Re: [PATCH v3 1/5] drm: Introduce plane and CRTC scaling filter properties

On Tue, Apr 07, 2020 at 08:28:02PM +0300, Ville Syrjälä wrote:

On Tue, Mar 31, 2020 at 12:08:53AM +0530, Pankaj Bharadiya wrote:

...

Introduce per-plane and per-CRTC scaling filter properties to allow userspace to select the driver's default scaling filter or Nearest-neighbor(NN) filter for upscaling operations on CRTC and plane.

Drivers can set up this property for a plane by calling drm_plane_create_scaling_filter() and for a CRTC by calling drm_crtc_create_scaling_filter().

NN filter works by filling in the missing color values in the upscaled image with that of the coordinate-mapped nearest source pixel value.

NN filter for integer multiple scaling can be particularly useful for for pixel art games that rely on sharp, blocky images to deliver their distinctive look.

changes since v2:

• Create per-plane and per-CRTC scaling filter property (Ville)

changes since v1:

• None

changes since RFC:

• Add separate properties for plane and CRTC (Ville)

Signed-off-by: Pankaj Bharadiya pankaj.laxminarayan.bharadiya@intel.com

drivers/gpu/drm/drm_atomic_uapi.c | 8 ++++ drivers/gpu/drm/drm_crtc.c | 78 +++++++++++++++++++++++++++++++ drivers/gpu/drm/drm_plane.c | 78 +++++++++++++++++++++++++++++++ include/drm/drm_crtc.h | 16 +++++++ include/drm/drm_plane.h | 21 +++++++++ 5 files changed, 201 insertions(+)

diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c index a1e5e262bae2..ac7dabbf0bcf 100644 --- a/drivers/gpu/drm/drm_atomic_uapi.c +++ b/drivers/gpu/drm/drm_atomic_uapi.c @@ -469,6 +469,8 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc, return -EFAULT;

set_out_fence_for_crtc(state->state, crtc, fence_ptr);

• } else if (property == crtc->scaling_filter_property) {

• state->scaling_filter = val;
  

  } else if (crtc->funcs->atomic_set_property) { return crtc->funcs->atomic_set_property(crtc, state, property, val); } else {

@@ -503,6 +505,8 @@ drm_atomic_crtc_get_property(struct drm_crtc *crtc, *val = (state->gamma_lut) ? state->gamma_lut->base.id : 0; else if (property == config->prop_out_fence_ptr) *val = 0;

• else if (property == crtc->scaling_filter_property)

Random side observation: Why do we have two different styles to naming these things (prop_foo vs. foo_property)? Would be nice to unify this one way or the other.

Need to handle this separately.

All per-plane props follow foo_property convention and we have mixed conventions for properties in struct drm_mode_config with majority being foo_property.

...

• *val = state->scaling_filter;
  

  else if (crtc->funcs->atomic_get_property) return crtc->funcs->atomic_get_property(crtc, state, property, val); else

@@ -583,6 +587,8 @@ static int drm_atomic_plane_set_property(struct drm_plane *plane, sizeof(struct drm_rect), &replaced); return ret;

• } else if (property == plane->scaling_filter_property) {

• state->scaling_filter = val;
  

  } else if (plane->funcs->atomic_set_property) { return plane->funcs->atomic_set_property(plane, state, property, val);

@@ -641,6 +647,8 @@ drm_atomic_plane_get_property(struct drm_plane *plane, } else if (property == config->prop_fb_damage_clips) { *val = (state->fb_damage_clips) ? state->fb_damage_clips->base.id : 0;

• } else if (property == plane->scaling_filter_property) {

• *val = state->scaling_filter;
  

  } else if (plane->funcs->atomic_get_property) { return plane->funcs->atomic_get_property(plane, state, property, val); } else {

diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 4936e1080e41..95502c88966b 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -748,3 +748,81 @@ int drm_mode_crtc_set_obj_prop(struct drm_mode_object *obj,

return ret; }

+/**

• • DOC: CRTC scaling filter property
• • SCALING_FILTER:
• • Indicates scaling filter to be used for CRTC scaler
• • The value of this property can be one of the following:
• • Default:

• • Driver's default scaling filter
    

• • Nearest Neighbor:

• • Nearest Neighbor scaling filter
    

• • Drivers can set up this property for a CRTC by calling
• • drm_crtc_create_scaling_filter_property
• */

+/**

• • drm_crtc_create_scaling_filter_property - create a new scaling filter
• • property
• • @crtc: drm CRTC
• • @supported_filters: bitmask of supported scaling filters, must include

• •        BIT(DRM_SCALING_FILTER_DEFAULT).
    

• • This function lets driver to enable the scaling filter property on a given
• • CRTC.
• • RETURNS:
• • Zero for success or -errno
• */

+int drm_crtc_create_scaling_filter_property(struct drm_crtc *crtc,

• 			    unsigned int supported_filters)
  

+{

• struct drm_device *dev = crtc->dev;
• struct drm_property *prop;
• static const struct drm_prop_enum_list props[] = {

• { DRM_SCALING_FILTER_DEFAULT, "Default" },
  

• { DRM_SCALING_FILTER_NEAREST_NEIGHBOR, "Nearest Neighbor" },
  

• };
• unsigned int valid_mode_mask = BIT(DRM_SCALING_FILTER_DEFAULT) |

• 		       BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR);
  

• int i;
• if (WARN_ON((supported_filters & ~valid_mode_mask) ||

•     ((supported_filters & BIT(DRM_SCALING_FILTER_DEFAULT)) == 0)))
  

• return -EINVAL;
  

• prop = drm_property_create(dev, DRM_MODE_PROP_ENUM,

• 		   "SCALING_FILTER",
  

• 		   hweight32(supported_filters));
  

• if (!prop)

• return -ENOMEM;
  

• for (i = 0; i < ARRAY_SIZE(props); i++) {

• int ret;
  

• if (!(BIT(props[i].type) & supported_filters))
  

• 	continue;
  

• ret = drm_property_add_enum(prop, props[i].type,
  

• 			    props[i].name);
  

• if (ret) {
  

• 	drm_property_destroy(dev, prop);
  

• 	return ret;
  

• }
  

• }
• drm_object_attach_property(&crtc->base, prop,

• 		   DRM_SCALING_FILTER_DEFAULT);
  

Everything up to here is identical between the crtc and plane. Needs a refactoring. In fact this whole thing seems pretty generic.

How about spliting code like below -

diff --git a/drivers/gpu/drm/drm_crtc_internal.h b/drivers/gpu/drm/drm_crtc_internal.h --- a/drivers/gpu/drm/drm_crtc_internal.h +++ b/drivers/gpu/drm/drm_crtc_internal.h @@ -72,6 +72,9 @@ int drm_crtc_force_disable(struct drm_crtc *crtc);

struct dma_fence *drm_crtc_create_fence(struct drm_crtc *crtc);

+struct drm_property * +drm_prepare_scaling_filter_prop(struct drm_device *dev, + unsigned int supported_filters); /* IOCTLs */ int drm_mode_getcrtc(struct drm_device *dev, void *data, struct drm_file *file_priv);

diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c index d6ad60ab0d38..e63614fe3eed 100644 --- a/drivers/gpu/drm/drm_plane.c +++ b/drivers/gpu/drm/drm_plane.c @@ -1221,3 +1221,93 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev,

return ret; } + +struct drm_property * +drm_prepare_scaling_filter_prop(struct drm_device *dev, + unsigned int supported_filters) +{ + struct drm_property *prop; + static const struct drm_prop_enum_list props[] = { + { DRM_SCALING_FILTER_DEFAULT, "Default" }, + { DRM_SCALING_FILTER_NEAREST_NEIGHBOR, "Nearest Neighbor" }, + }; + unsigned int valid_mode_mask = BIT(DRM_SCALING_FILTER_DEFAULT) | + BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR); + int i; + + if (WARN_ON((supported_filters & ~valid_mode_mask) || + ((supported_filters & BIT(DRM_SCALING_FILTER_DEFAULT)) == 0))) + return ERR_PTR(-EINVAL); + + prop = drm_property_create(dev, DRM_MODE_PROP_ENUM, + "SCALING_FILTER", + hweight32(supported_filters)); + if (!prop) + return ERR_PTR(-ENOMEM); + + for (i = 0; i < ARRAY_SIZE(props); i++) { + int ret; + + if (!(BIT(props[i].type) & supported_filters)) + continue; + + ret = drm_property_add_enum(prop, props[i].type, + props[i].name); + + if (ret) { + drm_property_destroy(dev, prop); + + return ERR_PTR(ret); + } + } + + return prop; +} + +/** + * drm_plane_create_scaling_filter_property - create a new scaling filter + * property + * + * @plane: drm plane + * @supported_filters: bitmask of supported scaling filters, must include + * BIT(DRM_SCALING_FILTER_DEFAULT). + * + * This function lets driver to enable the scaling filter property on a given + * plane. + * + * RETURNS: + * Zero for success or -errno + */ +int drm_plane_create_scaling_filter_property(struct drm_plane *plane, + unsigned int supported_filters) +{ + struct drm_property *prop = + drm_prepare_scaling_filter_prop(plane->dev, supported_filters); + + if (IS_ERR(prop)) + return PTR_ERR(prop); + + drm_object_attach_property(&plane->base, prop, + DRM_SCALING_FILTER_DEFAULT); + plane->scaling_filter_property = prop; + + return 0; +} +EXPORT_SYMBOL(drm_plane_create_scaling_filter_property);

index 4936e1080e41..b48e0bce8f60 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c

+int drm_crtc_create_scaling_filter_property(struct drm_crtc *crtc, + unsigned int supported_filters) +{ + struct drm_property *prop = + drm_prepare_scaling_filter_prop(crtc->dev, supported_filters); + + if (IS_ERR(prop)) + return PTR_ERR(prop); + + drm_object_attach_property(&crtc->base, prop, + DRM_SCALING_FILTER_DEFAULT); + crtc->scaling_filter_property = prop; + + return 0; +} +EXPORT_SYMBOL(drm_crtc_create_scaling_filter_property);

Should probably think about just adding that bitmask to drm_property_create_enum(). I suppose we could try to avoid having to change all the existing callers by keeping the current thing without the bitmask (though it could probably internally just call the version which takes the bitmask, assuming our enum values aren't too big for that.

As more filters can be added in future and different hardwares can have different capabilities, I think it make sense to provide a bitmask to the callers so that drivers can expose *only* filters which they support.

What do you think?

Thanks, Pankaj

Otherwise the patch seems reasonable.

...

• crtc->scaling_filter_property = prop;
• return 0;

+} +EXPORT_SYMBOL(drm_crtc_create_scaling_filter_property); diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c

[snip]

			state->fb_damage_clips->data : NULL);

...

}

+int drm_plane_create_scaling_filter_property(struct drm_plane *plane,

• 			     unsigned int supported_filters);
  

#endif

2.23.0

-- Ville Syrjälä Intel