[PATCH] io_uring: fix link timeout refs

[PATCH] io_uring: fix link timeout refs

[Pavel Begunkov]

[ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]

WARNING: CPU: 0 PID: 10242 at lib/refcount.c:28 refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
RIP: 0010:refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
Call Trace:
 __refcount_sub_and_test include/linux/refcount.h:283 [inline]
 __refcount_dec_and_test include/linux/refcount.h:315 [inline]
 refcount_dec_and_test include/linux/refcount.h:333 [inline]
 io_put_req fs/io_uring.c:2140 [inline]
 io_queue_linked_timeout fs/io_uring.c:6300 [inline]
 __io_queue_sqe+0xbef/0xec0 fs/io_uring.c:6354
 io_submit_sqe fs/io_uring.c:6534 [inline]
 io_submit_sqes+0x2bbd/0x7c50 fs/io_uring.c:6660
 __do_sys_io_uring_enter fs/io_uring.c:9240 [inline]
 __se_sys_io_uring_enter+0x256/0x1d60 fs/io_uring.c:9182

io_link_timeout_fn() should put only one reference of the linked timeout
request, however in case of racing with the master request's completion
first io_req_complete() puts one and then io_put_req_deferred() is
called.

Cc: stable@vger.kernel.org # 5.12+
Fixes: 9ae1f8dd372e0 ("io_uring: fix inconsistent lock state")
Reported-by: syzbot+a2910119328ce8e7996f@syzkaller.appspotmail.com
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/ff51018ff29de5ffa76f09273ef48cb24c720368.1620417627.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 42153106b7bc..42439838eaf7 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -6260,7 +6260,6 @@ static enum hrtimer_restart io_link_timeout_fn(struct hrtimer *timer)
 	if (prev) {
 		io_async_find_and_cancel(ctx, req, prev->user_data, -ETIME);
 		io_put_req_deferred(prev, 1);
-		io_put_req_deferred(req, 1);
 	} else {
 		io_cqring_add_event(req, -ETIME, 0);
 		io_put_req_deferred(req, 1);
-- 
2.32.0

Re: [PATCH] io_uring: fix link timeout refs

[Pavel Begunkov]

On 7/26/21 4:17 PM, Pavel Begunkov wrote:
> [ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]

Looking at it, it just reverts the backported patch,
i.e. 0b2a990e5d2f76d020cb840c456e6ec5f0c27530.
Wasn't needed in 5.10 in the first place.

Sudip, would be great if you can try it out


> WARNING: CPU: 0 PID: 10242 at lib/refcount.c:28 refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
> RIP: 0010:refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
> Call Trace:
>  __refcount_sub_and_test include/linux/refcount.h:283 [inline]
>  __refcount_dec_and_test include/linux/refcount.h:315 [inline]
>  refcount_dec_and_test include/linux/refcount.h:333 [inline]
>  io_put_req fs/io_uring.c:2140 [inline]
>  io_queue_linked_timeout fs/io_uring.c:6300 [inline]
>  __io_queue_sqe+0xbef/0xec0 fs/io_uring.c:6354
>  io_submit_sqe fs/io_uring.c:6534 [inline]
>  io_submit_sqes+0x2bbd/0x7c50 fs/io_uring.c:6660
>  __do_sys_io_uring_enter fs/io_uring.c:9240 [inline]
>  __se_sys_io_uring_enter+0x256/0x1d60 fs/io_uring.c:9182
> 
> io_link_timeout_fn() should put only one reference of the linked timeout
> request, however in case of racing with the master request's completion
> first io_req_complete() puts one and then io_put_req_deferred() is
> called.
> 
> Cc: stable@vger.kernel.org # 5.12+
> Fixes: 9ae1f8dd372e0 ("io_uring: fix inconsistent lock state")
> Reported-by: syzbot+a2910119328ce8e7996f@syzkaller.appspotmail.com
> Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
> Link: https://lore.kernel.org/r/ff51018ff29de5ffa76f09273ef48cb24c720368.1620417627.git.asml.silence@gmail.com
> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> ---
>  fs/io_uring.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index 42153106b7bc..42439838eaf7 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -6260,7 +6260,6 @@ static enum hrtimer_restart io_link_timeout_fn(struct hrtimer *timer)
>  	if (prev) {
>  		io_async_find_and_cancel(ctx, req, prev->user_data, -ETIME);
>  		io_put_req_deferred(prev, 1);
> -		io_put_req_deferred(req, 1);
>  	} else {
>  		io_cqring_add_event(req, -ETIME, 0);
>  		io_put_req_deferred(req, 1);
> 

-- 
Pavel Begunkov

Re: [PATCH] io_uring: fix link timeout refs

[Sudip Mukherjee]

On Mon, Jul 26, 2021 at 4:22 PM Pavel Begunkov <asml.silence@gmail.com> wrote:
>
> On 7/26/21 4:17 PM, Pavel Begunkov wrote:
> > [ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]
>
> Looking at it, it just reverts the backported patch,
> i.e. 0b2a990e5d2f76d020cb840c456e6ec5f0c27530.
> Wasn't needed in 5.10 in the first place.
>
> Sudip, would be great if you can try it out

Applied on top of v5.10.54-rc2 and tested, issue not reproduced. Thanks Pavel.

Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>


-- 
Regards
Sudip

Re: [PATCH] io_uring: fix link timeout refs

[Greg Kroah-Hartman]

On Mon, Jul 26, 2021 at 06:07:52PM +0100, Sudip Mukherjee wrote:
> On Mon, Jul 26, 2021 at 4:22 PM Pavel Begunkov <asml.silence@gmail.com> wrote:
> >
> > On 7/26/21 4:17 PM, Pavel Begunkov wrote:
> > > [ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]
> >
> > Looking at it, it just reverts the backported patch,
> > i.e. 0b2a990e5d2f76d020cb840c456e6ec5f0c27530.
> > Wasn't needed in 5.10 in the first place.
> >
> > Sudip, would be great if you can try it out
> 
> Applied on top of v5.10.54-rc2 and tested, issue not reproduced. Thanks Pavel.
> 
> Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>

So this is all good in the latest 5.10.54 release, right?

thanks,

greg k-h

Re: [PATCH] io_uring: fix link timeout refs

[Pavel Begunkov]

On 7/28/21 5:13 PM, Greg Kroah-Hartman wrote:
> On Mon, Jul 26, 2021 at 06:07:52PM +0100, Sudip Mukherjee wrote:
>> On Mon, Jul 26, 2021 at 4:22 PM Pavel Begunkov <asml.silence@gmail.com> wrote:
>>>
>>> On 7/26/21 4:17 PM, Pavel Begunkov wrote:
>>>> [ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]
>>>
>>> Looking at it, it just reverts the backported patch,
>>> i.e. 0b2a990e5d2f76d020cb840c456e6ec5f0c27530.
>>> Wasn't needed in 5.10 in the first place.
>>>
>>> Sudip, would be great if you can try it out
>>
>> Applied on top of v5.10.54-rc2 and tested, issue not reproduced. Thanks Pavel.
>>
>> Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>
> 
> So this is all good in the latest 5.10.54 release, right?

Right. And you can either take this or revert 0b2a990e5d2f76d0,
both ways are identical.

-- 
Pavel Begunkov

Re: [PATCH] io_uring: fix link timeout refs

[Greg Kroah-Hartman]

On Wed, Jul 28, 2021 at 05:30:22PM +0100, Pavel Begunkov wrote:
> On 7/28/21 5:13 PM, Greg Kroah-Hartman wrote:
> > On Mon, Jul 26, 2021 at 06:07:52PM +0100, Sudip Mukherjee wrote:
> >> On Mon, Jul 26, 2021 at 4:22 PM Pavel Begunkov <asml.silence@gmail.com> wrote:
> >>>
> >>> On 7/26/21 4:17 PM, Pavel Begunkov wrote:
> >>>> [ Upstream commit a298232ee6b9a1d5d732aa497ff8be0d45b5bd82 ]
> >>>
> >>> Looking at it, it just reverts the backported patch,
> >>> i.e. 0b2a990e5d2f76d020cb840c456e6ec5f0c27530.
> >>> Wasn't needed in 5.10 in the first place.
> >>>
> >>> Sudip, would be great if you can try it out
> >>
> >> Applied on top of v5.10.54-rc2 and tested, issue not reproduced. Thanks Pavel.
> >>
> >> Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>
> > 
> > So this is all good in the latest 5.10.54 release, right?
> 
> Right. And you can either take this or revert 0b2a990e5d2f76d0,
> both ways are identical.

Ah, I missed this, for some reason I thought this was already address.
Ok, now queued up.  Hopefully this is now fixed...

greg k-h