Re: [PATCH v2] Revert "virtio-blk: support completion batching for the IRQ path"

[Edward Liaw]

On Wed, Jun 21, 2023 at 11:14 PM Michael S. Tsirkin <mst@xxxxxxxxxx> wrote:
> Oh interesting, can you share how to reproduce the failures?
> Then maybe Suwan Kim can take a look at fixing up the patch.

I did not have a deterministic way to reproduce the failure.  I ran
these test suites 5x consecutively and that seemed to fail often
enough to let me bisect it.

I checked out the kernel at d645c62fd19af9c9246150a958450ffd26ad0bfa
Merge 84cc6674b76b ("Merge tag 'for_linus' of
git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost") into
android-mainline, built it, and ran cuttlefish with the following
configuration:
tools/bazel run
//common-modules/virtual-device:virtual_device_x86_64_dist  # builds
the kernel with bazel
acloud create --local-instance --local-kernel-image
$ANDROID_MAINLINE/out/virtual_device_x86_64/dist --branch aosp-master
--build-target aosp_cf_x86_64_phone-userdebug  # run cuttlefish with
custom kernel

And ran the following tests:
atest --rerun-until-failure 5 vab_legacy_tests vts_core_liblp_test
vts_libdm_test vts_libsnapshot_test vts_ltp_test_x86_64

I'm not sure if that will be helpful for you, though.

On Thu, Jun 22, 2023 at 10:53 AM Michael S. Tsirkin <mst@xxxxxxxxxx> wrote:
>
> Does Android have VIRTIO_BLK_F_ZONED ? Could you list the
> features this device has please? Thanks!

It does not.

$ cat /sys/block/vda/device/driver/virtio*/features
001000100100111000000000000001001000000000000000000000000000000
VIRTIO_BLK_F_SEG_MAX
VIRTIO_BLK_F_BLK_SIZE
VIRTIO_BLK_F_FLUSH
VIRTIO_BLK_F_MQ
VIRTIO_BLK_F_DISCARD
VIRTIO_BLK_F_WRITE_ZEROES