[igt-dev] [PATCH i-g-t 2/2] lib/intel_chipset: restore usage of has_flatccs flag

[Zbigniew Kempczyński]

On Mon, Jun 06, 2022 at 03:00:54PM +0300, Juha-Pekka Heikkila wrote:
> On 6.6.2022 14.03, Zbigniew Kempczyński wrote:
> > On Mon, Jun 06, 2022 at 12:05:11PM +0300, Juha-Pekka Heikkila wrote:
> > > When older patches had been merged on top of newer patches
> > > has_flatccs flag usage had gotten lost, put it back here.
> > > While at it also set has_flatccs flag to take only one bit
> > > since it is of boolean type.
> > > 
> > > CC: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
> > > Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
> > > ---
> > >   lib/intel_chipset.h | 5 ++---
> > >   1 file changed, 2 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/lib/intel_chipset.h b/lib/intel_chipset.h
> > > index 815474086..de4749e14 100644
> > > --- a/lib/intel_chipset.h
> > > +++ b/lib/intel_chipset.h
> > > @@ -41,7 +41,7 @@ struct intel_device_info {
> > >   	unsigned display_ver;
> > >   	unsigned gt; /* 0 if unknown */
> > >   	bool has_4tile : 1;
> > > -	bool has_flatccs;
> > > +	bool has_flatccs : 1;
> > >   	bool is_mobile : 1;
> > >   	bool is_whitney : 1;
> > >   	bool is_almador : 1;
> > > @@ -221,7 +221,6 @@ void intel_check_pch(void);
> > >   #define HAS_4TILE(devid)	(intel_get_device_info(devid)->has_4tile)
> > > -/* use HAS_4TILE here as all devices with 4-tile have flat ccs. */
> > > -#define HAS_FLATCCS(devid)	HAS_4TILE(devid)
> > > +#define HAS_FLATCCS(devid)	(intel_get_device_info(devid)->has_flatccs)
> > 
> > Uhm, dg1 has 4tile but this doesn't implicate it has flatccs, so:
> > 
> > Acked-by: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
> > 
> 
> Hi, thanks for the ack. About that dg1, you are not remembering some other
> platform by mistake? I didn't spot 4tile being enabled for dg1 on kernel, on
> platform definition I noticed it only enabled in DG2_FEATURES

Hmm, in block-copy-uncompressed executed:

./gem_ccs --run block-copy-uncompressed -p
md5sum *mid*

means tile4 is not exatly tile4 I expect. Format is "reversible" in sequence
linear -> tiled -> linear so maybe this deceived me. Still I don't know what
exactly this format is.

--
Zbigniew

> 
> /Juha-Pekka