Re: deadlock when using iothread during backup_clean()

[Fiona Ebner <f.ebner@proxmox.com>]

Am 19.10.23 um 15:53 schrieb Fiona Ebner:
> Am 19.10.23 um 14:14 schrieb Kevin Wolf:
>> Am 18.10.2023 um 11:42 hat Fiona Ebner geschrieben:
>>> Am 17.10.23 um 16:20 schrieb Kevin Wolf:
>>>> Am 17.10.2023 um 15:37 hat Fiona Ebner geschrieben:
>>>>> Am 17.10.23 um 14:12 schrieb Kevin Wolf:
>>>>>> Am 17.10.2023 um 12:18 hat Fiona Ebner geschrieben:
>>>>>>> I ran into similar issues now with mirror, (both deadlocks and stuck
>>>>>>> guest IO at other times), and interestingly, also during job start.
>>>>>>>
>>>>>>> Also had a backtrace similar to [0] once, so I took a closer look.
>>>>>>> Probably was obvious to others already, but for the record:
>>>>>>>
>>>>>>> 1. the graph is locked by the main thread
>>>>>>> 2. the iothread holds the AioContext lock
>>>>>>> 3. the main thread waits on the AioContext lock
>>>>>>> 4. the iothread waits for coroutine spawned by blk_is_available()
>>>>>>
>>>>>> Where does this blk_is_available() in the iothread come from? Having it
>>>>>> wait without dropping the AioContext lock sounds like something that
>>>>>> we'd want to avoid. Ideally, devices using iothreads shouldn't use
>>>>>> synchronous requests at all, but I think scsi-disk might have some of
>>>>>> them.
>>>>>>
>>>>>
>>>>> It's part of the request handling in virtio-scsi:
>>>>>
>>>>>> #0  0x00007ff7f5f55136 in __ppoll (fds=0x7ff7e40030c0, nfds=8, timeout=<optimized out>, sigmask=0x0) at ../sysdeps/unix/sysv/linux/ppoll.c:42
>>>>>> #1  0x00005587132615ab in qemu_poll_ns (fds=0x7ff7e40030c0, nfds=8, timeout=-1) at ../util/qemu-timer.c:339
>>>>>> #2  0x000055871323e8b1 in fdmon_poll_wait (ctx=0x55871598d5e0, ready_list=0x7ff7f288ebe0, timeout=-1) at ../util/fdmon-poll.c:79
>>>>>> #3  0x000055871323e1ed in aio_poll (ctx=0x55871598d5e0, blocking=true) at ../util/aio-posix.c:670
>>>>>> #4  0x0000558713089efa in bdrv_poll_co (s=0x7ff7f288ec90) at /home/febner/repos/qemu/block/block-gen.h:43
>>>>>> #5  0x000055871308c362 in blk_is_available (blk=0x55871599e2f0) at block/block-gen.c:1426
>>>>>> #6  0x0000558712f6843b in virtio_scsi_ctx_check (s=0x558716c049c0, d=0x55871581cd30) at ../hw/scsi/virtio-scsi.c:290
>>>>
>>>> Oh... So essentially for an assertion.
>>>>
>>>> I wonder if the blk_is_available() check introduced in 2a2d69f490c is
>>>> even necessary any more, because BlockBackend has its own AioContext
>>>> now. And if blk_bs(blk) != NULL isn't what we actually want to check if
>>>> the check is necessary, because calling bdrv_is_inserted() doesn't seem
>>>> to have been intended. blk_bs() wouldn't have to poll.
>>>>
>>>
>>> Could virtio_scsi_hotunplug() be an issue with removing or modifying
>>> the check? There's a call there which sets the blk's AioContext to
>>> qemu_get_aio_context(). Or are we sure that the assert in
>>> virtio_scsi_ctx_check() can't be reached after that?
>>
>> I think that would be the kind of bug that the assertion tries to
>> catch, because then we would be sending requests to blk from a thread
>> that doesn't match its AioContext (which will be allowed soon, but not
>> quite yet).
>>
>> Before resetting the AioContext, virtio_scsi_hotunplug() calls
>> qdev_simple_device_unplug_cb(), which unrealizes the SCSI device. This
>> calls scsi_qdev_unrealize() -> scsi_device_purge_requests(), which in
>> turn drains blk and cancels all pending requests. So there should be
>> nothing left that could call into virtio_scsi_ctx_check() any more.
>>
>> The other argument is that after unrealize, virtio_scsi_device_get()
>> would return NULL anyway, so even if a request were still pending, it
>> would just fail instead of accessing the unplugged device.
>>
> 
> Okay, sounds like a way to get around that deadlock issue then :)
> 

Unfortunately, scsi_dma_command() also has a blk_is_available() call and
I ran into a similar deadlock with that once.

> (...)
> 
>>>>
>>>> What does the stuck I/O look like? Is it stuck in the backend, i.e. the
>>>> device started requests that never complete? Or stuck from the guest
>>>> perspective, i.e. the device never checks for new requests?
>>>>
>>>
>>> AFAICT, from the guest perspective.
>>>
>>>> I don't really have an idea immediately, we'd have to find out where the
>>>> stuck I/O stops being processed.
>>>>
>>>
>>> I've described it in an earlier mail in this thread:
>>> https://lists.nongnu.org/archive/html/qemu-devel/2023-10/msg01900.html
>>>
>>> Quoting from there:
>>>
>>>> After the IO was stuck in the guest, I used bdrv_next_all_states() to
>>>> iterate over the states and there's only the bdrv_raw and the
>>>> bdrv_host_device. For both, tracked_requests was empty.
>>
>> And bs->in_flight and blk->in_flight are 0, too?
>>
> 
> Yes. And queued_requests in the BlockBackend is also empty.
> 
>> Is anything quiesced?
> 
> No. quiesce_counter is 0 for both BlockDriverState instances as well as
> for the BlockBackend. quiesced_parent is false for both parents (i.e.
> child_root for the bdrv_raw and child_of_bds for the bdrv_file (this
> time I used VirtIO SCSI, in the quote it was VirtIO block)).
> 
>>>> What is also very interesting is that the IO isn't always dead
>>>> immediately. It can be that the fio command still runs with lower speed
>>>> for a while (sometimes even up to about a minute, but most often about
>>>> 10-15 seconds or so). During that time, I still can see calls to
>>>> virtio_scsi_handle_cmd() and blk_aio_write_entry(). Then they suddenly stop.
>>>
>>> Noting again that (at least for backup) it happens with both virtio-blk
>>> and virtio-scsi and with both aio=io_uring and aio=threads. I also tried
>>> different host kernels 5.15, 6.2 and 6.5 and guest kernels 5.10 and 6.1.
>>
>> You say "at least for backup". Did you see the bug for other job types,
>> too?
>>
> 
> Yes, sorry. I meant to say that I only tested it for the backup
> canceling with all those configurations. I did run into the stuck guest
> IO issue in three different scenarios: canceling backup, canceling
> mirror and starting mirror:
> 
>>>>>> Am 17.10.2023 um 12:18 hat Fiona Ebner geschrieben:
>>>>>>> I ran into similar issues now with mirror, (both deadlocks and stuck
>>>>>>> guest IO at other times), and interestingly, also during job start.
> 
> But for mirror, I only used VirtIO SCSI and aio=io_uring and didn't test
> other configurations.
> 

I should also mention that I used drive-mirror and drive-backup. I can
also trigger the issue with blockdev-mirror when canceling, but didn't
run into it when starting with blockdev-mirror yet.

And I found that with drive-mirror, the issue during starting seems to
manifest with the bdrv_open() call. Adding a return before it, the guest
IO didn't get stuck in my testing, but adding a return after it, it can
get stuck. I'll try to see if I can further narrow it down next week,
but maybe that's already a useful hint?

> diff --git a/blockdev.c b/blockdev.c
> index a01c62596b..156915d15d 100644
> --- a/blockdev.c
> +++ b/blockdev.c
> @@ -3126,6 +3126,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
>      }
>      aio_context_release(aio_context);
>  
> +    // return; // GOOD
> +
>      /* Mirroring takes care of copy-on-write using the source's backing
>       * file.
>       */
> @@ -3133,6 +3135,8 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp)
>      target_bs = bdrv_open(arg->target, NULL, options, flags, errp);
>      aio_context_release(qemu_get_aio_context());
>  
> +    // return; // BAD
> +
>      if (!target_bs) {
>          return;
>      }


Best Regards,
Fiona