From: Joe Damato <joe@dama.to>
To: netdev@vger.kernel.org, Tariq Toukan <tariqt@nvidia.com>,
Andrew Lunn <andrew+netdev@lunn.ch>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@kernel.org>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: edumazet@google.com, horms@kernel.org,
linux-kernel@vger.kernel.org, daniel.machon@microchip.com,
Joe Damato <joe@dama.to>,
linux-rdma@vger.kernel.org
Subject: [PATCH net-next v2 4/5] net/mlx4: track whether a TX doorbell is pending
Date: Tue, 17 Nov 2026 07:53:49 -0800 [thread overview]
Message-ID: <20261117155353.2250540-5-joe@dama.to> (raw)
In-Reply-To: <20261117155353.2250540-1-joe@dama.to>
When the stack sets xmit_more, mlx4_en_xmit() skips the doorbell and
relies on a later call in the same batch to write it. Nothing records
that a doorbell is owed, so a path that ends the batch without writing
the doorbell leaves the WQEs posted before it unannounced to HW.
Add ring->doorbell_pending, set when the doorbell is deferred and
cleared whenever a doorbell or a BlueFlame write is issued, so such a
path can flush the deferred doorbell.
Place the flag in the cache line dirtied by mlx4_en_xmit(), next to
prod. That line has no room left, so move tx_dropped, which is only
written when an skb is dropped, to the end of the struct among the
fields not used in the fast path.
Signed-off-by: Joe Damato <joe@dama.to>
Reviewed-by: Daniel Machon <daniel.machon@microchip.com>
---
v2:
- doorbell_pending moved to the area of the struct marked as cache line
used and dirtied in mlx4_en_xmit. In exchange, tx_dropped is moved
among fields not written in the fast path. It is only written for
drops and is read for stats.
- Update the commit message.
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 9 +++++++--
drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 3 ++-
2 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 97f27e005810..123071a47bed 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -194,6 +194,7 @@ int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv,
ring->prod = 0;
ring->cons = 0xffffffff;
ring->last_nr_txbb = 1;
+ ring->doorbell_pending = false;
memset(ring->tx_info, 0, ring->size * sizeof(struct mlx4_en_tx_info));
memset(ring->buf, 0, ring->buf_size);
ring->free_tx_desc = mlx4_en_free_tx_desc;
@@ -757,6 +758,7 @@ void mlx4_en_xmit_doorbell(struct mlx4_en_tx_ring *ring)
iowrite32be(
#endif
(__force u32)ring->doorbell_qpn, ring->doorbell_address);
+ ring->doorbell_pending = false;
}
static void mlx4_en_tx_write_desc(struct mlx4_en_tx_ring *ring,
@@ -784,16 +786,19 @@ static void mlx4_en_tx_write_desc(struct mlx4_en_tx_ring *ring,
wmb();
ring->bf.offset ^= ring->bf.buf_size;
+ ring->doorbell_pending = false;
} else {
/* Ensure new descriptor hits memory
* before setting ownership of this descriptor to HW
*/
dma_wmb();
tx_desc->ctrl.owner_opcode = op_own;
- if (send_doorbell)
+ if (send_doorbell) {
mlx4_en_xmit_doorbell(ring);
- else
+ } else {
ring->xmit_more++;
+ ring->doorbell_pending = true;
+ }
}
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index aab97694f86b..c28d5213cccd 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -274,7 +274,7 @@ struct mlx4_en_tx_ring {
/* cache line used and dirtied in mlx4_en_xmit() */
u32 prod ____cacheline_aligned_in_smp;
- unsigned int tx_dropped;
+ bool doorbell_pending;
unsigned long bytes;
unsigned long packets;
unsigned long tx_csum;
@@ -311,6 +311,7 @@ struct mlx4_en_tx_ring {
enum mlx4_qp_state sp_qp_state;
u16 sp_stride;
u16 sp_cqn; /* index of port CQ associated with this ring */
+ unsigned int tx_dropped;
} ____cacheline_aligned_in_smp;
struct mlx4_en_rx_desc {
--
2.53.1-Meta
next prev parent reply other threads:[~2026-11-17 15:54 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-17 15:53 [PATCH net-next v2 0/5] net/mlx4: TX helpers and deferred doorbell fix Joe Damato
2026-11-17 15:53 ` [PATCH net-next v2 1/5] net/mlx4: factor VLAN parsing into mlx4_en_prep_vlan() Joe Damato
2026-11-17 15:53 ` [PATCH net-next v2 2/5] net/mlx4: factor csum flags into mlx4_en_csum_flags() Joe Damato
2026-11-17 15:53 ` [PATCH net-next v2 3/5] net/mlx4: factor HW loopback DMAC into mlx4_en_set_loopback_dmac() Joe Damato
2026-11-17 15:53 ` Joe Damato [this message]
2026-11-17 15:53 ` [PATCH net-next v2 5/5] net/mlx4: flush pending doorbell when dropping a TX skb Joe Damato
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=20261117155353.2250540-5-joe@dama.to \
--to=joe@dama.to \
--cc=andrew+netdev@lunn.ch \
--cc=daniel.machon@microchip.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=edumazet@kernel.org \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.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