Skip to contentSkip to conversation

Display Global Histogram

PATCH Superseded Arun R Murthy 4 participants 36 messages Jan 9, 2025 – Jan 21, 2025

Subsystems (5)

v7 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#0 lore raw

v7 cover Display Global Histogram

EDITME: Imported from [email protected]
Please review before sending.

Display histogram is a hardware functionality where a statistics for 'n' number of frames is generated to form a histogram data. This is notified to the user via histogram event. Compositor will then upon sensing the histogram event will read the histogram data from KMD via crtc property. User can use this histogram and apply various equilization techniques to enhance the image or use this histogram for shaders.

Display ImageEnhancemenT is a hardware that interpolates the LUT value to generate the enhanced output image. 1DLUT value is to be provided by the user via crtc property.

One such library Global Histogram Enhancement(GHE) will take the histogram as input and apply the algorithm to enhance the density and then return the enhanced LUT factor. This library can be located @
https://github.com/intel/ghe

The corresponding mutter changes to enable/disable histogram, read the
histogram data, communicate with the library and write the enhanced data
back to the KMD is also pushed for review at https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/3873
and https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/3873/diffs?commit_id=270808ca7c8be48513553d95b4a47541f5d40206
The IGT changes for validating the histogram event and reading the
histogram is also pushed for review at https://patchwork.freedesktop.org/series/135789/

NOTE: i915 driver changes for histogram and IET LUT is not fully tested and the series is pushed to get the inital feel of the histogram/IET LUT usage as well as to get started with the review.

Test-with: [email protected]

Arun R Murthy (10):
drm/crtc: Add histogram properties
drm/crtc: Expose API to create drm crtc property for histogram drm/i915/histogram: Define registers for histogram drm/i915/histogram: Add support for histogram drm/xe: Add histogram support to Xe builds drm/i915/histogram: histogram interrupt handling drm/i915/display: handle drm-crtc histogram property updates drm/i915/histogram: histogram delay counter doesnt reset drm/i915/histogram: Histogram changes for Display 20+ drm/i915/histogram: Enable pipe dithering

 drivers/gpu/drm/drm_atomic_state_helper.c     |   6 +
 drivers/gpu/drm/drm_atomic_uapi.c             |  17 +
 drivers/gpu/drm/drm_crtc.c                    |  30 ++
 drivers/gpu/drm/i915/Makefile                 |   1 +
 drivers/gpu/drm/i915/display/intel_atomic.c   |   1 +
 drivers/gpu/drm/i915/display/intel_crtc.c     |   7 +
 drivers/gpu/drm/i915/display/intel_display.c  |  17 +
 .../gpu/drm/i915/display/intel_display_irq.c  |   6 +-
 .../drm/i915/display/intel_display_types.h    |   4 +
 .../gpu/drm/i915/display/intel_histogram.c    | 380 ++++++++++++++++++
 .../gpu/drm/i915/display/intel_histogram.h    |  40 ++
 .../drm/i915/display/intel_histogram_regs.h   |  75 ++++
 drivers/gpu/drm/i915/i915_reg.h               |   5 +-
 drivers/gpu/drm/xe/Makefile                   |   1 +
 include/drm/drm_crtc.h                        |  49 +++
 include/uapi/drm/drm_mode.h                   |  11 +
 16 files changed, 647 insertions(+), 3 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/display/intel_histogram.c
 create mode 100644 drivers/gpu/drm/i915/display/intel_histogram.h
 create mode 100644 drivers/gpu/drm/i915/display/intel_histogram_regs.h

--
2.25.1

---
Arun R Murthy (4):
drm: Define histogram structures exposed to user
drm: Define ImageEnhancemenT LUT structures exposed to user drm/crtc: Expose API to create drm crtc property for histogram drm/crtc: Expose API to create drm crtc property for IET LUT

 drivers/gpu/drm/drm_atomic_state_helper.c |  23 +++++++
 drivers/gpu/drm/drm_atomic_uapi.c         |  28 ++++++++
 drivers/gpu/drm/drm_crtc.c                |  97 +++++++++++++++++++++++++++
 include/drm/drm_crtc.h                    |  77 ++++++++++++++++++++++
 include/uapi/drm/drm_mode.h               | 105 ++++++++++++++++++++++++++++++
 5 files changed, 330 insertions(+)

---
base-commit: 78526dfb8799485890dae3877fea308e9501879c
change-id: 20241218-dpst-c8ecf18062bb

Best regards,
--
Arun R Murthy <[email protected]>

---
Arun R Murthy (14):
drm: Define histogram structures exposed to user
drm: Define ImageEnhancemenT LUT structures exposed to user drm/crtc: Expose API to create drm crtc property for histogram drm/crtc: Expose API to create drm crtc property for IET LUT drm/i915/histogram: Define registers for histogram drm/i915/histogram: Add support for histogram drm/xe: Add histogram support to Xe builds drm/i915/histogram: histogram interrupt handling drm/i915/histogram: Hook i915 histogram with drm histogram drm/i915/iet: Add support to writing the IET LUT data drm/i915/crtc: Hook i915 IET LUT with the drm IET properties drm/i915/histogram: histogram delay counter doesnt reset drm/i915/histogram: Histogram changes for Display 20+ drm/i915/histogram: Enable pipe dithering

 drivers/gpu/drm/drm_atomic_state_helper.c          |  23 ++
 drivers/gpu/drm/drm_atomic_uapi.c                  |  28 ++
 drivers/gpu/drm/drm_crtc.c                         | 127 +++++++
 drivers/gpu/drm/i915/Makefile                      |   1 +
 drivers/gpu/drm/i915/display/intel_crtc.c          |  10 +
 drivers/gpu/drm/i915/display/intel_display.c       |  14 +
 drivers/gpu/drm/i915/display/intel_display_irq.c   |   6 +-
 drivers/gpu/drm/i915/display/intel_display_types.h |   2 +
 drivers/gpu/drm/i915/display/intel_histogram.c     | 396 +++++++++++++++++++++
 drivers/gpu/drm/i915/display/intel_histogram.h     |  55 +++
 .../gpu/drm/i915/display/intel_histogram_regs.h    |  75 ++++
 drivers/gpu/drm/i915/i915_reg.h                    |   5 +-
 drivers/gpu/drm/xe/Makefile                        |   1 +
 include/drm/drm_crtc.h                             |  80 +++++
 include/uapi/drm/drm_mode.h                        | 109 ++++++
 15 files changed, 929 insertions(+), 3 deletions(-)

---
base-commit: 78526dfb8799485890dae3877fea308e9501879c
change-id: 20241218-dpst-c8ecf18062bb

Best regards,

Signature · 1 line
Arun R Murthy <[email protected]>
#1 lore raw

PATCH v7 1/14 drm: Define histogram structures exposed to user

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800)
25 bits should be sufficient to represent this along with a buffer of 7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
	bin = max(RGB[i]);
	bin = bin >> 3;	/* consider the most significant bits */
	hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero.
Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
include/uapi/drm/drm_mode.h +59 -0
index c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bdaec42c18be34e2d05 100644
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
 	__u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.
+ * If the most significant 5 bits are considered, then bins = 0xff >> 3
+ * will be 32 bins.
+ * For illustration consider a full RED image of 10k resolution considering all
+ * 8 bits histogram would look like hist[255] = {0,0,....44236800}
+ */
+#define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 << 0)
+
+/**
+ * struct drm_histogram_caps
+ *
+ * @histogram_mode: histogram generation modes, defined in the above macros
+ * @bins_count: number of bins for a chosen histogram mode. For illustration
+ *		refer the above defined histogram mode.
+ */
+struct drm_histogram_caps {
+	u8 histogram_mode;
+	u32 bins_count;
+};
+
+/**
+ * struct drm_histogram_config
+ *
+ * @enable: flag to enable/disable histogram
+ * @hist_mode: histogram mode(HSV max(RGB), RGB, LUMA etc)
+ * @reserved1: Reserved for future use
+ * @reserved2: Reserved for future use
+ * @reserved3: Reserved for future use
+ * @reserved4: Reserved for future use
+ */
+struct drm_histogram_config {
+	bool enable;
+	u8 hist_mode;
+	u32 reserved1;
+	u32 reserved2;
+	u32 reserved3;
+	u32 reserved4;
+};
+
+/**
+ * struct drm_histogram
+ *
+ * @config: histogram configuration data pointed by struct drm_histogram_config
+ * @data_ptr: pointer to the array of histogram.
+ *	      Histogram is an array of bins. Data format for each bin depends
+ *	      on the histogram mode. Refer to the above histogram modes for
+ *	      more information.
+ * @nr_elements: number of bins in the histogram.
+ */
+struct drm_histogram {
+	struct drm_histogram_config config;
+	__u64 data_ptr;
+	__u32 nr_elements;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Signature · 1 line
2.25.1
#2 lore raw

Re: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 38 lines · Arun R Murthy

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800)
25 bits should be sufficient to represent this along with a buffer of 7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
	bin = max(RGB[i]);
	bin = bin >> 3;	/* consider the most significant bits */
	hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero.
Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
index c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bdaec42c18be34e2d05 100644

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1362

Quoted 8 lines
include/uapi/drm/drm_mode.h +5 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
 	__u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component.
Likewise even in an RGB24 case there are 256 possible values. It's not clear why 0 gets excluded.

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1364

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+ * If the most significant 5 bits are considered, then bins = 0xff >> 3
+ * will be 32 bins.

If 5 bits are considered, there will be 2^5 bins, no matter of 0xff >> 3.

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 4 lines
include/uapi/drm/drm_mode.h +4 -0
+ * For illustration consider a full RED image of 10k resolution considering all
+ * 8 bits histogram would look like hist[255] = {0,0,....44236800}
+ */
+#define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 << 0)

Why do you have a bitshift here?

Quoted 20 lines
include/uapi/drm/drm_mode.h +20 -0
+
+/**
+ * struct drm_histogram_caps
+ *
+ * @histogram_mode: histogram generation modes, defined in the above macros
+ * @bins_count: number of bins for a chosen histogram mode. For illustration
+ *		refer the above defined histogram mode.
+ */
+struct drm_histogram_caps {
+	u8 histogram_mode;
+	u32 bins_count;
+};
+
+/**
+ * struct drm_histogram_config
+ *
+ * @enable: flag to enable/disable histogram
+ * @hist_mode: histogram mode(HSV max(RGB), RGB, LUMA etc)
+ * @reserved1: Reserved for future use
+ * @reserved2: Reserved for future use

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1398

Quoted 10 lines
include/uapi/drm/drm_mode.h +10 -0
+ * @reserved3: Reserved for future use
+ * @reserved4: Reserved for future use
+ */
+struct drm_histogram_config {
+	bool enable;
+	u8 hist_mode;
+	u32 reserved1;
+	u32 reserved2;
+	u32 reserved3;
+	u32 reserved4;

What for? Also this struct leaves a 3-byte hole, which might be not so beneficial.

Quoted 25 lines
include/uapi/drm/drm_mode.h +18 -0
+};
+
+/**
+ * struct drm_histogram
+ *
+ * @config: histogram configuration data pointed by struct drm_histogram_config
+ * @data_ptr: pointer to the array of histogram.
+ *	      Histogram is an array of bins. Data format for each bin depends
+ *	      on the histogram mode. Refer to the above histogram modes for
+ *	      more information.
+ * @nr_elements: number of bins in the histogram.
+ */
+struct drm_histogram {
+	struct drm_histogram_config config;
+	__u64 data_ptr;
+	__u32 nr_elements;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Signature · 2 lines
-- 
2.25.1
Signature · 2 lines
With best wishes
Dmitry
#3 lore raw

RE: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 48 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of 7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
	bin = max(RGB[i]);
	bin = bin >> 3;	/* consider the most significant bits */
	hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero.
Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
Quoted 2 lines

index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda ec42

Quoted 7 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
 	__u32 pad;
 };
 
Quoted 11 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1362

Quoted 6 lines
include/uapi/drm/drm_mode.h +6 -0
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer each
+bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component.
Likewise even in an RGB24 case there are 256 possible values. It's not clear why 0 gets excluded.

This applies to only SDR and excludes HDR.
RGB in hex can have a maximum value of 0xFF { RGB (255, 255, 255) }

Quoted 6 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1366

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ * If the most significant 5 bits are considered, then bins = 0xff >>
+ 3
+ * will be 32 bins.

If 5 bits are considered, there will be 2^5 bins, no matter of 0xff >> 3.

Agree!

Quoted 9 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1371

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+ * For illustration consider a full RED image of 10k resolution
+considering all
+ * 8 bits histogram would look like hist[255] = {0,0,....44236800}
+*/
+#define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 <<

0)

Why do you have a bitshift here?

Bitwise notification is used to differentiate multiple histogram modes. Currently we have max(RGB), upon adding other histogram modes the same can be included here.

Quoted 26 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb {

Quoted 7 lines
include/uapi/drm/drm_mode.h +7 -0
+
+/**
+ * struct drm_histogram_caps
+ *
+ * @histogram_mode: histogram generation modes, defined in the above
+macros
+ * @bins_count: number of bins for a chosen histogram mode. For

illustration

Quoted 18 lines
include/uapi/drm/drm_mode.h +18 -0
+ *		refer the above defined histogram mode.
+ */
+struct drm_histogram_caps {
+	u8 histogram_mode;
+	u32 bins_count;
+};
+
+/**
+ * struct drm_histogram_config
+ *
+ * @enable: flag to enable/disable histogram
+ * @hist_mode: histogram mode(HSV max(RGB), RGB, LUMA etc)
+ * @reserved1: Reserved for future use
+ * @reserved2: Reserved for future use
+ * @reserved3: Reserved for future use
+ * @reserved4: Reserved for future use  */ struct
+drm_histogram_config {
+	bool enable;
Quoted 9 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1398

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+	u8 hist_mode;
+	u32 reserved1;
+	u32 reserved2;
+	u32 reserved3;
+	u32 reserved4;

What for? Also this struct leaves a 3-byte hole, which might be not so beneficial.

This is kept for future use. If weighted RGB mode is used for histogram generation then we need 3 variables to get the weightage. For any other new histogram modes or for future usage this is kept reserved.
Regarding the padding, will re-oder the elements in the struct.

Quoted 23 lines
Quoted 23 lines
include/uapi/drm/drm_mode.h +19 -0
+};
+
+/**
+ * struct drm_histogram
+ *
+ * @config: histogram configuration data pointed by struct
+drm_histogram_config
+ * @data_ptr: pointer to the array of histogram.
+ *	      Histogram is an array of bins. Data format for each bin depends
+ *	      on the histogram mode. Refer to the above histogram modes for
+ *	      more information.
+ * @nr_elements: number of bins in the histogram.
+ */
+struct drm_histogram {
+	struct drm_histogram_config config;
+	__u64 data_ptr;
+	__u32 nr_elements;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Quoted 4 lines
Quoted 3 lines

--
2.25.1

Thanks and Regards,
Arun R Murthy
--------------------

#4 lore raw

Re: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 60 lines · Arun R
Quoted 49 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of 7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero.
Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
Quoted 2 lines

index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda ec42

Quoted 7 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
Quoted 11 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1362

Quoted 6 lines
include/uapi/drm/drm_mode.h +6 -0
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer each
+bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component.
Likewise even in an RGB24 case there are 256 possible values. It's not clear why 0 gets excluded.

Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

Quoted 17 lines

RGB in hex can have a maximum value of 0xFF { RGB (255, 255, 255) }

Quoted 6 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1366

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ * If the most significant 5 bits are considered, then bins = 0xff >>
+ 3
+ * will be 32 bins.

If 5 bits are considered, there will be 2^5 bins, no matter of 0xff >> 3.

Agree!

Quoted 7 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1371

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+ * For illustration consider a full RED image of 10k resolution
+considering all
+ * 8 bits histogram would look like hist[255] = {0,0,....44236800}
+*/
+#define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 5 lines
Quoted 2 lines

Why do you have a bitshift here?

Bitwise notification is used to differentiate multiple histogram modes. Currently we have max(RGB), upon adding other histogram modes the same can be included here.

Define a normal enum. There is no need to use bitmasks for histogram modes.

Quoted 36 lines
Quoted 27 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb {

Quoted 8 lines
include/uapi/drm/drm_mode.h +7 -0
+
+/**
+ * struct drm_histogram_caps
+ *
+ * @histogram_mode: histogram generation modes, defined in the above
+macros
+ * @bins_count: number of bins for a chosen histogram mode. For

illustration

Quoted 18 lines
include/uapi/drm/drm_mode.h +18 -0
+ *         refer the above defined histogram mode.
+ */
+struct drm_histogram_caps {
+   u8 histogram_mode;
+   u32 bins_count;
+};
+
+/**
+ * struct drm_histogram_config
+ *
+ * @enable: flag to enable/disable histogram
+ * @hist_mode: histogram mode(HSV max(RGB), RGB, LUMA etc)
+ * @reserved1: Reserved for future use
+ * @reserved2: Reserved for future use
+ * @reserved3: Reserved for future use
+ * @reserved4: Reserved for future use  */ struct
+drm_histogram_config {
+   bool enable;
Quoted 9 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1398

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+   u8 hist_mode;
+   u32 reserved1;
+   u32 reserved2;
+   u32 reserved3;
+   u32 reserved4;

What for? Also this struct leaves a 3-byte hole, which might be not so beneficial.

Quoted 3 lines

This is kept for future use. If weighted RGB mode is used for histogram generation then we need 3 variables to get the weightage. For any other new histogram modes or for future usage this is kept reserved.

What for? Please don't pollute it with 'reserved' fields. Consider how the Color Pipelines data is defined: it has separate 'type' and 'data' objects.
Type can be selected via enum. Data is blob which is parsed according to the type.

Quoted 32 lines

Regarding the padding, will re-oder the elements in the struct.

Quoted 23 lines
Quoted 23 lines
include/uapi/drm/drm_mode.h +19 -0
+};
+
+/**
+ * struct drm_histogram
+ *
+ * @config: histogram configuration data pointed by struct
+drm_histogram_config
+ * @data_ptr: pointer to the array of histogram.
+ *       Histogram is an array of bins. Data format for each bin depends
+ *       on the histogram mode. Refer to the above histogram modes for
+ *       more information.
+ * @nr_elements: number of bins in the histogram.
+ */
+struct drm_histogram {
+   struct drm_histogram_config config;
+   __u64 data_ptr;
+   __u32 nr_elements;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Quoted 4 lines
Quoted 3 lines

--
2.25.1

Thanks and Regards,
Arun R Murthy
--------------------

Signature · 2 lines
With best wishes
Dmitry
#5 lore raw

RE: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 62 lines
Quoted 37 lines
Quoted 37 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 20 lines
Quoted 20 lines

ec42

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1363

Quoted 13 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +6 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer
+each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 4 lines
Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image.
We are trying to enhance the image quality of SDR with the support of histogram.

Quoted 18 lines
Quoted 18 lines

RGB in hex can have a maximum value of 0xFF { RGB (255, 255, 255) }

Quoted 7 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1367

Quoted 4 lines
include/uapi/drm/drm_mode.h +4 -0
+ * If the most significant 5 bits are considered, then bins =
+ 0xff >>
+ 3
+ * will be 32 bins.

If 5 bits are considered, there will be 2^5 bins, no matter of 0xff >> 3.

Agree!

Quoted 7 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1372

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+ * For illustration consider a full RED image of 10k resolution
+considering all
+ * 8 bits histogram would look like hist[255] =
+{0,0,....44236800} */
+#define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Why do you have a bitshift here?

Bitwise notification is used to differentiate multiple histogram modes. Currently we have max(RGB), upon adding other histogram modes the same can be included here.

Define a normal enum. There is no need to use bitmasks for histogram modes.

Ok Done!

Quoted 40 lines
Quoted 36 lines
Quoted 27 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb {

Quoted 8 lines
include/uapi/drm/drm_mode.h +7 -0
+
+/**
+ * struct drm_histogram_caps
+ *
+ * @histogram_mode: histogram generation modes, defined in the
+above macros
+ * @bins_count: number of bins for a chosen histogram mode. For

illustration

Quoted 18 lines
include/uapi/drm/drm_mode.h +18 -0
+ *         refer the above defined histogram mode.
+ */
+struct drm_histogram_caps {
+   u8 histogram_mode;
+   u32 bins_count;
+};
+
+/**
+ * struct drm_histogram_config
+ *
+ * @enable: flag to enable/disable histogram
+ * @hist_mode: histogram mode(HSV max(RGB), RGB, LUMA etc)
+ * @reserved1: Reserved for future use
+ * @reserved2: Reserved for future use
+ * @reserved3: Reserved for future use
+ * @reserved4: Reserved for future use  */ struct
+drm_histogram_config {
+   bool enable;
Quoted 9 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1398

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+   u8 hist_mode;
+   u32 reserved1;
+   u32 reserved2;
+   u32 reserved3;
+   u32 reserved4;

What for? Also this struct leaves a 3-byte hole, which might be not so beneficial.

Quoted 3 lines

This is kept for future use. If weighted RGB mode is used for histogram generation then we need 3 variables to get the weightage. For any other new histogram modes or for future usage this is kept reserved.

Quoted 6 lines

What for? Please don't pollute it with 'reserved' fields. Consider how the Color Pipelines data is defined: it has separate 'type' and 'data' objects.
Type can be selected via enum. Data is blob which is parsed according to the type.

Ok sure will re-check.
The challenging part over here in for some types we will have to specify the sub-type or the data format of the value derived the type.
Hence we are using a sub-type. Will remove these unused variables and re-frame the struct.

Thanks and Regards,
Arun R Murthy
--------------------

#6 lore raw

Re: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 62 lines · Arun R
Quoted 62 lines
Quoted 37 lines
Quoted 37 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 20 lines
Quoted 20 lines

ec42

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1363

Quoted 13 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +6 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in buffer
+each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image.
We are trying to enhance the image quality of SDR with the support of histogram.

You are defining generic API. It might be broader than your existing usecase. Please consider supporting HDR too.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

RGB in hex can have a maximum value of 0xFF { RGB (255, 255, 255) }

Signature · 2 lines
With best wishes
Dmitry
#7 lore raw

RE: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 68 lines · Arun R
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 24 lines
Quoted 24 lines
Quoted 20 lines
Quoted 8 lines

ec42

Quoted 7 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
Quoted 12 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1362

Quoted 6 lines
include/uapi/drm/drm_mode.h +6 -0
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in
+buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

Quoted 8 lines
Quoted 3 lines

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image. We are trying to enhance the image quality of SDR with the support of

histogram.

You are defining generic API. It might be broader than your existing usecase. Please consider supporting HDR too.

HDR image enhancement is very much complex including multiple stages such as image tone mapping and image denoising. Here for SDR planes, image enhancement is done by playing around the contrast and color.
Maybe at this stage we can focus on SDR and can take this HDR at the next stage.

Thanks and Regards,
Arun R Murthy
--------------------

#8 lore raw

Re: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 71 lines · Arun R
Quoted 71 lines · Arun R
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered.
max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 27 lines
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines

ec42

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1363

Quoted 13 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +6 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 = 44236800
+ * can be denoted in 25bits. With an additional 7 bits in
+buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 4 lines
Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image. We are trying to enhance the image quality of SDR with the support of

Quoted 8 lines
Quoted 5 lines

histogram.

You are defining generic API. It might be broader than your existing usecase. Please consider supporting HDR too.

HDR image enhancement is very much complex including multiple stages such as image tone mapping and image denoising. Here for SDR planes, image enhancement is done by playing around the contrast and color.
Maybe at this stage we can focus on SDR and can take this HDR at the next stage.

If you define max(colour) to be 255, then you can not expand it later. The API will have 8 bits for colour information everywhere.

Quoted 4 lines

Thanks and Regards,
Arun R Murthy
--------------------

Signature · 2 lines
With best wishes
Dmitry
#9 lore raw

RE: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 81 lines · Arun R
Quoted 38 lines
Quoted 38 lines · Arun R
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines
Quoted 37 lines · Arun R Murthy
Quoted 36 lines

Display Histogram is an array of bins and can be generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count.
For a maximum resolution of 10k (10240 x 4320 = 44236800) 25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered. max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is 255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 29 lines
Quoted 29 lines
Quoted 29 lines
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines

ec42

Quoted 4 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1364

Quoted 10 lines
include/uapi/drm/drm_mode.h +7 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
+/*
+ * Maximum resolution at present 10k, 10240x4320 =
+44236800
+ * can be denoted in 25bits. With an additional 7 bits in
+buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 4 lines
Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image. We are trying to enhance the image quality of SDR with the support of

Quoted 6 lines
Quoted 5 lines

histogram.

You are defining generic API. It might be broader than your existing usecase. Please consider supporting HDR too.

HDR image enhancement is very much complex including multiple stages such

Quoted 9 lines

as image tone mapping and image denoising.

Quoted 1 line

Here for SDR planes, image enhancement is done by playing around the

contrast and color.

Quoted 1 line

Maybe at this stage we can focus on SDR and can take this HDR at the next

stage.

If you define max(colour) to be 255, then you can not expand it later. The API will have 8 bits for colour information everywhere.

Ok will rework to consider the HDR planes as well.

Thanks and Regards,
Arun R Murthy
--------------------

#10 lore raw

RE: [PATCH v7 01/14] drm: Define histogram structures exposed to user

Quoted 89 lines
Quoted 41 lines · Arun R
Quoted 40 lines
Quoted 40 lines · Arun R
Quoted 39 lines
Quoted 39 lines
Quoted 39 lines
Quoted 39 lines · Arun R Murthy
Quoted 38 lines

Display Histogram is an array of bins and can be
generated in many ways referred to as modes.
Ex: HSV max(RGB), Wighted RGB etc.

Understanding the histogram data format(Ex: HSV max(RGB)) Histogram is just the pixel count. For a maximum resolution of 10k (10240 x 4320 = 44236800)
25 bits should be sufficient to represent this along with a buffer of
7 bits(future use) u32 is being considered. max(RGB) can be 255 i.e 0xFF 8 bit, considering the most significant 5 bits, hence 32 bins.
Below mentioned algorithm illustrates the histogram generation in hardware.

hist[32] = {0};

for (i = 0; i < resolution; i++) {
    bin = max(RGB[i]);
    bin = bin >> 3; /* consider the most significant bits */
    hist[bin]++;
}

If the entire image is Red color then max(255,0,0) is
255 so the pixel count of each pixels will be placed in the last bin. Hence except hist[31] all other bins will have a value zero. Generated histogram in this case would be hist[32] = {0,0,....44236800}

Description of the structures, properties defined are documented in the header file include/uapi/drm/drm_mode.h

---
 include/uapi/drm/drm_mode.h | 59
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
Quoted 7 lines
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

c082810c08a8b234ef2672ecf54fc8c05ddc2bd3..7a7039381142bb5dba269bda

Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 25 lines
Quoted 21 lines
Quoted 8 lines

ec42

Quoted 7 lines

c18be34e2d05 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1355,6 +1355,65 @@ struct drm_mode_closefb {
    __u32 pad;
 };
 
Quoted 13 lines

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1364

Quoted 7 lines
include/uapi/drm/drm_mode.h +7 -0
+/*
+ * Maximum resolution at present 10k, 10240x4320 =
+44236800
+ * can be denoted in 25bits. With an additional 7 bits
+in buffer each bin
+ * can be a u32 value.
+ * Maximum value of max(RGB) is 255, so max 255 bins.

HDR planes have higher max value for a component. Likewise even in an RGB24 case there are 256 possible values. It's not clear why
0 gets excluded.

Quoted 1 line

This applies to only SDR and excludes HDR.

Why?

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines

We are limiting to only SDR. HDR includes a broad range of color and finer details, which essentially means its an enhanced image. We are trying to enhance the image quality of SDR with the support of

histogram.

You are defining generic API. It might be broader than your existing

usecase.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

Please consider supporting HDR too.

HDR image enhancement is very much complex including multiple stages such

as image tone mapping and image denoising.

Quoted 1 line

Here for SDR planes, image enhancement is done by playing around the

contrast and color.

Quoted 9 lines
Quoted 7 lines
Quoted 2 lines

Maybe at this stage we can focus on SDR and can take this HDR at the next

stage.

If you define max(colour) to be 255, then you can not expand it later. The API will have 8 bits for colour information everywhere.

Ok will rework to consider the HDR planes as well.

This 255 is just an array count.
For HDR max(RGB) is 65535 so if for illustration considering 10k resolution generated histogram would look like hist[65535] = {0,0,.....44236800}

So the present structure and elements should be sufficient.

Thanks and Regards,
Arun R Murthy
--------------------

#11 lore raw

PATCH v7 2/14 drm: Define ImageEnhancemenT LUT structures exposed to user

In reply to Arun R Murthy

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc
The list of supported mode by hardware along with the format(exponent mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
include/uapi/drm/drm_mode.h +50 -0
index 7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc0089f1838c2547bcf8 100644
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 << 0)
 
+/* LUT values are points on exponential graph with x axis and y-axis y=f(x) */
+#define DRM_MODE_IET_LOOKUP_LUT				(1 << 0)
+/*
+ * LUT values, points on negative exponential graph with x-axis and y-axis
+ * y = y/x so upon multiplying x, y is obtained, hence multiplicative. The
+ * format of LUT can at max be 8.24(8integer 24 fractional) represented by
+ * u32. Depending on the hardware capability and exponent mantissa can be
+ * chosen.
+ */
+#define DRM_MODE_IET_MULTIPLICATIVE			(1 << 1)
+
 /**
  * struct drm_histogram_caps
  *
@@ -1414,6 +1425,45 @@ struct drm_histogram {
 	__u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above macros
+ * @iet_sample_format: holds the address of an array of u32 LUT sample formats
+ *		       depending on the hardware capability. Max being 8.24
+ *		       Doing a bitwise AND will get the present sample.
+ *		       Ex: for 1 integer 9 fraction AND with 0x10001FF
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported by the
+ *			   hardware
+ * @nr_iet_lut_entries: number of LUT entries
+ */
+struct drm_iet_caps {
+	__u8 iet_mode;
+	u64 iet_sample_format;
+	__u32 nr_iet_sample_formats;
+	__u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample
+ * @iet_mode: image enhancement mode, this will also convey the channel.
+ * @iet_format: LUT exponent and mantissa format, max being 8.24
+ * @data_ptr: pointer to the array of values which is of type u32.
+ *	      1 channel: 10 bit corrected value and remaining bits are reserved.
+ *	      multi channel: pointer to struct drm_color_lut
+ * @nr_elements: number of entries pointed by the data @data_ptr
+ * @reserved: reserved for future use
+ * @reserved1: reserved for future use
+ */
+struct drm_iet_1dlut_sample {
+	__u8 iet_mode;
+	__u32 iet_format;
+	__u64 data_ptr;
+	__u32 nr_elements;
+	__u32 reserved;
+	__u32 reserved1;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Signature · 1 line
2.25.1
#12 lore raw

Re: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 31 lines · Arun R Murthy

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc
The list of supported mode by hardware along with the format(exponent mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
index 7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc0089f1838c2547bcf8 100644

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1370

Quoted 4 lines
include/uapi/drm/drm_mode.h +1 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 << 0)
 
+/* LUT values are points on exponential graph with x axis and y-axis y=f(x) */

Huh?

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1371

Quoted 1 line
include/uapi/drm/drm_mode.h +1 -0
+#define DRM_MODE_IET_LOOKUP_LUT				(1 << 0)

Again, what is the reason for a shift? Can these values be OR'd?

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1374

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+/*
+ * LUT values, points on negative exponential graph with x-axis and y-axis
+ * y = y/x so upon multiplying x, y is obtained, hence multiplicative. The

Can't parse this sentence.

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1377

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ * format of LUT can at max be 8.24(8integer 24 fractional) represented by
+ * u32. Depending on the hardware capability and exponent mantissa can be
+ * chosen.

What does that mean? How is it choosen?

Quoted 8 lines
include/uapi/drm/drm_mode.h +3 -0
+ */
+#define DRM_MODE_IET_MULTIPLICATIVE			(1 << 1)
+
 /**
  * struct drm_histogram_caps
  *
@@ -1414,6 +1425,45 @@ struct drm_histogram {
 	__u32 nr_elements;

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1435

Quoted 10 lines
include/uapi/drm/drm_mode.h +8 -0
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above macros
+ * @iet_sample_format: holds the address of an array of u32 LUT sample formats
+ *		       depending on the hardware capability. Max being 8.24
+ *		       Doing a bitwise AND will get the present sample.
+ *		       Ex: for 1 integer 9 fraction AND with 0x10001FF

?? Can hardware support 16.16? 32.0?

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 13 lines
include/uapi/drm/drm_mode.h +13 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported by the
+ *			   hardware
+ * @nr_iet_lut_entries: number of LUT entries
+ */
+struct drm_iet_caps {
+	__u8 iet_mode;
+	u64 iet_sample_format;
+	__u32 nr_iet_sample_formats;
+	__u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

Quoted 25 lines
include/uapi/drm/drm_mode.h +18 -0
+ * @iet_mode: image enhancement mode, this will also convey the channel.
+ * @iet_format: LUT exponent and mantissa format, max being 8.24
+ * @data_ptr: pointer to the array of values which is of type u32.
+ *	      1 channel: 10 bit corrected value and remaining bits are reserved.
+ *	      multi channel: pointer to struct drm_color_lut
+ * @nr_elements: number of entries pointed by the data @data_ptr
+ * @reserved: reserved for future use
+ * @reserved1: reserved for future use
+ */
+struct drm_iet_1dlut_sample {
+	__u8 iet_mode;
+	__u32 iet_format;
+	__u64 data_ptr;
+	__u32 nr_elements;
+	__u32 reserved;
+	__u32 reserved1;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Signature · 2 lines
-- 
2.25.1
Signature · 2 lines
With best wishes
Dmitry
#13 lore raw

RE: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 41 lines · Arun R Murthy
Quoted 28 lines

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc The list of supported mode by hardware along with the format(exponent
mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
Quoted 2 lines

index

7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc00 89f1

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 6 lines

838c2547bcf8 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB			(1 <<

0)

Quoted 1 line
Quoted 5 lines

include/uapi/drm/drm_mode.h

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+/* LUT values are points on exponential graph with x axis and y-axis
+y=f(x) */

Huh?

This f(x) can be the algorithm defined by the user space algorithm to generate the lookup table. Generation of the LUT value is left to the user space algorithm. When this LUT table is passed to the hardware its just signifies how hardware should use this table to get the LUT value. In this mode it's a direct lookup table.

Quoted 5 lines
Quoted 1 line

+#define DRM_MODE_IET_LOOKUP_LUT (1 <<

0)

Again, what is the reason for a shift? Can these values be OR'd?

Yes can be OR'd values as well.
Let me know if this has to be changed?
Just chose bitwise shift to denote the multiple modes.

Quoted 8 lines

include/uapi/drm/drm_mode.h

Quoted 5 lines
include/uapi/drm/drm_mode.h +5 -0
+/*
+ * LUT values, points on negative exponential graph with x-axis and
+y-axis
+ * Y = y/x so upon multiplying x, y is obtained, hence
+multiplicative. The

Can't parse this sentence.

We need x and y points in the exponential graph.
For retrieving the value Y on the graph the value passed by the user is in the format y/x In order to get the Y points on the graph the value has to be multiplied by x. This is a floating point value when compared with an integer value with the direct lookup mode.

Quoted 8 lines
Quoted 5 lines

+ * format of LUT can at max be 8.24(8integer 24 fractional)
+ represented by
+ * u32. Depending on the hardware capability and exponent mantissa
+ can be
+ * chosen.

What does that mean? How is it choosen?

The max value that these kind of 1DLUT can be is 8.24
Hardware design can choose anything within this range. This depends on the accuracy required by hardware keeping in mind the hardware cost for implementation.
Just a precision for 32bit value.

Quoted 17 lines
Quoted 7 lines

+ */

include/uapi/drm/drm_mode.h +2 -0
+#define DRM_MODE_IET_MULTIPLICATIVE			(1 << 1)
+
 /**
  * struct drm_histogram_caps
  *

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1433

Quoted 9 lines
include/uapi/drm/drm_mode.h +6 -0
@@ -1414,6 +1425,45 @@ struct drm_histogram {
 	__u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above
+macros
+ * @iet_sample_format: holds the address of an array of u32 LUT sample

formats

Quoted 6 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1435

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ *		       depending on the hardware capability. Max being 8.24
+ *		       Doing a bitwise AND will get the present sample.
+ *		       Ex: for 1 integer 9 fraction AND with 0x10001FF

?? Can hardware support 16.16? 32.0?

No, for a 1D LUT maximum floating number can be 8.24 Hence hardware will have to adhere to anything within this range.

Quoted 18 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1437

Quoted 1 line
include/uapi/drm/drm_mode.h +1 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported by

the

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 11 lines
include/uapi/drm/drm_mode.h +11 -0
+ *			   hardware
+ * @nr_iet_lut_entries: number of LUT entries  */ struct drm_iet_caps
+{
+	__u8 iet_mode;
+	u64 iet_sample_format;
+	__u32 nr_iet_sample_formats;
+	__u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

The struct iet_1dlut_sample will be used for all the IET modes i.e direct lookup and multiplicative.
The element iet_sample_format will not be applicable for direct lookup. This will be used for multiplicative and the value what it can hold for multiplicative is mentioned in the above description.
I missed adding this info in the description, will add it in the next version.

Quoted 21 lines
Quoted 21 lines
include/uapi/drm/drm_mode.h +17 -0
+ * @iet_mode: image enhancement mode, this will also convey the channel.
+ * @iet_format: LUT exponent and mantissa format, max being 8.24
+ * @data_ptr: pointer to the array of values which is of type u32.
+ *	      1 channel: 10 bit corrected value and remaining bits are reserved.
+ *	      multi channel: pointer to struct drm_color_lut
+ * @nr_elements: number of entries pointed by the data @data_ptr
+ * @reserved: reserved for future use
+ * @reserved1: reserved for future use  */ struct
+drm_iet_1dlut_sample {
+	__u8 iet_mode;
+	__u32 iet_format;
+	__u64 data_ptr;
+	__u32 nr_elements;
+	__u32 reserved;
+	__u32 reserved1;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Quoted 4 lines
Quoted 3 lines

--
2.25.1

Thanks and Regards,
Arun R Murthy
--------------------

#14 lore raw

Re: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 47 lines · Arun R
Quoted 42 lines · Arun R Murthy
Quoted 28 lines

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc The list of supported mode by hardware along with the format(exponent
mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
include/uapi/drm/drm_mode.h +0 -0
Quoted 2 lines

index

7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc00 89f1

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 6 lines

838c2547bcf8 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 1 line
Quoted 5 lines

include/uapi/drm/drm_mode.h

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+/* LUT values are points on exponential graph with x axis and y-axis
+y=f(x) */

Huh?

Quoted 4 lines

This f(x) can be the algorithm defined by the user space algorithm to generate the lookup table. Generation of the LUT value is left to the user space algorithm. When this LUT table is passed to the hardware its just signifies how hardware should use this table to get the LUT value. In this mode it's a direct lookup table.

Your documentation should be describing what is expected from the userspace. What is y, x and f(x)? How is it being used?

Quoted 9 lines
Quoted 6 lines
Quoted 2 lines

+#define DRM_MODE_IET_LOOKUP_LUT (1 <<

0)

Again, what is the reason for a shift? Can these values be OR'd?

Yes can be OR'd values as well.
Let me know if this has to be changed?
Just chose bitwise shift to denote the multiple modes.

What does it mean if drm_iet_1dlut_sample.iet_mode contains OR of two values?

Quoted 9 lines
Quoted 9 lines

include/uapi/drm/drm_mode.h

Quoted 6 lines
include/uapi/drm/drm_mode.h +5 -0
+/*
+ * LUT values, points on negative exponential graph with x-axis and
+y-axis
+ * Y = y/x so upon multiplying x, y is obtained, hence
+multiplicative. The

Can't parse this sentence.

Quoted 5 lines

We need x and y points in the exponential graph.
For retrieving the value Y on the graph the value passed by the user is in the format y/x In order to get the Y points on the graph the value has to be multiplied by x. This is a floating point value when compared with an integer value with the direct lookup mode.

Again, what are x and y? Bin indices? Pixel counts? Number of CPUs in the current generation?

Quoted 11 lines
Quoted 10 lines
Quoted 7 lines

+ * format of LUT can at max be 8.24(8integer 24 fractional)
+ represented by
+ * u32. Depending on the hardware capability and exponent mantissa
+ can be
+ * chosen.

What does that mean? How is it choosen?

The max value that these kind of 1DLUT can be is 8.24

Why?

Quoted 22 lines

Hardware design can choose anything within this range. This depends on the accuracy required by hardware keeping in mind the hardware cost for implementation.
Just a precision for 32bit value.

Quoted 17 lines
Quoted 7 lines

+ */

include/uapi/drm/drm_mode.h +2 -0
+#define DRM_MODE_IET_MULTIPLICATIVE                        (1 << 1)
+
 /**
  * struct drm_histogram_caps
  *

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1433

Quoted 9 lines
include/uapi/drm/drm_mode.h +6 -0
@@ -1414,6 +1425,45 @@ struct drm_histogram {
    __u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above
+macros
+ * @iet_sample_format: holds the address of an array of u32 LUT sample

formats

Quoted 7 lines
Quoted 6 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1435

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ *                depending on the hardware capability. Max being 8.24
+ *                Doing a bitwise AND will get the present sample.
+ *                Ex: for 1 integer 9 fraction AND with 0x10001FF

?? Can hardware support 16.16? 32.0?

No, for a 1D LUT maximum floating number can be 8.24

Why? Is it a limitation of the Intel hardware or just a random API choice?

Quoted 20 lines

Hence hardware will have to adhere to anything within this range.

Quoted 18 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1437

Quoted 1 line
include/uapi/drm/drm_mode.h +1 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported by

the

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 11 lines
include/uapi/drm/drm_mode.h +11 -0
+ *                    hardware
+ * @nr_iet_lut_entries: number of LUT entries  */ struct drm_iet_caps
+{
+   __u8 iet_mode;
+   u64 iet_sample_format;
+   __u32 nr_iet_sample_formats;
+   __u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

Quoted 6 lines

The struct iet_1dlut_sample will be used for all the IET modes i.e direct lookup and multiplicative.
The element iet_sample_format will not be applicable for direct lookup. This will be used for multiplicative and the value what it can hold for multiplicative is mentioned in the above description.
I missed adding this info in the description, will add it in the next version.

And some other formats will also require additional data. This multi-format structure sounds bad from my POV.

Quoted 29 lines
Quoted 22 lines
Quoted 22 lines
include/uapi/drm/drm_mode.h +17 -0
+ * @iet_mode: image enhancement mode, this will also convey the channel.
+ * @iet_format: LUT exponent and mantissa format, max being 8.24
+ * @data_ptr: pointer to the array of values which is of type u32.
+ *       1 channel: 10 bit corrected value and remaining bits are reserved.
+ *       multi channel: pointer to struct drm_color_lut
+ * @nr_elements: number of entries pointed by the data @data_ptr
+ * @reserved: reserved for future use
+ * @reserved1: reserved for future use  */ struct
+drm_iet_1dlut_sample {
+   __u8 iet_mode;
+   __u32 iet_format;
+   __u64 data_ptr;
+   __u32 nr_elements;
+   __u32 reserved;
+   __u32 reserved1;
+};
+
 #if defined(__cplusplus)
 }
 #endif
Quoted 4 lines
Quoted 3 lines

--
2.25.1

Thanks and Regards,
Arun R Murthy
--------------------

Signature · 2 lines
With best wishes
Dmitry
#15 lore raw

RE: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 50 lines
Quoted 28 lines
Quoted 28 lines · Arun R Murthy
Quoted 27 lines

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc The list of supported mode by hardware along with the format(exponent mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc00

Quoted 9 lines
Quoted 9 lines

89f1

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 6 lines

838c2547bcf8 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 1 line
Quoted 8 lines
Quoted 6 lines

include/uapi/drm/drm_mode.h

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+/* LUT values are points on exponential graph with x axis and
+y-axis
+y=f(x) */

Huh?

This f(x) can be the algorithm defined by the user space algorithm to generate the lookup table. Generation of the LUT value is left to the user

Quoted 8 lines

space algorithm.

Quoted 2 lines

When this LUT table is passed to the hardware its just signifies how hardware should use this table to get the LUT value. In this mode it's a direct

lookup table.

Your documentation should be describing what is expected from the userspace. What is y, x and f(x)? How is it being used?

Sure will add the above explanation in the patch documentation.

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines

+#define DRM_MODE_IET_LOOKUP_LUT (1 <<

0)

Quoted 9 lines
Quoted 5 lines
Quoted 2 lines

Again, what is the reason for a shift? Can these values be OR'd?

Yes can be OR'd values as well.
Let me know if this has to be changed?
Just chose bitwise shift to denote the multiple modes.

What does it mean if drm_iet_1dlut_sample.iet_mode contains OR of two values?

iet_mode in struct drm_iet_caps can be OR of two such modes, which means that the hardware supports both of the modes. Drm_iet_1dlut_sample.iet_mode tells the hardware which iet mode is used in generating the LUT value. Because hardware will have to interpret the LUT value based on the mode.

Quoted 12 lines
Quoted 9 lines
Quoted 9 lines

include/uapi/drm/drm_mode.h

Quoted 6 lines
include/uapi/drm/drm_mode.h +5 -0
+/*
+ * LUT values, points on negative exponential graph with x-axis
+and y-axis
+ * Y = y/x so upon multiplying x, y is obtained, hence
+multiplicative. The

Can't parse this sentence.

Quoted 3 lines

We need x and y points in the exponential graph.
For retrieving the value Y on the graph the value passed by the user is in the format y/x In order to get the Y points on the graph the value has to

Quoted 7 lines

be multiplied by x.

Quoted 2 lines

This is a floating point value when compared with an integer value with the direct lookup mode.

Again, what are x and y? Bin indices? Pixel counts? Number of CPUs in the current generation?

It depends on the mode for direct lookup both x and y are pixels and for multiplicative mode X co-ordinate is proportional to the pixel value and the Y co-ordinate is the multiplier factor, i.e X-axis in pixels and Y-axis is OutPixel/InPixel

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

+ * format of LUT can at max be 8.24(8integer 24 fractional)
+ represented by
+ * u32. Depending on the hardware capability and exponent
+ mantissa can be
+ * chosen.

Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

What does that mean? How is it choosen?

The max value that these kind of 1DLUT can be is 8.24

Why?

32bit is the container and within this if we choose 16.16 then it doesn’t make sense to boost the pixel by 2^16
Hence set aside 8 bit for integer 2^8 thereby boosting the pixel by 255 and that’s a huge boost factor.
Remaining 24 bits out of the container 32 is fractional value. This is the optimal value for implementing in the hardware as well as per the hardware design.

Quoted 27 lines
Quoted 27 lines

Hardware design can choose anything within this range. This depends on the accuracy required by hardware keeping in mind the hardware cost for implementation.
Just a precision for 32bit value.

Quoted 22 lines
Quoted 7 lines

+ */

include/uapi/drm/drm_mode.h +2 -0
+#define DRM_MODE_IET_MULTIPLICATIVE                        (1 << 1)
+
 /**
  * struct drm_histogram_caps
  *

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1434

Quoted 10 lines
include/uapi/drm/drm_mode.h +7 -0
@@ -1414,6 +1425,45 @@ struct drm_histogram {
    __u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above
+macros
+ * @iet_sample_format: holds the address of an array of u32 LUT
+sample

formats

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ *                depending on the hardware capability. Max being 8.24
+ *                Doing a bitwise AND will get the present sample.
+ *                Ex: for 1 integer 9 fraction AND with 0x10001FF
Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

?? Can hardware support 16.16? 32.0?

No, for a 1D LUT maximum floating number can be 8.24

Why? Is it a limitation of the Intel hardware or just a random API choice?

As explained above this an optimal value yielding to a huge boost factor of 255.99. This is as per the hardware design.

Quoted 28 lines
Quoted 21 lines

Hence hardware will have to adhere to anything within this range.

Quoted 19 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1439

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported
+ by

the

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 11 lines
include/uapi/drm/drm_mode.h +11 -0
+ *                    hardware
+ * @nr_iet_lut_entries: number of LUT entries  */ struct
+drm_iet_caps {
+   __u8 iet_mode;
+   u64 iet_sample_format;
+   __u32 nr_iet_sample_formats;
+   __u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

Quoted 6 lines

The struct iet_1dlut_sample will be used for all the IET modes i.e direct lookup and multiplicative.
The element iet_sample_format will not be applicable for direct lookup. This will be used for multiplicative and the value what it can hold for multiplicative is mentioned in the above description.
I missed adding this info in the description, will add it in the next version.

Quoted 3 lines

And some other formats will also require additional data. This multi-format structure sounds bad from my POV.

Will try generalize this structure across the modes.
Its only for direct lookup mode we will not need any iet_sample_format. For other modes such as multiplicative, additive etc we will need to mention the iet_sample_format.
Top view of this LUT is just a lookup table which says for a particular pixel value in the LUT table, what is the output value and this output value is the pixel value that is replaced in the incoming image.
Now generation of this LUT can be done different methods referred to as modes over here.
So one variable to mention the type of mode and other to specify the internal details of the selected mode.
Will reframe accordingly.

Thanks and Regards,
Arun R Murthy
--------------------

#16 lore raw

Re: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 70 lines · Arun R
Quoted 50 lines
Quoted 28 lines
Quoted 28 lines · Arun R Murthy
Quoted 27 lines

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc The list of supported mode by hardware along with the format(exponent mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa.
For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc00

Quoted 9 lines
Quoted 9 lines

89f1

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 6 lines

838c2547bcf8 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 1 line
Quoted 8 lines
Quoted 6 lines

include/uapi/drm/drm_mode.h

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+/* LUT values are points on exponential graph with x axis and
+y-axis
+y=f(x) */

Huh?

This f(x) can be the algorithm defined by the user space algorithm to generate the lookup table. Generation of the LUT value is left to the user

Quoted 8 lines

space algorithm.

Quoted 2 lines

When this LUT table is passed to the hardware its just signifies how hardware should use this table to get the LUT value. In this mode it's a direct

lookup table.

Your documentation should be describing what is expected from the userspace. What is y, x and f(x)? How is it being used?

Sure will add the above explanation in the patch documentation.

Quoted 10 lines
Quoted 9 lines
Quoted 6 lines
Quoted 2 lines

+#define DRM_MODE_IET_LOOKUP_LUT (1 <<

0)

Again, what is the reason for a shift? Can these values be OR'd?

Yes can be OR'd values as well.
Let me know if this has to be changed?
Just chose bitwise shift to denote the multiple modes.

Quoted 8 lines
Quoted 3 lines

What does it mean if drm_iet_1dlut_sample.iet_mode contains OR of two values?

iet_mode in struct drm_iet_caps can be OR of two such modes, which means that the hardware supports both of the modes. Drm_iet_1dlut_sample.iet_mode tells the hardware which iet mode is used in generating the LUT value. Because hardware will have to interpret the LUT value based on the mode.

Yes. That's why I asked about the drm_iet_1dlut_sample.iet_mode, not the caps. It makes no sense to allow ORing several modes there. So the list of modes should be a simple enum and caps should use BIT(val).

Quoted 17 lines
Quoted 17 lines
Quoted 10 lines
Quoted 10 lines

include/uapi/drm/drm_mode.h

Quoted 7 lines
include/uapi/drm/drm_mode.h +5 -0
+/*
+ * LUT values, points on negative exponential graph with x-axis
+and y-axis
+ * Y = y/x so upon multiplying x, y is obtained, hence
+multiplicative. The

Can't parse this sentence.

Quoted 3 lines

We need x and y points in the exponential graph.
For retrieving the value Y on the graph the value passed by the user is in the format y/x In order to get the Y points on the graph the value has to

be multiplied by x.

Quoted 2 lines

This is a floating point value when compared with an integer value with the direct lookup mode.

Quoted 7 lines
Quoted 3 lines

Again, what are x and y? Bin indices? Pixel counts? Number of CPUs in the current generation?

It depends on the mode for direct lookup both x and y are pixels and for multiplicative mode X co-ordinate is proportional to the pixel value and the Y co-ordinate is the multiplier factor, i.e X-axis in pixels and Y-axis is OutPixel/InPixel

Please expand the description. An engineer, who has no Intel documentation, should be able to understand your description.

Quoted 12 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

+ * format of LUT can at max be 8.24(8integer 24 fractional)
+ represented by
+ * u32. Depending on the hardware capability and exponent
+ mantissa can be
+ * chosen.

Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

What does that mean? How is it choosen?

The max value that these kind of 1DLUT can be is 8.24

Why?

Quoted 7 lines

32bit is the container and within this if we choose 16.16 then it doesn’t make sense to boost the pixel by 2^16
Hence set aside 8 bit for integer 2^8 thereby boosting the pixel by 255 and that’s a huge boost factor.
Remaining 24 bits out of the container 32 is fractional value. This is the optimal value for implementing in the hardware as well as per the hardware design.

Generic API means that there is no particular hardware design. However the rest of the description makes sense. Please add it to the commit message.

Quoted 65 lines
Quoted 28 lines
Quoted 28 lines

Hardware design can choose anything within this range. This depends on the accuracy required by hardware keeping in mind the hardware cost for implementation.
Just a precision for 32bit value.

Quoted 22 lines
Quoted 7 lines

+ */

include/uapi/drm/drm_mode.h +2 -0
+#define DRM_MODE_IET_MULTIPLICATIVE                        (1 << 1)
+
 /**
  * struct drm_histogram_caps
  *

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1434

Quoted 10 lines
include/uapi/drm/drm_mode.h +7 -0
@@ -1414,6 +1425,45 @@ struct drm_histogram {
    __u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the above
+macros
+ * @iet_sample_format: holds the address of an array of u32 LUT
+sample

formats

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ *                depending on the hardware capability. Max being 8.24
+ *                Doing a bitwise AND will get the present sample.
+ *                Ex: for 1 integer 9 fraction AND with 0x10001FF
Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

?? Can hardware support 16.16? 32.0?

No, for a 1D LUT maximum floating number can be 8.24

Why? Is it a limitation of the Intel hardware or just a random API choice?

As explained above this an optimal value yielding to a huge boost factor of 255.99. This is as per the hardware design.

Quoted 28 lines
Quoted 27 lines

Hence hardware will have to adhere to anything within this range.

Quoted 19 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1439

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts supported
+ by

the

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 11 lines
include/uapi/drm/drm_mode.h +11 -0
+ *                    hardware
+ * @nr_iet_lut_entries: number of LUT entries  */ struct
+drm_iet_caps {
+   __u8 iet_mode;
+   u64 iet_sample_format;
+   __u32 nr_iet_sample_formats;
+   __u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

The struct iet_1dlut_sample will be used for all the IET modes i.e direct lookup and multiplicative.
The element iet_sample_format will not be applicable for direct lookup. This will be used for multiplicative and the value what it can hold for multiplicative is mentioned in the above description.
I missed adding this info in the description, will add it in the next version.

Quoted 7 lines
Quoted 3 lines

And some other formats will also require additional data. This multi-format structure sounds bad from my POV.

Will try generalize this structure across the modes.
Its only for direct lookup mode we will not need any iet_sample_format. For other modes such as multiplicative, additive etc we will need to mention the iet_sample_format.

There might be other modes which require more data.

Quoted 12 lines

Top view of this LUT is just a lookup table which says for a particular pixel value in the LUT table, what is the output value and this output value is the pixel value that is replaced in the incoming image.
Now generation of this LUT can be done different methods referred to as modes over here.
So one variable to mention the type of mode and other to specify the internal details of the selected mode.
Will reframe accordingly.

Thanks and Regards,
Arun R Murthy
--------------------

Signature · 2 lines
With best wishes
Dmitry
#17 lore raw

RE: [PATCH v7 02/14] drm: Define ImageEnhancemenT LUT structures exposed to user

Quoted 85 lines · Arun R
Quoted 29 lines
Quoted 29 lines
Quoted 29 lines
Quoted 29 lines · Arun R Murthy
Quoted 28 lines

ImageEnhancemenT(IET) hardware interpolates the LUT value to generate the enhanced output image. LUT takes an input value, outputs a new value based on the data within the LUT. 1D LUT can remap individual input values to new output values based on the LUT sample. LUT can be interpolated by the hardware by multiple modes Ex: Direct Lookup LUT, Multiplicative LUT etc The list of supported mode by hardware along with the format(exponent
mantissa) is exposed to user by the iet_lut_caps property. Maximum format being 8.24 i.e 8 exponent and 24 mantissa. For illustration a hardware supporting 1.9 format denotes this as 0x10001FF. In order to know the exponent do a bitwise AND with 0xF000000. The LUT value to be provided by user would be a 10bit value with 1 bit integer and 9 bit fractional value.

Multiple formats can be supported, hence pointer is used over here. User can then provide the LUT with any one of the supported modes in any of the supported formats.
The entries in the LUT can vary depending on the hardware capability with max being 255. This will also be exposed as iet_lut_caps so user can generate a LUT with the specified entries.

---
 include/uapi/drm/drm_mode.h | 50
+++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
(quoted diff) +0 -0

b/include/uapi/drm/drm_mode.h index

7a7039381142bb5dba269bdaec42c18be34e2d05..056c2efef1589848034afc00

Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 9 lines

89f1

include/uapi/drm/drm_mode.h struct drm_mode_closefb { · line 1368

Quoted 6 lines

838c2547bcf8 100644

include/uapi/drm/drm_mode.h +0 -0
@@ -1367,6 +1367,17 @@ struct drm_mode_closefb {
  */
 #define DRM_MODE_HISTOGRAM_HSV_MAX_RGB                     (1 <<

0)

Quoted 1 line
Quoted 6 lines

include/uapi/drm/drm_mode.h

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+/* LUT values are points on exponential graph with x axis and
+y-axis
+y=f(x) */

Huh?

This f(x) can be the algorithm defined by the user space algorithm to generate the lookup table. Generation of the LUT value is left to the user

Quoted 7 lines
Quoted 7 lines

space algorithm.

Quoted 3 lines

When this LUT table is passed to the hardware its just signifies how hardware should use this table to get the LUT value. In this mode it's a direct

lookup table.

Your documentation should be describing what is expected from the

userspace.

Quoted 14 lines
Quoted 2 lines

What is y, x and f(x)? How is it being used?

Sure will add the above explanation in the patch documentation.

Quoted 10 lines
Quoted 9 lines
Quoted 6 lines
Quoted 2 lines

+#define DRM_MODE_IET_LOOKUP_LUT (1 <<

0)

Again, what is the reason for a shift? Can these values be OR'd?

Yes can be OR'd values as well.
Let me know if this has to be changed?
Just chose bitwise shift to denote the multiple modes.

Quoted 8 lines
Quoted 3 lines

What does it mean if drm_iet_1dlut_sample.iet_mode contains OR of two values?

iet_mode in struct drm_iet_caps can be OR of two such modes, which means that the hardware supports both of the modes. Drm_iet_1dlut_sample.iet_mode tells the hardware which iet mode is used in generating the LUT value. Because hardware will have to interpret the LUT value based on the mode.

Quoted 4 lines

Yes. That's why I asked about the drm_iet_1dlut_sample.iet_mode, not the caps. It makes no sense to allow ORing several modes there. So the list of modes should be a simple enum and caps should use BIT(val).

Ok sure will change accordingly in the next revision.

Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 10 lines
Quoted 10 lines

include/uapi/drm/drm_mode.h

Quoted 7 lines
include/uapi/drm/drm_mode.h +5 -0
+/*
+ * LUT values, points on negative exponential graph with
+x-axis and y-axis
+ * Y = y/x so upon multiplying x, y is obtained, hence
+multiplicative. The

Can't parse this sentence.

Quoted 4 lines

We need x and y points in the exponential graph.
For retrieving the value Y on the graph the value passed by the user is in the format y/x In order to get the Y points on the graph the value has to

be multiplied by x.

Quoted 2 lines

This is a floating point value when compared with an integer value with the direct lookup mode.

Again, what are x and y? Bin indices? Pixel counts? Number of CPUs in the current generation?

Quoted 8 lines
Quoted 4 lines

It depends on the mode for direct lookup both x and y are pixels and for multiplicative mode X co-ordinate is proportional to the pixel value and the Y co-ordinate is the multiplier factor, i.e X-axis in pixels and Y-axis is OutPixel/InPixel

Please expand the description. An engineer, who has no Intel documentation, should be able to understand your description.

Sure will add this in the description and in the patch commit message as well in the next revision.

Quoted 20 lines
Quoted 12 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

+ * format of LUT can at max be 8.24(8integer 24 fractional)
+ represented by
+ * u32. Depending on the hardware capability and exponent
+ mantissa can be
+ * chosen.

Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

What does that mean? How is it choosen?

The max value that these kind of 1DLUT can be is 8.24

Why?

Quoted 7 lines

32bit is the container and within this if we choose 16.16 then it doesn’t make sense to boost the pixel by 2^16 Hence set aside 8 bit for integer 2^8 thereby boosting the pixel by 255 and that’s a huge boost factor.
Remaining 24 bits out of the container 32 is fractional value. This is the optimal value for implementing in the hardware as well as per the hardware design.

Quoted 3 lines

Generic API means that there is no particular hardware design. However the rest of the description makes sense. Please add it to the commit message.

Sure will do it in the next revision!

Quoted 68 lines
Quoted 68 lines
Quoted 28 lines
Quoted 28 lines

Hardware design can choose anything within this range. This depends on the accuracy required by hardware keeping in mind the hardware cost for implementation.
Just a precision for 32bit value.

Quoted 22 lines
Quoted 7 lines

+ */

include/uapi/drm/drm_mode.h +2 -0
+#define DRM_MODE_IET_MULTIPLICATIVE                        (1 << 1)
+
 /**
  * struct drm_histogram_caps
  *

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1434

Quoted 10 lines
include/uapi/drm/drm_mode.h +7 -0
@@ -1414,6 +1425,45 @@ struct drm_histogram {
    __u32 nr_elements;
 };
 
+/**
+ * struct drm_iet_caps
+ *
+ * @iet_mode: pixel factor enhancement modes defined in the
+above macros
+ * @iet_sample_format: holds the address of an array of u32
+LUT sample

formats

Quoted 3 lines
include/uapi/drm/drm_mode.h +3 -0
+ *                depending on the hardware capability. Max being 8.24
+ *                Doing a bitwise AND will get the present sample.
+ *                Ex: for 1 integer 9 fraction AND with 0x10001FF
Quoted 6 lines
Quoted 3 lines
Quoted 2 lines

?? Can hardware support 16.16? 32.0?

No, for a 1D LUT maximum floating number can be 8.24

Why? Is it a limitation of the Intel hardware or just a random API choice?

As explained above this an optimal value yielding to a huge boost factor of 255.99. This is as per the hardware design.

Quoted 31 lines
Quoted 21 lines

Hence hardware will have to adhere to anything within this range.

Quoted 19 lines

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1439

Quoted 2 lines
include/uapi/drm/drm_mode.h +2 -0
+ * @nr_iet_sample_formats: number of iet_sample_formsts
+ supported by

the

include/uapi/drm/drm_mode.h struct drm_histogram { · line 1448

Quoted 11 lines
include/uapi/drm/drm_mode.h +11 -0
+ *                    hardware
+ * @nr_iet_lut_entries: number of LUT entries  */ struct
+drm_iet_caps {
+   __u8 iet_mode;
+   u64 iet_sample_format;
+   __u32 nr_iet_sample_formats;
+   __u32 nr_iet_lut_entries;
+};
+
+/**
+ * struct drm_iet_1dlut_sample

Is it supposed to be used with DRM_MODE_IET_MULTIPLICATIVE only? Or is it supposed to be used with DRM_MODE_IET_LOOKUP_LUT? In the latter case what should be the iet_format value?

Quoted 6 lines

The struct iet_1dlut_sample will be used for all the IET modes i.e direct lookup and multiplicative.
The element iet_sample_format will not be applicable for direct lookup. This will be used for multiplicative and the value what it can hold for multiplicative is mentioned in the above description. I missed adding this info in the description, will add it in the next version.

And some other formats will also require additional data. This multi-format structure sounds bad from my POV.

Quoted 7 lines
Quoted 4 lines

Will try generalize this structure across the modes.
Its only for direct lookup mode we will not need any iet_sample_format. For other modes such as multiplicative, additive etc we will need to mention the iet_sample_format.

There might be other modes which require more data.

Yes agree.
That’s the reason iet_sample_format is u64 which holds the address of the format. For each individual mode that we add we will have to define the format as well. For now the modes that we add direct lookup and multiplicative will be added in this patch.
Will add these description in the next revision of the patch.

Thanks and Regards,
Arun R Murthy
--------------------

#18 lore raw

PATCH v7 3/14 drm/crtc: Expose API to create drm crtc property for histogram

In reply to Arun R Murthy

Add drm-crtc property for histogram and for the properties added add the corresponding get/set_property.

---
 drivers/gpu/drm/drm_atomic_state_helper.c | 14 ++++++
 drivers/gpu/drm/drm_atomic_uapi.c         | 15 +++++++
 drivers/gpu/drm/drm_crtc.c                | 73 +++++++++++++++++++++++++++++++
 include/drm/drm_crtc.h                    | 44 +++++++++++++++++++
 4 files changed, 146 insertions(+)
drivers/gpu/drm/drm_atomic_state_helper.c +14 -0
index 519228eb109533d2596e899a57b571fa0995824f..dfe6293f7a42d034da3de593094019ca15014a02 100644
@@ -143,6 +143,12 @@ void __drm_atomic_helper_crtc_duplicate_state(struct drm_crtc *crtc,
 		drm_property_blob_get(state->ctm);
 	if (state->gamma_lut)
 		drm_property_blob_get(state->gamma_lut);
+	if (state->histogram_caps)
+		drm_property_blob_get(state->histogram_caps);
+	if (state->histogram_enable)
+		drm_property_blob_get(state->histogram_enable);
+	if (state->histogram_data)
+		drm_property_blob_get(state->histogram_data);
 	state->mode_changed = false;
 	state->active_changed = false;
 	state->planes_changed = false;
@@ -156,6 +162,8 @@ void __drm_atomic_helper_crtc_duplicate_state(struct drm_crtc *crtc,
 	/* Self refresh should be canceled when a new update is available */
 	state->active = drm_atomic_crtc_effectively_active(state);
 	state->self_refresh_active = false;
+
+	state->histogram_updated = false;
 }
 EXPORT_SYMBOL(__drm_atomic_helper_crtc_duplicate_state);
 
@@ -215,6 +223,12 @@ void __drm_atomic_helper_crtc_destroy_state(struct drm_crtc_state *state)
 	drm_property_blob_put(state->degamma_lut);
 	drm_property_blob_put(state->ctm);
 	drm_property_blob_put(state->gamma_lut);
+	if (state->histogram_caps)
+		drm_property_blob_put(state->histogram_caps);
+	if (state->histogram_enable)
+		drm_property_blob_put(state->histogram_enable);
+	if (state->histogram_data)
+		drm_property_blob_put(state->histogram_data);
 }
 EXPORT_SYMBOL(__drm_atomic_helper_crtc_destroy_state);
 
drivers/gpu/drm/drm_atomic_uapi.c +15 -0
index 370dc676e3aa543c9827b50df20df78f02b738c9..459d30898196c94392a7f916b1fa9ca3a334eea8 100644
@@ -415,6 +415,15 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc,
 			return -EFAULT;
 
 		set_out_fence_for_crtc(state->state, crtc, fence_ptr);
+	} else if (property == crtc->histogram_enable_property) {
+		ret = drm_property_replace_blob_from_id(dev,
+							&state->histogram_enable,
+							val,
+							-1,
+							sizeof(struct drm_histogram_config),
+							&replaced);
+		state->histogram_updated |= replaced;
+		return ret;
 	} else if (property == crtc->scaling_filter_property) {
 		state->scaling_filter = val;
 	} else if (crtc->funcs->atomic_set_property) {
@@ -452,6 +461,12 @@ drm_atomic_crtc_get_property(struct drm_crtc *crtc,
 		*val = (state->gamma_lut) ? state->gamma_lut->base.id : 0;
 	else if (property == config->prop_out_fence_ptr)
 		*val = 0;
+	else if (property == crtc->histogram_caps_property)
+		*val = (state->histogram_caps) ? state->histogram_caps->base.id : 0;
+	else if (property == crtc->histogram_enable_property)
+		*val = (state->histogram_enable) ? state->histogram_enable->base.id : 0;
+	else if (property == crtc->histogram_data_property)
+		*val = (state->histogram_data) ? state->histogram_data->base.id : 0;
 	else if (property == crtc->scaling_filter_property)
 		*val = state->scaling_filter;
 	else if (crtc->funcs->atomic_get_property)
drivers/gpu/drm/drm_crtc.c +73 -0

76 lines not on this page · raw message

include/drm/drm_crtc.h +44 -0

61 lines not on this page · raw message

Signature · 1 line
2.25.1
#19 lore raw

PATCH v7 4/14 drm/crtc: Expose API to create drm crtc property for IET LUT

In reply to Arun R Murthy

Add drm-crtc property for IET 1DLUT and for the properties added add corresponding get/set_property.

---
 drivers/gpu/drm/drm_atomic_state_helper.c |  9 ++++++
 drivers/gpu/drm/drm_atomic_uapi.c         | 13 ++++++++
 drivers/gpu/drm/drm_crtc.c                | 54 +++++++++++++++++++++++++++++++
 include/drm/drm_crtc.h                    | 36 +++++++++++++++++++++
 4 files changed, 112 insertions(+)
drivers/gpu/drm/drm_atomic_state_helper.c +9 -0
index dfe6293f7a42d034da3de593094019ca15014a02..ceab90cec57cc580afcf334e275982827e9b0e0d 100644
@@ -149,6 +149,10 @@ void __drm_atomic_helper_crtc_duplicate_state(struct drm_crtc *crtc,
 		drm_property_blob_get(state->histogram_enable);
 	if (state->histogram_data)
 		drm_property_blob_get(state->histogram_data);
+	if (state->iet_lut_caps)
+		drm_property_blob_get(state->iet_lut_caps);
+	if (state->iet_lut)
+		drm_property_blob_get(state->iet_lut);
 	state->mode_changed = false;
 	state->active_changed = false;
 	state->planes_changed = false;
@@ -164,6 +168,7 @@ void __drm_atomic_helper_crtc_duplicate_state(struct drm_crtc *crtc,
 	state->self_refresh_active = false;
 
 	state->histogram_updated = false;
+	state->iet_lut_updated = false;
 }
 EXPORT_SYMBOL(__drm_atomic_helper_crtc_duplicate_state);
 
@@ -229,6 +234,10 @@ void __drm_atomic_helper_crtc_destroy_state(struct drm_crtc_state *state)
 		drm_property_blob_put(state->histogram_enable);
 	if (state->histogram_data)
 		drm_property_blob_put(state->histogram_data);
+	if (state->iet_lut_caps)
+		drm_property_blob_put(state->iet_lut_caps);
+	if (state->iet_lut)
+		drm_property_blob_put(state->iet_lut);
 }
 EXPORT_SYMBOL(__drm_atomic_helper_crtc_destroy_state);
 
drivers/gpu/drm/drm_atomic_uapi.c +13 -0
index 459d30898196c94392a7f916b1fa9ca3a334eea8..f31d24d80cc082b38c611b12f36f281fa7404869 100644
@@ -424,6 +424,15 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc,
 							&replaced);
 		state->histogram_updated |= replaced;
 		return ret;
+	} else if (property == crtc->iet_lut_property) {
+		ret = drm_property_replace_blob_from_id(dev,
+							&state->iet_lut,
+							val,
+							-1,
+							sizeof(struct drm_iet_1dlut_sample),
+							&replaced);
+		state->iet_lut_updated |= replaced;
+		return ret;
 	} else if (property == crtc->scaling_filter_property) {
 		state->scaling_filter = val;
 	} else if (crtc->funcs->atomic_set_property) {
@@ -467,6 +476,10 @@ drm_atomic_crtc_get_property(struct drm_crtc *crtc,
 		*val = (state->histogram_enable) ? state->histogram_enable->base.id : 0;
 	else if (property == crtc->histogram_data_property)
 		*val = (state->histogram_data) ? state->histogram_data->base.id : 0;
+	else if (property == crtc->iet_lut_caps_property)
+		*val = (state->iet_lut_caps) ? state->iet_lut_caps->base.id : 0;
+	else if (property == crtc->iet_lut_property)
+		*val = (state->iet_lut) ? state->iet_lut->base.id : 0;
 	else if (property == crtc->scaling_filter_property)
 		*val = state->scaling_filter;
 	else if (crtc->funcs->atomic_get_property)
drivers/gpu/drm/drm_crtc.c +54 -0

57 lines not on this page · raw message

include/drm/drm_crtc.h +36 -0
index bdca7a84e9a34c405fcda5377b93df1ed575f1dd..b2bb496cd093f828e5199f007ab5afec13f0a4ef 100644
@@ -300,6 +300,29 @@ struct drm_crtc_state {
 	struct drm_property_blob *histogram_data;
 	bool histogram_updated;
 
+	/**
+	 * @iet_lut_caps:
+	 *
+	 * The blob points to the structure drm_iet_lut_caps.
+	 * For more info on the elements of the struct drm_iet_lut_caps
+	 * see include/uapi/drm/drm_mode.h
+	 */
+	struct drm_property_blob *iet_lut_caps;
+	/**
+	 * @iet_lut:
+	 *
+	 * The blob points to the struct drm_lut_sample
+	 * For more information on the elements of struct drm_lut_sample
+	 * see include/uapi/drm/drm_mode.h
+	 */
+	struct drm_property_blob *iet_lut;
+	/**
+	 * @iet_lut_updates:
+	 *
+	 * Convey that the image enhanced data has been updated by the user
+	 */
+	bool iet_lut_updated;
+
 	/**
 	 * @target_vblank:
 	 *
@@ -1130,6 +1153,17 @@ struct drm_crtc {
 	 */
 	struct drm_property *histogram_data_property;
 
+	/**
+	 * @iet_lut_caps_property: Optional CRTC property for getting the
+	 * iet LUT hardware capability.
+	 */
+	struct drm_property *iet_lut_caps_property;
+	/**
+	 * @iet_lut_proeprty: Optional CRTC property for writing the
+	 * image enhanced LUT
+	 */
+	struct drm_property *iet_lut_property;
+
 	/**
 	 * @state:
 	 *
@@ -1367,5 +1401,7 @@ int drm_crtc_create_scaling_filter_property(struct drm_crtc *crtc,
 					    unsigned int supported_filters);
 int drm_crtc_create_histogram_property(struct drm_crtc *crtc,
 				       struct drm_histogram_caps *caps);
+int drm_crtc_create_iet_lut_property(struct drm_crtc *crtc,
+				     struct drm_iet_caps *caps);
 
 #endif /* __DRM_CRTC_H__ */
Signature · 1 line
2.25.1
#20 lore raw

PATCH v7 5/14 drm/i915/histogram: Define registers for histogram

In reply to Arun R Murthy

Add the register/bit definitions for global histogram.

v2: Intended the register contents, removed unused regs (Jani)

Bspec: 4270

---
 .../gpu/drm/i915/display/intel_histogram_regs.h    | 48 ++++++++++++++++++++++
 1 file changed, 48 insertions(+)
drivers/gpu/drm/i915/display/intel_histogram_regs.h new file +48 -0
new file mode 100644
index 0000000000000000000000000000000000000000..1252b4f339a63f70f44e249bdeae87805bee20fc
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2024 Intel Corporation
+ */
+
+#ifndef __INTEL_HISTOGRAM_REGS_H__
+#define __INTEL_HISTOGRAM_REGS_H__
+
+#include "intel_display_reg_defs.h"
+
+/* GLOBAL_HIST related registers */
+#define _DPST_CTL_A					0x490C0
+#define _DPST_CTL_B					0x491C0
+#define DPST_CTL(pipe)					_MMIO_PIPE(pipe, _DPST_CTL_A, _DPST_CTL_B)
+#define  DPST_CTL_IE_HIST_EN				REG_BIT(31)
+#define  DPST_CTL_RESTORE				REG_BIT(28)
+#define  DPST_CTL_IE_MODI_TABLE_EN			REG_BIT(27)
+#define  DPST_CTL_HIST_MODE				REG_BIT(24)
+#define  DPST_CTL_ENHANCEMENT_MODE_MASK			REG_GENMASK(14, 13)
+#define  DPST_CTL_EN_MULTIPLICATIVE			REG_FIELD_PREP(DPST_CTL_ENHANCEMENT_MODE_MASK, 2)
+#define  DPST_CTL_IE_TABLE_VALUE_FORMAT			REG_BIT(15)
+#define  DPST_CTL_BIN_REG_FUNC_SEL			REG_BIT(11)
+#define  DPST_CTL_BIN_REG_FUNC_TC			REG_FIELD_PREP(DPST_CTL_BIN_REG_FUNC_SEL, 0)
+#define  DPST_CTL_BIN_REG_FUNC_IE			REG_FIELD_PREP(DPST_CTL_BIN_REG_FUNC_SEL, 1)
+#define  DPST_CTL_BIN_REG_MASK				REG_GENMASK(6, 0)
+#define  DPST_CTL_BIN_REG_CLEAR				REG_FIELD_PREP(DPST_CTL_BIN_REG_MASK, 0)
+#define  DPST_CTL_IE_TABLE_VALUE_FORMAT_2INT_8FRAC	REG_FIELD_PREP(DPST_CTL_IE_TABLE_VALUE_FORMAT, 1)
+#define  DPST_CTL_IE_TABLE_VALUE_FORMAT_1INT_9FRAC	REG_FIELD_PREP(DPST_CTL_IE_TABLE_VALUE_FORMAT, 0)
+#define  DPST_CTL_HIST_MODE_YUV				REG_FIELD_PREP(DPST_CTL_HIST_MODE, 0)
+#define  DPST_CTL_HIST_MODE_HSV				REG_FIELD_PREP(DPST_CTL_HIST_MODE, 1)
+
+#define _DPST_GUARD_A					0x490C8
+#define _DPST_GUARD_B					0x491C8
+#define DPST_GUARD(pipe)				_MMIO_PIPE(pipe, _DPST_GUARD_A, _DPST_GUARD_B)
+#define  DPST_GUARD_HIST_INT_EN				REG_BIT(31)
+#define  DPST_GUARD_HIST_EVENT_STATUS			REG_BIT(30)
+#define  DPST_GUARD_INTERRUPT_DELAY_MASK			REG_GENMASK(29, 22)
+#define  DPST_GUARD_INTERRUPT_DELAY(val)			REG_FIELD_PREP(DPST_GUARD_INTERRUPT_DELAY_MASK, val)
+#define  DPST_GUARD_THRESHOLD_GB_MASK			REG_GENMASK(21, 0)
+#define  DPST_GUARD_THRESHOLD_GB(val)			REG_FIELD_PREP(DPST_GUARD_THRESHOLD_GB_MASK, val)
+
+#define _DPST_BIN_A					0x490C4
+#define _DPST_BIN_B					0x491C4
+#define DPST_BIN(pipe)					_MMIO_PIPE(pipe, _DPST_BIN_A, _DPST_BIN_B)
+#define  DPST_BIN_DATA_MASK				REG_GENMASK(23, 0)
+#define  DPST_BIN_BUSY					REG_BIT(31)
+
+#endif /* __INTEL_HISTOGRAM_REGS_H__ */
Signature · 1 line
2.25.1
#21 lore raw

PATCH v7 6/14 drm/i915/histogram: Add support for histogram

In reply to Arun R Murthy

Statistics is generated from the image frame that is coming to display and an event is sent to user after reading this histogram data.

v2: forward declaration in header file along with error handling (Jani)
v3: Replaced i915 with intel_display (Suraj)
v4: Removed dithering enable/disable (Vandita)
New patch for histogram register definitions (Suraj)
v5: IET LUT pgm follow the seq in spec and removed change to TC at end

    (Suraj)

v8: Retained only the Histogram part and move IET LUT to a different
patch.

---
 drivers/gpu/drm/i915/Makefile                      |   1 +
 drivers/gpu/drm/i915/display/intel_display_types.h |   2 +
 drivers/gpu/drm/i915/display/intel_histogram.c     | 149 +++++++++++++++++++++
 drivers/gpu/drm/i915/display/intel_histogram.h     |  48 +++++++
 4 files changed, 200 insertions(+)
drivers/gpu/drm/i915/Makefile +1 -0
index 3dda9f0eda82b7501193eb3fb9c0e5dd8efa71e4..8c4277cc63eb6c1e12a1714df2abf07c290aacc6 100644
@@ -270,6 +270,7 @@ i915-y += \
 	display/intel_hdcp.o \
 	display/intel_hdcp_gsc.o \
 	display/intel_hdcp_gsc_message.o \
+	display/intel_histogram.o \
 	display/intel_hotplug.o \
 	display/intel_hotplug_irq.o \
 	display/intel_hti.o \
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index eb9dd1125a4a09511936b81219e7f38fae106dfd..e17729f0f5f5e9e79a08c9b72ab3772a78e1fbc7 100644
@@ -1434,6 +1434,8 @@ struct intel_crtc {
 	/* for loading single buffered registers during vblank */
 	struct pm_qos_request vblank_pm_qos;
 
+	struct intel_histogram *histogram;
+
 #ifdef CONFIG_DEBUG_FS
 	struct intel_pipe_crc pipe_crc;
 #endif
drivers/gpu/drm/i915/display/intel_histogram.c new file +149 -0

149 lines not on this page · raw message

drivers/gpu/drm/i915/display/intel_histogram.h new file +48 -0

48 lines not on this page · raw message

Signature · 1 line
2.25.1
#22 lore raw

PATCH v7 7/14 drm/xe: Add histogram support to Xe builds

In reply to Arun R Murthy

Histogram added as part of i915/display driver. Adding the same for xe as well.

---
 drivers/gpu/drm/xe/Makefile | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/xe/Makefile +1 -0
index 5c97ad6ed7385616ecce3340ec74580f53a213e3..984def6077efb9b3fcedb2065414173691427e4a 100644
@@ -247,6 +247,7 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
 	i915-display/intel_hdcp.o \
 	i915-display/intel_hdcp_gsc_message.o \
 	i915-display/intel_hdmi.o \
+	i915-display/intel_histogram.o \
 	i915-display/intel_hotplug.o \
 	i915-display/intel_hotplug_irq.o \
 	i915-display/intel_hti.o \
Signature · 1 line
2.25.1
#23 lore raw

Re: [PATCH v7 07/14] drm/xe: Add histogram support to Xe builds

Quoted 5 lines · Arun R Murthy

Histogram added as part of i915/display driver. Adding the same for xe as well.

Is building of the Xe driver broken between the previous commit and this one? In such a case, it needs to be squashed into the previous commit.

Quoted 20 lines
---
 drivers/gpu/drm/xe/Makefile | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/xe/Makefile +1 -0
index 5c97ad6ed7385616ecce3340ec74580f53a213e3..984def6077efb9b3fcedb2065414173691427e4a 100644
@@ -247,6 +247,7 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
 	i915-display/intel_hdcp.o \
 	i915-display/intel_hdcp_gsc_message.o \
 	i915-display/intel_hdmi.o \
+	i915-display/intel_histogram.o \
 	i915-display/intel_hotplug.o \
 	i915-display/intel_hotplug_irq.o \
 	i915-display/intel_hti.o \
Signature · 2 lines
-- 
2.25.1
Signature · 2 lines
With best wishes
Dmitry
#24 lore raw

RE: [PATCH v7 07/14] drm/xe: Add histogram support to Xe builds

Quoted 10 lines · Arun R Murthy
Quoted 5 lines

Histogram added as part of i915/display driver. Adding the same for xe as well.

Is building of the Xe driver broken between the previous commit and this one? In such a case, it needs to be squashed into the previous commit.

No!
New file for histogram is added and compilation of this file for Xe builds is added in this patch. In the previous patch it was done for i915 builds.

Thanks and Regards,
Arun R Murthy
--------------------

#25 lore raw

Re: [PATCH v7 07/14] drm/xe: Add histogram support to Xe builds

Quoted 8 lines · Arun R
Quoted 8 lines · Arun R Murthy
Quoted 5 lines

Histogram added as part of i915/display driver. Adding the same for xe as well.

Quoted 6 lines
Quoted 3 lines

Is building of the Xe driver broken between the previous commit and this one? In such a case, it needs to be squashed into the previous commit.

No!
New file for histogram is added and compilation of this file for Xe builds is added in this patch. In the previous patch it was done for i915 builds.

Ack

Signature · 2 lines
With best wishes
Dmitry
#26 lore raw

PATCH v7 8/14 drm/i915/histogram: histogram interrupt handling

In reply to Arun R Murthy

Upon enabling histogram an interrupt is trigerred after the generation of the statistics. This patch registers the histogram interrupt and handles the interrupt.

v2: Added intel_crtc backpointer to intel_histogram struct (Jani)

    Removed histogram_wq and instead use dev_priv->unodered_eq (Jani)

v3: Replaced drm_i915_private with intel_display (Suraj)
Refactored the histogram read code (Jani)
v4: Rebased after addressing comments on patch 1
v5: removed the retry logic and moved to patch7 (Jani)

---
 drivers/gpu/drm/i915/display/intel_display_irq.c |   6 +-
 drivers/gpu/drm/i915/display/intel_histogram.c   | 106 ++++++++++++++++++++++-
 drivers/gpu/drm/i915/display/intel_histogram.h   |   3 +
 drivers/gpu/drm/i915/i915_reg.h                  |   5 +-
 4 files changed, 115 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/display/intel_display_irq.c +5 -1
index 069043f9d89450750f183d118e7a546fa5334e8e..ee3166d4c6561f1a9396979fbc5c95d53158ce8a 100644
@@ -20,6 +20,7 @@
 #include "intel_fdi_regs.h"
 #include "intel_fifo_underrun.h"
 #include "intel_gmbus.h"
+#include "intel_histogram.h"
 #include "intel_hotplug_irq.h"
 #include "intel_pipe_crc_regs.h"
 #include "intel_pmdemand.h"
@@ -1180,6 +1181,9 @@ void gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 		if (iir & GEN8_PIPE_FIFO_UNDERRUN)
 			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
 
+		if (iir & GEN9_PIPE_HISTOGRAM_EVENT)
+			intel_histogram_irq_handler(display, pipe);
+
 		fault_errors = iir & gen8_de_pipe_fault_mask(dev_priv);
 		if (fault_errors)
 			drm_err_ratelimited(&dev_priv->drm,
@@ -1773,7 +1777,7 @@ void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 	struct intel_uncore *uncore = &dev_priv->uncore;
 
 	u32 de_pipe_masked = gen8_de_pipe_fault_mask(dev_priv) |
-		GEN8_PIPE_CDCLK_CRC_DONE;
+		GEN8_PIPE_CDCLK_CRC_DONE | GEN9_PIPE_HISTOGRAM_EVENT;
 	u32 de_pipe_enables;
 	u32 de_port_masked = gen8_de_port_aux_mask(dev_priv);
 	u32 de_port_enables;
drivers/gpu/drm/i915/display/intel_histogram.c +104 -2

135 lines not on this page · raw message

drivers/gpu/drm/i915/display/intel_histogram.h +3 -0
index 5ea19ef2d3ecadf1ac159a784f51278fdde593de..b44ba3afc94f79f291f4e5ebdd04dcf9434b48a4 100644
@@ -14,6 +14,8 @@ struct drm_property_blob;
 struct drm_histogram_config;
 struct drm_histogram_caps;
 struct intel_crtc;
+struct intel_display;
+enum pipe;
 
 #define HISTOGRAM_BIN_COUNT                    32
 
@@ -39,6 +41,7 @@ enum intel_global_hist_lut {
 	INTEL_HISTOGRAM_PIXEL_FACTOR,
 };
 
+void intel_histogram_irq_handler(struct intel_display *display, enum pipe pipe);
 int intel_histogram_atomic_check(struct intel_crtc *intel_crtc);
 int intel_histogram_update(struct intel_crtc *intel_crtc,
 			   struct drm_histogram_config *config);
drivers/gpu/drm/i915/i915_reg.h +3 -2
index 765e6c0528fb0b5a894395b77a5edbf0b0c80009..297537dcbd704448d7b242070cac67ad9fd48a5f 100644
@@ -1598,7 +1598,7 @@
 #define   PIPE_HOTPLUG_INTERRUPT_ENABLE		(1UL << 26)
 #define   PIPE_VSYNC_INTERRUPT_ENABLE		(1UL << 25)
 #define   PIPE_DISPLAY_LINE_COMPARE_ENABLE	(1UL << 24)
-#define   PIPE_DPST_EVENT_ENABLE		(1UL << 23)
+#define   PIPE_HISTOGRAM_EVENT_ENABLE		(1UL << 23)
 #define   SPRITE0_FLIP_DONE_INT_EN_VLV		(1UL << 22)
 #define   PIPE_LEGACY_BLC_EVENT_ENABLE		(1UL << 22)
 #define   PIPE_ODD_FIELD_INTERRUPT_ENABLE	(1UL << 21)
@@ -1621,7 +1621,7 @@
 #define   PIPE_HOTPLUG_INTERRUPT_STATUS		(1UL << 10)
 #define   PIPE_VSYNC_INTERRUPT_STATUS		(1UL << 9)
 #define   PIPE_DISPLAY_LINE_COMPARE_STATUS	(1UL << 8)
-#define   PIPE_DPST_EVENT_STATUS		(1UL << 7)
+#define   PIPE_HISTOGRAM_EVENT_STATUS		(1UL << 7)
 #define   PIPE_A_PSR_STATUS_VLV			(1UL << 6)
 #define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL << 6)
 #define   PIPE_ODD_FIELD_INTERRUPT_STATUS	(1UL << 5)
@@ -2223,6 +2223,7 @@
 #define  GEN12_DSB_1_INT		REG_BIT(14) /* tgl+ */
 #define  GEN12_DSB_0_INT		REG_BIT(13) /* tgl+ */
 #define  GEN12_DSB_INT(dsb_id)		REG_BIT(13 + (dsb_id))
+#define  GEN9_PIPE_HISTOGRAM_EVENT	REG_BIT(12) /* skl+ */
 #define  GEN9_PIPE_CURSOR_FAULT		REG_BIT(11) /* skl+ */
 #define  GEN9_PIPE_PLANE4_FAULT		REG_BIT(10) /* skl+ */
 #define  GEN8_PIPE_CURSOR_FAULT		REG_BIT(10) /* bdw */
Signature · 1 line
2.25.1
#27 lore raw

PATCH v7 9/14 drm/i915/histogram: Hook i915 histogram with drm histogram

In reply to Arun R Murthy

Handle histogram caps and histogram config property in i915 driver. Fill the histogram hardware capability and act upon the histogram config property to enable/disable histogram in i915.

---
 drivers/gpu/drm/i915/display/intel_crtc.c    |  7 +++++++
 drivers/gpu/drm/i915/display/intel_display.c | 12 ++++++++++++
 2 files changed, 19 insertions(+)
drivers/gpu/drm/i915/display/intel_crtc.c +7 -0
index c910168602d28cd51a7fb376e8935dfe3922845b..619aa363724602d4084184bfdf5766b71aed1b9f 100644
@@ -28,6 +28,7 @@
 #include "intel_drrs.h"
 #include "intel_dsi.h"
 #include "intel_fifo_underrun.h"
+#include "intel_histogram.h"
 #include "intel_pipe_crc.h"
 #include "intel_psr.h"
 #include "intel_sprite.h"
@@ -211,6 +212,7 @@ static struct intel_crtc *intel_crtc_alloc(void)
 static void intel_crtc_free(struct intel_crtc *crtc)
 {
 	intel_crtc_destroy_state(&crtc->base, crtc->base.state);
+	intel_histogram_finish(crtc);
 	kfree(crtc);
 }
 
@@ -381,6 +383,11 @@ int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
 						BIT(DRM_SCALING_FILTER_DEFAULT) |
 						BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR));
 
+	intel_histogram_init(crtc);
+	if (drm_crtc_create_histogram_property(&crtc->base,
+					       crtc->histogram->caps))
+		drm_err(&dev_priv->drm, "Failed to initialize histogram properties\n");
+
 	intel_color_crtc_init(crtc);
 	intel_drrs_crtc_init(crtc);
 	intel_crtc_crc_init(crtc);
drivers/gpu/drm/i915/display/intel_display.c +12 -0
index 4271da219b4105630106a2bc7e1fa42015ede1e1..486992a2caadebf2d3deb200b01d2d0d26b26cb0 100644
@@ -93,6 +93,7 @@
 #include "intel_fifo_underrun.h"
 #include "intel_frontbuffer.h"
 #include "intel_hdmi.h"
+#include "intel_histogram.h"
 #include "intel_hotplug.h"
 #include "intel_link_bw.h"
 #include "intel_lvds.h"
@@ -4612,6 +4613,12 @@ static int intel_crtc_atomic_check(struct intel_atomic_state *state,
 	if (ret)
 		return ret;
 
+	if (crtc_state->uapi.histogram_updated) {
+		ret = intel_histogram_atomic_check(crtc);
+		if (ret)
+			return ret;
+	}
+
 	return 0;
 }
 
@@ -7878,6 +7885,11 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 		 */
 		old_crtc_state->dsb_color_vblank = fetch_and_zero(&new_crtc_state->dsb_color_vblank);
 		old_crtc_state->dsb_commit = fetch_and_zero(&new_crtc_state->dsb_commit);
+
+		if (new_crtc_state->uapi.histogram_updated)
+			intel_histogram_update(crtc,
+					       (struct drm_histogram_config *)
+					       new_crtc_state->uapi.histogram_enable->data);
 	}
 
 	/* Underruns don't always raise interrupts, so check manually */
Signature · 1 line
2.25.1
#28 lore raw

PATCH v7 10/14 drm/i915/iet: Add support to writing the IET LUT data

In reply to Arun R Murthy

User created LUT can be fed back to the hardware so that the hardware can apply this LUT data to see the enhancement in the image.

---
 drivers/gpu/drm/i915/display/intel_histogram.c | 70 ++++++++++++++++++++++++++
 drivers/gpu/drm/i915/display/intel_histogram.h |  4 ++
 2 files changed, 74 insertions(+)
drivers/gpu/drm/i915/display/intel_histogram.c +70 -0
index ea61a98efb18fcccce88a8a3b82fd373c47920df..499ea9157a338f5081c74dfc182371b2075634ea 100644
@@ -20,6 +20,7 @@
 /* Precision factor for threshold guardband */
 #define HISTOGRAM_GUARDBAND_PRECISION_FACTOR 10000
 #define HISTOGRAM_BIN_READ_RETRY_COUNT 5
+#define IET_SAMPLE_FORMAT_1_INT_9_FRACT 0x1000009
 
 static bool intel_histogram_get_data(struct intel_crtc *intel_crtc)
 {
@@ -215,6 +216,60 @@ int intel_histogram_update(struct intel_crtc *intel_crtc,
 	return 0;
 }
 
+int intel_histogram_set_iet_lut(struct intel_crtc *intel_crtc,
+				struct drm_property_blob *blob)
+{
+	struct intel_histogram *histogram = intel_crtc->histogram;
+	struct intel_display *display = to_intel_display(intel_crtc);
+	int pipe = intel_crtc->pipe;
+	int i = 0;
+	struct drm_iet_1dlut_sample *iet;
+	u32 *data;
+	int ret;
+
+	if (!histogram)
+		return -EINVAL;
+
+	if (!histogram->enable) {
+		drm_err(display->drm, "histogram not enabled");
+		return -EINVAL;
+	}
+
+	if (!data) {
+	drm_err(display->drm, "enhancement LUT data is NULL");
+		return -EINVAL;
+	}
+
+	/* Set DPST_CTL Bin Reg function select to IE & wait for a vblabk */
+	intel_de_rmw(display, DPST_CTL(pipe),
+		     DPST_CTL_BIN_REG_FUNC_SEL, DPST_CTL_BIN_REG_FUNC_IE);
+
+	drm_crtc_wait_one_vblank(&intel_crtc->base);
+
+	 /* Set DPST_CTL Bin Register Index to 0 */
+	intel_de_rmw(display, DPST_CTL(pipe),
+		     DPST_CTL_BIN_REG_MASK, DPST_CTL_BIN_REG_CLEAR);
+
+	iet = (struct drm_iet_1dlut_sample *)blob->data;
+	data = kzalloc(sizeof(data) * iet->nr_elements, GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+	ret = copy_from_user(data, (uint32_t __user *)(unsigned long)iet->data_ptr,
+			     sizeof(uint32_t) * iet->nr_elements);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < HISTOGRAM_IET_LENGTH; i++) {
+		intel_de_rmw(display, DPST_BIN(pipe),
+			     DPST_BIN_DATA_MASK, data[i]);
+		drm_dbg_atomic(display->drm, "iet_lut[%d]=%x\n", i, data[i]);
+	}
+	kfree(data);
+	drm_property_blob_put(intel_crtc->base.state->iet_lut);
+
+	return 0;
+}
+
 void intel_histogram_finish(struct intel_crtc *intel_crtc)
 {
 	struct intel_histogram *histogram = intel_crtc->histogram;
@@ -227,6 +282,8 @@ int intel_histogram_init(struct intel_crtc *crtc)
 {
 	struct intel_histogram *histogram;
 	struct drm_histogram_caps *histogram_caps;
+	struct drm_iet_caps *iet_caps;
+	u32 *iet_format;
 
 	/* Allocate histogram internal struct */
 	histogram = kzalloc(sizeof(*histogram), GFP_KERNEL);
@@ -239,10 +296,23 @@ int intel_histogram_init(struct intel_crtc *crtc)
 	histogram_caps->histogram_mode = DRM_MODE_HISTOGRAM_HSV_MAX_RGB;
 	histogram_caps->bins_count = HISTOGRAM_BIN_COUNT;
 
+	iet_caps = kzalloc(sizeof(*iet_caps), GFP_KERNEL);
+	if (!iet_caps)
+		return -ENOMEM;
+
+	iet_caps->iet_mode = DRM_MODE_IET_MULTIPLICATIVE;
+	iet_caps->nr_iet_sample_formats = 1;
+	iet_caps->nr_iet_lut_entries = HISTOGRAM_IET_LENGTH;
+	iet_format = kzalloc(sizeof(u32)*iet_caps->nr_iet_sample_formats,
+			     GFP_KERNEL);
+	*iet_format = IET_SAMPLE_FORMAT_1_INT_9_FRACT;
+	iet_caps->iet_sample_format = *iet_format;
+
 	crtc->histogram = histogram;
 	histogram->crtc = crtc;
 	histogram->can_enable = false;
 	histogram->caps = histogram_caps;
+	histogram->iet_caps = iet_caps;
 
 	INIT_DEFERRABLE_WORK(&histogram->work,
 			     intel_histogram_handle_int_work);
drivers/gpu/drm/i915/display/intel_histogram.h +4 -0
index b44ba3afc94f79f291f4e5ebdd04dcf9434b48a4..0999d1720c7abee8907c77896e4b1e6ff756160f 100644
@@ -18,9 +18,11 @@ struct intel_display;
 enum pipe;
 
 #define HISTOGRAM_BIN_COUNT                    32
+#define HISTOGRAM_IET_LENGTH                   33
 
 struct intel_histogram {
 	struct drm_histogram_caps *caps;
+	struct drm_iet_caps *iet_caps;
 	struct intel_crtc *crtc;
 	struct delayed_work work;
 	bool enable;
@@ -45,6 +47,8 @@ void intel_histogram_irq_handler(struct intel_display *display, enum pipe pipe);
 int intel_histogram_atomic_check(struct intel_crtc *intel_crtc);
 int intel_histogram_update(struct intel_crtc *intel_crtc,
 			   struct drm_histogram_config *config);
+int intel_histogram_set_iet_lut(struct intel_crtc *intel_crtc,
+				struct drm_property_blob *blob);
 int intel_histogram_init(struct intel_crtc *intel_crtc);
 void intel_histogram_finish(struct intel_crtc *intel_crtc);
 
Signature · 1 line
2.25.1
#29 lore raw

PATCH v7 11/14 drm/i915/crtc: Hook i915 IET LUT with the drm IET properties

In reply to Arun R Murthy

Upon drm getting the IET LUT value from the user through the IET_LUT property, i915 driver will write the LUT table to the hardware registers.

---
 drivers/gpu/drm/i915/display/intel_crtc.c    | 3 +++
 drivers/gpu/drm/i915/display/intel_display.c | 2 ++
 2 files changed, 5 insertions(+)
drivers/gpu/drm/i915/display/intel_crtc.c +3 -0
index 619aa363724602d4084184bfdf5766b71aed1b9f..8c1527ae348083c9d8af7cbbbe188ed18afb0a43 100644
@@ -387,6 +387,9 @@ int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
 	if (drm_crtc_create_histogram_property(&crtc->base,
 					       crtc->histogram->caps))
 		drm_err(&dev_priv->drm, "Failed to initialize histogram properties\n");
+	if (drm_crtc_create_iet_lut_property(&crtc->base,
+					     crtc->histogram->iet_caps))
+		drm_err(&dev_priv->drm, "Failed to initialize histogram properties\n");
 
 	intel_color_crtc_init(crtc);
 	intel_drrs_crtc_init(crtc);
drivers/gpu/drm/i915/display/intel_display.c +2 -0
index 486992a2caadebf2d3deb200b01d2d0d26b26cb0..dfa0b214a54b5336d45e965282b7d8017f68bafa 100644
@@ -7890,6 +7890,8 @@ static void intel_atomic_commit_tail(struct intel_atomic_state *state)
 			intel_histogram_update(crtc,
 					       (struct drm_histogram_config *)
 					       new_crtc_state->uapi.histogram_enable->data);
+		if (new_crtc_state->uapi.iet_lut_updated)
+			intel_histogram_set_iet_lut(crtc, new_crtc_state->uapi.iet_lut);
 	}
 
 	/* Underruns don't always raise interrupts, so check manually */
Signature · 1 line
2.25.1
#30 lore raw

PATCH v7 12/14 drm/i915/histogram: histogram delay counter doesnt reset

In reply to Arun R Murthy

The delay counter for histogram does not reset and as a result the histogram bin never gets updated. Workaround would be to use save and restore histogram register.

v2: Follow the seq in interrupt handler
Restore DPST bit 0
read/write dpst ctl rg
Restore DPST bit 1 and Guardband Delay Interrupt counter = 0 (Suraj)
v3: updated wa version for display 13 and 14

Wa: 14014889975

---
 drivers/gpu/drm/i915/display/intel_histogram.c      | 14 ++++++++++++++
 drivers/gpu/drm/i915/display/intel_histogram_regs.h |  2 ++
 2 files changed, 16 insertions(+)
drivers/gpu/drm/i915/display/intel_histogram.c +14 -0
index 499ea9157a338f5081c74dfc182371b2075634ea..039ca16023b1d56c0f1f91d3a1d8ed440e4ea675 100644
@@ -52,6 +52,11 @@ static void intel_histogram_handle_int_work(struct work_struct *work)
 	snprintf(pipe_id, sizeof(pipe_id),
 		 "PIPE=%u", intel_crtc->base.base.id);
 
+	/* Wa: 14014889975 */
+	if (IS_DISPLAY_VER(display, 13, 14))
+		intel_de_rmw(display, DPST_CTL(intel_crtc->pipe),
+			     DPST_CTL_RESTORE, 0);
+
 	/*
 	 * TODO: PSR to be exited while reading the Histogram data
 	 * Set DPST_CTL Bin Reg function select to TC
@@ -93,6 +98,15 @@ static void intel_histogram_handle_int_work(struct work_struct *work)
 		return;
 	}
 
+	/* Wa: 14014889975 */
+	if (IS_DISPLAY_VER(display, 13, 14))
+		/* Write the value read from DPST_CTL to DPST_CTL.Interrupt Delay Counter(bit 23:16) */
+		intel_de_rmw(display, DPST_CTL(intel_crtc->pipe),
+			     DPST_CTL_GUARDBAND_INTERRUPT_DELAY_CNT |
+			     DPST_CTL_RESTORE,
+			     DPST_CTL_GUARDBAND_INTERRUPT_DELAY(0x0) |
+			     DPST_CTL_RESTORE);
+
 	/* Enable histogram interrupt */
 	intel_de_rmw(display, DPST_GUARD(intel_crtc->pipe), DPST_GUARD_HIST_INT_EN,
 		     DPST_GUARD_HIST_INT_EN);
drivers/gpu/drm/i915/display/intel_histogram_regs.h +2 -0
index 1252b4f339a63f70f44e249bdeae87805bee20fc..213c9f483567cb19a47b44953749f6baf0afe9e7 100644
@@ -16,6 +16,8 @@
 #define  DPST_CTL_RESTORE				REG_BIT(28)
 #define  DPST_CTL_IE_MODI_TABLE_EN			REG_BIT(27)
 #define  DPST_CTL_HIST_MODE				REG_BIT(24)
+#define  DPST_CTL_GUARDBAND_INTERRUPT_DELAY_CNT		REG_GENMASK(23, 16)
+#define  DPST_CTL_GUARDBAND_INTERRUPT_DELAY(val)	REG_FIELD_PREP(DPST_CTL_GUARDBAND_INTERRUPT_DELAY_CNT, val)
 #define  DPST_CTL_ENHANCEMENT_MODE_MASK			REG_GENMASK(14, 13)
 #define  DPST_CTL_EN_MULTIPLICATIVE			REG_FIELD_PREP(DPST_CTL_ENHANCEMENT_MODE_MASK, 2)
 #define  DPST_CTL_IE_TABLE_VALUE_FORMAT			REG_BIT(15)
Signature · 1 line
2.25.1
#31 lore raw

PATCH v7 13/14 drm/i915/histogram: Histogram changes for Display 20+

In reply to Arun R Murthy

In Display 20+, new registers are added for setting index, reading histogram and writing the IET.

v2: Removed duplicate code (Jani)
v3: Moved histogram core changes to earlier patches (Jani/Suraj)
v4: Rebased after addressing comments on patch 1
v5: Added the retry logic from patch3 and rebased the patch series
v6: optimize wite_iet() (Suraj)

Bspec: 68895

---
 drivers/gpu/drm/i915/display/intel_histogram.c     | 108 +++++++++++++++------
 .../gpu/drm/i915/display/intel_histogram_regs.h    |  25 +++++
 2 files changed, 104 insertions(+), 29 deletions(-)
drivers/gpu/drm/i915/display/intel_histogram.c +79 -29
index 039ca16023b1d56c0f1f91d3a1d8ed440e4ea675..d015350b57ed5c8e9aaab71311159bf51e15e9c7 100644
@@ -22,6 +22,37 @@
 #define HISTOGRAM_BIN_READ_RETRY_COUNT 5
 #define IET_SAMPLE_FORMAT_1_INT_9_FRACT 0x1000009
 
+static void set_bin_index_0(struct intel_display *display, enum pipe pipe)
+{
+	if (DISPLAY_VER(display) >= 20)
+		intel_de_rmw(display, DPST_IE_INDEX(pipe),
+			     DPST_IE_BIN_INDEX_MASK, DPST_IE_BIN_INDEX(0));
+	else
+		intel_de_rmw(display, DPST_CTL(pipe),
+			     DPST_CTL_BIN_REG_MASK,
+			     DPST_CTL_BIN_REG_CLEAR);
+}
+
+static void write_iet(struct intel_display *display, enum pipe pipe,
+			      u32 *data)
+{
+	int i;
+
+	for (i = 0; i < HISTOGRAM_IET_LENGTH; i++) {
+		if (DISPLAY_VER(display) >= 20)
+			intel_de_rmw(display, DPST_IE_BIN(pipe),
+				     DPST_IE_BIN_DATA_MASK,
+				     DPST_IE_BIN_DATA(data[i]));
+		else
+			intel_de_rmw(display, DPST_BIN(pipe),
+				     DPST_BIN_DATA_MASK,
+				     DPST_BIN_DATA(data[i]));
+
+		drm_dbg_atomic(display->drm, "iet_lut[%d]=%x\n",
+			       i, data[i]);
+	}
+}
+
 static bool intel_histogram_get_data(struct intel_crtc *intel_crtc)
 {
 	struct intel_display *display = to_intel_display(intel_crtc);
@@ -29,12 +60,27 @@ static bool intel_histogram_get_data(struct intel_crtc *intel_crtc)
 	int index;
 	u32 dpstbin;
 
+	if (DISPLAY_VER(display) >= 20)
+		intel_de_rmw(display, DPST_HIST_INDEX(intel_crtc->pipe),
+			     DPST_HIST_BIN_INDEX_MASK,
+			     DPST_HIST_BIN_INDEX(0));
+	else
+		intel_de_rmw(display, DPST_CTL(intel_crtc->pipe),
+			     DPST_CTL_BIN_REG_FUNC_SEL | DPST_CTL_BIN_REG_MASK, 0);
+
 	for (index = 0; index < ARRAY_SIZE(histogram->bin_data); index++) {
-		dpstbin = intel_de_read(display, DPST_BIN(intel_crtc->pipe));
+		dpstbin = intel_de_read(display, (DISPLAY_VER(display) >= 20 ?
+					DPST_HIST_BIN(intel_crtc->pipe) :
+					DPST_BIN(intel_crtc->pipe)));
 		if (!(dpstbin & DPST_BIN_BUSY)) {
-			histogram->bin_data[index] = dpstbin & DPST_BIN_DATA_MASK;
-		} else
+			histogram->bin_data[index] = dpstbin & (DISPLAY_VER(display) >= 20 ?
+								DPST_HIST_BIN_DATA_MASK :
+								DPST_BIN_DATA_MASK);
+		} else {
+			drm_err(display->drm, "Histogram bin busy, retyring\n");
+			fsleep(2);
 			return false;
+		}
 	}
 	return true;
 }
@@ -62,8 +108,6 @@ static void intel_histogram_handle_int_work(struct work_struct *work)
 	 * Set DPST_CTL Bin Reg function select to TC
 	 * Set DPST_CTL Bin Register Index to 0
 	 */
-	intel_de_rmw(display, DPST_CTL(intel_crtc->pipe),
-		     DPST_CTL_BIN_REG_FUNC_SEL | DPST_CTL_BIN_REG_MASK, 0);
 	for (retry = 0; retry < HISTOGRAM_BIN_READ_RETRY_COUNT; retry++) {
 		if (intel_histogram_get_data(intel_crtc)) {
 			u32 *data;
@@ -156,17 +200,27 @@ static int intel_histogram_enable(struct intel_crtc *intel_crtc, u8 mode)
 
 	if (histogram->enable)
 		return 0;
-
-	 /* enable histogram, clear DPST_CTL bin reg func select to TC */
-	intel_de_rmw(display, DPST_CTL(pipe),
-		     DPST_CTL_BIN_REG_FUNC_SEL | DPST_CTL_IE_HIST_EN |
-		     DPST_CTL_HIST_MODE | DPST_CTL_IE_TABLE_VALUE_FORMAT |
-		     DPST_CTL_ENHANCEMENT_MODE_MASK | DPST_CTL_IE_MODI_TABLE_EN,
-		     ((mode == DRM_MODE_HISTOGRAM_HSV_MAX_RGB) ?
-		      DPST_CTL_BIN_REG_FUNC_TC : 0) | DPST_CTL_IE_HIST_EN |
-		     DPST_CTL_HIST_MODE_HSV |
-		     DPST_CTL_IE_TABLE_VALUE_FORMAT_1INT_9FRAC |
-		     DPST_CTL_EN_MULTIPLICATIVE | DPST_CTL_IE_MODI_TABLE_EN);
+	 /* enable histogram, clear DPST_BIN reg and select TC function */
+	if (DISPLAY_VER(display) >= 20)
+		intel_de_rmw(display, DPST_CTL(pipe),
+			     DPST_CTL_IE_HIST_EN |
+			     DPST_CTL_HIST_MODE,
+			     DPST_CTL_IE_HIST_EN |
+			     DPST_CTL_HIST_MODE_HSV);
+	else
+		 /* enable histogram, clear DPST_CTL bin reg func select to TC */
+		intel_de_rmw(display, DPST_CTL(pipe),
+			     DPST_CTL_BIN_REG_FUNC_SEL | DPST_CTL_IE_HIST_EN |
+			     DPST_CTL_HIST_MODE |
+			     DPST_CTL_IE_TABLE_VALUE_FORMAT |
+			     DPST_CTL_ENHANCEMENT_MODE_MASK |
+			     DPST_CTL_IE_MODI_TABLE_EN,
+			     ((mode == DRM_MODE_HISTOGRAM_HSV_MAX_RGB) ?
+			      DPST_CTL_BIN_REG_FUNC_TC : 0) |
+			     DPST_CTL_IE_HIST_EN |
+			     DPST_CTL_HIST_MODE_HSV |
+			     DPST_CTL_IE_TABLE_VALUE_FORMAT_1INT_9FRAC |
+			     DPST_CTL_EN_MULTIPLICATIVE | DPST_CTL_IE_MODI_TABLE_EN);
 
 	/* Re-Visit: check if wait for one vblank is required */
 	drm_crtc_wait_one_vblank(&intel_crtc->base);
@@ -236,7 +290,6 @@ int intel_histogram_set_iet_lut(struct intel_crtc *intel_crtc,
 	struct intel_histogram *histogram = intel_crtc->histogram;
 	struct intel_display *display = to_intel_display(intel_crtc);
 	int pipe = intel_crtc->pipe;
-	int i = 0;
 	struct drm_iet_1dlut_sample *iet;
 	u32 *data;
 	int ret;
@@ -254,15 +307,15 @@ int intel_histogram_set_iet_lut(struct intel_crtc *intel_crtc,
 		return -EINVAL;
 	}
 
-	/* Set DPST_CTL Bin Reg function select to IE & wait for a vblabk */
-	intel_de_rmw(display, DPST_CTL(pipe),
-		     DPST_CTL_BIN_REG_FUNC_SEL, DPST_CTL_BIN_REG_FUNC_IE);
 
-	drm_crtc_wait_one_vblank(&intel_crtc->base);
+	if (DISPLAY_VER(display) < 20) {
+		/* Set DPST_CTL Bin Reg function select to IE & wait for a vblabk */
+		intel_de_rmw(display, DPST_CTL(pipe),
+			     DPST_CTL_BIN_REG_FUNC_SEL,
+			     DPST_CTL_BIN_REG_FUNC_IE);
+	}
 
-	 /* Set DPST_CTL Bin Register Index to 0 */
-	intel_de_rmw(display, DPST_CTL(pipe),
-		     DPST_CTL_BIN_REG_MASK, DPST_CTL_BIN_REG_CLEAR);
+	set_bin_index_0(display, pipe);
 
 	iet = (struct drm_iet_1dlut_sample *)blob->data;
 	data = kzalloc(sizeof(data) * iet->nr_elements, GFP_KERNEL);
@@ -273,11 +326,8 @@ int intel_histogram_set_iet_lut(struct intel_crtc *intel_crtc,
 	if (ret)
 		return ret;
 
-	for (i = 0; i < HISTOGRAM_IET_LENGTH; i++) {
-		intel_de_rmw(display, DPST_BIN(pipe),
-			     DPST_BIN_DATA_MASK, data[i]);
-		drm_dbg_atomic(display->drm, "iet_lut[%d]=%x\n", i, data[i]);
-	}
+	write_iet(display, pipe, data);
+
 	kfree(data);
 	drm_property_blob_put(intel_crtc->base.state->iet_lut);
 
drivers/gpu/drm/i915/display/intel_histogram_regs.h +25 -0
index 213c9f483567cb19a47b44953749f6baf0afe9e7..3fbb9c2deaae6278d5a832dfb61ef860de0c6f21 100644
@@ -45,6 +45,31 @@
 #define _DPST_BIN_B					0x491C4
 #define DPST_BIN(pipe)					_MMIO_PIPE(pipe, _DPST_BIN_A, _DPST_BIN_B)
 #define  DPST_BIN_DATA_MASK				REG_GENMASK(23, 0)
+#define  DPST_BIN_DATA(val)				REG_FIELD_PREP(DPST_BIN_DATA_MASK, val)
 #define  DPST_BIN_BUSY					REG_BIT(31)
 
+#define _DPST_HIST_INDEX_A				0x490D8
+#define _DPST_HIST_INDEX_B				0x491D8
+#define DPST_HIST_INDEX(pipe)				_MMIO_PIPE(pipe, _DPST_HIST_INDEX_A, _DPST_HIST_INDEX_B)
+#define  DPST_HIST_BIN_INDEX_MASK			REG_GENMASK(4, 0)
+#define  DPST_HIST_BIN_INDEX(val)			REG_FIELD_PREP(DPST_HIST_BIN_INDEX_MASK, val)
+
+#define _DPST_HIST_BIN_A				0x490C4
+#define _DPST_HIST_BIN_B				0x491C4
+#define DPST_HIST_BIN(pipe)				_MMIO_PIPE(pipe, _DPST_HIST_BIN_A, _DPST_HIST_BIN_B)
+#define  DPST_HIST_BIN_BUSY				REG_BIT(31)
+#define  DPST_HIST_BIN_DATA_MASK				REG_GENMASK(30, 0)
+
+#define _DPST_IE_BIN_A					0x490CC
+#define _DPST_IE_BIN_B					0x491CC
+#define DPST_IE_BIN(pipe)				_MMIO_PIPE(pipe, _DPST_IE_BIN_A, _DPST_IE_BIN_B)
+#define	 DPST_IE_BIN_DATA_MASK				REG_GENMASK(9, 0)
+#define  DPST_IE_BIN_DATA(val)				REG_FIELD_PREP(DPST_IE_BIN_DATA_MASK, val)
+
+#define _DPST_IE_INDEX_A				0x490DC
+#define _DPST_IE_INDEX_B				0x491DC
+#define DPST_IE_INDEX(pipe)				_MMIO_PIPE(pipe, _DPST_IE_INDEX_A, _DPST_IE_INDEX_B)
+#define  DPST_IE_BIN_INDEX_MASK				REG_GENMASK(6, 0)
+#define  DPST_IE_BIN_INDEX(val)				REG_FIELD_PREP(DPST_IE_BIN_INDEX_MASK, val)
+
 #endif /* __INTEL_HISTOGRAM_REGS_H__ */
Signature · 1 line
2.25.1
#32 lore raw

PATCH v7 14/14 drm/i915/histogram: Enable pipe dithering

In reply to Arun R Murthy

Enable pipe dithering while enabling histogram to overcome some atrifacts seen on the screen.

---
 drivers/gpu/drm/i915/display/intel_histogram.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/gpu/drm/i915/display/intel_histogram.c +11 -0
index d015350b57ed5c8e9aaab71311159bf51e15e9c7..7d0c5d07042c5eb0e33c95e7cadac5c0d1fda379 100644
@@ -22,6 +22,13 @@
 #define HISTOGRAM_BIN_READ_RETRY_COUNT 5
 #define IET_SAMPLE_FORMAT_1_INT_9_FRACT 0x1000009
 
+static void intel_histogram_enable_dithering(struct intel_display *display,
+					     enum pipe pipe)
+{
+	intel_de_rmw(display, PIPE_MISC(pipe), PIPE_MISC_DITHER_ENABLE,
+		     PIPE_MISC_DITHER_ENABLE);
+}
+
 static void set_bin_index_0(struct intel_display *display, enum pipe pipe)
 {
 	if (DISPLAY_VER(display) >= 20)
@@ -200,6 +207,10 @@ static int intel_histogram_enable(struct intel_crtc *intel_crtc, u8 mode)
 
 	if (histogram->enable)
 		return 0;
+
+	/* Pipe Dithering should be enabled with histogram */
+	intel_histogram_enable_dithering(display, pipe);
+
 	 /* enable histogram, clear DPST_BIN reg and select TC function */
 	if (DISPLAY_VER(display) >= 20)
 		intel_de_rmw(display, DPST_CTL(pipe),
Signature · 1 line
2.25.1
Reviewed-by on v7 14/14
#33 lore raw

RE: [PATCH v7 14/14] drm/i915/histogram: Enable pipe dithering

Quoted 14 lines

-----Original Message-----
From: Murthy, Arun R <[email protected]>
Sent: Friday, January 10, 2025 1:16 AM
To: [email protected]; [email protected]; intel- [email protected]
Cc: [email protected]; Kandpal, Suraj <[email protected]>; Shankar, Uma <[email protected]>; "Imported from f20241218-dpst- v7-0-81bfe7d08c2d"@intel.com; 20240705091333.328322-1- [email protected]; Murthy, Arun R <[email protected]>
Subject: [PATCH v7 14/14] drm/i915/histogram: Enable pipe dithering

Quoted 4 lines

Enable pipe dithering while enabling histogram to overcome some atrifacts seen on the screen.

LGTM,

Quoted 39 lines
---
 drivers/gpu/drm/i915/display/intel_histogram.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/gpu/drm/i915/display/intel_histogram.c +0 -0

index d015350b57ed5c8e9aaab71311159bf51e15e9c7..7d0c5d07042c5eb0e33c 95e7cadac5c0d1fda379 100644

drivers/gpu/drm/i915/display/intel_histogram.c +0 -0
@@ -22,6 +22,13 @@
 #define HISTOGRAM_BIN_READ_RETRY_COUNT 5  #define

IET_SAMPLE_FORMAT_1_INT_9_FRACT 0x1000009

drivers/gpu/drm/i915/display/intel_histogram.c +11 -0
+static void intel_histogram_enable_dithering(struct intel_display *display,
+					     enum pipe pipe)
+{
+	intel_de_rmw(display, PIPE_MISC(pipe), PIPE_MISC_DITHER_ENABLE,
+		     PIPE_MISC_DITHER_ENABLE);
+}
+
 static void set_bin_index_0(struct intel_display *display, enum pipe pipe)  {
 	if (DISPLAY_VER(display) >= 20)
@@ -200,6 +207,10 @@ static int intel_histogram_enable(struct intel_crtc
 *intel_crtc, u8 mode)
 
 	if (histogram->enable)
 		return 0;
+
+	/* Pipe Dithering should be enabled with histogram */
+	intel_histogram_enable_dithering(display, pipe);
+
 	 /* enable histogram, clear DPST_BIN reg and select TC function */
 	if (DISPLAY_VER(display) >= 20)

intel_de_rmw(display, DPST_CTL(pipe),

--
2.25.1

#34 lore raw

✗ Fi.CI.BUILD: failure for Display Global Histogram (rev12)

In reply to Arun R Murthy
== Series Details ==

== Series Details ==

Series: Display Global Histogram (rev12)
URL : https://patchwork.freedesktop.org/series/135793/
State : failure

== Summary ==

Error: patch https://patchwork.freedesktop.org/api/1.0/series/135793/revisions/12/mbox/ not applied
Applying: drm: Define histogram structures exposed to user
Applying: drm: Define ImageEnhancemenT LUT structures exposed to user
Applying: drm/crtc: Expose API to create drm crtc property for histogram
Using index info to reconstruct a base tree...
M drivers/gpu/drm/drm_crtc.c
M include/drm/drm_crtc.h
Falling back to patching base and 3-way merge...
Auto-merging include/drm/drm_crtc.h
CONFLICT (content): Merge conflict in include/drm/drm_crtc.h
Auto-merging drivers/gpu/drm/drm_crtc.c
CONFLICT (content): Merge conflict in drivers/gpu/drm/drm_crtc.c
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0003 drm/crtc: Expose API to create drm crtc property for histogram When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
Build failed, no error log produced

#35 lore raw

✗ CI.Patch_applied: failure for Display Global Histogram (rev8)

In reply to Arun R Murthy
== Series Details ==

== Series Details ==

Series: Display Global Histogram (rev8)
URL : https://patchwork.freedesktop.org/series/138867/
State : failure

== Summary ==

=== Applying kernel patches on branch 'drm-tip' with base: ===
Base commit: cf64efebee4a drm-tip: 2025y-01m-09d-21h-54m-27s UTC integration manifest === git am output follows ===
error: patch failed: drivers/gpu/drm/drm_crtc.c:939
error: drivers/gpu/drm/drm_crtc.c: patch does not apply
error: patch failed: include/drm/drm_crtc.h:1323
error: include/drm/drm_crtc.h: patch does not apply
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Applying: drm: Define histogram structures exposed to user
Applying: drm: Define ImageEnhancemenT LUT structures exposed to user
Applying: drm/crtc: Expose API to create drm crtc property for histogram Patch failed at 0003 drm/crtc: Expose API to create drm crtc property for histogram When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".