Re: [PATCH v1 4/4] iotests: add test for virtio-scsi and virtio-blk machine type settings

[Cleber Rosa <crosa@redhat.com>]

On Wed, Nov 06, 2019 at 04:26:41PM -0300, Eduardo Habkost wrote:
> On Wed, Nov 06, 2019 at 11:04:16AM +0100, Max Reitz wrote:
> > On 06.11.19 10:24, Stefan Hajnoczi wrote:
> > > On Tue, Nov 05, 2019 at 07:11:05PM +0300, Denis Plotnikov wrote:
> > >> It tests proper queue size settings for all available machine types.
> > >>
> > >> Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
> > >> ---
> > >>  tests/qemu-iotests/267     | 154 +++++++++++++++++++++++++++++++++++++
> > >>  tests/qemu-iotests/267.out |   1 +
> > >>  tests/qemu-iotests/group   |   1 +
> > >>  3 files changed, 156 insertions(+)
> > >>  create mode 100755 tests/qemu-iotests/267
> > >>  create mode 100644 tests/qemu-iotests/267.out
> > > 
> > > The qemu-iotests maintainers might prefer for this to be at the
> > > top-level in tests/ since it's not really an iotest, but the code itself
> > > looks fine to me:
> > > 
> > > Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
> > 
> > Good question.  I don’t really mind, but it would be weird if started
> > adding all kinds of “external” qemu tests (i.e. that use QMP) in the
> > iotests directory.
> > 
> > What is the alternative?  Just putting it in a different directory
> > doesn’t sound that appealing to me either, because it would still depend
> > on the iotests infrastructure, right?  (i.e., iotests.py and check)
> 
> We do have tests/acceptance for simple test cases written in
> Python.  What's the reason for this test case to depend on the
> iotests infrastructure?
> 
> -- 
> Eduardo

This test does look similar in spirit to "tests/acceptance/virtio_version.py".

Denis,

If you think this is more of a generic test than an IO test, and would
rather want to have it a more agnostic location, I can provide you
with tips (or a patch) to do so.

Cheers,
- Cleber.