Re: [PATCH v3 1/3] drm/i915/display: Convert intel_bw.c internally to intel_display

[Gustavo Sousa <gustavo.sousa@intel.com>]

Quoting Jani Nikula (2025-03-11 06:26:32-03:00)
>On Mon, 10 Mar 2025, Gustavo Sousa <gustavo.sousa@intel.com> wrote:
>> Quoting Ville Syrjälä (2025-03-10 13:47:57-03:00)
>>>On Fri, Mar 07, 2025 at 04:25:11PM -0300, Gustavo Sousa wrote:
>>>> Update intel_bw.c internally use intel_display. Conversion of the public
>>>> interface will come as a follow-up.
>>>> 
>>>> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
>>>> Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
>>>> ---
>>>>  drivers/gpu/drm/i915/display/intel_bw.c | 416 +++++++++++++++++---------------
>>>>  1 file changed, 217 insertions(+), 199 deletions(-)
>>>> 
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_bw.c b/drivers/gpu/drm/i915/display/intel_bw.c
>>>> index 048be287224774110d94fe2944daa580d8dc20a6..d71dc796f1793f546fe04146e5987a9be56bae9b 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_bw.c
>>>> +++ b/drivers/gpu/drm/i915/display/intel_bw.c
>>>> @@ -11,6 +11,7 @@
>>>>  #include "intel_atomic.h"
>>>>  #include "intel_bw.h"
>>>>  #include "intel_cdclk.h"
>>>> +#include "intel_de.h"
>>>>  #include "intel_display_core.h"
>>>>  #include "intel_display_types.h"
>>>>  #include "skl_watermark.h"
>>>> @@ -39,14 +40,14 @@ struct intel_qgv_info {
>>>>          u8 deinterleave;
>>>>  };
>>>>  
>>>> -static int dg1_mchbar_read_qgv_point_info(struct drm_i915_private *dev_priv,
>>>> +static int dg1_mchbar_read_qgv_point_info(struct intel_display *display,
>>>>                                            struct intel_qgv_point *sp,
>>>>                                            int point)
>>>>  {
>>>>          u32 dclk_ratio, dclk_reference;
>>>>          u32 val;
>>>>  
>>>> -        val = intel_uncore_read(&dev_priv->uncore, SA_PERF_STATUS_0_0_0_MCHBAR_PC);
>>>> +        val = intel_de_read(display, SA_PERF_STATUS_0_0_0_MCHBAR_PC);
>>>
>>>Not entirely convinced we want to start using intel_de_*() on
>>>mchbar registers.
>>>
>>>Apart from that the rest looks fine.
>>
>> Okay. I sent v4 with reverting this function to use intel_uncore_read().
>
>The question becomes, *what* are we supposed to do with this stuff in
>the future? Kind of want to get rid of all intel_uncore_* in display/.

The first thing that comes to mind is a common interface that is
provided to display by both xe and i915...

--
Gustavo Sousa