Hi,
I hit WARNING below after update to 3.19-rc6, when I running fio with libaio.
Any one see this warning before, or maybe know how to fix it?
>From the call trace:
(gdb) list *read_events+0x1e8
0xffffffff81225ad8 is in read_events (fs/aio.c:1259).
1254 * something to be aware of when touching this code.
1255 */
1256 if (until.tv64 == 0)
1257 aio_read_events(ctx, min_nr, nr, event, &ret);
1258 else
1259 wait_event_interruptible_hrtimeout(ctx->wait,
1260 aio_read_events(ctx, min_nr, nr, event, &ret),
1261 until);
1262
1263 if (!ret && signal_pending(current))
(gdb)
looks like the WARNING come from wait_event_interruptible_hrtimeout?
[ 165.011157] ------------[ cut here ]------------
[ 165.011182] WARNING: CPU: 27 PID: 6295 at kernel/sched/core.c:7300
__might_sleep+0xbd/0xd0()
[ 165.011197] do not call blocking ops when !TASK_RUNNING; state=1
set at [<ffffffff810d17f9>] prepare_to_wait_event+0x59/0x110
[ 165.011199] Modules linked in: ibnbd_client(O) ibtrs_client(O)
ib_ipoib ib_uverbs ib_umad mlx4_ib mlx4_en null_blk rdma_cm iw_cm
ib_cm ib_sa ib_mad ib_core ib_addr mlx4_core
[ 165.011224] CPU: 27 PID: 6295 Comm: fio Tainted: G O
3.19.0-rc6debug #3
[ 165.011226] Hardware name: Supermicro SBA-7142G-T4/BHQGE, BIOS 3.00
10/24/2012
[ 165.011228] ffffffff81a56455 ffff88081173bc58 ffffffff817d232b
0000000000000007
[ 165.011233] ffff88081173bca8 ffff88081173bc98 ffffffff8109237a
ffff88081173bc88
[ 165.011236] ffffffff81a573d8 000000000000026d 0000000000000000
ffff881011e70000
[ 165.011240] Call Trace:
[ 165.011247] [<ffffffff817d232b>] dump_stack+0x4c/0x65
[ 165.011252] [<ffffffff8109237a>] warn_slowpath_common+0x8a/0xc0
[ 165.011255] [<ffffffff810923f6>] warn_slowpath_fmt+0x46/0x50
[ 165.011260] [<ffffffff810daab5>] ? mark_held_locks+0x75/0xa0
[ 165.011264] [<ffffffff817d73b9>] ? mutex_lock_nested+0x289/0x3c0
[ 165.011268] [<ffffffff810d17f9>] ? prepare_to_wait_event+0x59/0x110
[ 165.011271] [<ffffffff810d17f9>] ? prepare_to_wait_event+0x59/0x110
[ 165.011274] [<ffffffff810bb6ed>] __might_sleep+0xbd/0xd0
[ 165.011277] [<ffffffff817d715f>] mutex_lock_nested+0x2f/0x3c0
[ 165.011280] [<ffffffff810c5a65>] ? local_clock+0x25/0x30
[ 165.011284] [<ffffffff817da7d6>] ? _raw_spin_unlock_irqrestore+0x36/0x60
[ 165.011290] [<ffffffff812255fc>] aio_read_events+0x4c/0x340
[ 165.011293] [<ffffffff810d1837>] ? prepare_to_wait_event+0x97/0x110
[ 165.011296] [<ffffffff81225ad8>] read_events+0x1e8/0x300
[ 165.011300] [<ffffffff810d1b60>] ? wait_woken+0xc0/0xc0
[ 165.011305] [<ffffffff810fa7c0>] ? __run_hrtimer+0x5e0/0x5e0
[ 165.011308] [<ffffffff81224d30>] ? lookup_ioctx+0x1c0/0x2a0
[ 165.011312] [<ffffffff81227982>] SyS_io_getevents+0x62/0x1a0
[ 165.011317] [<ffffffff813b490b>] ? trace_hardirqs_on_thunk+0x3a/0x3f
[ 165.011321] [<ffffffff817db392>] system_call_fastpath+0x12/0x17
[ 165.011323] ---[ end trace 5e2ce0c3bdb141cd ]---
Best Regards,
Jack Wang
On Wed, Feb 4, 2015 at 1:57 AM, Jinpu Wang <[email protected]> wrote:
> Hi,
>
> I hit WARNING below after update to 3.19-rc6, when I running fio with libaio.
> Any one see this warning before, or maybe know how to fix it?
>
> From the call trace:
> (gdb) list *read_events+0x1e8
> 0xffffffff81225ad8 is in read_events (fs/aio.c:1259).
> 1254 * something to be aware of when touching this code.
> 1255 */
> 1256 if (until.tv64 == 0)
> 1257 aio_read_events(ctx, min_nr, nr, event, &ret);
> 1258 else
> 1259 wait_event_interruptible_hrtimeout(ctx->wait,
> 1260 aio_read_events(ctx, min_nr, nr, event, &ret),
> 1261 until);
> 1262
> 1263 if (!ret && signal_pending(current))
> (gdb)
>
> looks like the WARNING come from wait_event_interruptible_hrtimeout?
Probably fixed by:
http://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/commit/?id=3e87523897e18a3e17fc8955ed795188be737ff1
Thanks Cong,
I found another thread:
https://lkml.org/lkml/2015/2/3/690
Looks the patch fix the same WARNING.
On Thu, Feb 5, 2015 at 7:33 AM, Cong Wang <[email protected]> wrote:
> On Wed, Feb 4, 2015 at 1:57 AM, Jinpu Wang <[email protected]> wrote:
>> Hi,
>>
>> I hit WARNING below after update to 3.19-rc6, when I running fio with libaio.
>> Any one see this warning before, or maybe know how to fix it?
>>
>> From the call trace:
>> (gdb) list *read_events+0x1e8
>> 0xffffffff81225ad8 is in read_events (fs/aio.c:1259).
>> 1254 * something to be aware of when touching this code.
>> 1255 */
>> 1256 if (until.tv64 == 0)
>> 1257 aio_read_events(ctx, min_nr, nr, event, &ret);
>> 1258 else
>> 1259 wait_event_interruptible_hrtimeout(ctx->wait,
>> 1260 aio_read_events(ctx, min_nr, nr, event, &ret),
>> 1261 until);
>> 1262
>> 1263 if (!ret && signal_pending(current))
>> (gdb)
>>
>> looks like the WARNING come from wait_event_interruptible_hrtimeout?
>
> Probably fixed by:
> http://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/commit/?id=3e87523897e18a3e17fc8955ed795188be737ff1
--
Mit freundlichen Grüßen,
Best Regards,
Jack Wang