[PATCH net v2] net: stmmac: fix handling of zero coalescing tx-usecs

[PATCH net v2] net: stmmac: fix handling of zero coalescing tx-usecs

[Vincent Whitchurch]

Setting ethtool -C eth0 tx-usecs 0 is supposed to disable the use of the
coalescing timer but currently it gets programmed with zero delay
instead.

Disable the use of the coalescing timer if tx-usecs is zero by
preventing it from being restarted.  Note that to keep things simple we
don't start/stop the timer when the coalescing settings are changed, but
just let that happen on the next transmit or timer expiry.

Fixes: 8fce33317023 ("net: stmmac: Rework coalesce timer and fix multi-queue races")
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---
Changes in v2:
- Rebased on current net/main.
- Link to v1: https://lore.kernel.org/r/20230905-stmmac-coaloff-v1-1-e29820e8ff6d@axis.com
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index 9a3182b9e767..2206789802bf 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -2704,9 +2704,7 @@ static int stmmac_tx_clean(struct stmmac_priv *priv, int budget, u32 queue)
 
 	/* We still have pending packets, let's call for a new scheduling */
 	if (tx_q->dirty_tx != tx_q->cur_tx)
-		hrtimer_start(&tx_q->txtimer,
-			      STMMAC_COAL_TIMER(priv->tx_coal_timer[queue]),
-			      HRTIMER_MODE_REL);
+		stmmac_tx_timer_arm(priv, queue);
 
 	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
 	tx_q->txq_stats.tx_packets += tx_packets;
@@ -2995,9 +2993,13 @@ static int stmmac_init_dma_engine(struct stmmac_priv *priv)
 static void stmmac_tx_timer_arm(struct stmmac_priv *priv, u32 queue)
 {
 	struct stmmac_tx_queue *tx_q = &priv->dma_conf.tx_queue[queue];
+	u32 tx_coal_timer = priv->tx_coal_timer[queue];
+
+	if (!tx_coal_timer)
+		return;
 
 	hrtimer_start(&tx_q->txtimer,
-		      STMMAC_COAL_TIMER(priv->tx_coal_timer[queue]),
+		      STMMAC_COAL_TIMER(tx_coal_timer),
 		      HRTIMER_MODE_REL);
 }
 

---
base-commit: 35494b0d61e44b517178aa1c6f5a69168b086940
change-id: 20230904-stmmac-coaloff-aca6befbd24b

Best regards,
-- 
Vincent Whitchurch <vincent.whitchurch@axis.com>

Re: [PATCH net v2] net: stmmac: fix handling of zero coalescing tx-usecs

[Jakub Kicinski]

On Thu, 7 Sep 2023 12:46:31 +0200 Vincent Whitchurch wrote:
> Setting ethtool -C eth0 tx-usecs 0 is supposed to disable the use of the
> coalescing timer but currently it gets programmed with zero delay
> instead.
> 
> Disable the use of the coalescing timer if tx-usecs is zero by
> preventing it from being restarted.  Note that to keep things simple we
> don't start/stop the timer when the coalescing settings are changed, but
> just let that happen on the next transmit or timer expiry.
> 
> Fixes: 8fce33317023 ("net: stmmac: Rework coalesce timer and fix multi-queue races")
> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>

Felix, good enough?

Re: [PATCH net v2] net: stmmac: fix handling of zero coalescing tx-usecs

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by David S. Miller <davem@davemloft.net>:

On Thu, 7 Sep 2023 12:46:31 +0200 you wrote:
> Setting ethtool -C eth0 tx-usecs 0 is supposed to disable the use of the
> coalescing timer but currently it gets programmed with zero delay
> instead.
> 
> Disable the use of the coalescing timer if tx-usecs is zero by
> preventing it from being restarted.  Note that to keep things simple we
> don't start/stop the timer when the coalescing settings are changed, but
> just let that happen on the next transmit or timer expiry.
> 
> [...]

Here is the summary with links:
  - [net,v2] net: stmmac: fix handling of zero coalescing tx-usecs
    https://git.kernel.org/netdev/net/c/fa60b8163816

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html