Re: [PATCH v3 7/7] media: ipu-bridge: Remove unneeded conditional compilations

[Mauro Carvalho Chehab <mchehab+huawei@kernel.org>]

Em Wed, 11 Dec 2024 09:37:07 +0100
Ricardo Ribalda <ribalda@chromium.org> escreveu:

> On Wed, 11 Dec 2024 at 09:32, Sakari Ailus <sakari.ailus@linux.intel.com> wrote:
> >
> > Hi Mauro,
> >
> > On Wed, Dec 11, 2024 at 09:19:54AM +0100, Mauro Carvalho Chehab wrote:  
> > > > @@ -768,15 +755,10 @@ static int ipu_bridge_ivsc_is_ready(void)
> > > >     unsigned int i;
> > > >
> > > >     for (i = 0; i < ARRAY_SIZE(ipu_supported_sensors); i++) {
> > > > -#if IS_ENABLED(CONFIG_ACPI)
> > > >             const struct ipu_sensor_config *cfg =
> > > >                     &ipu_supported_sensors[i];
> > > >
> > > >             for_each_acpi_dev_match(sensor_adev, cfg->hid, NULL, -1) {
> > > > -#else
> > > > -           while (false) {
> > > > -                   sensor_adev = NULL;
> > > > -#endif
> > > >                     if (!ACPI_PTR(sensor_adev->status.enabled))
> > > >                             continue;
> > > >
> > > >  
> > >
> > > Considering that you drop patch 1, and keep the ACPI dependencies
> > > at the header, as proposed by patches 2-6:
> > >
> > > Reviewed-by: Mauro Carvalho Chehab <mchehab+huawei@kerenel.org>  
> >
> > The 1st patch fixes a compilation warning when CONFIG_ACPI is disabled.
> > Merging that patch as a temporary solution is simply easier than making
> > arragements for merging the ACPI patches to the Media tree so the last
> > patch may be merged, too.

If this is breaking compilation with W=0, then this is a different
matter: this one shall be submitted in separate, against fixes branch,
and the patch title shall be something like:

	media: ipu6: fix compilation when ACPI is disabled

And have cc stable ad fixes.

Once merged upstream, a separate patch series (without patch 1)
should be placed against the -rc kernel where the fix land.

Yet, based on the title, enforced by its description:

	> One of the quirks that we introduced to build with !ACPI && COMPILE_TEST
	> throws the following smatch warning:
	> drivers/media/pci/intel/ipu-bridge.c:752 ipu_bridge_ivsc_is_ready() warn: iterator 'i' not incremented

I don't think it makes sense to c/c stable, as this is just a smatch
warning, for a configuration that will never be used in production.


> > Besides, the fix should also be backported.
> >
> > Ricardo: how about adding Cc: stable for that one?  
> 
> Adding:
>     Cc: stable@kernel.org
>     Fixes: 8810e055b575 ("media: intel/ipu6: Fix build with !ACPI")
> to v4

Thanks,
Mauro