Re: [PATCH 0/8] tools/vhost: Reset virtqueue on tests

[Michael S. Tsirkin]

On Mon, Apr 13, 2020 at 04:50:06PM +0200, Eugenio Perez Martin wrote:
> On Mon, Apr 13, 2020 at 1:13 PM Michael S. Tsirkin <mst@xxxxxxxxxx> wrote:
> >
> > On Fri, Apr 03, 2020 at 06:51:11PM +0200, Eugenio Pérez wrote:
> > > This series add the tests used to validate the "vhost: Reset batched
> > > descriptors on SET_VRING_BASE call" series, with a small change on the
> > > reset code (delete an extra unneded reset on VHOST_SET_VRING_BASE).
> > >
> > > They are based on the tests sent back them, the ones that were not
> > > included (reasons in that thread). This series changes:
> > >
> > > * Delete need to export the ugly function in virtio_ring, now all the
> > > code is added in tools/virtio (except the one line fix).
> > > * Add forgotten uses of vhost_vq_set_backend. Fix bad usage order in
> > > vhost_test_set_backend.
> > > * Drop random reset, not really needed.
> > > * Minor changes updating tests code.
> > >
> > > This serie is meant to be applied on top of
> > > 5de4e0b7068337cf0d4ca48a4011746410115aae in
> > > git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git.
> >
> > Is this still needed?
> 
> ("tools/virtio: fix virtio_test.c") indentation is actually cosmetic.
> ("vhost: Not cleaning batched descs in VHOST_SET_VRING_BASE ioctl")
> just avoid to clean batches descriptors for a third time (they are
> cleaned on backend removal and addition).
> 
> ("vhost: Fix bad order in vhost_test_set_backend at enable") is
> actually a fix, the test does not work properly without it. And
> ("tools/virtio: Reset index in virtio_test --reset.") Makes the test
> work more similar than the actual VM does in a reset.
> 
> ("tools/virtio: Use __vring_new_virtqueue in virtio_test.c") and
> ("tools/virtio: Extract virtqueue initialization in vq_reset") are
> convenience commits to reach the previous two.
> 
> Lastly, ("tools/virtio: Use tools/include/list.h instead of stubs")
> just removes stub code, I did it when I try to test vdpa code and it
> seems to me a nice to have, but we can drop it from the patchset if
> you don't see that way.
> 
> > The patches lack Signed-off-by and
> > commit log descriptions, reference commit Ids without subject.
> > See Documentation/process/submitting-patches.rst
> >
> 
> Sorry, I will try to keep an eye on that from now on. I will send a v2
> with Signed-off-by and extended descriptions if you see it ok.
> 
> Thanks!

Sure, pls go ahead.

> > > Eugenio Pérez (8):
> > >   tools/virtio: fix virtio_test.c indentation
> > >   vhost: Not cleaning batched descs in VHOST_SET_VRING_BASE ioctl
> > >   vhost: Replace vq->private_data access by backend accesors
> > >   vhost: Fix bad order in vhost_test_set_backend at enable
> > >   tools/virtio: Use __vring_new_virtqueue in virtio_test.c
> > >   tools/virtio: Extract virtqueue initialization in vq_reset
> > >   tools/virtio: Reset index in virtio_test --reset.
> > >   tools/virtio: Use tools/include/list.h instead of stubs
> > >
> > >  drivers/vhost/test.c        |  8 ++---
> > >  drivers/vhost/vhost.c       |  1 -
> > >  tools/virtio/linux/kernel.h |  7 +----
> > >  tools/virtio/linux/virtio.h |  5 ++--
> > >  tools/virtio/virtio_test.c  | 58 +++++++++++++++++++++++++++----------
> > >  tools/virtio/vringh_test.c  |  2 ++
> > >  6 files changed, 51 insertions(+), 30 deletions(-)
> > >
> > > --
> > > 2.18.1
> >