Dumazet <edumazet@google.com>
Cc: Saeed Mahameed <saeedm@nvidia.com>,
netdev@vger.kernel.org, Maxim Mikityanskiy <maximmi@nvidia.com>,
Tariq Toukan <tariqt@nvidia.com>
Subject: [net 9/9] net/mlx5e: Ring the TX doorbell on DMA errors
Date: Wed, 6 Jul 2022 16:13:09 -0700 [thread overview]
Message-ID: <20220706231309.38579-10-saeed@kernel.org> (raw)
In-Reply-To: <20220706231309.38579-1-saeed@kernel.org>
From: Maxim Mikityanskiy <maximmi@nvidia.com>
TX doorbells may be postponed, because sometimes the driver knows that
another packet follows (for example, when xmit_more is true, or when a
MPWQE session is closed before transmitting a packet).
However, the DMA mapping may fail for the next packet, in which case a
new WQE is not posted, the doorbell isn't updated either, and the
transmission of the previous packet will be delayed indefinitely.
This commit fixes the described rare error flow by posting a NOP and
ringing the doorbell on errors to flush all the previous packets. The
MPWQE session is closed before that. DMA mapping in the MPWQE flow is
moved to the beginning of mlx5e_sq_xmit_mpwqe, because empty sessions
are not allowed. Stop room always has enough space for a NOP, because
the actual TX WQE is not posted.
Fixes: e586b3b0baee ("net/mlx5: Ethernet Datapath files")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en_tx.c | 39 ++++++++++++++-----
1 file changed, 30 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index 50d14cec4894..9a7250be229f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -341,6 +341,26 @@ static void mlx5e_tx_check_stop(struct mlx5e_txqsq *sq)
}
}
+static void mlx5e_tx_flush(struct mlx5e_txqsq *sq)
+{
+ struct mlx5e_tx_wqe_info *wi;
+ struct mlx5e_tx_wqe *wqe;
+ u16 pi;
+
+ /* Must not be called when a MPWQE session is active but empty. */
+ mlx5e_tx_mpwqe_ensure_complete(sq);
+
+ pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
+ wi = &sq->db.wqe_info[pi];
+
+ *wi = (struct mlx5e_tx_wqe_info) {
+ .num_wqebbs = 1,
+ };
+
+ wqe = mlx5e_post_nop(&sq->wq, sq->sqn, &sq->pc);
+ mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, &wqe->ctrl);
+}
+
static inline void
mlx5e_txwqe_complete(struct mlx5e_txqsq *sq, struct sk_buff *skb,
const struct mlx5e_tx_attr *attr,
@@ -459,6 +479,7 @@ mlx5e_sq_xmit_wqe(struct mlx5e_txqsq *sq, struct sk_buff *skb,
err_drop:
stats->dropped++;
dev_kfree_skb_any(skb);
+ mlx5e_tx_flush(sq);
}
static bool mlx5e_tx_skb_supports_mpwqe(struct sk_buff *skb, struct mlx5e_tx_attr *attr)
@@ -560,6 +581,13 @@ mlx5e_sq_xmit_mpwqe(struct mlx5e_txqsq *sq, struct sk_buff *skb,
struct mlx5_wqe_ctrl_seg *cseg;
struct mlx5e_xmit_data txd;
+ txd.data = skb->data;
+ txd.len = skb->len;
+
+ txd.dma_addr = dma_map_single(sq->pdev, txd.data, txd.len, DMA_TO_DEVICE);
+ if (unlikely(dma_mapping_error(sq->pdev, txd.dma_addr)))
+ goto err_unmap;
+
if (!mlx5e_tx_mpwqe_session_is_active(sq)) {
mlx5e_tx_mpwqe_session_start(sq, eseg);
} else if (!mlx5e_tx_mpwqe_same_eseg(sq, eseg)) {
@@ -569,18 +597,9 @@ mlx5e_sq_xmit_mpwqe(struct mlx5e_txqsq *sq, struct sk_buff *skb,
sq->stats->xmit_more += xmit_more;
- txd.data = skb->data;
- txd.len = skb->len;
-
- txd.dma_addr = dma_map_single(sq->pdev, txd.data, txd.len, DMA_TO_DEVICE);
- if (unlikely(dma_mapping_error(sq->pdev, txd.dma_addr)))
- goto err_unmap;
mlx5e_dma_push(sq, txd.dma_addr, txd.len, MLX5E_DMA_MAP_SINGLE);
-
mlx5e_skb_fifo_push(&sq->db.skb_fifo, skb);
-
mlx5e_tx_mpwqe_add_dseg(sq, &txd);
-
mlx5e_tx_skb_update_hwts_flags(skb);
if (unlikely(mlx5e_tx_mpwqe_is_full(&sq->mpwqe, sq->max_sq_mpw_wqebbs))) {
@@ -602,6 +621,7 @@ mlx5e_sq_xmit_mpwqe(struct mlx5e_txqsq *sq, struct sk_buff *skb,
mlx5e_dma_unmap_wqe_err(sq, 1);
sq->stats->dropped++;
dev_kfree_skb_any(skb);
+ mlx5e_tx_flush(sq);
}
void mlx5e_tx_mpwqe_ensure_complete(struct mlx5e_txqsq *sq)
@@ -1006,5 +1026,6 @@ void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
err_drop:
stats->dropped++;
dev_kfree_skb_any(skb);
+ mlx5e_tx_flush(sq);
}
#endif
--
2.36.1
prev parent reply other threads:[~2022-07-06 23:13 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-06 23:13 [pull request][net 0/9] mlx5 fixes 2022-07-06 Saeed Mahameed
2022-07-06 23:13 ` [net 1/9] net/mlx5: TC, allow offload from uplink to other PF's VF Saeed Mahameed
2022-07-08 1:00 ` patchwork-bot+netdevbpf
2022-07-06 23:13 ` [net 2/9] net/mlx5: Lag, decouple FDB selection and shared FDB Saeed Mahameed
2022-07-06 23:13 ` [net 3/9] net/mlx5e: kTLS, Fix build time constant test in TX Saeed Mahameed
2022-07-06 23:13 ` [net 4/9] net/mlx5e: kTLS, Fix build time constant test in RX Saeed Mahameed
2022-07-06 23:13 ` [net 5/9] net/mlx5e: Fix enabling sriov while tc nic rules are offloaded Saeed Mahameed
2022-07-06 23:13 ` [net 6/9] net/mlx5: Lag, correct get the port select mode str Saeed Mahameed
2022-07-06 23:13 ` [net 7/9] net/mlx5e: CT: Use own workqueue instead of mlx5e priv Saeed Mahameed
2022-07-06 23:13 ` [net 8/9] net/mlx5e: Fix capability check for updating vnic env counters Saeed Mahameed
2022-07-06 23:13 ` Saeed Mahameed [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220706231309.38579-10-saeed@kernel.org \
--to=saeed@kernel.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=kuba@kernel.org \
--cc=maximmi@nvidia.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=saeedm@nvidia.com \
--cc=tariqt@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).