Add Multi Segment Gamma Support
Subsystems (3)
v2 cover Add Multi Segment Gamma Support
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1. GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.
2. GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode property
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structure
drivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465 ++++++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)
Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH v2 1/7 drm: Add gamma mode caps property
From: Ville Syrjälä <[email protected]>
Add a gamma mode capability property to enable various kind of gamma modes supported by platforms like: Interpolated, Split, Multi Segmented etc. Userspace can get this property and should be able to get the platform capabilties wrt various gamma modes possible and the possible ranges.
It can then create the LUT and send it to driver using another gamma mode property as blob.
--- drivers/gpu/drm/drm_atomic_uapi.c | 3 ++ drivers/gpu/drm/drm_color_mgmt.c | 78 +++++++++++++++++++++++++++++++++++++++ include/drm/drm_color_mgmt.h | 8 ++++ include/drm/drm_crtc.h | 7 ++++ include/drm/drm_mode_config.h | 5 +++ include/uapi/drm/drm_mode.h | 38 +++++++++++++++++++ 6 files changed, 139 insertions(+)
drivers/gpu/drm/drm_atomic_uapi.c +3 -0
@@ -495,6 +495,9 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc, *val = (state->mode_blob) ? state->mode_blob->base.id :0 ; else if (property == config->prop_vrr_enabled) *val = state->vrr_enabled; + else if (property == config->gamma_mode_caps_property) + *val = (state->gamma_mode_caps) ? + state->gamma_mode_caps->base.id :0 ; else if (property == config->degamma_lut_property) *val = (state->degamma_lut) ? state->degamma_lut->base.id : 0; else if (property == config->ctm_property)
drivers/gpu/drm/drm_color_mgmt.c +78 -0
84 lines not on this page · raw message
include/drm/drm_color_mgmt.h +8 -0
@@ -51,6 +51,14 @@ static inline int drm_color_lut_size(const struct drm_property_blob *blob)return blob->length /sizeof (struct drm_color_lut ); } +int drm_color_create_gamma_mode_caps_property (struct drm_device *dev, +int num_values); +void drm_crtc_attach_gamma_mode_caps_property (struct drm_crtc *crtc); +int drm_color_add_gamma_mode_range (struct drm_device *dev, +const char *name, +const struct drm_color_lut_range *ranges, +size_t length); +enum drm_color_encoding {DRM_COLOR_YCBCR_BT601 ,DRM_COLOR_YCBCR_BT709 ,
include/drm/drm_crtc.h +7 -0
@@ -249,6 +249,13 @@ struct drm_crtc_state {struct drm_property_blob *mode_blob;/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode_caps; + +/** * @degamma_lut: * * Lookup table for converting framebuffer pixel data before apply the
include/drm/drm_mode_config.h +5 -0
@@ -761,6 +761,11 @@ struct drm_mode_config {*/ struct drm_property *content_type_property;/** +* @gamma_mode_property: Optional CRTC property to enumerate and +* select the mode of the crtc gamma/degmama LUTs. +*/ +struct drm_property *gamma_mode_caps_property; +/** * @degamma_lut_property: Optional CRTC property to set the LUT used to * convert the framebuffer's colors to linear gamma. */
include/uapi/drm/drm_mode.h +38 -0
@@ -629,6 +629,44 @@ struct drm_color_lut {__u16 reserved; }; +/* +* DRM_MODE_LUT_GAMMA|DRM_MODE_LUT_DEGAMMA is legal and means the LUT +* can be used for either purpose, but not simultaneously. To expose +* modes that support gamma and degamma simultaneously the gamma mode +* must declare distinct DRM_MODE_LUT_GAMMA and DRM_MODE_LUT_DEGAMMA +* ranges. +*/ +/* LUT is for gamma (after CTM) */ +#define DRM_MODE_LUT_GAMMA (1 << 0) +/* LUT is for degamma (before CTM) */ +#define DRM_MODE_LUT_DEGAMMA (1 << 1) +/* linearly interpolate between the points */ +#define DRM_MODE_LUT_INTERPOLATE (1 << 2) +/* +* the last value of the previous range is the +* first value of the current range. +*/ +#define DRM_MODE_LUT_REUSE_LAST (1 << 3) +/* the curve must be non-decreasing */ +#define DRM_MODE_LUT_NON_DECREASING (1 << 4) +/* the curve is reflected across origin for negative inputs */ +#define DRM_MODE_LUT_REFLECT_NEGATIVE (1 << 5) +/* the same curve (red) is used for blue and green channels as well */ +#define DRM_MODE_LUT_SINGLE_CHANNEL (1 << 6) + +struct drm_color_lut_range { +/* DRM_MODE_LUT_* */ +__u32 flags; +/* number of points on the curve */ +__u16 count; +/* input/output bits per component */ +__u8 input_bpc, output_bpc; +/* input start/end values */ +__s32 start, end; +/* output min/max values */ +__s32 min, max; +}; +#define DRM_MODE_PAGE_FLIP_EVENT 0x01#define DRM_MODE_PAGE_FLIP_ASYNC 0x02#define DRM_MODE_PAGE_FLIP_TARGET_ABSOLUTE 0x4
Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 1/7] drm: Add gamma mode caps property
Hi Uma.
include/drm/drm_crtc.h struct drm_crtc_state { · line 256
Quoted 13 lines
include/drm/drm_crtc.h +5
-0@@ -249,6 +249,13 @@ struct drm_crtc_state {struct drm_property_blob *mode_blob;/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode_caps;
This looks wrong as "gamma_mode" != "gamma_mode_caps". Could you fix it, and also add some documentation to help the next reader.
include/drm/drm_crtc.h struct drm_crtc_state {
Quoted 13 lines
include/drm/drm_crtc.h +13
-0+/* +* DRM_MODE_LUT_GAMMA|DRM_MODE_LUT_DEGAMMA is legal and means the LUT +* can be used for either purpose, but not simultaneously. To expose +* modes that support gamma and degamma simultaneously the gamma mode +* must declare distinct DRM_MODE_LUT_GAMMA and DRM_MODE_LUT_DEGAMMA +* ranges. +*/ +/* LUT is for gamma (after CTM) */ +#define DRM_MODE_LUT_GAMMA (1 << 0) +/* LUT is for degamma (before CTM) */ +#define DRM_MODE_LUT_DEGAMMA (1 << 1) +/* linearly interpolate between the points */ +#define DRM_MODE_LUT_INTERPOLATE (1 << 2)
Use BIT(2), and same for the other bits above and below?
include/drm/drm_crtc.h
Quoted 11 lines
include/drm/drm_crtc.h +11
-0+/* +* the last value of the previous range is the +* first value of the current range. +*/ +#define DRM_MODE_LUT_REUSE_LAST (1 << 3) +/* the curve must be non-decreasing */ +#define DRM_MODE_LUT_NON_DECREASING (1 << 4) +/* the curve is reflected across origin for negative inputs */ +#define DRM_MODE_LUT_REFLECT_NEGATIVE (1 << 5) +/* the same curve (red) is used for blue and green channels as well */ +#define DRM_MODE_LUT_SINGLE_CHANNEL (1 << 6)
Sam
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 1/7] drm: Add gamma mode caps property
Quoted 22 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Sam Ravnborg
Sent: Tuesday, April 2, 2019 12:03 AM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [v2 1/7] drm: Add gamma mode caps propertyHi Uma.
Quoted 7 lines
include/drm/drm_crtc.h index renamed from include/drm/drm_crtc.h +0
-058ad983..cdfda90 100644
include/drm/drm_crtc.h +0
-0@@ -249,6 +249,13 @@ struct drm_crtc_state {struct drm_property_blob *mode_blob;
Quoted 9 lines
include/drm/drm_crtc.h struct drm_crtc_state { · line 256
Quoted 6 lines
include/drm/drm_crtc.h +5
-0/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode_caps;This looks wrong as "
gamma_mode" != "gamma_mode_caps".
Could you fix it, and also add some documentation to help the next reader.
Hi Sam,
Thanks for the review.
Sure will update the naming and add more documentation once we have the design agreement on this. Idea is to converge to just one property.
Regards,
Uma Shankar
Quoted 8 lines
include/drm/drm_crtc.h struct drm_crtc_state { · line 258
Quoted 3 lines
include/drm/drm_crtc.h +2
-0+/* +* DRM_MODE_LUT_GAMMA|DRM_MODE_LUT_DEGAMMA is legal and means the LUT
include/drm/drm_crtc.h struct drm_crtc_state { · line 261
Quoted 3 lines
include/drm/drm_crtc.h +3
-0+* can be used for either purpose, but not simultaneously. To expose +* modes that support gamma and degamma simultaneously the gamma mode +* must declare distinct DRM_MODE_LUT_GAMMA and
DRM_MODE_LUT_DEGAMMA
Quoted 10 lines
include/drm/drm_crtc.h struct drm_crtc_state {
Quoted 8 lines
include/drm/drm_crtc.h +8
-0+* ranges. +*/ +/* LUT is for gamma (after CTM) */ +#define DRM_MODE_LUT_GAMMA (1 << 0) +/* LUT is for degamma (before CTM) */ #define DRM_MODE_LUT_DEGAMMA (1 +<< 1) +/* linearly interpolate between the points */ #define +DRM_MODE_LUT_INTERPOLATE (1 << 2)Use BIT(2), and same for the other bits above and below?
Sure, will update this.
Quoted 14 lines
include/drm/drm_crtc.h
Quoted 11 lines
include/drm/drm_crtc.h +11
-0+/* +* the last value of the previous range is the +* first value of the current range. +*/ +#define DRM_MODE_LUT_REUSE_LAST (1 << 3) +/* the curve must be non-decreasing */ #define +DRM_MODE_LUT_NON_DECREASING (1 << 4) +/* the curve is reflected across origin for negative inputs */ +#define DRM_MODE_LUT_REFLECT_NEGATIVE (1 << 5) +/* the same curve (red) is used for blue and green channels as well +*/ #define DRM_MODE_LUT_SINGLE_CHANNEL (1 << 6)Sam
Quoted 4 lines
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH v2 2/7 drm/i915: Define color lut range structure
In reply to Uma ShankarFrom: Ville Syrjälä <[email protected]>
This defines the color lut ranges for 10bit and multi segmented gamma range for ICL.
--- drivers/gpu/drm/i915/intel_color.c | 301 ++++++++++++++++++++++++++++++++++++- 1 file changed, 297 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/intel_color.c +297 -4
@@ -1083,9 +1083,279 @@ static int icl_color_check(struct intel_crtc_state *crtc_state)return 0 ; } +enum { +I9XX_LUT_SIZE_8BIT =256 , +I9XX_LUT_SIZE_10BIT =129 , + +ILK_LUT_SIZE_10BIT =1024 , +ILK_LUT_SIZE_12BIT =513 , + +IVB_LUT_SIZE_SPLIT =512 , + +CHV_LUT_SIZE_CGM_DEGAMMA =65 , +CHV_LUT_SIZE_CGM_GAMMA =257 , +}; + +#define I9XX_GAMMA_8 \ + { \ + .flags = DRM_MODE_LUT_GAMMA, \ + .count = 256, \ + .input_bpc = 8, .output_bpc = 8, \ + .start = 0, .end = (1 << 8) - 1, \ + .min = 0, .max = (1 << 8) - 1, \ + } + +static const struct drm_color_lut_range i9xx_gamma_8[] = { +I9XX_GAMMA_8 , +}; + +static const struct drm_color_lut_range i9xx_gamma_10_slope[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =129 , + .input_bpc =10 , .output_bpc =10 , + .start =0 , .end =1 <<10 , + .min =0 , .max = (1 <<10 ) -1 , + }, +}; + +#define I965_GAMMA_10 \ + { \ + .flags = (DRM_MODE_LUT_GAMMA | \ + DRM_MODE_LUT_INTERPOLATE | \ + DRM_MODE_LUT_NON_DECREASING), \ + .count = 128, \ + .input_bpc = 10, .output_bpc = 16, \ + .start = 0, .end = (1 << 10) - (1 << 10) / 128, \ + .min = 0, .max = (1 << 16) - 1, \ + }, \ + /* PIPEGCMAX */ \ + { \ + .flags = (DRM_MODE_LUT_GAMMA | \ + DRM_MODE_LUT_INTERPOLATE | \ + DRM_MODE_LUT_REUSE_LAST | \ + DRM_MODE_LUT_NON_DECREASING), \ + .count = 1, \ + .input_bpc = 10, .output_bpc = 16, \ + .start = (1 << 10) - (1 << 10) / 128, .end = 1 << 10, \ + .min = 0, .max = 1 << 16, \ + } + +static const struct drm_color_lut_range i965_gamma_10[] = { +I965_GAMMA_10 , +}; + +static const struct drm_color_lut_range ilk_gamma_degamma_8[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_DEGAMMA ), + .count =256 , + .input_bpc =8 , .output_bpc =8 , + .start =0 , .end = (1 <<8 ) -1 , + .min =0 , .max = (1 <<8 ) -1 , + }, +}; + +static const struct drm_color_lut_range ilk_gamma_degamma_10[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_DEGAMMA ), + .count =1024 , + .input_bpc =10 , .output_bpc =10 , + .start =0 , .end = (1 <<10 ) -1 , + .min =0 , .max = (1 <<10 ) -1 , + }, +}; + +static const struct drm_color_lut_range ilk_gamma_degamma_12p4[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_DEGAMMA | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =512 , + .input_bpc =12 , .output_bpc =16 , + .start =0 , .end = (1 <<12 ) - (1 <<12 ) /512 , + .min =0 , .max = (1 <<16 ) -1 , + }, +/* PIPEGCMAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_DEGAMMA | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =10 , .output_bpc =16 , + .start = (1 <<12 ) - (1 <<12 ) /512 , .end =1 <<12 , + .min =0 , .max =1 <<16 , + }, +}; + +static const struct drm_color_lut_range glk_gamma_10[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE ), + .count =1024 , + .input_bpc =10 , .output_bpc =10 , + .start =0 , .end = (1 <<10 ) -1 , + .min =0 , .max = (1 <<10 ) -1 , + }, +/* PAL_EXT_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =10 , .output_bpc =16 , + .start =1 <<10 , .end =3 <<10 , + .min =0 , .max = (8 <<16 ) -1 , + }, +/* PAL_EXT2_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =10 , .output_bpc =16 , + .start =3 <<12 , .end =7 <<12 , + .min =0 , .max = (8 <<16 ) -1 , + }, +}; + +/* FIXME input bpc? */ +static const struct drm_color_lut_range glk_gamma_12p4[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =512 , + .input_bpc =16 , .output_bpc =16 , + .start =0 , .end = (1 <<16 ) - (1 <<16 ) /512 , + .min =0 , .max = (1 <<16 ) -1 , + }, +/* PAL_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =16 , .output_bpc =16 , + .start = (1 <<16 ) - (1 <<16 ) /512 , .end =1 <<16 , + .min =0 , .max =1 <<16 , + }, +/* PAL_EXT_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =16 , .output_bpc =16 , + .start =1 <<16 , .end =3 <<16 , + .min =0 , .max = (8 <<16 ) -1 , + }, +/* PAL_EXT2_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =16 , .output_bpc =16 , + .start =3 <<16 , .end =7 <<16 , + .min =0 , .max = (8 <<16 ) -1 , + }, +}; + +/* FIXME input bpc? */ +static const struct drm_color_lut_range icl_multi_seg_gamma[] = { +/* segment 1 aka. super fine segment */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =9 , + .input_bpc =24 , .output_bpc =16 , + .start =0 , .end = (1 <<24 ) / (128 *256 ), + .min =0 , .max = (1 <<16 ) -1 , + }, +/* segment 2 aka. fine segment */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =257 , + .input_bpc =24 , .output_bpc =16 , + .start =0 , .end = (1 <<24 ) /128 , + .min =0 , .max = (1 <<16 ) -1 , + }, +/* segment 3 aka. coarse segment */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =257 , + .input_bpc =24 , .output_bpc =16 , + .start =0 , .end = (1 <<24 ) - (1 <<24 ) /256 , + .min =0 , .max = (1 <<16 ) -1 , + }, +/* segment 3 aka. coarse segment / PAL_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =24 , .output_bpc =16 , + .start = (1 <<24 ) - (1 <<24 ) /256 , .end =1 <<24 , + .min =0 , .max =1 <<16 , + }, +/* PAL_EXT_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =24 , .output_bpc =16 , + .start =1 <<24 , .end =3 <<24 , + .min =0 , .max = (8 <<16 ) -1 , + }, +/* PAL_EXT2_GC_MAX */ + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_REFLECT_NEGATIVE | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_REUSE_LAST | +DRM_MODE_LUT_NON_DECREASING ), + .count =1 , + .input_bpc =24 , .output_bpc =16 , + .start =3 <<24 , .end =7 <<24 , + .min =0 , .max = (8 <<16 ) -1 , + }, +}; +void intel_color_init (struct intel_crtc *crtc) {struct drm_i915_private *dev_priv =to_i915 (crtc->base.dev); +int degamma_lut_size, gamma_lut_size;drm_mode_crtc_set_gamma_size (&crtc->base,256 );
@@ -1100,14 +1370,37 @@ void intel_color_init(struct intel_crtc *crtc) dev_priv->display.load_luts = i9xx_load_luts; } }else { -if (INTEL_GEN (dev_priv) >=11 ) +if (INTEL_GEN (dev_priv) >=11 ) { dev_priv->display.color_check = icl_color_check; -else if (INTEL_GEN (dev_priv) >=10 ||IS_GEMINILAKE (dev_priv)) + +/* don't advertize the >= 1.0 entries */ + degamma_lut_size =0 ; + gamma_lut_size =ILK_LUT_SIZE_10BIT ; + +drm_color_add_gamma_mode_range (&dev_priv->drm, +"8bit gamma" , + i9xx_gamma_8, +sizeof (i9xx_gamma_8)); +drm_color_add_gamma_mode_range (&dev_priv->drm, +"10bit gamma" , + glk_gamma_10, +sizeof (glk_gamma_10)); +drm_color_add_gamma_mode_range (&dev_priv->drm, +"interpolated gamma" , + glk_gamma_12p4, +sizeof (glk_gamma_12p4)); +drm_color_add_gamma_mode_range (&dev_priv->drm, +"multi-segmented gamma" , + icl_multi_seg_gamma, +sizeof (icl_multi_seg_gamma)); + }else if (INTEL_GEN (dev_priv) >=10 || +IS_GEMINILAKE (dev_priv)) { dev_priv->display.color_check = glk_color_check; -else if (INTEL_GEN (dev_priv) >=8 ) + }else if (INTEL_GEN (dev_priv) >=8 ) { dev_priv->display.color_check = bdw_color_check; -else + }else { dev_priv->display.color_check = ilk_color_check; + }if (INTEL_GEN (dev_priv) >=9 ) dev_priv->display.color_commit = skl_color_commit;
Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 2/7] drm/i915: Define color lut range structure
Quoted 46 lines · Uma Shankar
From: Ville Syrjälä <[email protected]>
This defines the color lut ranges for 10bit and multi segmented gamma range for ICL.
Signed-off-by Ville Syrjälä <[email protected]>Signed-off-by Uma Shankar <[email protected]>--- drivers/gpu/drm/i915/intel_color.c | 301 ++++++++++++++++++++++++++++++++++++- 1 file changed, 297 insertions(+), 4 deletions(-)drivers/gpu/drm/i915/intel_color.c +27
-0@@ -1083,9 +1083,279 @@ static int icl_color_check(struct intel_crtc_state *crtc_state)return 0 ; } +enum { +I9XX_LUT_SIZE_8BIT =256 , +I9XX_LUT_SIZE_10BIT =129 , + +ILK_LUT_SIZE_10BIT =1024 , +ILK_LUT_SIZE_12BIT =513 , + +IVB_LUT_SIZE_SPLIT =512 , + +CHV_LUT_SIZE_CGM_DEGAMMA =65 , +CHV_LUT_SIZE_CGM_GAMMA =257 , +}; + +#define I9XX_GAMMA_8 \ + { \ + .flags = DRM_MODE_LUT_GAMMA, \ + .count = 256, \ + .input_bpc = 8, .output_bpc = 8, \ + .start = 0, .end = (1 << 8) - 1, \ + .min = 0, .max = (1 << 8) - 1, \ + } + +static const struct drm_color_lut_range i9xx_gamma_8[] = { +I9XX_GAMMA_8 , +}; + +static const struct drm_color_lut_range i9xx_gamma_10_slope[] = {
drivers/gpu/drm/i915/intel_color.c static int icl_color_check(struct intel_crtc_state *crtc_state) · line 1122
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +10
-0+ { + .flags = (DRM_MODE_LUT_GAMMA | + DRM_MODE_LUT_INTERPOLATE | + DRM_MODE_LUT_NON_DECREASING), + .count = 129, + .input_bpc = 10, .output_bpc = 10, + .start = 0, .end = 1 << 10, + .min = 0, .max = (1 << 10) - 1, + }, +};
Step 1 should probably be to advertise these for the gamma modes we currently have, on every platform.
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 2/7] drm/i915: Define color lut range structure
Quoted 68 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 3:39 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; Lankhorst, Maarten <[email protected]>; Syrjala, Ville <[email protected]>; Sharma, Shashank <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Roper, Matthew D <[email protected]>
Subject: Re: [v2 2/7] drm/i915: Define color lut range structuredrivers/gpu/drm/i915/intel_color.c static int icl_color_check(struct intel_crtc_state
Quoted 18 lines · Uma Shankar
From: Ville Syrjälä <[email protected]>
This defines the color lut ranges for 10bit and multi segmented gamma range for ICL.
Signed-off-by Ville Syrjälä <[email protected]>Signed-off-by Uma Shankar <[email protected]>--- drivers/gpu/drm/i915/intel_color.c | 301 ++++++++++++++++++++++++++++++++++++- 1 file changed, 297 insertions(+), 4 deletions(-)drivers/gpu/drm/i915/intel_color.c +0
-0@@ -1083,9 +1083,279 @@ static int icl_color_check(struct intel_crtc_state
*crtc_state)Quoted 35 lines
drivers/gpu/drm/i915/intel_color.c +32
-0return 0 ; } +enum { +I9XX_LUT_SIZE_8BIT =256 , +I9XX_LUT_SIZE_10BIT =129 , + +ILK_LUT_SIZE_10BIT =1024 , +ILK_LUT_SIZE_12BIT =513 , + +IVB_LUT_SIZE_SPLIT =512 , + +CHV_LUT_SIZE_CGM_DEGAMMA =65 , +CHV_LUT_SIZE_CGM_GAMMA =257 , +}; + +#define I9XX_GAMMA_8 \ + { \ + .flags = DRM_MODE_LUT_GAMMA, \ + .count = 256, \ + .input_bpc = 8, .output_bpc = 8, \ + .start = 0, .end = (1 << 8) - 1, \ + .min = 0, .max = (1 << 8) - 1, \ + } + +static const struct drm_color_lut_range i9xx_gamma_8[] = { +I9XX_GAMMA_8 , +}; + +static const struct drm_color_lut_range i9xx_gamma_10_slope[] = { + { + .flags = (DRM_MODE_LUT_GAMMA | +DRM_MODE_LUT_INTERPOLATE | +DRM_MODE_LUT_NON_DECREASING ), + .count =129 ,
Quoted 8 lines
drivers/gpu/drm/i915/intel_color.c static int icl_color_check(struct intel_crtc_state *crtc_state) · line 1122
Quoted 5 lines
drivers/gpu/drm/i915/intel_color.c +5
-0+ .input_bpc = 10, .output_bpc = 10, + .start = 0, .end = 1 << 10, + .min = 0, .max = (1 << 10) - 1, + }, +};Step 1 should probably be to advertise these for the gamma modes we currently have, on every platform.
Ok, focussed here mainly on multi-segmented gamma. But will add all the gamma modes supported on various platforms.
Quoted 3 lines
--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH v2 3/7 drm: Add gamma mode property
In reply to Uma ShankarAdd Gamma Mode property to set the gamma mode (Interploated, Split, Multi Segmented etc) from the list obtained through the gamma mode caps property.
Create the blob and send to driver for programming the luts to the appropriate registers and setting the chosen gamma mode.
--- drivers/gpu/drm/drm_atomic_uapi.c | 10 ++++++++++ drivers/gpu/drm/drm_color_mgmt.c | 32 ++++++++++++++++++++++++++++++++ include/drm/drm_color_mgmt.h | 3 +++ include/drm/drm_crtc.h | 7 +++++++ include/drm/drm_mode_config.h | 5 +++++ include/uapi/drm/drm_mode.h | 11 +++++++++++ 6 files changed, 68 insertions(+)
drivers/gpu/drm/drm_atomic_uapi.c +10 -0
@@ -459,6 +459,14 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc, &replaced); state->color_mgmt_changed |= replaced;return ret; + }else if (property == config->gamma_mode_property) { + ret =drm_atomic_replace_property_blob_from_id (dev, + &state->gamma_mode, + val, +-1 ,sizeof (struct drm_color_mode_lut ), + &replaced); + state->color_mgmt_changed |= replaced; +return ret; }else if (property == config->prop_out_fence_ptr) {s32 __user *fence_ptr =u64_to_user_ptr (val);
@@ -498,6 +506,8 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc,else if (property == config->gamma_mode_caps_property) *val = (state->gamma_mode_caps) ? state->gamma_mode_caps->base.id :0 ; +else if (property == config->gamma_mode_property) + *val = (state->gamma_mode) ? state->gamma_mode->base.id :0 ;else if (property == config->degamma_lut_property) *val = (state->degamma_lut) ? state->degamma_lut->base.id :0 ;else if (property == config->ctm_property)
drivers/gpu/drm/drm_color_mgmt.c +32 -0
@@ -176,6 +176,38 @@ void drm_crtc_enable_color_mgmt(struct drm_crtc *crtc, }EXPORT_SYMBOL (drm_crtc_enable_color_mgmt); +void drm_crtc_attach_gamma_mode_property (struct drm_crtc *crtc) +{ +struct drm_device *dev = crtc->dev; +struct drm_mode_config *config = &dev->mode_config; + +if (!config->gamma_mode_property) +return ; + +drm_object_attach_property (&crtc->base, + config->gamma_mode_property,0 ); +} +EXPORT_SYMBOL (drm_crtc_attach_gamma_mode_property); + +int drm_color_create_gamma_mode_property (struct drm_device *dev, +int num_values) +{ +struct drm_mode_config *config = &dev->mode_config; +struct drm_property *prop; + + prop =drm_property_create (dev, +DRM_MODE_PROP_BLOB | +DRM_MODE_PROP_ATOMIC , +"GAMMA_MODE" , num_values); +if (!prop) +return -ENOMEM ; + + config->gamma_mode_property = prop; + +return 0 ; +} +EXPORT_SYMBOL (drm_color_create_gamma_mode_property); +void drm_crtc_attach_gamma_mode_caps_property (struct drm_crtc *crtc) {struct drm_device *dev = crtc->dev;
include/drm/drm_color_mgmt.h +3 -0
@@ -54,6 +54,9 @@ static inline int drm_color_lut_size(const struct drm_property_blob *blob)int drm_color_create_gamma_mode_caps_property (struct drm_device *dev,int num_values);void drm_crtc_attach_gamma_mode_caps_property (struct drm_crtc *crtc); +int drm_color_create_gamma_mode_property (struct drm_device *dev, +int num_values); +void drm_crtc_attach_gamma_mode_property (struct drm_crtc *crtc); int drm_color_add_gamma_mode_range(struct drm_device *dev, const char *name, const struct drm_color_lut_range *ranges,
include/drm/drm_crtc.h +7 -0
@@ -256,6 +256,13 @@ struct drm_crtc_state {struct drm_property_blob *gamma_mode_caps;/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode; + +/** * @degamma_lut: * * Lookup table for converting framebuffer pixel data before apply the
include/drm/drm_mode_config.h +5 -0
@@ -766,6 +766,11 @@ struct drm_mode_config {*/ struct drm_property *gamma_mode_caps_property;/** +* @gamma_mode_property: Optional CRTC property to enumerate and +* select the mode of the crtc gamma/degmama LUTs. +*/ +struct drm_property *gamma_mode_property; +/** * @degamma_lut_property: Optional CRTC property to set the LUT used to * convert the framebuffer's colors to linear gamma. */
include/uapi/drm/drm_mode.h +11 -0
@@ -667,6 +667,17 @@ struct drm_color_lut_range {__s32 min, max; }; +struct drm_color_mode_lut { +/* DRM_MODE_LUT_* */ +__u32 flags; +/* number of points on the curve */ +__u32 count; +/* Name of Gamma Mode */ +char name[DRM_PROP_NAME_LEN ]; +/* Pointer to Lut elements */ +__u64 lut; +}; +#define DRM_MODE_PAGE_FLIP_EVENT 0x01#define DRM_MODE_PAGE_FLIP_ASYNC 0x02#define DRM_MODE_PAGE_FLIP_TARGET_ABSOLUTE 0x4
Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 3/7] drm: Add gamma mode property
Hi Uma.
include/drm/drm_crtc.h struct drm_crtc_state { · line 263
Quoted 11 lines
include/drm/drm_crtc.h +5
-0@@ -256,6 +256,13 @@ struct drm_crtc_state {struct drm_property_blob *gamma_mode_caps;/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode;
HEre the name matches, but please add documentation too.
include/drm/drm_crtc.h struct drm_crtc_state {
Quoted 11 lines
include/drm/drm_crtc.h +10
-0+struct drm_color_mode_lut { +/* DRM_MODE_LUT_* */ +__u32 flags; +/* number of points on the curve */ +__u32 count; +/* Name of Gamma Mode */ +char name[DRM_PROP_NAME_LEN ]; +/* Pointer to Lut elements */ +__u64 lut; +};
From an alignment point of view the __u64 should come before the char name[]. But the above may be fine depending on DRM_PROP_NAME_LEN
Sam
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
RE: [v2 3/7] drm: Add gamma mode property
Quoted 19 lines
-----Original Message-----
From: Sam Ravnborg [mailto:[email protected]]
Sent: Tuesday, April 2, 2019 12:07 AM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [v2 3/7] drm: Add gamma mode propertyHi Uma.
Quoted 5 lines
include/drm/drm_crtc.h +0
-0@@ -256,6 +256,13 @@ struct drm_crtc_state {struct drm_property_blob *gamma_mode_caps;
Quoted 8 lines
include/drm/drm_crtc.h struct drm_crtc_state { · line 263
Quoted 6 lines
include/drm/drm_crtc.h +5
-0/** +* @gamma_mode: +* +* FIXME +*/ +struct drm_property_blob *gamma_mode;HEre the name matches, but please add documentation too.
Sure, will add documentation explaining how this should get used once we have the agreement on design and approach.
Quoted 13 lines
include/drm/drm_crtc.h struct drm_crtc_state {
Quoted 11 lines
include/drm/drm_crtc.h +10
-0+struct drm_color_mode_lut { +/* DRM_MODE_LUT_* */ +__u32 flags; +/* number of points on the curve */ +__u32 count; +/* Name of Gamma Mode */ +char name[DRM_PROP_NAME_LEN ]; +/* Pointer to Lut elements */ +__u64 lut; +};From an alignment point of view the
__u64should come before the char name[]. But the above may be fine depending onDRM_PROP_NAME_LEN
Yeah, but can re-order this. Thanks for the review.
Regards,
Uma Shankar
Quoted 2 lines
Sam
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH v2 4/7 drm/i915/icl: Add register definitions for Multi Segmented gamma
In reply to Uma ShankarAdd macros to define multi segmented gamma registers
--- drivers/gpu/drm/i915/i915_reg.h | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +17 -0
@@ -7218,6 +7218,7 @@ enum {#define GAMMA_MODE_MODE_10BIT (1 << 0)#define GAMMA_MODE_MODE_12BIT (2 << 0)#define GAMMA_MODE_MODE_SPLIT (3 << 0) +#define GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED (3 << 0)/* DMC/CSR */ #define CSR_PROGRAM (i) _MMIO(0x80000 + (i) * 4)
@@ -10157,6 +10158,22 @@ enum skl_power_gate {#define PRE_CSC_GAMC_INDEX (pipe) _MMIO_PIPE(pipe, _PRE_CSC_GAMC_INDEX_A, _PRE_CSC_GAMC_INDEX_B)#define PRE_CSC_GAMC_DATA (pipe) _MMIO_PIPE(pipe, _PRE_CSC_GAMC_DATA_A, _PRE_CSC_GAMC_DATA_B) +/* Add registers for Gen11 Multi Segmented Gamma Mode */ +#define _PAL_PREC_MULTI_SEG_INDEX_A 0x4A408 +#define _PAL_PREC_MULTI_SEG_INDEX_B 0x4AC08 +#define PAL_PREC_MULTI_SEGMENT_AUTO_INCREMENT BIT(15) +#define PAL_PREC_MULTI_SEGMENT_INDEX_VALUE_MASK (0x1f << 0) + +#define _PAL_PREC_MULTI_SEG_DATA_A 0x4A40C +#define _PAL_PREC_MULTI_SEG_DATA_B 0x4AC0C + +#define PREC_PAL_MULTI_SEG_INDEX (pipe) _MMIO_PIPE(pipe, \ + _PAL_PREC_MULTI_SEG_INDEX_A, \ + _PAL_PREC_MULTI_SEG_INDEX_B) +#define PREC_PAL_MULTI_SEG_DATA (pipe) _MMIO_PIPE(pipe, \ + _PAL_PREC_MULTI_SEG_DATA_A, \ + _PAL_PREC_MULTI_SEG_DATA_B) +/* pipe CSC & degamma/gamma LUTs on CHV */ #define _CGM_PIPE_A_CSC_COEFF01 (VLV_DISPLAY_BASE + 0x67900)#define _CGM_PIPE_A_CSC_COEFF23 (VLV_DISPLAY_BASE + 0x67904)
Signature · 6 lines
1.9.1 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH v2 5/7 drm/i915/icl: Add support for multi segmented gamma mode
In reply to Uma ShankarGen11 introduced a new gamma mode i.e, multi segmented gamma mode. Added support for the same.
--- drivers/gpu/drm/i915/intel_color.c | 161 ++++++++++++++++++++++++++++++++++++- include/drm/drm_crtc.h | 3 + 2 files changed, 161 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_color.c +158 -3
@@ -57,6 +57,12 @@#define ILK_CSC_POSTOFF_LIMITED_RANGE (16 * (1 << 12) / 255) +#define LEGACY_PALETTE_MODE_8BIT BIT(0) +#define PRECISION_PALETTE_MODE_10BIT BIT(1) +#define INTERPOLATED_GAMMA_MODE_12BIT BIT(2) +#define MULTI_SEGMENTED_GAMMA_MODE_12BIT BIT(3) +#define SPLIT_GAMMA_MODE_12BIT BIT(4) +static const u16 ilk_csc_off_zero[3 ] = {};static const u16 ilk_csc_coeff_identity[9 ] = {
@@ -92,6 +98,9 @@ 0x0800, 0x0100, 0x0800, }; +#define GEN11_GET_MS10BITS_OF_LUT (lut) (((lut) >> 6) & 0x3FF) +#define GEN11_GET_LS6BITS_OF_LUT (lut) ((lut) & 0x3F) +static bool lut_is_legacy (const struct drm_property_blob *lut) {return drm_color_lut_size (lut) ==LEGACY_LUT_LENGTH ;
@@ -670,16 +679,149 @@ static void glk_load_luts(const struct intel_crtc_state *crtc_state)bdw_load_gamma_lut (crtc_state,0 ); } +static void icl_program_coarse_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset) +{ +struct intel_crtc *crtc =to_intel_crtc (crtc_state->base.crtc); +struct drm_i915_private *dev_priv =to_i915 (crtc->base.dev); +const struct drm_color_lut *lut = gamma_lut; +enum pipe pipe = crtc->pipe; +u32 i, lut_size, word; + +WARN_ON (offset & ~PAL_PREC_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_INDEX (pipe), + (offset ?PAL_PREC_SPLIT_MODE :0 ) | +PAL_PREC_AUTO_INCREMENT | + offset); + +if (lut && crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { + lut_size =9 +514 ; +for (i =9 ; i < lut_size; i++) { +/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + } + } + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 1.0 +*/ +I915_WRITE (PREC_PAL_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_GC_MAX (pipe,2 ), (1 <<16 )); + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 1.0 to 3.0 +*/ +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,2 ), (1 <<16 )); + +/* +* Program the gc max 2 register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 3.0 to 7.0 +*/ +if (INTEL_GEN (dev_priv) >=10 ||IS_GEMINILAKE (dev_priv)) { +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,2 ), (1 <<16 )); + } +} + +static void icl_program_fine_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset) +{ +struct drm_crtc *crtc = crtc_state->base.crtc; +struct drm_device *dev = crtc_state->base.crtc->dev; +struct drm_i915_private *dev_priv =to_i915 (dev); +enum pipe pipe =to_intel_crtc (crtc)->pipe; +u32 i, word, lut_size =9 ; + +WARN_ON (offset & ~PAL_PREC_MULTI_SEGMENT_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_MULTI_SEG_INDEX (pipe), + (PAL_PREC_AUTO_INCREMENT | offset)); + +if (gamma_lut) { +struct drm_color_lut *lut = + (struct drm_color_lut *)gamma_lut; + +for (i =0 ; i < lut_size; i++) { +/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + } + } +} + +static void icl_load_gamma_multi_segmented_lut (const struct intel_crtc_state + *crtc_state,u32 offset) +{ +const struct drm_property_blob *gamma_mode_blob = + crtc_state->base.gamma_mode; +struct drm_color_lut *gamma_lut; +int ret; + +if (gamma_mode_blob) { +struct drm_color_mode_lut *arg = gamma_mode_blob->data; +u64 __user *props_ptr = (u64 __user *)(unsigned long )(arg->lut); + +DRM_INFO ("Gamma Mode=%s\n" , arg->name); + gamma_lut =kmalloc ((sizeof (struct drm_color_lut ) * arg->count), +GFP_KERNEL ); + ret =copy_from_user (gamma_lut, props_ptr, +sizeof (struct drm_color_lut ) * arg->count); + } + +icl_program_fine_segment_lut (crtc_state, gamma_lut,0 ); +icl_program_coarse_segment_lut (crtc_state, gamma_lut,0 ); +} +static void icl_load_luts (const struct intel_crtc_state *crtc_state) {if (crtc_state->base.degamma_lut)glk_load_degamma_lut (crtc_state); -if (crtc_state_is_legacy_gamma (crtc_state)) +if (crtc_state_is_legacy_gamma (crtc_state)) {i9xx_load_luts (crtc_state); -else + }else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { +icl_load_gamma_multi_segmented_lut (crtc_state,0 ); + }else {/* ToDo: Add support for multi segment gamma LUT */ bdw_load_gamma_lut (crtc_state,0 ); + } }static void cherryview_load_luts (const struct intel_crtc_state *crtc_state)
@@ -1034,10 +1176,20 @@ static int glk_color_check(struct intel_crtc_state *crtc_state)return 0 ; } -static u32 icl_gamma_mode (const struct intel_crtc_state *crtc_state) +static u32 icl_gamma_mode (struct intel_crtc_state *crtc_state) { +const struct drm_property_blob *gamma_mode_blob = + crtc_state->base.gamma_mode; +struct drm_color_mode_lut *arg;u32 gamma_mode =0 ; +if (gamma_mode_blob) { + arg = gamma_mode_blob->data; +if (!strcmp (arg->name,"multi-segmented gamma" )) + crtc_state->base.gamma_mode_type = +MULTI_SEGMENTED_GAMMA_MODE_12BIT ; + } +if (crtc_state->base.degamma_lut) gamma_mode |=PRE_CSC_GAMMA_ENABLE ;
@@ -1048,6 +1200,9 @@ static u32 icl_gamma_mode(const struct intel_crtc_state *crtc_state)if (!crtc_state->base.gamma_lut ||crtc_state_is_legacy_gamma (crtc_state)) gamma_mode |=GAMMA_MODE_MODE_8BIT ; +else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) + gamma_mode |=GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED ;else gamma_mode |=GAMMA_MODE_MODE_10BIT ;
include/drm/drm_crtc.h +3 -0
@@ -262,6 +262,9 @@ struct drm_crtc_state {*/ struct drm_property_blob *gamma_mode; +/* Gamma mode type programmed on the pipe */ +u32 gamma_mode_type; +/** * @degamma_lut: *
Signature · 6 lines
1.9.1 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: [v2 5/7] drm/i915/icl: Add support for multi segmented gamma mode
Quoted 27 lines · Uma Shankar
Gen11 introduced a new gamma mode i.e, multi segmented gamma mode. Added support for the same.
Signed-off-by Uma Shankar <[email protected]>--- drivers/gpu/drm/i915/intel_color.c | 161 ++++++++++++++++++++++++++++++++++++- include/drm/drm_crtc.h | 3 + 2 files changed, 161 insertions(+), 3 deletions(-)drivers/gpu/drm/i915/intel_color.c +6
-0@@ -57,6 +57,12 @@#define ILK_CSC_POSTOFF_LIMITED_RANGE (16 * (1 << 12) / 255) +#define LEGACY_PALETTE_MODE_8BIT BIT(0) +#define PRECISION_PALETTE_MODE_10BIT BIT(1) +#define INTERPOLATED_GAMMA_MODE_12BIT BIT(2) +#define MULTI_SEGMENTED_GAMMA_MODE_12BIT BIT(3) +#define SPLIT_GAMMA_MODE_12BIT BIT(4) +static const u16 ilk_csc_off_zero[3 ] = {};static const u16 ilk_csc_coeff_identity[9 ] = {
drivers/gpu/drm/i915/intel_color.c line 102
Quoted 5 lines
drivers/gpu/drm/i915/intel_color.c +2
-0@@ -92,6 +98,9 @@ 0x0800, 0x0100, 0x0800, }; +#define GEN11_GET_MS10BITS_OF_LUT (lut) (((lut) >> 6) & 0x3FF) +#define GEN11_GET_LS6BITS_OF_LUT (lut) ((lut) & 0x3F)
I think this is just the ilk+ 12.4 format. IIRC I had some kind of ilk_lut_12p4_ldw/udw() funcs for these in my branch somewhere. I'd like to see somehting like that in this patch.
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 685
Quoted 12 lines
drivers/gpu/drm/i915/intel_color.c +5
-0+static bool lut_is_legacy (const struct drm_property_blob *lut) {return drm_color_lut_size (lut) ==LEGACY_LUT_LENGTH ;@@ -670,16 +679,149 @@ static void glk_load_luts(const struct intel_crtc_state *crtc_state)bdw_load_gamma_lut (crtc_state,0 ); } +static void icl_program_coarse_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset)
I think to match the current pattern we should have something like:
icl_load_lut_coarse_segment(struct intel_crtc *crtc,
const struct drm_property_blob *blob);
icl_load_lut_fine_segment(struct intel_crtc *crtc,
const struct drm_property_blob *blob);Quoted 41 lines
drivers/gpu/drm/i915/intel_color.c +41
-0+{ +struct intel_crtc *crtc =to_intel_crtc (crtc_state->base.crtc); +struct drm_i915_private *dev_priv =to_i915 (crtc->base.dev); +const struct drm_color_lut *lut = gamma_lut; +enum pipe pipe = crtc->pipe; +u32 i, lut_size, word; + +WARN_ON (offset & ~PAL_PREC_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_INDEX (pipe), + (offset ?PAL_PREC_SPLIT_MODE :0 ) | +PAL_PREC_AUTO_INCREMENT | + offset); + +if (lut && crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { + lut_size =9 +514 ; +for (i =9 ; i < lut_size; i++) { +/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + } + } + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 1.0 +*/ +I915_WRITE (PREC_PAL_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_GC_MAX (pipe,1 ), (1 <<16 ));
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 736
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +10
-0+I915_WRITE (PREC_PAL_GC_MAX (pipe,2 ), (1 <<16 )); + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 1.0 to 3.0 +*/ +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,2 ), (1 <<16 ));
Doubled thing. But we can just drop these and call ivb_load_lut_10_max() from higher up instead.
Quoted 59 lines
drivers/gpu/drm/i915/intel_color.c +59
-0+ +/* +* Program the gc max 2 register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 3.0 to 7.0 +*/ +if (INTEL_GEN (dev_priv) >=10 ||IS_GEMINILAKE (dev_priv)) { +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,2 ), (1 <<16 )); + } +} + +static void icl_program_fine_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset) +{ +struct drm_crtc *crtc = crtc_state->base.crtc; +struct drm_device *dev = crtc_state->base.crtc->dev; +struct drm_i915_private *dev_priv =to_i915 (dev); +enum pipe pipe =to_intel_crtc (crtc)->pipe; +u32 i, word, lut_size =9 ; + +WARN_ON (offset & ~PAL_PREC_MULTI_SEGMENT_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_MULTI_SEG_INDEX (pipe), + (PAL_PREC_AUTO_INCREMENT | offset)); + +if (gamma_lut) { +struct drm_color_lut *lut = + (struct drm_color_lut *)gamma_lut; + +for (i =0 ; i < lut_size; i++) { +/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + } + } +} + +static void icl_load_gamma_multi_segmented_lut (const struct intel_crtc_state + *crtc_state,u32 offset) +{ +const struct drm_property_blob *gamma_mode_blob = + crtc_state->base.gamma_mode; +struct drm_color_lut *gamma_lut; +int ret; +
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 805
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +10
-0+if (gamma_mode_blob) { +struct drm_color_mode_lut *arg = gamma_mode_blob->data; +u64 __user *props_ptr = (u64 __user *)(unsigned long )(arg->lut); + +DRM_INFO ("Gamma Mode=%s\n" , arg->name); + gamma_lut =kmalloc ((sizeof (struct drm_color_lut ) * arg->count), +GFP_KERNEL ); + ret =copy_from_user (gamma_lut, props_ptr, +sizeof (struct drm_color_lut ) * arg->count); + }
What's this doing here?
Quoted 32 lines
drivers/gpu/drm/i915/intel_color.c +13
-3+ +icl_program_fine_segment_lut (crtc_state, gamma_lut,0 ); +icl_program_coarse_segment_lut (crtc_state, gamma_lut,0 ); +} +static void icl_load_luts (const struct intel_crtc_state *crtc_state) {if (crtc_state->base.degamma_lut)glk_load_degamma_lut (crtc_state); -if (crtc_state_is_legacy_gamma (crtc_state)) +if (crtc_state_is_legacy_gamma (crtc_state)) {i9xx_load_luts (crtc_state); -else + }else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { +icl_load_gamma_multi_segmented_lut (crtc_state,0 ); + }else {/* ToDo: Add support for multi segment gamma LUT */ bdw_load_gamma_lut (crtc_state,0 ); + } }static void cherryview_load_luts (const struct intel_crtc_state *crtc_state)@@ -1034,10 +1176,20 @@ static int glk_color_check(struct intel_crtc_state *crtc_state)return 0 ; } -static u32 icl_gamma_mode (const struct intel_crtc_state *crtc_state) +static u32 icl_gamma_mode (struct intel_crtc_state *crtc_state) { + const struct drm_property_blob *gamma_mode_blob =
drivers/gpu/drm/i915/intel_color.c static int glk_color_check(struct intel_crtc_state *crtc_state) · line 1191
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +8
-0+ crtc_state->base.gamma_mode; +struct drm_color_mode_lut *arg;u32 gamma_mode =0 ; +if (gamma_mode_blob) { + arg = gamma_mode_blob->data; +if (!strcmp (arg->name,"multi-segmented gamma" )) + crtc_state->base.gamma_mode_type = +MULTI_SEGMENTED_GAMMA_MODE_12BIT ; + }
With the blob_enum this kind of stuff should disappear I think.
Quoted 35 lines
drivers/gpu/drm/i915/intel_color.c +4
-0+if (crtc_state->base.degamma_lut) gamma_mode |=PRE_CSC_GAMMA_ENABLE ;@@ -1048,6 +1200,9 @@ static u32 icl_gamma_mode(const struct intel_crtc_state *crtc_state)if (!crtc_state->base.gamma_lut ||crtc_state_is_legacy_gamma (crtc_state)) gamma_mode |=GAMMA_MODE_MODE_8BIT ; +else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) + gamma_mode |=GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED ;else gamma_mode |=GAMMA_MODE_MODE_10BIT ;include/drm/drm_crtc.h +3
-0@@ -262,6 +262,9 @@ struct drm_crtc_state {*/ struct drm_property_blob *gamma_mode; +/* Gamma mode type programmed on the pipe */ +u32 gamma_mode_type; +/** * @degamma_lut: * Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [v2 5/7] drm/i915/icl: Add support for multi segmented gamma mode
Quoted 45 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 3:50 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Syrjala, Ville <[email protected]>; [email protected]; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 5/7] drm/i915/icl: Add support for multi segmented gamma modeQuoted 6 lines · Uma Shankar
Gen11 introduced a new gamma mode i.e, multi segmented gamma mode. Added support for the same.
Signed-off-by Uma Shankar <[email protected]>--- drivers/gpu/drm/i915/intel_color.c | 161++++++++++++++++++++++++++++++++++++-
Quoted 24 lines
include/drm/drm_crtc.h | 3 + 2 files changed, 161 insertions(+), 3 deletions(-)drivers/gpu/drm/i915/intel_color.c +6
-0@@ -57,6 +57,12 @@#define ILK_CSC_POSTOFF_LIMITED_RANGE (16 * (1 << 12) / 255) +#define LEGACY_PALETTE_MODE_8BIT BIT(0) +#define PRECISION_PALETTE_MODE_10BIT BIT(1) +#define INTERPOLATED_GAMMA_MODE_12BIT BIT(2) +#define MULTI_SEGMENTED_GAMMA_MODE_12BIT BIT(3) +#define SPLIT_GAMMA_MODE_12BIT BIT(4) +static const u16 ilk_csc_off_zero[3 ] = {}; static const u16 ilk_csc_coeff_identity[9] = { @@ -92,6 +98,9 @@
0x0800,0x0100,0x0800,
};
Quoted 6 lines
drivers/gpu/drm/i915/intel_color.c
Quoted 2 lines
drivers/gpu/drm/i915/intel_color.c +2
-0+#define GEN11_GET_MS10BITS_OF_LUT (lut) (((lut) >> 6) & 0x3FF) #define +GEN11_GET_LS6BITS_OF_LUT (lut) ((lut) &0x3F )I think this is just the ilk+ 12.4 format. IIRC I had some kind of ilk_lut_12p4_ldw/udw() funcs for these in my branch somewhere. I'd like to see somehting like that in this patch.
Sure, I will add those and use the same. Something similar to (i965_lut_10p6_udw)
Quoted 14 lines
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 685
Quoted 11 lines
drivers/gpu/drm/i915/intel_color.c +6
-0+static bool lut_is_legacy (const struct drm_property_blob *lut) { return drm_color_lut_size(lut) == LEGACY_LUT_LENGTH; @@ -670,16 +679,149 @@ static void glk_load_luts(const struct intel_crtc_state *crtc_state) bdw_load_gamma_lut(crtc_state, 0); } +static void icl_program_coarse_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset)I think to match the current pattern we should have something like:
Quoted 4 lines
icl_load_lut_coarse_segment(struct
intel_crtc*crtc,const struct drm_property_blob *blob); icl_load_lut_fine_segment(struct intel_crtc *crtc, const struct drm_property_blob *blob);
Ok, will update this.
Quoted 47 lines
Quoted 18 lines
drivers/gpu/drm/i915/intel_color.c +18
-0+{ +struct intel_crtc *crtc =to_intel_crtc (crtc_state->base.crtc); +struct drm_i915_private *dev_priv =to_i915 (crtc->base.dev); +const struct drm_color_lut *lut = gamma_lut; +enum pipe pipe = crtc->pipe; +u32 i, lut_size, word; + +WARN_ON (offset & ~PAL_PREC_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_INDEX (pipe), + (offset ?PAL_PREC_SPLIT_MODE :0 ) | +PAL_PREC_AUTO_INCREMENT | + offset); + +if (lut && crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { + lut_size =9 +514 ; +for (i =9 ; i < lut_size; i++) {drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state)
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +10
-0+/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 )|
Quoted 18 lines
drivers/gpu/drm/i915/intel_color.c +18
-0+GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_DATA (pipe), word); + } + } + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 1.0 +*/ +I915_WRITE (PREC_PAL_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_GC_MAX (pipe,2 ), (1 <<16 )); + +/* +* Program the max register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values
Quoted 8 lines
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 736
Quoted 5 lines
drivers/gpu/drm/i915/intel_color.c +5
-0+* from 1.0 to 3.0 +*/ +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT_GC_MAX (pipe,2 ), (1 <<16 ));Doubled thing. But we can just drop these and call
ivb_load_lut_10_max()from higher up instead.
Actually one is GC_MAX and another EXT_GC_MAX. There seem to be 3 set of registers to be programed here. Sure, I will use the ivb helper here to optimize this.
Quoted 64 lines
Quoted 34 lines
drivers/gpu/drm/i915/intel_color.c +34
-0+ +/* +* Program the gc max 2 register to clamp values > 1.0. +* ToDo: Extend the ABI to be able to program values +* from 3.0 to 7.0 +*/ +if (INTEL_GEN (dev_priv) >=10 ||IS_GEMINILAKE (dev_priv)) { +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,0 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,1 ), (1 <<16 )); +I915_WRITE (PREC_PAL_EXT2_GC_MAX (pipe,2 ), (1 <<16 )); + } +} + +static void icl_program_fine_segment_lut (const struct intel_crtc_state + *crtc_state, +struct drm_color_lut *gamma_lut, +u32 offset) +{ +struct drm_crtc *crtc = crtc_state->base.crtc; +struct drm_device *dev = crtc_state->base.crtc->dev; +struct drm_i915_private *dev_priv =to_i915 (dev); +enum pipe pipe =to_intel_crtc (crtc)->pipe; +u32 i, word, lut_size =9 ; + +WARN_ON (offset & ~PAL_PREC_MULTI_SEGMENT_INDEX_VALUE_MASK ); + +I915_WRITE (PREC_PAL_MULTI_SEG_INDEX (pipe), + (PAL_PREC_AUTO_INCREMENT | offset)); + +if (gamma_lut) { +struct drm_color_lut *lut = + (struct drm_color_lut *)gamma_lut; + +for (i =0 ; i < lut_size; i++) {drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state)
Quoted 10 lines
drivers/gpu/drm/i915/intel_color.c +10
-0+/* For Even Index */ + word = (GEN11_GET_LS6BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_LS6BITS_OF_LUT (lut[i].green) <<10 ) | +GEN11_GET_LS6BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + +/* For ODD index */ + word = (GEN11_GET_MS10BITS_OF_LUT (lut[i].red) <<20 ) | + (GEN11_GET_MS10BITS_OF_LUT (lut[i].green) <<10 )|
Quoted 19 lines
drivers/gpu/drm/i915/intel_color.c +19
-0+GEN11_GET_MS10BITS_OF_LUT (lut[i].blue); + +I915_WRITE (PREC_PAL_MULTI_SEG_DATA (pipe), word); + } + } +} + +static void icl_load_gamma_multi_segmented_lut (const struct intel_crtc_state + *crtc_state,u32 offset) +{ +const struct drm_property_blob *gamma_mode_blob = + crtc_state->base.gamma_mode; +struct drm_color_lut *gamma_lut; +int ret; + +if (gamma_mode_blob) { +struct drm_color_mode_lut *arg = gamma_mode_blob->data; +u64 __user *props_ptr = (u64 __user *)(unsigned long )(arg->lut); +
Quoted 8 lines
drivers/gpu/drm/i915/intel_color.c static void glk_load_luts(const struct intel_crtc_state *crtc_state) · line 805
Quoted 6 lines
drivers/gpu/drm/i915/intel_color.c +6
-0+DRM_INFO ("Gamma Mode=%s\n" , arg->name); + gamma_lut =kmalloc ((sizeof (struct drm_color_lut ) * arg->count), +GFP_KERNEL ); + ret =copy_from_user (gamma_lut, props_ptr, +sizeof (struct drm_color_lut ) * arg->count); + }What's this doing here?
This is to get the lut values in driver from userspace. The blob passed from user is
struct drm_color_mode_lut {
/* DRM_MODE_LUT_* */
__u32 flags;
/* number of points on the curve */
__u32 count;
/* Name of Gamma Mode */
char name[DRM_PROP_NAME_LEN];
/* Pointer to Lut elements */
__u64 lut;
};The lut values are passed as pointer in "lut", we are copying the same to kernel memory space here.
Quoted 36 lines
Quoted 23 lines
drivers/gpu/drm/i915/intel_color.c +4
-0+ +icl_program_fine_segment_lut (crtc_state, gamma_lut,0 ); +icl_program_coarse_segment_lut (crtc_state, gamma_lut,0 ); } +static void icl_load_luts (const struct intel_crtc_state *crtc_state){
drivers/gpu/drm/i915/intel_color.c +6
-2if (crtc_state->base.degamma_lut)glk_load_degamma_lut (crtc_state); -if (crtc_state_is_legacy_gamma (crtc_state)) +if (crtc_state_is_legacy_gamma (crtc_state)) {i9xx_load_luts (crtc_state); -else + }else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) { +icl_load_gamma_multi_segmented_lut (crtc_state,0 ); + }else {/* ToDo: Add support for multi segment gamma LUT */ bdw_load_gamma_lut (crtc_state,0 ); + } } static void cherryview_load_luts(const struct intel_crtc_stateQuoted 1 line
*crtc_state) @@ -1034,10 +1176,20 @@ static int glk_color_check(struct
intel_crtc_state*crtc_state)Quoted 11 lines
drivers/gpu/drm/i915/intel_color.c +4
-1return 0 ; } -static u32 icl_gamma_mode (const struct intel_crtc_state *crtc_state) +static u32 icl_gamma_mode (struct intel_crtc_state *crtc_state) { +const struct drm_property_blob *gamma_mode_blob = + crtc_state->base.gamma_mode; +struct drm_color_mode_lut *arg;u32 gamma_mode =0 ;
Quoted 8 lines
drivers/gpu/drm/i915/intel_color.c static int glk_color_check(struct intel_crtc_state *crtc_state) · line 1191
Quoted 6 lines
drivers/gpu/drm/i915/intel_color.c +6
-0+if (gamma_mode_blob) { + arg = gamma_mode_blob->data; +if (!strcmp (arg->name,"multi-segmented gamma" )) + crtc_state->base.gamma_mode_type = +MULTI_SEGMENTED_GAMMA_MODE_12BIT ; + }With the
blob_enumthis kind of stuff should disappear I think.
I feel keeping 2 separate properties should be good as we can keep the capabilities immutable. And with the above structure userspace should be able to set any supported gamma_mode. Let me know your opinion on the same.
Regards,
Uma Shankar
Quoted 33 lines
drivers/gpu/drm/i915/intel_color.c static u32 icl_gamma_mode(const struct intel_crtc_state
Quoted 5 lines
drivers/gpu/drm/i915/intel_color.c +1
-0+if (crtc_state->base.degamma_lut) gamma_mode |=PRE_CSC_GAMMA_ENABLE ;@@ -1048,6 +1200,9 @@ static u32 icl_gamma_mode(const struct intel_crtc_state
*crtc_state)drivers/gpu/drm/i915/intel_color.c static u32 icl_gamma_mode(const struct intel_crtc_state *crtc_state)
Quoted 6 lines
drivers/gpu/drm/i915/intel_color.c +3
-0if (!crtc_state->base.gamma_lut ||crtc_state_is_legacy_gamma (crtc_state)) gamma_mode |=GAMMA_MODE_MODE_8BIT ; +else if (crtc_state->base.gamma_mode_type == +MULTI_SEGMENTED_GAMMA_MODE_12BIT ) + gamma_mode |=
GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED;Quoted 20 lines
drivers/gpu/drm/i915/intel_color.c +0
-0else gamma_mode |=GAMMA_MODE_MODE_10BIT ;include/drm/drm_crtc.h index renamed from include/drm/drm_crtc.h +0
-0bc8a2e7..a8d0b4c 100644
include/drm/drm_crtc.h +3
-1@@ -262,6 +262,9 @@ struct drm_crtc_state {*/ struct drm_property_blob *gamma_mode; +/* Gamma mode type programmed on the pipe */ +u32 gamma_mode_type; +/** * @degamma_lut: * -- 1.9.1
Quoted 8 lines
Quoted 4 lines
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH v2 6/7 drm/i915: Add gamma mode caps property
In reply to Uma ShankarCreate the gamma mode caps property and attach
to crtc.
--- drivers/gpu/drm/i915/intel_color.c | 2 ++ drivers/gpu/drm/i915/intel_display.c | 3 +++ 2 files changed, 5 insertions(+)
drivers/gpu/drm/i915/intel_color.c +2 -0
@@ -1528,6 +1528,8 @@ void intel_color_init(struct intel_crtc *crtc)if (INTEL_GEN (dev_priv) >=11 ) { dev_priv->display.color_check = icl_color_check; +drm_crtc_attach_gamma_mode_caps_property (&crtc->base); +/* don't advertize the >= 1.0 entries */ degamma_lut_size =0 ; gamma_lut_size =ILK_LUT_SIZE_10BIT ;
drivers/gpu/drm/i915/intel_display.c +3 -0
@@ -15606,6 +15606,9 @@ int intel_modeset_init(struct drm_device *dev)INTEL_INFO (dev_priv)->num_pipes, INTEL_INFO(dev_priv)->num_pipes > 1 ? "s" : ""); +if (INTEL_GEN (dev_priv) >=11 ) +drm_color_create_gamma_mode_caps_property (&dev_priv->drm,4 ); +for_each_pipe (dev_priv, pipe) { ret =intel_crtc_init (dev_priv, pipe);if (ret) {
Signature · 6 lines
1.9.1 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH v2 7/7 drm/i915: Attach gamma mode property
In reply to Uma ShankarAttach the gamma mode property to allow userspace
set the gamma mode and provide the luts for the
same.
--- drivers/gpu/drm/i915/intel_color.c | 1 + drivers/gpu/drm/i915/intel_display.c | 4 +++- 2 files changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_color.c +1 -0
@@ -1529,6 +1529,7 @@ void intel_color_init(struct intel_crtc *crtc) dev_priv->display.color_check = icl_color_check;drm_crtc_attach_gamma_mode_caps_property (&crtc->base); +drm_crtc_attach_gamma_mode_property (&crtc->base);/* don't advertize the >= 1.0 entries */ degamma_lut_size =0 ;
drivers/gpu/drm/i915/intel_display.c +3 -1
@@ -15606,8 +15606,10 @@ int intel_modeset_init(struct drm_device *dev)INTEL_INFO (dev_priv)->num_pipes, INTEL_INFO(dev_priv)->num_pipes > 1 ? "s" : ""); -if (INTEL_GEN (dev_priv) >=11 ) +if (INTEL_GEN (dev_priv) >=11 ) {drm_color_create_gamma_mode_caps_property (&dev_priv->drm,4 ); +drm_color_create_gamma_mode_property (&dev_priv->drm,0 ); + }for_each_pipe (dev_priv, pipe) { ret =intel_crtc_init (dev_priv, pipe);
Signature · 6 lines
1.9.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✗ Fi.CI.CHECKPATCH: warning for Add Multi Segment Gamma Support (rev2)
In reply to Uma Shankar== Series Details ==
== Series Details ==
Series: Add Multi Segment Gamma Support (rev2)
URL : https://patchwork.freedesktop.org/series/58169/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
75ad704a2b21 drm: Add gamma mode caps property
659fe8427470 drm/i915: Define color lut range structure
-:61: ERROR:COMPLEX_MACRO: Macros with complex values should be enclosed in parentheses
#61: FILE: drivers/gpu/drm/i915/intel_color.c:1124:
+#define I965_GAMMA_10 \
+ { \
+ .flags = (DRM_MODE_LUT_GAMMA | \
+ DRM_MODE_LUT_INTERPOLATE | \
+ DRM_MODE_LUT_NON_DECREASING), \
+ .count = 128, \
+ .input_bpc = 10, .output_bpc = 16, \
+ .start = 0, .end = (1 << 10) - (1 << 10) / 128, \
+ .min = 0, .max = (1 << 16) - 1, \
+ }, \
+ /* PIPEGCMAX */ \
+ { \
+ .flags = (DRM_MODE_LUT_GAMMA | \
+ DRM_MODE_LUT_INTERPOLATE | \
+ DRM_MODE_LUT_REUSE_LAST | \
+ DRM_MODE_LUT_NON_DECREASING), \
+ .count = 1, \
+ .input_bpc = 10, .output_bpc = 16, \
+ .start = (1 << 10) - (1 << 10) / 128, .end = 1 << 10, \
+ .min = 0, .max = 1 << 16, \
+ }total: 1 errors, 0 warnings, 0 checks, 320 lines checked a059b6bf4fe1 drm: Add gamma mode property
-:26: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis #26: FILE: drivers/gpu/drm/drm_atomic_uapi.c:464:
+ ret = drm_atomic_replace_property_blob_from_id(dev,
+ &state->gamma_mode,
total: 0 errors, 0 warnings, 1 checks, 110 lines checked a5deda67bb81 drm/i915/icl: Add register definitions for Multi Segmented gamma 3b03f07ba428 drm/i915/icl: Add support for multi segmented gamma mode -:79: ERROR:TRAILING_WHITESPACE: trailing whitespace
#79: FILE: drivers/gpu/drm/i915/intel_color.c:719:
+^I$
-:102: WARNING:BLOCK_COMMENT_STYLE: Block comments should align the * on each line #102: FILE: drivers/gpu/drm/i915/intel_color.c:742:
+ * from 3.0 to 7.0
+ */
total: 1 errors, 1 warnings, 0 checks, 211 lines checked 7e97b366c05e drm/i915: Add gamma mode caps property a7d35ed03858 drm/i915: Attach gamma mode property
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✗ Fi.CI.BAT: failure for Add Multi Segment Gamma Support (rev2)
In reply to Uma Shankar== Series Details ==
== Series Details ==
Series: Add Multi Segment Gamma Support (rev2)
URL : https://patchwork.freedesktop.org/series/58169/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_5854 -> Patchwork_12653
====================================================
Summary
-------
**FAILURE**
Serious unknown changes coming with Patchwork_12653 absolutely need to be verified manually.
If you think the reported changes have nothing to do with the changes introduced in Patchwork_12653, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.
External URL: https://patchwork.freedesktop.org/api/1.0/series/58169/revisions/2/mbox/
Possible new issues
-------------------
Here are the unknown changes that may have been introduced in Patchwork_12653:
### IGT changes ###
#### Possible regressions ####
* igt@runner@aborted:
- fi-icl-u3: NOTRUN -> FAIL
- fi-icl-y: NOTRUN -> FAIL
Known issues
------------
Here are the changes found in Patchwork_12653 that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@kms_pipe_crc_basic@read-crc-pipe-b:
- fi-byt-clapper: PASS -> FAIL [fdo#107362]
* igt@runner@aborted:
- fi-icl-u2: NOTRUN -> FAIL [k.org#202973]
- fi-apl-guc: NOTRUN -> FAIL [fdo#108622] / [fdo#109720]
#### Possible fixes ####
* igt@i915_selftest@live_contexts:
- fi-skl-gvtdvm: DMESG-FAIL [fdo#110235 ] -> PASS
* igt@i915_selftest@live_uncore:
- fi-ivb-3770: DMESG-FAIL [fdo#110210] -> PASS
* igt@kms_pipe_crc_basic@nonblocking-crc-pipe-a-frame-sequence:
- fi-byt-clapper: FAIL [fdo#103191] / [fdo#107362] -> PASS
[fdo#103191]: https://bugs.freedesktop.org/show_bug.cgi?id=103191
[fdo#107362]: https://bugs.freedesktop.org/show_bug.cgi?id=107362
[fdo#108622]: https://bugs.freedesktop.org/show_bug.cgi?id=108622
[fdo#109720]: https://bugs.freedesktop.org/show_bug.cgi?id=109720
[fdo#110210]: https://bugs.freedesktop.org/show_bug.cgi?id=110210
[fdo#110235 ]: https://bugs.freedesktop.org/show_bug.cgi?id=110235
[k.org#202973]: https://bugzilla.kernel.org/show_bug.cgi?id=202973
Participating hosts (51 -> 40)
------------------------------
Missing (11): fi-kbl-soraka fi-ilk-m540 fi-hsw-4200u fi-byt-j1900 fi-hsw-peppy fi-byt-squawks fi-bsw-cyan fi-ctg-p8600 fi-gdg-551 fi-pnv-d510 fi-bdw-samus
Build changes
-------------
* Linux: CI_DRM_5854 -> Patchwork_12653
CI_DRM_5854: 3b59d4fb5ba3fb53a87c7a4495d81b0a25aa96eb @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_4919: f539e21e934019f0196fee646f351b4e30a8c341 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_12653: a7d35ed03858e1ceb89578d7bf0bb7d9196e0a9a @ git://anongit.freedesktop.org/gfx-ci/linux
== Linux commits ==
a7d35ed03858 drm/i915: Attach gamma mode property7e97b366c05e drm/i915: Add gamma mode caps property3b03f07ba428 drm/i915/icl: Add support for multi segmented gamma mode a5deda67bb81 drm/i915/icl: Add register definitions for Multi Segmented gamma a059b6bf4fe1 drm: Add gamma mode property659fe8427470 drm/i915: Define color lut range structure 75ad704a2b21 drm: Add gamma mode caps property
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_12653/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 12 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.
Quoted 2 lines
2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.
I think we should just go for the BLOB_ENUM prop type instead. Then the possible values and the current value are all part of the same prop.
Quoted 34 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyVille Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465 ++++++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)Signature · 6 lines
1.9.1 _______________________________________________ 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 24 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 11 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.
Quoted 5 lines
Quoted 2 lines
2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.
Hi Ville,
With the current approach, we have enum property with values as blob_ids (representing platform capabilities). This should not get modified and needs to be immutable.
Userspace can query the property and get the blob using the blob_ids. Thereby getting all the platform capabilities.
Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable LUT_SIZE property (indicating number of luts) and another blob property for actual lut values.
Regards,
Uma Shankar
Quoted 35 lines
Quoted 11 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 7 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)Quoted 7 lines
--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 27 lines · Uma
Quoted 27 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.
Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other drivers/hardware).
Quoted 54 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of luts) and another blob property for actual lut values.Regards,
Uma ShankarQuoted 42 lines
Quoted 11 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 7 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)Quoted 7 lines
--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 48 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 28 lines · Uma
Quoted 28 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.
Quoted 14 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode.
Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for otherdrivers/hardware).Quoted 8 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of luts) and another blob property for actual lut values..
Hi Ville,
Just to clarify and clear some doubts :)
We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and select one and do a setprop to set a mode. Driver will get the blob_id and will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this one property. I hope my understanding is correct.
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob (gamma_lut_property)" property to send the lut values. The challenge there is that it currently uses 16 bit lut values
struct drm_color_lut {
__u16 red;
__u16 green;
__u16 blue;
__u16 reserved;
};which is not sufficient for HDR usecases. Or should we need a new property for advance lut/extended lut like below:
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7
What do you suggest ?
Note: I have currently used 16bit values only to get the feedback on the approach.
Regards,
Uma Shankar
Quoted 44 lines
Quoted 36 lines
Quoted 36 lines
Quoted 12 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 14 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfxQuoted 7 lines
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 3 lines
--
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 80 lines · Uma
Quoted 48 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 28 lines · Uma
Quoted 28 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 14 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode.
Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for otherdrivers/hardware).Quoted 8 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of luts) and another blob property for actual lut values..Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and select one and do a setprop to set a mode. Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this one property. I hope my understanding is correct.Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob (
gamma_lut_property)" property to send the lut values. The challenge there is that it currently uses 16 bit lut valuesstruct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };
Quoted 4 lines
which is not sufficient for HDR usecases. Or should we need a new property for advance lut/extended lut like below:
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when gamma_mode is set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and the kernel would then refuse the operation with with both props being set.
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older client.
Quoted 53 lines
Note: I have currently used 16bit values only to get the feedback on the approach.
Regards,
Uma ShankarQuoted 47 lines
Quoted 36 lines
Quoted 36 lines
Quoted 12 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 14 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfxQuoted 7 lines
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel--
Ville Syrjälä
Intel
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 110 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 64 lines · Uma
Quoted 49 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 28 lines · Uma
Quoted 28 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 14 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 8 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of luts) and another blob property for actual lut values..Quoted 5 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 2 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 3 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 8 lines
The challenge there is that it currently uses 16 bit lut values struct
drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 3 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7
What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and the kernel would then refuse the operation with with both props being set.
Quoted 4 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older client.
We could also say that if a legacy gamma_mode_property is set (which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earlier gamma_mode and fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicit unset may not be needed. What do you think ?
Quoted 50 lines
Quoted 50 lines
Note: I have currently used 16bit values only to get the feedback on the approach.
Regards,
Uma ShankarQuoted 44 lines
Quoted 43 lines
Quoted 43 lines
Quoted 12 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 7 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)Quoted 7 lines
--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
--
Ville Syrjälä
Intel--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 115 lines · Uma
Quoted 110 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 64 lines · Uma
Quoted 49 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 28 lines · Uma
Quoted 28 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 14 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 8 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of luts) and another blob property for actual lut values..Quoted 5 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 2 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 3 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 8 lines
The challenge there is that it currently uses 16 bit lut values struct
drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 3 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7
What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and the kernel would then refuse the operation with with both props being set.Quoted 4 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older client.
Quoted 5 lines
We could also say that if a legacy
gamma_mode_propertyis set (which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicit unset may not be needed. What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Quoted 58 lines
Quoted 58 lines
Quoted 54 lines
Note: I have currently used 16bit values only to get the feedback on the approach.
Regards,
Uma ShankarQuoted 47 lines
Quoted 36 lines
Quoted 36 lines
Quoted 12 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 14 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfxQuoted 7 lines
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel--
Ville Syrjälä
Intel--
Ville Syrjälä
Intel
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 142 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:15 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 110 lines · Uma
Quoted 110 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 67 lines · Uma
Quoted 51 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 30 lines · Uma
Quoted 30 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 15 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 9 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of
luts) and another blob property for actual lut values..Quoted 6 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 3 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 4 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 9 lines
The challenge there is that it currently uses 16 bit lut values
struct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 3 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7
What do you suggest ?
Quoted 7 lines
Quoted 7 lines
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and thekernel would then refuse the operation with with both props being set.
Quoted 11 lines
Quoted 5 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older client.
We could also say that if a legacy
gamma_mode_propertyis set (which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicit unset may not be needed.
Quoted 4 lines
What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Oh ok. So do you suggest that we add some kind of flag to be set by user, based on which we take either legacy or advance_gamma path. Is my understanding correct ?
Quoted 60 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Note: I have currently used 16bit values only to get the feedback on the
approach.
Quoted 48 lines
Quoted 48 lines
Quoted 48 lines
Regards,
Uma ShankarQuoted 44 lines
Quoted 43 lines
Quoted 43 lines
Quoted 12 lines
v2: Used Ville's design and approach to define the interfaces. Addressed Matt Roper's review feedback and re-ordered the patches.
Uma Shankar (5):
drm: Add gamma mode property
drm/i915/icl: Add register definitions for Multi Segmented gamma drm/i915/icl: Add support for multi segmented gamma mode drm/i915: Add gamma mode caps property
drm/i915: Attach gamma mode propertyQuoted 8 lines
Ville Syrjälä (2):
drm: Add gamma mode caps property
drm/i915: Define color lut range structuredrivers/gpu/drm/drm_atomic_uapi.c | 13 + drivers/gpu/drm/drm_color_mgmt.c | 110 +++++++++ drivers/gpu/drm/i915/i915_reg.h | 17 ++ drivers/gpu/drm/i915/intel_color.c | 465++++++++++++++++++++++++++++++++++-
Quoted 7 lines
drivers/gpu/drm/i915/intel_display.c | 5 + include/drm/drm_color_mgmt.h | 11 + include/drm/drm_crtc.h | 17 ++ include/drm/drm_mode_config.h | 10 + include/uapi/drm/drm_mode.h | 49 ++++ 9 files changed, 690 insertions(+), 7 deletions(-)Quoted 7 lines
--
1.9.1_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-develQuoted 7 lines
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
--
Ville Syrjälä
Intel--
Ville Syrjälä
Intel
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 142 lines · Uma
Quoted 142 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:15 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 110 lines · Uma
Quoted 110 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 67 lines · Uma
Quoted 51 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 30 lines · Uma
Quoted 30 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 13 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lut coefficients.2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.Quoted 7 lines
Quoted 2 lines
I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the same prop.Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 15 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 9 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of
luts) and another blob property for actual lut values..Quoted 6 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 3 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 4 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 9 lines
The challenge there is that it currently uses 16 bit lut values
struct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 3 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev=7
What do you suggest ?
Quoted 7 lines
Quoted 7 lines
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and thekernel would then refuse the operation with with both props being set.
Quoted 11 lines
Quoted 5 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older client.
We could also say that if a legacy
gamma_mode_propertyis set (which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicit unset may not be needed.
Quoted 7 lines
Quoted 4 lines
What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Oh ok. So do you suggest that we add some kind of flag to be set by user, based on which we take either legacy or
advance_gammapath. Is my understanding correct ?
Yeah, just another client cap. I can't immediately think of a nicer way to extend the precision.
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
Quoted 168 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:38 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 61 lines · Uma
Quoted 61 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:15 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 47 lines · Uma
Quoted 47 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 34 lines · Uma
Quoted 34 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 20 lines · Uma
Quoted 20 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 4 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lutcoefficients.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 3 lines
2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the sameprop.
Quoted 61 lines
Quoted 61 lines
Quoted 61 lines
Quoted 61 lines
Quoted 22 lines
Quoted 6 lines
Quoted 5 lines
Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 15 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 9 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of
luts) and another blob property for actual lut values..Quoted 6 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 3 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 4 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 9 lines
The challenge there is that it currently uses 16 bit lut values
struct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 4 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev
=7What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and theQuoted 6 lines
Quoted 6 lines
kernel would then refuse the operation with with both props being set.
Quoted 5 lines
Quoted 5 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older
client.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
We could also say that if a legacy
gamma_mode_propertyis set
(which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicitunset may not be needed.
Quoted 5 lines
Quoted 4 lines
What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Quoted 6 lines
Quoted 2 lines
Oh ok. So do you suggest that we add some kind of flag to be set by user, based on which we take either legacy or
advance_gammapath. Is myunderstanding correct ?
Yeah, just another client cap. I can't immediately think of a nicer way to extend the precision.
Sure Ville, I am implementing based on this suggestion. Basically will expose a new advance_gamma_mode flag as a client cap. Something like:
#define DRM_CLIENT_CAP_ADVANCE_GAMMA_MODES 6
The new users will set this and we will assume that advance gamma mode paths is active. If the flag is not set, driver will take the legacy path and enable default gamma_mode for that particular platform.
In both cases, we will use the existing GAMMA_LUT blob property to send the lut values from userspace.
Will send out the next version soon. Thanks for your valuable feedback and suggestions.
Regards,
Uma Shankar
Quoted 8 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 175 lines · Uma
Quoted 168 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:38 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 61 lines · Uma
Quoted 61 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:15 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 47 lines · Uma
Quoted 47 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 34 lines · Uma
Quoted 34 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 20 lines · Uma
Quoted 20 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 4 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lutcoefficients.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 3 lines
2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the sameprop.
Quoted 61 lines
Quoted 61 lines
Quoted 61 lines
Quoted 61 lines
Quoted 22 lines
Quoted 6 lines
Quoted 5 lines
Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.Quoted 15 lines
That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma mode. Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for i915 but may not work for other
drivers/hardware).Quoted 9 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of
luts) and another blob property for actual lut values..Quoted 6 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name embedded) and
select one and do a setprop to set a mode.
Quoted 3 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 4 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 9 lines
The challenge there is that it currently uses 16 bit lut values
struct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 4 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&rev
=7What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and theQuoted 6 lines
Quoted 6 lines
kernel would then refuse the operation with with both props being set.
Quoted 5 lines
Quoted 5 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older
client.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
We could also say that if a legacy
gamma_mode_propertyis set
(which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicitunset may not be needed.
Quoted 5 lines
Quoted 4 lines
What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Quoted 6 lines
Quoted 2 lines
Oh ok. So do you suggest that we add some kind of flag to be set by user, based on which we take either legacy or
advance_gammapath. Is myunderstanding correct ?
Yeah, just another client cap. I can't immediately think of a nicer way to extend the precision.
Quoted 8 lines
Sure Ville, I am implementing based on this suggestion. Basically will expose a new
advance_gamma_modeflag as a client cap. Something like:#define
DRM_CLIENT_CAP_ADVANCE_GAMMA_MODES6The new users will set this and we will assume that advance gamma mode paths is active. If the flag is not set, driver will take the legacy path and enable default
gamma_modefor that particular platform.
It's maybe a still a bit nasty because we basically have to ignore the gamma mode prop entirely in that case. Not quite sure what is the best way to shoehorn this in. I guess the main problem is what we would report to userspace via the (DE)GAMMA_MODE props if the previous client left gamma_mode set in a way that conflicts with what (DE)GAMMA_LUT_SIZE suggests it should be.
Signature · 6 lines
Ville Syrjälä Intel _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 196 lines
Quoted 96 lines
Quoted 96 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:38 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 64 lines · Uma
Quoted 64 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Monday, April 8, 2019 9:15 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 50 lines · Uma
Quoted 50 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 8:27 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 36 lines · Uma
Quoted 36 lines
-----Original Message-----
From: Ville Syrjälä [mailto:[email protected]]
Sent: Monday, April 8, 2019 6:01 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected]; dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 22 lines · Uma
Quoted 22 lines
-----Original Message-----
From: dri-devel [mailto:[email protected]] On Behalf Of Ville Syrjälä
Sent: Friday, April 5, 2019 9:42 PM
To: Shankar, Uma <[email protected]>
Cc: [email protected]; [email protected];
dri- [email protected]; [email protected]; Syrjala, Ville <[email protected]>; Lankhorst, Maarten <[email protected]>
Subject: Re: [Intel-gfx] [v2 0/7] Add Multi Segment Gamma SupportQuoted 5 lines · Uma Shankar
This series adds support for programmable gamma modes and exposes a property interface for the same. Also added, support for multi segment gamma mode introduced in ICL+
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
It creates 2 property interfaces :
1.GAMMA_MODE_CAPS: This is immutable property and exposes the various gamma modes supported and the lut ranges. This is an enum property with element as blob id. Getting the blob id in userspace, user can get the mode supported and also the range of gamma mode supported with number of lutcoefficients.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 3 lines
2.
GAMMA_MODE: This is for user to set the gamma mode and send the lut values for that particular mode.I think we should just go for the
BLOB_ENUMprop type instead. Then the possible values and the current value are all part of the sameprop.
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Hi Ville,
With the current approach, we have enum property with values asblob_ids(representing platform capabilities). This should not get modified and needs to be immutable.That's not quite what we want. We want to let the user modify the current value so that they can actually select the gamma
mode.
Quoted 85 lines
Quoted 77 lines
Quoted 56 lines
Quoted 56 lines
Quoted 56 lines
Quoted 56 lines
Quoted 15 lines
Quoted 14 lines
Otherwise we need yet another prop for it, or we have to deduce it from the LUT size (that apporach would actually work for
i915 but may not work for otherdrivers/hardware).Quoted 9 lines
Userspace can query the property and get the blob using the
blob_ids. Thereby getting all the platform capabilities.Now to set the LUT values, he can use another blob property and pass the luts. This is inline to how gamma/degamma is implemented where we have one immutable
LUT_SIZEproperty (indicating number of
luts) and another blob property for actual lut values..Quoted 7 lines
Hi Ville,
Just to clarify and clear some doubts :)We should be able to set the gamma mode using the blob enum value. Userspace will get the list enum vals (blob ids with mode name
embedded) andselect one and do a setprop to set a mode.
Quoted 3 lines
Driver will get the
blob_idand will be able to get the mode to be set. So exposing capabilities and setting the mode should be possible with this oneproperty. I hope my understanding is correct.
Quoted 4 lines
Now to send the actual blob values to be set, we need to use some other property interface. Should we use the currently available "gamma blob
(
gamma_lut_property)" property to send the lut values.Quoted 9 lines
The challenge there is that it currently uses 16 bit lut
values struct drm_color_lut { __u16 red; __u16 green; __u16 blue; __u16 reserved; };which is not sufficient for HDR usecases. Or should we need a new property for
advance lut/extended lut like below:
Quoted 5 lines
https://patchwork.freedesktop.org/patch/294732/?series=30875&
rev
=7What do you suggest ?
I was thinking that we might get away with reusing the current props and just change the interpretation of the data when
gamma_modeis set. But I'm not sure that's going to work out so well if one client sets this up and then another client comes along that doesn't understand the new props at all. But even with separate props I think we might still end up in a mess because the new client wouldn't know how to unset the higher precision LUT before setting up the old style prop and theQuoted 7 lines
Quoted 7 lines
kernel would then refuse the operation with with both props being set.
Quoted 6 lines
Quoted 6 lines
So I think we might need a client cap for this which simply changes how the data in the existing props is represented. So internally we could always store things in the new high precision format, but we'd convert to/from the old format when dealing with an older
client.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
We could also say that if a legacy
gamma_mode_propertyis set (which will be used by legacy apps or apps not aware of new interface), in driver we will simply unset the earliergamma_modeand fallback to legacy mode (whatever it was for a particular platform). This way we should be able to deal with this situation and an explicitunset may not be needed.
Quoted 5 lines
Quoted 4 lines
What do you think ?
I don't want (non-immutable) properties that magically change values. That way lies madness.
Quoted 7 lines
Quoted 3 lines
Oh ok. So do you suggest that we add some kind of flag to be set by user, based on which we take either legacy or
advance_gammapath. Is myunderstanding correct ?
Yeah, just another client cap. I can't immediately think of a nicer way to extend the precision.
Quoted 6 lines
Sure Ville, I am implementing based on this suggestion. Basically will expose a new
advance_gamma_modeflag as a client cap. Something like:#define
DRM_CLIENT_CAP_ADVANCE_GAMMA_MODES6The new users will set this and we will assume that advance gamma mode paths is
active.
Quoted 8 lines
Quoted 2 lines
If the flag is not set, driver will take the legacy path and enable default
gamma_modefor that particular platform.It's maybe a still a bit nasty because we basically have to ignore the gamma mode prop entirely in that case. Not quite sure what is the best way to shoehorn this in. I guess the main problem is what we would report to userspace via the (DE)GAMMA_MODE props if the previous client left
gamma_modeset in a way that conflicts with what (DE)GAMMA_LUT_SIZE suggests it should be.
If user sets the Client Cap then he will ignore the DE)GAMMA_LUT_SIZE properties and will rely on the new property interface. Data passed as blob in already available GAMMA_LUT property will be taken in conjunction with the mode set by new GAMMA_MODE property . Now if this app finishes and new user comes up, and if it doesn't sets the Client CAP that means its not planning to use new interface, as part of gamma_lut programming we will fallback to linear lut with default mode.
This anyways would have been the state if the first app was not there. And if the new app wants to use legacy method to change (DE)GAMMA lut he can get the default size in LUT_SIZE property which will still be good for legacy and work with that. I feel this will solve the issues we may get and both can co-exist.
Regards,
Uma Shankar
Quoted 7 lines
--
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel