[pve-devel] [PATCH qemu] block/io_uring: resubmit when result is -EAGAIN

[pve-devel] [PATCH qemu] block/io_uring: resubmit when result is -EAGAIN

[Fabian Ebner]

Quoting from [0]:

 Some setups, like SCSI, can throw spurious -EAGAIN off the softirq
 completion path. Normally we expect this to happen inline as part
 of submission, but apparently SCSI has a weird corner case where it
 can happen as part of normal completions.

Host kernels without patch [0] can panic when this happens [1], and
resubmitting makes the panic more likely. On the other hand, for
kernels with patch [0], resubmitting ensures that a block job is not
aborted just because of such spurious errors. In particular, this
should fix the problem reported in [2] with PBS backups of LVM-backed
VMs.

[0]: https://lore.kernel.org/io-uring/20210727165811.284510-3-axboe@kernel.dk/T/#u

[1]:
  #9 [ffffb732000c8b70] asm_exc_page_fault at ffffffffa4800ade
 #10 [ffffb732000c8bf8] io_prep_async_work at ffffffffa3d89c16
 #11 [ffffb732000c8c50] io_rw_reissue at ffffffffa3d8b2e1
 #12 [ffffb732000c8c78] io_complete_rw at ffffffffa3d8baa8
 #13 [ffffb732000c8c98] blkdev_bio_end_io at ffffffffa3d62a80
 #14 [ffffb732000c8cc8] bio_endio at ffffffffa3f4e800
 #15 [ffffb732000c8ce8] dec_pending at ffffffffa432f854
 #16 [ffffb732000c8d30] clone_endio at ffffffffa433170c
 #17 [ffffb732000c8d70] bio_endio at ffffffffa3f4e800
 #18 [ffffb732000c8d90] blk_update_request at ffffffffa3f53a37
 #19 [ffffb732000c8dd0] scsi_end_request at ffffffffa4233a5c
 #20 [ffffb732000c8e08] scsi_io_completion at ffffffffa423432c
 #21 [ffffb732000c8e58] scsi_finish_command at ffffffffa422c527
 #22 [ffffb732000c8e88] scsi_softirq_done at ffffffffa42341e4

[2]: https://forum.proxmox.com/threads/backup-job-failed-with-err-11-on-2-of-6-vms.92568/

Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
---

Mail sent upstream:
https://lists.nongnu.org/archive/html/qemu-devel/2021-07/msg06765.html

 block/io_uring.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/io_uring.c b/block/io_uring.c
index 00a3ee9fb8..77d162cb24 100644
--- a/block/io_uring.c
+++ b/block/io_uring.c
@@ -165,7 +165,7 @@ static void luring_process_completions(LuringState *s)
         total_bytes = ret + luringcb->total_read;
 
         if (ret < 0) {
-            if (ret == -EINTR) {
+            if (ret == -EINTR || ret == -EAGAIN) {
                 luring_resubmit(s, luringcb);
                 continue;
             }
-- 
2.30.2

Re: [pve-devel] [PATCH qemu] block/io_uring: resubmit when result is -EAGAIN

[Thomas Lamprecht]

On 28/07/2021 12:55, Fabian Ebner wrote:
> Quoting from [0]:
> 
>  Some setups, like SCSI, can throw spurious -EAGAIN off the softirq
>  completion path. Normally we expect this to happen inline as part
>  of submission, but apparently SCSI has a weird corner case where it
>  can happen as part of normal completions.
> 
> Host kernels without patch [0] can panic when this happens [1], and
> resubmitting makes the panic more likely. On the other hand, for
> kernels with patch [0], resubmitting ensures that a block job is not
> aborted just because of such spurious errors. In particular, this
> should fix the problem reported in [2] with PBS backups of LVM-backed
> VMs.
> 
> [0]: https://lore.kernel.org/io-uring/20210727165811.284510-3-axboe@kernel.dk/T/#u
> 
> [1]:
>   #9 [ffffb732000c8b70] asm_exc_page_fault at ffffffffa4800ade
>  #10 [ffffb732000c8bf8] io_prep_async_work at ffffffffa3d89c16
>  #11 [ffffb732000c8c50] io_rw_reissue at ffffffffa3d8b2e1
>  #12 [ffffb732000c8c78] io_complete_rw at ffffffffa3d8baa8
>  #13 [ffffb732000c8c98] blkdev_bio_end_io at ffffffffa3d62a80
>  #14 [ffffb732000c8cc8] bio_endio at ffffffffa3f4e800
>  #15 [ffffb732000c8ce8] dec_pending at ffffffffa432f854
>  #16 [ffffb732000c8d30] clone_endio at ffffffffa433170c
>  #17 [ffffb732000c8d70] bio_endio at ffffffffa3f4e800
>  #18 [ffffb732000c8d90] blk_update_request at ffffffffa3f53a37
>  #19 [ffffb732000c8dd0] scsi_end_request at ffffffffa4233a5c
>  #20 [ffffb732000c8e08] scsi_io_completion at ffffffffa423432c
>  #21 [ffffb732000c8e58] scsi_finish_command at ffffffffa422c527
>  #22 [ffffb732000c8e88] scsi_softirq_done at ffffffffa42341e4
> 
> [2]: https://forum.proxmox.com/threads/backup-job-failed-with-err-11-on-2-of-6-vms.92568/
> 
> Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
> ---
> 
> Mail sent upstream:
> https://lists.nongnu.org/archive/html/qemu-devel/2021-07/msg06765.html

I see that it's already reviewed (at least after the maintainers got CC'd ;)) and OK
besides some missing comments, so great work!

Can you please send this here as "patch of a patch" to be applied on pve-qemu though?
Would make it easier and commit the actual author to that git history too.

I'd put it into the `debian/patches/extra` folder there, it'll be gone again when
updating the submodule to a future QEMU 6.1 anyway, and it's not PVE specific.