Support 64 bpp half float formats
Subsystems (3)
cover Support 64 bpp half float formats
This series defines new formats and adds a plane property to be used for floating point framebuffer content. Implementation is then added to i915.
I have shared an IGT branch which adds test coverage for the new formats:
https://github.com/strassek/xorg-intel-gpu-tools/tree/fp16
Kevin Strasser (3):
drm/fourcc: Add 64 bpp half float formats
drm: Add optional PIXEL_NORMALIZE_RANGE property to drm_plane drm/i915: Implement half float formats and pixel normalize property
drivers/gpu/drm/drm_atomic.c | 2 + drivers/gpu/drm/drm_atomic_uapi.c | 4 ++ drivers/gpu/drm/drm_color_mgmt.c | 67 +++++++++++++++++++++++ drivers/gpu/drm/drm_crtc_internal.h | 1 + drivers/gpu/drm/drm_fourcc.c | 4 ++ drivers/gpu/drm/i915/i915_reg.h | 15 ++++- drivers/gpu/drm/i915/intel_display.c | 47 ++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 5 ++ drivers/gpu/drm/i915/intel_sprite.c | 82 ++++++++++++++++++++++++++-- include/drm/drm_color_mgmt.h | 9 +++ include/drm/drm_fourcc.h | 3 + include/drm/drm_plane.h | 14 +++++ include/uapi/drm/drm_fourcc.h | 6 ++ 13 files changed, 252 insertions(+), 7 deletions(-)
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 1/3 drm/fourcc: Add 64 bpp half float formats
Add 64 bpp 16:16:16:16 half float pixel formats. Each 16 bit component is formatted in IEEE-754 half-precision float (binary16) 1:5:10 MSb-sign:exponent:fraction form.
An 'is_fp' attribute is added to drm_format_info so that drivers can easily distinguish these formats from those that might contain uint pixel data.
This patch attempts to address the feedback provided when 2 of these formats were previosly proposed:
https://patchwork.kernel.org/patch/10072545/
--- drivers/gpu/drm/drm_fourcc.c | 4 ++++ include/drm/drm_fourcc.h | 3 +++ include/uapi/drm/drm_fourcc.h | 6 ++++++ 3 files changed, 13 insertions(+)
drivers/gpu/drm/drm_fourcc.c +4 -0
@@ -198,6 +198,10 @@ const struct drm_format_info *__drm_format_info(u32 format) { .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_RGBA8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_BGRA8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, + { .format = DRM_FORMAT_XRGB16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_XBGR16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_ARGB16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true }, + { .format = DRM_FORMAT_ABGR16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true }, { .format = DRM_FORMAT_RGB888_A8, .depth = 32, .num_planes = 2, .cpp = { 3, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_BGR888_A8, .depth = 32, .num_planes = 2, .cpp = { 3, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_XRGB8888_A8, .depth = 32, .num_planes = 2, .cpp = { 4, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true },
include/drm/drm_fourcc.h +3 -0
@@ -133,6 +133,9 @@ struct drm_format_info {/** @is_yuv: Is it a YUV format? */ bool is_yuv; + +/** @is_fp: Is it a floating point format? */ +bool is_fp; };/**
include/uapi/drm/drm_fourcc.h +6 -0
@@ -144,6 +144,12 @@ extern "C" {#define DRM_FORMAT_RGBA1010102 fourcc_code('R', 'A', '3', '0')/* [31:0] R:G:B:A 10:10:10:2 little endian */ #define DRM_FORMAT_BGRA1010102 fourcc_code('B', 'A', '3', '0')/* [31:0] B:G:R:A 10:10:10:2 little endian */ +/* 64 bpp RGB IEEE-754 half-precision float (binary16) */ +#define DRM_FORMAT_XBGR16161616H fourcc_code('X', 'B', '4', 'H')/* [63:0] x:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_ABGR16161616H fourcc_code('A', 'B', '4', 'H')/* [63:0] A:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_XRGB16161616H fourcc_code('X', 'R', '4', 'H')/* [63:0] x:R:G:B 16:16:16:16 little endian */ +#define DRM_FORMAT_ARGB16161616H fourcc_code('A', 'R', '4', 'H')/* [63:0] A:R:G:B 16:16:16:16 little endian */ +/* packed YCbCr */ #define DRM_FORMAT_YUYV fourcc_code('Y', 'U', 'Y', 'V')/* [31:0] Cr0:Y1:Cb0:Y0 8:8:8:8 little endian */ #define DRM_FORMAT_YVYU fourcc_code('Y', 'V', 'Y', 'U')/* [31:0] Cb0:Y1:Cr0:Y0 8:8:8:8 little endian */
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 1/3] drm/fourcc: Add 64 bpp half float formats
Quoted 30 lines · Kevin Strasser
Add 64 bpp 16:16:16:16 half float pixel formats. Each 16 bit component is formatted in IEEE-754 half-precision float (binary16) 1:5:10 MSb-sign:exponent:fraction form.
An '
is_fp' attribute is added todrm_format_infoso that drivers can easily distinguish these formats from those that might contain uint pixel data.This patch attempts to address the feedback provided when 2 of these formats were previosly proposed:
https://patchwork.kernel.org/patch/10072545/Signed-off-by Kevin Strasser <[email protected]>Cc Tina Zhang <[email protected]>Cc Uma Shankar <[email protected]>Cc Shashank Sharma <[email protected]>Cc Ville Syrjälä <[email protected]>Cc David Airlie <[email protected]>Cc Daniel Vetter <[email protected]>--- drivers/gpu/drm/drm_fourcc.c | 4 ++++ include/drm/drm_fourcc.h | 3 +++ include/uapi/drm/drm_fourcc.h | 6 ++++++ 3 files changed, 13 insertions(+)drivers/gpu/drm/drm_fourcc.c +0
-0
drivers/gpu/drm/drm_fourcc.c const struct drm_format_info *__drm_format_info(u32 format) · line 204
Quoted 7 lines
drivers/gpu/drm/drm_fourcc.c +4
-0@@ -198,6 +198,10 @@ const struct drm_format_info *__drm_format_info(u32 format) { .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_RGBA8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_BGRA8888, .depth = 32, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, + { .format = DRM_FORMAT_XRGB16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_XBGR16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_ARGB16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true }, + { .format = DRM_FORMAT_ABGR16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true },
cpp is wrong for all of these.
Quoted 22 lines
drivers/gpu/drm/drm_fourcc.c +0
-0{ .format = DRM_FORMAT_RGB888_A8, .depth = 32, .num_planes = 2, .cpp = { 3, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_BGR888_A8, .depth = 32, .num_planes = 2, .cpp = { 3, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true }, { .format = DRM_FORMAT_XRGB8888_A8, .depth = 32, .num_planes = 2, .cpp = { 4, 1, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true },include/drm/drm_fourcc.h +3
-0@@ -133,6 +133,9 @@ struct drm_format_info {/** @is_yuv: Is it a YUV format? */ bool is_yuv; + +/** @is_fp: Is it a floating point format? */ +bool is_fp; };/** include/uapi/drm/drm_fourcc.h +0
-0
include/uapi/drm/drm_fourcc.h extern "C" { · line 147
Quoted 4 lines
include/uapi/drm/drm_fourcc.h +1
-0@@ -144,6 +144,12 @@ extern "C" {#define DRM_FORMAT_RGBA1010102 fourcc_code('R', 'A', '3', '0')/* [31:0] R:G:B:A 10:10:10:2 little endian */ #define DRM_FORMAT_BGRA1010102 fourcc_code('B', 'A', '3', '0')/* [31:0] B:G:R:A 10:10:10:2 little endian */ +/* 64 bpp RGB IEEE-754 half-precision float (binary16) */
Might as well document the bits:
* [15:0] sign:exponent:mantissa 1:5:10
include/uapi/drm/drm_fourcc.h extern "C" { · line 151
Quoted 4 lines
include/uapi/drm/drm_fourcc.h +4
-0+#define DRM_FORMAT_XBGR16161616H fourcc_code('X', 'B', '4', 'H')/* [63:0] x:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_ABGR16161616H fourcc_code('A', 'B', '4', 'H')/* [63:0] A:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_XRGB16161616H fourcc_code('X', 'R', '4', 'H')/* [63:0] x:R:G:B 16:16:16:16 little endian */ +#define DRM_FORMAT_ARGB16161616H fourcc_code('A', 'R', '4', 'H')/* [63:0] A:R:G:B 16:16:16:16 little endian */
ocd: the other formats have separate groups for X vs. A separate. And RGB before BGR.
Quoted 6 lines
include/uapi/drm/drm_fourcc.h +1
-0+/* packed YCbCr */ #define DRM_FORMAT_YUYV fourcc_code('Y', 'U', 'Y', 'V')/* [31:0] Cr0:Y1:Cb0:Y0 8:8:8:8 little endian */ #define DRM_FORMAT_YVYU fourcc_code('Y', 'V', 'Y', 'U')/* [31:0] Cb0:Y1:Cr0:Y0 8:8:8:8 little endian */ Signature · 1 line
2.7.4
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: [PATCH 1/3] drm/fourcc: Add 64 bpp half float formats
Quoted 7 lines · Ville Syrjälä
drivers/gpu/drm/drm_fourcc.c const struct drm_format_info *__drm_format_info(u32 format) · line 204
Quoted 4 lines · Kevin Strasser
drivers/gpu/drm/drm_fourcc.c +4
-0+ { .format = DRM_FORMAT_XRGB16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_XBGR16161616H, .depth = 48, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .is_fp = true }, + { .format = DRM_FORMAT_ARGB16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true }, + { .format = DRM_FORMAT_ABGR16161616H, .depth = 64, .num_planes = 1, .cpp = { 4, 0, 0 }, .hsub = 1, .vsub = 1, .has_alpha = true, .is_fp = true },cpp is wrong for all of these.
I'll change it to .cpp = { 8, 0, 0 }
Quoted 4 lines
Quoted 1 line
+/* 64 bpp RGB IEEE-754 half-precision float (binary16) */
Might as well document the bits:
* [15:0] sign:exponent:mantissa 1:5:10
Sure.
Quoted 7 lines
Review
Quoted 4 lines
(quoted diff) +4
-0+#define DRM_FORMAT_XBGR16161616H fourcc_code('X', 'B', '4', 'H') /* [63:0] >> x:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_ABGR16161616H fourcc_code('A', 'B', '4', 'H') /* [63:0] >> A:B:G:R 16:16:16:16 little endian */ +#define DRM_FORMAT_XRGB16161616H fourcc_code('X', 'R', '4', 'H') /* [63:0] >> x:R:G:B 16:16:16:16 little endian */ +#define DRM_FORMAT_ARGB16161616H fourcc_code('A', 'R', '4', 'H') /* [63:0] >> A:R:G:B 16:16:16:16 little endian */ocd: the other formats have separate groups for X vs. A separate. And RGB before BGR.
Agreed, I'll fix these up.
Thanks,
Kevin
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 2/3 drm: Add optional PIXEL_NORMALIZE_RANGE property to drm_plane
In reply to Kevin StrasserAdd an optional property to allow applications to indicate what range their floating point pixel data is normalized to. Drivers are free to choose what ranges they want to support and can attach this property to each plane that actually supports floating point formats
--- drivers/gpu/drm/drm_atomic.c | 2 ++ drivers/gpu/drm/drm_atomic_uapi.c | 4 +++ drivers/gpu/drm/drm_color_mgmt.c | 68 +++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/drm_crtc_internal.h | 1 + include/drm/drm_color_mgmt.h | 9 +++++ include/drm/drm_plane.h | 14 ++++++++ 6 files changed, 98 insertions(+)
drivers/gpu/drm/drm_atomic.c +2 -0
@@ -624,6 +624,8 @@ static void drm_atomic_plane_print_state(struct drm_printer *p, drm_get_color_encoding_name(state->color_encoding));drm_printf (p,"\tcolor-range=%s\n" ,drm_get_color_range_name (state->color_range)); +drm_printf (p,"\tpixel-normalize-range=%s\n" , +drm_get_pixel_normalize_range_name (state->pixel_normalize_range));if (plane->funcs->atomic_print_state) plane->funcs->atomic_print_state (p, state);
drivers/gpu/drm/drm_atomic_uapi.c +4 -0
@@ -566,6 +566,8 @@ static int drm_atomic_plane_set_property(struct drm_plane *plane, state->color_encoding = val; }else if (property == plane->color_range_property) { state->color_range = val; + }else if (property == plane->pixel_normalize_range_property) { + state->pixel_normalize_range = val; }else if (plane->funcs->atomic_set_property) {return plane->funcs->atomic_set_property (plane, state, property, val);
@@ -621,6 +623,8 @@ drm_atomic_plane_get_property(struct drm_plane *plane, *val = state->color_encoding; }else if (property == plane->color_range_property) { *val = state->color_range; + }else if (property == plane->pixel_normalize_range_property) { + *val = state->pixel_normalize_range; }else if (plane->funcs->atomic_get_property) {return plane->funcs->atomic_get_property (plane, state, property, val); }else {
drivers/gpu/drm/drm_color_mgmt.c +68 -0
71 lines not on this page · raw message
drivers/gpu/drm/drm_crtc_internal.h +1 -0
@@ -90,6 +90,7 @@ int drm_mode_destroy_dumb_ioctl(struct drm_device *dev,/* drm_color_mgmt.c */ const char *drm_get_color_encoding_name (enum drm_color_encoding encoding);const char *drm_get_color_range_name (enum drm_color_range range); +const char *drm_get_pixel_normalize_range_name (enum drm_pixel_normalize_range range);/* IOCTLs */ int drm_mode_gamma_get_ioctl(struct drm_device *dev,
include/drm/drm_color_mgmt.h +9 -0
@@ -64,9 +64,18 @@ enum drm_color_range { DRM_COLOR_RANGE_MAX, }; +enum drm_pixel_normalize_range { +DRM_PIXEL_NORMALIZE_RANGE_0_1 , +DRM_PIXEL_NORMALIZE_RANGE_0_255 , +DRM_PIXEL_NORMALIZE_RANGE_MAX +}; +int drm_plane_create_color_properties (struct drm_plane *plane,u32 supported_encodings,u32 supported_ranges,enum drm_color_encoding default_encoding,enum drm_color_range default_range); +int drm_plane_create_pixel_normalize_range_property (struct drm_plane *plane, +u32 supported_ranges, +enum drm_pixel_normalize_range default_range);#endif
include/drm/drm_plane.h +14 -0
@@ -130,6 +130,13 @@ struct drm_plane_state {uint16_t pixel_blend_mode;/** +* @pixel_normalize_range: +* The range to use for floating point pixel data normalization. See +* drm_plane_create_pixel_normalize_range_property() for more details. +*/ +enum drm_pixel_normalize_range pixel_normalize_range; + +/** * @rotation: * Rotation of the plane. See drm_plane_create_rotation_property() for * more details.
@@ -680,6 +687,13 @@ struct drm_plane {struct drm_property *blend_mode_property;/** +* @pixel_normalize_range_property: +* Optional "PIXEL_NORMALIZE_RANGE" property for this plane. See +* drm_plane_create_pixel_normalize_range_property(). +*/ +struct drm_property *pixel_normalize_range_property; + +/** * @color_encoding_property: * * Optional "COLOR_ENCODING" enum property for specifying
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/3] drm: Add optional PIXEL_NORMALIZE_RANGE property to drm_plane
Quoted 4 lines · Kevin Strasser
Add an optional property to allow applications to indicate what range their floating point pixel data is normalized to. Drivers are free to choose what ranges they want to support and can attach this property to each plane that actually supports floating point formats
Do we have a plan to actually use this? Earlier platforms didn't have anything like this IIRC.
Quoted 197 lines
Signed-off-by Kevin Strasser <[email protected]>Cc Uma Shankar <[email protected]>Cc Shashank Sharma <[email protected]>Cc Ville Syrjälä <[email protected]>Cc David Airlie <[email protected]>Cc Daniel Vetter <[email protected]>--- drivers/gpu/drm/drm_atomic.c | 2 ++ drivers/gpu/drm/drm_atomic_uapi.c | 4 +++ drivers/gpu/drm/drm_color_mgmt.c | 68 +++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/drm_crtc_internal.h | 1 + include/drm/drm_color_mgmt.h | 9 +++++ include/drm/drm_plane.h | 14 ++++++++ 6 files changed, 98 insertions(+)drivers/gpu/drm/drm_atomic.c +2
-0@@ -624,6 +624,8 @@ static void drm_atomic_plane_print_state(struct drm_printer *p, drm_get_color_encoding_name(state->color_encoding));drm_printf (p,"\tcolor-range=%s\n" ,drm_get_color_range_name (state->color_range)); +drm_printf (p,"\tpixel-normalize-range=%s\n" , +drm_get_pixel_normalize_range_name (state->pixel_normalize_range));if (plane->funcs->atomic_print_state) plane->funcs->atomic_print_state (p, state);drivers/gpu/drm/drm_atomic_uapi.c +4
-0@@ -566,6 +566,8 @@ static int drm_atomic_plane_set_property(struct drm_plane *plane, state->color_encoding = val; }else if (property == plane->color_range_property) { state->color_range = val; + }else if (property == plane->pixel_normalize_range_property) { + state->pixel_normalize_range = val; }else if (plane->funcs->atomic_set_property) {return plane->funcs->atomic_set_property (plane, state, property, val);@@ -621,6 +623,8 @@ drm_atomic_plane_get_property(struct drm_plane *plane, *val = state->color_encoding; }else if (property == plane->color_range_property) { *val = state->color_range; + }else if (property == plane->pixel_normalize_range_property) { + *val = state->pixel_normalize_range; }else if (plane->funcs->atomic_get_property) {return plane->funcs->atomic_get_property (plane, state, property, val); }else {drivers/gpu/drm/drm_color_mgmt.c +68
-0@@ -472,3 +472,71 @@ int drm_plane_create_color_properties(struct drm_plane *plane,return 0 ; }EXPORT_SYMBOL (drm_plane_create_color_properties); + +static const char *const pixel_normalize_range_name[] = { + [DRM_PIXEL_NORMALIZE_RANGE_0_1 ] ="0.0 - 1.0" , + [DRM_PIXEL_NORMALIZE_RANGE_0_255 ] ="0.0 - 255.0" , +}; + +/** +* drm_get_pixel_normalize_range_name - return a string for pixel normalize +* range +* @range: pixel normalize range to compute name of +* +* In contrast to the other drm_get_*_name functions this one here returns a +* const pointer and hence is threadsafe. +*/ +const char *drm_get_pixel_normalize_range_name (enum drm_pixel_normalize_range range) +{ +if (WARN_ON (range >=ARRAY_SIZE (pixel_normalize_range_name))) +return "unknown" ; + +return pixel_normalize_range_name[range]; +} + +/** +* drm_plane_create_pixel_normalize_range_property - pixel normalize range +* property +* @plane: plane object +* @supported_ranges: bitfield indicating supported normalize ranges +* @default_range: default normalize range +* +* Create and attach plane specific PIXEL_NORMALIZE_RANGE property to @plane. +* The supported ranges should be provided in supported_ranges bitmask. Eeach +* bit set in the bitmask indicates that its number as enum value is supported. +*/ +int drm_plane_create_pixel_normalize_range_property (struct drm_plane *plane, +u32 supported_ranges,enum drm_pixel_normalize_range default_range) +{ +struct drm_property *prop; +struct drm_prop_enum_list enum_list[DRM_PIXEL_NORMALIZE_RANGE_MAX ]; +int i, len =0 ; + +if (WARN_ON (supported_ranges ==0 || + (supported_ranges & -BIT (DRM_PIXEL_NORMALIZE_RANGE_MAX )) !=0 || + (supported_ranges &BIT (default_range)) ==0 )) +return -EINVAL ; + +for (i =0 ; i <DRM_PIXEL_NORMALIZE_RANGE_MAX ; i++) { +if ((supported_ranges &BIT (i)) ==0 ) +continue ; + + enum_list[len].type = i; + enum_list[len].name = pixel_normalize_range_name[i]; + len++; + } + + prop =drm_property_create_enum (plane->dev,0 ,"PIXEL_NORMALIZE_RANGE" , + enum_list, len); +if (!prop) +return -ENOMEM ; + + plane->pixel_normalize_range_property = prop; +drm_object_attach_property (&plane->base, prop, default_range); + +if (plane->state) + plane->state->pixel_normalize_range = default_range; + +return 0 ; +} +EXPORT_SYMBOL (drm_plane_create_pixel_normalize_range_property);drivers/gpu/drm/drm_crtc_internal.h +1
-0@@ -90,6 +90,7 @@ int drm_mode_destroy_dumb_ioctl(struct drm_device *dev,/* drm_color_mgmt.c */ const char *drm_get_color_encoding_name (enum drm_color_encoding encoding);const char *drm_get_color_range_name (enum drm_color_range range); +const char *drm_get_pixel_normalize_range_name (enum drm_pixel_normalize_range range);/* IOCTLs */ int drm_mode_gamma_get_ioctl(struct drm_device *dev,include/drm/drm_color_mgmt.h +9
-0@@ -64,9 +64,18 @@ enum drm_color_range { DRM_COLOR_RANGE_MAX, }; +enum drm_pixel_normalize_range { +DRM_PIXEL_NORMALIZE_RANGE_0_1 , +DRM_PIXEL_NORMALIZE_RANGE_0_255 , +DRM_PIXEL_NORMALIZE_RANGE_MAX +}; +int drm_plane_create_color_properties (struct drm_plane *plane,u32 supported_encodings,u32 supported_ranges,enum drm_color_encoding default_encoding,enum drm_color_range default_range); +int drm_plane_create_pixel_normalize_range_property (struct drm_plane *plane, +u32 supported_ranges, +enum drm_pixel_normalize_range default_range);#endif include/drm/drm_plane.h +14
-0@@ -130,6 +130,13 @@ struct drm_plane_state {uint16_t pixel_blend_mode;/** +* @pixel_normalize_range: +* The range to use for floating point pixel data normalization. See +* drm_plane_create_pixel_normalize_range_property() for more details. +*/ +enum drm_pixel_normalize_range pixel_normalize_range; + +/** * @rotation: * Rotation of the plane. See drm_plane_create_rotation_property() for * more details. @@ -680,6 +687,13 @@ struct drm_plane {struct drm_property *blend_mode_property;/** +* @pixel_normalize_range_property: +* Optional "PIXEL_NORMALIZE_RANGE" property for this plane. See +* drm_plane_create_pixel_normalize_range_property(). +*/ +struct drm_property *pixel_normalize_range_property; + +/** * @color_encoding_property: * * Optional "COLOR_ENCODING" enum property for specifying Signature · 1 line
2.7.4
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
RE: [PATCH 2/3] drm: Add optional PIXEL_NORMALIZE_RANGE property to drm_plane
Quoted 8 lines · Ville Syrjälä
Quoted 4 lines · Kevin Strasser
Add an optional property to allow applications to indicate what range their floating point pixel data is normalized to. Drivers are free to choose what ranges they want to support and can attach this property to each plane that actually supports floating point formats
Do we have a plan to actually use this? Earlier platforms didn't have anything like this IIRC.
I don't know about any *real* use case, but I did find it useful in the IGT conversion code. It isn't strictly required to be enabled if the data is already normalized.
Thanks,
Kevin
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH 3/3 drm/i915: Implement half float formats and pixel normalize property
In reply to Kevin Strasser64 bpp half float formats are supported on hdr planes only and are subject to the following restrictions:
* 90/270 rotation not supported
* Yf Tiling not supported
* Frame Buffer Compression not supported
* Color Keying not supported
The behavior of pixel normalize with non-float formats is currently undefined. As such, the pixel normalize register is enabled iff the framebuffer contains floating point pixel data.
--- drivers/gpu/drm/i915/i915_reg.h | 15 ++++++- drivers/gpu/drm/i915/intel_display.c | 48 +++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 5 +++ drivers/gpu/drm/i915/intel_sprite.c | 82 +++++++++++++++++++++++++++++++++--- 4 files changed, 143 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/i915_reg.h +14 -1
@@ -6563,6 +6563,10 @@ enum {#define _PLANE_KEYMAX_1_A 0x701a0#define _PLANE_KEYMAX_2_A 0x702a0#define PLANE_KEYMAX_ALPHA (a) ((a) << 24) +#define _PLANE_PIXEL_NORMALIZE_1_A 0x701a8 +#define _PLANE_PIXEL_NORMALIZE_2_A 0x702a8 +#define PLANE_PIXEL_NORMALIZE_ENABLE (1 << 31) +#define PLANE_PIXEL_NORMALIZE_FACTOR_MASK 0xffff#define _PLANE_AUX_DIST_1_A 0x701c0#define _PLANE_AUX_DIST_2_A 0x702c0#define _PLANE_AUX_OFFSET_1_A 0x701c4
@@ -6786,7 +6790,16 @@ enum {#define PLANE_COLOR_CTL (pipe, plane) \ _MMIO_PLANE(plane, _PLANE_COLOR_CTL_1(pipe), _PLANE_COLOR_CTL_2(pipe)) -#/* SKL new cursor registers */ +#define _PLANE_PIXEL_NORMALIZE_1_B 0x711a8 +#define _PLANE_PIXEL_NORMALIZE_2_B 0x712a8 +#define _PLANE_PIXEL_NORMALIZE_1 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_1_A, _PLANE_PIXEL_NORMALIZE_1_B) +#define _PLANE_PIXEL_NORMALIZE_2 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_2_A, _PLANE_PIXEL_NORMALIZE_2_B) +#define PLANE_PIXEL_NORMALIZE (pipe, plane) \ + _MMIO_PLANE(plane, _PLANE_PIXEL_NORMALIZE_1(pipe), _PLANE_PIXEL_NORMALIZE_2(pipe)) + +/* SKL new cursor registers */ #define _CUR_BUF_CFG_A 0x7017c#define _CUR_BUF_CFG_B 0x7117c#define CUR_BUF_CFG (pipe) _MMIO_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)
drivers/gpu/drm/i915/intel_display.c +48 -0
72 lines not on this page · raw message
drivers/gpu/drm/i915/intel_drv.h +5 -0
@@ -565,6 +565,9 @@ struct intel_plane_state {u32 slave;struct drm_intel_sprite_colorkey ckey; + +/* plane pixel normalize register */ +u32 pixel_normalize; };struct intel_initial_plane_config {
@@ -1738,6 +1741,8 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state,u32 glk_color_ctl (const struct intel_plane_state *plane_state);u32 skl_plane_stride (const struct intel_plane_state *plane_state,int plane); +u32 icl_plane_pixel_normalize (uint32_t pixel_format, +enum drm_pixel_normalize_range range);int skl_check_plane_surface (struct intel_plane_state *plane_state);int i9xx_check_plane_surface (struct intel_plane_state *plane_state);int skl_format_to_fourcc (int format,bool rgb_order,bool alpha);
drivers/gpu/drm/i915/intel_sprite.c +76 -6
131 lines not on this page · raw message
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 3/3] drm/i915: Implement half float formats and pixel normalize property
Quoted 11 lines · Kevin Strasser
64 bpp half float formats are supported on hdr planes only and are subject to the following restrictions:
* 90/270 rotation not supported
* Yf Tiling not supported
* Frame Buffer Compression not supported
* Color Keying not supportedThe behavior of pixel normalize with non-float formats is currently undefined. As such, the pixel normalize register is enabled iff the framebuffer contains floating point pixel data.
Quoted 7 lines
Signed-off-by Kevin Strasser <[email protected]>Cc Uma Shankar <[email protected]>Cc Shashank Sharma <[email protected]>Cc Ville Syrjälä <[email protected]>Cc David Airlie <[email protected]>Cc Daniel Vetter <[email protected]>
Do we have end-to-end userspace for this?
-Daniel
Quoted 289 lines
--- drivers/gpu/drm/i915/i915_reg.h | 15 ++++++- drivers/gpu/drm/i915/intel_display.c | 48 +++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 5 +++ drivers/gpu/drm/i915/intel_sprite.c | 82 +++++++++++++++++++++++++++++++++--- 4 files changed, 143 insertions(+), 7 deletions(-)drivers/gpu/drm/i915/i915_reg.h +14
-1@@ -6563,6 +6563,10 @@ enum {#define _PLANE_KEYMAX_1_A 0x701a0#define _PLANE_KEYMAX_2_A 0x702a0#define PLANE_KEYMAX_ALPHA (a) ((a) << 24) +#define _PLANE_PIXEL_NORMALIZE_1_A 0x701a8 +#define _PLANE_PIXEL_NORMALIZE_2_A 0x702a8 +#define PLANE_PIXEL_NORMALIZE_ENABLE (1 << 31) +#define PLANE_PIXEL_NORMALIZE_FACTOR_MASK 0xffff#define _PLANE_AUX_DIST_1_A 0x701c0#define _PLANE_AUX_DIST_2_A 0x702c0#define _PLANE_AUX_OFFSET_1_A 0x701c4@@ -6786,7 +6790,16 @@ enum {#define PLANE_COLOR_CTL (pipe, plane) \ _MMIO_PLANE(plane, _PLANE_COLOR_CTL_1(pipe), _PLANE_COLOR_CTL_2(pipe)) -#/* SKL new cursor registers */ +#define _PLANE_PIXEL_NORMALIZE_1_B 0x711a8 +#define _PLANE_PIXEL_NORMALIZE_2_B 0x712a8 +#define _PLANE_PIXEL_NORMALIZE_1 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_1_A, _PLANE_PIXEL_NORMALIZE_1_B) +#define _PLANE_PIXEL_NORMALIZE_2 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_2_A, _PLANE_PIXEL_NORMALIZE_2_B) +#define PLANE_PIXEL_NORMALIZE (pipe, plane) \ + _MMIO_PLANE(plane, _PLANE_PIXEL_NORMALIZE_1(pipe), _PLANE_PIXEL_NORMALIZE_2(pipe)) + +/* SKL new cursor registers */ #define _CUR_BUF_CFG_A 0x7017c#define _CUR_BUF_CFG_B 0x7117c#define CUR_BUF_CFG (pipe) _MMIO_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)drivers/gpu/drm/i915/intel_display.c +48
-0@@ -2618,6 +2618,18 @@ int skl_format_to_fourcc(int format, bool rgb_order, bool alpha)return DRM_FORMAT_RGB565 ;case PLANE_CTL_FORMAT_NV12 :return DRM_FORMAT_NV12 ; +case PLANE_CTL_FORMAT_XRGB_16161616F : +if (rgb_order) { +if (alpha) +return DRM_FORMAT_ABGR16161616H ; +else +return DRM_FORMAT_XBGR16161616H ; + }else { +if (alpha) +return DRM_FORMAT_ARGB16161616H ; +else +return DRM_FORMAT_XRGB16161616H ; + }default :case PLANE_CTL_FORMAT_XRGB_8888 :if (rgb_order) {@@ -3505,6 +3517,12 @@ static u32 skl_plane_ctl_format(uint32_t pixel_format)return PLANE_CTL_FORMAT_YUV422 |PLANE_CTL_YUV422_VYUY ;case DRM_FORMAT_NV12 :return PLANE_CTL_FORMAT_NV12 ; +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +return PLANE_CTL_FORMAT_XRGB_16161616F |PLANE_CTL_ORDER_RGBX ; +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H : +return PLANE_CTL_FORMAT_XRGB_16161616F ;default :MISSING_CASE (pixel_format); }@@ -3680,6 +3698,32 @@ u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state,return plane_color_ctl; } +u32 icl_plane_pixel_normalize (uint32_t pixel_format, +enum drm_pixel_normalize_range range) +{ +/* 1.0 in half float */ +u16 half_float_1 =0x3c00 ; +/* 3.92E-3 in half float */ +u16 half_float_255 =0x1c04 ; + +switch (pixel_format) { +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ARGB16161616H : +case DRM_FORMAT_ABGR16161616H : +switch (range) { +case DRM_PIXEL_NORMALIZE_RANGE_0_1 : +return PLANE_PIXEL_NORMALIZE_ENABLE | half_float_1; +case DRM_PIXEL_NORMALIZE_RANGE_0_255 : +return PLANE_PIXEL_NORMALIZE_ENABLE | half_float_255; +default : +return 0 ; + } +default : +return 0 ; + } +} + static int __intel_display_resume(struct drm_device *dev, struct drm_atomic_state *state,@@ -4998,6 +5042,10 @@ static int skl_update_scaler_plane(struct intel_crtc_state *crtc_state,case DRM_FORMAT_UYVY :case DRM_FORMAT_VYUY :case DRM_FORMAT_NV12 : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H :break ;default : DRM_DEBUG_KMS("[PLANE:%d:%s] FB:%d unsupported scaling format 0x%x\n",drivers/gpu/drm/i915/intel_drv.h +5
-0@@ -565,6 +565,9 @@ struct intel_plane_state {u32 slave;struct drm_intel_sprite_colorkey ckey; + +/* plane pixel normalize register */ +u32 pixel_normalize; };struct intel_initial_plane_config {@@ -1738,6 +1741,8 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state,u32 glk_color_ctl (const struct intel_plane_state *plane_state);u32 skl_plane_stride (const struct intel_plane_state *plane_state,int plane); +u32 icl_plane_pixel_normalize (uint32_t pixel_format, +enum drm_pixel_normalize_range range);int skl_check_plane_surface (struct intel_plane_state *plane_state);int i9xx_check_plane_surface (struct intel_plane_state *plane_state);int skl_format_to_fourcc (int format,bool rgb_order,bool alpha);drivers/gpu/drm/i915/intel_sprite.c +76
-6@@ -515,6 +515,10 @@ skl_program_plane(struct intel_plane *plane,if (fb->format->is_yuv &&icl_is_hdr_plane (plane))icl_program_input_csc_coeff (crtc_state, plane_state); +if (fb->format->is_fp &&icl_is_hdr_plane (plane)) +I915_WRITE_FW (PLANE_PIXEL_NORMALIZE (pipe, plane_id), + plane_state->pixel_normalize); +I915_WRITE_FW (PLANE_KEYVAL (pipe, plane_id), key->min_value);I915_WRITE_FW (PLANE_KEYMAX (pipe, plane_id), keymax);I915_WRITE_FW (PLANE_KEYMSK (pipe, plane_id), keymsk);@@ -1417,8 +1421,6 @@ static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,/* * 90/270 is not allowed with RGB64 16:16:16:16 and * Indexed 8-bit. RGB 16-bit 5:6:5 is allowed gen11 onwards. -* TBD: Add RGB64 case once its added in supported format -* list. */ switch (fb->format->format) {case DRM_FORMAT_RGB565 :@@ -1426,6 +1428,10 @@ static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,break ;/* fall through */ case DRM_FORMAT_C8 : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ARGB16161616H : +case DRM_FORMAT_ABGR16161616H :DRM_DEBUG_KMS ("Unsupported pixel format %s for 90/270!\n" ,drm_get_format_name (fb->format->format, &format_name));@@ -1552,6 +1558,11 @@ static int skl_plane_check(struct intel_crtc_state *crtc_state, plane_state->color_ctl =glk_plane_color_ctl (crtc_state, plane_state); +if (icl_is_hdr_plane (plane)) + plane_state->pixel_normalize = +icl_plane_pixel_normalize (plane_state->base.fb->format->format, + plane_state->base.pixel_normalize_range); +return 0 ; }@@ -1741,6 +1752,45 @@ static const uint32_t skl_planar_formats[] = { DRM_FORMAT_NV12, }; +static const uint32_t icl_hdr_plane_formats[] = { +DRM_FORMAT_C8 , +DRM_FORMAT_RGB565 , +DRM_FORMAT_XRGB8888 , +DRM_FORMAT_XBGR8888 , +DRM_FORMAT_ARGB8888 , +DRM_FORMAT_ABGR8888 , +DRM_FORMAT_XRGB2101010 , +DRM_FORMAT_XBGR2101010 , +DRM_FORMAT_XRGB16161616H , +DRM_FORMAT_XBGR16161616H , +DRM_FORMAT_ARGB16161616H , +DRM_FORMAT_ABGR16161616H , +DRM_FORMAT_YUYV , +DRM_FORMAT_YVYU , +DRM_FORMAT_UYVY , +DRM_FORMAT_VYUY , +}; + +static const uint32_t icl_hdr_planar_formats[] = { +DRM_FORMAT_C8 , +DRM_FORMAT_RGB565 , +DRM_FORMAT_XRGB8888 , +DRM_FORMAT_XBGR8888 , +DRM_FORMAT_ARGB8888 , +DRM_FORMAT_ABGR8888 , +DRM_FORMAT_XRGB2101010 , +DRM_FORMAT_XBGR2101010 , +DRM_FORMAT_XRGB16161616H , +DRM_FORMAT_XBGR16161616H , +DRM_FORMAT_ARGB16161616H , +DRM_FORMAT_ABGR16161616H , +DRM_FORMAT_YUYV , +DRM_FORMAT_YVYU , +DRM_FORMAT_UYVY , +DRM_FORMAT_VYUY , +DRM_FORMAT_NV12 , +}; +static const uint64_t skl_plane_format_modifiers_noccs[] = { I915_FORMAT_MOD_Yf_TILED,I915_FORMAT_MOD_Y_TILED ,@@ -1884,6 +1934,10 @@ static bool skl_plane_format_mod_supported(struct drm_plane *_plane,return true;/* fall through */ case DRM_FORMAT_C8 : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H :if (modifier ==DRM_FORMAT_MOD_LINEAR || modifier ==I915_FORMAT_MOD_X_TILED || modifier ==I915_FORMAT_MOD_Y_TILED )@@ -2020,11 +2074,21 @@ skl_universal_plane_create(struct drm_i915_private *dev_priv, plane->update_slave = icl_update_slave;if (skl_plane_has_planar (dev_priv, pipe, plane_id)) { - formats = skl_planar_formats; - num_formats =ARRAY_SIZE (skl_planar_formats); +if (icl_is_hdr_plane (plane)) { + formats = icl_hdr_planar_formats; + num_formats =ARRAY_SIZE (icl_hdr_planar_formats); + }else { + formats = skl_planar_formats; + num_formats =ARRAY_SIZE (skl_planar_formats); + } }else { - formats = skl_plane_formats; - num_formats =ARRAY_SIZE (skl_plane_formats); +if (icl_is_hdr_plane (plane)) { + formats = icl_hdr_plane_formats; + num_formats =ARRAY_SIZE (icl_hdr_plane_formats); + }else { + formats = skl_plane_formats; + num_formats =ARRAY_SIZE (skl_plane_formats); + } } plane->has_ccs =skl_plane_has_ccs (dev_priv, pipe, plane_id);@@ -2074,6 +2138,12 @@ skl_universal_plane_create(struct drm_i915_private *dev_priv,BIT (DRM_MODE_BLEND_PREMULTI ) | BIT(DRM_MODE_BLEND_COVERAGE)); +if (icl_is_hdr_plane (plane)) +drm_plane_create_pixel_normalize_range_property (&plane->base, +BIT (DRM_PIXEL_NORMALIZE_RANGE_0_1 ) | +BIT (DRM_PIXEL_NORMALIZE_RANGE_0_255 ), +DRM_PIXEL_NORMALIZE_RANGE_0_1 ); +drm_plane_helper_add (&plane->base, &intel_plane_helper_funcs);return plane;Signature · 1 line
2.7.4
Signature · 7 lines
Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: [PATCH 3/3] drm/i915: Implement half float formats and pixel normalize property
Quoted 1 line · Daniel Vetter
Do we have end-to-end userspace for this?
I have patches for IGT and I'm planning on adding usage code to Weston. Apart from that there is a Windows use case that Tina mentioned previously. I take it that you will need to see the Weston part before accepting this.
Thanks,
Kevin
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 3/3] drm/i915: Implement half float formats and pixel normalize property
Quoted 6 lines · Kevin
Quoted 1 line · Daniel Vetter
Do we have end-to-end userspace for this?
I have patches for IGT and I'm planning on adding usage code to Weston. Apart from that there is a Windows use case that Tina mentioned previously. I take it that you will need to see the Weston part before accepting this.
Yup.
-Daniel
Signature · 7 lines
Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 3/3] drm/i915: Implement half float formats and pixel normalize property
In reply to Kevin StrasserQuoted 6 lines · Kevin Strasser
64 bpp half float formats are supported on hdr planes only and are subject to the following restrictions:
* 90/270 rotation not supported
* Yf Tiling not supported
* Frame Buffer Compression not supported
* Color Keying not supported
Is scaling supported? It wasn't on earlier platforms.
Quoted 300 lines
The behavior of pixel normalize with non-float formats is currently undefined. As such, the pixel normalize register is enabled iff the framebuffer contains floating point pixel data.
Signed-off-by Kevin Strasser <[email protected]>Cc Uma Shankar <[email protected]>Cc Shashank Sharma <[email protected]>Cc Ville Syrjälä <[email protected]>Cc David Airlie <[email protected]>Cc Daniel Vetter <[email protected]>--- drivers/gpu/drm/i915/i915_reg.h | 15 ++++++- drivers/gpu/drm/i915/intel_display.c | 48 +++++++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 5 +++ drivers/gpu/drm/i915/intel_sprite.c | 82 +++++++++++++++++++++++++++++++++--- 4 files changed, 143 insertions(+), 7 deletions(-)drivers/gpu/drm/i915/i915_reg.h +14
-1@@ -6563,6 +6563,10 @@ enum {#define _PLANE_KEYMAX_1_A 0x701a0#define _PLANE_KEYMAX_2_A 0x702a0#define PLANE_KEYMAX_ALPHA (a) ((a) << 24) +#define _PLANE_PIXEL_NORMALIZE_1_A 0x701a8 +#define _PLANE_PIXEL_NORMALIZE_2_A 0x702a8 +#define PLANE_PIXEL_NORMALIZE_ENABLE (1 << 31) +#define PLANE_PIXEL_NORMALIZE_FACTOR_MASK 0xffff#define _PLANE_AUX_DIST_1_A 0x701c0#define _PLANE_AUX_DIST_2_A 0x702c0#define _PLANE_AUX_OFFSET_1_A 0x701c4@@ -6786,7 +6790,16 @@ enum {#define PLANE_COLOR_CTL (pipe, plane) \ _MMIO_PLANE(plane, _PLANE_COLOR_CTL_1(pipe), _PLANE_COLOR_CTL_2(pipe)) -#/* SKL new cursor registers */ +#define _PLANE_PIXEL_NORMALIZE_1_B 0x711a8 +#define _PLANE_PIXEL_NORMALIZE_2_B 0x712a8 +#define _PLANE_PIXEL_NORMALIZE_1 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_1_A, _PLANE_PIXEL_NORMALIZE_1_B) +#define _PLANE_PIXEL_NORMALIZE_2 (pipe) \ + _PIPE(pipe, _PLANE_PIXEL_NORMALIZE_2_A, _PLANE_PIXEL_NORMALIZE_2_B) +#define PLANE_PIXEL_NORMALIZE (pipe, plane) \ + _MMIO_PLANE(plane, _PLANE_PIXEL_NORMALIZE_1(pipe), _PLANE_PIXEL_NORMALIZE_2(pipe)) + +/* SKL new cursor registers */ #define _CUR_BUF_CFG_A 0x7017c#define _CUR_BUF_CFG_B 0x7117c#define CUR_BUF_CFG (pipe) _MMIO_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)drivers/gpu/drm/i915/intel_display.c +48
-0@@ -2618,6 +2618,18 @@ int skl_format_to_fourcc(int format, bool rgb_order, bool alpha)return DRM_FORMAT_RGB565 ;case PLANE_CTL_FORMAT_NV12 :return DRM_FORMAT_NV12 ; +case PLANE_CTL_FORMAT_XRGB_16161616F : +if (rgb_order) { +if (alpha) +return DRM_FORMAT_ABGR16161616H ; +else +return DRM_FORMAT_XBGR16161616H ; + }else { +if (alpha) +return DRM_FORMAT_ARGB16161616H ; +else +return DRM_FORMAT_XRGB16161616H ; + }default :case PLANE_CTL_FORMAT_XRGB_8888 :if (rgb_order) {@@ -3505,6 +3517,12 @@ static u32 skl_plane_ctl_format(uint32_t pixel_format)return PLANE_CTL_FORMAT_YUV422 |PLANE_CTL_YUV422_VYUY ;case DRM_FORMAT_NV12 :return PLANE_CTL_FORMAT_NV12 ; +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +return PLANE_CTL_FORMAT_XRGB_16161616F |PLANE_CTL_ORDER_RGBX ; +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H : +return PLANE_CTL_FORMAT_XRGB_16161616F ;default :MISSING_CASE (pixel_format); }@@ -3680,6 +3698,32 @@ u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state,return plane_color_ctl; } +u32 icl_plane_pixel_normalize (uint32_t pixel_format, +enum drm_pixel_normalize_range range) +{ +/* 1.0 in half float */ +u16 half_float_1 =0x3c00 ; +/* 3.92E-3 in half float */ +u16 half_float_255 =0x1c04 ; + +switch (pixel_format) { +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ARGB16161616H : +case DRM_FORMAT_ABGR16161616H : +switch (range) { +case DRM_PIXEL_NORMALIZE_RANGE_0_1 : +return PLANE_PIXEL_NORMALIZE_ENABLE | half_float_1; +case DRM_PIXEL_NORMALIZE_RANGE_0_255 : +return PLANE_PIXEL_NORMALIZE_ENABLE | half_float_255; +default : +return 0 ; + } +default : +return 0 ; + } +} + static int __intel_display_resume(struct drm_device *dev, struct drm_atomic_state *state,@@ -4998,6 +5042,10 @@ static int skl_update_scaler_plane(struct intel_crtc_state *crtc_state,case DRM_FORMAT_UYVY :case DRM_FORMAT_VYUY :case DRM_FORMAT_NV12 : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H :break ;default : DRM_DEBUG_KMS("[PLANE:%d:%s] FB:%d unsupported scaling format 0x%x\n",drivers/gpu/drm/i915/intel_drv.h +5
-0@@ -565,6 +565,9 @@ struct intel_plane_state {u32 slave;struct drm_intel_sprite_colorkey ckey; + +/* plane pixel normalize register */ +u32 pixel_normalize; };struct intel_initial_plane_config {@@ -1738,6 +1741,8 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state,u32 glk_color_ctl (const struct intel_plane_state *plane_state);u32 skl_plane_stride (const struct intel_plane_state *plane_state,int plane); +u32 icl_plane_pixel_normalize (uint32_t pixel_format, +enum drm_pixel_normalize_range range);int skl_check_plane_surface (struct intel_plane_state *plane_state);int i9xx_check_plane_surface (struct intel_plane_state *plane_state);int skl_format_to_fourcc (int format,bool rgb_order,bool alpha);drivers/gpu/drm/i915/intel_sprite.c +76
-6@@ -515,6 +515,10 @@ skl_program_plane(struct intel_plane *plane,if (fb->format->is_yuv &&icl_is_hdr_plane (plane))icl_program_input_csc_coeff (crtc_state, plane_state); +if (fb->format->is_fp &&icl_is_hdr_plane (plane)) +I915_WRITE_FW (PLANE_PIXEL_NORMALIZE (pipe, plane_id), + plane_state->pixel_normalize); +I915_WRITE_FW (PLANE_KEYVAL (pipe, plane_id), key->min_value);I915_WRITE_FW (PLANE_KEYMAX (pipe, plane_id), keymax);I915_WRITE_FW (PLANE_KEYMSK (pipe, plane_id), keymsk);@@ -1417,8 +1421,6 @@ static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,/* * 90/270 is not allowed with RGB64 16:16:16:16 and * Indexed 8-bit. RGB 16-bit 5:6:5 is allowed gen11 onwards. -* TBD: Add RGB64 case once its added in supported format -* list. */ switch (fb->format->format) {case DRM_FORMAT_RGB565 :@@ -1426,6 +1428,10 @@ static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,break ;/* fall through */ case DRM_FORMAT_C8 : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ARGB16161616H : +case DRM_FORMAT_ABGR16161616H :DRM_DEBUG_KMS ("Unsupported pixel format %s for 90/270!\n" ,drm_get_format_name (fb->format->format, &format_name));@@ -1552,6 +1558,11 @@ static int skl_plane_check(struct intel_crtc_state *crtc_state, plane_state->color_ctl =glk_plane_color_ctl (crtc_state, plane_state); +if (icl_is_hdr_plane (plane)) + plane_state->pixel_normalize = +icl_plane_pixel_normalize (plane_state->base.fb->format->format, + plane_state->base.pixel_normalize_range); +return 0 ; }@@ -1741,6 +1752,45 @@ static const uint32_t skl_planar_formats[] = { DRM_FORMAT_NV12, }; +static const uint32_t icl_hdr_plane_formats[] = { +DRM_FORMAT_C8 , +DRM_FORMAT_RGB565 , +DRM_FORMAT_XRGB8888 , +DRM_FORMAT_XBGR8888 , +DRM_FORMAT_ARGB8888 , +DRM_FORMAT_ABGR8888 , +DRM_FORMAT_XRGB2101010 , +DRM_FORMAT_XBGR2101010 , +DRM_FORMAT_XRGB16161616H , +DRM_FORMAT_XBGR16161616H , +DRM_FORMAT_ARGB16161616H , +DRM_FORMAT_ABGR16161616H , +DRM_FORMAT_YUYV , +DRM_FORMAT_YVYU , +DRM_FORMAT_UYVY , +DRM_FORMAT_VYUY , +}; + +static const uint32_t icl_hdr_planar_formats[] = { +DRM_FORMAT_C8 , +DRM_FORMAT_RGB565 , +DRM_FORMAT_XRGB8888 , +DRM_FORMAT_XBGR8888 , +DRM_FORMAT_ARGB8888 , +DRM_FORMAT_ABGR8888 , +DRM_FORMAT_XRGB2101010 , +DRM_FORMAT_XBGR2101010 , +DRM_FORMAT_XRGB16161616H , +DRM_FORMAT_XBGR16161616H , +DRM_FORMAT_ARGB16161616H , +DRM_FORMAT_ABGR16161616H , +DRM_FORMAT_YUYV , +DRM_FORMAT_YVYU , +DRM_FORMAT_UYVY , +DRM_FORMAT_VYUY , +DRM_FORMAT_NV12 , +}; +static const uint64_t skl_plane_format_modifiers_noccs[] = { I915_FORMAT_MOD_Yf_TILED,I915_FORMAT_MOD_Y_TILED ,@@ -1884,6 +1934,10 @@ static bool skl_plane_format_mod_supported(struct drm_plane *_plane,return true;/* fall through */ case DRM_FORMAT_C8 : +case DRM_FORMAT_XBGR16161616H : +case DRM_FORMAT_ABGR16161616H : +case DRM_FORMAT_XRGB16161616H : +case DRM_FORMAT_ARGB16161616H :if (modifier ==DRM_FORMAT_MOD_LINEAR || modifier ==I915_FORMAT_MOD_X_TILED || modifier ==I915_FORMAT_MOD_Y_TILED )@@ -2020,11 +2074,21 @@ skl_universal_plane_create(struct drm_i915_private *dev_priv, plane->update_slave = icl_update_slave;if (skl_plane_has_planar (dev_priv, pipe, plane_id)) { - formats = skl_planar_formats; - num_formats =ARRAY_SIZE (skl_planar_formats); +if (icl_is_hdr_plane (plane)) { + formats = icl_hdr_planar_formats; + num_formats =ARRAY_SIZE (icl_hdr_planar_formats); + }else { + formats = skl_planar_formats; + num_formats =ARRAY_SIZE (skl_planar_formats); + } }else { - formats = skl_plane_formats; - num_formats =ARRAY_SIZE (skl_plane_formats); +if (icl_is_hdr_plane (plane)) { + formats = icl_hdr_plane_formats; + num_formats =ARRAY_SIZE (icl_hdr_plane_formats); + }else { + formats = skl_plane_formats; + num_formats =ARRAY_SIZE (skl_plane_formats); + } } plane->has_ccs =skl_plane_has_ccs (dev_priv, pipe, plane_id);@@ -2074,6 +2138,12 @@ skl_universal_plane_create(struct drm_i915_private *dev_priv,BIT (DRM_MODE_BLEND_PREMULTI ) | BIT(DRM_MODE_BLEND_COVERAGE)); +if (icl_is_hdr_plane (plane)) +drm_plane_create_pixel_normalize_range_property (&plane->base, +BIT (DRM_PIXEL_NORMALIZE_RANGE_0_1 ) | +BIT (DRM_PIXEL_NORMALIZE_RANGE_0_255 ), +DRM_PIXEL_NORMALIZE_RANGE_0_1 ); +drm_plane_helper_add (&plane->base, &intel_plane_helper_funcs);return plane;Signature · 1 line
2.7.4
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 3/3] drm/i915: Implement half float formats and pixel normalize property
Quoted 9 lines · Ville Syrjälä
Quoted 6 lines · Kevin Strasser
64 bpp half float formats are supported on hdr planes only and are subject to the following restrictions:
* 90/270 rotation not supported
* Yf Tiling not supported
* Frame Buffer Compression not supported
* Color Keying not supportedIs scaling supported? It wasn't on earlier platforms.
Yes, ICL does support scaling for these formats.
Thanks,
Kevin
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✗ Fi.CI.BAT: failure for Support 64 bpp half float formats
In reply to Kevin Strasser== Series Details ==
== Series Details ==
Series: Support 64 bpp half float formats
URL : https://patchwork.freedesktop.org/series/53212/
State : failure
== Summary ==
Applying: drm/fourcc: Add 64 bpp half float formats
Applying: drm: Add optional PIXEL_NORMALIZE_RANGE property to drm_plane
Applying: drm/i915: Implement half float formats and pixel normalize property Using index info to reconstruct a base tree...
M drivers/gpu/drm/i915/intel_display.c
M drivers/gpu/drm/i915/intel_drv.h
M drivers/gpu/drm/i915/intel_sprite.c
Falling back to patching base and 3-way merge...
Auto-merging drivers/gpu/drm/i915/intel_sprite.c
CONFLICT (content): Merge conflict in drivers/gpu/drm/i915/intel_sprite.c Auto-merging drivers/gpu/drm/i915/intel_drv.h
Auto-merging drivers/gpu/drm/i915/intel_display.c
error: Failed to merge in the changes.
Patch failed at 0003 drm/i915: Implement half float formats and pixel normalize property Use 'git am --show-current-patch' to see the failed patch
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 5 lines · Kevin Strasser
This series defines new formats and adds a plane property to be used for floating point framebuffer content. Implementation is then added to i915.
I have shared an IGT branch which adds test coverage for the new formats:
https://github.com/strassek/xorg-intel-gpu-tools/tree/fp16
Looks about similar as what I had written. I wrote my half<->full conversion thing from scratch which probably means it has more rounding errors and whatnot. The speed of mine wasn't exactly stellar and looks like your version probably has the same issue. So I was actually thinking of using the sse<something> instructions meant for this could provide a nice speedup. I guess we might want the pure c version as a backup though. Hmm. Now I also seem to recall that I noticed there being a compiler intrinsic even for single value half<->full precision conversion. Did you look into using that (if I didn't imagine it)?
BTW I just rebased my fp16 for pre-icl platforms:git://github.com/vsyrjala/linux.git fp16_scanout_2
Apart from the ivb/hsw w/a there isn't all that much unexpected when it comes to fp16 on those platforms either.
Quoted 28 lines
Kevin Strasser (3):
drm/fourcc: Add 64 bpp half float formats
drm: Add optionalPIXEL_NORMALIZE_RANGEproperty todrm_planedrm/i915: Implement half float formats and pixel normalize propertydrivers/gpu/drm/drm_atomic.c | 2 + drivers/gpu/drm/drm_atomic_uapi.c | 4 ++ drivers/gpu/drm/drm_color_mgmt.c | 67 +++++++++++++++++++++++ drivers/gpu/drm/drm_crtc_internal.h | 1 + drivers/gpu/drm/drm_fourcc.c | 4 ++ drivers/gpu/drm/i915/i915_reg.h | 15 ++++- drivers/gpu/drm/i915/intel_display.c | 47 ++++++++++++++++ drivers/gpu/drm/i915/intel_drv.h | 5 ++ drivers/gpu/drm/i915/intel_sprite.c | 82 ++++++++++++++++++++++++++-- include/drm/drm_color_mgmt.h | 9 +++ include/drm/drm_fourcc.h | 3 + include/drm/drm_plane.h | 14 +++++ include/uapi/drm/drm_fourcc.h | 6 ++ 13 files changed, 252 insertions(+), 7 deletions(-)Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 7 lines · Ville Syrjälä
Quoted 5 lines · Kevin Strasser
This series defines new formats and adds a plane property to be used for floating point framebuffer content. Implementation is then added to i915.
I have shared an IGT branch which adds test coverage for the new formats:
https://github.com/strassek/xorg-intel-gpu-tools/tree/fp16
Quoted 10 lines
Looks about similar as what I had written. I wrote my half<->full conversion thing from scratch which probably means it has more rounding errors and whatnot. The speed of mine wasn't exactly stellar and looks like your version probably has the same issue. So I was actually thinking of using the sse<something> instructions meant for this could provide a nice speedup. I guess we might want the pure c version as a backup though. Hmm. Now I also seem to recall that I noticed there being a compiler intrinsic even for single value half<->full precision conversion. Did you look into using that (if I didn't imagine it)?
You are thinking of vcvtps2ph and vcvtph2ps, I haven't yet had a chance to give them a try, but I agree it seems like a good idea.
Quoted 5 lines
BTW I just rebased my fp16 for pre-icl platforms:
git://github.com/vsyrjala/linux.gitfp16_scanout_2Apart from the ivb/hsw w/a there isn't all that much unexpected when it comes to fp16 on those platforms either.
I don't mean to step on your toes with this series, were you waiting for / working on a real usecase before pushing that code?
Thanks,
Kevin
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 22 lines · Kevin
Quoted 7 lines · Ville Syrjälä
Quoted 5 lines · Kevin Strasser
This series defines new formats and adds a plane property to be used for floating point framebuffer content. Implementation is then added to i915.
I have shared an IGT branch which adds test coverage for the new formats:
https://github.com/strassek/xorg-intel-gpu-tools/tree/fp16Quoted 10 lines
Looks about similar as what I had written. I wrote my half<->full conversion thing from scratch which probably means it has more rounding errors and whatnot. The speed of mine wasn't exactly stellar and looks like your version probably has the same issue. So I was actually thinking of using the sse<something> instructions meant for this could provide a nice speedup. I guess we might want the pure c version as a backup though. Hmm. Now I also seem to recall that I noticed there being a compiler intrinsic even for single value half<->full precision conversion. Did you look into using that (if I didn't imagine it)?
You are thinking of vcvtps2ph and vcvtph2ps, I haven't yet had a chance to give them a try, but I agree it seems like a good idea.
Quoted 8 lines
Quoted 5 lines
BTW I just rebased my fp16 for pre-icl platforms:
git://github.com/vsyrjala/linux.gitfp16_scanout_2Apart from the ivb/hsw w/a there isn't all that much unexpected when it comes to fp16 on those platforms either.
I don't mean to step on your toes with this series, were you waiting for / working on a real usecase before pushing that code?
I pretty much just did it so that I could test >10bpc gamma LUTs. But I got sidetracked by other things so I didn't really get even that far. Also another problem is that igt depends on cairo which didn't support rendering at >10bpc, so I couldn't really test that stuff properly even if I wanted to. Maarten has patches to wire up floats into cairo but I think he just said that it still kinda uses 8bpc precision only :(
Anyways, the fact that you did icl and I did pre-icl is pretty good division of labour. Sometimes things work out by accident :)
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx