[PATCH 2/2] io_uring: fix shared sqpoll cancellation hangs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



[  736.982891] INFO: task iou-sqp-4294:4295 blocked for more than 122 seconds.
[  736.982897] Call Trace:
[  736.982901]  schedule+0x68/0xe0
[  736.982903]  io_uring_cancel_sqpoll+0xdb/0x110
[  736.982908]  io_sqpoll_cancel_cb+0x24/0x30
[  736.982911]  io_run_task_work_head+0x28/0x50
[  736.982913]  io_sq_thread+0x4e3/0x720

We call io_uring_cancel_sqpoll() one by one for each ctx either in
sq_thread() itself or via task works, and it's intended to cancel all
requests of a specified context. However the function uses per-task
counters to track the number of inflight requests, so it counts more
requests than available via currect io_uring ctx and goes to sleep for
them to appear (e.g. from IRQ), that will never happen.

Cancel a bit more than before, i.e. all ctxs that share sqpoll
and continue to use shared counters. Don't forget that we should not
remove ctx from the list before running that task_work sqpoll-cancel,
otherwise the function wouldn't be able to find the context and will
hang.

Reported-by: Joakim Hassila <joj@xxxxxxx>
Reported-by: Jens Axboe <axboe@xxxxxxxxx>
Fixes: 37d1e2e3642e2 ("io_uring: move SQPOLL thread io-wq forked worker")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
---
 fs/io_uring.c | 27 ++++++++++++++-------------
 1 file changed, 14 insertions(+), 13 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index fb41725204f0..99e55f7f6c34 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1022,7 +1022,7 @@ static void io_uring_del_task_file(unsigned long index);
 static void io_uring_try_cancel_requests(struct io_ring_ctx *ctx,
 					 struct task_struct *task,
 					 struct files_struct *files);
-static void io_uring_cancel_sqpoll(struct io_ring_ctx *ctx);
+static void io_uring_cancel_sqpoll(struct io_sq_data *sqd);
 static struct io_rsrc_node *io_rsrc_node_alloc(struct io_ring_ctx *ctx);
 
 static bool io_cqring_fill_event(struct io_kiocb *req, long res, unsigned cflags);
@@ -6867,15 +6867,14 @@ static int io_sq_thread(void *data)
 		timeout = jiffies + sqd->sq_thread_idle;
 	}
 
-	list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
-		io_uring_cancel_sqpoll(ctx);
+	io_uring_cancel_sqpoll(sqd);
 	sqd->thread = NULL;
 	list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
 		io_ring_set_wakeup_flag(ctx);
-	mutex_unlock(&sqd->lock);
-
 	io_run_task_work();
 	io_run_task_work_head(&sqd->park_task_work);
+	mutex_unlock(&sqd->lock);
+
 	complete(&sqd->exited);
 	do_exit(0);
 }
@@ -8867,11 +8866,11 @@ static s64 tctx_inflight(struct io_uring_task *tctx, bool tracked)
 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_ring_ctx *ctx = work->ctx;
-	struct io_sq_data *sqd = ctx->sq_data;
+	struct io_sq_data *sqd = work->ctx->sq_data;
 
 	if (sqd->thread)
-		io_uring_cancel_sqpoll(ctx);
+		io_uring_cancel_sqpoll(sqd);
+	list_del_init(&work->ctx->sqd_list);
 	complete(&work->completion);
 }
 
@@ -8882,7 +8881,6 @@ static void io_sqpoll_cancel_sync(struct io_ring_ctx *ctx)
 	struct task_struct *task;
 
 	io_sq_thread_park(sqd);
-	list_del_init(&ctx->sqd_list);
 	io_sqd_update_thread_idle(sqd);
 	task = sqd->thread;
 	if (task) {
@@ -8890,6 +8888,8 @@ static void io_sqpoll_cancel_sync(struct io_ring_ctx *ctx)
 		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_sq_thread_unpark(sqd);
 
@@ -8915,14 +8915,14 @@ static void io_uring_try_cancel(struct files_struct *files)
 }
 
 /* should only be called by SQPOLL task */
-static void io_uring_cancel_sqpoll(struct io_ring_ctx *ctx)
+static void io_uring_cancel_sqpoll(struct io_sq_data *sqd)
 {
-	struct io_sq_data *sqd = ctx->sq_data;
 	struct io_uring_task *tctx = current->io_uring;
+	struct io_ring_ctx *ctx;
 	s64 inflight;
 	DEFINE_WAIT(wait);
 
-	WARN_ON_ONCE(!sqd || ctx->sq_data->thread != current);
+	WARN_ON_ONCE(!sqd || sqd->thread != current);
 
 	atomic_inc(&tctx->in_idle);
 	do {
@@ -8930,7 +8930,8 @@ static void io_uring_cancel_sqpoll(struct io_ring_ctx *ctx)
 		inflight = tctx_inflight(tctx, false);
 		if (!inflight)
 			break;
-		io_uring_try_cancel_requests(ctx, current, NULL);
+		list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
+			io_uring_try_cancel_requests(ctx, current, NULL);
 
 		prepare_to_wait(&tctx->wait, &wait, TASK_UNINTERRUPTIBLE);
 		/*
-- 
2.31.1




[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux