diff mbox series

[net-next,v4] mlx5/core: Schedule EQ comp tasklet only if necessary

Message ID 20241105204000.1807095-1-csander@purestorage.com (mailing list archive)
State Accepted
Delegated to: Netdev Maintainers
Headers show
Series [net-next,v4] mlx5/core: Schedule EQ comp tasklet only if necessary | expand

Checks

Context Check Description
netdev/series_format success Single patches do not need cover letters
netdev/tree_selection success Clearly marked for net-next
netdev/ynl success Generated files up to date; no warnings/errors; no diff in generated;
netdev/fixes_present success Fixes tag not required for -next series
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 3 this patch: 3
netdev/build_tools success No tools touched, skip
netdev/cc_maintainers success CCed 9 of 9 maintainers
netdev/build_clang success Errors and warnings before: 3 this patch: 3
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/deprecated_api success None detected
netdev/check_selftest success No net selftest shell script
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 4 this patch: 4
netdev/checkpatch success total: 0 errors, 0 warnings, 0 checks, 46 lines checked
netdev/build_clang_rust success No Rust files in patch. Skipping build
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0
netdev/contest success net-next-2024-11-07--12-00 (tests: 787)

Commit Message

Caleb Sander Nov. 5, 2024, 8:39 p.m. UTC 
Currently, the mlx5_eq_comp_int() interrupt handler schedules a tasklet
to call mlx5_cq_tasklet_cb() if it processes any completions. For CQs
whose completions don't need to be processed in tasklet context, this
adds unnecessary overhead. In a heavy TCP workload, we see 4% of CPU
time spent on the tasklet_trylock() in tasklet_action_common(), with a
smaller amount spent on the atomic operations in tasklet_schedule(),
tasklet_clear_sched(), and locking the spinlock in mlx5_cq_tasklet_cb().
TCP completions are handled by mlx5e_completion_event(), which schedules
NAPI to poll the queue, so they don't need tasklet processing.

Schedule the tasklet in mlx5_add_cq_to_tasklet() instead to avoid this
overhead. mlx5_add_cq_to_tasklet() is responsible for enqueuing the CQs
to be processed in tasklet context, so it can schedule the tasklet. CQs
that need tasklet processing have their interrupt comp handler set to
mlx5_add_cq_to_tasklet(), so they will schedule the tasklet. CQs that
don't need tasklet processing won't schedule the tasklet. To avoid
scheduling the tasklet multiple times during the same interrupt, only
schedule the tasklet in mlx5_add_cq_to_tasklet() if the tasklet work
queue was empty before the new CQ was pushed to it.

The additional branch in mlx5_add_cq_to_tasklet(), called for each EQE,
may add a small cost for the userspace Infiniband CQs whose completions
are processed in tasklet context. But this seems worth it to avoid the
tasklet overhead for CQs that don't need it.

Note that the mlx4 driver works the same way: it schedules the tasklet
in mlx4_add_cq_to_tasklet() and only if the work queue was empty before.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Parav Pandit <parav@nvidia.com>
Acked-by: Tariq Toukan <tariqt@nvidia.com>
Acked-by: Saeed Mahameed <saeedm@nvidia.com>
---
v4: add comment describing tasklet schedule condition
v3: revise commit message
v2: reorder variable declarations, describe CPU profile results

 drivers/net/ethernet/mellanox/mlx5/core/cq.c | 11 +++++++++++
 drivers/net/ethernet/mellanox/mlx5/core/eq.c |  5 +----
 2 files changed, 12 insertions(+), 4 deletions(-)

Comments

Jakub Kicinski Nov. 11, 2024, 10:13 p.m. UTC | #1 
On Tue,  5 Nov 2024 13:39:59 -0700 Caleb Sander Mateos wrote:
> Currently, the mlx5_eq_comp_int() interrupt handler schedules a tasklet
> to call mlx5_cq_tasklet_cb() if it processes any completions. For CQs
> whose completions don't need to be processed in tasklet context, this
> adds unnecessary overhead. In a heavy TCP workload, we see 4% of CPU
> time spent on the tasklet_trylock() in tasklet_action_common(), with a
> smaller amount spent on the atomic operations in tasklet_schedule(),
> tasklet_clear_sched(), and locking the spinlock in mlx5_cq_tasklet_cb().
> TCP completions are handled by mlx5e_completion_event(), which schedules
> NAPI to poll the queue, so they don't need tasklet processing.

Applied, thanks
diff mbox series

Patch

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c
index 4caa1b6f40ba..1fd403713baf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/cq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c
@@ -69,22 +69,33 @@  void mlx5_cq_tasklet_cb(struct tasklet_struct *t)
 static void mlx5_add_cq_to_tasklet(struct mlx5_core_cq *cq,
 				   struct mlx5_eqe *eqe)
 {
 	unsigned long flags;
 	struct mlx5_eq_tasklet *tasklet_ctx = cq->tasklet_ctx.priv;
+	bool schedule_tasklet = false;
 
 	spin_lock_irqsave(&tasklet_ctx->lock, flags);
 	/* When migrating CQs between EQs will be implemented, please note
 	 * that you need to sync this point. It is possible that
 	 * while migrating a CQ, completions on the old EQs could
 	 * still arrive.
 	 */
 	if (list_empty_careful(&cq->tasklet_ctx.list)) {
 		mlx5_cq_hold(cq);
+		/* If the tasklet CQ work list isn't empty, mlx5_cq_tasklet_cb()
+		 * is scheduled/running and hasn't processed the list yet, so it
+		 * will see this added CQ when it runs. If the list is empty,
+		 * the tasklet needs to be scheduled to pick up the CQ. The
+		 * spinlock avoids any race with the tasklet accessing the list.
+		 */
+		schedule_tasklet = list_empty(&tasklet_ctx->list);
 		list_add_tail(&cq->tasklet_ctx.list, &tasklet_ctx->list);
 	}
 	spin_unlock_irqrestore(&tasklet_ctx->lock, flags);
+
+	if (schedule_tasklet)
+		tasklet_schedule(&tasklet_ctx->task);
 }
 
 /* Callers must verify outbox status in case of err */
 int mlx5_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
 		   u32 *in, int inlen, u32 *out, int outlen)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
index 859dcf09b770..3fd2091c11c8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -112,14 +112,14 @@  static int mlx5_eq_comp_int(struct notifier_block *nb,
 	struct mlx5_eq_comp *eq_comp =
 		container_of(nb, struct mlx5_eq_comp, irq_nb);
 	struct mlx5_eq *eq = &eq_comp->core;
 	struct mlx5_eqe *eqe;
 	int num_eqes = 0;
-	u32 cqn = -1;
 
 	while ((eqe = next_eqe_sw(eq))) {
 		struct mlx5_core_cq *cq;
+		u32 cqn;
 
 		/* Make sure we read EQ entry contents after we've
 		 * checked the ownership bit.
 		 */
 		dma_rmb();
@@ -142,13 +142,10 @@  static int mlx5_eq_comp_int(struct notifier_block *nb,
 			break;
 	}
 
 	eq_update_ci(eq, 1);
 
-	if (cqn != -1)
-		tasklet_schedule(&eq_comp->tasklet_ctx.task);
-
 	return 0;
 }
 
 /* Some architectures don't latch interrupts when they are disabled, so using
  * mlx5_eq_poll_irq_disabled could end up losing interrupts while trying to