This is a note to let you know that I've just added the patch titled
io_uring: ensure local task_work is run on wait timeout
to the 6.7-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
io_uring-ensure-local-task_work-is-run-on-wait-timeout.patch
and it can be found in the queue-6.7 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From 6ff1407e24e6fdfa4a16ba9ba551e3d253a26391 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@xxxxxxxxx>
Date: Thu, 4 Jan 2024 12:21:08 -0700
Subject: io_uring: ensure local task_work is run on wait timeout
From: Jens Axboe <axboe@xxxxxxxxx>
commit 6ff1407e24e6fdfa4a16ba9ba551e3d253a26391 upstream.
A previous commit added an earlier break condition here, which is fine if
we're using non-local task_work as it'll be run on return to userspace.
However, if DEFER_TASKRUN is used, then we could be leaving local
task_work that is ready to process in the ctx list until next time that
we enter the kernel to wait for events.
Move the break condition to _after_ we have run task_work.
Cc: stable@xxxxxxxxxxxxxxx
Fixes: 846072f16eed ("io_uring: mimimise io_cqring_wait_schedule")
Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
io_uring/io_uring.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -2637,8 +2637,6 @@ static int io_cqring_wait(struct io_ring
__set_current_state(TASK_RUNNING);
atomic_set(&ctx->cq_wait_nr, 0);
- if (ret < 0)
- break;
/*
* Run task_work after scheduling and before io_should_wake().
* If we got woken because of task_work being processed, run it
@@ -2648,6 +2646,18 @@ static int io_cqring_wait(struct io_ring
if (!llist_empty(&ctx->work_llist))
io_run_local_work(ctx);
+ /*
+ * Non-local task_work will be run on exit to userspace, but
+ * if we're using DEFER_TASKRUN, then we could have waited
+ * with a timeout for a number of requests. If the timeout
+ * hits, we could have some requests ready to process. Ensure
+ * this break is _after_ we have run task_work, to avoid
+ * deferring running potentially pending requests until the
+ * next time we wait for events.
+ */
+ if (ret < 0)
+ break;
+
check_cq = READ_ONCE(ctx->check_cq);
if (unlikely(check_cq)) {
/* let the caller flush overflows, retry */
Patches currently in stable-queue which might be from axboe@xxxxxxxxx are
queue-6.7/blk-cgroup-fix-rcu-lockdep-warning-in-blkg_lookup.patch
queue-6.7/block-fix-iterating-over-an-empty-bio-with-bio_for_each_folio_all.patch
queue-6.7/io_uring-don-t-check-iopoll-if-request-completes.patch
queue-6.7/null_blk-don-t-cap-max_hw_sectors-to-blk_def_max_sec.patch
queue-6.7/block-add-check-of-minors-and-first_minor-in-device_.patch
queue-6.7/io_uring-rw-ensure-io-bytes_done-is-always-initialized.patch
queue-6.7/md-raid1-use-blk_opf_t-for-read-and-write-operations.patch
queue-6.7/block-remove-special-casing-of-compound-pages.patch
queue-6.7/block-set-memalloc_noio-to-false-on-device_add_disk-.patch
queue-6.7/io_uring-ensure-local-task_work-is-run-on-wait-timeout.patch
queue-6.7/scsi-target-core-add-missing-file_-start-end-_write.patch
queue-6.7/block-add-check-that-partition-length-needs-to-be-aligned-with-block-size.patch
