diff mbox series

[v3,5/5] drm/i915/dg2: extend Wa_1409120013 to DG2

Message ID 20211116174818.2128062-6-matthew.d.roper@intel.com (mailing list archive)
State New, archived
Headers show
Series i915: Additional DG2 workarounds | expand

Commit Message

Matt Roper Nov. 16, 2021, 5:48 p.m. UTC 
From: Matt Atwood <matthew.s.atwood@intel.com>

Extend existing workaround 1409120013 to DG2.

Cc: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Ville Syrjälä Nov. 17, 2021, 6:43 p.m. UTC | #1 
On Tue, Nov 16, 2021 at 09:48:18AM -0800, Matt Roper wrote:
> From: Matt Atwood <matthew.s.atwood@intel.com>
> 
> Extend existing workaround 1409120013 to DG2.

I don't see this listed for DG2.

> 
> Cc: José Roberto de Souza <jose.souza@intel.com>
> Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> ---
>  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 89dc7f69baf3..e721c421cc58 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
>  
>  static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
>  {
> -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
>  	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
>  		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
>  				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
>  
> -- 
> 2.33.0
Matt Roper Nov. 17, 2021, 6:51 p.m. UTC | #2 
On Wed, Nov 17, 2021 at 08:43:19PM +0200, Ville Syrjälä wrote:
> On Tue, Nov 16, 2021 at 09:48:18AM -0800, Matt Roper wrote:
> > From: Matt Atwood <matthew.s.atwood@intel.com>
> > 
> > Extend existing workaround 1409120013 to DG2.
> 
> I don't see this listed for DG2.

This seems to be problem with the DG2 query since for some reason they
marked this workaround as 'driver_change_required' rather than
'driver_permanent_wa' in the database and that prevents it from showing
up in some of the queries properly.  The DG2-specific ID number
to check is 1409222275.


Matt

> 
> > 
> > Cc: José Roberto de Souza <jose.souza@intel.com>
> > Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> > index 89dc7f69baf3..e721c421cc58 100644
> > --- a/drivers/gpu/drm/i915/intel_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
> >  
> >  static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
> >  {
> > -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> > +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
> >  	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> > -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> > +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
> >  		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
> >  				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
> >  
> > -- 
> > 2.33.0
> 
> -- 
> Ville Syrjälä
> Intel
Ville Syrjälä Nov. 17, 2021, 6:54 p.m. UTC | #3 
On Wed, Nov 17, 2021 at 08:43:19PM +0200, Ville Syrjälä wrote:
> On Tue, Nov 16, 2021 at 09:48:18AM -0800, Matt Roper wrote:
> > From: Matt Atwood <matthew.s.atwood@intel.com>
> > 
> > Extend existing workaround 1409120013 to DG2.
> 
> I don't see this listed for DG2.
> 
> > 
> > Cc: José Roberto de Souza <jose.souza@intel.com>
> > Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> > index 89dc7f69baf3..e721c421cc58 100644
> > --- a/drivers/gpu/drm/i915/intel_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
> >  
> >  static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
> >  {
> > -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> > +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
> >  	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> > -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> > +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))

AFAIK we're not even calling this function on dg2, so this is just dead
code. And in fact without dg2 this seems to be the same as DISPLAY_VER==12
so we shuld stop calling it on adl-p as well. We could then rip out most
of the platform checks in here.

> >  		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
> >  				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
> >  
> > -- 
> > 2.33.0
> 
> -- 
> Ville Syrjälä
> Intel
Ville Syrjälä Nov. 17, 2021, 7:01 p.m. UTC | #4 
On Wed, Nov 17, 2021 at 10:51:39AM -0800, Matt Roper wrote:
> On Wed, Nov 17, 2021 at 08:43:19PM +0200, Ville Syrjälä wrote:
> > On Tue, Nov 16, 2021 at 09:48:18AM -0800, Matt Roper wrote:
> > > From: Matt Atwood <matthew.s.atwood@intel.com>
> > > 
> > > Extend existing workaround 1409120013 to DG2.
> > 
> > I don't see this listed for DG2.
> 
> This seems to be problem with the DG2 query since for some reason they
> marked this workaround as 'driver_change_required' rather than
> 'driver_permanent_wa' in the database and that prevents it from showing
> up in some of the queries properly.  The DG2-specific ID number
> to check is 1409222275.

Bit of mes that one. I can't really figure out if dg2 is the only
d13 platform that needs this or might there be others?
Souza, Jose Nov. 19, 2021, 4:36 p.m. UTC | #5 
On Tue, 2021-11-16 at 09:48 -0800, Matt Roper wrote:
> From: Matt Atwood <matthew.s.atwood@intel.com>
> 
> Extend existing workaround 1409120013 to DG2.
> 
> Cc: José Roberto de Souza <jose.souza@intel.com>
> Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> ---
>  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 89dc7f69baf3..e721c421cc58 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
>  
>  static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
>  {
> -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */

I'm not finding this workaround in the DG2 WA spec page, maybe it was removed because it is not necessary anymore?

>  	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
>  		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
>  				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
>
Matt Roper Nov. 19, 2021, 5:11 p.m. UTC | #6 
On Fri, Nov 19, 2021 at 08:36:56AM -0800, Souza, Jose wrote:
> On Tue, 2021-11-16 at 09:48 -0800, Matt Roper wrote:
> > From: Matt Atwood <matthew.s.atwood@intel.com>
> > 
> > Extend existing workaround 1409120013 to DG2.
> > 
> > Cc: José Roberto de Souza <jose.souza@intel.com>
> > Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> > index 89dc7f69baf3..e721c421cc58 100644
> > --- a/drivers/gpu/drm/i915/intel_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_pm.c
> > @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
> >  
> >  static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
> >  {
> > -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> > +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
> 
> I'm not finding this workaround in the DG2 WA spec page, maybe it was removed because it is not necessary anymore?

Ville raised the same question; I believe this is just an issue with the
query that generates the bspec page from the database; here's my earlier
response:

        >> This seems to be problem with the DG2 query since for some
        >> reason they marked this workaround as
        >> 'driver_change_required' rather than 'driver_permanent_wa' in
        >> the database and that prevents it from showing up in some of
        >> the queries properly.  The DG2-specific ID number to check is
        >> 1409222275.

Since it appears this is currently needed on every version 12 and
version 13 display platform _except_ for ADL-P, I did send a question to
the HW team to confirm that the lack of ADL-P isn't an oversight, but I
haven't heard back yet.


Matt

> 
> >  	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> > -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> > +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
> >  		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
> >  				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
> >  
>
Clint Taylor Dec. 2, 2021, 11 p.m. UTC | #7 
Reviewed-by: Clint Taylor <Clinton.A.Taylor@intel.com>

-Clint

On 11/16/21 9:48 AM, Matt Roper wrote:
> From: Matt Atwood <matthew.s.atwood@intel.com>
>
> Extend existing workaround 1409120013 to DG2.
>
> Cc: José Roberto de Souza <jose.souza@intel.com>
> Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> ---
>   drivers/gpu/drm/i915/intel_pm.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 89dc7f69baf3..e721c421cc58 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -7444,9 +7444,9 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
>   
>   static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
>   {
> -	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
> +	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
>   	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
> -	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
> +	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
>   		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
>   				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
>
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 89dc7f69baf3..e721c421cc58 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -7444,9 +7444,9 @@  static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
 
 static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
 {
-	/* Wa_1409120013:tgl,rkl,adl-s,dg1 */
+	/* Wa_1409120013:tgl,rkl,adl-s,dg1,dg2 */
 	if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv) ||
-	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv))
+	    IS_ALDERLAKE_S(dev_priv) || IS_DG1(dev_priv) || IS_DG2(dev_priv))
 		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN,
 				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);