io_uring: don't submit sqes when ctx->refs is dying (6b668c9b) · Commits · 戴 / test

When IORING_SETUP_SQPOLL is enabled, io_ring_ctx_wait_and_kill() will wait for sq thread to idle by busy loop: while (ctx->sqo_thread && !wq_has_sleeper(&ctx->sqo_wait)) cond_resched(); Above loop isn't very CPU friendly, it may introduce a short cpu burst on the current cpu. If ctx->refs is dying, we forbid sq_thread from submitting any further SQEs. Instead they just get discarded when we exit. Signed-off-by:

Xiaoguang Wang <xiaoguang.wang@linux.alibaba.com> Signed-off-by:

Jens Axboe <axboe@kernel.dk>

fs/io_uring.c

+2 −11

Original line number	Diff line number	Diff line
		@@ -6175,6 +6175,7 @@ static int io_sq_thread(void *data)
		}

		mutex_lock(&ctx->uring_lock);
		if (likely(!percpu_ref_is_dying(&ctx->refs)))
		ret = io_submit_sqes(ctx, to_submit, NULL, -1);
		mutex_unlock(&ctx->uring_lock);
		timeout = jiffies + ctx->sq_thread_idle;
		@@ -7465,16 +7466,6 @@ static void io_ring_ctx_wait_and_kill(struct io_ring_ctx *ctx)
		percpu_ref_kill(&ctx->refs);
		mutex_unlock(&ctx->uring_lock);

		/*
		* Wait for sq thread to idle, if we have one. It won't spin on new
		* work after we've killed the ctx ref above. This is important to do
		* before we cancel existing commands, as the thread could otherwise
		* be queueing new work post that. If that's work we need to cancel,
		* it could cause shutdown to hang.
		*/
		while (ctx->sqo_thread && !wq_has_sleeper(&ctx->sqo_wait))
		cond_resched();

		io_kill_timeouts(ctx);
		io_poll_remove_all(ctx);

Admin message