Re: [igt-dev] [PATCH i-g-t] xe/xe_huc_copy: use IS_TIGERLAKE macro

[Mauro Carvalho Chehab <mauro.chehab@linux.intel.com>]

Hi Lucas,

On Tue, 4 Apr 2023 08:29:58 -0700
Lucas De Marchi <lucas.demarchi@intel.com> wrote:

> On Tue, Apr 04, 2023 at 09:52:00AM +0200, Mauro Carvalho Chehab wrote:
> >On Mon, 3 Apr 2023 23:27:01 -0700
> >Lucas De Marchi <lucas.demarchi@intel.com> wrote:
> >  
> >> On Tue, Mar 28, 2023 at 10:39:08AM +0200, Mauro Carvalho Chehab wrote:  
> >> >From: Mauro Carvalho Chehab <mchehab@kernel.org>
> >> >
> >> >Instead of hardcoding the PCI IDs at the test, use a macro
> >> >to check if the platform is compatible with the test.
> >> >
> >> >Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>  
> >>
> >> neither this or the previous solution scale well. Can we get this info
> >> from huc_info in debugfs?
> >>
> >> We are already loading HuC in platforms other than TGL  
> >
> >Changing IGT to not use IS_platform macros anymore is not an easy task,
> >as this is used on lots of place, for both i915 and Xe drivers.  
> 
> I'm not talking generically everywhere in igt. I'm talking specifically
> about huc. For i915 there is a query, for xe we don't have it. But the
> same info can be obtained from the huc_info file in debugfs.

Ah, I see. I was thinking that you wanted to replace the IS_platform
macros with something else based on HuC debugfs data.

Yeah, for this specific sense, I fully agree that we need a better
solution to detect HuC and if it has support for whatever functionality
huc-specific tests are exercising.

Regards,
Mauro