Merge ~sergiodj/ubuntu/+source/qemu:bug1970737-stale-io-sysbench-jammy into ubuntu/+source/qemu:ubuntu/jammy-devel

[Sergio Durigan Junior]

This MP fixes bug #1970737, which reports an stalled I/O scenario on qemu when using an NMVe storage device.

This bug has been discussed and fixed upstream; this is the backport of the patches to Jammy. Note that qemu on Kinetic hasn't been fixed yet, because there are some security fixes that need to be uploaded there first.

I've discussed this with Christian and he's OK with the changes. There's a PPA with the proposed package here:

https://launchpad.net/~sergiodj/+archive/ubuntu/qemu-bug1970737/+packages

qemu doesn't have autopkgtests, but Christian runs his own internal tests.

[Christian Ehrhardt  (paelzer)]

LGTM just like what I've seen before.
Qemu has a few rough edges, but you are touching none of them - so that should really be fine.

I've pinged mdeslaur as there seems to be non security upload for Kinetic.
Depending on the answer you can upload both soon.

P.S. yes I'll run the tests once it is in jammy-proposed.

review: Approve

[git-ubuntu bot (git-ubuntu-bot)]

Approvers: sergiodj, paelzer
Uploaders: sergiodj, paelzer
MP auto-approved

review: Approve

[Sergio Durigan Junior (sergiodj)]

On Wednesday, June 22 2022, Christian Ehrhardt  wrote:

> LGTM just like what I've seen before.
> Qemu has a few rough edges, but you are touching none of them - so that should really be fine.

Thanks, Christian.

> I've pinged mdeslaur as there seems to be non security upload for Kinetic.
> Depending on the answer you can upload both soon.

Yeah, I've seen the ping, thanks. I'm now waiting for mdeslaur to send
me the changes. Once he does, I will file the MP for Kinetic.

Cheers,

--
Sergio
GPG key ID: E92F D0B3 6B14 F1F4 D8E0 EB2F 106D A1C8 C3CB BF14

[Sergio Durigan Junior (sergiodj)]

On Wednesday, June 22 2022, Sergio Durigan Junior wrote:

> On Wednesday, June 22 2022, Christian Ehrhardt  wrote:
>> I've pinged mdeslaur as there seems to be non security upload for Kinetic.
>> Depending on the answer you can upload both soon.
>
> Yeah, I've seen the ping, thanks. I'm now waiting for mdeslaur to send
> me the changes. Once he does, I will file the MP for Kinetic.

I've prepared and uploaded the changes to Kinetic. Therefore, this has
also been uploaded:

$ dput qemu_6.2+dfsg-2ubuntu6.3_source.changes
Trying to upload package to ubuntu
Checking signature on .changes
gpg: /home/sergio/work/qemu/qemu_6.2+dfsg-2ubuntu6.3_source.changes: Valid signature from 106DA1C8C3CBBF14
Checking signature on .dsc
gpg: /home/sergio/work/qemu/qemu_6.2+dfsg-2ubuntu6.3.dsc: Valid signature from 106DA1C8C3CBBF14
Uploading to ubuntu (via ftp to upload.ubuntu.com):
  Uploading qemu_6.2+dfsg-2ubuntu6.3.dsc: done.
  Uploading qemu_6.2+dfsg-2ubuntu6.3.debian.tar.xz: done.
  Uploading qemu_6.2+dfsg-2ubuntu6.3_source.buildinfo: done.
  Uploading qemu_6.2+dfsg-2ubuntu6.3_source.changes: done.
Successfully uploaded packages.

--
Sergio
GPG key ID: E92F D0B3 6B14 F1F4 D8E0 EB2F 106D A1C8 C3CB BF14

diff --git a/debian/changelog b/debian/changelog
index 98f991c..d765eb3 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,10 @@
+qemu (1:6.2+dfsg-2ubuntu6.3) jammy; urgency=medium
+
+  * Fix unbalanced plugged counter in laio_io_unplug (LP: #1970737)
+    - d/p/lp1970737-linux-aio-*.patch: Upstream patches.
+
+ -- Sergio Durigan Junior <[e-mail]>  Tue, 21 Jun 2022 17:07:50 -0400
+
 qemu (1:6.2+dfsg-2ubuntu6.2) jammy-security; urgency=medium
 
   * SECURITY UPDATE: heap overflow in floppy disk emulator
diff --git a/debian/patches/series b/debian/patches/series
index 8d71725..a6da6ee 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -37,3 +37,5 @@ CVE-2021-4207.patch
 CVE-2022-0358.patch
 CVE-2022-26353.patch
 CVE-2022-26354.patch
+ubuntu/lp1970737-linux-aio-fix-unbalanced-plugged-counter-in-laio_io_.patch
+ubuntu/lp1970737-linux-aio-explain-why-max-batch-is-checked-in-laio_i.patch
diff --git a/debian/patches/ubuntu/lp1970737-linux-aio-explain-why-max-batch-is-checked-in-laio_i.patch b/debian/patches/ubuntu/lp1970737-linux-aio-explain-why-max-batch-is-checked-in-laio_i.patch
new file mode 100644
index 0000000..187435c
--- /dev/null
+++ b/debian/patches/ubuntu/lp1970737-linux-aio-explain-why-max-batch-is-checked-in-laio_i.patch
@@ -0,0 +1,37 @@
+From: Stefan Hajnoczi <[e-mail]>
+Date: Thu, 9 Jun 2022 17:47:12 +0100
+Subject: linux-aio: explain why max batch is checked in laio_io_unplug()
+
+It may not be obvious why laio_io_unplug() checks max batch. I discussed
+this with Stefano and have added a comment summarizing the reason.
+
+Cc: Stefano Garzarella <[e-mail]>
+Cc: Kevin Wolf <[e-mail]>
+Signed-off-by: Stefan Hajnoczi <[e-mail]>
+Reviewed-by: Stefano Garzarella <[e-mail]>
+Message-id: [e-mail]
+Signed-off-by: Stefan Hajnoczi <[e-mail]>
+
+Origin: upstream, https://gitlab.com/qemu-project/qemu/-/commit/99b969fbe105117f5af6060d3afef40ca39cc9c1
+Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/1970737
+---
+ block/linux-aio.c | 6 ++++++
+ 1 file changed, 6 insertions(+)
+
+diff --git a/block/linux-aio.c b/block/linux-aio.c
+index 77f17ad..85650c4 100644
+--- a/block/linux-aio.c
++++ b/block/linux-aio.c
+@@ -362,6 +362,12 @@ void laio_io_unplug(BlockDriverState *bs, LinuxAioState *s,
+     assert(s->io_q.plugged);
+     s->io_q.plugged--;
+ 
++    /*
++     * Why max batch checking is performed here:
++     * Another BDS may have queued requests with a higher dev_max_batch and
++     * therefore in_queue could now exceed our dev_max_batch. Re-check the max
++     * batch so we can honor our device's dev_max_batch.
++     */
+     if (s->io_q.in_queue >= laio_max_batch(s, dev_max_batch) ||
+         (!s->io_q.plugged &&
+          !s->io_q.blocked && !QSIMPLEQ_EMPTY(&s->io_q.pending))) {
diff --git a/debian/patches/ubuntu/lp1970737-linux-aio-fix-unbalanced-plugged-counter-in-laio_io_.patch b/debian/patches/ubuntu/lp1970737-linux-aio-fix-unbalanced-plugged-counter-in-laio_io_.patch
new file mode 100644
index 0000000..c555b4d
--- /dev/null
+++ b/debian/patches/ubuntu/lp1970737-linux-aio-fix-unbalanced-plugged-counter-in-laio_io_.patch
@@ -0,0 +1,44 @@
+From: Stefan Hajnoczi <[e-mail]>
+Date: Thu, 9 Jun 2022 17:47:11 +0100
+Subject: linux-aio: fix unbalanced plugged counter in laio_io_unplug()
+
+Every laio_io_plug() call has a matching laio_io_unplug() call. There is
+a plugged counter that tracks the number of levels of plugging and
+allows for nesting.
+
+The plugged counter must reflect the balance between laio_io_plug() and
+laio_io_unplug() calls accurately. Otherwise I/O stalls occur since
+io_submit(2) calls are skipped while plugged.
+
+Reported-by: Nikolay Tenev <[e-mail]>
+Signed-off-by: Stefan Hajnoczi <[e-mail]>
+Reviewed-by: Stefano Garzarella <[e-mail]>
+Message-id: [e-mail]
+Cc: Stefano Garzarella <[e-mail]>
+Fixes: 68d7946648 ("linux-aio: add `dev_max_batch` parameter to laio_io_unplug()")
+[Stefano Garzarella suggested adding a Fixes tag.
+--Stefan]
+Signed-off-by: Stefan Hajnoczi <[e-mail]>
+
+Origin: upstream, https://gitlab.com/qemu-project/qemu/-/commit/f387cac5af030a58ac5a0dacf64cab5e5a4fe5c7
+Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/qemu/+bug/1970737
+---
+ block/linux-aio.c | 4 +++-
+ 1 file changed, 3 insertions(+), 1 deletion(-)
+
+diff --git a/block/linux-aio.c b/block/linux-aio.c
+index f53ae72..77f17ad 100644
+--- a/block/linux-aio.c
++++ b/block/linux-aio.c
+@@ -360,8 +360,10 @@ void laio_io_unplug(BlockDriverState *bs, LinuxAioState *s,
+                     uint64_t dev_max_batch)
+ {
+     assert(s->io_q.plugged);
++    s->io_q.plugged--;
++
+     if (s->io_q.in_queue >= laio_max_batch(s, dev_max_batch) ||
+-        (--s->io_q.plugged == 0 &&
++        (!s->io_q.plugged &&
+          !s->io_q.blocked && !QSIMPLEQ_EMPTY(&s->io_q.pending))) {
+         ioq_submit(s);
+     }