* [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver
@ 2025-07-14 15:06 Daniel Scally
2025-07-14 15:06 ` [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops Daniel Scally
` (17 more replies)
0 siblings, 18 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus,
Krzysztof Kozlowski
Hello all
Quick note at the top, this series now uses the v4l2 extensible parameters
series that was recently posted [1] so is based on top of that set. It also
takes in the first two patches from the RZV2H IVC driver series [2] - those
new functions are also used here, and the IVC driver actually depends on
patches #1 and #2 of this set so it makes sense to have them here and merge
this set first.
Dan
---------
Original cover message:
This patchset introduces a driver for Arm's Mali-C55 Image Signal Processor.
The driver uses the V4L2 / media controller API and implements both of the ISP's
capture pipelines allowing a range of output formats plus downscaling and
cropping. The capture pipelines are named "Full resolution" and "Downscale" and
so abbreviated FR and DS throughout the driver.
The driver exposes 4 V4L2 subdevices:
- mali-c55 isp: input data formatting
- mali-c55 tpg: test pattern generator (modeled as a camera sensor entity)
- mali-c55 resizer fr: downscale / crop and format setting for the FR pipe
- mali-c55 resizer ds: downscale / crop and format setting for the DS pipe
Along with 4 V4L2 Video devices:
- mali-c55 fr: Capture device for the full resolution pipe
- mali-c55 ds: Capture device for the downscale pipe
- mali-c55 3a stats: Capture device for statistics to support 3A algorithms
- mali-c55 3a params: Output device for parameter buffers to configure the ISP
Support is implemented in the parameters video device code for many of the ISP'S
hardware blocks, but not yet all of them. The buffer format is (as far as I am
aware anyway) a novel design that we intend to be extensible so that support for
the C55's remaining hardware blocks can be added later.
Patches 1, 4, 5, 6 and 7 have already had versions 1-4 on the mailing list...I
decided to post the additional work on the driver as extra patches rather than
merge them all into the existing series as it's already a lot of code to review
and I hoped that that might make it a little easier...if I'm wrong and that's
not liked I can just squash them into a much smaller series.
Previous versions:
v10: https://lore.kernel.org/r/20250624-c55-v10-0-54f3d4196990@ideasonboard.com
v9: https://lore.kernel.org/linux-media/20250519143409.451100-1-dan.scally@ideasonboard.com/#t
v8: https://lore.kernel.org/linux-media/20241106100534.768400-1-dan.scally@ideasonboard.com/
v7: https://lore.kernel.org/linux-media/20240906153406.650105-1-dan.scally@ideasonboard.com/
v6: https://lore.kernel.org/linux-media/20240709132906.3198927-1-dan.scally@ideasonboard.com/
v5: https://lore.kernel.org/linux-media/20240529152858.183799-1-dan.scally@ideasonboard.com/
v4: https://lore.kernel.org/linux-media/20240418090825.624747-1-dan.scally@ideasonboard.com/
v3: https://lore.kernel.org/linux-media/20240305164832.2055437-1-dan.scally@ideasonboard.com/
v2: https://lore.kernel.org/linux-media/20240214141906.245685-1-dan.scally@ideasonboard.com/
v1: https://lore.kernel.org/linux-media/20240131164709.810587-1-dan.scally@ideasonboard.com/
The rest of this message comprises the v4l2-compliance report for the driver.
The tool reports 4 failures, which are because height and width are set to 0 for
MEDIA_BUS_FMT_METADATA_FIXED in the ISP subdev, but this is as directed by the
documentation [3] so I suspect that v4l2-compliance needs amending.
[1] https://lore.kernel.org/linux-media/20250710-extensible-parameters-validation-v2-0-7ec8918ec443@ideasonboard.com/
[2] https://lore.kernel.org/linux-media/20250704-ivc-v3-0-5c45d936ef2e@ideasonboard.com/T/
[3] Documentation/userspace-api/media/v4l-subdev-formats.rst
v4l2-compliance 1.29.0-5246, 64 bits, 64-bit time_t
v4l2-compliance SHA: 201ccf743cde 2024-09-06 07:16:29
Compliance test for mali-c55 device /dev/media0:
Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Required ioctls:
test MEDIA_IOC_DEVICE_INFO: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/media0 open: OK
test MEDIA_IOC_DEVICE_INFO: OK
test for unlimited opens: OK
Media Controller ioctls:
test MEDIA_IOC_G_TOPOLOGY: OK
Entities: 8 Interfaces: 8 Pads: 15 Links: 16
test MEDIA_IOC_ENUM_ENTITIES/LINKS: OK
test MEDIA_IOC_SETUP_LINK: OK
Total for mali-c55 device /dev/media0: 8, Succeeded: 8, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/video0:
Driver Info:
Driver name : mali-c55
Card type : ARM Mali-C55 ISP
Bus info : platform:16080000.isp
Driver version : 6.14.0
Capabilities : 0xa4201000
Video Capture Multiplanar
I/O MC
Streaming
Extended Pix Format
Device Capabilities
Device Caps : 0x24201000
Video Capture Multiplanar
I/O MC
Streaming
Extended Pix Format
Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x0300000f
Type : V4L Video
Entity Info:
ID : 0x0000000d (13)
Name : mali-c55 fr
Function : V4L2 I/O
Pad 0x0100000e : 0: Sink
Link 0x0200001f: from remote pad 0x100000b of entity 'mali-c55 resizer fr' (Video Scaler): Data, Enabled
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/video0 open: OK
test VIDIOC_QUERYCAP: OK
test VIDIOC_G/S_PRIORITY: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_DBG_G/S_REGISTER: OK (Not Supported)
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 1 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Control ioctls (Input 0):
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls (Input 0):
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK
test VIDIOC_TRY_FMT: OK
test VIDIOC_S_FMT: OK
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK
Codec ioctls (Input 0):
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls (Input 0):
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK
test Requests: OK (Not Supported)
test blocking wait: OK
Total for mali-c55 device /dev/video0: 49, Succeeded: 49, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/video1:
Driver Info:
Driver name : mali-c55
Card type : ARM Mali-C55 ISP
Bus info : platform:16080000.isp
Driver version : 6.14.0
Capabilities : 0xac200000
Metadata Output
I/O MC
Streaming
Extended Pix Format
Device Capabilities
Device Caps : 0x2c200000
Metadata Output
I/O MC
Streaming
Extended Pix Format
Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x03000013
Type : V4L Video
Entity Info:
ID : 0x00000011 (17)
Name : mali-c55 3a params
Function : V4L2 I/O
Pad 0x01000012 : 0: Source
Link 0x02000023: to remote pad 0x1000008 of entity 'mali-c55 isp' (Image Signal Processor): Data, Enabled
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/video1 open: OK
test VIDIOC_QUERYCAP: OK
test VIDIOC_G/S_PRIORITY: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_DBG_G/S_REGISTER: OK (Not Supported)
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 1 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Control ioctls (Output 0):
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls (Output 0):
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK
test VIDIOC_TRY_FMT: OK
test VIDIOC_S_FMT: OK
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls (Output 0):
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls (Output 0):
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK
test Requests: OK (Not Supported)
test blocking wait: OK
Total for mali-c55 device /dev/video1: 49, Succeeded: 49, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/video2:
Driver Info:
Driver name : mali-c55
Card type : ARM Mali-C55 ISP
Bus info : platform:16080000.isp
Driver version : 6.14.0
Capabilities : 0x84a00000
Metadata Capture
Streaming
Extended Pix Format
Device Capabilities
Device Caps : 0x04a00000
Metadata Capture
Streaming
Extended Pix Format
Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x03000017
Type : V4L Video
Entity Info:
ID : 0x00000015 (21)
Name : mali-c55 3a stats
Function : V4L2 I/O
Pad 0x01000016 : 0: Sink
Link 0x02000021: from remote pad 0x1000007 of entity 'mali-c55 isp' (Image Signal Processor): Data, Enabled
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/video2 open: OK
test VIDIOC_QUERYCAP: OK
test VIDIOC_G/S_PRIORITY: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_DBG_G/S_REGISTER: OK (Not Supported)
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK
test VIDIOC_TRY_FMT: OK
test VIDIOC_S_FMT: OK
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK
test Requests: OK (Not Supported)
test blocking wait: OK
Total for mali-c55 device /dev/video2: 49, Succeeded: 49, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for rzv2h-ivc device /dev/video3:
Driver Info:
Driver name : rzv2h-ivc
Card type : Renesas Input Video Control
Bus info : platform:16080000.isp
Driver version : 6.14.0
Capabilities : 0x84200002
Video Output
Streaming
Extended Pix Format
Device Capabilities
Device Caps : 0x04200002
Video Output
Streaming
Extended Pix Format
Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x0300002a
Type : V4L Video
Entity Info:
ID : 0x00000028 (40)
Name : rzv2h-ivc
Function : V4L2 I/O
Pad 0x01000029 : 0: Source
Link 0x02000036: to remote pad 0x1000026 of entity 'rzv2h ivc block' (V4L2 I/O): Data, Enabled, Immutable
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/video3 open: OK
test VIDIOC_QUERYCAP: OK
test VIDIOC_G/S_PRIORITY: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_DBG_G/S_REGISTER: OK (Not Supported)
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK
test VIDIOC_TRY_FMT: OK
test VIDIOC_S_FMT: OK
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK
test Requests: OK (Not Supported)
test blocking wait: OK
Total for rzv2h-ivc device /dev/video3: 49, Succeeded: 49, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/v4l-subdev0:
Driver Info:
Driver version : 6.14.0
Capabilities : 0x00000000
Client Capabilities: 0x0000000000000003
streams interval-uses-which Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x0300002e
Type : V4L Sub-Device
Entity Info:
ID : 0x00000001 (1)
Name : mali-c55 tpg
Function : Camera Sensor
Pad 0x01000002 : 0: Source
Link 0x02000019: to remote pad 0x1000004 of entity 'mali-c55 isp' (Image Signal Processor): Data
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_SUDBEV_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/v4l-subdev0 open: OK
test VIDIOC_SUBDEV_QUERYCAP: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Sub-Device ioctls (Source Pad 0):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK
test VIDIOC_QUERYCTRL: OK
test VIDIOC_G/S_CTRL: OK
test VIDIOC_G/S/TRY_EXT_CTRLS: OK
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 6 Private Controls: 0
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK (Not Supported)
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK (Not Supported)
test VIDIOC_TRY_FMT: OK (Not Supported)
test VIDIOC_S_FMT: OK (Not Supported)
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK (Not Supported)
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK (Not Supported)
test Requests: OK (Not Supported)
test blocking wait: OK (Not Supported)
Total for mali-c55 device /dev/v4l-subdev0: 54, Succeeded: 54, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/v4l-subdev1:
Driver Info:
Driver version : 6.14.0
Capabilities : 0x00000000
Client Capabilities: 0x0000000000000003
streams interval-uses-which Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x03000030
Type : V4L Sub-Device
Entity Info:
ID : 0x00000003 (3)
Name : mali-c55 isp
Function : Image Signal Processor
Pad 0x01000004 : 0: Sink, Must Connect
Link 0x02000019: from remote pad 0x1000002 of entity 'mali-c55 tpg' (Camera Sensor): Data
Link 0x0200002c: from remote pad 0x1000027 of entity 'rzv2h ivc block' (V4L2 I/O): Data, Enabled
Pad 0x01000005 : 1: Source
Link 0x0200001b: to remote pad 0x100000a of entity 'mali-c55 resizer fr' (Video Scaler): Data, Enabled, Immutable
Pad 0x01000006 : 2: Source
Link 0x0200001d: to remote pad 0x100000c of entity 'mali-c55 resizer fr' (Video Scaler): Data, Enabled, Immutable
Pad 0x01000007 : 3: Source
Link 0x02000021: to remote pad 0x1000016 of entity 'mali-c55 3a stats' (V4L2 I/O): Data, Enabled
Pad 0x01000008 : 4: Sink
Link 0x02000023: from remote pad 0x1000012 of entity 'mali-c55 3a params' (V4L2 I/O): Data, Enabled
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_SUDBEV_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/v4l-subdev1 open: OK
test VIDIOC_SUBDEV_QUERYCAP: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Sub-Device ioctls (Sink Pad 0):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Source Pad 1):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Source Pad 2):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Source Pad 3):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(369): fmt.width == 0 || fmt.width > 65536
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(418): checkMBusFrameFmt(node, fmt.format)
test Try VIDIOC_SUBDEV_G/S_FMT: FAIL
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(369): fmt.width == 0 || fmt.width > 65536
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(418): checkMBusFrameFmt(node, fmt.format)
test Active VIDIOC_SUBDEV_G/S_FMT: FAIL
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Sink Pad 4):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(369): fmt.width == 0 || fmt.width > 65536
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(418): checkMBusFrameFmt(node, fmt.format)
test Try VIDIOC_SUBDEV_G/S_FMT: FAIL
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(369): fmt.width == 0 || fmt.width > 65536
fail: ../utils/v4l2-compliance/v4l2-test-subdevs.cpp(418): checkMBusFrameFmt(node, fmt.format)
test Active VIDIOC_SUBDEV_G/S_FMT: FAIL
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK
test VIDIOC_QUERYCTRL: OK
test VIDIOC_G/S_CTRL: OK
test VIDIOC_G/S/TRY_EXT_CTRLS: OK
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 1 Private Controls: 1
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK (Not Supported)
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK (Not Supported)
test VIDIOC_TRY_FMT: OK (Not Supported)
test VIDIOC_S_FMT: OK (Not Supported)
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK (Not Supported)
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK (Not Supported)
test Requests: OK (Not Supported)
test blocking wait: OK (Not Supported)
Total for mali-c55 device /dev/v4l-subdev1: 82, Succeeded: 78, Failed: 4, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/v4l-subdev2:
Driver Info:
Driver version : 6.14.0
Capabilities : 0x00000002
Streams Support
Client Capabilities: 0x0000000000000003
streams interval-uses-which Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x03000032
Type : V4L Sub-Device
Entity Info:
ID : 0x00000009 (9)
Name : mali-c55 resizer fr
Function : Video Scaler
Pad 0x0100000a : 0: Sink
Link 0x0200001b: from remote pad 0x1000005 of entity 'mali-c55 isp' (Image Signal Processor): Data, Enabled, Immutable
Pad 0x0100000b : 1: Source
Link 0x0200001f: to remote pad 0x100000e of entity 'mali-c55 fr' (V4L2 I/O): Data, Enabled
Pad 0x0100000c : 2: Sink
Link 0x0200001d: from remote pad 0x1000006 of entity 'mali-c55 isp' (Image Signal Processor): Data, Enabled, Immutable
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_SUDBEV_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/v4l-subdev2 open: OK
test VIDIOC_SUBDEV_QUERYCAP: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Sub-Device routing ioctls:
test Try VIDIOC_SUBDEV_G_ROUTING/VIDIOC_SUBDEV_S_ROUTING: OK
test Active VIDIOC_SUBDEV_G_ROUTING/VIDIOC_SUBDEV_S_ROUTING: OK
Sub-Device ioctls (Sink Pad 0):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Source Pad 1):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Sink Pad 2):
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK (Not Supported)
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK (Not Supported)
test VIDIOC_TRY_FMT: OK (Not Supported)
test VIDIOC_S_FMT: OK (Not Supported)
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK (Not Supported)
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK (Not Supported)
test Requests: OK (Not Supported)
test blocking wait: OK (Not Supported)
Total for mali-c55 device /dev/v4l-subdev2: 63, Succeeded: 63, Failed: 0, Warnings: 0
--------------------------------------------------------------------------------
Compliance test for mali-c55 device /dev/v4l-subdev3:
Driver Info:
Driver version : 6.14.0
Capabilities : 0x00000000
Client Capabilities: 0x0000000000000003
streams interval-uses-which Media Driver Info:
Driver name : mali-c55
Model : ARM Mali-C55 ISP
Serial :
Bus info : platform:16080000.isp
Media version : 6.14.0
Hardware revision: 0x01d982d6 (31032022)
Driver version : 6.14.0
Interface Info:
ID : 0x03000034
Type : V4L Sub-Device
Entity Info:
ID : 0x00000025 (37)
Name : rzv2h ivc block
Function : V4L2 I/O
Pad 0x01000026 : 0: Sink
Link 0x02000036: from remote pad 0x1000029 of entity 'rzv2h-ivc' (V4L2 I/O): Data, Enabled, Immutable
Pad 0x01000027 : 1: Source
Link 0x0200002c: to remote pad 0x1000004 of entity 'mali-c55 isp' (Image Signal Processor): Data, Enabled
Required ioctls:
test MC information (see 'Media Driver Info' above): OK
test VIDIOC_SUDBEV_QUERYCAP: OK
test invalid ioctls: OK
Allow for multiple opens:
test second /dev/v4l-subdev3 open: OK
test VIDIOC_SUBDEV_QUERYCAP: OK
test for unlimited opens: OK
Debug ioctls:
test VIDIOC_LOG_STATUS: OK (Not Supported)
Input ioctls:
test VIDIOC_G/S_TUNER/ENUM_FREQ_BANDS: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_S_HW_FREQ_SEEK: OK (Not Supported)
test VIDIOC_ENUMAUDIO: OK (Not Supported)
test VIDIOC_G/S/ENUMINPUT: OK (Not Supported)
test VIDIOC_G/S_AUDIO: OK (Not Supported)
Inputs: 0 Audio Inputs: 0 Tuners: 0
Output ioctls:
test VIDIOC_G/S_MODULATOR: OK (Not Supported)
test VIDIOC_G/S_FREQUENCY: OK (Not Supported)
test VIDIOC_ENUMAUDOUT: OK (Not Supported)
test VIDIOC_G/S/ENUMOUTPUT: OK (Not Supported)
test VIDIOC_G/S_AUDOUT: OK (Not Supported)
Outputs: 0 Audio Outputs: 0 Modulators: 0
Input/Output configuration ioctls:
test VIDIOC_ENUM/G/S/QUERY_STD: OK (Not Supported)
test VIDIOC_ENUM/G/S/QUERY_DV_TIMINGS: OK (Not Supported)
test VIDIOC_DV_TIMINGS_CAP: OK (Not Supported)
test VIDIOC_G/S_EDID: OK (Not Supported)
Sub-Device ioctls (Sink Pad 0):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Sub-Device ioctls (Source Pad 1):
Try Stream 0
test Try VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Try VIDIOC_SUBDEV_G/S_FMT: OK
test Try VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
Active Stream 0
test Active VIDIOC_SUBDEV_ENUM_MBUS_CODE/FRAME_SIZE/FRAME_INTERVAL: OK
test Active VIDIOC_SUBDEV_G/S_FMT: OK
test Active VIDIOC_SUBDEV_G/S_SELECTION/CROP: OK (Not Supported)
test Active VIDIOC_SUBDEV_G/S_FRAME_INTERVAL: OK (Not Supported)
Control ioctls:
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK (Not Supported)
test VIDIOC_QUERYCTRL: OK (Not Supported)
test VIDIOC_G/S_CTRL: OK (Not Supported)
test VIDIOC_G/S/TRY_EXT_CTRLS: OK (Not Supported)
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK (Not Supported)
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 0 Private Controls: 0
Format ioctls:
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: OK (Not Supported)
test VIDIOC_G/S_PARM: OK (Not Supported)
test VIDIOC_G_FBUF: OK (Not Supported)
test VIDIOC_G_FMT: OK (Not Supported)
test VIDIOC_TRY_FMT: OK (Not Supported)
test VIDIOC_S_FMT: OK (Not Supported)
test VIDIOC_G_SLICED_VBI_CAP: OK (Not Supported)
test Cropping: OK (Not Supported)
test Composing: OK (Not Supported)
test Scaling: OK (Not Supported)
Codec ioctls:
test VIDIOC_(TRY_)ENCODER_CMD: OK (Not Supported)
test VIDIOC_G_ENC_INDEX: OK (Not Supported)
test VIDIOC_(TRY_)DECODER_CMD: OK (Not Supported)
Buffer ioctls:
test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK (Not Supported)
test CREATE_BUFS maximum buffers: OK
test VIDIOC_REMOVE_BUFS: OK
test VIDIOC_EXPBUF: OK (Not Supported)
test Requests: OK (Not Supported)
test blocking wait: OK (Not Supported)
Total for mali-c55 device /dev/v4l-subdev3: 61, Succeeded: 61, Failed: 0, Warnings: 0
Grand Total for mali-c55 device /dev/media0: 464, Succeeded: 460, Failed: 4, Warnings: 0
---
Daniel Scally (18):
media: mc: entity: Add pipeline_started/stopped ops
media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]()
media: uapi: Add MEDIA_BUS_FMT_RGB202020_1X60 format code
media: uapi: Add 20-bit bayer formats
media: v4l2-common: Add RAW16 format info
media: v4l2-common: Add RAW14 format info
dt-bindings: media: Add bindings for ARM mali-c55
media: uapi: Add controls for Mali-C55 ISP
media: mali-c55: Add Mali-C55 ISP driver
media: Documentation: Add Mali-C55 ISP Documentation
MAINTAINERS: Add entry for mali-c55 driver
media: Add MALI_C55_3A_STATS meta format
media: uapi: Add 3a stats buffer for mali-c55
media: platform: Add mali-c55 3a stats devnode
Documentation: mali-c55: Add Statistics documentation
media: uapi: Add parameters structs to mali-c55-config.h
media: platform: Add mali-c55 parameters video node
Documentation: mali-c55: Document the mali-c55 parameter setting
Jacopo Mondi (1):
media: mali-c55: Add image formats for Mali-C55 parameters buffer
Documentation/admin-guide/media/mali-c55-graph.dot | 19 +
Documentation/admin-guide/media/mali-c55.rst | 413 +++++++
Documentation/admin-guide/media/v4l-drivers.rst | 1 +
.../devicetree/bindings/media/arm,mali-c55.yaml | 89 ++
.../userspace-api/media/drivers/index.rst | 1 +
.../userspace-api/media/drivers/mali-c55.rst | 55 +
.../userspace-api/media/v4l/meta-formats.rst | 1 +
.../media/v4l/metafmt-arm-mali-c55.rst | 89 ++
.../userspace-api/media/v4l/subdev-formats.rst | 420 ++++++-
MAINTAINERS | 14 +
drivers/media/mc/mc-entity.c | 46 +
drivers/media/platform/Kconfig | 1 +
drivers/media/platform/Makefile | 1 +
drivers/media/platform/arm/Kconfig | 5 +
drivers/media/platform/arm/Makefile | 2 +
drivers/media/platform/arm/mali-c55/Kconfig | 17 +
drivers/media/platform/arm/mali-c55/Makefile | 11 +
.../media/platform/arm/mali-c55/mali-c55-capture.c | 959 ++++++++++++++++
.../media/platform/arm/mali-c55/mali-c55-common.h | 311 ++++++
.../media/platform/arm/mali-c55/mali-c55-core.c | 1046 ++++++++++++++++++
drivers/media/platform/arm/mali-c55/mali-c55-isp.c | 689 ++++++++++++
.../media/platform/arm/mali-c55/mali-c55-params.c | 825 ++++++++++++++
.../platform/arm/mali-c55/mali-c55-registers.h | 449 ++++++++
.../media/platform/arm/mali-c55/mali-c55-resizer.c | 1156 ++++++++++++++++++++
.../media/platform/arm/mali-c55/mali-c55-stats.c | 428 ++++++++
drivers/media/platform/arm/mali-c55/mali-c55-tpg.c | 438 ++++++++
drivers/media/v4l2-core/v4l2-common.c | 8 +
drivers/media/v4l2-core/v4l2-dev.c | 57 +
drivers/media/v4l2-core/v4l2-ioctl.c | 2 +
include/media/media-entity.h | 29 +
include/media/v4l2-dev.h | 52 +
include/uapi/linux/media-bus-format.h | 9 +-
include/uapi/linux/media/arm/mali-c55-config.h | 796 ++++++++++++++
include/uapi/linux/v4l2-controls.h | 6 +
include/uapi/linux/videodev2.h | 4 +
35 files changed, 8445 insertions(+), 4 deletions(-)
---
base-commit: a8598c7de1bcd94461ca54c972efa9b4ea501fb9
change-id: 20250624-c55-b3c36b2e1d8c
prerequisite-change-id: 20250701-extensible-parameters-validation-c831f7f5cc0b:v2
prerequisite-patch-id: eedc860a79e00a7b875b51c1b27369b7d9343ea7
prerequisite-patch-id: 92fbff7664257963606b4b528a9ce517462a9b6c
prerequisite-patch-id: e7986bc6bf23aced55153711a840568810638aed
prerequisite-patch-id: b5e5b4110ff198562face15fdb99813d3792ebe1
prerequisite-patch-id: e6ebc9baad32ef8286b05f9ebc77b25b2c53d39c
prerequisite-patch-id: 2b002f36ec01afc3571f19cf6b6474e59cd8ed65
prerequisite-patch-id: 64d6914de4c8c877567752ba4fec82328a8a9ea9
prerequisite-patch-id: 35f7062835d60a07930bc8733b3f47771f68c074
Best regards,
--
Daniel Scally <dan.scally@ideasonboard.com>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-08-05 7:45 ` Jacopo Mondi
2025-07-14 15:06 ` [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]() Daniel Scally
` (16 subsequent siblings)
17 siblings, 1 reply; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add two new members to struct media_entity_operations, along with new
functions in media-entity.c to traverse a media pipeline and call the
new operations. The new functions are intended to be used to signal
to a media pipeline that it has fully started, with the entity ops
allowing drivers to define some action to be taken when those
conditions are met.
The combination of the new functions and operations allows drivers
which are part of a multi-driver pipeline to delay actually starting
streaming until all of the conditions for streaming succcessfully are
met across all drivers.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v5:
- Update kerneldoc comments with Optional statement in the
right place
Changes in v4:
- Reverted to having the iter variable
Changes in v3:
- Dropped the iter variable now that the pipeline entity
iterator functions don't need it.
- Updated documentation to specify Optional and return
values
Changes in v2:
- Refactored media_pipeline_started() such that the cleanup
function for media_pipeline_entity_iter is unconditionally
called
- Avoided using media_entity_call() helper for operation that
has return type void to avoid compiler warnings
---
drivers/media/mc/mc-entity.c | 46 ++++++++++++++++++++++++++++++++++++++++++++
include/media/media-entity.h | 29 ++++++++++++++++++++++++++++
2 files changed, 75 insertions(+)
diff --git a/drivers/media/mc/mc-entity.c b/drivers/media/mc/mc-entity.c
index 045590905582054c46656e20463271b1f93fa6b4..d3443537d4304e12cb015630101efba22375c011 100644
--- a/drivers/media/mc/mc-entity.c
+++ b/drivers/media/mc/mc-entity.c
@@ -1053,6 +1053,52 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
}
EXPORT_SYMBOL_GPL(__media_pipeline_entity_iter_next);
+int media_pipeline_started(struct media_pipeline *pipe)
+{
+ struct media_pipeline_entity_iter iter;
+ struct media_entity *entity;
+ int ret;
+
+ ret = media_pipeline_entity_iter_init(pipe, &iter);
+ if (ret)
+ return ret;
+
+ media_pipeline_for_each_entity(pipe, &iter, entity) {
+ ret = media_entity_call(entity, pipeline_started);
+ if (ret && ret != -ENOIOCTLCMD)
+ break;
+ }
+
+ media_pipeline_entity_iter_cleanup(&iter);
+
+ ret = ret == -ENOIOCTLCMD ? 0 : ret;
+ if (ret)
+ media_pipeline_stopped(pipe);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(media_pipeline_started);
+
+int media_pipeline_stopped(struct media_pipeline *pipe)
+{
+ struct media_pipeline_entity_iter iter;
+ struct media_entity *entity;
+ int ret;
+
+ ret = media_pipeline_entity_iter_init(pipe, &iter);
+ if (ret)
+ return ret;
+
+ media_pipeline_for_each_entity(pipe, &iter, entity)
+ if (entity->ops && entity->ops->pipeline_stopped)
+ entity->ops->pipeline_stopped(entity);
+
+ media_pipeline_entity_iter_cleanup(&iter);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(media_pipeline_stopped);
+
/* -----------------------------------------------------------------------------
* Links management
*/
diff --git a/include/media/media-entity.h b/include/media/media-entity.h
index 64cf590b11343f68a456c5870ca2f32917c122f9..1e1026f65f2050bb9aa39bde68794da8d2d0a669 100644
--- a/include/media/media-entity.h
+++ b/include/media/media-entity.h
@@ -269,6 +269,10 @@ struct media_pad {
* media_entity_has_pad_interdep().
* Optional: If the operation isn't implemented all pads
* will be considered as interdependent.
+ * @pipeline_started: Optional: Notify this entity that the pipeline it is a
+ * part of has been started.
+ * @pipeline_stopped: Optional: Notify this entity that the pipeline it is a
+ * part of has been stopped.
*
* .. note::
*
@@ -284,6 +288,8 @@ struct media_entity_operations {
int (*link_validate)(struct media_link *link);
bool (*has_pad_interdep)(struct media_entity *entity, unsigned int pad0,
unsigned int pad1);
+ int (*pipeline_started)(struct media_entity *entity);
+ void (*pipeline_stopped)(struct media_entity *entity);
};
/**
@@ -1261,6 +1267,29 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
entity != NULL; \
entity = __media_pipeline_entity_iter_next((pipe), iter, entity))
+/**
+ * media_pipeline_started - Inform entities in a pipeline that it has started
+ * @pipe: The pipeline
+ *
+ * Iterate on all entities in a media pipeline and call their pipeline_started
+ * member of media_entity_operations.
+ *
+ * Return: zero on success, or a negative error code passed through from an
+ * entity's .pipeline_started() operation.
+ */
+int media_pipeline_started(struct media_pipeline *pipe);
+
+/**
+ * media_pipeline_stopped - Inform entities in a pipeline that it has stopped
+ * @pipe: The pipeline
+ *
+ * Iterate on all entities in a media pipeline and call their pipeline_stopped
+ * member of media_entity_operations.
+ *
+ * Return: zero on success, or -ENOMEM if the iterator initialisation failed.
+ */
+int media_pipeline_stopped(struct media_pipeline *pipe);
+
/**
* media_pipeline_alloc_start - Mark a pipeline as streaming
* @pad: Starting pad
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]()
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
2025-07-14 15:06 ` [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-08-05 8:52 ` Jacopo Mondi
2025-07-14 15:06 ` [PATCH v11 03/19] media: uapi: Add MEDIA_BUS_FMT_RGB202020_1X60 format code Daniel Scally
` (15 subsequent siblings)
17 siblings, 1 reply; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add helpers to run the new media_pipeline_started() and
media_pipeline_stopped() functions. The helpers iterate over the
entities in the pipeline and count the number of video devices and
compare that to the pipeline's start_count() before acting. This
allows us to only run the media pipeline callbacks in the event that
the pipeline has had video_pipeline_start() called for each video
device.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v3:
- Revised the documentation a bit, detailing the specific error
codes and rewording the descriptions
- Renamed __video_device_pipeline_started()
- List the possible return values explicitly
Changes in v2:
- Adapted now media_pipeline_for_each_entity() takes an iter
variable
- Fixed the Return: section of the kerneldoc comments
---
drivers/media/v4l2-core/v4l2-dev.c | 57 ++++++++++++++++++++++++++++++++++++++
include/media/v4l2-dev.h | 52 ++++++++++++++++++++++++++++++++++
2 files changed, 109 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-dev.c b/drivers/media/v4l2-core/v4l2-dev.c
index c369235113d98ae26c30a1aa386e7d60d541a66e..a25038f3c9cc9213b840e2b96d1ef49d17d22fb8 100644
--- a/drivers/media/v4l2-core/v4l2-dev.c
+++ b/drivers/media/v4l2-core/v4l2-dev.c
@@ -1200,6 +1200,63 @@ struct media_pipeline *video_device_pipeline(struct video_device *vdev)
}
EXPORT_SYMBOL_GPL(video_device_pipeline);
+static int video_device_pipeline_unstarted_vdevs(struct media_pipeline *pipe)
+{
+ struct media_pipeline_entity_iter iter;
+ unsigned int n_video_devices = 0;
+ struct media_entity *entity;
+ int ret;
+
+ ret = media_pipeline_entity_iter_init(pipe, &iter);
+ if (ret)
+ return ret;
+
+ media_pipeline_for_each_entity(pipe, &iter, entity) {
+ if (entity->obj_type == MEDIA_ENTITY_TYPE_VIDEO_DEVICE)
+ n_video_devices++;
+ }
+
+ media_pipeline_entity_iter_cleanup(&iter);
+
+ return n_video_devices - pipe->start_count;
+}
+
+int video_device_pipeline_started(struct video_device *vdev)
+{
+ struct media_pipeline *pipe;
+ int ret;
+
+ pipe = video_device_pipeline(vdev);
+ if (!pipe)
+ return -ENODEV;
+
+ ret = video_device_pipeline_unstarted_vdevs(pipe);
+ if (ret)
+ return ret;
+
+ return media_pipeline_started(pipe);
+}
+EXPORT_SYMBOL_GPL(video_device_pipeline_started);
+
+int video_device_pipeline_stopped(struct video_device *vdev)
+{
+ struct media_pipeline *pipe;
+ int ret;
+
+ pipe = video_device_pipeline(vdev);
+ if (!pipe)
+ return -ENODEV;
+
+ ret = video_device_pipeline_unstarted_vdevs(pipe);
+ if (ret)
+ return ret;
+
+ media_pipeline_stopped(pipe);
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(video_device_pipeline_stopped);
+
#endif /* CONFIG_MEDIA_CONTROLLER */
/*
diff --git a/include/media/v4l2-dev.h b/include/media/v4l2-dev.h
index 1b6222fab24eda96cbe459b435431c01f7259366..45a87a5a76e8782c79da9d4890b063b8f0cbb3d8 100644
--- a/include/media/v4l2-dev.h
+++ b/include/media/v4l2-dev.h
@@ -654,6 +654,58 @@ __must_check int video_device_pipeline_alloc_start(struct video_device *vdev);
*/
struct media_pipeline *video_device_pipeline(struct video_device *vdev);
+/**
+ * video_device_pipeline_started - Run the pipeline_started() entity operation
+ * for a fully-started media pipeline
+ * @vdev: A video device that's part of the pipeline
+ *
+ * This function checks whether all MEDIA_ENTITY_TYPE_VIDEO_DEVICE entities
+ * connected to a given video device through enabled links, either directly or
+ * indirectly, have been marked as streaming through the use of
+ * video_device_pipeline_start() or one of its equivalent functions. If so,
+ * media_pipeline_started() is called to inform entities in the pipeline of that
+ * fact. The intention is to provide drivers with a mechanism for checking
+ * whether their pipeline is fully ready to start processing data and call the
+ * .pipeline_started() media entity operation on all the entities in the
+ * pipeline if so.
+ *
+ * Return: The number of video devices in the pipeline remaining to be started,
+ * or a negative error number on failure:
+ *
+ * * 0 - Success
+ * * n > 0 - n video devices in the pipeline remain unstarted
+ * * -ENODEV - The video device has no pipeline
+ * * -ENOMEM - failed to allocate pipeline iterator
+ * * n < 0 - Error passed through from a media entity's
+ * .pipeline_started() operation
+ */
+int video_device_pipeline_started(struct video_device *vdev);
+
+/**
+ * video_device_pipeline_stopped - Run the pipeline_stopped() entity operation
+ * for a fully-started media pipeline
+ * @vdev: A video device that's part of the pipeline
+ *
+ * This function checks whether the MEDIA_ENTITY_TYPE_VIDEO_DEVICE entities
+ * connected to a given video device through enabled links, either directly or
+ * indirectly, have been marked as not streaming through the use of
+ * video_device_pipeline_stop() or one of its equivalent functions. If none of
+ * the video devices in the pipeline have been stopped, then the function
+ * media_pipeline_stopped() is called. The intention is to provide drivers with
+ * a mechanism for checking whether this video device is the first device in the
+ * pipeline to be stopped and call the .pipeline_stopped() media entity
+ * operation on all the entities in the pipeline if so.
+ *
+ * Return: The number of video devices in the pipeline remaining to be started,
+ * or a negative error number on failure:
+ *
+ * * 0 - Success
+ * * n > 0 - n video devices in the pipeline already stopped
+ * * -ENODEV - The video device has no pipeline
+ * * -ENOMEM - failed to allocate pipeline iterator
+ */
+int video_device_pipeline_stopped(struct video_device *vdev);
+
#endif /* CONFIG_MEDIA_CONTROLLER */
#endif /* _V4L2_DEV_H */
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 03/19] media: uapi: Add MEDIA_BUS_FMT_RGB202020_1X60 format code
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
2025-07-14 15:06 ` [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops Daniel Scally
2025-07-14 15:06 ` [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]() Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 04/19] media: uapi: Add 20-bit bayer formats Daniel Scally
` (14 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
The Mali-C55 ISP by ARM requires 20-bits per colour channel input on
the bus. Add a new media bus format code to represent it.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- Fixed the table formatting
Changes in v5:
- none
Changes in v4:
- None
Changes in v3:
- None
Changes in v2:
- none
---
.../userspace-api/media/v4l/subdev-formats.rst | 168 +++++++++++++++++++++
include/uapi/linux/media-bus-format.h | 3 +-
2 files changed, 170 insertions(+), 1 deletion(-)
diff --git a/Documentation/userspace-api/media/v4l/subdev-formats.rst b/Documentation/userspace-api/media/v4l/subdev-formats.rst
index 2a94371448dc07e5c7097421bd82f42dcd7e21aa..57194180fe42f31f88aa0aeb4a4650bf4fb32bcc 100644
--- a/Documentation/userspace-api/media/v4l/subdev-formats.rst
+++ b/Documentation/userspace-api/media/v4l/subdev-formats.rst
@@ -2224,6 +2224,174 @@ The following table list existing packed 48bit wide RGB formats.
\endgroup
+The following table list existing packed 60bit wide RGB formats.
+
+.. tabularcolumns:: |p{4.0cm}|p{0.7cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|
+
+.. _v4l2-mbus-pixelcode-rgb-60:
+
+.. raw:: latex
+
+ \begingroup
+ \tiny
+ \setlength{\tabcolsep}{2pt}
+
+.. flat-table:: 60bit RGB formats
+ :header-rows: 3
+ :stub-columns: 0
+ :widths: 36 7 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
+
+ * - Identifier
+ - Code
+ -
+ - :cspan:`31` Data organization
+ * -
+ -
+ - Bit
+ -
+ -
+ -
+ -
+ - 59
+ - 58
+ - 57
+ - 56
+ - 55
+ - 54
+ - 53
+ - 52
+ - 51
+ - 50
+ - 49
+ - 48
+ - 47
+ - 46
+ - 45
+ - 44
+ - 43
+ - 42
+ - 41
+ - 40
+ - 39
+ - 38
+ - 37
+ - 36
+ - 35
+ - 34
+ - 33
+ - 32
+ * -
+ -
+ -
+ - 31
+ - 30
+ - 29
+ - 28
+ - 27
+ - 26
+ - 25
+ - 24
+ - 23
+ - 22
+ - 21
+ - 20
+ - 19
+ - 18
+ - 17
+ - 16
+ - 15
+ - 14
+ - 13
+ - 12
+ - 11
+ - 10
+ - 9
+ - 8
+ - 7
+ - 6
+ - 5
+ - 4
+ - 3
+ - 2
+ - 1
+ - 0
+ * .. _MEDIA-BUS-FMT-RGB202020-1X60:
+
+ - MEDIA_BUS_FMT_RGB202020_1X60
+ - 0x1026
+ -
+ -
+ -
+ -
+ -
+ - r\ :sub:`19`
+ - r\ :sub:`18`
+ - r\ :sub:`17`
+ - r\ :sub:`16`
+ - r\ :sub:`15`
+ - r\ :sub:`14`
+ - r\ :sub:`13`
+ - r\ :sub:`12`
+ - r\ :sub:`11`
+ - r\ :sub:`10`
+ - r\ :sub:`9`
+ - r\ :sub:`8`
+ - r\ :sub:`7`
+ - r\ :sub:`6`
+ - r\ :sub:`5`
+ - r\ :sub:`4`
+ - r\ :sub:`3`
+ - r\ :sub:`2`
+ - r\ :sub:`1`
+ - r\ :sub:`0`
+ - g\ :sub:`19`
+ - g\ :sub:`18`
+ - g\ :sub:`17`
+ - g\ :sub:`16`
+ - g\ :sub:`15`
+ - g\ :sub:`14`
+ - g\ :sub:`13`
+ - g\ :sub:`12`
+ * -
+ -
+ -
+ - g\ :sub:`11`
+ - g\ :sub:`10`
+ - g\ :sub:`9`
+ - g\ :sub:`8`
+ - g\ :sub:`7`
+ - g\ :sub:`6`
+ - g\ :sub:`5`
+ - g\ :sub:`4`
+ - g\ :sub:`3`
+ - g\ :sub:`2`
+ - g\ :sub:`1`
+ - g\ :sub:`0`
+ - b\ :sub:`19`
+ - b\ :sub:`18`
+ - b\ :sub:`17`
+ - b\ :sub:`16`
+ - b\ :sub:`15`
+ - b\ :sub:`14`
+ - b\ :sub:`13`
+ - b\ :sub:`12`
+ - b\ :sub:`11`
+ - b\ :sub:`10`
+ - b\ :sub:`9`
+ - b\ :sub:`8`
+ - b\ :sub:`7`
+ - b\ :sub:`6`
+ - b\ :sub:`5`
+ - b\ :sub:`4`
+ - b\ :sub:`3`
+ - b\ :sub:`2`
+ - b\ :sub:`1`
+ - b\ :sub:`0`
+
+.. raw:: latex
+
+ \endgroup
+
On LVDS buses, usually each sample is transferred serialized in seven
time slots per pixel clock, on three (18-bit) or four (24-bit) or five (30-bit)
differential data pairs at the same time. The remaining bits are used
diff --git a/include/uapi/linux/media-bus-format.h b/include/uapi/linux/media-bus-format.h
index ff62056feed5b6588bfcfdff178f5b68eecd3a26..62ad82fd285a8ce24bd16c207c51c0e75ea95070 100644
--- a/include/uapi/linux/media-bus-format.h
+++ b/include/uapi/linux/media-bus-format.h
@@ -34,7 +34,7 @@
#define MEDIA_BUS_FMT_FIXED 0x0001
-/* RGB - next is 0x1028 */
+/* RGB - next is 0x1029 */
#define MEDIA_BUS_FMT_RGB444_1X12 0x1016
#define MEDIA_BUS_FMT_RGB444_2X8_PADHI_BE 0x1001
#define MEDIA_BUS_FMT_RGB444_2X8_PADHI_LE 0x1002
@@ -74,6 +74,7 @@
#define MEDIA_BUS_FMT_RGB888_1X36_CPADLO 0x1021
#define MEDIA_BUS_FMT_RGB121212_1X36 0x1019
#define MEDIA_BUS_FMT_RGB161616_1X48 0x101a
+#define MEDIA_BUS_FMT_RGB202020_1X60 0x1028
/* YUV (including grey) - next is 0x202f */
#define MEDIA_BUS_FMT_Y8_1X8 0x2001
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 04/19] media: uapi: Add 20-bit bayer formats
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (2 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 03/19] media: uapi: Add MEDIA_BUS_FMT_RGB202020_1X60 format code Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 05/19] media: v4l2-common: Add RAW16 format info Daniel Scally
` (13 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
The Mali-C55 requires input data be in 20-bit format, MSB aligned.
Add some new media bus format macros to represent that input format.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- Fixed the table formatting
Changes in v5:
- New patch
---
.../userspace-api/media/v4l/subdev-formats.rst | 252 ++++++++++++++++++++-
include/uapi/linux/media-bus-format.h | 6 +-
2 files changed, 255 insertions(+), 3 deletions(-)
diff --git a/Documentation/userspace-api/media/v4l/subdev-formats.rst b/Documentation/userspace-api/media/v4l/subdev-formats.rst
index 57194180fe42f31f88aa0aeb4a4650bf4fb32bcc..ecbe83dcc4b716c77ce37f55fbb8a59b2eed17ab 100644
--- a/Documentation/userspace-api/media/v4l/subdev-formats.rst
+++ b/Documentation/userspace-api/media/v4l/subdev-formats.rst
@@ -2816,7 +2816,7 @@ organization is given as an example for the first pixel only.
\tiny
\setlength{\tabcolsep}{2pt}
-.. tabularcolumns:: |p{6.0cm}|p{0.7cm}|p{0.3cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|
+.. tabularcolumns:: |p{6.0cm}|p{0.7cm}|p{0.3cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|p{0.22cm}|
.. _v4l2-mbus-pixelcode-bayer:
@@ -2829,10 +2829,14 @@ organization is given as an example for the first pixel only.
* - Identifier
- Code
-
- - :cspan:`15` Data organization
+ - :cspan:`19` Data organization
* -
-
- Bit
+ - 19
+ - 18
+ - 17
+ - 16
- 15
- 14
- 13
@@ -2862,6 +2866,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`7`
- b\ :sub:`6`
- b\ :sub:`5`
@@ -2883,6 +2891,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -2904,6 +2916,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -2925,6 +2941,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`7`
- r\ :sub:`6`
- r\ :sub:`5`
@@ -2946,6 +2966,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`7`
- b\ :sub:`6`
- b\ :sub:`5`
@@ -2967,6 +2991,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -2988,6 +3016,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -3009,6 +3041,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`7`
- r\ :sub:`6`
- r\ :sub:`5`
@@ -3030,6 +3066,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`7`
- b\ :sub:`6`
- b\ :sub:`5`
@@ -3051,6 +3091,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -3072,6 +3116,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`7`
- g\ :sub:`6`
- g\ :sub:`5`
@@ -3093,6 +3141,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`7`
- r\ :sub:`6`
- r\ :sub:`5`
@@ -3114,6 +3166,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- 0
- 0
- 0
@@ -3133,6 +3189,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`7`
- b\ :sub:`6`
- b\ :sub:`5`
@@ -3154,6 +3214,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`7`
- b\ :sub:`6`
- b\ :sub:`5`
@@ -3173,6 +3237,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- 0
- 0
- 0
@@ -3194,6 +3262,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`9`
- b\ :sub:`8`
- b\ :sub:`7`
@@ -3213,6 +3285,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`1`
- b\ :sub:`0`
- 0
@@ -3234,6 +3310,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`1`
- b\ :sub:`0`
- 0
@@ -3253,6 +3333,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`9`
- b\ :sub:`8`
- b\ :sub:`7`
@@ -3272,6 +3356,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`9`
- b\ :sub:`8`
- b\ :sub:`7`
@@ -3293,6 +3381,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`9`
- g\ :sub:`8`
- g\ :sub:`7`
@@ -3314,6 +3406,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`9`
- g\ :sub:`8`
- g\ :sub:`7`
@@ -3335,6 +3431,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`9`
- r\ :sub:`8`
- r\ :sub:`7`
@@ -3354,6 +3454,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`11`
- b\ :sub:`10`
- b\ :sub:`9`
@@ -3375,6 +3479,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`11`
- g\ :sub:`10`
- g\ :sub:`9`
@@ -3396,6 +3504,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`11`
- g\ :sub:`10`
- g\ :sub:`9`
@@ -3417,6 +3529,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`11`
- r\ :sub:`10`
- r\ :sub:`9`
@@ -3436,6 +3552,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- b\ :sub:`13`
- b\ :sub:`12`
- b\ :sub:`11`
@@ -3457,6 +3577,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`13`
- g\ :sub:`12`
- g\ :sub:`11`
@@ -3478,6 +3602,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- g\ :sub:`13`
- g\ :sub:`12`
- g\ :sub:`11`
@@ -3499,6 +3627,10 @@ organization is given as an example for the first pixel only.
-
-
-
+ -
+ -
+ -
+ -
- r\ :sub:`13`
- r\ :sub:`12`
- r\ :sub:`11`
@@ -3518,6 +3650,10 @@ organization is given as an example for the first pixel only.
- MEDIA_BUS_FMT_SBGGR16_1X16
- 0x301d
-
+ -
+ -
+ -
+ -
- b\ :sub:`15`
- b\ :sub:`14`
- b\ :sub:`13`
@@ -3539,6 +3675,10 @@ organization is given as an example for the first pixel only.
- MEDIA_BUS_FMT_SGBRG16_1X16
- 0x301e
-
+ -
+ -
+ -
+ -
- g\ :sub:`15`
- g\ :sub:`14`
- g\ :sub:`13`
@@ -3560,6 +3700,10 @@ organization is given as an example for the first pixel only.
- MEDIA_BUS_FMT_SGRBG16_1X16
- 0x301f
-
+ -
+ -
+ -
+ -
- g\ :sub:`15`
- g\ :sub:`14`
- g\ :sub:`13`
@@ -3581,6 +3725,110 @@ organization is given as an example for the first pixel only.
- MEDIA_BUS_FMT_SRGGB16_1X16
- 0x3020
-
+ -
+ -
+ -
+ -
+ - r\ :sub:`15`
+ - r\ :sub:`14`
+ - r\ :sub:`13`
+ - r\ :sub:`12`
+ - r\ :sub:`11`
+ - r\ :sub:`10`
+ - r\ :sub:`9`
+ - r\ :sub:`8`
+ - r\ :sub:`7`
+ - r\ :sub:`6`
+ - r\ :sub:`5`
+ - r\ :sub:`4`
+ - r\ :sub:`3`
+ - r\ :sub:`2`
+ - r\ :sub:`1`
+ - r\ :sub:`0`
+ * .. _MEDIA-BUS-FMT-SBGGR20-1X20:
+
+ - MEDIA_BUS_FMT_SBGGR20_1X20
+ - 0x3021
+ -
+ - b\ :sub:`19`
+ - b\ :sub:`18`
+ - b\ :sub:`17`
+ - b\ :sub:`16`
+ - b\ :sub:`15`
+ - b\ :sub:`14`
+ - b\ :sub:`13`
+ - b\ :sub:`12`
+ - b\ :sub:`11`
+ - b\ :sub:`10`
+ - b\ :sub:`9`
+ - b\ :sub:`8`
+ - b\ :sub:`7`
+ - b\ :sub:`6`
+ - b\ :sub:`5`
+ - b\ :sub:`4`
+ - b\ :sub:`3`
+ - b\ :sub:`2`
+ - b\ :sub:`1`
+ - b\ :sub:`0`
+ * .. _MEDIA-BUS-FMT-SGBRG20-1X20:
+
+ - MEDIA_BUS_FMT_SGBRG20_1X20
+ - 0x3022
+ -
+ - g\ :sub:`19`
+ - g\ :sub:`18`
+ - g\ :sub:`17`
+ - g\ :sub:`16`
+ - g\ :sub:`15`
+ - g\ :sub:`14`
+ - g\ :sub:`13`
+ - g\ :sub:`12`
+ - g\ :sub:`11`
+ - g\ :sub:`10`
+ - g\ :sub:`9`
+ - g\ :sub:`8`
+ - g\ :sub:`7`
+ - g\ :sub:`6`
+ - g\ :sub:`5`
+ - g\ :sub:`4`
+ - g\ :sub:`3`
+ - g\ :sub:`2`
+ - g\ :sub:`1`
+ - g\ :sub:`0`
+ * .. _MEDIA-BUS-FMT-SGRBG20-1X20:
+
+ - MEDIA_BUS_FMT_SGRBG20_1X20
+ - 0x3023
+ -
+ - g\ :sub:`19`
+ - g\ :sub:`18`
+ - g\ :sub:`17`
+ - g\ :sub:`16`
+ - g\ :sub:`15`
+ - g\ :sub:`14`
+ - g\ :sub:`13`
+ - g\ :sub:`12`
+ - g\ :sub:`11`
+ - g\ :sub:`10`
+ - g\ :sub:`9`
+ - g\ :sub:`8`
+ - g\ :sub:`7`
+ - g\ :sub:`6`
+ - g\ :sub:`5`
+ - g\ :sub:`4`
+ - g\ :sub:`3`
+ - g\ :sub:`2`
+ - g\ :sub:`1`
+ - g\ :sub:`0`
+ * .. _MEDIA-BUS-FMT-SRGGB20-1X20:
+
+ - MEDIA_BUS_FMT_SRGGB20_1X20
+ - 0x3024
+ -
+ - r\ :sub:`19`
+ - r\ :sub:`18`
+ - r\ :sub:`17`
+ - r\ :sub:`16`
- r\ :sub:`15`
- r\ :sub:`14`
- r\ :sub:`13`
diff --git a/include/uapi/linux/media-bus-format.h b/include/uapi/linux/media-bus-format.h
index 62ad82fd285a8ce24bd16c207c51c0e75ea95070..6005f033e62c04bdd3dff7ac905f5b921341cd47 100644
--- a/include/uapi/linux/media-bus-format.h
+++ b/include/uapi/linux/media-bus-format.h
@@ -124,7 +124,7 @@
#define MEDIA_BUS_FMT_YUV16_1X48 0x202a
#define MEDIA_BUS_FMT_UYYVYY16_0_5X48 0x202b
-/* Bayer - next is 0x3021 */
+/* Bayer - next is 0x3025 */
#define MEDIA_BUS_FMT_SBGGR8_1X8 0x3001
#define MEDIA_BUS_FMT_SGBRG8_1X8 0x3013
#define MEDIA_BUS_FMT_SGRBG8_1X8 0x3002
@@ -157,6 +157,10 @@
#define MEDIA_BUS_FMT_SGBRG16_1X16 0x301e
#define MEDIA_BUS_FMT_SGRBG16_1X16 0x301f
#define MEDIA_BUS_FMT_SRGGB16_1X16 0x3020
+#define MEDIA_BUS_FMT_SBGGR20_1X20 0x3021
+#define MEDIA_BUS_FMT_SGBRG20_1X20 0x3022
+#define MEDIA_BUS_FMT_SGRBG20_1X20 0x3023
+#define MEDIA_BUS_FMT_SRGGB20_1X20 0x3024
/* JPEG compressed formats - next is 0x4002 */
#define MEDIA_BUS_FMT_JPEG_1X8 0x4001
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 05/19] media: v4l2-common: Add RAW16 format info
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (3 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 04/19] media: uapi: Add 20-bit bayer formats Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 06/19] media: v4l2-common: Add RAW14 " Daniel Scally
` (12 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add entries to v4l2_format_info describing the 16-bit bayer
formats so that they can be used in drivers.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- New patch
---
drivers/media/v4l2-core/v4l2-common.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
index df40756ac8f6c5103d61cd1d4b2ff99b8fbab641..5bac5d45c5ad19f2e76eda127338409d400d5aca 100644
--- a/drivers/media/v4l2-core/v4l2-common.c
+++ b/drivers/media/v4l2-core/v4l2-common.c
@@ -344,6 +344,10 @@ const struct v4l2_format_info *v4l2_format_info(u32 format)
{ .format = V4L2_PIX_FMT_SGBRG12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SGRBG12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SRGGB12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SBGGR16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SGBRG16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SGRBG16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SRGGB16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
/* Renesas Camera Data Receiver Unit formats, bayer order agnostic */
{ .format = V4L2_PIX_FMT_RAW_CRU10, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 8, 0, 0, 0 }, .bpp_div = { 6, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 06/19] media: v4l2-common: Add RAW14 format info
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (4 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 05/19] media: v4l2-common: Add RAW16 format info Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55 Daniel Scally
` (11 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add entries to v4l2_format_info describing the 14-bit bayer
formats so that they can be used in drivers.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- New patch
---
drivers/media/v4l2-core/v4l2-common.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
index 5bac5d45c5ad19f2e76eda127338409d400d5aca..209e73a715fa8346a15eeb19cd2caa26a6b3e9e1 100644
--- a/drivers/media/v4l2-core/v4l2-common.c
+++ b/drivers/media/v4l2-core/v4l2-common.c
@@ -344,6 +344,10 @@ const struct v4l2_format_info *v4l2_format_info(u32 format)
{ .format = V4L2_PIX_FMT_SGBRG12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SGRBG12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SRGGB12, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SBGGR14, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SGBRG14, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SGRBG14, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
+ { .format = V4L2_PIX_FMT_SRGGB14, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SBGGR16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SGBRG16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
{ .format = V4L2_PIX_FMT_SGRBG16, .pixel_enc = V4L2_PIXEL_ENC_BAYER, .mem_planes = 1, .comp_planes = 1, .bpp = { 2, 0, 0, 0 }, .bpp_div = { 1, 1, 1, 1 }, .hdiv = 1, .vdiv = 1 },
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (5 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 06/19] media: v4l2-common: Add RAW14 " Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 21:25 ` Rob Herring
2025-07-31 7:11 ` [PATCH v11.1 " Daniel Scally
2025-07-14 15:06 ` [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP Daniel Scally
` (10 subsequent siblings)
17 siblings, 2 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus,
Krzysztof Kozlowski
Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Dropped in arm,inline_mode property. This is now identical to the
reviewed version 8, so I have kept the tags on there.
Changes in v10:
- None
Changes in v9:
- Added the arm,inline_mode property to differentiate between inline and
memory input configurations
Changes in v8:
- Added the video clock back in. Now that we have actual hardware it's
clear that it's necessary.
- Added reset lines
- Dropped R-bs
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- None
Changes in v4:
- Switched to port instead of ports
Changes in v3:
- Dropped the video clock as suggested by Laurent. I didn't retain it
for the purposes of the refcount since this driver will call .s_stream()
for the sensor driver which will refcount the clock anyway.
- Clarified that the port is a parallel input port rather (Sakari)
Changes in v2:
- Added clocks information
- Fixed the warnings raised by Rob
---
.../devicetree/bindings/media/arm,mali-c55.yaml | 89 ++++++++++++++++++++++
1 file changed, 89 insertions(+)
diff --git a/Documentation/devicetree/bindings/media/arm,mali-c55.yaml b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..539fa8163bd07cf8a80c361012400cbcc7d073eb
--- /dev/null
+++ b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
@@ -0,0 +1,89 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/media/arm,mali-c55.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: ARM Mali-C55 Image Signal Processor
+
+maintainers:
+ - Daniel Scally <dan.scally@ideasonboard.com>
+ - Jacopo Mondi <jacopo.mondi@ideasonboard.com>
+
+properties:
+ compatible:
+ const: arm,mali-c55
+
+ reg:
+ maxItems: 1
+
+ interrupts:
+ maxItems: 1
+
+ clocks:
+ items:
+ - description: ISP Video Clock
+ - description: ISP AXI clock
+ - description: ISP AHB-lite clock
+
+ clock-names:
+ items:
+ - const: vclk
+ - const: aclk
+ - const: hclk
+
+ resets:
+ items:
+ - description: vclk domain reset
+ - description: aclk domain reset
+ - description: hclk domain reset
+
+ reset-names:
+ items:
+ - const: vresetn
+ - const: aresetn
+ - const: hresetn
+
+ port:
+ $ref: /schemas/graph.yaml#/properties/port
+ description: Input parallel video bus
+
+ properties:
+ endpoint:
+ $ref: /schemas/graph.yaml#/properties/endpoint
+
+ arm,inline_mode:
+ description:
+ The ISP can be either electrically connected to sensor and CSI-2 receiver
+ or driven through a DMA input device. This property declares the ISP as
+ being electrically connected to the source of image data.
+ type: boolean
+
+required:
+ - compatible
+ - reg
+ - interrupts
+ - clocks
+ - clock-names
+ - port
+
+additionalProperties: false
+
+examples:
+ - |
+ mali_c55: isp@400000 {
+ compatible = "arm,mali-c55";
+ reg = <0x400000 0x200000>;
+ clocks = <&clk 0>, <&clk 1>, <&clk 2>;
+ clock-names = "vclk", "aclk", "hclk";
+ resets = <&resets 0>, <&resets 1>, <&resets 2>;
+ reset-names = "vresetn", "aresetn", "hresetn";
+ interrupts = <0>;
+
+ port {
+ isp_in: endpoint {
+ remote-endpoint = <&csi2_rx_out>;
+ };
+ };
+ };
+...
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (6 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55 Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-08-05 11:27 ` Jacopo Mondi
2025-07-14 15:06 ` [PATCH v11 10/19] media: Documentation: Add Mali-C55 ISP Documentation Daniel Scally
` (9 subsequent siblings)
17 siblings, 1 reply; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add definitions and documentation for the custom control that will
be needed by the Mali-C55 ISP driver. This will be a read only
bitmask of the driver's capabilities, informing userspace of which
blocks are fitted and which are absent.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Reserve 16 controls in line with other drivers
Changes in v10:
- None
Changes in v9:
- New patch
---
.../userspace-api/media/drivers/index.rst | 1 +
.../userspace-api/media/drivers/mali-c55.rst | 55 ++++++++++++++++++++++
include/uapi/linux/media/arm/mali-c55-config.h | 26 ++++++++++
include/uapi/linux/v4l2-controls.h | 6 +++
4 files changed, 88 insertions(+)
diff --git a/Documentation/userspace-api/media/drivers/index.rst b/Documentation/userspace-api/media/drivers/index.rst
index d706cb47b1122b6e145a02ab826eb3ecc7997c2b..02967c9b18d6e90f414ccc1329c09bffee895e68 100644
--- a/Documentation/userspace-api/media/drivers/index.rst
+++ b/Documentation/userspace-api/media/drivers/index.rst
@@ -32,6 +32,7 @@ For more details see the file COPYING in the source distribution of Linux.
cx2341x-uapi
dw100
imx-uapi
+ mali-c55
max2175
npcm-video
omap3isp-uapi
diff --git a/Documentation/userspace-api/media/drivers/mali-c55.rst b/Documentation/userspace-api/media/drivers/mali-c55.rst
new file mode 100644
index 0000000000000000000000000000000000000000..e8519da77d737b91a931bbe47920af707eebf110
--- /dev/null
+++ b/Documentation/userspace-api/media/drivers/mali-c55.rst
@@ -0,0 +1,55 @@
+.. SPDX-License-Identifier: GPL-2.0-only
+
+Arm Mali-C55 ISP driver
+=======================
+
+The Arm Mali-C55 ISP driver implements a single driver-specific control:
+
+``V4L2_CID_MALI_C55_CAPABILITIES``
+ Detail the capabilities of the ISP by giving detail about the fitted blocks.
+
+ .. flat-table:: Bitmask meaning definitions
+ :header-rows: 1
+ :widths: 2 4 8
+
+ * - Bit
+ - Macro
+ - Meaning
+ * - 0
+ - MALI_C55_GPS_PONG_FITTED
+ - Pong configuration space is fitted in the ISP
+ * - 1
+ - MALI_C55_GPS_WDR_FITTED
+ - WDR Framestitch, offset and gain is fitted in the ISP
+ * - 2
+ - MALI_C55_GPS_COMPRESSION_FITTED
+ - Temper compression is fitted in the ISP
+ * - 3
+ - MALI_C55_GPS_TEMPER_FITTED
+ - Temper is fitted in the ISP
+ * - 4
+ - MALI_C55_GPS_SINTER_LITE_FITTED
+ - Sinter Lite is fitted in the ISP instead of the full Sinter version
+ * - 5
+ - MALI_C55_GPS_SINTER_FITTED
+ - Sinter is fitted in the ISP
+ * - 6
+ - MALI_C55_GPS_IRIDIX_LTM_FITTED
+ - Iridix local tone mappine is fitted in the ISP
+ * - 7
+ - MALI_C55_GPS_IRIDIX_GTM_FITTED
+ - Iridix global tone mapping is fitted in the ISP
+ * - 8
+ - MALI_C55_GPS_CNR_FITTED
+ - Colour noise reduction is fitted in the ISP
+ * - 9
+ - MALI_C55_GPS_FRSCALER_FITTED
+ - The full resolution pipe scaler is fitted in the ISP
+ * - 10
+ - MALI_C55_GPS_DS_PIPE_FITTED
+ - The downscale pipe is fitted in the ISP
+
+ The Mali-C55 ISP can be configured in a number of ways to include or exclude
+ blocks which may not be necessary. This control provides a way for the
+ driver to communicate to userspace which of the blocks are fitted in the
+ design.
\ No newline at end of file
diff --git a/include/uapi/linux/media/arm/mali-c55-config.h b/include/uapi/linux/media/arm/mali-c55-config.h
new file mode 100644
index 0000000000000000000000000000000000000000..2bd60a0d78786be368c2e51b1a0a63fd2a5f690b
--- /dev/null
+++ b/include/uapi/linux/media/arm/mali-c55-config.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+/*
+ * ARM Mali-C55 ISP Driver - Userspace API
+ *
+ * Copyright (C) 2023 Ideas on Board Oy
+ */
+
+#ifndef __UAPI_MALI_C55_CONFIG_H
+#define __UAPI_MALI_C55_CONFIG_H
+
+#include <linux/v4l2-controls.h>
+
+#define V4L2_CID_MALI_C55_CAPABILITIES (V4L2_CID_USER_MALI_C55_BASE + 0x0)
+#define MALI_C55_GPS_PONG_FITTED BIT(0)
+#define MALI_C55_GPS_WDR_FITTED BIT(1)
+#define MALI_C55_GPS_COMPRESSION_FITTED BIT(2)
+#define MALI_C55_GPS_TEMPER_FITTED BIT(3)
+#define MALI_C55_GPS_SINTER_LITE_FITTED BIT(4)
+#define MALI_C55_GPS_SINTER_FITTED BIT(5)
+#define MALI_C55_GPS_IRIDIX_LTM_FITTED BIT(6)
+#define MALI_C55_GPS_IRIDIX_GTM_FITTED BIT(7)
+#define MALI_C55_GPS_CNR_FITTED BIT(8)
+#define MALI_C55_GPS_FRSCALER_FITTED BIT(9)
+#define MALI_C55_GPS_DS_PIPE_FITTED BIT(10)
+
+#endif /* __UAPI_MALI_C55_CONFIG_H */
diff --git a/include/uapi/linux/v4l2-controls.h b/include/uapi/linux/v4l2-controls.h
index f836512e9debbc65d62a9fe04069b056be42f7b2..9f07d80d7af1dfb347fca6e2c65b25bc2cb86af1 100644
--- a/include/uapi/linux/v4l2-controls.h
+++ b/include/uapi/linux/v4l2-controls.h
@@ -228,6 +228,12 @@ enum v4l2_colorfx {
*/
#define V4L2_CID_USER_RKISP1_BASE (V4L2_CID_USER_BASE + 0x1220)
+/*
+ * The base for the Arm Mali-C55 ISP driver controls.
+ * We reserve 8 controls for this driver
+ */
+#define V4L2_CID_USER_MALI_C55_BASE (V4L2_CID_USER_BASE + 0x1230)
+
/* MPEG-class control IDs */
/* The MPEG controls are applicable to all codec controls
* and the 'MPEG' part of the define is historical */
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 10/19] media: Documentation: Add Mali-C55 ISP Documentation
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (7 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 11/19] MAINTAINERS: Add entry for mali-c55 driver Daniel Scally
` (8 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add a documentation page for the mali-c55 driver, which gives a brief
overview of the hardware and explains how to use the driver's capture
devices and the crop/scaler functions.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- Minor formatting changes
Changes in v5:
- None
Changes in v4:
- None
Changes in v3:
- Documented the synchronised buffer sequence numbers (Sakari)
- Clarified that the downscale pipe cannot output raw data, the ISP'S
resolution limits and choice of media bus format code (Kieran)
Changes in v2:
- none
---
Documentation/admin-guide/media/mali-c55-graph.dot | 19 ++
Documentation/admin-guide/media/mali-c55.rst | 340 +++++++++++++++++++++
Documentation/admin-guide/media/v4l-drivers.rst | 1 +
3 files changed, 360 insertions(+)
diff --git a/Documentation/admin-guide/media/mali-c55-graph.dot b/Documentation/admin-guide/media/mali-c55-graph.dot
new file mode 100644
index 0000000000000000000000000000000000000000..0775ba42bf4cfacc3911dc96b3c20a687bd25e4b
--- /dev/null
+++ b/Documentation/admin-guide/media/mali-c55-graph.dot
@@ -0,0 +1,19 @@
+digraph board {
+ rankdir=TB
+ n00000001 [label="{{} | mali-c55 tpg\n/dev/v4l-subdev0 | {<port0> 0}}", shape=Mrecord, style=filled, fillcolor=green]
+ n00000001:port0 -> n00000003:port0 [style=dashed]
+ n00000003 [label="{{<port0> 0} | mali-c55 isp\n/dev/v4l-subdev1 | {<port1> 1 | <port2> 2}}", shape=Mrecord, style=filled, fillcolor=green]
+ n00000003:port1 -> n00000007:port0 [style=bold]
+ n00000003:port2 -> n00000007:port2 [style=bold]
+ n00000003:port1 -> n0000000b:port0 [style=bold]
+ n00000007 [label="{{<port0> 0 | <port2> 2} | mali-c55 resizer fr\n/dev/v4l-subdev2 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green]
+ n00000007:port1 -> n0000000e [style=bold]
+ n0000000b [label="{{<port0> 0} | mali-c55 resizer ds\n/dev/v4l-subdev3 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green]
+ n0000000b:port1 -> n00000012 [style=bold]
+ n0000000e [label="mali-c55 fr\n/dev/video0", shape=box, style=filled, fillcolor=yellow]
+ n00000012 [label="mali-c55 ds\n/dev/video1", shape=box, style=filled, fillcolor=yellow]
+ n00000022 [label="{{<port0> 0} | csi2-rx\n/dev/v4l-subdev4 | {<port1> 1}}", shape=Mrecord, style=filled, fillcolor=green]
+ n00000022:port1 -> n00000003:port0
+ n00000027 [label="{{} | imx415 1-001a\n/dev/v4l-subdev5 | {<port0> 0}}", shape=Mrecord, style=filled, fillcolor=green]
+ n00000027:port0 -> n00000022:port0 [style=bold]
+}
\ No newline at end of file
diff --git a/Documentation/admin-guide/media/mali-c55.rst b/Documentation/admin-guide/media/mali-c55.rst
new file mode 100644
index 0000000000000000000000000000000000000000..72cdded507b3a7918e28d9bd9a1c6f72ddf9e664
--- /dev/null
+++ b/Documentation/admin-guide/media/mali-c55.rst
@@ -0,0 +1,340 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+==========================================
+ARM Mali-C55 Image Signal Processor driver
+==========================================
+
+Introduction
+============
+
+This file documents the driver for ARM's Mali-C55 Image Signal Processor. The
+driver is located under drivers/media/platform/arm/mali-c55.
+
+The Mali-C55 ISP receives data in either raw Bayer format or RGB/YUV format from
+sensors through either a parallel interface or a memory bus before processing it
+and outputting it through an internal DMA engine. Two output pipelines are
+possible (though one may not be fitted, depending on the implementation). These
+are referred to as "Full resolution" and "Downscale", but the naming is historic
+and both pipes are capable of cropping/scaling operations. The full resolution
+pipe is also capable of outputting RAW data, bypassing much of the ISP's
+processing. The downscale pipe cannot output RAW data. An integrated test
+pattern generator can be used to drive the ISP and produce image data in the
+absence of a connected camera sensor. The driver module is named mali_c55, and
+is enabled through the CONFIG_VIDEO_MALI_C55 config option.
+
+The driver implements V4L2, Media Controller and V4L2 Subdevice interfaces and
+expects camera sensors connected to the ISP to have V4L2 subdevice interfaces.
+
+Mali-C55 ISP hardware
+=====================
+
+A high level functional view of the Mali-C55 ISP is presented below. The ISP
+takes input from either a live source or through a DMA engine for memory input,
+depending on the SoC integration.::
+
+ +---------+ +----------+ +--------+
+ | Sensor |--->| CSI-2 Rx | "Full Resolution" | DMA |
+ +---------+ +----------+ |\ Output +--->| Writer |
+ | | \ | +--------+
+ | | \ +----------+ +------+---> Streaming I/O
+ +------------+ +------->| | | | |
+ | | | |-->| Mali-C55 |--+
+ | DMA Reader |--------------->| | | ISP | |
+ | | | / | | | +---> Streaming I/O
+ +------------+ | / +----------+ | |
+ |/ +------+
+ | +--------+
+ +--->| DMA |
+ "Downscaled" | Writer |
+ Output +--------+
+
+Media Controller Topology
+=========================
+
+An example of the ISP's topology (as implemented in a system with an IMX415
+camera sensor and generic CSI-2 receiver) is below:
+
+
+.. kernel-figure:: mali-c55-graph.dot
+ :alt: mali-c55-graph.dot
+ :align: center
+
+The driver has 4 V4L2 subdevices:
+
+- `mali_c55 isp`: Responsible for configuring input crop and color space
+ conversion
+- `mali_c55 tpg`: The test pattern generator, emulating a camera sensor.
+- `mali_c55 resizer fr`: The Full-Resolution pipe resizer
+- `mali_c55 resizer ds`: The Downscale pipe resizer
+
+The driver has 2 V4L2 video devices:
+
+- `mali-c55 fr`: The full-resolution pipe's capture device
+- `mali-c55 ds`: The downscale pipe's capture device
+
+Frame sequences are synchronised across to two capture devices, meaning if one
+pipe is started later than the other the sequence numbers returned in its
+buffers will match those of the other pipe rather than starting from zero.
+
+Idiosyncrasies
+--------------
+
+**mali-c55 isp**
+The `mali-c55 isp` subdevice has a single sink pad to which all sources of data
+should be connected. The active source is selected by enabling the appropriate
+media link and disabling all others. The ISP has two source pads, reflecting the
+different paths through which it can internally route data. Tap points within
+the ISP allow users to divert data to avoid processing by some or all of the
+hardware's processing steps. The diagram below is intended only to highlight how
+the bypassing works and is not a true reflection of those processing steps; for
+a high-level functional block diagram see ARM's developer page for the
+ISP [3]_::
+
+ +--------------------------------------------------------------+
+ | Possible Internal ISP Data Routes |
+ | +------------+ +----------+ +------------+ |
+ +---+ | | | | | Colour | +---+
+ | 0 |--+-->| Processing |->| Demosaic |->| Space |--->| 1 |
+ +---+ | | | | | | Conversion | +---+
+ | | +------------+ +----------+ +------------+ |
+ | | +---+
+ | +---------------------------------------------------| 2 |
+ | +---+
+ | |
+ +--------------------------------------------------------------+
+
+
+.. flat-table::
+ :header-rows: 1
+
+ * - Pad
+ - Direction
+ - Purpose
+
+ * - 0
+ - sink
+ - Data input, connected to the TPG and camera sensors
+
+ * - 1
+ - source
+ - RGB/YUV data, connected to the FR and DS V4L2 subdevices
+
+ * - 2
+ - source
+ - RAW bayer data, connected to the FR V4L2 subdevices
+
+The ISP is limited to both input and output resolutions between 640x480 and
+8192x8192, and this is reflected in the ISP and resizer subdevice's .set_fmt()
+operations.
+
+**mali-c55 resizer fr**
+The `mali-c55 resizer fr` subdevice has two _sink_ pads to reflect the different
+insertion points in the hardware (either RAW or demosaiced data):
+
+.. flat-table::
+ :header-rows: 1
+
+ * - Pad
+ - Direction
+ - Purpose
+
+ * - 0
+ - sink
+ - Data input connected to the ISP's demosaiced stream.
+
+ * - 1
+ - source
+ - Data output connected to the capture video device
+
+ * - 2
+ - sink
+ - Data input connected to the ISP's raw data stream
+
+The data source in use is selected through the routing API; two routes each of a
+single stream are available:
+
+.. flat-table::
+ :header-rows: 1
+
+ * - Sink Pad
+ - Source Pad
+ - Purpose
+
+ * - 0
+ - 1
+ - Demosaiced data route
+
+ * - 2
+ - 1
+ - Raw data route
+
+
+If the demosaiced route is active then the FR pipe is only capable of output
+in RGB/YUV formats. If the raw route is active then the output reflects the
+input (which may be either Bayer or RGB/YUV data).
+
+Using the driver to capture video
+=================================
+
+Using the media controller APIs we can configure the input source and ISP to
+capture images in a variety of formats. In the examples below, configuring the
+media graph is done with the v4l-utils [1]_ package's media-ctl utility.
+Capturing the images is done with yavta [2]_.
+
+Configuring the input source
+----------------------------
+
+The first step is to set the input source that we wish by enabling the correct
+media link. Using the example topology above, we can select the TPG as follows:
+
+.. code-block:: none
+
+ media-ctl -l "'lte-csi2-rx':1->'mali-c55 isp':0[0]"
+ media-ctl -l "'mali-c55 tpg':0->'mali-c55 isp':0[1]"
+
+Configuring which video devices will stream data
+------------------------------------------------
+
+The driver will wait for all video devices to have their VIDIOC_STREAMON ioctl
+called before it tells the sensor to start streaming. To facilitate this we need
+to enable links to the video devices that we want to use. In the example below
+we enable the links to both of the image capture video devices
+
+.. code-block:: none
+
+ media-ctl -l "'mali-c55 resizer fr':1->'mali-c55 fr':0[1]"
+ media-ctl -l "'mali-c55 resizer ds':1->'mali-c55 ds':0[1]"
+
+Capturing bayer data from the source and processing to RGB/YUV
+--------------------------------------------------------------
+
+To capture 1920x1080 bayer data from the source and push it through the ISP's
+full processing pipeline, we configure the data formats appropriately on the
+source, ISP and resizer subdevices and set the FR resizer's routing to select
+processed data. The media bus format on the resizer's source pad will be either
+RGB121212_1X36 or YUV10_1X30, depending on whether you want to capture RGB or
+YUV. The ISP's debayering block outputs RGB data natively, setting the source
+pad format to YUV10_1X30 enables the colour space conversion block.
+
+In this example we target RGB565 output, so select RGB121212_1X36 as the resizer
+source pad's format:
+
+.. code-block:: none
+
+ # Set formats on the TPG and ISP
+ media-ctl -V "'mali-c55 tpg':0[fmt:SRGGB20_1X20/1920x1080]"
+ media-ctl -V "'mali-c55 isp':0[fmt:SRGGB20_1X20/1920x1080]"
+ media-ctl -V "'mali-c55 isp':1[fmt:SRGGB20_1X20/1920x1080]"
+
+ # Set routing on the FR resizer
+ media-ctl -R "'mali-c55 resizer fr'[0/0->1/0[1],2/0->1/0[0]]"
+
+ # Set format on the resizer, must be done AFTER the routing.
+ media-ctl -V "'mali-c55 resizer fr':1[fmt:RGB121212_1X36/1920x1080]"
+
+The downscale output can also be used to stream data at the same time. In this
+case since only processed data can be captured through the downscale output no
+routing need be set:
+
+.. code-block:: none
+
+ # Set format on the resizer
+ media-ctl -V "'mali-c55 resizer ds':1[fmt:RGB121212_1X36/1920x1080]"
+
+Following which images can be captured from both the FR and DS output's video
+devices (simultaneously, if desired):
+
+.. code-block:: none
+
+ yavta -f RGB565 -s 1920x1080 -c10 /dev/video0
+ yavta -f RGB565 -s 1920x1080 -c10 /dev/video1
+
+Cropping the image
+~~~~~~~~~~~~~~~~~~
+
+Both the full resolution and downscale pipes can crop to a minimum resolution of
+640x480. To crop the image simply configure the resizer's sink pad's crop and
+compose rectangles and set the format on the video device:
+
+.. code-block:: none
+
+ media-ctl -V "'mali-c55 resizer fr':0[fmt:RGB121212_1X36/1920x1080 crop:(480,270)/640x480 compose:(0,0)/640x480]"
+ media-ctl -V "'mali-c55 resizer fr':1[fmt:RGB121212_1X36/640x480]"
+ yavta -f RGB565 -s 640x480 -c10 /dev/video0
+
+Downscaling the image
+~~~~~~~~~~~~~~~~~~~~~
+
+Both the full resolution and downscale pipes can downscale the image by up to 8x
+provided the minimum 640x480 output resolution is adhered to. For the best image
+result the scaling ratio for each direction should be the same. To configure
+scaling we use the compose rectangle on the resizer's sink pad:
+
+.. code-block:: none
+
+ media-ctl -V "'mali-c55 resizer fr':0[fmt:RGB121212_1X36/1920x1080 crop:(0,0)/1920x1080 compose:(0,0)/640x480]"
+ media-ctl -V "'mali-c55 resizer fr':1[fmt:RGB121212_1X36/640x480]"
+ yavta -f RGB565 -s 640x480 -c10 /dev/video0
+
+Capturing images in YUV formats
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+If we need to output YUV data rather than RGB the color space conversion block
+needs to be active, which is achieved by setting MEDIA_BUS_FMT_YUV10_1X30 on the
+resizer's source pad. We can then configure a capture format like NV12 (here in
+its multi-planar variant)
+
+.. code-block:: none
+
+ media-ctl -V "'mali-c55 resizer fr':1[fmt:YUV10_1X30/1920x1080]"
+ yavta -f NV12M -s 1920x1080 -c10 /dev/video0
+
+Capturing RGB data from the source and processing it with the resizers
+----------------------------------------------------------------------
+
+The Mali-C55 ISP can work with sensors capable of outputting RGB data. In this
+case although none of the image quality blocks would be used it can still
+crop/scale the data in the usual way. For this reason RGB data input to the ISP
+still goes through the ISP subdevice's pad 1 to the resizer.
+
+To achieve this, the ISP's sink pad's format is set to
+MEDIA_BUS_FMT_RGB202020_1X60 - this reflects the format that data must be in to
+work with the ISP. Converting the camera sensor's output to that format is the
+responsibility of external hardware.
+
+In this example we ask the test pattern generator to give us RGB data instead of
+bayer.
+
+.. code-block:: none
+
+ media-ctl -V "'mali-c55 tpg':0[fmt:RGB202020_1X60/1920x1080]"
+ media-ctl -V "'mali-c55 isp':0[fmt:RGB202020_1X60/1920x1080]"
+
+Cropping or scaling the data can be done in exactly the same way as outlined
+earlier.
+
+Capturing raw data from the source and outputting it unmodified
+-----------------------------------------------------------------
+
+The ISP can additionally capture raw data from the source and output it on the
+full resolution pipe only, completely unmodified. In this case the downscale
+pipe can still process the data normally and be used at the same time.
+
+To configure raw bypass the FR resizer's subdevice's routing table needs to be
+configured, followed by formats in the appropriate places:
+
+.. code-block:: none
+
+ media-ctl -R "'mali-c55 resizer fr'[0/0->1/0[0],2/0->1/0[1]]"
+ media-ctl -V "'mali-c55 isp':0[fmt:RGB202020_1X60/1920x1080]"
+ media-ctl -V "'mali-c55 resizer fr':2[fmt:RGB202020_1X60/1920x1080]"
+ media-ctl -V "'mali-c55 resizer fr':1[fmt:RGB202020_1X60/1920x1080]"
+
+ # Set format on the video device and stream
+ yavta -f RGB565 -s 1920x1080 -c10 /dev/video0
+
+References
+==========
+.. [1] https://git.linuxtv.org/v4l-utils.git/
+.. [2] https://git.ideasonboard.org/yavta.git
+.. [3] https://developer.arm.com/Processors/Mali-C55
diff --git a/Documentation/admin-guide/media/v4l-drivers.rst b/Documentation/admin-guide/media/v4l-drivers.rst
index 3bac5165b13416ee94d9f98ccc22d07d334ae90e..ebd9b15efa903ecb62869d4b7e0b0d6f3ed73589 100644
--- a/Documentation/admin-guide/media/v4l-drivers.rst
+++ b/Documentation/admin-guide/media/v4l-drivers.rst
@@ -19,6 +19,7 @@ Video4Linux (V4L) driver-specific documentation
ipu3
ipu6-isys
ivtv
+ mali-c55
mgb4
omap3isp
philips
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 11/19] MAINTAINERS: Add entry for mali-c55 driver
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (8 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 10/19] media: Documentation: Add Mali-C55 ISP Documentation Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 12/19] media: Add MALI_C55_3A_STATS meta format Daniel Scally
` (7 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add a MAINTAINERS entry for the mali-c55 driver and its associated
documentation.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- Added a File line for
Documentation/userspace-api/media/drivers/mali-c55.rst
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- None
Changes in v4:
- None
Changes in v3:
- none
Changes in v2:
- none
---
MAINTAINERS | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index f03c10092a891a06052484b691409f0c459de87d..7d0b268ed7588fbffb767cd962df6b0e31a9b842 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2028,6 +2028,18 @@ F: drivers/gpu/drm/ci/xfails/panfrost*
F: drivers/gpu/drm/panfrost/
F: include/uapi/drm/panfrost_drm.h
+ARM MALI-C55 ISP DRIVER
+M: Daniel Scally <dan.scally@ideasonboard.com>
+M: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
+L: linux-media@vger.kernel.org
+S: Maintained
+T: git git://linuxtv.org/media_tree.git
+F: Documentation/admin-guide/media/mali-c55-graph.dot
+F: Documentation/admin-guide/media/mali-c55.rst
+F: Documentation/devicetree/bindings/media/arm,mali-c55.yaml
+F: Documentation/userspace-api/media/drivers/mali-c55.rst
+F: drivers/media/platform/arm/mali-c55/
+
ARM MALI PANTHOR DRM DRIVER
M: Boris Brezillon <boris.brezillon@collabora.com>
M: Steven Price <steven.price@arm.com>
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 12/19] media: Add MALI_C55_3A_STATS meta format
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (9 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 11/19] MAINTAINERS: Add entry for mali-c55 driver Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 13/19] media: uapi: Add 3a stats buffer for mali-c55 Daniel Scally
` (6 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add a new meta format for the Mali-C55 ISP's 3A Statistics along
with a new descriptor entry.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- New patch
Changes in v5:
- New patch
---
drivers/media/v4l2-core/v4l2-ioctl.c | 1 +
include/uapi/linux/videodev2.h | 3 +++
2 files changed, 4 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
index be94a79b976e3de4eb957f5d2584ec6d4230469e..d2e7d6b115b53980fef959cbd9c8e758819d20e9 100644
--- a/drivers/media/v4l2-core/v4l2-ioctl.c
+++ b/drivers/media/v4l2-core/v4l2-ioctl.c
@@ -1470,6 +1470,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
case V4L2_META_FMT_RK_ISP1_EXT_PARAMS: descr = "Rockchip ISP1 Ext 3A Params"; break;
case V4L2_META_FMT_C3ISP_PARAMS: descr = "Amlogic C3 ISP Parameters"; break;
case V4L2_META_FMT_C3ISP_STATS: descr = "Amlogic C3 ISP Statistics"; break;
+ case V4L2_META_FMT_MALI_C55_STATS: descr = "ARM Mali-C55 ISP 3A Statistics"; break;
case V4L2_PIX_FMT_NV12_8L128: descr = "NV12 (8x128 Linear)"; break;
case V4L2_PIX_FMT_NV12M_8L128: descr = "NV12M (8x128 Linear)"; break;
case V4L2_PIX_FMT_NV12_10BE_8L128: descr = "10-bit NV12 (8x128 Linear, BE)"; break;
diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
index 6f7bd38dd5aa4b1b2084685512512a380d76a5e4..752d4d6ae5ca4ca924dd8a58aed6c4af3304879e 100644
--- a/include/uapi/linux/videodev2.h
+++ b/include/uapi/linux/videodev2.h
@@ -883,6 +883,9 @@ struct v4l2_pix_format {
#define V4L2_META_FMT_RPI_FE_CFG v4l2_fourcc('R', 'P', 'F', 'C') /* PiSP FE configuration */
#define V4L2_META_FMT_RPI_FE_STATS v4l2_fourcc('R', 'P', 'F', 'S') /* PiSP FE stats */
+/* Vendor specific - used for Arm Mali-C55 ISP */
+#define V4L2_META_FMT_MALI_C55_STATS v4l2_fourcc('C', '5', '5', 'S') /* ARM Mali-C55 3A Statistics */
+
#ifdef __KERNEL__
/*
* Line-based metadata formats. Remember to update v4l_fill_fmtdesc() when
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 13/19] media: uapi: Add 3a stats buffer for mali-c55
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (10 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 12/19] media: Add MALI_C55_3A_STATS meta format Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 14/19] media: platform: Add mali-c55 3a stats devnode Daniel Scally
` (5 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Describe the format of the 3A statistics buffers in the userspace API
header for the mali-c55 ISP.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- New patch
Changes in v5:
- New patch
---
MAINTAINERS | 1 +
include/uapi/linux/media/arm/mali-c55-config.h | 170 +++++++++++++++++++++++++
2 files changed, 171 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 7d0b268ed7588fbffb767cd962df6b0e31a9b842..87255112240dbce86bca8d9a6683015ce12a3bce 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2039,6 +2039,7 @@ F: Documentation/admin-guide/media/mali-c55.rst
F: Documentation/devicetree/bindings/media/arm,mali-c55.yaml
F: Documentation/userspace-api/media/drivers/mali-c55.rst
F: drivers/media/platform/arm/mali-c55/
+F: include/uapi/linux/media/arm/mali-c55-config.h
ARM MALI PANTHOR DRM DRIVER
M: Boris Brezillon <boris.brezillon@collabora.com>
diff --git a/include/uapi/linux/media/arm/mali-c55-config.h b/include/uapi/linux/media/arm/mali-c55-config.h
index 2bd60a0d78786be368c2e51b1a0a63fd2a5f690b..8d31764140a0ef6aa8b0c771a4ed7565c9c96670 100644
--- a/include/uapi/linux/media/arm/mali-c55-config.h
+++ b/include/uapi/linux/media/arm/mali-c55-config.h
@@ -8,6 +8,7 @@
#ifndef __UAPI_MALI_C55_CONFIG_H
#define __UAPI_MALI_C55_CONFIG_H
+#include <linux/types.h>
#include <linux/v4l2-controls.h>
#define V4L2_CID_MALI_C55_CAPABILITIES (V4L2_CID_USER_MALI_C55_BASE + 0x0)
@@ -23,4 +24,173 @@
#define MALI_C55_GPS_FRSCALER_FITTED BIT(9)
#define MALI_C55_GPS_DS_PIPE_FITTED BIT(10)
+/*
+ * Frames are split into zones of almost equal width and height - a zone is a
+ * rectangular tile of a frame. The metering blocks within the ISP collect
+ * aggregated statistics per zone. A maximum of 15x15 zones can be configured,
+ * and so the statistics buffer within the hardware is sized to accommodate
+ * that.
+ *
+ * The utilised number of zones is runtime configurable.
+ */
+#define MALI_C55_MAX_ZONES (15 * 15)
+
+/**
+ * struct mali_c55_ae_1024bin_hist - Auto Exposure 1024-bin histogram statistics
+ *
+ * @bins: 1024 element array of 16-bit pixel counts.
+ *
+ * The 1024-bin histogram module collects image-global but zone-weighted
+ * intensity distributions of pixels in fixed-width bins. The modules can be
+ * configured into different "plane modes" which affect the contents of the
+ * collected statistics. In plane mode 0, pixel intensities are taken regardless
+ * of colour plane into a single 1024-bin histogram with a bin width of 4. In
+ * plane mode 1, four 256-bin histograms with a bin width of 16 are collected -
+ * one for each CFA colour plane. In plane modes 4, 5, 6 and 7 two 512-bin
+ * histograms with a bin width of 8 are collected - in each mode one of the
+ * colour planes is collected into the first histogram and all the others are
+ * combined into the second. The histograms are stored consecutively in the bins
+ * array.
+ *
+ * The 16-bit pixel counts are stored as a 4-bit exponent in the most
+ * significant bits followed by a 12-bit mantissa. Conversion to a usable
+ * format can be done according to the following pseudo-code::
+ *
+ * if (e == 0) {
+ * bin = m * 2;
+ * } else {
+ * bin = (m + 4096) * 2^e
+ * }
+ *
+ * where
+ * e is the exponent value in range 0..15
+ * m is the mantissa value in range 0..4095
+ *
+ * The pixels used in calculating the statistics can be masked using three
+ * methods:
+ *
+ * 1. Pixels can be skipped in X and Y directions independently.
+ * 2. Minimum/Maximum intensities can be configured
+ * 3. Zones can be differentially weighted, including 0 weighted to mask them
+ *
+ * The data for this histogram can be collected from different tap points in the
+ * ISP depending on configuration - after the white balance or digital gain
+ * blocks, or immediately after the input crossbar.
+ */
+struct mali_c55_ae_1024bin_hist {
+ __u16 bins[1024];
+} __attribute__((packed));
+
+/**
+ * struct mali_c55_ae_5bin_hist - Auto Exposure 5-bin histogram statistics
+ *
+ * @hist0: 16-bit normalised pixel count for the 0th intensity bin
+ * @hist1: 16-bit normalised pixel count for the 1st intensity bin
+ * @hist3: 16-bit normalised pixel count for the 3rd intensity bin
+ * @hist4: 16-bit normalised pixel count for the 4th intensity bin
+ *
+ * The ISP generates a 5-bin histogram of normalised pixel counts within bins of
+ * pixel intensity for each of 225 possible zones within a frame. The centre bin
+ * of the histogram for each zone is not available from the hardware and must be
+ * calculated by subtracting the values of hist0, hist1, hist3 and hist4 from
+ * 0xffff as in the following equation:
+ *
+ * hist2 = 0xffff - (hist0 + hist1 + hist3 + hist4)
+ */
+struct mali_c55_ae_5bin_hist {
+ __u16 hist0;
+ __u16 hist1;
+ __u16 hist3;
+ __u16 hist4;
+} __attribute__((packed));
+
+/**
+ * struct mali_c55_awb_average_ratios - Auto White Balance colour ratios
+ *
+ * @avg_rg_gr: Average R/G or G/R ratio in Q4.8 format.
+ * @avg_bg_br: Average B/G or B/R ratio in Q4.8 format.
+ * @num_pixels: The number of pixels used in the AWB calculation
+ *
+ * The ISP calculates and collects average colour ratios for each zone in an
+ * image and stores them in Q4.8 format (the lowest 8 bits are fractional, with
+ * bits [11:8] representing the integer). The exact ratios collected (either
+ * R/G, B/G or G/R, B/R) are configurable through the parameters buffer. The
+ * value of the 4 high bits is undefined.
+ */
+struct mali_c55_awb_average_ratios {
+ __u16 avg_rg_gr;
+ __u16 avg_bg_br;
+ __u32 num_pixels;
+} __attribute__((packed));
+
+/**
+ * struct mali_c55_af_statistics - Auto Focus edge and intensity statistics
+ *
+ * @intensity_stats: Packed mantissa and exponent value for pixel intensity
+ * @edge_stats: Packed mantissa and exponent values for edge intensity
+ *
+ * The ISP collects the squared sum of pixel intensities for each zone within a
+ * configurable Region of Interest on the frame. Additionally, the same data are
+ * collected after being passed through a bandpass filter which removes high and
+ * low frequency components - these are referred to as the edge statistics.
+ *
+ * The intensity and edge statistics for a zone can be used to calculate the
+ * contrast information for a zone
+ *
+ * C = E2 / I2
+ *
+ * Where I2 is the intensity statistic for a zone and E2 is the edge statistic
+ * for that zone. Optimum focus is reached when C is at its maximum.
+ *
+ * The intensity and edge statistics are stored packed into a non-standard 16
+ * bit floating point format, where the 7 most significant bits represent the
+ * exponent and the 9 least significant bits the mantissa. This format can be
+ * unpacked with the following pseudocode::
+ *
+ * if (e == 0) {
+ * x = m;
+ * } else {
+ * x = 2^e-1 * (m + 2^9)
+ * }
+ *
+ * where
+ * e is the exponent value in range 0..127
+ * m is the mantissa value in range 0..511
+ */
+struct mali_c55_af_statistics {
+ __u16 intensity_stats;
+ __u16 edge_stats;
+} __attribute__((packed));
+
+/**
+ * struct mali_c55_stats_buffer - 3A statistics for the mali-c55 ISP
+ *
+ * @ae_1024bin_hist: 1024-bin frame-global pixel intensity histogram
+ * @iridix_1024bin_hist: Post-Iridix block 1024-bin histogram
+ * @ae_5bin_hists: 5-bin pixel intensity histograms for AEC
+ * @reserved1: Undefined buffer space
+ * @awb_ratios: Color balance ratios for Auto White Balance
+ * @reserved2: Undefined buffer space
+ * @af_statistics: Pixel intensity statistics for Auto Focus
+ * @reserved3: Undefined buffer space
+ *
+ * This struct describes the metering statistics space in the Mali-C55 ISP's
+ * hardware in its entirety. The space between each defined area is marked as
+ * "unknown" and may not be 0, but should not be used. The @ae_5bin_hists,
+ * @awb_ratios and @af_statistics members are arrays of statistics per-zone.
+ * The zones are arranged in the array in raster order starting from the top
+ * left corner of the image.
+ */
+
+struct mali_c55_stats_buffer {
+ struct mali_c55_ae_1024bin_hist ae_1024bin_hist;
+ struct mali_c55_ae_1024bin_hist iridix_1024bin_hist;
+ struct mali_c55_ae_5bin_hist ae_5bin_hists[MALI_C55_MAX_ZONES];
+ __u32 reserved1[14];
+ struct mali_c55_awb_average_ratios awb_ratios[MALI_C55_MAX_ZONES];
+ __u32 reserved2[14];
+ struct mali_c55_af_statistics af_statistics[MALI_C55_MAX_ZONES];
+ __u32 reserved3[15];
+} __attribute__((packed));
+
#endif /* __UAPI_MALI_C55_CONFIG_H */
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 14/19] media: platform: Add mali-c55 3a stats devnode
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (11 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 13/19] media: uapi: Add 3a stats buffer for mali-c55 Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 15/19] Documentation: mali-c55: Add Statistics documentation Daniel Scally
` (4 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add a new code file to govern the 3a statistics capture node.
On ISP_START, fill the stats buffer by reading out the metering space
in the ISP's memory. This is done for the non-active config just as
the dma transfer of the registers is. To acheive that, move the
checking of the current config outside of mali_c55_swap_next_config()
so we can use it for both functions.
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Don't use the media jobs framework, we'll add it separately
later.
- Update copyright year
- Use new video_device_pipeline_started() helper
Changes in v10:
- Account for v2 of media jobs framework
Changes in v9:
- Squashed the originally separate patch that filled the stats buffer on
the ISP start interrupt into this patch.
- A couple of differences accounting for the different way of operating
in inline and memory input mode
- Infrastructure for the media jobs framework.
- Include this video device in the synchronised start/stop across the
pipeline
Changes in v8:
- None
Changes in v7:
- Set width and height of MEDIA_BUS_FMT_METADATA_FIXED to 0.
- Added cpu fallback to stats in the event of no DMA channel acquired
- Terminate ISP streaming in the statistics vb2 stop streaming function
Changes in v6:
- Fixed mising includes
- Minor renames and formatting
- Reworked mali_c55_stats_metering_complete() so it could only return
buffers when both halves of the DMA read were done
- Terminate dma transfers on streamoff
Changes in v5:
- New patch
---
drivers/media/platform/arm/mali-c55/Makefile | 1 +
.../media/platform/arm/mali-c55/mali-c55-common.h | 29 ++
.../media/platform/arm/mali-c55/mali-c55-core.c | 40 +-
drivers/media/platform/arm/mali-c55/mali-c55-isp.c | 11 +
.../platform/arm/mali-c55/mali-c55-registers.h | 3 +
.../media/platform/arm/mali-c55/mali-c55-stats.c | 428 +++++++++++++++++++++
6 files changed, 505 insertions(+), 7 deletions(-)
diff --git a/drivers/media/platform/arm/mali-c55/Makefile b/drivers/media/platform/arm/mali-c55/Makefile
index 9178ac35e50efdc42289f5ed5e3ef5af55e8ac32..b5a22d414479dd1eb6435f474cb95aaa14879aa8 100644
--- a/drivers/media/platform/arm/mali-c55/Makefile
+++ b/drivers/media/platform/arm/mali-c55/Makefile
@@ -4,6 +4,7 @@ mali-c55-y := mali-c55-capture.o \
mali-c55-core.o \
mali-c55-isp.o \
mali-c55-resizer.o \
+ mali-c55-stats.o \
mali-c55-tpg.o
obj-$(CONFIG_VIDEO_MALI_C55) += mali-c55.o
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-common.h b/drivers/media/platform/arm/mali-c55/mali-c55-common.h
index 7970fabf0b6e1caa6461a8737097edd52d1262ce..3970ace9ee8143b65639db1111b120e6c5fec70f 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-common.h
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-common.h
@@ -52,6 +52,7 @@ enum mali_c55_isp_pads {
MALI_C55_ISP_PAD_SINK_VIDEO,
MALI_C55_ISP_PAD_SOURCE_VIDEO,
MALI_C55_ISP_PAD_SOURCE_BYPASS,
+ MALI_C55_ISP_PAD_SOURCE_STATS,
MALI_C55_ISP_NUM_PADS,
};
@@ -161,6 +162,29 @@ struct mali_c55_cap_dev {
} buffers;
};
+struct mali_c55_stats_buf {
+ struct vb2_v4l2_buffer vb;
+ unsigned int segments_remaining;
+ struct list_head queue;
+ bool failed;
+};
+
+struct mali_c55_stats {
+ struct mali_c55 *mali_c55;
+ struct video_device vdev;
+ struct dma_chan *channel;
+ struct vb2_queue queue;
+ struct media_pad pad;
+ /* Mutex to provide to vb2 */
+ struct mutex lock;
+
+ struct {
+ /* Spinlock to guard buffer queue */
+ spinlock_t lock;
+ struct list_head queue;
+ } buffers;
+};
+
enum mali_c55_config_spaces {
MALI_C55_CONFIG_PING,
MALI_C55_CONFIG_PONG,
@@ -205,6 +229,7 @@ struct mali_c55 {
struct mali_c55_isp isp;
struct mali_c55_resizer resizers[MALI_C55_NUM_RSZS];
struct mali_c55_cap_dev cap_devs[MALI_C55_NUM_CAP_DEVS];
+ struct mali_c55_stats stats;
struct mali_c55_context context;
u32 next_config;
@@ -233,6 +258,8 @@ int mali_c55_register_resizers(struct mali_c55 *mali_c55);
void mali_c55_unregister_resizers(struct mali_c55 *mali_c55);
int mali_c55_register_capture_devs(struct mali_c55 *mali_c55);
void mali_c55_unregister_capture_devs(struct mali_c55 *mali_c55);
+int mali_c55_register_stats(struct mali_c55 *mali_c55);
+void mali_c55_unregister_stats(struct mali_c55 *mali_c55);
struct mali_c55_context *mali_c55_get_active_context(struct mali_c55 *mali_c55);
void mali_c55_set_plane_done(struct mali_c55_cap_dev *cap_dev,
enum mali_c55_planes plane);
@@ -250,5 +277,7 @@ mali_c55_isp_get_mbus_config_by_shifted_code(u32 code);
const struct mali_c55_isp_format_info *
mali_c55_isp_get_mbus_config_by_index(u32 index);
bool mali_c55_pipeline_ready(struct media_pipeline *pipe);
+void mali_c55_stats_fill_buffer(struct mali_c55 *mali_c55,
+ enum mali_c55_config_spaces cfg_space);
#endif /* _MALI_C55_COMMON_H */
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-core.c b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
index 4c9e769bcd2fbae76a8f81d2acca8159b5cad656..4b71725f956648595400dc57d590383bcea989e8 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-core.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
@@ -377,6 +377,16 @@ static int mali_c55_create_links(struct mali_c55 *mali_c55)
}
}
+ ret = media_create_pad_link(&mali_c55->isp.sd.entity,
+ MALI_C55_ISP_PAD_SOURCE_STATS,
+ &mali_c55->stats.vdev.entity, 0,
+ MEDIA_LNK_FL_ENABLED);
+ if (ret) {
+ dev_err(mali_c55->dev,
+ "failed to link ISP and 3a stats node\n");
+ goto err_remove_links;
+ }
+
return 0;
err_remove_links:
@@ -391,19 +401,13 @@ static void mali_c55_unregister_entities(struct mali_c55 *mali_c55)
mali_c55_unregister_isp(mali_c55);
mali_c55_unregister_resizers(mali_c55);
mali_c55_unregister_capture_devs(mali_c55);
+ mali_c55_unregister_stats(mali_c55);
}
static void mali_c55_swap_next_config(struct mali_c55 *mali_c55)
{
struct mali_c55_context *ctx = mali_c55_get_active_context(mali_c55);
- u32 curr_config;
-
- curr_config = mali_c55_read(mali_c55, MALI_C55_REG_PING_PONG_READ);
- curr_config = (curr_config & MALI_C55_REG_PING_PONG_READ_MASK)
- >> (ffs(MALI_C55_REG_PING_PONG_READ_MASK) - 1);
- mali_c55->next_config = curr_config ^ 1;
-
mali_c55_config_write(ctx, mali_c55->next_config ?
MALI_C55_CONFIG_PING : MALI_C55_CONFIG_PONG,
false);
@@ -439,6 +443,10 @@ static int mali_c55_register_entities(struct mali_c55 *mali_c55)
if (ret)
goto err_unregister_entities;
+ ret = mali_c55_register_stats(mali_c55);
+ if (ret)
+ goto err_unregister_entities;
+
ret = mali_c55_create_links(mali_c55);
if (ret)
goto err_unregister_entities;
@@ -639,6 +647,7 @@ static irqreturn_t mali_c55_isr(int irq, void *context)
struct device *dev = context;
struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
unsigned long interrupt_status;
+ u32 curr_config;
unsigned int i;
interrupt_status = mali_c55_read(mali_c55,
@@ -660,6 +669,23 @@ static irqreturn_t mali_c55_isr(int irq, void *context)
if (mali_c55->capabilities & MALI_C55_GPS_DS_PIPE_FITTED)
mali_c55_set_next_buffer(&mali_c55->cap_devs[MALI_C55_CAP_DEV_DS]);
+ /*
+ * When the ISP starts a frame we have some work to do:
+ *
+ * 1. Copy over the config for the **next** frame
+ * 2. Read out the metering stats for the **last** frame
+ */
+
+ curr_config = mali_c55_read(mali_c55,
+ MALI_C55_REG_PING_PONG_READ);
+ curr_config &= MALI_C55_REG_PING_PONG_READ_MASK;
+ curr_config >>= ffs(MALI_C55_REG_PING_PONG_READ_MASK) - 1;
+ mali_c55->next_config = curr_config ^ 1;
+
+ mali_c55_stats_fill_buffer(mali_c55,
+ mali_c55->next_config ? MALI_C55_CONFIG_PING :
+ MALI_C55_CONFIG_PONG);
+
mali_c55_swap_next_config(mali_c55);
break;
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-isp.c b/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
index 1d310c77998b73d9a301538a8b66b73b5f5cc579..3095a9b9ceacebb688f17cd49de490385f417e12 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
@@ -5,6 +5,8 @@
* Copyright (C) 2025 Ideas on Board Oy
*/
+#include <linux/media/arm/mali-c55-config.h>
+
#include <linux/delay.h>
#include <linux/iopoll.h>
#include <linux/property.h>
@@ -512,6 +514,14 @@ static int mali_c55_isp_init_state(struct v4l2_subdev *sd,
in_crop->width = MALI_C55_DEFAULT_WIDTH;
in_crop->height = MALI_C55_DEFAULT_HEIGHT;
+ src_fmt = v4l2_subdev_state_get_format(state,
+ MALI_C55_ISP_PAD_SOURCE_STATS);
+
+ src_fmt->width = 0;
+ src_fmt->height = 0;
+ src_fmt->field = V4L2_FIELD_NONE;
+ src_fmt->code = MEDIA_BUS_FMT_METADATA_FIXED;
+
return 0;
}
@@ -610,6 +620,7 @@ int mali_c55_register_isp(struct mali_c55 *mali_c55)
MEDIA_PAD_FL_MUST_CONNECT;
isp->pads[MALI_C55_ISP_PAD_SOURCE_VIDEO].flags = MEDIA_PAD_FL_SOURCE;
isp->pads[MALI_C55_ISP_PAD_SOURCE_BYPASS].flags = MEDIA_PAD_FL_SOURCE;
+ isp->pads[MALI_C55_ISP_PAD_SOURCE_STATS].flags = MEDIA_PAD_FL_SOURCE;
ret = media_entity_pads_init(&sd->entity, MALI_C55_ISP_NUM_PADS,
isp->pads);
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-registers.h b/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
index 66b144c6fe906b56ea98302d2d8515db883411f0..8b04aff1990a962ddb7671d2f66f70c0638737a7 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
@@ -103,6 +103,9 @@ enum mali_c55_interrupts {
#define MALI_C55_VC_START(v) ((v) & 0xffff)
#define MALI_C55_VC_SIZE(v) (((v) & 0xffff) << 16)
+#define MALI_C55_REG_1024BIN_HIST 0x054a8
+#define MALI_C55_1024BIN_HIST_SIZE 4096
+
/* Ping/Pong Configuration Space */
#define MALI_C55_REG_BASE_ADDR 0x18e88
#define MALI_C55_REG_BYPASS_0 0x18eac
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-stats.c b/drivers/media/platform/arm/mali-c55/mali-c55-stats.c
new file mode 100644
index 0000000000000000000000000000000000000000..857f8734f78e6ed203d6fd47607617e2d57f42d7
--- /dev/null
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-stats.c
@@ -0,0 +1,428 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * ARM Mali-C55 ISP Driver - 3A Statistics capture device
+ *
+ * Copyright (C) 2025 Ideas on Board Oy
+ */
+
+#include <linux/container_of.h>
+#include <linux/dev_printk.h>
+#include <linux/dmaengine.h>
+#include <linux/list.h>
+#include <linux/media/arm/mali-c55-config.h>
+#include <linux/mutex.h>
+#include <linux/pm_runtime.h>
+#include <linux/spinlock.h>
+#include <linux/string.h>
+
+#include <media/media-entity.h>
+#include <media/v4l2-dev.h>
+#include <media/v4l2-event.h>
+#include <media/v4l2-fh.h>
+#include <media/v4l2-ioctl.h>
+#include <media/videobuf2-core.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "mali-c55-common.h"
+#include "mali-c55-registers.h"
+
+static const unsigned int metering_space_addrs[] = {
+ [MALI_C55_CONFIG_PING] = 0x095ac,
+ [MALI_C55_CONFIG_PONG] = 0x2156c,
+};
+
+static int mali_c55_stats_enum_fmt_meta_cap(struct file *file, void *fh,
+ struct v4l2_fmtdesc *f)
+{
+ if (f->index)
+ return -EINVAL;
+
+ f->pixelformat = V4L2_META_FMT_MALI_C55_STATS;
+
+ return 0;
+}
+
+static int mali_c55_stats_g_fmt_meta_cap(struct file *file, void *fh,
+ struct v4l2_format *f)
+{
+ static const struct v4l2_meta_format mfmt = {
+ .dataformat = V4L2_META_FMT_MALI_C55_STATS,
+ .buffersize = sizeof(struct mali_c55_stats_buffer)
+ };
+
+ f->fmt.meta = mfmt;
+
+ return 0;
+}
+
+static int mali_c55_stats_querycap(struct file *file,
+ void *priv, struct v4l2_capability *cap)
+{
+ strscpy(cap->driver, MALI_C55_DRIVER_NAME, sizeof(cap->driver));
+ strscpy(cap->card, "ARM Mali-C55 ISP", sizeof(cap->card));
+
+ return 0;
+}
+
+static const struct v4l2_ioctl_ops mali_c55_stats_v4l2_ioctl_ops = {
+ .vidioc_reqbufs = vb2_ioctl_reqbufs,
+ .vidioc_querybuf = vb2_ioctl_querybuf,
+ .vidioc_create_bufs = vb2_ioctl_create_bufs,
+ .vidioc_qbuf = vb2_ioctl_qbuf,
+ .vidioc_expbuf = vb2_ioctl_expbuf,
+ .vidioc_dqbuf = vb2_ioctl_dqbuf,
+ .vidioc_prepare_buf = vb2_ioctl_prepare_buf,
+ .vidioc_streamon = vb2_ioctl_streamon,
+ .vidioc_streamoff = vb2_ioctl_streamoff,
+ .vidioc_enum_fmt_meta_cap = mali_c55_stats_enum_fmt_meta_cap,
+ .vidioc_g_fmt_meta_cap = mali_c55_stats_g_fmt_meta_cap,
+ .vidioc_s_fmt_meta_cap = mali_c55_stats_g_fmt_meta_cap,
+ .vidioc_try_fmt_meta_cap = mali_c55_stats_g_fmt_meta_cap,
+ .vidioc_querycap = mali_c55_stats_querycap,
+ .vidioc_subscribe_event = v4l2_ctrl_subscribe_event,
+ .vidioc_unsubscribe_event = v4l2_event_unsubscribe,
+};
+
+static const struct v4l2_file_operations mali_c55_stats_v4l2_fops = {
+ .owner = THIS_MODULE,
+ .unlocked_ioctl = video_ioctl2,
+ .open = v4l2_fh_open,
+ .release = vb2_fop_release,
+ .poll = vb2_fop_poll,
+ .mmap = vb2_fop_mmap,
+};
+
+static int
+mali_c55_stats_queue_setup(struct vb2_queue *q, unsigned int *num_buffers,
+ unsigned int *num_planes, unsigned int sizes[],
+ struct device *alloc_devs[])
+{
+ struct mali_c55_stats *stats = vb2_get_drv_priv(q);
+
+ if (*num_planes && *num_planes > 1)
+ return -EINVAL;
+
+ if (sizes[0] && sizes[0] < sizeof(struct mali_c55_stats_buffer))
+ return -EINVAL;
+
+ *num_planes = 1;
+
+ if (!sizes[0])
+ sizes[0] = sizeof(struct mali_c55_stats_buffer);
+
+ if (stats->channel)
+ alloc_devs[0] = stats->channel->device->dev;
+
+ return 0;
+}
+
+static void mali_c55_stats_buf_queue(struct vb2_buffer *vb)
+{
+ struct mali_c55_stats *stats = vb2_get_drv_priv(vb->vb2_queue);
+ struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+ struct mali_c55_stats_buf *buf = container_of(vbuf,
+ struct mali_c55_stats_buf, vb);
+
+ vb2_set_plane_payload(vb, 0, sizeof(struct mali_c55_stats_buffer));
+ buf->segments_remaining = 2;
+ buf->failed = false;
+
+ spin_lock(&stats->buffers.lock);
+ list_add_tail(&buf->queue, &stats->buffers.queue);
+ spin_unlock(&stats->buffers.lock);
+}
+
+static void mali_c55_stats_return_buffers(struct mali_c55_stats *stats,
+ enum vb2_buffer_state state)
+{
+ struct mali_c55_stats_buf *buf, *tmp;
+
+ guard(spinlock)(&stats->buffers.lock);
+
+ list_for_each_entry_safe(buf, tmp, &stats->buffers.queue, queue) {
+ list_del(&buf->queue);
+ vb2_buffer_done(&buf->vb.vb2_buf, state);
+ }
+}
+
+static int mali_c55_stats_start_streaming(struct vb2_queue *q,
+ unsigned int count)
+{
+ struct mali_c55_stats *stats = vb2_get_drv_priv(q);
+ struct mali_c55 *mali_c55 = stats->mali_c55;
+ int ret;
+
+ ret = pm_runtime_resume_and_get(mali_c55->dev);
+ if (ret)
+ goto err_return_buffers;
+
+ ret = video_device_pipeline_alloc_start(&stats->vdev);
+ if (ret)
+ goto err_pm_put;
+
+ ret = video_device_pipeline_started(&stats->vdev);
+ if (ret < 0)
+ goto err_stop_pipeline;
+
+ return 0;
+
+err_stop_pipeline:
+ video_device_pipeline_stop(&stats->vdev);
+err_pm_put:
+ pm_runtime_put(mali_c55->dev);
+err_return_buffers:
+ mali_c55_stats_return_buffers(stats, VB2_BUF_STATE_QUEUED);
+
+ return ret;
+}
+
+static void mali_c55_stats_stop_streaming(struct vb2_queue *q)
+{
+ struct mali_c55_stats *stats = vb2_get_drv_priv(q);
+ struct media_pipeline *pipe;
+
+ if (stats->channel)
+ dmaengine_terminate_sync(stats->channel);
+
+ pipe = video_device_pipeline(&stats->vdev);
+ if (mali_c55_pipeline_ready(pipe))
+ media_pipeline_stopped(pipe);
+
+ video_device_pipeline_stop(&stats->vdev);
+ mali_c55_stats_return_buffers(stats, VB2_BUF_STATE_ERROR);
+}
+
+static const struct vb2_ops mali_c55_stats_vb2_ops = {
+ .queue_setup = mali_c55_stats_queue_setup,
+ .buf_queue = mali_c55_stats_buf_queue,
+ .wait_prepare = vb2_ops_wait_prepare,
+ .wait_finish = vb2_ops_wait_finish,
+ .start_streaming = mali_c55_stats_start_streaming,
+ .stop_streaming = mali_c55_stats_stop_streaming,
+};
+
+static void
+mali_c55_stats_metering_complete(void *param,
+ const struct dmaengine_result *result)
+{
+ struct mali_c55_stats_buf *buf = param;
+
+ if (result->result != DMA_TRANS_NOERROR)
+ buf->failed = true;
+
+ if (!--buf->segments_remaining)
+ vb2_buffer_done(&buf->vb.vb2_buf, buf->failed ?
+ VB2_BUF_STATE_ERROR : VB2_BUF_STATE_DONE);
+}
+
+static int mali_c55_stats_dma_xfer(struct mali_c55_stats *stats, dma_addr_t src,
+ dma_addr_t dst,
+ struct mali_c55_stats_buf *buf,
+ size_t length)
+{
+ struct dma_async_tx_descriptor *tx;
+ dma_cookie_t cookie;
+
+ tx = dmaengine_prep_dma_memcpy(stats->channel, dst, src, length, 0);
+ if (!tx) {
+ dev_err(stats->mali_c55->dev, "failed to prep stats DMA\n");
+ return -EIO;
+ }
+
+ tx->callback_result = mali_c55_stats_metering_complete;
+ tx->callback_param = buf;
+
+ cookie = dmaengine_submit(tx);
+ if (dma_submit_error(cookie)) {
+ dev_err(stats->mali_c55->dev, "failed to submit stats DMA\n");
+ return -EIO;
+ }
+
+ dma_async_issue_pending(stats->channel);
+ return 0;
+}
+
+static int mali_c55_stats_dma_read(struct mali_c55_stats *stats,
+ struct mali_c55_stats_buf *buf,
+ enum mali_c55_config_spaces cfg_space)
+{
+ struct mali_c55 *mali_c55 = stats->mali_c55;
+ struct mali_c55_context *ctx = mali_c55_get_active_context(mali_c55);
+ dma_addr_t src, dst;
+ size_t length;
+ int ret;
+
+ /*
+ * There are in fact two noncontiguous sections of the ISP's
+ * memory space that hold statistics for 3a algorithms to use: A
+ * section in each config space and a global section holding
+ * histograms which is double buffered and so holds data for the
+ * last frame. We need to read both.
+ */
+ src = ctx->base + MALI_C55_REG_1024BIN_HIST;
+ dst = vb2_dma_contig_plane_dma_addr(&buf->vb.vb2_buf, 0);
+
+ ret = mali_c55_stats_dma_xfer(stats, src, dst, buf,
+ MALI_C55_1024BIN_HIST_SIZE);
+ if (ret)
+ return ret;
+
+ src = ctx->base + metering_space_addrs[cfg_space];
+ dst += MALI_C55_1024BIN_HIST_SIZE;
+
+ length = sizeof(struct mali_c55_stats_buffer) - MALI_C55_1024BIN_HIST_SIZE;
+ ret = mali_c55_stats_dma_xfer(stats, src, dst, buf, length);
+ if (ret) {
+ dmaengine_terminate_sync(stats->channel);
+ return ret;
+ }
+
+ return 0;
+}
+
+static void mali_c55_stats_cpu_read(struct mali_c55_stats *stats,
+ struct mali_c55_stats_buf *buf,
+ enum mali_c55_config_spaces cfg_space)
+{
+ struct mali_c55 *mali_c55 = stats->mali_c55;
+ size_t length;
+ void *src;
+ void *dst;
+
+ src = mali_c55->base + MALI_C55_REG_1024BIN_HIST;
+ dst = vb2_plane_vaddr(&buf->vb.vb2_buf, 0);
+ memcpy_fromio(dst, src, MALI_C55_1024BIN_HIST_SIZE);
+
+ src = mali_c55->base + metering_space_addrs[cfg_space];
+ dst += MALI_C55_1024BIN_HIST_SIZE;
+ length = sizeof(struct mali_c55_stats_buffer) - MALI_C55_1024BIN_HIST_SIZE;
+ memcpy_fromio(dst, src, length);
+}
+
+void mali_c55_stats_fill_buffer(struct mali_c55 *mali_c55,
+ enum mali_c55_config_spaces cfg_space)
+{
+ struct mali_c55_stats *stats = &mali_c55->stats;
+ struct mali_c55_stats_buf *buf = NULL;
+ int ret;
+
+ spin_lock(&stats->buffers.lock);
+ if (!list_empty(&stats->buffers.queue)) {
+ buf = list_first_entry(&stats->buffers.queue,
+ struct mali_c55_stats_buf, queue);
+ list_del(&buf->queue);
+ }
+ spin_unlock(&stats->buffers.lock);
+
+ if (!buf)
+ return;
+
+ buf->vb.sequence = mali_c55->isp.frame_sequence;
+ buf->vb.vb2_buf.timestamp = ktime_get_boottime_ns();
+
+ if (stats->channel) {
+ ret = mali_c55_stats_dma_read(stats, buf, cfg_space);
+ if (ret)
+ vb2_buffer_done(&buf->vb.vb2_buf, VB2_BUF_STATE_ERROR);
+ } else {
+ mali_c55_stats_cpu_read(stats, buf, cfg_space);
+ vb2_buffer_done(&buf->vb.vb2_buf, VB2_BUF_STATE_DONE);
+ }
+}
+
+void mali_c55_unregister_stats(struct mali_c55 *mali_c55)
+{
+ struct mali_c55_stats *stats = &mali_c55->stats;
+
+ if (!video_is_registered(&stats->vdev))
+ return;
+
+ vb2_video_unregister_device(&stats->vdev);
+ media_entity_cleanup(&stats->vdev.entity);
+
+ if (stats->channel)
+ dma_release_channel(stats->channel);
+
+ mutex_destroy(&stats->lock);
+}
+
+int mali_c55_register_stats(struct mali_c55 *mali_c55)
+{
+ struct mali_c55_stats *stats = &mali_c55->stats;
+ struct video_device *vdev = &stats->vdev;
+ struct vb2_queue *vb2q = &stats->queue;
+ dma_cap_mask_t mask;
+ int ret;
+
+ mutex_init(&stats->lock);
+ INIT_LIST_HEAD(&stats->buffers.queue);
+ spin_lock_init(&stats->buffers.lock);
+
+ dma_cap_zero(mask);
+ dma_cap_set(DMA_MEMCPY, mask);
+
+ /*
+ * No failure here, because we will just fallback on memcpy if there is
+ * no usable DMA channel on the system.
+ */
+ stats->channel = dma_request_channel(mask, 0, NULL);
+ if (!stats->channel)
+ dev_dbg(mali_c55->dev,
+ "No DMA channel for stats, falling back to memcpy\n");
+
+ stats->pad.flags = MEDIA_PAD_FL_SINK;
+ ret = media_entity_pads_init(&stats->vdev.entity, 1, &stats->pad);
+ if (ret)
+ goto err_release_dma_channel;
+
+ vb2q->type = V4L2_BUF_TYPE_META_CAPTURE;
+ vb2q->io_modes = VB2_MMAP | VB2_DMABUF;
+ vb2q->drv_priv = stats;
+ vb2q->mem_ops = &vb2_dma_contig_memops;
+ vb2q->ops = &mali_c55_stats_vb2_ops;
+ vb2q->buf_struct_size = sizeof(struct mali_c55_stats_buf);
+ vb2q->min_queued_buffers = 1;
+ vb2q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
+ vb2q->lock = &stats->lock;
+ vb2q->dev = stats->channel ? stats->channel->device->dev : mali_c55->dev;
+
+ ret = vb2_queue_init(vb2q);
+ if (ret) {
+ dev_err(mali_c55->dev, "stats vb2 queue init failed\n");
+ goto err_cleanup_entity;
+ }
+
+ strscpy(stats->vdev.name, "mali-c55 3a stats", sizeof(stats->vdev.name));
+ vdev->release = video_device_release_empty;
+ vdev->fops = &mali_c55_stats_v4l2_fops;
+ vdev->ioctl_ops = &mali_c55_stats_v4l2_ioctl_ops;
+ vdev->lock = &stats->lock;
+ vdev->v4l2_dev = &mali_c55->v4l2_dev;
+ vdev->queue = &stats->queue;
+ vdev->device_caps = V4L2_CAP_META_CAPTURE | V4L2_CAP_STREAMING;
+ vdev->vfl_dir = VFL_DIR_RX;
+ video_set_drvdata(vdev, stats);
+
+ ret = video_register_device(vdev, VFL_TYPE_VIDEO, -1);
+ if (ret) {
+ dev_err(mali_c55->dev,
+ "failed to register stats video device\n");
+ goto err_release_vb2q;
+ }
+
+ stats->mali_c55 = mali_c55;
+
+ return 0;
+
+err_release_vb2q:
+ vb2_queue_release(vb2q);
+err_cleanup_entity:
+ media_entity_cleanup(&stats->vdev.entity);
+err_release_dma_channel:
+ if (stats->channel)
+ dma_release_channel(stats->channel);
+
+ mutex_destroy(&stats->lock);
+
+ return ret;
+}
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 15/19] Documentation: mali-c55: Add Statistics documentation
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (12 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 14/19] media: platform: Add mali-c55 3a stats devnode Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 16/19] media: mali-c55: Add image formats for Mali-C55 parameters buffer Daniel Scally
` (3 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add documentation explaining the ability to capture statistics from
the mali-c55 driver's new V4L2 device, as well as the various tap
points from which those statistics can be drawn in the ISP's
processing flow. Additionally add a page detailing the new V4L2
meta format for the mali-c55 statistics.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- New patch
---
Documentation/admin-guide/media/mali-c55.rst | 60 +++++++++++++++++++++-
.../userspace-api/media/v4l/meta-formats.rst | 1 +
.../media/v4l/metafmt-arm-mali-c55.rst | 29 +++++++++++
MAINTAINERS | 1 +
4 files changed, 90 insertions(+), 1 deletion(-)
diff --git a/Documentation/admin-guide/media/mali-c55.rst b/Documentation/admin-guide/media/mali-c55.rst
index 72cdded507b3a7918e28d9bd9a1c6f72ddf9e664..7eaeac63ddf713730fac266d81866539310cc7e2 100644
--- a/Documentation/admin-guide/media/mali-c55.rst
+++ b/Documentation/admin-guide/media/mali-c55.rst
@@ -67,10 +67,11 @@ The driver has 4 V4L2 subdevices:
- `mali_c55 resizer fr`: The Full-Resolution pipe resizer
- `mali_c55 resizer ds`: The Downscale pipe resizer
-The driver has 2 V4L2 video devices:
+The driver has 3 V4L2 video devices:
- `mali-c55 fr`: The full-resolution pipe's capture device
- `mali-c55 ds`: The downscale pipe's capture device
+- `mali-c55 3a stats`: The 3A statistics capture device
Frame sequences are synchronised across to two capture devices, meaning if one
pipe is started later than the other the sequence numbers returned in its
@@ -333,6 +334,63 @@ configured, followed by formats in the appropriate places:
# Set format on the video device and stream
yavta -f RGB565 -s 1920x1080 -c10 /dev/video0
+.. _mali-c55-3a-stats:
+
+Capturing ISP Statistics
+========================
+
+The ISP is capable of producing statistics for consumption by image processing
+algorithms running in userspace. These statistics can be captured by queueing
+buffers to the `mali-c55 3a stats` V4L2 Device whilst the ISP is streaming. Only
+the :ref:`V4L2_META_FMT_MALI_C55_STATS <v4l2-meta-fmt-mali-c55-3a-stats>`
+format is supported, so no format-setting need be done:
+
+.. code-block:: none
+
+ # We assume the media graph has been configured to support RGB565 capture
+ # from the mali-c55 fr V4L2 Device, which is at /dev/video0. The statistics
+ # V4L2 device is at /dev/video3
+
+ yavta -f RGB565 -s 1920x1080 -c32 /dev/video0 && \
+ yavta -c10 -F /dev/video3
+
+The layout of the buffer is described by :c:type:`mali_c55_stats_buffer`,
+but broadly statistics are generated to support three image processing
+algorithms; AEXP (Auto-Exposure), AWB (Auto-White Balance) and AF (Auto-Focus).
+These stats can be drawn from various places in the Mali C55 ISP pipeline, known
+as "tap points". This high-level block diagram is intended to explain where in
+the processing flow the statistics can be drawn from::
+
+ +--> AEXP-2 +----> AEXP-1 +--> AF-0
+ | +----> AF-1 |
+ | | |
+ +---------+ | +--------------+ | +--------------+ |
+ | Input +-+-->+ Digital Gain +---+-->+ Black Level +---+---+
+ +---------+ +--------------+ +--------------+ |
+ +-----------------------------------------------------------------+
+ |
+ | +--------------+ +---------+ +----------------+
+ +-->| Sinter Noise +-+ White +--+--->| Lens Shading +--+---------------+
+ | Reduction | | Balance | | | | | |
+ +--------------+ +---------+ | +----------------+ | |
+ +---> AEXP-0 (A) +--> AEXP-0 (B) |
+ +--------------------------------------------------------------------------+
+ |
+ | +----------------+ +--------------+ +----------------+
+ +-->| Tone mapping +-+--->| Demosaicing +->+ Purple Fringe +-+-----------+
+ | | | +--------------+ | Correction | | |
+ +----------------+ +-> AEXP-IRIDIX +----------------+ +---> AWB-0 |
+ +----------------------------------------------------------------------------+
+ | +-------------+ +-------------+
+ +------------------->| Colour +---+--->| Output |
+ | Correction | | | Pipelines |
+ +-------------+ | +-------------+
+ +--> AWB-1
+
+At present all statistics are drawn from the 0th tap point for each algorithm;
+I.E. AEXP statistics from AEXP-0 (A), AWB statistics from AWB-0 and AF
+statistics from AF-0. In the future this will be configurable.
+
References
==========
.. [1] https://git.linuxtv.org/v4l-utils.git/
diff --git a/Documentation/userspace-api/media/v4l/meta-formats.rst b/Documentation/userspace-api/media/v4l/meta-formats.rst
index 58eb3c9c962bee008eee27d9c16678213c47baa9..4d4b897b03d951ab4f9f7e9ba37f72d826208e1a 100644
--- a/Documentation/userspace-api/media/v4l/meta-formats.rst
+++ b/Documentation/userspace-api/media/v4l/meta-formats.rst
@@ -13,6 +13,7 @@ These formats are used for the :ref:`metadata` interface only.
:maxdepth: 1
extensible-parameters
+ metafmt-arm-mali-c55
metafmt-c3-isp
metafmt-d4xx
metafmt-generic
diff --git a/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst b/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
new file mode 100644
index 0000000000000000000000000000000000000000..186e0deb9ece70ab1d2b22bb5ccb69196264a793
--- /dev/null
+++ b/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
@@ -0,0 +1,29 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+.. _v4l2-meta-fmt-mali-c55-3a-stats:
+
+*************************************
+V4L2_META_FMT_MALI_C55_STATS ('C55S')
+*************************************
+
+3A Statistics
+=============
+
+The ISP device collects different statistics over an input bayer frame. Those
+statistics can be obtained by userspace from the
+:ref:`mali-c55 3a stats <mali-c55-3a-stats>` metadata capture video node, using
+the :c:type:`v4l2_meta_format` interface. The buffer contains a single instance
+of the C structure :c:type:`mali_c55_stats_buffer` defined in
+``mali-c55-config.h``, so the structure can be obtained from the buffer by:
+
+.. code-block:: C
+
+ struct mali_c55_stats_buffer *stats =
+ (struct mali_c55_stats_buffer *)buf;
+
+For details of the statistics see :c:type:`mali_c55_stats_buffer`.
+
+Arm Mali-C55 uAPI data types
+============================
+
+.. kernel-doc:: include/uapi/linux/media/arm/mali-c55-config.h
diff --git a/MAINTAINERS b/MAINTAINERS
index 87255112240dbce86bca8d9a6683015ce12a3bce..3a53565aa34897e5df13f0420908598add5b28fb 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2038,6 +2038,7 @@ F: Documentation/admin-guide/media/mali-c55-graph.dot
F: Documentation/admin-guide/media/mali-c55.rst
F: Documentation/devicetree/bindings/media/arm,mali-c55.yaml
F: Documentation/userspace-api/media/drivers/mali-c55.rst
+F: Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
F: drivers/media/platform/arm/mali-c55/
F: include/uapi/linux/media/arm/mali-c55-config.h
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 16/19] media: mali-c55: Add image formats for Mali-C55 parameters buffer
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (13 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 15/19] Documentation: mali-c55: Add Statistics documentation Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 17/19] media: uapi: Add parameters structs to mali-c55-config.h Daniel Scally
` (2 subsequent siblings)
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
From: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Add a new V4L2 meta format code for the Mali-C55 parameters.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- New patch
---
drivers/media/v4l2-core/v4l2-ioctl.c | 1 +
include/uapi/linux/videodev2.h | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
index d2e7d6b115b53980fef959cbd9c8e758819d20e9..bd65cc771d682dd3856a7a0a38dc63a268a01e7b 100644
--- a/drivers/media/v4l2-core/v4l2-ioctl.c
+++ b/drivers/media/v4l2-core/v4l2-ioctl.c
@@ -1470,6 +1470,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
case V4L2_META_FMT_RK_ISP1_EXT_PARAMS: descr = "Rockchip ISP1 Ext 3A Params"; break;
case V4L2_META_FMT_C3ISP_PARAMS: descr = "Amlogic C3 ISP Parameters"; break;
case V4L2_META_FMT_C3ISP_STATS: descr = "Amlogic C3 ISP Statistics"; break;
+ case V4L2_META_FMT_MALI_C55_PARAMS: descr = "ARM Mali-C55 ISP Parameters"; break;
case V4L2_META_FMT_MALI_C55_STATS: descr = "ARM Mali-C55 ISP 3A Statistics"; break;
case V4L2_PIX_FMT_NV12_8L128: descr = "NV12 (8x128 Linear)"; break;
case V4L2_PIX_FMT_NV12M_8L128: descr = "NV12M (8x128 Linear)"; break;
diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
index 752d4d6ae5ca4ca924dd8a58aed6c4af3304879e..9da42c528a04398e05c06b5e9da4ef22a0b2c8d0 100644
--- a/include/uapi/linux/videodev2.h
+++ b/include/uapi/linux/videodev2.h
@@ -884,6 +884,7 @@ struct v4l2_pix_format {
#define V4L2_META_FMT_RPI_FE_STATS v4l2_fourcc('R', 'P', 'F', 'S') /* PiSP FE stats */
/* Vendor specific - used for Arm Mali-C55 ISP */
+#define V4L2_META_FMT_MALI_C55_PARAMS v4l2_fourcc('C', '5', '5', 'P') /* ARM Mali-C55 Parameters */
#define V4L2_META_FMT_MALI_C55_STATS v4l2_fourcc('C', '5', '5', 'S') /* ARM Mali-C55 3A Statistics */
#ifdef __KERNEL__
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 17/19] media: uapi: Add parameters structs to mali-c55-config.h
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (14 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 16/19] media: mali-c55: Add image formats for Mali-C55 parameters buffer Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-14 15:06 ` [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node Daniel Scally
2025-07-14 15:06 ` [PATCH v11 19/19] Documentation: mali-c55: Document the mali-c55 parameter setting Daniel Scally
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add structures describing the ISP parameters to mali-c55-config.h
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Used the new v4l2-extensible-params structs instead of
defining driver specific ones
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- Switched to using a @flags bitmask field instead of @enabled bool
field in the block header struct.
- Ditched the block type sentinel
- Removed the union from this header
Changes in v6:
- Flagged which struct goes with which enum value from
enum mali_c55_param_block_type
- Used only types with well defined sizes in the structs
- Expanded the documentation for skipping in the AEXP histogram config
- Aligned the header struct to 64 bits
- Added a new union type to hold pointers to the parameter structs
Changes in v5:
- New patch
---
include/uapi/linux/media/arm/mali-c55-config.h | 600 +++++++++++++++++++++++++
1 file changed, 600 insertions(+)
diff --git a/include/uapi/linux/media/arm/mali-c55-config.h b/include/uapi/linux/media/arm/mali-c55-config.h
index 8d31764140a0ef6aa8b0c771a4ed7565c9c96670..5afcf108e230cd48fec9305797b0be98d8b6e631 100644
--- a/include/uapi/linux/media/arm/mali-c55-config.h
+++ b/include/uapi/linux/media/arm/mali-c55-config.h
@@ -11,6 +11,9 @@
#include <linux/types.h>
#include <linux/v4l2-controls.h>
+#define _UAPI_V4L2_EXTENSIBLE_PARAMS_GUARD_
+#include <linux/media/v4l2-extensible-params.h>
+
#define V4L2_CID_MALI_C55_CAPABILITIES (V4L2_CID_USER_MALI_C55_BASE + 0x0)
#define MALI_C55_GPS_PONG_FITTED BIT(0)
#define MALI_C55_GPS_WDR_FITTED BIT(1)
@@ -193,4 +196,601 @@ struct mali_c55_stats_buffer {
__u32 reserved3[15];
} __attribute__((packed));
+/**
+ * enum mali_c55_param_buffer_version - Mali-C55 parameters block versioning
+ *
+ * @MALI_C55_PARAM_BUFFER_V1: First version of Mali-C55 parameters block
+ */
+enum mali_c55_param_buffer_version {
+ MALI_C55_PARAM_BUFFER_V1,
+};
+
+/**
+ * enum mali_c55_param_block_type - Enumeration of Mali-C55 parameter blocks
+ *
+ * This enumeration defines the types of Mali-C55 parameters block. Each block
+ * configures a specific processing block of the Mali-C55 ISP. The block
+ * type allows the driver to correctly interpret the parameters block data.
+ *
+ * It is the responsibility of userspace to correctly set the type of each
+ * parameters block.
+ *
+ * @MALI_C55_PARAM_BLOCK_SENSOR_OFFS: Sensor pre-shading black level offset
+ * @MALI_C55_PARAM_BLOCK_AEXP_HIST: Auto-exposure 1024-bin histogram
+ * configuration
+ * @MALI_C55_PARAM_BLOCK_AEXP_IHIST: Post-Iridix auto-exposure 1024-bin
+ * histogram configuration
+ * @MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS: Auto-exposure 1024-bin histogram
+ * weighting
+ * @MALI_C55_PARAM_BLOCK_AEXP_IHIST_WEIGHTS: Post-Iridix auto-exposure 1024-bin
+ * histogram weighting
+ * @MALI_C55_PARAM_BLOCK_DIGITAL_GAIN: Digital gain
+ * @MALI_C55_PARAM_BLOCK_AWB_GAINS: Auto-white balance gains
+ * @MALI_C55_PARAM_BLOCK_AWB_CONFIG: Auto-white balance statistics config
+ * @MALI_C55_PARAM_BLOCK_AWB_GAINS_AEXP: Auto-white balance gains for AEXP-0 tap
+ * @MALI_C55_PARAM_MESH_SHADING_CONFIG : Mesh shading tables configuration
+ * @MALI_C55_PARAM_MESH_SHADING_SELECTION: Mesh shading table selection
+ */
+enum mali_c55_param_block_type {
+ MALI_C55_PARAM_BLOCK_SENSOR_OFFS,
+ MALI_C55_PARAM_BLOCK_AEXP_HIST,
+ MALI_C55_PARAM_BLOCK_AEXP_IHIST,
+ MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS,
+ MALI_C55_PARAM_BLOCK_AEXP_IHIST_WEIGHTS,
+ MALI_C55_PARAM_BLOCK_DIGITAL_GAIN,
+ MALI_C55_PARAM_BLOCK_AWB_GAINS,
+ MALI_C55_PARAM_BLOCK_AWB_CONFIG,
+ MALI_C55_PARAM_BLOCK_AWB_GAINS_AEXP,
+ MALI_C55_PARAM_MESH_SHADING_CONFIG,
+ MALI_C55_PARAM_MESH_SHADING_SELECTION,
+};
+
+/**
+ * struct mali_c55_params_sensor_off_preshading - offset subtraction for each
+ * color channel
+ *
+ * Provides removal of the sensor black level from the sensor data. Separate
+ * offsets are provided for each of the four Bayer component color channels
+ * which are defaulted to R, Gr, Gb, B.
+ *
+ * header.type should be set to MALI_C55_PARAM_BLOCK_SENSOR_OFFS from
+ * :c:type:`mali_c55_param_block_type` for this block.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @chan00: Offset for color channel 00 (default: R)
+ * @chan01: Offset for color channel 01 (default: Gr)
+ * @chan10: Offset for color channel 10 (default: Gb)
+ * @chan11: Offset for color channel 11 (default: B)
+ */
+struct mali_c55_params_sensor_off_preshading {
+ struct v4l2_params_block_header header;
+ __u32 chan00;
+ __u32 chan01;
+ __u32 chan10;
+ __u32 chan11;
+};
+
+/**
+ * enum mali_c55_aexp_hist_tap_points - Tap points for the AEXP histogram
+ * @MALI_C55_AEXP_HIST_TAP_WB: After static white balance
+ * @MALI_C55_AEXP_HIST_TAP_FS: After WDR Frame Stitch
+ * @MALI_C55_AEXP_HIST_TAP_TPG: After the test pattern generator
+ */
+enum mali_c55_aexp_hist_tap_points {
+ MALI_C55_AEXP_HIST_TAP_WB = 0,
+ MALI_C55_AEXP_HIST_TAP_FS,
+ MALI_C55_AEXP_HIST_TAP_TPG,
+};
+
+/**
+ * enum mali_c55_aexp_skip_x - Horizontal pixel skipping
+ * @MALI_C55_AEXP_SKIP_X_EVERY_2ND: Collect every 2nd pixel horizontally
+ * @MALI_C55_AEXP_SKIP_X_EVERY_3RD: Collect every 3rd pixel horizontally
+ * @MALI_C55_AEXP_SKIP_X_EVERY_4TH: Collect every 4th pixel horizontally
+ * @MALI_C55_AEXP_SKIP_X_EVERY_5TH: Collect every 5th pixel horizontally
+ * @MALI_C55_AEXP_SKIP_X_EVERY_8TH: Collect every 8th pixel horizontally
+ * @MALI_C55_AEXP_SKIP_X_EVERY_9TH: Collect every 9th pixel horizontally
+ */
+enum mali_c55_aexp_skip_x {
+ MALI_C55_AEXP_SKIP_X_EVERY_2ND,
+ MALI_C55_AEXP_SKIP_X_EVERY_3RD,
+ MALI_C55_AEXP_SKIP_X_EVERY_4TH,
+ MALI_C55_AEXP_SKIP_X_EVERY_5TH,
+ MALI_C55_AEXP_SKIP_X_EVERY_8TH,
+ MALI_C55_AEXP_SKIP_X_EVERY_9TH
+};
+
+/**
+ * enum mali_c55_aexp_skip_y - Vertical pixel skipping
+ * @MALI_C55_AEXP_SKIP_Y_ALL: Collect every single pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_2ND: Collect every 2nd pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_3RD: Collect every 3rd pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_4TH: Collect every 4th pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_5TH: Collect every 5th pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_8TH: Collect every 8th pixel vertically
+ * @MALI_C55_AEXP_SKIP_Y_EVERY_9TH: Collect every 9th pixel vertically
+ */
+enum mali_c55_aexp_skip_y {
+ MALI_C55_AEXP_SKIP_Y_ALL,
+ MALI_C55_AEXP_SKIP_Y_EVERY_2ND,
+ MALI_C55_AEXP_SKIP_Y_EVERY_3RD,
+ MALI_C55_AEXP_SKIP_Y_EVERY_4TH,
+ MALI_C55_AEXP_SKIP_Y_EVERY_5TH,
+ MALI_C55_AEXP_SKIP_Y_EVERY_8TH,
+ MALI_C55_AEXP_SKIP_Y_EVERY_9TH
+};
+
+/**
+ * enum mali_c55_aexp_row_column_offset - Start from the first or second row or
+ * column
+ * @MALI_C55_AEXP_FIRST_ROW_OR_COL: Start from the first row / column
+ * @MALI_C55_AEXP_SECOND_ROW_OR_COL: Start from the second row / column
+ */
+enum mali_c55_aexp_row_column_offset {
+ MALI_C55_AEXP_FIRST_ROW_OR_COL = 1,
+ MALI_C55_AEXP_SECOND_ROW_OR_COL = 2,
+};
+
+/**
+ * enum mali_c55_aexp_hist_plane_mode - Mode for the AEXP Histograms
+ * @MALI_C55_AEXP_HIST_COMBINED: All color planes in one 1024-bin histogram
+ * @MALI_C55_AEXP_HIST_SEPARATE: Each color plane in one 256-bin histogram with a bin width of 16
+ * @MALI_C55_AEXP_HIST_FOCUS_00: Top left plane in the first bank, rest in second bank
+ * @MALI_C55_AEXP_HIST_FOCUS_01: Top right plane in the first bank, rest in second bank
+ * @MALI_C55_AEXP_HIST_FOCUS_10: Bottom left plane in the first bank, rest in second bank
+ * @MALI_C55_AEXP_HIST_FOCUS_11: Bottom right plane in the first bank, rest in second bank
+ *
+ * In the "focus" modes statistics are collected into two 512-bin histograms
+ * with a bin width of 8. One colour plane is in the first histogram with the
+ * remainder combined into the second. The four options represent which of the
+ * four positions in a bayer pattern are the focused plane.
+ */
+enum mali_c55_aexp_hist_plane_mode {
+ MALI_C55_AEXP_HIST_COMBINED = 0,
+ MALI_C55_AEXP_HIST_SEPARATE = 1,
+ MALI_C55_AEXP_HIST_FOCUS_00 = 4,
+ MALI_C55_AEXP_HIST_FOCUS_01 = 5,
+ MALI_C55_AEXP_HIST_FOCUS_10 = 6,
+ MALI_C55_AEXP_HIST_FOCUS_11 = 7,
+};
+
+/**
+ * struct mali_c55_params_aexp_hist - configuration for AEXP metering hists
+ *
+ * This struct allows users to configure the 1024-bin AEXP histograms. Broadly
+ * speaking the parameters allow you to mask particular regions of the image and
+ * to select different kinds of histogram.
+ *
+ * The skip_x, offset_x, skip_y and offset_y fields allow users to ignore or
+ * mask pixels in the frame by their position relative to the top left pixel.
+ * First, the skip_y, offset_x and offset_y fields define which of the pixels
+ * within each 2x2 region will be counted in the statistics.
+ *
+ * If skip_y == 0 then two pixels from each covered region will be counted. If
+ * both offset_x and offset_y are zero, then the two left-most pixels in each
+ * 2x2 pixel region will be counted. Setting offset_x = 1 will discount the top
+ * left pixel and count the top right pixel. Setting offset_y = 1 will discount
+ * the bottom left pixel and count the bottom right pixel.
+ *
+ * If skip_y != 0 then only a single pixel from each region covered by the
+ * pattern will be counted. In this case offset_x controls whether the pixel
+ * that's counted is in the left (if offset_x == 0) or right (if offset_x == 1)
+ * column and offset_y controls whether the pixel that's counted is in the top
+ * (if offset_y == 0) or bottom (if offset_y == 1) row.
+ *
+ * The skip_x and skip_y fields control how the 2x2 pixel region is repeated
+ * across the image data. The first instance of the region is always in the top
+ * left of the image data. The skip_x field controls how many pixels are ignored
+ * in the x direction before the pixel masking region is repeated. The skip_y
+ * field controls how many pixels are ignored in the y direction before the
+ * pixel masking region is repeated.
+ *
+ * These fields can be used to reduce the number of pixels counted for the
+ * statistics, but it's important to be careful to configure them correctly.
+ * Some combinations of values will result in colour components from the input
+ * data being ignored entirely, for example in the following configuration:
+ *
+ * skip_x = 0
+ * offset_x = 0
+ * skip_y = 0
+ * offset_y = 0
+ *
+ * Only the R and Gb components of RGGB data that was input would be collected.
+ * Similarly in the following configuration:
+ *
+ * skip_x = 0
+ * offset_x = 0
+ * skip_y = 1
+ * offset_y = 1
+ *
+ * Only the Gb component of RGGB data that was input would be collected. To
+ * correct things such that all 4 colour components were included it would be
+ * necessary to set the skip_x and skip_y fields in a way that resulted in all
+ * four colour components being collected:
+ *
+ * skip_x = 1
+ * offset_x = 0
+ * skip_y = 1
+ * offset_y = 1
+ *
+ * header.type should be set to one of either MALI_C55_PARAM_BLOCK_AEXP_HIST or
+ * MALI_C55_PARAM_BLOCK_AEXP_IHIST from :c:type:`mali_c55_param_block_type`.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @skip_x: Horizontal decimation. See enum mali_c55_aexp_skip_x
+ * @offset_x: Skip the first column, or not. See enum mali_c55_aexp_row_column_offset
+ * @skip_y: Vertical decimation. See enum mali_c55_aexp_skip_y
+ * @offset_y: Skip the first row, or not. See enum mali_c55_aexp_row_column_offset
+ * @scale_bottom: Scale pixels in bottom half of intensity range: 0=1x ,1=2x, 2=4x, 4=8x, 4=16x
+ * @scale_top: scale pixels in top half of intensity range: 0=1x ,1=2x, 2=4x, 4=8x, 4=16x
+ * @plane_mode: Plane separation mode. See enum mali_c55_aexp_hist_plane_mode
+ * @tap_point: Tap point for histogram from enum mali_c55_aexp_hist_tap_points.
+ * This parameter is unused for the post-Iridix Histogram
+ */
+struct mali_c55_params_aexp_hist {
+ struct v4l2_params_block_header header;
+ __u8 skip_x;
+ __u8 offset_x;
+ __u8 skip_y;
+ __u8 offset_y;
+ __u8 scale_bottom;
+ __u8 scale_top;
+ __u8 plane_mode;
+ __u8 tap_point;
+};
+
+/**
+ * struct mali_c55_params_aexp_weights - Array of weights for AEXP metering
+ *
+ * This struct allows users to configure the weighting for both of the 1024-bin
+ * AEXP histograms. The pixel data collected for each zone is multiplied by the
+ * corresponding weight from this array, which may be zero if the intention is
+ * to mask off the zone entirely.
+ *
+ * header.type should be set to one of either MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS
+ * or MALI_C55_PARAM_BLOCK_AEXP_IHIST_WEIGHTS from :c:type:`mali_c55_param_block_type`.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @nodes_used_horiz: Number of active zones horizontally [0..15]
+ * @nodes_used_vert: Number of active zones vertically [0..15]
+ * @zone_weights: Zone weighting. Index is row*col where 0,0 is the top
+ * left zone continuing in raster order. Each zone can be
+ * weighted in the range [0..15]. The number of rows and
+ * columns is defined by @nodes_used_vert and
+ * @nodes_used_horiz
+ */
+struct mali_c55_params_aexp_weights {
+ struct v4l2_params_block_header header;
+ __u8 nodes_used_horiz;
+ __u8 nodes_used_vert;
+ __u8 zone_weights[MALI_C55_MAX_ZONES];
+};
+
+/**
+ * struct mali_c55_params_digital_gain - Digital gain value
+ *
+ * This struct carries a digital gain value to set in the ISP.
+ *
+ * header.type should be set to MALI_C55_PARAM_BLOCK_DIGITAL_GAIN from
+ * :c:type:`mali_c55_param_block_type` for this block.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @gain: The digital gain value to apply, in Q5.8 format.
+ */
+struct mali_c55_params_digital_gain {
+ struct v4l2_params_block_header header;
+ __u16 gain;
+};
+
+/**
+ * enum mali_c55_awb_stats_mode - Statistics mode for AWB
+ * @MALI_C55_AWB_MODE_GRBR: Statistics collected as Green/Red and Blue/Red ratios
+ * @MALI_C55_AWB_MODE_RGBG: Statistics collected as Red/Green and Blue/Green ratios
+ */
+enum mali_c55_awb_stats_mode {
+ MALI_C55_AWB_MODE_GRBR = 0,
+ MALI_C55_AWB_MODE_RGBG,
+};
+
+/**
+ * struct mali_c55_params_awb_gains - Gain settings for auto white balance
+ *
+ * This struct allows users to configure the gains for auto-white balance. There
+ * are four gain settings corresponding to each colour channel in the bayer
+ * domain. Although named generically, the association between the gain applied
+ * and the colour channel is done automatically within the ISP depending on the
+ * input format, and so the following mapping always holds true::
+ *
+ * gain00 = R
+ * gain01 = Gr
+ * gain10 = Gb
+ * gain11 = B
+ *
+ * All of the gains are stored in Q4.8 format.
+ *
+ * header.type should be set to one of either MALI_C55_PARAM_BLOCK_AWB_GAINS or
+ * MALI_C55_PARAM_BLOCK_AWB_GAINS_AEXP from :c:type:`mali_c55_param_block_type`.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @gain00: Multiplier for colour channel 00
+ * @gain01: Multiplier for colour channel 01
+ * @gain10: Multiplier for colour channel 10
+ * @gain11: Multiplier for colour channel 11
+ */
+struct mali_c55_params_awb_gains {
+ struct v4l2_params_block_header header;
+ __u16 gain00;
+ __u16 gain01;
+ __u16 gain10;
+ __u16 gain11;
+};
+
+/**
+ * enum mali_c55_params_awb_tap_points - Tap points for the AWB statistics
+ * @MALI_C55_AWB_STATS_TAP_PF: Immediately after the Purple Fringe block
+ * @MALI_C55_AWB_STATS_TAP_CNR: Immediately after the CNR block
+ */
+enum mali_c55_params_awb_tap_points {
+ MALI_C55_AWB_STATS_TAP_PF = 0,
+ MALI_C55_AWB_STATS_TAP_CNR,
+};
+
+/**
+ * struct mali_c55_params_awb_config - Stats settings for auto-white balance
+ *
+ * This struct allows the configuration of the statistics generated for auto
+ * white balance. Pixel intensity limits can be set to exclude overly bright or
+ * dark regions of an image from the statistics entirely. Colour ratio minima
+ * and maxima can be set to discount pixels who's ratios fall outside the
+ * defined boundaries; there are two sets of registers to do this - the
+ * "min/max" ratios which bound a region and the "high/low" ratios which further
+ * trim the upper and lower ratios. For example with the boundaries configured
+ * as follows, only pixels whos colour ratios falls into the region marked "A"
+ * would be counted::
+ *
+ * cr_high
+ * 2.0 | |
+ * | cb_max --> _________________________v_____
+ * 1.8 | | \ |
+ * | | \ |
+ * 1.6 | | \ |
+ * | | \ |
+ * c 1.4 | cb_low -->|\ A \|<-- cb_high
+ * b | | \ |
+ * 1.2 | | \ |
+ * r | | \ |
+ * a 1.0 | cb_min --> |____\_________________________|
+ * t | ^ ^ ^
+ * i 0.8 | | | |
+ * o | cr_min | cr_max
+ * s 0.6 | |
+ * | cr_low
+ * 0.4 |
+ * |
+ * 0.2 |
+ * |
+ * 0.0 |_______________________________________________________________
+ * 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
+ * cr ratios
+ *
+ * header.type should be set to MALI_C55_PARAM_BLOCK_AWB_CONFIG from
+ * :c:type:`mali_c55_param_block_type` for this block.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @tap_point: The tap point from enum mali_c55_params_awb_tap_points
+ * @stats_mode: AWB statistics collection mode, see :c:type:`mali_c55_awb_stats_mode`
+ * @white_level: Upper pixel intensity (I.E. raw pixel values) limit
+ * @black_level: Lower pixel intensity (I.E. raw pixel values) limit
+ * @cr_max: Maximum R/G ratio (Q4.8 format)
+ * @cr_min: Minimum R/G ratio (Q4.8 format)
+ * @cb_max: Maximum B/G ratio (Q4.8 format)
+ * @cb_min: Minimum B/G ratio (Q4.8 format)
+ * @nodes_used_horiz: Number of active zones horizontally [0..15]
+ * @nodes_used_vert: Number of active zones vertically [0..15]
+ * @cr_high: R/G ratio trim high (Q4.8 format)
+ * @cr_low: R/G ratio trim low (Q4.8 format)
+ * @cb_high: B/G ratio trim high (Q4.8 format)
+ * @cb_low: B/G ratio trim low (Q4.8 format)
+ */
+struct mali_c55_params_awb_config {
+ struct v4l2_params_block_header header;
+ __u8 tap_point;
+ __u8 stats_mode;
+ __u16 white_level;
+ __u16 black_level;
+ __u16 cr_max;
+ __u16 cr_min;
+ __u16 cb_max;
+ __u16 cb_min;
+ __u8 nodes_used_horiz;
+ __u8 nodes_used_vert;
+ __u16 cr_high;
+ __u16 cr_low;
+ __u16 cb_high;
+ __u16 cb_low;
+};
+
+#define MALI_C55_NUM_MESH_SHADING_ELEMENTS 3072
+
+/**
+ * struct mali_c55_params_mesh_shading_config - Mesh shading configuration
+ *
+ * The mesh shading correction module allows programming a separate table of
+ * either 16x16 or 32x32 node coefficients for 3 different light sources. The
+ * final correction coefficients applied are computed by blending the
+ * coefficients from two tables together.
+ *
+ * A page of 1024 32-bit integers is associated to each colour channel, with
+ * pages stored consecutively in memory. Each 32-bit integer packs 3 8-bit
+ * correction coefficients for a single node, one for each of the three light
+ * sources. The 8 most significant bits are unused. The following table
+ * describes the layout::
+ *
+ * +----------- Page (Colour Plane) 0 -------------+
+ * | @mesh[i] | Mesh Point | Bits | Light Source |
+ * +-----------+------------+-------+--------------+
+ * | 0 | 0,0 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | 1 | 0,1 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | ... | ... | ... | ... |
+ * +-----------+------------+-------+--------------+
+ * | 1023 | 31,31 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +----------- Page (Colour Plane) 1 -------------+
+ * | @mesh[i] | Mesh Point | Bits | Light Source |
+ * +-----------+------------+-------+--------------+
+ * | 1024 | 0,0 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | 1025 | 0,1 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | ... | ... | ... | ... |
+ * +-----------+------------+-------+--------------+
+ * | 2047 | 31,31 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +----------- Page (Colour Plane) 2 -------------+
+ * | @mesh[i] | Mesh Point | Bits | Light Source |
+ * +-----------+------------+-------+--------------+
+ * | 2048 | 0,0 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | 2049 | 0,1 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ * | ... | ... | ... | ... |
+ * +-----------+------------+-------+--------------+
+ * | 3071 | 31,31 | 16,23 | LS2 |
+ * | | | 08-15 | LS1 |
+ * | | | 00-07 | LS0 |
+ * +-----------+------------+-------+--------------+
+ *
+ * The @mesh_scale member determines the precision and minimum and maximum gain.
+ * For example if @mesh_scale is 0 and therefore selects 0 - 2x gain, a value of
+ * 0 in a coefficient means 0.0 gain, a value of 128 means 1.0 gain and 255
+ * means 2.0 gain.
+ *
+ * header.type should be set to MALI_C55_PARAM_MESH_SHADING_CONFIG from
+ * :c:type:`mali_c55_param_block_type` for this block.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @mesh_show: Output the mesh data rather than image data
+ * @mesh_scale: Set the precision and maximum gain range of mesh shading
+ * - 0 = 0-2x gain
+ * - 1 = 0-4x gain
+ * - 2 = 0-8x gain
+ * - 3 = 0-16x gain
+ * - 4 = 1-2x gain
+ * - 5 = 1-3x gain
+ * - 6 = 1-5x gain
+ * - 7 = 1-9x gain
+ * @mesh_page_r: Mesh page select for red colour plane [0..2]
+ * @mesh_page_g: Mesh page select for green colour plane [0..2]
+ * @mesh_page_b: Mesh page select for blue colour plane [0..2]
+ * @mesh_width: Number of horizontal nodes minus 1 [15,31]
+ * @mesh_height: Number of vertical nodes minus 1 [15,31]
+ * @mesh: Mesh shading correction tables
+ */
+struct mali_c55_params_mesh_shading_config {
+ struct v4l2_params_block_header header;
+ __u8 mesh_show;
+ __u8 mesh_scale;
+ __u8 mesh_page_r;
+ __u8 mesh_page_g;
+ __u8 mesh_page_b;
+ __u8 mesh_width;
+ __u8 mesh_height;
+ __u32 mesh[MALI_C55_NUM_MESH_SHADING_ELEMENTS];
+};
+
+/** enum mali_c55_params_mesh_alpha_bank - Mesh shading table bank selection
+ * @MALI_C55_MESH_ALPHA_BANK_LS0_AND_LS1 - Select Light Sources 0 and 1
+ * @MALI_C55_MESH_ALPHA_BANK_LS1_AND_LS2 - Select Light Sources 1 and 2
+ * @MALI_C55_MESH_ALPHA_BANK_LS0_AND_LS2 - Select Light Sources 0 and 2
+ */
+enum mali_c55_params_mesh_alpha_bank {
+ MALI_C55_MESH_ALPHA_BANK_LS0_AND_LS1 = 0,
+ MALI_C55_MESH_ALPHA_BANK_LS1_AND_LS2 = 1,
+ MALI_C55_MESH_ALPHA_BANK_LS0_AND_LS2 = 4
+};
+
+/**
+ * struct mali_c55_params_mesh_shading_selection - Mesh table selection
+ *
+ * The module computes the final correction coefficients by blending the ones
+ * from two light source tables, which are selected (independently for each
+ * colour channel) by the @mesh_alpha_bank_r/g/b fields.
+ *
+ * The final blended coefficients for each node are calculated using the
+ * following equation:
+ *
+ * Final coefficient = (a * LS\ :sub:`b`\ + (256 - a) * LS\ :sub:`a`\) / 256
+ *
+ * Where a is the @mesh_alpha_r/g/b value, and LS\ :sub:`a`\ and LS\ :sub:`b`\
+ * are the node cofficients for the two tables selected by the
+ * @mesh_alpha_bank_r/g/b value.
+ *
+ * The scale of the applied correction may also be controlled by tuning the
+ * @mesh_strength member. This is a modifier to the final coefficients which can
+ * be used to globally reduce the gains applied.
+ *
+ * header.type should be set to MALI_C55_PARAM_MESH_SHADING_SELECTION from
+ * :c:type:`mali_c55_param_block_type` for this block.
+ *
+ * @header: The Mali-C55 parameters block header
+ * @mesh_alpha_bank_r: Red mesh table select (c:type:`enum mali_c55_params_mesh_alpha_bank`)
+ * @mesh_alpha_bank_g: Green mesh table select (c:type:`enum mali_c55_params_mesh_alpha_bank`)
+ * @mesh_alpha_bank_b: Blue mesh table select (c:type:`enum mali_c55_params_mesh_alpha_bank`)
+ * @mesh_alpha_r: Blend coefficient for R [0..255]
+ * @mesh_alpha_g: Blend coefficient for G [0..255]
+ * @mesh_alpha_b: Blend coefficient for B [0..255]
+ * @mesh_strength: Mesh strength in Q4.12 format [0..4096]
+ */
+struct mali_c55_params_mesh_shading_selection {
+ struct v4l2_params_block_header header;
+ __u8 mesh_alpha_bank_r;
+ __u8 mesh_alpha_bank_g;
+ __u8 mesh_alpha_bank_b;
+ __u8 mesh_alpha_r;
+ __u8 mesh_alpha_g;
+ __u8 mesh_alpha_b;
+ __u16 mesh_strength;
+};
+
+/**
+ * define MALI_C55_PARAMS_MAX_SIZE - Maximum size of all Mali C55 Parameters
+ *
+ * Though the parameters for the Mali-C55 are passed as optional blocks, the
+ * driver still needs to know the absolute maximum size so that it can allocate
+ * a buffer sized appropriately to accommodate userspace attempting to set all
+ * possible parameters in a single frame.
+ *
+ * Some structs are in this list multiple times. Where that's the case, it just
+ * reflects the fact that the same struct can be used with multiple different
+ * header types from :c:type:`mali_c55_param_block_type`.
+ */
+#define MALI_C55_PARAMS_MAX_SIZE \
+ (sizeof(struct mali_c55_params_sensor_off_preshading) + \
+ sizeof(struct mali_c55_params_aexp_hist) + \
+ sizeof(struct mali_c55_params_aexp_weights) + \
+ sizeof(struct mali_c55_params_aexp_hist) + \
+ sizeof(struct mali_c55_params_aexp_weights) + \
+ sizeof(struct mali_c55_params_digital_gain) + \
+ sizeof(struct mali_c55_params_awb_gains) + \
+ sizeof(struct mali_c55_params_awb_config) + \
+ sizeof(struct mali_c55_params_awb_gains) + \
+ sizeof(struct mali_c55_params_mesh_shading_config) + \
+ sizeof(struct mali_c55_params_mesh_shading_selection))
+
#endif /* __UAPI_MALI_C55_CONFIG_H */
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (15 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 17/19] media: uapi: Add parameters structs to mali-c55-config.h Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
2025-07-30 14:36 ` Sakari Ailus
2025-07-14 15:06 ` [PATCH v11 19/19] Documentation: mali-c55: Document the mali-c55 parameter setting Daniel Scally
17 siblings, 1 reply; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Add a new code file to the mali-c55 driver that registers an output
video node for userspace to queue buffers of parameters to. Handlers
are included to program the statistics generation plus the white
balance, black level correction and mesh shading correction blocks.
Update the rest of the driver to register and link the new video node
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Remove the media jobs / scheduler functionality. We'll add
it in a separate series.
- Update copyright year
- Use the new v4l2-params functions and objects
- Use new video_device_pipeline_started() helper
Changes in v10:
- Account for v2 of media jobs framework
Changes in v9:
- Support memory input mode through the media jobs framework
- Include this video device in the synchronised start/stop across the
pipeline
Changes in v8:
- None
Changes in v7:
- Swapped to use a distinct struct as the buffer for vb2 so that we
didn't have lots of unused fields.
- Defined a union for easy handling of the different block types
- Handled the new @flags field instead of the @enabled field in each
handler.
- Copied the userspace buffer to a scratch buffer in the driver for each
buffer to protect against userspace modifying it whilst the driver is
processing it
- Validated the buffer from userspace when the buffer is queued rather
than processed.
- Helper function to return all buffers to userspace
- Stopped the ISP when stream stop is called.
Changes in v6:
- Used a union to generalise the block pointer rather than resorting to
casting everywhere - fantastic idea Sakari, this made it much cleaner.
- Reworked the loop in mali_c55_params_write_config() so that we can be
sure there's remaining space for the next block header.
Changes in v5:
- New patch
---
drivers/media/platform/arm/mali-c55/Makefile | 1 +
.../media/platform/arm/mali-c55/mali-c55-common.h | 28 +
.../media/platform/arm/mali-c55/mali-c55-core.c | 24 +
drivers/media/platform/arm/mali-c55/mali-c55-isp.c | 23 +-
.../media/platform/arm/mali-c55/mali-c55-params.c | 825 +++++++++++++++++++++
.../platform/arm/mali-c55/mali-c55-registers.h | 128 ++++
6 files changed, 1028 insertions(+), 1 deletion(-)
diff --git a/drivers/media/platform/arm/mali-c55/Makefile b/drivers/media/platform/arm/mali-c55/Makefile
index b5a22d414479dd1eb6435f474cb95aaa14879aa8..d5718b0b23e0ec1ced76837cb7be08df71fc91a7 100644
--- a/drivers/media/platform/arm/mali-c55/Makefile
+++ b/drivers/media/platform/arm/mali-c55/Makefile
@@ -3,6 +3,7 @@
mali-c55-y := mali-c55-capture.o \
mali-c55-core.o \
mali-c55-isp.o \
+ mali-c55-params.o \
mali-c55-resizer.o \
mali-c55-stats.o \
mali-c55-tpg.o
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-common.h b/drivers/media/platform/arm/mali-c55/mali-c55-common.h
index 3970ace9ee8143b65639db1111b120e6c5fec70f..9267a2e9344e96f1102cfa8f2e88c67b80b1b764 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-common.h
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-common.h
@@ -21,6 +21,7 @@
#include <media/v4l2-ctrls.h>
#include <media/v4l2-dev.h>
#include <media/v4l2-device.h>
+#include <media/v4l2-params.h>
#include <media/v4l2-subdev.h>
#include <media/videobuf2-core.h>
#include <media/videobuf2-v4l2.h>
@@ -45,6 +46,7 @@ struct dma_chan;
struct mali_c55;
struct mali_c55_cap_dev;
struct media_pipeline;
+struct mali_c55_params_buffer;
struct platform_device;
struct resource;
@@ -53,6 +55,7 @@ enum mali_c55_isp_pads {
MALI_C55_ISP_PAD_SOURCE_VIDEO,
MALI_C55_ISP_PAD_SOURCE_BYPASS,
MALI_C55_ISP_PAD_SOURCE_STATS,
+ MALI_C55_ISP_PAD_SINK_PARAMS,
MALI_C55_ISP_NUM_PADS,
};
@@ -169,6 +172,12 @@ struct mali_c55_stats_buf {
bool failed;
};
+struct mali_c55_params_buf {
+ struct vb2_v4l2_buffer vb;
+ struct list_head queue;
+ struct v4l2_params_buffer *config;
+};
+
struct mali_c55_stats {
struct mali_c55 *mali_c55;
struct video_device vdev;
@@ -185,6 +194,21 @@ struct mali_c55_stats {
} buffers;
};
+struct mali_c55_params {
+ struct mali_c55 *mali_c55;
+ struct video_device vdev;
+ struct vb2_queue queue;
+ struct media_pad pad;
+ /* Mutex to provide to vb2 */
+ struct mutex lock;
+
+ struct {
+ /* Spinlock to guard buffer queue */
+ spinlock_t lock;
+ struct list_head queue;
+ } buffers;
+};
+
enum mali_c55_config_spaces {
MALI_C55_CONFIG_PING,
MALI_C55_CONFIG_PONG,
@@ -229,6 +253,7 @@ struct mali_c55 {
struct mali_c55_isp isp;
struct mali_c55_resizer resizers[MALI_C55_NUM_RSZS];
struct mali_c55_cap_dev cap_devs[MALI_C55_NUM_CAP_DEVS];
+ struct mali_c55_params params;
struct mali_c55_stats stats;
struct mali_c55_context context;
@@ -260,6 +285,8 @@ int mali_c55_register_capture_devs(struct mali_c55 *mali_c55);
void mali_c55_unregister_capture_devs(struct mali_c55 *mali_c55);
int mali_c55_register_stats(struct mali_c55 *mali_c55);
void mali_c55_unregister_stats(struct mali_c55 *mali_c55);
+int mali_c55_register_params(struct mali_c55 *mali_c55);
+void mali_c55_unregister_params(struct mali_c55 *mali_c55);
struct mali_c55_context *mali_c55_get_active_context(struct mali_c55 *mali_c55);
void mali_c55_set_plane_done(struct mali_c55_cap_dev *cap_dev,
enum mali_c55_planes plane);
@@ -279,5 +306,6 @@ mali_c55_isp_get_mbus_config_by_index(u32 index);
bool mali_c55_pipeline_ready(struct media_pipeline *pipe);
void mali_c55_stats_fill_buffer(struct mali_c55 *mali_c55,
enum mali_c55_config_spaces cfg_space);
+void mali_c55_params_write_config(struct mali_c55 *mali_c55);
#endif /* _MALI_C55_COMMON_H */
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-core.c b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
index 4b71725f956648595400dc57d590383bcea989e8..40ec62486de152b86c72e3f6057f98240756bca0 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-core.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
@@ -387,6 +387,16 @@ static int mali_c55_create_links(struct mali_c55 *mali_c55)
goto err_remove_links;
}
+ ret = media_create_pad_link(&mali_c55->params.vdev.entity, 0,
+ &mali_c55->isp.sd.entity,
+ MALI_C55_ISP_PAD_SINK_PARAMS,
+ MEDIA_LNK_FL_ENABLED);
+ if (ret) {
+ dev_err(mali_c55->dev,
+ "failed to link ISP and parameters video node\n");
+ goto err_remove_links;
+ }
+
return 0;
err_remove_links:
@@ -401,6 +411,7 @@ static void mali_c55_unregister_entities(struct mali_c55 *mali_c55)
mali_c55_unregister_isp(mali_c55);
mali_c55_unregister_resizers(mali_c55);
mali_c55_unregister_capture_devs(mali_c55);
+ mali_c55_unregister_params(mali_c55);
mali_c55_unregister_stats(mali_c55);
}
@@ -443,6 +454,10 @@ static int mali_c55_register_entities(struct mali_c55 *mali_c55)
if (ret)
goto err_unregister_entities;
+ ret = mali_c55_register_params(mali_c55);
+ if (ret)
+ goto err_unregister_entities;
+
ret = mali_c55_register_stats(mali_c55);
if (ret)
goto err_unregister_entities;
@@ -532,6 +547,7 @@ static int mali_c55_media_frameworks_init(struct mali_c55 *mali_c55)
sizeof(mali_c55->media_dev.model));
media_device_init(&mali_c55->media_dev);
+
ret = media_device_register(&mali_c55->media_dev);
if (ret)
goto err_cleanup_media_device;
@@ -682,6 +698,14 @@ static irqreturn_t mali_c55_isr(int irq, void *context)
curr_config >>= ffs(MALI_C55_REG_PING_PONG_READ_MASK) - 1;
mali_c55->next_config = curr_config ^ 1;
+ /*
+ * Write the configuration parameters received from
+ * userspace into the configuration buffer, which will
+ * be transferred to the 'next' active config space at
+ * by mali_c55_swap_next_config().
+ */
+ mali_c55_params_write_config(mali_c55);
+
mali_c55_stats_fill_buffer(mali_c55,
mali_c55->next_config ? MALI_C55_CONFIG_PING :
MALI_C55_CONFIG_PONG);
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-isp.c b/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
index 3095a9b9ceacebb688f17cd49de490385f417e12..020580fb2bac748bc4431d1f7288f1a7e652628d 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-isp.c
@@ -151,6 +151,7 @@ static int mali_c55_isp_start(struct mali_c55 *mali_c55,
cfg->bypass ? MALI_C55_ISP_RAW_BYPASS_BYPASS_MASK :
0x00);
+ mali_c55_params_write_config(mali_c55);
ret = mali_c55_config_write(ctx, MALI_C55_CONFIG_PING, true);
if (ret) {
dev_err(mali_c55->dev, "failed to write ISP config\n");
@@ -516,12 +517,19 @@ static int mali_c55_isp_init_state(struct v4l2_subdev *sd,
src_fmt = v4l2_subdev_state_get_format(state,
MALI_C55_ISP_PAD_SOURCE_STATS);
+ sink_fmt = v4l2_subdev_state_get_format(state,
+ MALI_C55_ISP_PAD_SINK_PARAMS);
src_fmt->width = 0;
src_fmt->height = 0;
src_fmt->field = V4L2_FIELD_NONE;
src_fmt->code = MEDIA_BUS_FMT_METADATA_FIXED;
+ sink_fmt->width = 0;
+ sink_fmt->height = 0;
+ sink_fmt->field = V4L2_FIELD_NONE;
+ sink_fmt->code = MEDIA_BUS_FMT_METADATA_FIXED;
+
return 0;
}
@@ -529,8 +537,20 @@ static const struct v4l2_subdev_internal_ops mali_c55_isp_internal_ops = {
.init_state = mali_c55_isp_init_state,
};
+static int mali_c55_subdev_link_validate(struct media_link *link)
+{
+ /*
+ * Skip validation for the parameters sink pad, as the source is not
+ * a subdevice.
+ */
+ if (link->sink->index == MALI_C55_ISP_PAD_SINK_PARAMS)
+ return 0;
+
+ return v4l2_subdev_link_validate(link);
+}
+
static const struct media_entity_operations mali_c55_isp_media_ops = {
- .link_validate = v4l2_subdev_link_validate,
+ .link_validate = mali_c55_subdev_link_validate,
.pipeline_started = mali_c55_pipeline_started,
.pipeline_stopped = mali_c55_pipeline_stopped,
};
@@ -621,6 +641,7 @@ int mali_c55_register_isp(struct mali_c55 *mali_c55)
isp->pads[MALI_C55_ISP_PAD_SOURCE_VIDEO].flags = MEDIA_PAD_FL_SOURCE;
isp->pads[MALI_C55_ISP_PAD_SOURCE_BYPASS].flags = MEDIA_PAD_FL_SOURCE;
isp->pads[MALI_C55_ISP_PAD_SOURCE_STATS].flags = MEDIA_PAD_FL_SOURCE;
+ isp->pads[MALI_C55_ISP_PAD_SINK_PARAMS].flags = MEDIA_PAD_FL_SINK;
ret = media_entity_pads_init(&sd->entity, MALI_C55_ISP_NUM_PADS,
isp->pads);
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-params.c b/drivers/media/platform/arm/mali-c55/mali-c55-params.c
new file mode 100644
index 0000000000000000000000000000000000000000..0b47ecf5de8c32c062028a0de59915aea8461052
--- /dev/null
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-params.c
@@ -0,0 +1,825 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * ARM Mali-C55 ISP Driver - Configuration parameters output device
+ *
+ * Copyright (C) 2025 Ideas on Board Oy
+ */
+#include <linux/media/arm/mali-c55-config.h>
+#include <linux/pm_runtime.h>
+
+#include <media/media-entity.h>
+#include <media/v4l2-dev.h>
+#include <media/v4l2-event.h>
+#include <media/v4l2-fh.h>
+#include <media/v4l2-ioctl.h>
+#include <media/v4l2-params.h>
+#include <media/videobuf2-core.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "mali-c55-common.h"
+#include "mali-c55-registers.h"
+
+/**
+ * union mali_c55_params_block - Generalisation of a parameter block
+ *
+ * This union allows the driver to treat a block as a generic pointer to this
+ * union and safely access the header and block-specific struct without having
+ * to resort to casting. The header member is accessed first, and the type field
+ * checked which allows the driver to determine which of the other members
+ * should be used. The data member at the end allows a pointer to an address
+ * within the data member of :c:type:`mali_c55_params_buffer` to initialise a
+ * union variable.
+ *
+ * @header: Pointer to the shared header struct embedded as the
+ * first member of all the possible other members (except
+ * @data). This member would be accessed first and the type
+ * field checked to determine which of the other members
+ * should be accessed.
+ * @sensor_offs: For header->type == MALI_C55_PARAM_BLOCK_SENSOR_OFFS
+ * @aexp_hist: For header->type == MALI_C55_PARAM_BLOCK_AEXP_HIST and
+ * header->type == MALI_C55_PARAM_BLOCK_AEXP_IHIST
+ * @aexp_weights: For header->type == MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS
+ * and header->type = MALI_C55_PARAM_BLOCK_AEXP_IHIST_WEIGHTS
+ * @digital_gain: For header->type == MALI_C55_PARAM_BLOCK_DIGITAL_GAIN
+ * @awb_gains: For header->type == MALI_C55_PARAM_BLOCK_AWB_GAINS and
+ * header->type = MALI_C55_PARAM_BLOCK_AWB_GAINS_AEXP
+ * @awb_config: For header->type == MALI_C55_PARAM_MESH_SHADING_CONFIG
+ * @shading_config: For header->type == MALI_C55_PARAM_MESH_SHADING_SELECTION
+ * @shading_selection: For header->type == MALI_C55_PARAM_BLOCK_SENSOR_OFFS
+ * @data: Allows easy initialisation of a union variable with a
+ * pointer into a __u8 array.
+ */
+union mali_c55_params_block {
+ const struct v4l2_params_block_header *header;
+ const struct mali_c55_params_sensor_off_preshading *sensor_offs;
+ const struct mali_c55_params_aexp_hist *aexp_hist;
+ const struct mali_c55_params_aexp_weights *aexp_weights;
+ const struct mali_c55_params_digital_gain *digital_gain;
+ const struct mali_c55_params_awb_gains *awb_gains;
+ const struct mali_c55_params_awb_config *awb_config;
+ const struct mali_c55_params_mesh_shading_config *shading_config;
+ const struct mali_c55_params_mesh_shading_selection *shading_selection;
+ const __u8 *data;
+};
+
+#define to_mali_c55_params_buf(vbuf) \
+ container_of(vbuf, struct mali_c55_params_buf, vb)
+
+static void mali_c55_params_sensor_offs(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_sensor_off_preshading *p;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+ __u32 global_offset;
+
+ block.header = header;
+ p = block.sensor_offs;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE) {
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_BYPASS_3,
+ MALI_C55_REG_BYPASS_3_SENSOR_OFFSET_PRE_SH,
+ MALI_C55_REG_BYPASS_3_SENSOR_OFFSET_PRE_SH);
+ return;
+ }
+
+ if (!(p->chan00 || p->chan01 || p->chan10 || p->chan11))
+ return;
+
+ mali_c55_ctx_write(mali_c55, MALI_C55_REG_SENSOR_OFF_PRE_SHA_00,
+ p->chan00 & MALI_C55_SENSOR_OFF_PRE_SHA_MASK);
+ mali_c55_ctx_write(mali_c55, MALI_C55_REG_SENSOR_OFF_PRE_SHA_01,
+ p->chan01 & MALI_C55_SENSOR_OFF_PRE_SHA_MASK);
+ mali_c55_ctx_write(mali_c55, MALI_C55_REG_SENSOR_OFF_PRE_SHA_10,
+ p->chan10 & MALI_C55_SENSOR_OFF_PRE_SHA_MASK);
+ mali_c55_ctx_write(mali_c55, MALI_C55_REG_SENSOR_OFF_PRE_SHA_11,
+ p->chan11 & MALI_C55_SENSOR_OFF_PRE_SHA_MASK);
+
+ /*
+ * The average offset is applied as a global offset for the digital
+ * gain block
+ */
+ global_offset = (p->chan00 + p->chan01 + p->chan10 + p->chan11) >> 2;
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_DIGITAL_GAIN_OFFSET,
+ MALI_C55_DIGITAL_GAIN_OFFSET_MASK,
+ global_offset);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_BYPASS_3,
+ MALI_C55_REG_BYPASS_3_SENSOR_OFFSET_PRE_SH,
+ 0x00);
+}
+
+static void mali_c55_params_aexp_hist(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_aexp_hist *params;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+ u32 disable_mask;
+ u32 disable_val;
+ u32 base;
+
+ block.header = header;
+
+ if (block.header->type == MALI_C55_PARAM_BLOCK_AEXP_HIST) {
+ disable_mask = MALI_C55_AEXP_HIST_DISABLE_MASK;
+ disable_val = MALI_C55_AEXP_HIST_DISABLE;
+ base = MALI_C55_REG_AEXP_HIST_BASE;
+ } else {
+ disable_mask = MALI_C55_AEXP_IHIST_DISABLE_MASK;
+ disable_val = MALI_C55_AEXP_IHIST_DISABLE;
+ base = MALI_C55_REG_AEXP_IHIST_BASE;
+ }
+
+ params = block.aexp_hist;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE) {
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ disable_mask, disable_val);
+ return;
+ }
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ disable_mask, false);
+
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SKIP_OFFSET,
+ MALI_C55_AEXP_HIST_SKIP_X_MASK, params->skip_x);
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SKIP_OFFSET,
+ MALI_C55_AEXP_HIST_OFFSET_X_MASK,
+ MALI_C55_AEXP_HIST_OFFSET_X(params->offset_x));
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SKIP_OFFSET,
+ MALI_C55_AEXP_HIST_SKIP_Y_MASK,
+ MALI_C55_AEXP_HIST_SKIP_Y(params->skip_y));
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SKIP_OFFSET,
+ MALI_C55_AEXP_HIST_OFFSET_Y_MASK,
+ MALI_C55_AEXP_HIST_OFFSET_Y(params->offset_y));
+
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SCALE_OFFSET,
+ MALI_C55_AEXP_HIST_SCALE_BOTTOM_MASK,
+ params->scale_bottom);
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_SCALE_OFFSET,
+ MALI_C55_AEXP_HIST_SCALE_TOP_MASK,
+ MALI_C55_AEXP_HIST_SCALE_TOP(params->scale_top));
+
+ mali_c55_ctx_update_bits(mali_c55, base + MALI_C55_AEXP_HIST_PLANE_MODE_OFFSET,
+ MALI_C55_AEXP_HIST_PLANE_MODE_MASK,
+ params->plane_mode);
+
+ if (block.header->type == MALI_C55_PARAM_BLOCK_AEXP_HIST)
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ MALI_C55_AEXP_HIST_SWITCH_MASK,
+ MALI_C55_AEXP_HIST_SWITCH(params->tap_point));
+}
+
+static void
+mali_c55_params_aexp_hist_weights(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_aexp_weights *params;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+ u32 base;
+
+ block.header = header;
+ params = block.aexp_weights;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE)
+ return;
+
+ base = block.header->type == MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS ?
+ MALI_C55_REG_AEXP_HIST_BASE :
+ MALI_C55_REG_AEXP_IHIST_BASE;
+
+ mali_c55_ctx_update_bits(mali_c55,
+ base + MALI_C55_AEXP_HIST_NODES_USED_OFFSET,
+ MALI_C55_AEXP_HIST_NODES_USED_HORIZ_MASK,
+ params->nodes_used_horiz);
+ mali_c55_ctx_update_bits(mali_c55,
+ base + MALI_C55_AEXP_HIST_NODES_USED_OFFSET,
+ MALI_C55_AEXP_HIST_NODES_USED_VERT_MASK,
+ MALI_C55_AEXP_HIST_NODES_USED_VERT(params->nodes_used_vert));
+
+ /*
+ * The zone weights array is a 225-element array of u8 values, but that
+ * is a bit annoying to handle given the ISP expects 32-bit writes. We
+ * just reinterpret it as a 57-element array of 32-bit values for the
+ * purposes of this transaction (the 3 bytes of additional space at the
+ * end of the write is just padding for the array of weights in the ISP
+ * memory space anyway, so there's no risk of overwriting other
+ * registers).
+ */
+ for (unsigned int i = 0; i < 57; i++) {
+ u32 val = ((u32 *)params->zone_weights)[i]
+ & MALI_C55_AEXP_HIST_ZONE_WEIGHT_MASK;
+ u32 addr = base + MALI_C55_AEXP_HIST_ZONE_WEIGHTS_OFFSET + (4 * i);
+
+ mali_c55_ctx_write(mali_c55, addr, val);
+ }
+}
+
+static void mali_c55_params_digital_gain(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_digital_gain *dgain;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+ u32 gain;
+
+ block.header = header;
+ dgain = block.digital_gain;
+
+ /*
+ * If the block is flagged as disabled we write a gain of 1.0, which in
+ * Q5.8 format is 256.
+ */
+ gain = block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE ?
+ 256 : dgain->gain;
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_DIGITAL_GAIN,
+ MALI_C55_DIGITAL_GAIN_MASK,
+ gain);
+}
+
+static void mali_c55_params_awb_gains(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_awb_gains *gains;
+ u32 gain00, gain01, gain10, gain11;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+
+ block.header = header;
+ gains = block.awb_gains;
+
+ /*
+ * There are two places AWB gains can be set in the ISP; one affects the
+ * image output data and the other affects the statistics for the
+ * AEXP-0 tap point.
+ */
+ u32 addr1 = block.header->type == MALI_C55_PARAM_BLOCK_AWB_GAINS ?
+ MALI_C55_REG_AWB_GAINS1 :
+ MALI_C55_REG_AWB_GAINS1_AEXP;
+ u32 addr2 = block.header->type == MALI_C55_PARAM_BLOCK_AWB_GAINS ?
+ MALI_C55_REG_AWB_GAINS2 :
+ MALI_C55_REG_AWB_GAINS2_AEXP;
+
+ /* If the block is flagged disabled, set all of the gains to 1.0 */
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE) {
+ gain00 = 256;
+ gain01 = 256;
+ gain10 = 256;
+ gain11 = 256;
+ } else {
+ gain00 = gains->gain00;
+ gain01 = gains->gain01;
+ gain10 = gains->gain10;
+ gain11 = gains->gain11;
+ }
+
+ mali_c55_ctx_update_bits(mali_c55, addr1, MALI_C55_AWB_GAIN00_MASK,
+ gain00);
+ mali_c55_ctx_update_bits(mali_c55, addr1, MALI_C55_AWB_GAIN01_MASK,
+ MALI_C55_AWB_GAIN01(gain01));
+ mali_c55_ctx_update_bits(mali_c55, addr2, MALI_C55_AWB_GAIN10_MASK,
+ gain10);
+ mali_c55_ctx_update_bits(mali_c55, addr2, MALI_C55_AWB_GAIN11_MASK,
+ MALI_C55_AWB_GAIN11(gain11));
+}
+
+static void mali_c55_params_awb_config(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_awb_config *params;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+
+ block.header = header;
+ params = block.awb_config;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE) {
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ MALI_C55_AWB_DISABLE_MASK,
+ MALI_C55_AWB_DISABLE_MASK);
+ return;
+ }
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ MALI_C55_AWB_DISABLE_MASK, false);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_STATS_MODE,
+ MALI_C55_AWB_STATS_MODE_MASK, params->stats_mode);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_WHITE_LEVEL,
+ MALI_C55_AWB_WHITE_LEVEL_MASK, params->white_level);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_BLACK_LEVEL,
+ MALI_C55_AWB_BLACK_LEVEL_MASK, params->black_level);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CR_MAX,
+ MALI_C55_AWB_CR_MAX_MASK, params->cr_max);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CR_MIN,
+ MALI_C55_AWB_CR_MIN_MASK, params->cr_min);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CB_MAX,
+ MALI_C55_AWB_CB_MAX_MASK, params->cb_max);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CB_MIN,
+ MALI_C55_AWB_CB_MIN_MASK, params->cb_min);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_NODES_USED,
+ MALI_C55_AWB_NODES_USED_HORIZ_MASK,
+ params->nodes_used_horiz);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_NODES_USED,
+ MALI_C55_AWB_NODES_USED_VERT_MASK,
+ MALI_C55_AWB_NODES_USED_VERT(params->nodes_used_vert));
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CR_HIGH,
+ MALI_C55_AWB_CR_HIGH_MASK, params->cr_high);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CR_LOW,
+ MALI_C55_AWB_CR_LOW_MASK, params->cr_low);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CB_HIGH,
+ MALI_C55_AWB_CB_HIGH_MASK, params->cb_high);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_AWB_CB_LOW,
+ MALI_C55_AWB_CB_LOW_MASK, params->cb_low);
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_METERING_CONFIG,
+ MALI_C55_AWB_SWITCH_MASK,
+ MALI_C55_AWB_SWITCH(params->tap_point));
+}
+
+static void mali_c55_params_lsc_config(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_mesh_shading_config *params;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+ unsigned int i;
+ u32 addr;
+
+ block.header = header;
+ params = block.shading_config;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE) {
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_ENABLE_MASK,
+ false);
+ return;
+ }
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_ENABLE_MASK, true);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_MESH_SHOW_MASK,
+ MALI_C55_MESH_SHADING_MESH_SHOW(params->mesh_show));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_SCALE_MASK,
+ MALI_C55_MESH_SHADING_SCALE(params->mesh_scale));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_PAGE_R_MASK,
+ MALI_C55_MESH_SHADING_PAGE_R(params->mesh_page_r));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_PAGE_G_MASK,
+ MALI_C55_MESH_SHADING_PAGE_G(params->mesh_page_g));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_PAGE_B_MASK,
+ MALI_C55_MESH_SHADING_PAGE_B(params->mesh_page_b));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_MESH_WIDTH_MASK,
+ MALI_C55_MESH_SHADING_MESH_WIDTH(params->mesh_width));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_CONFIG,
+ MALI_C55_MESH_SHADING_MESH_HEIGHT_MASK,
+ MALI_C55_MESH_SHADING_MESH_HEIGHT(params->mesh_height));
+
+ for (i = 0; i < MALI_C55_NUM_MESH_SHADING_ELEMENTS; i++) {
+ addr = MALI_C55_REG_MESH_SHADING_TABLES + (i * 4);
+ mali_c55_ctx_write(mali_c55, addr, params->mesh[i]);
+ }
+}
+
+static void mali_c55_params_lsc_selection(void *arg, const struct v4l2_params_block_header *header)
+{
+ const struct mali_c55_params_mesh_shading_selection *params;
+ union mali_c55_params_block block;
+ struct mali_c55 *mali_c55 = arg;
+
+ block.header = header;
+ params = block.shading_selection;
+
+ if (block.header->flags & V4L2_PARAMS_FL_BLOCK_DISABLE)
+ return;
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA_BANK,
+ MALI_C55_MESH_SHADING_ALPHA_BANK_R_MASK,
+ params->mesh_alpha_bank_r);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA_BANK,
+ MALI_C55_MESH_SHADING_ALPHA_BANK_G_MASK,
+ MALI_C55_MESH_SHADING_ALPHA_BANK_G(params->mesh_alpha_bank_g));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA_BANK,
+ MALI_C55_MESH_SHADING_ALPHA_BANK_B_MASK,
+ MALI_C55_MESH_SHADING_ALPHA_BANK_B(params->mesh_alpha_bank_b));
+
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA,
+ MALI_C55_MESH_SHADING_ALPHA_R_MASK,
+ params->mesh_alpha_r);
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA,
+ MALI_C55_MESH_SHADING_ALPHA_G_MASK,
+ MALI_C55_MESH_SHADING_ALPHA_G(params->mesh_alpha_g));
+ mali_c55_ctx_update_bits(mali_c55, MALI_C55_REG_MESH_SHADING_ALPHA,
+ MALI_C55_MESH_SHADING_ALPHA_B_MASK,
+ MALI_C55_MESH_SHADING_ALPHA_B(params->mesh_alpha_b));
+
+ mali_c55_ctx_update_bits(mali_c55,
+ MALI_C55_REG_MESH_SHADING_MESH_STRENGTH,
+ MALI_c55_MESH_STRENGTH_MASK,
+ params->mesh_strength);
+}
+
+static const struct v4l2_params_handler mali_c55_block_handlers[] = {
+ [MALI_C55_PARAM_BLOCK_SENSOR_OFFS] = {
+ .size = sizeof(struct mali_c55_params_sensor_off_preshading),
+ .handler = &mali_c55_params_sensor_offs,
+ },
+ [MALI_C55_PARAM_BLOCK_AEXP_HIST] = {
+ .size = sizeof(struct mali_c55_params_aexp_hist),
+ .handler = &mali_c55_params_aexp_hist,
+ },
+ [MALI_C55_PARAM_BLOCK_AEXP_IHIST] = {
+ .size = sizeof(struct mali_c55_params_aexp_hist),
+ .handler = &mali_c55_params_aexp_hist,
+ },
+ [MALI_C55_PARAM_BLOCK_AEXP_HIST_WEIGHTS] = {
+ .size = sizeof(struct mali_c55_params_aexp_weights),
+ .handler = &mali_c55_params_aexp_hist_weights,
+ },
+ [MALI_C55_PARAM_BLOCK_AEXP_IHIST_WEIGHTS] = {
+ .size = sizeof(struct mali_c55_params_aexp_weights),
+ .handler = &mali_c55_params_aexp_hist_weights,
+ },
+ [MALI_C55_PARAM_BLOCK_DIGITAL_GAIN] = {
+ .size = sizeof(struct mali_c55_params_digital_gain),
+ .handler = &mali_c55_params_digital_gain,
+ },
+ [MALI_C55_PARAM_BLOCK_AWB_GAINS] = {
+ .size = sizeof(struct mali_c55_params_awb_gains),
+ .handler = &mali_c55_params_awb_gains,
+ },
+ [MALI_C55_PARAM_BLOCK_AWB_CONFIG] = {
+ .size = sizeof(struct mali_c55_params_awb_config),
+ .handler = &mali_c55_params_awb_config,
+ },
+ [MALI_C55_PARAM_BLOCK_AWB_GAINS_AEXP] = {
+ .size = sizeof(struct mali_c55_params_awb_gains),
+ .handler = &mali_c55_params_awb_gains,
+ },
+ [MALI_C55_PARAM_MESH_SHADING_CONFIG] = {
+ .size = sizeof(struct mali_c55_params_mesh_shading_config),
+ .handler = &mali_c55_params_lsc_config,
+ },
+ [MALI_C55_PARAM_MESH_SHADING_SELECTION] = {
+ .size = sizeof(struct mali_c55_params_mesh_shading_selection),
+ .handler = &mali_c55_params_lsc_selection,
+ },
+};
+
+static int mali_c55_params_enum_fmt_meta_out(struct file *file, void *fh,
+ struct v4l2_fmtdesc *f)
+{
+ if (f->index)
+ return -EINVAL;
+
+ if (f->mbus_code && f->mbus_code != MEDIA_BUS_FMT_METADATA_FIXED)
+ return -EINVAL;
+
+ f->pixelformat = V4L2_META_FMT_MALI_C55_PARAMS;
+
+ return 0;
+}
+
+static int mali_c55_params_g_fmt_meta_out(struct file *file, void *fh,
+ struct v4l2_format *f)
+{
+ static const struct v4l2_meta_format mfmt = {
+ .dataformat = V4L2_META_FMT_MALI_C55_PARAMS,
+ .buffersize = v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
+ };
+
+ f->fmt.meta = mfmt;
+
+ return 0;
+}
+
+static int mali_c55_params_querycap(struct file *file,
+ void *priv, struct v4l2_capability *cap)
+{
+ strscpy(cap->driver, MALI_C55_DRIVER_NAME, sizeof(cap->driver));
+ strscpy(cap->card, "ARM Mali-C55 ISP", sizeof(cap->card));
+
+ return 0;
+}
+
+static const struct v4l2_ioctl_ops mali_c55_params_v4l2_ioctl_ops = {
+ .vidioc_reqbufs = vb2_ioctl_reqbufs,
+ .vidioc_querybuf = vb2_ioctl_querybuf,
+ .vidioc_create_bufs = vb2_ioctl_create_bufs,
+ .vidioc_qbuf = vb2_ioctl_qbuf,
+ .vidioc_expbuf = vb2_ioctl_expbuf,
+ .vidioc_dqbuf = vb2_ioctl_dqbuf,
+ .vidioc_prepare_buf = vb2_ioctl_prepare_buf,
+ .vidioc_streamon = vb2_ioctl_streamon,
+ .vidioc_streamoff = vb2_ioctl_streamoff,
+ .vidioc_enum_fmt_meta_out = mali_c55_params_enum_fmt_meta_out,
+ .vidioc_g_fmt_meta_out = mali_c55_params_g_fmt_meta_out,
+ .vidioc_s_fmt_meta_out = mali_c55_params_g_fmt_meta_out,
+ .vidioc_try_fmt_meta_out = mali_c55_params_g_fmt_meta_out,
+ .vidioc_querycap = mali_c55_params_querycap,
+ .vidioc_subscribe_event = v4l2_ctrl_subscribe_event,
+ .vidioc_unsubscribe_event = v4l2_event_unsubscribe,
+};
+
+static const struct v4l2_file_operations mali_c55_params_v4l2_fops = {
+ .owner = THIS_MODULE,
+ .unlocked_ioctl = video_ioctl2,
+ .open = v4l2_fh_open,
+ .release = vb2_fop_release,
+ .poll = vb2_fop_poll,
+ .mmap = vb2_fop_mmap,
+};
+
+static int
+mali_c55_params_queue_setup(struct vb2_queue *q, unsigned int *num_buffers,
+ unsigned int *num_planes, unsigned int sizes[],
+ struct device *alloc_devs[])
+{
+ if (*num_planes && *num_planes > 1)
+ return -EINVAL;
+
+ if (sizes[0] && sizes[0] < v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE))
+ return -EINVAL;
+
+ *num_planes = 1;
+
+ if (!sizes[0])
+ sizes[0] = v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE);
+
+ return 0;
+}
+
+static int mali_c55_params_buf_init(struct vb2_buffer *vb)
+{
+ struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+ struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
+
+ buf->config = kvmalloc(v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
+ GFP_KERNEL);
+ if (!buf->config)
+ return -ENOMEM;
+
+ return 0;
+}
+
+static void mali_c55_params_buf_cleanup(struct vb2_buffer *vb)
+{
+ struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+ struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
+
+ kvfree(buf->config);
+ buf->config = NULL;
+}
+
+static int
+mali_c55_params_validate_buffer(struct device *dev,
+ const struct v4l2_params_buffer *buffer)
+{
+ /* Only v1 is supported at the moment. */
+ if (buffer->version != MALI_C55_PARAM_BUFFER_V1) {
+ dev_dbg(dev, "Unsupported extensible format version: %u\n",
+ buffer->version);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int mali_c55_params_buf_prepare(struct vb2_buffer *vb)
+{
+ struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
+ struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+ struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
+ struct mali_c55 *mali_c55 = params->mali_c55;
+ struct v4l2_params_buffer *config;
+ int ret;
+
+ ret = v4l2_params_buffer_validate(
+ mali_c55->dev, vb,
+ v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
+ mali_c55_params_validate_buffer);
+ if (ret)
+ return ret;
+
+ /*
+ * Copy the parameters buffer provided by userspace to the internal
+ * scratch buffer. This protects against the chance of userspace making
+ * changed to the buffer content whilst the driver processes it.
+ */
+ config = vb2_plane_vaddr(vb, 0);
+ memcpy(buf->config, config, v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE));
+
+ return v4l2_params_blocks_validate(mali_c55->dev, buf->config,
+ mali_c55_block_handlers,
+ ARRAY_SIZE(mali_c55_block_handlers),
+ NULL);
+}
+
+static void mali_c55_params_buf_queue(struct vb2_buffer *vb)
+{
+ struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
+ struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+ struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
+
+ spin_lock(¶ms->buffers.lock);
+ list_add_tail(&buf->queue, ¶ms->buffers.queue);
+ spin_unlock(¶ms->buffers.lock);
+}
+
+static void mali_c55_params_return_buffers(struct mali_c55_params *params,
+ enum vb2_buffer_state state)
+{
+ struct mali_c55_params_buf *buf, *tmp;
+
+ guard(spinlock)(¶ms->buffers.lock);
+
+ list_for_each_entry_safe(buf, tmp, ¶ms->buffers.queue, queue) {
+ list_del(&buf->queue);
+ vb2_buffer_done(&buf->vb.vb2_buf, state);
+ }
+}
+
+static int mali_c55_params_start_streaming(struct vb2_queue *q,
+ unsigned int count)
+{
+ struct mali_c55_params *params = vb2_get_drv_priv(q);
+ struct mali_c55 *mali_c55 = params->mali_c55;
+ int ret;
+
+ ret = pm_runtime_resume_and_get(mali_c55->dev);
+ if (ret)
+ goto err_return_buffers;
+
+ ret = video_device_pipeline_alloc_start(¶ms->vdev);
+ if (ret)
+ goto err_pm_put;
+
+ ret = video_device_pipeline_started(¶ms->vdev);
+ if (ret < 0)
+ goto err_stop_pipeline;
+
+ return 0;
+
+err_stop_pipeline:
+ video_device_pipeline_stop(¶ms->vdev);
+err_pm_put:
+ pm_runtime_put(mali_c55->dev);
+err_return_buffers:
+ mali_c55_params_return_buffers(params, VB2_BUF_STATE_QUEUED);
+
+ return ret;
+}
+
+static void mali_c55_params_stop_streaming(struct vb2_queue *q)
+{
+ struct mali_c55_params *params = vb2_get_drv_priv(q);
+ struct media_pipeline *pipe;
+
+ pipe = video_device_pipeline(¶ms->vdev);
+ if (mali_c55_pipeline_ready(pipe))
+ media_pipeline_stopped(pipe);
+
+ video_device_pipeline_stop(¶ms->vdev);
+ mali_c55_params_return_buffers(params, VB2_BUF_STATE_ERROR);
+}
+
+static const struct vb2_ops mali_c55_params_vb2_ops = {
+ .queue_setup = mali_c55_params_queue_setup,
+ .buf_init = mali_c55_params_buf_init,
+ .buf_cleanup = mali_c55_params_buf_cleanup,
+ .buf_queue = mali_c55_params_buf_queue,
+ .buf_prepare = mali_c55_params_buf_prepare,
+ .wait_prepare = vb2_ops_wait_prepare,
+ .wait_finish = vb2_ops_wait_finish,
+ .start_streaming = mali_c55_params_start_streaming,
+ .stop_streaming = mali_c55_params_stop_streaming,
+};
+
+void mali_c55_params_write_config(struct mali_c55 *mali_c55)
+{
+ struct mali_c55_params *params = &mali_c55->params;
+ struct v4l2_params_buffer *config;
+ struct mali_c55_params_buf *buf;
+ size_t block_offset = 0;
+ size_t max_offset;
+
+ spin_lock(¶ms->buffers.lock);
+
+ buf = list_first_entry_or_null(¶ms->buffers.queue,
+ struct mali_c55_params_buf, queue);
+ if (buf)
+ list_del(&buf->queue);
+ spin_unlock(¶ms->buffers.lock);
+
+ if (!buf)
+ return;
+
+ buf->vb.sequence = mali_c55->isp.frame_sequence;
+ config = buf->config;
+
+ max_offset = config->data_size - sizeof(struct v4l2_params_block_header);
+
+ /*
+ * Walk the list of parameter blocks and process them. No validation is
+ * done here, as the contents of the config buffer are already checked
+ * when the buffer is queued.
+ */
+ while (block_offset < max_offset) {
+ const struct v4l2_params_handler *block_handler;
+ union mali_c55_params_block block;
+
+ block.data = &config->data[block_offset];
+
+ /* We checked the array index already in .buf_queue() */
+ block_handler = &mali_c55_block_handlers[block.header->type];
+ block_handler->handler(mali_c55, block.header);
+
+ block_offset += block.header->size;
+ }
+
+ vb2_buffer_done(&buf->vb.vb2_buf, VB2_BUF_STATE_DONE);
+}
+
+void mali_c55_unregister_params(struct mali_c55 *mali_c55)
+{
+ struct mali_c55_params *params = &mali_c55->params;
+
+ if (!video_is_registered(¶ms->vdev))
+ return;
+
+ vb2_video_unregister_device(¶ms->vdev);
+ media_entity_cleanup(¶ms->vdev.entity);
+ mutex_destroy(¶ms->lock);
+}
+
+int mali_c55_register_params(struct mali_c55 *mali_c55)
+{
+ struct mali_c55_params *params = &mali_c55->params;
+ struct video_device *vdev = ¶ms->vdev;
+ struct vb2_queue *vb2q = ¶ms->queue;
+ int ret;
+
+ mutex_init(¶ms->lock);
+ INIT_LIST_HEAD(¶ms->buffers.queue);
+ spin_lock_init(¶ms->buffers.lock);
+
+ params->pad.flags = MEDIA_PAD_FL_SOURCE;
+ ret = media_entity_pads_init(¶ms->vdev.entity, 1, ¶ms->pad);
+ if (ret)
+ goto err_destroy_mutex;
+
+ vb2q->type = V4L2_BUF_TYPE_META_OUTPUT;
+ vb2q->io_modes = VB2_MMAP | VB2_DMABUF;
+ vb2q->drv_priv = params;
+ vb2q->mem_ops = &vb2_dma_contig_memops;
+ vb2q->ops = &mali_c55_params_vb2_ops;
+ vb2q->buf_struct_size = sizeof(struct mali_c55_params_buf);
+ vb2q->min_queued_buffers = 1;
+ vb2q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
+ vb2q->lock = ¶ms->lock;
+ vb2q->dev = mali_c55->dev;
+
+ ret = vb2_queue_init(vb2q);
+ if (ret) {
+ dev_err(mali_c55->dev, "params vb2 queue init failed\n");
+ goto err_cleanup_entity;
+ }
+
+ strscpy(params->vdev.name, "mali-c55 3a params",
+ sizeof(params->vdev.name));
+ vdev->release = video_device_release_empty;
+ vdev->fops = &mali_c55_params_v4l2_fops;
+ vdev->ioctl_ops = &mali_c55_params_v4l2_ioctl_ops;
+ vdev->lock = ¶ms->lock;
+ vdev->v4l2_dev = &mali_c55->v4l2_dev;
+ vdev->queue = ¶ms->queue;
+ vdev->device_caps = V4L2_CAP_META_OUTPUT | V4L2_CAP_STREAMING |
+ V4L2_CAP_IO_MC;
+ vdev->vfl_dir = VFL_DIR_TX;
+ video_set_drvdata(vdev, params);
+
+ ret = video_register_device(vdev, VFL_TYPE_VIDEO, -1);
+ if (ret) {
+ dev_err(mali_c55->dev,
+ "failed to register params video device\n");
+ goto err_release_vb2q;
+ }
+
+ params->mali_c55 = mali_c55;
+
+ return 0;
+
+err_release_vb2q:
+ vb2_queue_release(vb2q);
+err_cleanup_entity:
+ media_entity_cleanup(¶ms->vdev.entity);
+err_destroy_mutex:
+ mutex_destroy(¶ms->lock);
+
+ return ret;
+}
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-registers.h b/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
index 8b04aff1990a962ddb7671d2f66f70c0638737a7..f5a148add1c811f8dc49170db67d151bd54a4307 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-registers.h
@@ -159,6 +159,23 @@ enum mali_c55_interrupts {
#define MALI_C55_BAYER_ORDER_GBRG 2
#define MALI_C55_BAYER_ORDER_BGGR 3
+#define MALI_C55_REG_METERING_CONFIG 0x18ed0
+#define MALI_C55_5BIN_HIST_DISABLE_MASK BIT(0)
+#define MALI_C55_5BIN_HIST_SWITCH_MASK GENMASK(2, 1)
+#define MALI_C55_5BIN_HIST_SWITCH(x) ((x) << 1)
+#define MALI_C55_AF_DISABLE_MASK BIT(4)
+#define MALI_C55_AF_SWITCH_MASK BIT(5)
+#define MALI_C55_AWB_DISABLE_MASK BIT(8)
+#define MALI_C55_AWB_SWITCH_MASK BIT(9)
+#define MALI_C55_AWB_SWITCH(x) ((x) << 9)
+#define MALI_C55_AEXP_HIST_DISABLE_MASK BIT(12)
+#define MALI_C55_AEXP_HIST_DISABLE (0x01 << 12)
+#define MALI_C55_AEXP_HIST_SWITCH_MASK GENMASK(14, 13)
+#define MALI_C55_AEXP_HIST_SWITCH(x) ((x) << 13)
+#define MALI_C55_AEXP_IHIST_DISABLE_MASK BIT(16)
+#define MALI_C55_AEXP_IHIST_DISABLE (0x01 << 12)
+#define MALI_C55_AEXP_SRC_MASK BIT(24)
+
#define MALI_C55_REG_TPG_CH0 0x18ed8
#define MALI_C55_TEST_PATTERN_ON_OFF BIT(0)
#define MALI_C55_TEST_PATTERN_RGB_MASK BIT(1)
@@ -179,6 +196,11 @@ enum mali_c55_interrupts {
#define MALI_C55_REG_CONFIG_SPACES_OFFSET 0x0ab6c
#define MALI_C55_CONFIG_SPACE_SIZE 0x1231c
+#define MALI_C55_REG_DIGITAL_GAIN 0x1926c
+#define MALI_C55_DIGITAL_GAIN_MASK GENMASK(12, 0)
+#define MALI_C55_REG_DIGITAL_GAIN_OFFSET 0x19270
+#define MALI_C55_DIGITAL_GAIN_OFFSET_MASK GENMASK(19, 0)
+
#define MALI_C55_REG_SINTER_CONFIG 0x19348
#define MALI_C55_SINTER_VIEW_FILTER_MASK GENMASK(1, 0)
#define MALI_C55_SINTER_SCALE_MODE_MASK GENMASK(3, 2)
@@ -192,6 +214,59 @@ enum mali_c55_interrupts {
#define MALI_C55_TEMPER_DMA_WRITE_ON BIT(0)
#define MALI_C55_TEMPER_DMA_READ_ON BIT(1)
+/* Black Level Correction Configuration */
+#define MALI_C55_REG_SENSOR_OFF_PRE_SHA_00 0x1abcc
+#define MALI_C55_REG_SENSOR_OFF_PRE_SHA_01 0x1abd0
+#define MALI_C55_REG_SENSOR_OFF_PRE_SHA_10 0x1abd4
+#define MALI_C55_REG_SENSOR_OFF_PRE_SHA_11 0x1abd8
+#define MALI_C55_SENSOR_OFF_PRE_SHA_MASK 0xfffff
+
+/* Lens Mesh Shading Configuration */
+#define MALI_C55_REG_MESH_SHADING_TABLES 0x13074
+#define MALI_C55_REG_MESH_SHADING_CONFIG 0x1abfc
+#define MALI_C55_MESH_SHADING_ENABLE_MASK BIT(0)
+#define MALI_C55_MESH_SHADING_MESH_SHOW_MASK BIT(1)
+#define MALI_C55_MESH_SHADING_MESH_SHOW(x) ((x) << 1)
+#define MALI_C55_MESH_SHADING_SCALE_MASK GENMASK(4, 2)
+#define MALI_C55_MESH_SHADING_SCALE(x) ((x) << 2)
+#define MALI_C55_MESH_SHADING_PAGE_R_MASK GENMASK(9, 8)
+#define MALI_C55_MESH_SHADING_PAGE_R(x) ((x) << 8)
+#define MALI_C55_MESH_SHADING_PAGE_G_MASK GENMASK(11, 10)
+#define MALI_C55_MESH_SHADING_PAGE_G(x) ((x) << 10)
+#define MALI_C55_MESH_SHADING_PAGE_B_MASK GENMASK(13, 12)
+#define MALI_C55_MESH_SHADING_PAGE_B(x) ((x) << 12)
+#define MALI_C55_MESH_SHADING_MESH_WIDTH_MASK GENMASK(21, 16)
+#define MALI_C55_MESH_SHADING_MESH_WIDTH(x) ((x) << 16)
+#define MALI_C55_MESH_SHADING_MESH_HEIGHT_MASK GENMASK(29, 24)
+#define MALI_C55_MESH_SHADING_MESH_HEIGHT(x) ((x) << 24)
+
+#define MALI_C55_REG_MESH_SHADING_ALPHA_BANK 0x1ac04
+#define MALI_C55_MESH_SHADING_ALPHA_BANK_R_MASK GENMASK(2, 0)
+#define MALI_C55_MESH_SHADING_ALPHA_BANK_G_MASK GENMASK(5, 3)
+#define MALI_C55_MESH_SHADING_ALPHA_BANK_G(x) ((x) << 3)
+#define MALI_C55_MESH_SHADING_ALPHA_BANK_B_MASK GENMASK(8, 6)
+#define MALI_C55_MESH_SHADING_ALPHA_BANK_B(x) ((x) << 6)
+#define MALI_C55_REG_MESH_SHADING_ALPHA 0x1ac08
+#define MALI_C55_MESH_SHADING_ALPHA_R_MASK GENMASK(7, 0)
+#define MALI_C55_MESH_SHADING_ALPHA_G_MASK GENMASK(15, 8)
+#define MALI_C55_MESH_SHADING_ALPHA_G(x) ((x) << 8)
+#define MALI_C55_MESH_SHADING_ALPHA_B_MASK GENMASK(23, 16)
+#define MALI_C55_MESH_SHADING_ALPHA_B(x) ((x) << 16)
+#define MALI_C55_REG_MESH_SHADING_MESH_STRENGTH 0x1ac0c
+#define MALI_c55_MESH_STRENGTH_MASK GENMASK(15, 0)
+
+/* AWB Gains Configuration */
+#define MALI_C55_REG_AWB_GAINS1 0x1ac10
+#define MALI_C55_AWB_GAIN00_MASK GENMASK(11, 0)
+#define MALI_C55_AWB_GAIN01_MASK GENMASK(27, 16)
+#define MALI_C55_AWB_GAIN01(x) ((x) << 16)
+#define MALI_C55_REG_AWB_GAINS2 0x1ac14
+#define MALI_C55_AWB_GAIN10_MASK GENMASK(11, 0)
+#define MALI_C55_AWB_GAIN11_MASK GENMASK(27, 16)
+#define MALI_C55_AWB_GAIN11(x) ((x) << 16)
+#define MALI_C55_REG_AWB_GAINS1_AEXP 0x1ac18
+#define MALI_C55_REG_AWB_GAINS2_AEXP 0x1ac1c
+
/* Colour Correction Matrix Configuration */
#define MALI_C55_REG_CCM_ENABLE 0x1b07c
#define MALI_C55_CCM_ENABLE_MASK BIT(0)
@@ -214,6 +289,59 @@ enum mali_c55_interrupts {
#define MALI_C55_REG_CCM_ANTIFOG_OFFSET_B 0x1b0c8
#define MALI_C55_CCM_ANTIFOG_OFFSET_MASK GENMASK(11, 0)
+/* AWB Statistics Configuration */
+#define MALI_C55_REG_AWB_STATS_MODE 0x1b29c
+#define MALI_C55_AWB_STATS_MODE_MASK BIT(0)
+#define MALI_C55_REG_AWB_WHITE_LEVEL 0x1b2a0
+#define MALI_C55_AWB_WHITE_LEVEL_MASK GENMASK(9, 0)
+#define MALI_C55_REG_AWB_BLACK_LEVEL 0x1b2a4
+#define MALI_C55_AWB_BLACK_LEVEL_MASK GENMASK(9, 0)
+#define MALI_C55_REG_AWB_CR_MAX 0x1b2a8
+#define MALI_C55_AWB_CR_MAX_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CR_MIN 0x1b2ac
+#define MALI_C55_AWB_CR_MIN_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CB_MAX 0x1b2b0
+#define MALI_C55_AWB_CB_MAX_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CB_MIN 0x1b2b4
+#define MALI_C55_AWB_CB_MIN_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_NODES_USED 0x1b2c4
+#define MALI_C55_AWB_NODES_USED_HORIZ_MASK GENMASK(7, 0)
+#define MALI_C55_AWB_NODES_USED_VERT_MASK GENMASK(15, 8)
+#define MALI_C55_AWB_NODES_USED_VERT(x) ((x) << 8)
+#define MALI_C55_REG_AWB_CR_HIGH 0x1b2c8
+#define MALI_C55_AWB_CR_HIGH_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CR_LOW 0x1b2cc
+#define MALI_C55_AWB_CR_LOW_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CB_HIGH 0x1b2d0
+#define MALI_C55_AWB_CB_HIGH_MASK GENMASK(11, 0)
+#define MALI_C55_REG_AWB_CB_LOW 0x1b2d4
+#define MALI_C55_AWB_CB_LOW_MASK GENMASK(11, 0)
+
+/* AEXP Metering Histogram Configuration */
+#define MALI_C55_REG_AEXP_HIST_BASE 0x1b730
+#define MALI_C55_REG_AEXP_IHIST_BASE 0x1bbac
+#define MALI_C55_AEXP_HIST_SKIP_OFFSET 0
+#define MALI_C55_AEXP_HIST_SKIP_X_MASK GENMASK(2, 0)
+#define MALI_C55_AEXP_HIST_SKIP_X(x) ((x) << 0)
+#define MALI_C55_AEXP_HIST_OFFSET_X_MASK BIT(3)
+#define MALI_C55_AEXP_HIST_OFFSET_X(x) ((x) << 3)
+#define MALI_C55_AEXP_HIST_SKIP_Y_MASK GENMASK(6, 4)
+#define MALI_C55_AEXP_HIST_SKIP_Y(x) ((x) << 4)
+#define MALI_C55_AEXP_HIST_OFFSET_Y_MASK BIT(7)
+#define MALI_C55_AEXP_HIST_OFFSET_Y(x) ((x) << 7)
+#define MALI_C55_AEXP_HIST_SCALE_OFFSET 4
+#define MALI_C55_AEXP_HIST_SCALE_BOTTOM_MASK GENMASK(3, 0)
+#define MALI_C55_AEXP_HIST_SCALE_TOP_MASK GENMASK(7, 4)
+#define MALI_C55_AEXP_HIST_SCALE_TOP(x) ((x) << 4)
+#define MALI_C55_AEXP_HIST_PLANE_MODE_OFFSET 16
+#define MALI_C55_AEXP_HIST_PLANE_MODE_MASK GENMASK(2, 0)
+#define MALI_C55_AEXP_HIST_NODES_USED_OFFSET 52
+#define MALI_C55_AEXP_HIST_NODES_USED_HORIZ_MASK GENMASK(7, 0)
+#define MALI_C55_AEXP_HIST_NODES_USED_VERT_MASK GENMASK(15, 8)
+#define MALI_C55_AEXP_HIST_NODES_USED_VERT(x) ((x) << 8)
+#define MALI_C55_AEXP_HIST_ZONE_WEIGHTS_OFFSET 56
+#define MALI_C55_AEXP_HIST_ZONE_WEIGHT_MASK 0x0f0f0f0f
+
/*
* The Mali-C55 ISP has up to two output pipes; known as full resolution and
* down scaled. The register space for these is laid out identically, but offset
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* [PATCH v11 19/19] Documentation: mali-c55: Document the mali-c55 parameter setting
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
` (16 preceding siblings ...)
2025-07-14 15:06 ` [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node Daniel Scally
@ 2025-07-14 15:06 ` Daniel Scally
17 siblings, 0 replies; 32+ messages in thread
From: Daniel Scally @ 2025-07-14 15:06 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, dan.scally, Sakari Ailus
Document the mali-c55 parameter setting by expanding the relevant
pages in both admin-guide/ and userspace-api/.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Changes in v11:
- Updated the documentation to account for the use of the
V4L2 extensible params
Changes in v10:
- None
Changes in v9:
- None
Changes in v8:
- None
Changes in v7:
- None
Changes in v7:
- None
Changes in v6:
- Minor rewording
Changes in v5:
- New patch
---
Documentation/admin-guide/media/mali-c55.rst | 19 ++++++-
.../media/v4l/metafmt-arm-mali-c55.rst | 66 +++++++++++++++++++++-
2 files changed, 80 insertions(+), 5 deletions(-)
diff --git a/Documentation/admin-guide/media/mali-c55.rst b/Documentation/admin-guide/media/mali-c55.rst
index 7eaeac63ddf713730fac266d81866539310cc7e2..dbed5179d5f84fb69465beee9281ac9fb572a37f 100644
--- a/Documentation/admin-guide/media/mali-c55.rst
+++ b/Documentation/admin-guide/media/mali-c55.rst
@@ -387,9 +387,24 @@ the processing flow the statistics can be drawn from::
+-------------+ | +-------------+
+--> AWB-1
-At present all statistics are drawn from the 0th tap point for each algorithm;
+By default all statistics are drawn from the 0th tap point for each algorithm;
I.E. AEXP statistics from AEXP-0 (A), AWB statistics from AWB-0 and AF
-statistics from AF-0. In the future this will be configurable.
+statistics from AF-0. This is configurable for AEXP and AWB statsistics through
+programming the ISP's parameters.
+
+.. _mali-c55-3a-params:
+
+Programming ISP Parameters
+==========================
+
+The ISP can be programmed with various parameters from userspace to apply to the
+hardware before and during video stream. This allows userspace to dynamically
+change values such as black level, white balance and lens shading gains and so
+on.
+
+The buffer format and how to populate it are described by the
+:ref:`V4L2_META_FMT_MALI_C55_PARAMS <v4l2-meta-fmt-mali-c55-params>` format,
+which should be set as the data format for the `mali-c55 3a params` video node.
References
==========
diff --git a/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst b/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
index 186e0deb9ece70ab1d2b22bb5ccb69196264a793..f21f4f8cff49874821cea7d9ad8d6a41e57d6ff8 100644
--- a/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
+++ b/Documentation/userspace-api/media/v4l/metafmt-arm-mali-c55.rst
@@ -1,10 +1,11 @@
.. SPDX-License-Identifier: GPL-2.0
+.. _v4l2-meta-fmt-mali-c55-params:
.. _v4l2-meta-fmt-mali-c55-3a-stats:
-*************************************
-V4L2_META_FMT_MALI_C55_STATS ('C55S')
-*************************************
+*****************************************************************************
+V4L2_META_FMT_MALI_C55_STATS ('C55S'), V4L2_META_FMT_MALI_C55_PARAMS ('C55P')
+*****************************************************************************
3A Statistics
=============
@@ -23,6 +24,65 @@ of the C structure :c:type:`mali_c55_stats_buffer` defined in
For details of the statistics see :c:type:`mali_c55_stats_buffer`.
+Configuration Parameters
+========================
+
+The configuration parameters are passed to the
+:ref:`mali-c55 3a params <mali-c55-3a-params>` metadata output video node, using
+the :c:type:`v4l2_meta_format` interface. Rather than a single struct containing
+sub-structs for each configurable area of the ISP, parameters for the Mali-C55
+use the V4L2 extensible parameters system, through which groups of parameters
+are defined as distinct structs or "blocks" which may be added to the data
+member of :c:type:`v4l2_params_buffer`. Userspace is responsible for populating
+the data member with the blocks that need to be configured by the driver, but
+need not populate it with **all** the blocks, or indeed with any at all if there
+are no configuration changes to make. Populated blocks **must** be consecutive
+in the buffer. To assist both userspace and the driver in identifying the
+blocks each block-specific struct embeds :c:type:`v4l2_params_block_header` as
+its first member and userspace must populate the type member with a value from
+:c:type:`mali_c55_param_block_type`. Once the blocks have been populated
+into the data buffer, the combined size of all populated blocks shall be set in
+the data_size member of :c:type:`v4l2_params_buffer`. For example:
+
+.. code-block:: c
+
+ struct v4l2_params_buffer *params =
+ (struct v4l2_params_buffer *)buffer;
+
+ params->version = MALI_C55_PARAM_BUFFER_V1;
+ params->data_size = 0;
+
+ void *data = (void *)params->data;
+
+ struct mali_c55_params_awb_gains *gains =
+ (struct mali_c55_params_awb_gains *)data;
+
+ gains->header.type = MALI_C55_PARAM_BLOCK_AWB_GAINS;
+ gains->header.flags |= V4L2_PARAMS_FL_BLOCK_ENABLE;
+ gains->header.size = sizeof(struct mali_c55_params_awb_gains);
+
+ gains->gain00 = 256;
+ gains->gain00 = 256;
+ gains->gain00 = 256;
+ gains->gain00 = 256;
+
+ data += sizeof(struct mali_c55_params_awb_gains);
+ params->data_size += sizeof(struct mali_c55_params_awb_gains);
+
+ struct mali_c55_params_sensor_off_preshading *blc =
+ (struct mali_c55_params_sensor_off_preshading *)data;
+
+ blc->header.type = MALI_C55_PARAM_BLOCK_SENSOR_OFFS;
+ blc->header.flags |= V4L2_PARAMS_FL_BLOCK_ENABLE;
+ blc->header.size = sizeof(struct mali_c55_params_sensor_off_preshading);
+
+ blc->chan00 = 51200;
+ blc->chan01 = 51200;
+ blc->chan10 = 51200;
+ blc->chan11 = 51200;
+
+ params->total_size += sizeof(struct mali_c55_params_sensor_off_preshading);
+
Arm Mali-C55 uAPI data types
============================
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* Re: [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-14 15:06 ` [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55 Daniel Scally
@ 2025-07-14 21:25 ` Rob Herring
2025-07-15 6:28 ` Dan Scally
2025-07-31 7:11 ` [PATCH v11.1 " Daniel Scally
1 sibling, 1 reply; 32+ messages in thread
From: Rob Herring @ 2025-07-14 21:25 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
Sakari Ailus, Krzysztof Kozlowski
On Mon, Jul 14, 2025 at 04:06:33PM +0100, Daniel Scally wrote:
> Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
> Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
> ---
> Changes in v11:
>
> - Dropped in arm,inline_mode property. This is now identical to the
> reviewed version 8, so I have kept the tags on there.
[...]
> + arm,inline_mode:
> + description:
> + The ISP can be either electrically connected to sensor and CSI-2 receiver
> + or driven through a DMA input device. This property declares the ISP as
> + being electrically connected to the source of image data.
> + type: boolean
Except you didn't drop it.
Rob
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-14 21:25 ` Rob Herring
@ 2025-07-15 6:28 ` Dan Scally
2025-07-15 8:53 ` Krzysztof Kozlowski
0 siblings, 1 reply; 32+ messages in thread
From: Dan Scally @ 2025-07-15 6:28 UTC (permalink / raw)
To: Rob Herring
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
Sakari Ailus, Krzysztof Kozlowski
On 14/07/2025 22:25, Rob Herring wrote:
> On Mon, Jul 14, 2025 at 04:06:33PM +0100, Daniel Scally wrote:
>> Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
>>
>> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>> Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
>> Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
>> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
>> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>> ---
>> Changes in v11:
>>
>> - Dropped in arm,inline_mode property. This is now identical to the
>> reviewed version 8, so I have kept the tags on there.
> [...]
>
>> + arm,inline_mode:
>> + description:
>> + The ISP can be either electrically connected to sensor and CSI-2 receiver
>> + or driven through a DMA input device. This property declares the ISP as
>> + being electrically connected to the source of image data.
>> + type: boolean
> Except you didn't drop it.
Argh, how did I miss that. Sorry (and sorry Krzysztof too!). Thanks for the heads up...I'll re-send
this patch.
Thanks
Dan
>
> Rob
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-15 6:28 ` Dan Scally
@ 2025-07-15 8:53 ` Krzysztof Kozlowski
0 siblings, 0 replies; 32+ messages in thread
From: Krzysztof Kozlowski @ 2025-07-15 8:53 UTC (permalink / raw)
To: Dan Scally, Rob Herring
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
Sakari Ailus
On 15/07/2025 08:28, Dan Scally wrote:
>
> On 14/07/2025 22:25, Rob Herring wrote:
>> On Mon, Jul 14, 2025 at 04:06:33PM +0100, Daniel Scally wrote:
>>> Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
>>>
>>> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>> Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
>>> Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
>>> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
>>> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>>> ---
>>> Changes in v11:
>>>
>>> - Dropped in arm,inline_mode property. This is now identical to the
>>> reviewed version 8, so I have kept the tags on there.
>> [...]
>>
>>> + arm,inline_mode:
>>> + description:
>>> + The ISP can be either electrically connected to sensor and CSI-2 receiver
>>> + or driven through a DMA input device. This property declares the ISP as
>>> + being electrically connected to the source of image data.
>>> + type: boolean
>> Except you didn't drop it.
>
>
> Argh, how did I miss that. Sorry (and sorry Krzysztof too!). Thanks for the heads up...I'll re-send
> this patch.
No worries, happens. I appreciate the work.
Best regards,
Krzysztof
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node
2025-07-14 15:06 ` [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node Daniel Scally
@ 2025-07-30 14:36 ` Sakari Ailus
2025-07-30 21:11 ` Dan Scally
0 siblings, 1 reply; 32+ messages in thread
From: Sakari Ailus @ 2025-07-30 14:36 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart
Hi Daniel,
Thanks for the update.
On Mon, Jul 14, 2025 at 04:06:44PM +0100, Daniel Scally wrote:
> +static int
> +mali_c55_params_validate_buffer(struct device *dev,
> + const struct v4l2_params_buffer *buffer)
> +{
> + /* Only v1 is supported at the moment. */
> + if (buffer->version != MALI_C55_PARAM_BUFFER_V1) {
> + dev_dbg(dev, "Unsupported extensible format version: %u\n",
> + buffer->version);
> + return -EINVAL;
> + }
Is there anything else to validate here?
I guess nothing is done with the information other than the value is being
checked above, but if it had an effect on something, one would need to copy
the information to memory not accessible to the user.
> +
> + return 0;
> +}
> +
> +static int mali_c55_params_buf_prepare(struct vb2_buffer *vb)
> +{
> + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
> + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
> + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
> + struct mali_c55 *mali_c55 = params->mali_c55;
> + struct v4l2_params_buffer *config;
> + int ret;
> +
> + ret = v4l2_params_buffer_validate(
> + mali_c55->dev, vb,
> + v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
> + mali_c55_params_validate_buffer);
> + if (ret)
> + return ret;
> +
> + /*
> + * Copy the parameters buffer provided by userspace to the internal
> + * scratch buffer. This protects against the chance of userspace making
> + * changed to the buffer content whilst the driver processes it.
> + */
> + config = vb2_plane_vaddr(vb, 0);
> + memcpy(buf->config, config, v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE));
> +
> + return v4l2_params_blocks_validate(mali_c55->dev, buf->config,
> + mali_c55_block_handlers,
> + ARRAY_SIZE(mali_c55_block_handlers),
> + NULL);
> +}
> +
> +static void mali_c55_params_buf_queue(struct vb2_buffer *vb)
> +{
> + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
> + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
> + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
> +
> + spin_lock(¶ms->buffers.lock);
> + list_add_tail(&buf->queue, ¶ms->buffers.queue);
> + spin_unlock(¶ms->buffers.lock);
> +}
> +
> +static void mali_c55_params_return_buffers(struct mali_c55_params *params,
> + enum vb2_buffer_state state)
> +{
> + struct mali_c55_params_buf *buf, *tmp;
> +
> + guard(spinlock)(¶ms->buffers.lock);
> +
> + list_for_each_entry_safe(buf, tmp, ¶ms->buffers.queue, queue) {
> + list_del(&buf->queue);
> + vb2_buffer_done(&buf->vb.vb2_buf, state);
> + }
> +}
> +
> +static int mali_c55_params_start_streaming(struct vb2_queue *q,
> + unsigned int count)
> +{
> + struct mali_c55_params *params = vb2_get_drv_priv(q);
> + struct mali_c55 *mali_c55 = params->mali_c55;
> + int ret;
> +
> + ret = pm_runtime_resume_and_get(mali_c55->dev);
> + if (ret)
> + goto err_return_buffers;
> +
> + ret = video_device_pipeline_alloc_start(¶ms->vdev);
> + if (ret)
> + goto err_pm_put;
> +
> + ret = video_device_pipeline_started(¶ms->vdev);
> + if (ret < 0)
> + goto err_stop_pipeline;
> +
> + return 0;
> +
> +err_stop_pipeline:
> + video_device_pipeline_stop(¶ms->vdev);
> +err_pm_put:
> + pm_runtime_put(mali_c55->dev);
> +err_return_buffers:
> + mali_c55_params_return_buffers(params, VB2_BUF_STATE_QUEUED);
> +
> + return ret;
> +}
> +
> +static void mali_c55_params_stop_streaming(struct vb2_queue *q)
> +{
> + struct mali_c55_params *params = vb2_get_drv_priv(q);
> + struct media_pipeline *pipe;
> +
> + pipe = video_device_pipeline(¶ms->vdev);
> + if (mali_c55_pipeline_ready(pipe))
> + media_pipeline_stopped(pipe);
> +
> + video_device_pipeline_stop(¶ms->vdev);
> + mali_c55_params_return_buffers(params, VB2_BUF_STATE_ERROR);
> +}
> +
> +static const struct vb2_ops mali_c55_params_vb2_ops = {
> + .queue_setup = mali_c55_params_queue_setup,
> + .buf_init = mali_c55_params_buf_init,
> + .buf_cleanup = mali_c55_params_buf_cleanup,
> + .buf_queue = mali_c55_params_buf_queue,
> + .buf_prepare = mali_c55_params_buf_prepare,
> + .wait_prepare = vb2_ops_wait_prepare,
> + .wait_finish = vb2_ops_wait_finish,
> + .start_streaming = mali_c55_params_start_streaming,
> + .stop_streaming = mali_c55_params_stop_streaming,
> +};
> +
> +void mali_c55_params_write_config(struct mali_c55 *mali_c55)
> +{
> + struct mali_c55_params *params = &mali_c55->params;
> + struct v4l2_params_buffer *config;
> + struct mali_c55_params_buf *buf;
> + size_t block_offset = 0;
> + size_t max_offset;
> +
> + spin_lock(¶ms->buffers.lock);
> +
> + buf = list_first_entry_or_null(¶ms->buffers.queue,
> + struct mali_c55_params_buf, queue);
> + if (buf)
> + list_del(&buf->queue);
> + spin_unlock(¶ms->buffers.lock);
> +
> + if (!buf)
> + return;
> +
> + buf->vb.sequence = mali_c55->isp.frame_sequence;
> + config = buf->config;
> +
> + max_offset = config->data_size - sizeof(struct v4l2_params_block_header);
> +
> + /*
> + * Walk the list of parameter blocks and process them. No validation is
> + * done here, as the contents of the config buffer are already checked
> + * when the buffer is queued.
> + */
> + while (block_offset < max_offset) {
> + const struct v4l2_params_handler *block_handler;
> + union mali_c55_params_block block;
> +
> + block.data = &config->data[block_offset];
> +
> + /* We checked the array index already in .buf_queue() */
Not a lot seems to be done there in terms of validation.
Even if that had been done, you can't trust the buffer contents as it
remains mapped to the userspace.
> + block_handler = &mali_c55_block_handlers[block.header->type];
> + block_handler->handler(mali_c55, block.header);
> +
> + block_offset += block.header->size;
> + }
> +
> + vb2_buffer_done(&buf->vb.vb2_buf, VB2_BUF_STATE_DONE);
> +}
--
Kind regards,
Sakari Ailus
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node
2025-07-30 14:36 ` Sakari Ailus
@ 2025-07-30 21:11 ` Dan Scally
2025-07-30 21:22 ` Sakari Ailus
0 siblings, 1 reply; 32+ messages in thread
From: Dan Scally @ 2025-07-30 21:11 UTC (permalink / raw)
To: Sakari Ailus
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart
Hi Sakari, thanks for the comments
On 30/07/2025 15:36, Sakari Ailus wrote:
> Hi Daniel,
>
> Thanks for the update.
>
> On Mon, Jul 14, 2025 at 04:06:44PM +0100, Daniel Scally wrote:
>> +static int
>> +mali_c55_params_validate_buffer(struct device *dev,
>> + const struct v4l2_params_buffer *buffer)
>> +{
>> + /* Only v1 is supported at the moment. */
>> + if (buffer->version != MALI_C55_PARAM_BUFFER_V1) {
>> + dev_dbg(dev, "Unsupported extensible format version: %u\n",
>> + buffer->version);
>> + return -EINVAL;
>> + }
> Is there anything else to validate here?
Not at this stage; the rest of the validation is done by Jacopo's new framework
>
> I guess nothing is done with the information other than the value is being
> checked above, but if it had an effect on something, one would need to copy
> the information to memory not accessible to the user.
Thats done later in .buf_prepare()
>
>> +
>> + return 0;
>> +}
>> +
>> +static int mali_c55_params_buf_prepare(struct vb2_buffer *vb)
>> +{
>> + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
>> + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
>> + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
>> + struct mali_c55 *mali_c55 = params->mali_c55;
>> + struct v4l2_params_buffer *config;
>> + int ret;
>> +
>> + ret = v4l2_params_buffer_validate(
>> + mali_c55->dev, vb,
>> + v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
>> + mali_c55_params_validate_buffer);
>> + if (ret)
>> + return ret;
>> +
>> + /*
>> + * Copy the parameters buffer provided by userspace to the internal
>> + * scratch buffer. This protects against the chance of userspace making
>> + * changed to the buffer content whilst the driver processes it.
>> + */
>> + config = vb2_plane_vaddr(vb, 0);
>> + memcpy(buf->config, config, v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE));
>> +
>> + return v4l2_params_blocks_validate(mali_c55->dev, buf->config,
>> + mali_c55_block_handlers,
>> + ARRAY_SIZE(mali_c55_block_handlers),
>> + NULL);
>> +}
>> +
>> +static void mali_c55_params_buf_queue(struct vb2_buffer *vb)
>> +{
>> + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
>> + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
>> + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
>> +
>> + spin_lock(¶ms->buffers.lock);
>> + list_add_tail(&buf->queue, ¶ms->buffers.queue);
>> + spin_unlock(¶ms->buffers.lock);
>> +}
>> +
>> +static void mali_c55_params_return_buffers(struct mali_c55_params *params,
>> + enum vb2_buffer_state state)
>> +{
>> + struct mali_c55_params_buf *buf, *tmp;
>> +
>> + guard(spinlock)(¶ms->buffers.lock);
>> +
>> + list_for_each_entry_safe(buf, tmp, ¶ms->buffers.queue, queue) {
>> + list_del(&buf->queue);
>> + vb2_buffer_done(&buf->vb.vb2_buf, state);
>> + }
>> +}
>> +
>> +static int mali_c55_params_start_streaming(struct vb2_queue *q,
>> + unsigned int count)
>> +{
>> + struct mali_c55_params *params = vb2_get_drv_priv(q);
>> + struct mali_c55 *mali_c55 = params->mali_c55;
>> + int ret;
>> +
>> + ret = pm_runtime_resume_and_get(mali_c55->dev);
>> + if (ret)
>> + goto err_return_buffers;
>> +
>> + ret = video_device_pipeline_alloc_start(¶ms->vdev);
>> + if (ret)
>> + goto err_pm_put;
>> +
>> + ret = video_device_pipeline_started(¶ms->vdev);
>> + if (ret < 0)
>> + goto err_stop_pipeline;
>> +
>> + return 0;
>> +
>> +err_stop_pipeline:
>> + video_device_pipeline_stop(¶ms->vdev);
>> +err_pm_put:
>> + pm_runtime_put(mali_c55->dev);
>> +err_return_buffers:
>> + mali_c55_params_return_buffers(params, VB2_BUF_STATE_QUEUED);
>> +
>> + return ret;
>> +}
>> +
>> +static void mali_c55_params_stop_streaming(struct vb2_queue *q)
>> +{
>> + struct mali_c55_params *params = vb2_get_drv_priv(q);
>> + struct media_pipeline *pipe;
>> +
>> + pipe = video_device_pipeline(¶ms->vdev);
>> + if (mali_c55_pipeline_ready(pipe))
>> + media_pipeline_stopped(pipe);
>> +
>> + video_device_pipeline_stop(¶ms->vdev);
>> + mali_c55_params_return_buffers(params, VB2_BUF_STATE_ERROR);
>> +}
>> +
>> +static const struct vb2_ops mali_c55_params_vb2_ops = {
>> + .queue_setup = mali_c55_params_queue_setup,
>> + .buf_init = mali_c55_params_buf_init,
>> + .buf_cleanup = mali_c55_params_buf_cleanup,
>> + .buf_queue = mali_c55_params_buf_queue,
>> + .buf_prepare = mali_c55_params_buf_prepare,
>> + .wait_prepare = vb2_ops_wait_prepare,
>> + .wait_finish = vb2_ops_wait_finish,
>> + .start_streaming = mali_c55_params_start_streaming,
>> + .stop_streaming = mali_c55_params_stop_streaming,
>> +};
>> +
>> +void mali_c55_params_write_config(struct mali_c55 *mali_c55)
>> +{
>> + struct mali_c55_params *params = &mali_c55->params;
>> + struct v4l2_params_buffer *config;
>> + struct mali_c55_params_buf *buf;
>> + size_t block_offset = 0;
>> + size_t max_offset;
>> +
>> + spin_lock(¶ms->buffers.lock);
>> +
>> + buf = list_first_entry_or_null(¶ms->buffers.queue,
>> + struct mali_c55_params_buf, queue);
>> + if (buf)
>> + list_del(&buf->queue);
>> + spin_unlock(¶ms->buffers.lock);
>> +
>> + if (!buf)
>> + return;
>> +
>> + buf->vb.sequence = mali_c55->isp.frame_sequence;
>> + config = buf->config;
>> +
>> + max_offset = config->data_size - sizeof(struct v4l2_params_block_header);
>> +
>> + /*
>> + * Walk the list of parameter blocks and process them. No validation is
>> + * done here, as the contents of the config buffer are already checked
>> + * when the buffer is queued.
>> + */
>> + while (block_offset < max_offset) {
>> + const struct v4l2_params_handler *block_handler;
>> + union mali_c55_params_block block;
>> +
>> + block.data = &config->data[block_offset];
>> +
>> + /* We checked the array index already in .buf_queue() */
> Not a lot seems to be done there in terms of validation.
>
> Even if that had been done, you can't trust the buffer contents as it
> remains mapped to the userspace.
Ah; a hangover of an earlier iteration before using the new v4l2-params helpers; the validation is
done by those functions in .buf_prepare() now and the userspace buffer is memcopied to a scratch buffer.
Thanks
Dan
>> + block_handler = &mali_c55_block_handlers[block.header->type];
>> + block_handler->handler(mali_c55, block.header);
>> +
>> + block_offset += block.header->size;
>> + }
>> +
>> + vb2_buffer_done(&buf->vb.vb2_buf, VB2_BUF_STATE_DONE);
>> +}
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node
2025-07-30 21:11 ` Dan Scally
@ 2025-07-30 21:22 ` Sakari Ailus
0 siblings, 0 replies; 32+ messages in thread
From: Sakari Ailus @ 2025-07-30 21:22 UTC (permalink / raw)
To: Dan Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart
Hi Dan,
On Wed, Jul 30, 2025 at 10:11:22PM +0100, Dan Scally wrote:
> Hi Sakari, thanks for the comments
>
> On 30/07/2025 15:36, Sakari Ailus wrote:
> > Hi Daniel,
> >
> > Thanks for the update.
> >
> > On Mon, Jul 14, 2025 at 04:06:44PM +0100, Daniel Scally wrote:
> > > +static int
> > > +mali_c55_params_validate_buffer(struct device *dev,
> > > + const struct v4l2_params_buffer *buffer)
> > > +{
> > > + /* Only v1 is supported at the moment. */
> > > + if (buffer->version != MALI_C55_PARAM_BUFFER_V1) {
> > > + dev_dbg(dev, "Unsupported extensible format version: %u\n",
> > > + buffer->version);
> > > + return -EINVAL;
> > > + }
> > Is there anything else to validate here?
>
>
> Not at this stage; the rest of the validation is done by Jacopo's new
> framework
>
> >
> > I guess nothing is done with the information other than the value is being
> > checked above, but if it had an effect on something, one would need to copy
> > the information to memory not accessible to the user.
>
>
> Thats done later in .buf_prepare()
Ah, thanks; I missed that earlier.
>
> >
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static int mali_c55_params_buf_prepare(struct vb2_buffer *vb)
> > > +{
> > > + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
> > > + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
> > > + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
> > > + struct mali_c55 *mali_c55 = params->mali_c55;
> > > + struct v4l2_params_buffer *config;
> > > + int ret;
> > > +
> > > + ret = v4l2_params_buffer_validate(
> > > + mali_c55->dev, vb,
> > > + v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE),
> > > + mali_c55_params_validate_buffer);
> > > + if (ret)
> > > + return ret;
> > > +
> > > + /*
> > > + * Copy the parameters buffer provided by userspace to the internal
> > > + * scratch buffer. This protects against the chance of userspace making
> > > + * changed to the buffer content whilst the driver processes it.
> > > + */
> > > + config = vb2_plane_vaddr(vb, 0);
> > > + memcpy(buf->config, config, v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE));
> > > +
> > > + return v4l2_params_blocks_validate(mali_c55->dev, buf->config,
> > > + mali_c55_block_handlers,
> > > + ARRAY_SIZE(mali_c55_block_handlers),
> > > + NULL);
> > > +}
> > > +
> > > +static void mali_c55_params_buf_queue(struct vb2_buffer *vb)
> > > +{
> > > + struct mali_c55_params *params = vb2_get_drv_priv(vb->vb2_queue);
> > > + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
> > > + struct mali_c55_params_buf *buf = to_mali_c55_params_buf(vbuf);
> > > +
> > > + spin_lock(¶ms->buffers.lock);
> > > + list_add_tail(&buf->queue, ¶ms->buffers.queue);
> > > + spin_unlock(¶ms->buffers.lock);
> > > +}
> > > +
> > > +static void mali_c55_params_return_buffers(struct mali_c55_params *params,
> > > + enum vb2_buffer_state state)
> > > +{
> > > + struct mali_c55_params_buf *buf, *tmp;
> > > +
> > > + guard(spinlock)(¶ms->buffers.lock);
> > > +
> > > + list_for_each_entry_safe(buf, tmp, ¶ms->buffers.queue, queue) {
> > > + list_del(&buf->queue);
> > > + vb2_buffer_done(&buf->vb.vb2_buf, state);
> > > + }
> > > +}
> > > +
> > > +static int mali_c55_params_start_streaming(struct vb2_queue *q,
> > > + unsigned int count)
> > > +{
> > > + struct mali_c55_params *params = vb2_get_drv_priv(q);
> > > + struct mali_c55 *mali_c55 = params->mali_c55;
> > > + int ret;
> > > +
> > > + ret = pm_runtime_resume_and_get(mali_c55->dev);
> > > + if (ret)
> > > + goto err_return_buffers;
> > > +
> > > + ret = video_device_pipeline_alloc_start(¶ms->vdev);
> > > + if (ret)
> > > + goto err_pm_put;
> > > +
> > > + ret = video_device_pipeline_started(¶ms->vdev);
> > > + if (ret < 0)
> > > + goto err_stop_pipeline;
> > > +
> > > + return 0;
> > > +
> > > +err_stop_pipeline:
> > > + video_device_pipeline_stop(¶ms->vdev);
> > > +err_pm_put:
> > > + pm_runtime_put(mali_c55->dev);
> > > +err_return_buffers:
> > > + mali_c55_params_return_buffers(params, VB2_BUF_STATE_QUEUED);
> > > +
> > > + return ret;
> > > +}
> > > +
> > > +static void mali_c55_params_stop_streaming(struct vb2_queue *q)
> > > +{
> > > + struct mali_c55_params *params = vb2_get_drv_priv(q);
> > > + struct media_pipeline *pipe;
> > > +
> > > + pipe = video_device_pipeline(¶ms->vdev);
> > > + if (mali_c55_pipeline_ready(pipe))
> > > + media_pipeline_stopped(pipe);
> > > +
> > > + video_device_pipeline_stop(¶ms->vdev);
> > > + mali_c55_params_return_buffers(params, VB2_BUF_STATE_ERROR);
> > > +}
> > > +
> > > +static const struct vb2_ops mali_c55_params_vb2_ops = {
> > > + .queue_setup = mali_c55_params_queue_setup,
> > > + .buf_init = mali_c55_params_buf_init,
> > > + .buf_cleanup = mali_c55_params_buf_cleanup,
> > > + .buf_queue = mali_c55_params_buf_queue,
> > > + .buf_prepare = mali_c55_params_buf_prepare,
> > > + .wait_prepare = vb2_ops_wait_prepare,
> > > + .wait_finish = vb2_ops_wait_finish,
> > > + .start_streaming = mali_c55_params_start_streaming,
> > > + .stop_streaming = mali_c55_params_stop_streaming,
> > > +};
> > > +
> > > +void mali_c55_params_write_config(struct mali_c55 *mali_c55)
> > > +{
> > > + struct mali_c55_params *params = &mali_c55->params;
> > > + struct v4l2_params_buffer *config;
> > > + struct mali_c55_params_buf *buf;
> > > + size_t block_offset = 0;
> > > + size_t max_offset;
> > > +
> > > + spin_lock(¶ms->buffers.lock);
> > > +
> > > + buf = list_first_entry_or_null(¶ms->buffers.queue,
> > > + struct mali_c55_params_buf, queue);
> > > + if (buf)
> > > + list_del(&buf->queue);
> > > + spin_unlock(¶ms->buffers.lock);
> > > +
> > > + if (!buf)
> > > + return;
> > > +
> > > + buf->vb.sequence = mali_c55->isp.frame_sequence;
> > > + config = buf->config;
> > > +
> > > + max_offset = config->data_size - sizeof(struct v4l2_params_block_header);
> > > +
> > > + /*
> > > + * Walk the list of parameter blocks and process them. No validation is
> > > + * done here, as the contents of the config buffer are already checked
> > > + * when the buffer is queued.
> > > + */
> > > + while (block_offset < max_offset) {
> > > + const struct v4l2_params_handler *block_handler;
> > > + union mali_c55_params_block block;
> > > +
> > > + block.data = &config->data[block_offset];
> > > +
> > > + /* We checked the array index already in .buf_queue() */
> > Not a lot seems to be done there in terms of validation.
> >
> > Even if that had been done, you can't trust the buffer contents as it
> > remains mapped to the userspace.
>
> Ah; a hangover of an earlier iteration before using the new v4l2-params
> helpers; the validation is done by those functions in .buf_prepare() now and
> the userspace buffer is memcopied to a scratch buffer.
Ack, I'll take a look at Jacopo's set and get back to this. It looks nice
overall.
--
Kind regards,
Sakari Ailus
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH v11.1 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-14 15:06 ` [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55 Daniel Scally
2025-07-14 21:25 ` Rob Herring
@ 2025-07-31 7:11 ` Daniel Scally
2025-07-31 13:35 ` Rob Herring
1 sibling, 1 reply; 32+ messages in thread
From: Daniel Scally @ 2025-07-31 7:11 UTC (permalink / raw)
To: linux-media, devicetree, linux-arm-kernel
Cc: Anthony.McGivern, jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, sakari.ailus, Daniel Scally,
Krzysztof Kozlowski
Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
---
Resending this patch with the change that I had forgotten to make,
actually made.
Changes in v11.1:
- _Actually_ dropped the arm,inline property mode, having forgotten to
do so in v11.
Changes in v11:
- Dropped in arm,inline_mode property. This is now identical to the
reviewed version 8, so I have kept the tags on there.
Changes in v10:
- None
Changes in v9:
- Added the arm,inline_mode property to differentiate between inline and
memory input configurations
Changes in v8:
- Added the video clock back in. Now that we have actual hardware it's
clear that it's necessary.
- Added reset lines
- Dropped R-bs
Changes in v7:
- None
Changes in v6:
- None
Changes in v5:
- None
Changes in v4:
- Switched to port instead of ports
Changes in v3:
- Dropped the video clock as suggested by Laurent. I didn't retain it
for the purposes of the refcount since this driver will call .s_stream()
for the sensor driver which will refcount the clock anyway.
- Clarified that the port is a parallel input port rather (Sakari)
Changes in v2:
- Added clocks information
- Fixed the warnings raised by Rob
---
.../bindings/media/arm,mali-c55.yaml | 82 +++++++++++++++++++
1 file changed, 82 insertions(+)
create mode 100644 Documentation/devicetree/bindings/media/arm,mali-c55.yaml
diff --git a/Documentation/devicetree/bindings/media/arm,mali-c55.yaml b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
new file mode 100644
index 000000000000..efc88fd2c447
--- /dev/null
+++ b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
@@ -0,0 +1,82 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/media/arm,mali-c55.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: ARM Mali-C55 Image Signal Processor
+
+maintainers:
+ - Daniel Scally <dan.scally@ideasonboard.com>
+ - Jacopo Mondi <jacopo.mondi@ideasonboard.com>
+
+properties:
+ compatible:
+ const: arm,mali-c55
+
+ reg:
+ maxItems: 1
+
+ interrupts:
+ maxItems: 1
+
+ clocks:
+ items:
+ - description: ISP Video Clock
+ - description: ISP AXI clock
+ - description: ISP AHB-lite clock
+
+ clock-names:
+ items:
+ - const: vclk
+ - const: aclk
+ - const: hclk
+
+ resets:
+ items:
+ - description: vclk domain reset
+ - description: aclk domain reset
+ - description: hclk domain reset
+
+ reset-names:
+ items:
+ - const: vresetn
+ - const: aresetn
+ - const: hresetn
+
+ port:
+ $ref: /schemas/graph.yaml#/properties/port
+ description: Input parallel video bus
+
+ properties:
+ endpoint:
+ $ref: /schemas/graph.yaml#/properties/endpoint
+
+required:
+ - compatible
+ - reg
+ - interrupts
+ - clocks
+ - clock-names
+ - port
+
+additionalProperties: false
+
+examples:
+ - |
+ mali_c55: isp@400000 {
+ compatible = "arm,mali-c55";
+ reg = <0x400000 0x200000>;
+ clocks = <&clk 0>, <&clk 1>, <&clk 2>;
+ clock-names = "vclk", "aclk", "hclk";
+ resets = <&resets 0>, <&resets 1>, <&resets 2>;
+ reset-names = "vresetn", "aresetn", "hresetn";
+ interrupts = <0>;
+
+ port {
+ isp_in: endpoint {
+ remote-endpoint = <&csi2_rx_out>;
+ };
+ };
+ };
+...
--
2.34.1
^ permalink raw reply related [flat|nested] 32+ messages in thread
* Re: [PATCH v11.1 07/19] dt-bindings: media: Add bindings for ARM mali-c55
2025-07-31 7:11 ` [PATCH v11.1 " Daniel Scally
@ 2025-07-31 13:35 ` Rob Herring
0 siblings, 0 replies; 32+ messages in thread
From: Rob Herring @ 2025-07-31 13:35 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
sakari.ailus, Krzysztof Kozlowski
On Thu, Jul 31, 2025 at 08:11:51AM +0100, Daniel Scally wrote:
> Add the yaml binding for ARM's Mali-C55 Image Signal Processor.
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> Acked-by: Nayden Kanchev <nayden.kanchev@arm.com>
> Co-developed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
> ---
> Resending this patch with the change that I had forgotten to make,
> actually made.
>
> Changes in v11.1:
Creative versioning like this and sending 1 patch of the series doesn't
work with the tools (i.e. b4). You're going to have to send the whole
thing as v12. But do that after rc1 is out.
Rob
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops
2025-07-14 15:06 ` [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops Daniel Scally
@ 2025-08-05 7:45 ` Jacopo Mondi
2025-08-20 10:47 ` Dan Scally
0 siblings, 1 reply; 32+ messages in thread
From: Jacopo Mondi @ 2025-08-05 7:45 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, Sakari Ailus
Hi Dan
On Mon, Jul 14, 2025 at 04:06:27PM +0100, Daniel Scally wrote:
> Add two new members to struct media_entity_operations, along with new
> functions in media-entity.c to traverse a media pipeline and call the
> new operations. The new functions are intended to be used to signal
> to a media pipeline that it has fully started, with the entity ops
> allowing drivers to define some action to be taken when those
> conditions are met.
>
> The combination of the new functions and operations allows drivers
> which are part of a multi-driver pipeline to delay actually starting
> streaming until all of the conditions for streaming succcessfully are
> met across all drivers.
>
> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
> ---
> Changes in v5:
>
> - Update kerneldoc comments with Optional statement in the
> right place
>
> Changes in v4:
>
> - Reverted to having the iter variable
>
> Changes in v3:
>
> - Dropped the iter variable now that the pipeline entity
> iterator functions don't need it.
> - Updated documentation to specify Optional and return
> values
>
> Changes in v2:
>
> - Refactored media_pipeline_started() such that the cleanup
> function for media_pipeline_entity_iter is unconditionally
> called
> - Avoided using media_entity_call() helper for operation that
> has return type void to avoid compiler warnings
> ---
> drivers/media/mc/mc-entity.c | 46 ++++++++++++++++++++++++++++++++++++++++++++
> include/media/media-entity.h | 29 ++++++++++++++++++++++++++++
> 2 files changed, 75 insertions(+)
>
> diff --git a/drivers/media/mc/mc-entity.c b/drivers/media/mc/mc-entity.c
> index 045590905582054c46656e20463271b1f93fa6b4..d3443537d4304e12cb015630101efba22375c011 100644
> --- a/drivers/media/mc/mc-entity.c
> +++ b/drivers/media/mc/mc-entity.c
> @@ -1053,6 +1053,52 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
> }
> EXPORT_SYMBOL_GPL(__media_pipeline_entity_iter_next);
>
> +int media_pipeline_started(struct media_pipeline *pipe)
> +{
> + struct media_pipeline_entity_iter iter;
> + struct media_entity *entity;
> + int ret;
> +
> + ret = media_pipeline_entity_iter_init(pipe, &iter);
> + if (ret)
> + return ret;
> +
> + media_pipeline_for_each_entity(pipe, &iter, entity) {
> + ret = media_entity_call(entity, pipeline_started);
> + if (ret && ret != -ENOIOCTLCMD)
> + break;
> + }
> +
> + media_pipeline_entity_iter_cleanup(&iter);
> +
> + ret = ret == -ENOIOCTLCMD ? 0 : ret;
> + if (ret)
> + media_pipeline_stopped(pipe);
> +
> + return ret;
> +}
> +EXPORT_SYMBOL_GPL(media_pipeline_started);
> +
> +int media_pipeline_stopped(struct media_pipeline *pipe)
> +{
> + struct media_pipeline_entity_iter iter;
> + struct media_entity *entity;
> + int ret;
> +
> + ret = media_pipeline_entity_iter_init(pipe, &iter);
> + if (ret)
> + return ret;
> +
> + media_pipeline_for_each_entity(pipe, &iter, entity)
> + if (entity->ops && entity->ops->pipeline_stopped)
> + entity->ops->pipeline_stopped(entity);
I was sure I asked this already, but I wasn't able to find any
reference to this in the review of the previous version, so I'll
re-ask (sorry if it's the second time):
why can't you use media_entity_call() here as well ?
> +
> + media_pipeline_entity_iter_cleanup(&iter);
> +
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(media_pipeline_stopped);
> +
> /* -----------------------------------------------------------------------------
> * Links management
> */
> diff --git a/include/media/media-entity.h b/include/media/media-entity.h
> index 64cf590b11343f68a456c5870ca2f32917c122f9..1e1026f65f2050bb9aa39bde68794da8d2d0a669 100644
> --- a/include/media/media-entity.h
> +++ b/include/media/media-entity.h
> @@ -269,6 +269,10 @@ struct media_pad {
> * media_entity_has_pad_interdep().
> * Optional: If the operation isn't implemented all pads
> * will be considered as interdependent.
> + * @pipeline_started: Optional: Notify this entity that the pipeline it is a
> + * part of has been started.
> + * @pipeline_stopped: Optional: Notify this entity that the pipeline it is a
> + * part of has been stopped.
Why not use the same style as the other entries ?
* @get_fwnode_pad: Return the pad number based on a fwnode endpoint or
* a negative value on error. This operation can be used
* to map a fwnode to a media pad number. Optional.
Or
* @has_pad_interdep: Return whether two pads of the entity are
* interdependent. If two pads are interdependent they are
* part of the same pipeline and enabling one of the pads
* means that the other pad will become "locked" and
* doesn't allow configuration changes. pad0 and pad1 are
* guaranteed to not both be sinks or sources. Never call
* the .has_pad_interdep() operation directly, always use
* media_entity_has_pad_interdep().
* Optional: If the operation isn't implemented all pads
* will be considered as interdependent.
Also, the existing doc uses "the entity" and not "this entity"
These would then be
* @pipeline_started: Notify the entity that the pipeline it is a
* part of has been started. Optional.
* @pipeline_stopped: Notify the entity that the pipeline it is a
* part of has been stopped. Optional
Question from a non-native speaker: "it is a part of" or "it is part
of" ?
> *
> * .. note::
> *
> @@ -284,6 +288,8 @@ struct media_entity_operations {
> int (*link_validate)(struct media_link *link);
> bool (*has_pad_interdep)(struct media_entity *entity, unsigned int pad0,
> unsigned int pad1);
> + int (*pipeline_started)(struct media_entity *entity);
> + void (*pipeline_stopped)(struct media_entity *entity);
> };
>
> /**
> @@ -1261,6 +1267,29 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
> entity != NULL; \
> entity = __media_pipeline_entity_iter_next((pipe), iter, entity))
>
> +/**
> + * media_pipeline_started - Inform entities in a pipeline that it has started
> + * @pipe: The pipeline
> + *
> + * Iterate on all entities in a media pipeline and call their pipeline_started
> + * member of media_entity_operations.
> + *
> + * Return: zero on success, or a negative error code passed through from an
> + * entity's .pipeline_started() operation.
If you don't have specific return codes to document you could consider
a simpler
* Returns zero on success or a negative error code otherwise.
Up to you on this one.
With the above documentation aligned to the existing one and a
clarification on the media_entity_call usage:
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Thanks
> + */
> +int media_pipeline_started(struct media_pipeline *pipe);
> +
> +/**
> + * media_pipeline_stopped - Inform entities in a pipeline that it has stopped
> + * @pipe: The pipeline
> + *
> + * Iterate on all entities in a media pipeline and call their pipeline_stopped
> + * member of media_entity_operations.
> + *
> + * Return: zero on success, or -ENOMEM if the iterator initialisation failed.
> + */
> +int media_pipeline_stopped(struct media_pipeline *pipe);
> +
> /**
> * media_pipeline_alloc_start - Mark a pipeline as streaming
> * @pad: Starting pad
>
> --
> 2.34.1
>
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]()
2025-07-14 15:06 ` [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]() Daniel Scally
@ 2025-08-05 8:52 ` Jacopo Mondi
0 siblings, 0 replies; 32+ messages in thread
From: Jacopo Mondi @ 2025-08-05 8:52 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, Sakari Ailus
Hi Dan
On Mon, Jul 14, 2025 at 04:06:28PM +0100, Daniel Scally wrote:
> Add helpers to run the new media_pipeline_started() and
> media_pipeline_stopped() functions. The helpers iterate over the
> entities in the pipeline and count the number of video devices and
> compare that to the pipeline's start_count() before acting. This
> allows us to only run the media pipeline callbacks in the event that
> the pipeline has had video_pipeline_start() called for each video
> device.
>
> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>
> ---
> Changes in v3:
> - Revised the documentation a bit, detailing the specific error
> codes and rewording the descriptions
> - Renamed __video_device_pipeline_started()
> - List the possible return values explicitly
>
> Changes in v2:
>
> - Adapted now media_pipeline_for_each_entity() takes an iter
> variable
> - Fixed the Return: section of the kerneldoc comments
> ---
> drivers/media/v4l2-core/v4l2-dev.c | 57 ++++++++++++++++++++++++++++++++++++++
> include/media/v4l2-dev.h | 52 ++++++++++++++++++++++++++++++++++
> 2 files changed, 109 insertions(+)
>
> diff --git a/drivers/media/v4l2-core/v4l2-dev.c b/drivers/media/v4l2-core/v4l2-dev.c
> index c369235113d98ae26c30a1aa386e7d60d541a66e..a25038f3c9cc9213b840e2b96d1ef49d17d22fb8 100644
> --- a/drivers/media/v4l2-core/v4l2-dev.c
> +++ b/drivers/media/v4l2-core/v4l2-dev.c
> @@ -1200,6 +1200,63 @@ struct media_pipeline *video_device_pipeline(struct video_device *vdev)
> }
> EXPORT_SYMBOL_GPL(video_device_pipeline);
>
> +static int video_device_pipeline_unstarted_vdevs(struct media_pipeline *pipe)
> +{
> + struct media_pipeline_entity_iter iter;
> + unsigned int n_video_devices = 0;
> + struct media_entity *entity;
> + int ret;
> +
> + ret = media_pipeline_entity_iter_init(pipe, &iter);
> + if (ret)
> + return ret;
> +
> + media_pipeline_for_each_entity(pipe, &iter, entity) {
> + if (entity->obj_type == MEDIA_ENTITY_TYPE_VIDEO_DEVICE)
> + n_video_devices++;
> + }
> +
> + media_pipeline_entity_iter_cleanup(&iter);
> +
> + return n_video_devices - pipe->start_count;
> +}
> +
> +int video_device_pipeline_started(struct video_device *vdev)
> +{
> + struct media_pipeline *pipe;
> + int ret;
> +
> + pipe = video_device_pipeline(vdev);
> + if (!pipe)
> + return -ENODEV;
> +
> + ret = video_device_pipeline_unstarted_vdevs(pipe);
> + if (ret)
> + return ret;
> +
> + return media_pipeline_started(pipe);
> +}
> +EXPORT_SYMBOL_GPL(video_device_pipeline_started);
> +
> +int video_device_pipeline_stopped(struct video_device *vdev)
> +{
> + struct media_pipeline *pipe;
> + int ret;
> +
> + pipe = video_device_pipeline(vdev);
> + if (!pipe)
> + return -ENODEV;
> +
> + ret = video_device_pipeline_unstarted_vdevs(pipe);
> + if (ret)
> + return ret;
> +
> + media_pipeline_stopped(pipe);
> +
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(video_device_pipeline_stopped);
> +
> #endif /* CONFIG_MEDIA_CONTROLLER */
>
> /*
> diff --git a/include/media/v4l2-dev.h b/include/media/v4l2-dev.h
> index 1b6222fab24eda96cbe459b435431c01f7259366..45a87a5a76e8782c79da9d4890b063b8f0cbb3d8 100644
> --- a/include/media/v4l2-dev.h
> +++ b/include/media/v4l2-dev.h
> @@ -654,6 +654,58 @@ __must_check int video_device_pipeline_alloc_start(struct video_device *vdev);
> */
> struct media_pipeline *video_device_pipeline(struct video_device *vdev);
>
> +/**
> + * video_device_pipeline_started - Run the pipeline_started() entity operation
> + * for a fully-started media pipeline
> + * @vdev: A video device that's part of the pipeline
> + *
> + * This function checks whether all MEDIA_ENTITY_TYPE_VIDEO_DEVICE entities
> + * connected to a given video device through enabled links, either directly or
> + * indirectly, have been marked as streaming through the use of
> + * video_device_pipeline_start() or one of its equivalent functions. If so,
> + * media_pipeline_started() is called to inform entities in the pipeline of that
> + * fact. The intention is to provide drivers with a mechanism for checking
> + * whether their pipeline is fully ready to start processing data and call the
> + * .pipeline_started() media entity operation on all the entities in the
> + * pipeline if so.
> + *
> + * Return: The number of video devices in the pipeline remaining to be started,
> + * or a negative error number on failure:
> + *
> + * * 0 - Success
> + * * n > 0 - n video devices in the pipeline remain unstarted
> + * * -ENODEV - The video device has no pipeline
> + * * -ENOMEM - failed to allocate pipeline iterator
> + * * n < 0 - Error passed through from a media entity's
> + * .pipeline_started() operation
> + */
> +int video_device_pipeline_started(struct video_device *vdev);
> +
> +/**
> + * video_device_pipeline_stopped - Run the pipeline_stopped() entity operation
> + * for a fully-started media pipeline
> + * @vdev: A video device that's part of the pipeline
> + *
> + * This function checks whether the MEDIA_ENTITY_TYPE_VIDEO_DEVICE entities
> + * connected to a given video device through enabled links, either directly or
> + * indirectly, have been marked as not streaming through the use of
> + * video_device_pipeline_stop() or one of its equivalent functions. If none of
> + * the video devices in the pipeline have been stopped, then the function
> + * media_pipeline_stopped() is called. The intention is to provide drivers with
> + * a mechanism for checking whether this video device is the first device in the
> + * pipeline to be stopped and call the .pipeline_stopped() media entity
> + * operation on all the entities in the pipeline if so.
> + *
> + * Return: The number of video devices in the pipeline remaining to be started,
> + * or a negative error number on failure:
> + *
> + * * 0 - Success
> + * * n > 0 - n video devices in the pipeline already stopped
> + * * -ENODEV - The video device has no pipeline
> + * * -ENOMEM - failed to allocate pipeline iterator
> + */
> +int video_device_pipeline_stopped(struct video_device *vdev);
I still have troubles here,
media_pipeline_stopped() is documented as
* media_pipeline_stopped - Inform entities in a pipeline that it has stopped
and the entity operation pipeline_stopped() is documented as
* @pipeline_stopped: Optional: Notify this entity that the pipeline it is a
* part of has been stopped.
However video_device_pipeline_stopped() calls media_pipeline_stopped()
(and consequentially the entities' .pipeline_stopped() op) on the first call
to streamoff on any video devices in the pipeline.
At this point what confuses me is:
- If we're on the first call to streamoff, has the pipeline actually
"stopped" ?
- if the fist call to streamoff triggers a pipeline halt, what happens
to the non-yet-stopped video devices ? Will they stop sending
buffers back ?
- I understand a post-pipeline-start hook is desirable, but I fail to
clearly find a use case for this one, that to me looks like a
pre-pipeline-stop hook.
> +
> #endif /* CONFIG_MEDIA_CONTROLLER */
>
> #endif /* _V4L2_DEV_H */
>
> --
> 2.34.1
>
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP
2025-07-14 15:06 ` [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP Daniel Scally
@ 2025-08-05 11:27 ` Jacopo Mondi
2025-08-21 11:14 ` Dan Scally
0 siblings, 1 reply; 32+ messages in thread
From: Jacopo Mondi @ 2025-08-05 11:27 UTC (permalink / raw)
To: Daniel Scally
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
jacopo.mondi, nayden.kanchev, robh+dt, mchehab,
krzysztof.kozlowski+dt, conor+dt, jerome.forissier,
kieran.bingham, laurent.pinchart, Sakari Ailus
Hi Dan
On Mon, Jul 14, 2025 at 04:06:34PM +0100, Daniel Scally wrote:
> Add definitions and documentation for the custom control that will
> be needed by the Mali-C55 ISP driver. This will be a read only
> bitmask of the driver's capabilities, informing userspace of which
> blocks are fitted and which are absent.
>
> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
> ---
> Changes in v11:
>
> - Reserve 16 controls in line with other drivers
>
> Changes in v10:
>
> - None
>
> Changes in v9:
>
> - New patch
> ---
> .../userspace-api/media/drivers/index.rst | 1 +
> .../userspace-api/media/drivers/mali-c55.rst | 55 ++++++++++++++++++++++
> include/uapi/linux/media/arm/mali-c55-config.h | 26 ++++++++++
> include/uapi/linux/v4l2-controls.h | 6 +++
> 4 files changed, 88 insertions(+)
>
> diff --git a/Documentation/userspace-api/media/drivers/index.rst b/Documentation/userspace-api/media/drivers/index.rst
> index d706cb47b1122b6e145a02ab826eb3ecc7997c2b..02967c9b18d6e90f414ccc1329c09bffee895e68 100644
> --- a/Documentation/userspace-api/media/drivers/index.rst
> +++ b/Documentation/userspace-api/media/drivers/index.rst
> @@ -32,6 +32,7 @@ For more details see the file COPYING in the source distribution of Linux.
> cx2341x-uapi
> dw100
> imx-uapi
> + mali-c55
> max2175
> npcm-video
> omap3isp-uapi
> diff --git a/Documentation/userspace-api/media/drivers/mali-c55.rst b/Documentation/userspace-api/media/drivers/mali-c55.rst
> new file mode 100644
> index 0000000000000000000000000000000000000000..e8519da77d737b91a931bbe47920af707eebf110
> --- /dev/null
> +++ b/Documentation/userspace-api/media/drivers/mali-c55.rst
> @@ -0,0 +1,55 @@
> +.. SPDX-License-Identifier: GPL-2.0-only
> +
> +Arm Mali-C55 ISP driver
> +=======================
> +
> +The Arm Mali-C55 ISP driver implements a single driver-specific control:
> +
> +``V4L2_CID_MALI_C55_CAPABILITIES``
> + Detail the capabilities of the ISP by giving detail about the fitted blocks.
This doesn't tell what is the control type. Just add "(bitmask)" to the
control name
``V4L2_CID_MALI_C55_CAPABILITIES (bitmask)``
> +
> + .. flat-table:: Bitmask meaning definitions
> + :header-rows: 1
> + :widths: 2 4 8
> +
> + * - Bit
> + - Macro
> + - Meaning
> + * - 0
> + - MALI_C55_GPS_PONG_FITTED
Do we want "_FITTED" in the control values ?
> + - Pong configuration space is fitted in the ISP
> + * - 1
> + - MALI_C55_GPS_WDR_FITTED
> + - WDR Framestitch, offset and gain is fitted in the ISP
> + * - 2
> + - MALI_C55_GPS_COMPRESSION_FITTED
> + - Temper compression is fitted in the ISP
> + * - 3
> + - MALI_C55_GPS_TEMPER_FITTED
> + - Temper is fitted in the ISP
> + * - 4
> + - MALI_C55_GPS_SINTER_LITE_FITTED
> + - Sinter Lite is fitted in the ISP instead of the full Sinter version
> + * - 5
> + - MALI_C55_GPS_SINTER_FITTED
> + - Sinter is fitted in the ISP
> + * - 6
> + - MALI_C55_GPS_IRIDIX_LTM_FITTED
> + - Iridix local tone mappine is fitted in the ISP
> + * - 7
> + - MALI_C55_GPS_IRIDIX_GTM_FITTED
> + - Iridix global tone mapping is fitted in the ISP
> + * - 8
> + - MALI_C55_GPS_CNR_FITTED
> + - Colour noise reduction is fitted in the ISP
> + * - 9
> + - MALI_C55_GPS_FRSCALER_FITTED
> + - The full resolution pipe scaler is fitted in the ISP
> + * - 10
> + - MALI_C55_GPS_DS_PIPE_FITTED
> + - The downscale pipe is fitted in the ISP
> +
> + The Mali-C55 ISP can be configured in a number of ways to include or exclude
> + blocks which may not be necessary. This control provides a way for the
> + driver to communicate to userspace which of the blocks are fitted in the
> + design.
> \ No newline at end of file
> diff --git a/include/uapi/linux/media/arm/mali-c55-config.h b/include/uapi/linux/media/arm/mali-c55-config.h
> new file mode 100644
> index 0000000000000000000000000000000000000000..2bd60a0d78786be368c2e51b1a0a63fd2a5f690b
> --- /dev/null
> +++ b/include/uapi/linux/media/arm/mali-c55-config.h
> @@ -0,0 +1,26 @@
> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
> +/*
> + * ARM Mali-C55 ISP Driver - Userspace API
> + *
> + * Copyright (C) 2023 Ideas on Board Oy
> + */
> +
> +#ifndef __UAPI_MALI_C55_CONFIG_H
> +#define __UAPI_MALI_C55_CONFIG_H
> +
> +#include <linux/v4l2-controls.h>
> +
> +#define V4L2_CID_MALI_C55_CAPABILITIES (V4L2_CID_USER_MALI_C55_BASE + 0x0)
> +#define MALI_C55_GPS_PONG_FITTED BIT(0)
afaik you can't use BIT() in uapi headers. Use (1U << 0)
> +#define MALI_C55_GPS_WDR_FITTED BIT(1)
> +#define MALI_C55_GPS_COMPRESSION_FITTED BIT(2)
> +#define MALI_C55_GPS_TEMPER_FITTED BIT(3)
> +#define MALI_C55_GPS_SINTER_LITE_FITTED BIT(4)
> +#define MALI_C55_GPS_SINTER_FITTED BIT(5)
> +#define MALI_C55_GPS_IRIDIX_LTM_FITTED BIT(6)
> +#define MALI_C55_GPS_IRIDIX_GTM_FITTED BIT(7)
> +#define MALI_C55_GPS_CNR_FITTED BIT(8)
> +#define MALI_C55_GPS_FRSCALER_FITTED BIT(9)
> +#define MALI_C55_GPS_DS_PIPE_FITTED BIT(10)
> +
> +#endif /* __UAPI_MALI_C55_CONFIG_H */
> diff --git a/include/uapi/linux/v4l2-controls.h b/include/uapi/linux/v4l2-controls.h
> index f836512e9debbc65d62a9fe04069b056be42f7b2..9f07d80d7af1dfb347fca6e2c65b25bc2cb86af1 100644
> --- a/include/uapi/linux/v4l2-controls.h
> +++ b/include/uapi/linux/v4l2-controls.h
> @@ -228,6 +228,12 @@ enum v4l2_colorfx {
> */
> #define V4L2_CID_USER_RKISP1_BASE (V4L2_CID_USER_BASE + 0x1220)
>
> +/*
> + * The base for the Arm Mali-C55 ISP driver controls.
> + * We reserve 8 controls for this driver
Doesn't the commit message says 16 ?
> + */
> +#define V4L2_CID_USER_MALI_C55_BASE (V4L2_CID_USER_BASE + 0x1230)
> +
> /* MPEG-class control IDs */
> /* The MPEG controls are applicable to all codec controls
> * and the 'MPEG' part of the define is historical */
>
> --
> 2.34.1
>
>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops
2025-08-05 7:45 ` Jacopo Mondi
@ 2025-08-20 10:47 ` Dan Scally
0 siblings, 0 replies; 32+ messages in thread
From: Dan Scally @ 2025-08-20 10:47 UTC (permalink / raw)
To: Jacopo Mondi
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
nayden.kanchev, robh+dt, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
Sakari Ailus
Hi Jacopo
On 05/08/2025 08:45, Jacopo Mondi wrote:
> Hi Dan
>
> On Mon, Jul 14, 2025 at 04:06:27PM +0100, Daniel Scally wrote:
>> Add two new members to struct media_entity_operations, along with new
>> functions in media-entity.c to traverse a media pipeline and call the
>> new operations. The new functions are intended to be used to signal
>> to a media pipeline that it has fully started, with the entity ops
>> allowing drivers to define some action to be taken when those
>> conditions are met.
>>
>> The combination of the new functions and operations allows drivers
>> which are part of a multi-driver pipeline to delay actually starting
>> streaming until all of the conditions for streaming succcessfully are
>> met across all drivers.
>>
>> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>> ---
>> Changes in v5:
>>
>> - Update kerneldoc comments with Optional statement in the
>> right place
>>
>> Changes in v4:
>>
>> - Reverted to having the iter variable
>>
>> Changes in v3:
>>
>> - Dropped the iter variable now that the pipeline entity
>> iterator functions don't need it.
>> - Updated documentation to specify Optional and return
>> values
>>
>> Changes in v2:
>>
>> - Refactored media_pipeline_started() such that the cleanup
>> function for media_pipeline_entity_iter is unconditionally
>> called
>> - Avoided using media_entity_call() helper for operation that
>> has return type void to avoid compiler warnings
>> ---
>> drivers/media/mc/mc-entity.c | 46 ++++++++++++++++++++++++++++++++++++++++++++
>> include/media/media-entity.h | 29 ++++++++++++++++++++++++++++
>> 2 files changed, 75 insertions(+)
>>
>> diff --git a/drivers/media/mc/mc-entity.c b/drivers/media/mc/mc-entity.c
>> index 045590905582054c46656e20463271b1f93fa6b4..d3443537d4304e12cb015630101efba22375c011 100644
>> --- a/drivers/media/mc/mc-entity.c
>> +++ b/drivers/media/mc/mc-entity.c
>> @@ -1053,6 +1053,52 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
>> }
>> EXPORT_SYMBOL_GPL(__media_pipeline_entity_iter_next);
>>
>> +int media_pipeline_started(struct media_pipeline *pipe)
>> +{
>> + struct media_pipeline_entity_iter iter;
>> + struct media_entity *entity;
>> + int ret;
>> +
>> + ret = media_pipeline_entity_iter_init(pipe, &iter);
>> + if (ret)
>> + return ret;
>> +
>> + media_pipeline_for_each_entity(pipe, &iter, entity) {
>> + ret = media_entity_call(entity, pipeline_started);
>> + if (ret && ret != -ENOIOCTLCMD)
>> + break;
>> + }
>> +
>> + media_pipeline_entity_iter_cleanup(&iter);
>> +
>> + ret = ret == -ENOIOCTLCMD ? 0 : ret;
>> + if (ret)
>> + media_pipeline_stopped(pipe);
>> +
>> + return ret;
>> +}
>> +EXPORT_SYMBOL_GPL(media_pipeline_started);
>> +
>> +int media_pipeline_stopped(struct media_pipeline *pipe)
>> +{
>> + struct media_pipeline_entity_iter iter;
>> + struct media_entity *entity;
>> + int ret;
>> +
>> + ret = media_pipeline_entity_iter_init(pipe, &iter);
>> + if (ret)
>> + return ret;
>> +
>> + media_pipeline_for_each_entity(pipe, &iter, entity)
>> + if (entity->ops && entity->ops->pipeline_stopped)
>> + entity->ops->pipeline_stopped(entity);
>
> I was sure I asked this already, but I wasn't able to find any
> reference to this in the review of the previous version, so I'll
> re-ask (sorry if it's the second time):
>
> why can't you use media_entity_call() here as well ?
It causes compilation errors because media_entity_call() explicitly assumes that the callback will
return an integer, and .pipeline_stopped() returns void.
>
>> +
>> + media_pipeline_entity_iter_cleanup(&iter);
>> +
>> + return 0;
>> +}
>> +EXPORT_SYMBOL_GPL(media_pipeline_stopped);
>> +
>> /* -----------------------------------------------------------------------------
>> * Links management
>> */
>> diff --git a/include/media/media-entity.h b/include/media/media-entity.h
>> index 64cf590b11343f68a456c5870ca2f32917c122f9..1e1026f65f2050bb9aa39bde68794da8d2d0a669 100644
>> --- a/include/media/media-entity.h
>> +++ b/include/media/media-entity.h
>> @@ -269,6 +269,10 @@ struct media_pad {
>> * media_entity_has_pad_interdep().
>> * Optional: If the operation isn't implemented all pads
>> * will be considered as interdependent.
>> + * @pipeline_started: Optional: Notify this entity that the pipeline it is a
>> + * part of has been started.
>> + * @pipeline_stopped: Optional: Notify this entity that the pipeline it is a
>> + * part of has been stopped.
>
> Why not use the same style as the other entries ?
Poor reading comprehension - I thought I had! Thanks, I'll adopt your suggestion.
>
> * @get_fwnode_pad: Return the pad number based on a fwnode endpoint or
> * a negative value on error. This operation can be used
> * to map a fwnode to a media pad number. Optional.
>
> Or
> * @has_pad_interdep: Return whether two pads of the entity are
> * interdependent. If two pads are interdependent they are
> * part of the same pipeline and enabling one of the pads
> * means that the other pad will become "locked" and
> * doesn't allow configuration changes. pad0 and pad1 are
> * guaranteed to not both be sinks or sources. Never call
> * the .has_pad_interdep() operation directly, always use
> * media_entity_has_pad_interdep().
> * Optional: If the operation isn't implemented all pads
> * will be considered as interdependent.
>
> Also, the existing doc uses "the entity" and not "this entity"
>
>
> These would then be
>
> * @pipeline_started: Notify the entity that the pipeline it is a
> * part of has been started. Optional.
> * @pipeline_stopped: Notify the entity that the pipeline it is a
> * part of has been stopped. Optional
>
> Question from a non-native speaker: "it is a part of" or "it is part
> of" ?
I think either would be acceptable; I can switch if it's clearer to you?
>
>> *
>> * .. note::
>> *
>> @@ -284,6 +288,8 @@ struct media_entity_operations {
>> int (*link_validate)(struct media_link *link);
>> bool (*has_pad_interdep)(struct media_entity *entity, unsigned int pad0,
>> unsigned int pad1);
>> + int (*pipeline_started)(struct media_entity *entity);
>> + void (*pipeline_stopped)(struct media_entity *entity);
>> };
>>
>> /**
>> @@ -1261,6 +1267,29 @@ __media_pipeline_entity_iter_next(struct media_pipeline *pipe,
>> entity != NULL; \
>> entity = __media_pipeline_entity_iter_next((pipe), iter, entity))
>>
>> +/**
>> + * media_pipeline_started - Inform entities in a pipeline that it has started
>> + * @pipe: The pipeline
>> + *
>> + * Iterate on all entities in a media pipeline and call their pipeline_started
>> + * member of media_entity_operations.
>> + *
>> + * Return: zero on success, or a negative error code passed through from an
>> + * entity's .pipeline_started() operation.
>
> If you don't have specific return codes to document you could consider
> a simpler
>
> * Returns zero on success or a negative error code otherwise.
>
> Up to you on this one.
>
> With the above documentation aligned to the existing one and a
> clarification on the media_entity_call usage:
>
> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Thanks!
>
> Thanks
>
>> + */
>> +int media_pipeline_started(struct media_pipeline *pipe);
>> +
>> +/**
>> + * media_pipeline_stopped - Inform entities in a pipeline that it has stopped
>> + * @pipe: The pipeline
>> + *
>> + * Iterate on all entities in a media pipeline and call their pipeline_stopped
>> + * member of media_entity_operations.
>> + *
>> + * Return: zero on success, or -ENOMEM if the iterator initialisation failed.
>> + */
>> +int media_pipeline_stopped(struct media_pipeline *pipe);
>> +
>> /**
>> * media_pipeline_alloc_start - Mark a pipeline as streaming
>> * @pad: Starting pad
>>
>> --
>> 2.34.1
>>
>>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP
2025-08-05 11:27 ` Jacopo Mondi
@ 2025-08-21 11:14 ` Dan Scally
0 siblings, 0 replies; 32+ messages in thread
From: Dan Scally @ 2025-08-21 11:14 UTC (permalink / raw)
To: Jacopo Mondi
Cc: linux-media, devicetree, linux-arm-kernel, Anthony.McGivern,
nayden.kanchev, robh+dt, mchehab, krzysztof.kozlowski+dt,
conor+dt, jerome.forissier, kieran.bingham, laurent.pinchart,
Sakari Ailus
Hi Jacopo
On 05/08/2025 12:27, Jacopo Mondi wrote:
> Hi Dan
>
> On Mon, Jul 14, 2025 at 04:06:34PM +0100, Daniel Scally wrote:
>> Add definitions and documentation for the custom control that will
>> be needed by the Mali-C55 ISP driver. This will be a read only
>> bitmask of the driver's capabilities, informing userspace of which
>> blocks are fitted and which are absent.
>>
>> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>> ---
>> Changes in v11:
>>
>> - Reserve 16 controls in line with other drivers
>>
>> Changes in v10:
>>
>> - None
>>
>> Changes in v9:
>>
>> - New patch
>> ---
>> .../userspace-api/media/drivers/index.rst | 1 +
>> .../userspace-api/media/drivers/mali-c55.rst | 55 ++++++++++++++++++++++
>> include/uapi/linux/media/arm/mali-c55-config.h | 26 ++++++++++
>> include/uapi/linux/v4l2-controls.h | 6 +++
>> 4 files changed, 88 insertions(+)
>>
>> diff --git a/Documentation/userspace-api/media/drivers/index.rst b/Documentation/userspace-api/media/drivers/index.rst
>> index d706cb47b1122b6e145a02ab826eb3ecc7997c2b..02967c9b18d6e90f414ccc1329c09bffee895e68 100644
>> --- a/Documentation/userspace-api/media/drivers/index.rst
>> +++ b/Documentation/userspace-api/media/drivers/index.rst
>> @@ -32,6 +32,7 @@ For more details see the file COPYING in the source distribution of Linux.
>> cx2341x-uapi
>> dw100
>> imx-uapi
>> + mali-c55
>> max2175
>> npcm-video
>> omap3isp-uapi
>> diff --git a/Documentation/userspace-api/media/drivers/mali-c55.rst b/Documentation/userspace-api/media/drivers/mali-c55.rst
>> new file mode 100644
>> index 0000000000000000000000000000000000000000..e8519da77d737b91a931bbe47920af707eebf110
>> --- /dev/null
>> +++ b/Documentation/userspace-api/media/drivers/mali-c55.rst
>> @@ -0,0 +1,55 @@
>> +.. SPDX-License-Identifier: GPL-2.0-only
>> +
>> +Arm Mali-C55 ISP driver
>> +=======================
>> +
>> +The Arm Mali-C55 ISP driver implements a single driver-specific control:
>> +
>> +``V4L2_CID_MALI_C55_CAPABILITIES``
>> + Detail the capabilities of the ISP by giving detail about the fitted blocks.
>
> This doesn't tell what is the control type. Just add "(bitmask)" to the
> control name
>
> ``V4L2_CID_MALI_C55_CAPABILITIES (bitmask)``
Ack, thanks
>
>> +
>> + .. flat-table:: Bitmask meaning definitions
>> + :header-rows: 1
>> + :widths: 2 4 8
>> +
>> + * - Bit
>> + - Macro
>> + - Meaning
>> + * - 0
>> + - MALI_C55_GPS_PONG_FITTED
>
> Do we want "_FITTED" in the control values ?
I'll drop the suffix, I basically just copied it from the documentation but it should be clear from
the control name that that's the meaning
>
>> + - Pong configuration space is fitted in the ISP
>> + * - 1
>> + - MALI_C55_GPS_WDR_FITTED
>> + - WDR Framestitch, offset and gain is fitted in the ISP
>> + * - 2
>> + - MALI_C55_GPS_COMPRESSION_FITTED
>> + - Temper compression is fitted in the ISP
>> + * - 3
>> + - MALI_C55_GPS_TEMPER_FITTED
>> + - Temper is fitted in the ISP
>> + * - 4
>> + - MALI_C55_GPS_SINTER_LITE_FITTED
>> + - Sinter Lite is fitted in the ISP instead of the full Sinter version
>> + * - 5
>> + - MALI_C55_GPS_SINTER_FITTED
>> + - Sinter is fitted in the ISP
>> + * - 6
>> + - MALI_C55_GPS_IRIDIX_LTM_FITTED
>> + - Iridix local tone mappine is fitted in the ISP
>> + * - 7
>> + - MALI_C55_GPS_IRIDIX_GTM_FITTED
>> + - Iridix global tone mapping is fitted in the ISP
>> + * - 8
>> + - MALI_C55_GPS_CNR_FITTED
>> + - Colour noise reduction is fitted in the ISP
>> + * - 9
>> + - MALI_C55_GPS_FRSCALER_FITTED
>> + - The full resolution pipe scaler is fitted in the ISP
>> + * - 10
>> + - MALI_C55_GPS_DS_PIPE_FITTED
>> + - The downscale pipe is fitted in the ISP
>> +
>> + The Mali-C55 ISP can be configured in a number of ways to include or exclude
>> + blocks which may not be necessary. This control provides a way for the
>> + driver to communicate to userspace which of the blocks are fitted in the
>> + design.
>> \ No newline at end of file
>> diff --git a/include/uapi/linux/media/arm/mali-c55-config.h b/include/uapi/linux/media/arm/mali-c55-config.h
>> new file mode 100644
>> index 0000000000000000000000000000000000000000..2bd60a0d78786be368c2e51b1a0a63fd2a5f690b
>> --- /dev/null
>> +++ b/include/uapi/linux/media/arm/mali-c55-config.h
>> @@ -0,0 +1,26 @@
>> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
>> +/*
>> + * ARM Mali-C55 ISP Driver - Userspace API
>> + *
>> + * Copyright (C) 2023 Ideas on Board Oy
>> + */
>> +
>> +#ifndef __UAPI_MALI_C55_CONFIG_H
>> +#define __UAPI_MALI_C55_CONFIG_H
>> +
>> +#include <linux/v4l2-controls.h>
>> +
>> +#define V4L2_CID_MALI_C55_CAPABILITIES (V4L2_CID_USER_MALI_C55_BASE + 0x0)
>> +#define MALI_C55_GPS_PONG_FITTED BIT(0)
>
> afaik you can't use BIT() in uapi headers. Use (1U << 0)
Ack
>
>> +#define MALI_C55_GPS_WDR_FITTED BIT(1)
>> +#define MALI_C55_GPS_COMPRESSION_FITTED BIT(2)
>> +#define MALI_C55_GPS_TEMPER_FITTED BIT(3)
>> +#define MALI_C55_GPS_SINTER_LITE_FITTED BIT(4)
>> +#define MALI_C55_GPS_SINTER_FITTED BIT(5)
>> +#define MALI_C55_GPS_IRIDIX_LTM_FITTED BIT(6)
>> +#define MALI_C55_GPS_IRIDIX_GTM_FITTED BIT(7)
>> +#define MALI_C55_GPS_CNR_FITTED BIT(8)
>> +#define MALI_C55_GPS_FRSCALER_FITTED BIT(9)
>> +#define MALI_C55_GPS_DS_PIPE_FITTED BIT(10)
>> +
>> +#endif /* __UAPI_MALI_C55_CONFIG_H */
>> diff --git a/include/uapi/linux/v4l2-controls.h b/include/uapi/linux/v4l2-controls.h
>> index f836512e9debbc65d62a9fe04069b056be42f7b2..9f07d80d7af1dfb347fca6e2c65b25bc2cb86af1 100644
>> --- a/include/uapi/linux/v4l2-controls.h
>> +++ b/include/uapi/linux/v4l2-controls.h
>> @@ -228,6 +228,12 @@ enum v4l2_colorfx {
>> */
>> #define V4L2_CID_USER_RKISP1_BASE (V4L2_CID_USER_BASE + 0x1220)
>>
>> +/*
>> + * The base for the Arm Mali-C55 ISP driver controls.
>> + * We reserve 8 controls for this driver
>
> Doesn't the commit message says 16 ?
Yes...that was my intention; thanks!
>
>> + */
>> +#define V4L2_CID_USER_MALI_C55_BASE (V4L2_CID_USER_BASE + 0x1230)
>> +
>> /* MPEG-class control IDs */
>> /* The MPEG controls are applicable to all codec controls
>> * and the 'MPEG' part of the define is historical */
>>
>> --
>> 2.34.1
>>
>>
^ permalink raw reply [flat|nested] 32+ messages in thread
end of thread, other threads:[~2025-08-21 15:30 UTC | newest]
Thread overview: 32+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-07-14 15:06 [PATCH v11 00/19] Add Arm Mali-C55 Image Signal Processor Driver Daniel Scally
2025-07-14 15:06 ` [PATCH v11 01/19] media: mc: entity: Add pipeline_started/stopped ops Daniel Scally
2025-08-05 7:45 ` Jacopo Mondi
2025-08-20 10:47 ` Dan Scally
2025-07-14 15:06 ` [PATCH v11 02/19] media: v4l2-dev: Add helpers to run media_pipeline_[started|stopped]() Daniel Scally
2025-08-05 8:52 ` Jacopo Mondi
2025-07-14 15:06 ` [PATCH v11 03/19] media: uapi: Add MEDIA_BUS_FMT_RGB202020_1X60 format code Daniel Scally
2025-07-14 15:06 ` [PATCH v11 04/19] media: uapi: Add 20-bit bayer formats Daniel Scally
2025-07-14 15:06 ` [PATCH v11 05/19] media: v4l2-common: Add RAW16 format info Daniel Scally
2025-07-14 15:06 ` [PATCH v11 06/19] media: v4l2-common: Add RAW14 " Daniel Scally
2025-07-14 15:06 ` [PATCH v11 07/19] dt-bindings: media: Add bindings for ARM mali-c55 Daniel Scally
2025-07-14 21:25 ` Rob Herring
2025-07-15 6:28 ` Dan Scally
2025-07-15 8:53 ` Krzysztof Kozlowski
2025-07-31 7:11 ` [PATCH v11.1 " Daniel Scally
2025-07-31 13:35 ` Rob Herring
2025-07-14 15:06 ` [PATCH v11 08/19] media: uapi: Add controls for Mali-C55 ISP Daniel Scally
2025-08-05 11:27 ` Jacopo Mondi
2025-08-21 11:14 ` Dan Scally
2025-07-14 15:06 ` [PATCH v11 10/19] media: Documentation: Add Mali-C55 ISP Documentation Daniel Scally
2025-07-14 15:06 ` [PATCH v11 11/19] MAINTAINERS: Add entry for mali-c55 driver Daniel Scally
2025-07-14 15:06 ` [PATCH v11 12/19] media: Add MALI_C55_3A_STATS meta format Daniel Scally
2025-07-14 15:06 ` [PATCH v11 13/19] media: uapi: Add 3a stats buffer for mali-c55 Daniel Scally
2025-07-14 15:06 ` [PATCH v11 14/19] media: platform: Add mali-c55 3a stats devnode Daniel Scally
2025-07-14 15:06 ` [PATCH v11 15/19] Documentation: mali-c55: Add Statistics documentation Daniel Scally
2025-07-14 15:06 ` [PATCH v11 16/19] media: mali-c55: Add image formats for Mali-C55 parameters buffer Daniel Scally
2025-07-14 15:06 ` [PATCH v11 17/19] media: uapi: Add parameters structs to mali-c55-config.h Daniel Scally
2025-07-14 15:06 ` [PATCH v11 18/19] media: platform: Add mali-c55 parameters video node Daniel Scally
2025-07-30 14:36 ` Sakari Ailus
2025-07-30 21:11 ` Dan Scally
2025-07-30 21:22 ` Sakari Ailus
2025-07-14 15:06 ` [PATCH v11 19/19] Documentation: mali-c55: Document the mali-c55 parameter setting Daniel Scally
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).