[PATCH 2/2] device_reset: Require a PCI device

[Rob Clark]

On Mon, Aug 19, 2024 at 10:37 AM Kamil Konieczny
<kamil.konieczny at linux.intel.com> wrote:
>
> Hi Rob,
> On 2024-08-14 at 11:54:26 -0700, Rob Clark wrote:
> > From: Rob Clark <robdclark at chromium.org>
>
> One nit about subject, please use 'tests/' prefix:
>
> [PATCH 2/2] device_reset: Require a PCI device
>
> should be:
>
> [PATCH 2/2] tests/device_reset: Require a PCI device
>
> >
> > Too much baked in assumptions that the drm device is a PCI device.
>
> Do you have examples where it is not true (e.g. not a PCI device)
> and reset is possible? Just curious.

maybe usb devices?

Platform devices cannot be reset at the bus level, since there is no bus.

BR,
-R

> Regards,
> Kamil
>
> >
> > Signed-off-by: Rob Clark <robdclark at chromium.org>
> > ---
> >  tests/device_reset.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/tests/device_reset.c b/tests/device_reset.c
> > index 8ed3ef5220fc..c49f94e2faed 100644
> > --- a/tests/device_reset.c
> > +++ b/tests/device_reset.c
> > @@ -200,6 +200,7 @@ static void init_device_fds(struct device_fds *dev)
> >               igt_require_gem(dev->fds.dev);
> >
> >       igt_assert(device_sysfs_path(dev->fds.dev, dev_path));
> > +     igt_require(strstr(dev_path, "/sys/devices/pci") == dev_path);
> >       addr_pos = strrchr(dev_path, '/');
> >       igt_assert(addr_pos);
> >       igt_assert_eq(sizeof(dev->dev_bus_addr) - 1,
> > --
> > 2.46.0
> >