[PATCH] drm/i915/display: Don't allow tile4 framebuffer to do hflip on Xe2

[Juha-Pekka Heikkila]

On 12.9.2024 20.14, Lucas De Marchi wrote:
> On Thu, Sep 12, 2024 at 05:46:06PM GMT, Juha-Pekka Heikkila wrote:
>> On Intel Xe2 hw tile4 is not supported with horizontal flip
>>
>> bspec 69853
> 
> Usual spelling is: "Bspec: 69853" and as part of the trailers below.
> 
>>
>> Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
>> ---
>> drivers/gpu/drm/i915/display/intel_fb.c            | 13 +++++++++++++
>> drivers/gpu/drm/i915/display/intel_fb.h            |  1 +
>> drivers/gpu/drm/i915/display/skl_universal_plane.c | 12 ++++++++++++
>> 3 files changed, 26 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/ 
>> drm/i915/display/intel_fb.c
>> index d2ff21e98545..c9038d239eb2 100644
>> --- a/drivers/gpu/drm/i915/display/intel_fb.c
>> +++ b/drivers/gpu/drm/i915/display/intel_fb.c
>> @@ -439,6 +439,19 @@ bool intel_fb_needs_64k_phys(u64 modifier)
>>                       INTEL_PLANE_CAP_NEED64K_PHYS);
>> }
>>
>> +/**
>> + * intel_fb_is_tile4_modifier: Check if a modifier is a tile4 
>> modifier type
>> + * @modifier: Modifier to check
>> + *
>> + * Returns:
>> + * Returns %true if @modifier is a tile4 modifier.
> 
> double Returns

This same form is on all those helpers in same file

> 
>> + */
>> +bool intel_fb_is_tile4_modifier(u64 modifier)
>> +{
>> +    return plane_caps_contain_any(lookup_modifier(modifier)->plane_caps,
>> +                      INTEL_PLANE_CAP_TILING_4);
>> +}
>> +
>> static bool check_modifier_display_ver_range(const struct 
>> intel_modifier_desc *md,
>>                          u8 display_ver_from, u8 display_ver_until)
>> {
>> diff --git a/drivers/gpu/drm/i915/display/intel_fb.h b/drivers/gpu/ 
>> drm/i915/display/intel_fb.h
>> index 068f3aee99aa..bff87994cf2c 100644
>> --- a/drivers/gpu/drm/i915/display/intel_fb.h
>> +++ b/drivers/gpu/drm/i915/display/intel_fb.h
>> @@ -35,6 +35,7 @@ bool intel_fb_is_ccs_modifier(u64 modifier);
>> bool intel_fb_is_rc_ccs_cc_modifier(u64 modifier);
>> bool intel_fb_is_mc_ccs_modifier(u64 modifier);
>> bool intel_fb_needs_64k_phys(u64 modifier);
>> +bool intel_fb_is_tile4_modifier(u64 modifier);
>>
>> bool intel_fb_is_ccs_aux_plane(const struct drm_framebuffer *fb, int 
>> color_plane);
>> int intel_fb_rc_ccs_cc_plane(const struct drm_framebuffer *fb);
>> diff --git a/drivers/gpu/drm/i915/display/skl_universal_plane.c b/ 
>> drivers/gpu/drm/i915/display/skl_universal_plane.c
>> index 17d4c880ecc4..4de41ab5060a 100644
>> --- a/drivers/gpu/drm/i915/display/skl_universal_plane.c
>> +++ b/drivers/gpu/drm/i915/display/skl_universal_plane.c
>> @@ -1591,6 +1591,18 @@ static int skl_plane_check_fb(const struct 
>> intel_crtc_state *crtc_state,
>>         return -EINVAL;
>>     }
>>
>> +    /*
>> +     * Starting with LNL and BMG tile4 hflip is not supported
>> +     */
>> +    if (rotation & DRM_MODE_REFLECT_X &&
>> +        intel_fb_is_tile4_modifier(fb->modifier) &&
>> +        ((DISPLAY_VER(dev_priv) >= 14 && IS_DGFX(dev_priv)) ||
>> +         (DISPLAY_VER(dev_priv) >= 20 && !IS_DGFX(dev_priv)))) {
> 
> the correct bspec for BMG is 50387 where the only documented restriction is
> on linear surface formats. This can rather be simplified with:
> 
>      if (rotation & DRM_MODE_REFLECT_X &&
>          intel_fb_is_tile4_modifier(fb->modifier) &&
>          DISPLAY_VER(dev_priv) >= 20)
> 

Yea, you're right. Seems I can drop bmg from here. What did look like 
similar bug was not really so.

/Juha-Pekka

>> +        drm_dbg_kms(&dev_priv->drm,
>> +                "horizontal flip is not supported with tile4 surface 
>> formats\n");
>> +        return -EINVAL;
>> +    }
>> +
>>     if (drm_rotation_90_or_270(rotation)) {
>>         if (! 
>> intel_fb_supports_90_270_rotation(to_intel_framebuffer(fb))) {
>>             drm_dbg_kms(&dev_priv->drm,
>> -- 
>> 2.45.2
>>