Commit 9f59a9d8 authored by Pavel Begunkov's avatar Pavel Begunkov Committed by Jens Axboe
Browse files

io_uring: simplify SQPOLL cancellations 



All sqpoll rings (even sharing sqpoll task) are currently dead bound
to the task that created them, iow when owner task dies it kills all
its SQPOLL rings and their inflight requests via task_work infra. It's
neither the nicist way nor the most convenient as adds extra
locking/waiting and dependencies.

Leave it alone and rely on SIGKILL being delivered on its thread group
exit, so there are only two cases left:

1) thread group is dying, so sqpoll task gets a signal and exit itself
   cancelling all requests.

2) an sqpoll ring is dying. Because refs_kill() is called the sqpoll not
   going to submit any new request, and that's what we need. And
   io_ring_exit_work() will do all the cancellation itself before
   actually killing ctx, so sqpoll doesn't need to worry about it.

Signed-off-by: default avatarPavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/3cd7f166b9c326a2c932b70e71a655b03257b366.1619389911.git.asml.silence@gmail.com


Signed-off-by: default avatarJens Axboe <axboe@kernel.dk>
parent 28090c13

fs/io_uring.c

+3 −42
Original line number Diff line number Diff line
@@ -9021,41 +9021,6 @@ static s64 tctx_inflight(struct io_uring_task *tctx, bool tracked) 
	return percpu_counter_sum(&tctx->inflight);

}



static void io_sqpoll_cancel_cb(struct callback_head *cb)

{

	struct io_tctx_exit *work = container_of(cb, struct io_tctx_exit, task_work);

	struct io_sq_data *sqd = work->ctx->sq_data;



	if (sqd->thread)

		io_uring_cancel_sqpoll(sqd);

	list_del_init(&work->ctx->sqd_list);

	io_sqd_update_thread_idle(sqd);

	complete(&work->completion);

}



static void io_sqpoll_cancel_sync(struct io_ring_ctx *ctx)

{

	struct io_sq_data *sqd = ctx->sq_data;

	struct io_tctx_exit work = { .ctx = ctx, };

	struct task_struct *task;



	io_sq_thread_park(sqd);

	task = sqd->thread;

	if (task) {

		init_completion(&work.completion);

		init_task_work(&work.task_work, io_sqpoll_cancel_cb);

		io_task_work_add_head(&sqd->park_task_work, &work.task_work);

		wake_up_process(task);

	} else {

		list_del_init(&ctx->sqd_list);

		io_sqd_update_thread_idle(sqd);

	}

	io_sq_thread_unpark(sqd);



	if (task)

		wait_for_completion(&work.completion);

}



static void io_uring_try_cancel(struct files_struct *files)

{

	struct io_uring_task *tctx = current->io_uring;
@@ -9065,10 +9030,8 @@ static void io_uring_try_cancel(struct files_struct *files) 
	xa_for_each(&tctx->xa, index, node) {

		struct io_ring_ctx *ctx = node->ctx;



		if (ctx->sq_data) {

			io_sqpoll_cancel_sync(ctx);

			continue;

		}

		/* sqpoll task will cancel all its requests */

		if (!ctx->sq_data)

			io_uring_try_cancel_requests(ctx, current, files);

	}

}
@@ -9117,8 +9080,6 @@ void __io_uring_cancel(struct files_struct *files) 


	/* make sure overflow events are dropped */

	atomic_inc(&tctx->in_idle);

	io_uring_try_cancel(files);



	do {

		/* read completions before cancelations */

		inflight = tctx_inflight(tctx, !!files);