[Mesa-dev] [PATCH 02/12] i965/cnl: Add gen10.xml

[Anuj Phogat]

On Sun, Apr 16, 2017 at 8:32 PM, Jason Ekstrand <jason at jlekstrand.net> wrote:
> On Fri, Apr 14, 2017 at 5:35 PM, Anuj Phogat <anuj.phogat at gmail.com> wrote:
>>
>> From: Jason Ekstrand <jason.ekstrand at intel.com>
>>
>> Signed-off-by: Anuj Phogat <anuj.phogat at gmail.com>
>> ---
>>  src/intel/Makefile.sources |    3 +-
>>  src/intel/genxml/gen10.xml | 3557
>> ++++++++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 3559 insertions(+), 1 deletion(-)
>>  create mode 100644 src/intel/genxml/gen10.xml
>>
>> diff --git a/src/intel/Makefile.sources b/src/intel/Makefile.sources
>> index c568916..57eb157 100644
>> --- a/src/intel/Makefile.sources
>> +++ b/src/intel/Makefile.sources
>> @@ -117,7 +117,8 @@ GENXML_XML_FILES = \
>>         genxml/gen7.xml \
>>         genxml/gen75.xml \
>>         genxml/gen8.xml \
>> -       genxml/gen9.xml
>> +       genxml/gen9.xml \
>> +       genxml/gen10.xml
>
>
> I don't see any patches in this series that update genX_macros.h,
> genX_pack.h, isl, blorp, anv or anything else for that matter to actually
> use the gen10 XML.  How is isl_surf_fill_state_s, for instance, not failing?
As I mentioned in the cover letter we are still using gen9 paths for isl, blorp,
anv for cannonlake. We've glxgears and many piglit tests rendering correctly
with these patches. Using isl_gen9_surf_fill_state_s is not causing issues
possibly because of little changes in surface state from gen9 to gen10.
I have patches rebased on master that replace the gen9 paths and to update
genX_macros.h, genX_pack.h. But, things are not rendering as expected.
So, I sent this series out to start the review process while I work on the
rendering issues with cnl patches on master.