Devicetree
 help / color / mirror / Atom feed
From: Atanas Filipov <atanas.filipov@oss.qualcomm.com>
To: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Cc: linux-media@vger.kernel.org, mchehab@kernel.org, robh@kernel.org,
	krzysztof.kozlowski+dt@linaro.org, conor+dt@kernel.org,
	andersson@kernel.org, quic_vgarodia@quicinc.com,
	quic_jesszhan@quicinc.com, linux-arm-msm@vger.kernel.org,
	devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 4/4] media: qcom: jpeg: Add Qualcomm JPEG V4L2 encoder
Date: Tue, 30 Jun 2026 17:14:14 +0300	[thread overview]
Message-ID: <2a7cd484-ee0f-40bd-bba7-3faa86bb8247@oss.qualcomm.com> (raw)
In-Reply-To: <ts6esusuua2khweva6iyfel2yowislwwx6zb2fl2azcefownie@cg7jegn334fj>

On 6/30/2026 4:06 PM, Dmitry Baryshkov wrote:
> On Mon, Jun 29, 2026 at 03:17:50PM +0300, Atanas Filipov wrote:
>> Add a Qualcomm JPEG encoder driver implemented on top of the
>> V4L2 mem2mem framework.
>>
>> The driver wires vb2 queue handling, format negotiation, JPEG header
>> handling, interrupt-driven job completion, and runtime PM/clock/ICC
>> integration for the standalone JPEG encode hardware block.
>>
>> This series targets SM8250 (Kona) platforms.
>>
>> The jpeg-encoder node is described as a child node of the CAMSS block
>> and is probed automatically via of_platform_populate() in camss_probe().
>>
>> Signed-off-by: Atanas Filipov <atanas.filipov@oss.qualcomm.com>
>>
>> media: qcom: jpeg: restore JPEG_ENCODE_H1V1 for CAPTURE format lookup
>>
>> V4L2_PIX_FMT_JPEG must remain in jpeg_encode_fmt[] with JPEG_ENCODE_H1V1
>> so that jpeg_get_encode_fmt() and jpeg_get_memory_fmt() resolve correctly
>> for the Write Engine (CAPTURE queue) path.
>>
>> The corresponding jpeg_mcu_blocks[H1V1] entry is also restored.
>>
>> V4L2_PIX_FMT_JPEG is not present in jpeg_src_formats[] so it cannot be
>> set as a SOURCE format by userspace.
>>
>> Signed-off-by: Atanas Filipov <atanas.filipov@oss.qualcomm.com>
>>
>> fixup! media: qcom: jpeg: Add Qualcomm JPEG V4L2 encoder
>>
>> Signed-off-by: Atanas Filipov <atanas.filipov@oss.qualcomm.com>
> 
> You definitely haven't read your commit message. Sad.
> 
>> ---
>>   drivers/media/platform/qcom/Kconfig           |    1 +
>>   drivers/media/platform/qcom/Makefile          |    1 +
>>   drivers/media/platform/qcom/jpeg/Kconfig      |   17 +
>>   drivers/media/platform/qcom/jpeg/Makefile     |    9 +
>>   .../media/platform/qcom/jpeg/qcom_jenc_defs.h |  310 ++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_dev.c  |  337 ++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_dev.h  |  111 ++
>>   .../media/platform/qcom/jpeg/qcom_jenc_hdr.c  |  354 ++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_hdr.h  |  119 ++
>>   .../media/platform/qcom/jpeg/qcom_jenc_ops.c  | 1524 +++++++++++++++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_ops.h  |   52 +
>>   .../media/platform/qcom/jpeg/qcom_jenc_res.c  |   39 +
>>   .../media/platform/qcom/jpeg/qcom_jenc_res.h  |   30 +
>>   .../qcom/jpeg/qcom_jenc_v420_hw_info.h        |  527 ++++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_v4l2.c | 1150 +++++++++++++
>>   .../media/platform/qcom/jpeg/qcom_jenc_v4l2.h |   25 +
>>   16 files changed, 4606 insertions(+)
>>   create mode 100644 drivers/media/platform/qcom/jpeg/Kconfig
>>   create mode 100644 drivers/media/platform/qcom/jpeg/Makefile
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_defs.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_dev.c
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_dev.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_hdr.c
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_hdr.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_ops.c
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_ops.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_res.c
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_res.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_v420_hw_info.h
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_v4l2.c
>>   create mode 100644 drivers/media/platform/qcom/jpeg/qcom_jenc_v4l2.h
>>
>> diff --git a/drivers/media/platform/qcom/Kconfig b/drivers/media/platform/qcom/Kconfig
>> index 4f4d3a68e6e5..f33d53a754a0 100644
>> --- a/drivers/media/platform/qcom/Kconfig
>> +++ b/drivers/media/platform/qcom/Kconfig
>> @@ -5,3 +5,4 @@ comment "Qualcomm media platform drivers"
>>   source "drivers/media/platform/qcom/camss/Kconfig"
>>   source "drivers/media/platform/qcom/iris/Kconfig"
>>   source "drivers/media/platform/qcom/venus/Kconfig"
>> +source "drivers/media/platform/qcom/jpeg/Kconfig"
>> diff --git a/drivers/media/platform/qcom/Makefile b/drivers/media/platform/qcom/Makefile
>> index ea2221a202c0..011b5d237193 100644
>> --- a/drivers/media/platform/qcom/Makefile
>> +++ b/drivers/media/platform/qcom/Makefile
>> @@ -1,4 +1,5 @@
>>   # SPDX-License-Identifier: GPL-2.0-only
>>   obj-y += camss/
>>   obj-y += iris/
>> +obj-y += jpeg/
>>   obj-y += venus/
>> diff --git a/drivers/media/platform/qcom/jpeg/Kconfig b/drivers/media/platform/qcom/jpeg/Kconfig
>> new file mode 100644
>> index 000000000000..8a96b37b919b
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/jpeg/Kconfig
>> @@ -0,0 +1,17 @@
>> +# SPDX-License-Identifier: GPL-2.0-only
>> +config VIDEO_QCOM_JENC
>> +	tristate "Qualcomm V4L2 JPEG Encoder driver"
>> +	depends on V4L_MEM2MEM_DRIVERS
>> +	depends on (ARCH_QCOM && IOMMU_DMA) || COMPILE_TEST
>> +	depends on VIDEO_DEV
>> +	select VIDEOBUF2_DMA_SG
>> +	select V4L2_JPEG_HELPER
>> +	select V4L2_MEM2MEM_DEV
>> +	help
>> +	  Qualcomm JPEG memory-to-memory V4L2 encoder driver.
>> +
>> +	  Provides:
>> +	    - qcom-jenc (encode)
> 
> What does this mean? How is it different from the next sentence?
> 
>> +
>> +	  To compile this driver as a module, choose M here: the
>> +	  module will be called qcom-jenc
>> diff --git a/drivers/media/platform/qcom/jpeg/Makefile b/drivers/media/platform/qcom/jpeg/Makefile
>> new file mode 100644
>> index 000000000000..310f6c3c1f19
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/jpeg/Makefile
>> @@ -0,0 +1,9 @@
>> +# SPDX-License-Identifier: GPL-2.0-only
>> +obj-$(CONFIG_VIDEO_QCOM_JENC) += qcom-jenc.o
>> +
>> +qcom-jenc-objs += \
>> +	qcom_jenc_dev.o \
>> +	qcom_jenc_v4l2.o \
>> +	qcom_jenc_ops.o \
>> +	qcom_jenc_res.o \
>> +	qcom_jenc_hdr.o
>> diff --git a/drivers/media/platform/qcom/jpeg/qcom_jenc_defs.h b/drivers/media/platform/qcom/jpeg/qcom_jenc_defs.h
>> new file mode 100644
>> index 000000000000..cd6216fc3424
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/jpeg/qcom_jenc_defs.h
>> @@ -0,0 +1,310 @@
>> +/* SPDX-License-Identifier: GPL-2.0-only */
>> +/*
>> + * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
>> + */
>> +
>> +#ifndef QCOM_JENC_DEFS_H
>> +#define QCOM_JENC_DEFS_H
>> +
>> +#include <linux/bitfield.h>
>> +#include <linux/io.h>
>> +#include <linux/types.h>
>> +#include <linux/videodev2.h>
>> +#include <media/videobuf2-core.h>
> 
> Please keep only the headers which are required for this header file.
> All other headers should be included from the source files.
> 
>> +
>> +/* Offline JPEG encoder constraints */
>> +#define QCOM_JPEG_HW_MAX_WIDTH	8192
>> +#define QCOM_JPEG_HW_MAX_HEIGHT	8192
>> +#define QCOM_JPEG_HW_MIN_WIDTH	256
>> +#define QCOM_JPEG_HW_MIN_HEIGHT	256
>> +
>> +#define QCOM_JPEG_HW_DEF_HSTEP	16
>> +#define QCOM_JPEG_HW_DEF_VSTEP	16
>> +
>> +#define QCOM_JPEG_HW_DEF_WIDTH	1920
>> +#define QCOM_JPEG_HW_DEF_HEIGHT	1088
>> +
>> +#define QCOM_JPEG_MAX_PLANES	3
>> +
>> +#define QCOM_JPEG_QUALITY_MIN	1
>> +#define QCOM_JPEG_QUALITY_DEF	98
>> +#define QCOM_JPEG_QUALITY_MAX	100
>> +#define QCOM_JPEG_QUALITY_MID	(QCOM_JPEG_QUALITY_MAX / 2)
>> +#define QCOM_JPEG_QUALITY_UNT	1
>> +
>> +/*
>> + * V4L2_CID_QCOM_JPEG_PERF_LEVEL_AUTO - enable adaptive performance scaling.
>> + *
>> + * When set to 1 the driver selects the core clock OPP level based on the
>> + * encoded frame resolution and quality factor.  When set to 0 (default) the
>> + * driver always runs at NOMINAL (highest) OPP level, trading power for
>> + * deterministic latency.
>> + *
>> + * Accessible via v4l2-ctl:
>> + *   v4l2-ctl --set-ctrl=perf_level_auto=1
>> + */
>> +#define V4L2_CID_QCOM_JPEG_PERF_LEVEL_AUTO \
>> +	(V4L2_CID_USER_BASE + 0x1240)
>> +
>> +/*
>> + * V4L2_CID_QCOM_JPEG_FPS_TARGET - target encode rate in frames per second.
>> + *
>> + * Used together with V4L2_CID_QCOM_JPEG_PERF_LEVEL_AUTO to select the lowest
>> + * OPP level whose throughput is sufficient for the requested frame rate.
>> + * Has no effect when perf_level_auto is 0.
>> + *
>> + * Range: 1 to 240 fps. Default: 30.
>> + *
>> + * Accessible via v4l2-ctl:
>> + *   v4l2-ctl --set-ctrl=fps_target=60
>> + */
>> +#define V4L2_CID_QCOM_JPEG_FPS_TARGET \
>> +	(V4L2_CID_USER_BASE + 0x1241)
> 
> These (and all other relevant defines) should go to include/uapi/
> 
>> +
>> +#define QCOM_JPEG_FPS_MIN	1
>> +#define QCOM_JPEG_FPS_MAX	240
>> +#define QCOM_JPEG_FPS_DEF	30
>> +#define QCOM_JPEG_FPS_UNT	1
>> +
>> +#define JPEG_CLK_LOWSVS_HZ	300000000ULL
>> +#define JPEG_CLK_SVS_HZ		400000000ULL
>> +#define JPEG_CLK_SVS_L1_HZ	480000000ULL
>> +#define JPEG_CLK_NOMINAL_HZ	600000000ULL
> 
> These are platform peculiarities. Move them to the platform-specific
> definitions. Or, better, drop them completely.
> 
>> +
>> +/*
>> + * Reference measurements:
>> + *
>> + *  - JPEG clock      : 600 MHz
>> + *  - Input           : Worst-case (NOISE)
>> + *
>> + * Although ~120 MPixel/s was measured, use a conservative value of
>> + * 110 MPixel/s to provide operating margin.
>> + */
>> +#define JPEG_REF_CLK_HZ             JPEG_CLK_NOMINAL_HZ
> 
> I assume all of these are related to SM8250 and don't apply to other
> devices. Don't make the life of people adding other platforms harder
> than it should be. Remove it and read from the OPP.
> 
>> +#define JPEG_REF_THROUGHPUT_MPPS    110ULL
>> +#define JPEG_REF_PIXEL_RATE         (JPEG_REF_THROUGHPUT_MPPS * 1000000ULL)
>> +
>> +/*
>> + * Performance levels for the JPEG core clock, used as level values.
>> + * The numeric values must match the opp-level entries in the DTS node:
>> + *
>> + *   LOWSVS  = 0  -> opp-level = <0>  (e.g. 300 MHz on SM8250)
>> + *   SVS     = 1  -> opp-level = <1>  (e.g. 400 MHz on SM8250)
>> + *   SVS_L1  = 2  -> opp-level = <2>  (e.g. 480 MHz on SM8250)
>> + *   NOMINAL = 3  -> opp-level = <3>  (e.g. 600 MHz on SM8250)
> 
> Definitely no. Drop this alltogether. Use OPP tables and OPP-related
> functions.
> 
>> + */
>> +enum qcom_soc_perf_level {
>> +	QCOM_SOC_PERF_LOWSVS = 0,
>> +	QCOM_SOC_PERF_SVS,
>> +	QCOM_SOC_PERF_SVS_L1,
>> +	QCOM_SOC_PERF_NOMINAL
>> +};
>> +
>> +/* hardware register field mask identifiers */
>> +enum qcom_jpeg_mask_id {
> 
> Why do you need to keep all of them abstracted? Just use the registers
> as is, assuming that the registers don't change a lot between platforms.
>

Hi Dmitry,

Regarding the register abstraction — I'd like to make a case for keeping it.

The Qualcomm JPEG hardware exists in multiple versions across different 
SoCs. I don't have access to all of them, but Qualcomm tends to reuse 
and recombine IP blocks across platforms, and assuming register-level 
uniformity seems premature.

Adding a new platform only requires a new ops/mask table — existing ones 
remain untouched. Would you be open to keeping this abstraction as-is?

Thanks,
Atanas

>> +	/* hW version fields */
>> +	JMSK_HW_VER_STEP,
>> +	JMSK_HW_VER_MINOR,
>> +	JMSK_HW_VER_MAJOR,
>> +
>> +	/* hW capability fields */
>> +	JMSK_HW_CAP_ENCODE,
>> +	JMSK_HW_CAP_DECODE,
>> +	JMSK_HW_CAP_UPSCALE,
>> +	JMSK_HW_CAP_DOWNSCALE,
>> +
>> +	/* reset command fields for individual JPEG subsystems */
>> +	JMSK_RST_CMD_COMMON,
>> +	JMSK_RST_CMD_FE_RESET,
>> +	JMSK_RST_CMD_WE_RESET,
>> +	JMSK_RST_CMD_ENCODER_RESET,
>> +	JMSK_RST_CMD_DECODER_RESET,
>> +	JMSK_RST_CMD_BLOCK_FORMATTER_RST,
>> +	JMSK_RST_CMD_SCALE_RESET,
>> +	JMSK_RST_CMD_REGISTER_RESET,
>> +	JMSK_RST_CMD_MISR_RESET,
>> +	JMSK_RST_CMD_CORE_RESET,
>> +	JMSK_RST_CMD_JMSK_DOMAIN_RESET,
>> +	JMSK_RST_CMD_RESET_BYPASS,
>> +
>> +	/* hW command fields */
>> +	JMSK_CMD_HW_START,
>> +	JMSK_CMD_HW_STOP,
>> +	JMSK_CMD_CLR_RD_PLNS_QUEUE,
>> +	JMSK_CMD_CLR_WR_PLNS_QUEUE,
>> +	JMSK_CMD_APPLY_SWC_RD_PARAMS,
>> +
>> +	/* core configuration fields */
>> +	JMSK_CORE_CFG_FE_ENABLE,
>> +	JMSK_CORE_CFG_WE_ENABLE,
>> +	JMSK_CORE_CFG_ENC_ENABLE,
>> +	JMSK_CORE_CFG_SCALE_ENABLE,
>> +	JMSK_CORE_CFG_TESTBUS_ENABLE,
>> +	JMSK_CORE_CFG_MODE,
>> +	JMSK_CORE_CFG_CGC_DISABLE,
>> +
>> +	/* core status fields */
>> +	JMSK_CORE_STATUS_ENCODE_STATE,
>> +	JMSK_CORE_STATUS_SCALE_STATE,
>> +	JMSK_CORE_STATUS_RT_STATE,
>> +	JMSK_CORE_STATUS_BUS_STATE,
>> +	JMSK_CORE_STATUS_CGC_STATE,
>> +
>> +	/* interrupt status fields */
>> +	JMSK_IRQ_STATUS_SESSION_DONE,
>> +	JMSK_IRQ_STATUS_RD_BUF_PLN0_DONE,
>> +	JMSK_IRQ_STATUS_RD_BUF_PLN1_DONE,
>> +	JMSK_IRQ_STATUS_RD_BUF_PLN2_DONE,
>> +	JMSK_IRQ_STATUS_RD_BUF_PLNS_ATTN,
>> +	JMSK_IRQ_STATUS_WR_BUF_PLN0_DONE,
>> +	JMSK_IRQ_STATUS_WR_BUF_PLN1_DONE,
>> +	JMSK_IRQ_STATUS_WR_BUF_PLN2_DONE,
>> +	JMSK_IRQ_STATUS_WR_BUF_PLNS_ATTN,
>> +	JMSK_IRQ_STATUS_SESSION_ERROR,
>> +	JMSK_IRQ_STATUS_STOP_ACK,
>> +	JMSK_IRQ_STATUS_RESET_ACK,
>> +
>> +	/* combined mask for clearing all interrupt status bits */
>> +	JMSK_IRQ_STATUS_ALL_BITS,
>> +
>> +	/* fetch engine (FE) configuration fields */
>> +	JMSK_FE_CFG_BYTE_ORDERING,
>> +	JMSK_FE_CFG_BURST_LENGTH_MAX,
>> +	JMSK_FE_CFG_MEMORY_FORMAT,
>> +	JMSK_FE_CFG_CBCR_ORDER,
>> +	JMSK_FE_CFG_BOTTOM_VPAD_EN,
>> +	JMSK_FE_CFG_PLN0_EN,
>> +	JMSK_FE_CFG_PLN1_EN,
>> +	JMSK_FE_CFG_PLN2_EN,
>> +	JMSK_FE_CFG_SIXTEEN_MCU_EN,
>> +	JMSK_FE_CFG_MCUS_PER_BLOCK,
>> +	JMSK_FE_CFG_MAL_BOUNDARY,
>> +	JMSK_FE_CFG_MAL_EN,
>> +
>> +	JMSK_FE_VBPAD_CFG_BLOCK_ROW,
>> +	JMSK_FE_VBPAD_CFG,
>> +
>> +	/* fetch engine plane address/geometry fields */
>> +	JMSK_PLNS_RD_OFFSET,
>> +	JMSK_PLNS_RD_BUF_SIZE_WIDTH,
>> +	JMSK_PLNS_RD_BUF_SIZE_HEIGHT,
>> +	JMSK_PLNS_RD_STRIDE,
>> +	JMSK_PLNS_RD_HINIT,
>> +	JMSK_PLNS_RD_VINIT,
>> +	JMSK_PLN0_RD_HINIT_INT,
>> +	JMSK_PLN1_RD_HINIT_INT,
>> +	JMSK_PLN2_RD_HINIT_INT,
>> +	JMSK_PLN0_RD_VINIT_INT,
>> +	JMSK_PLN1_RD_VINIT_INT,
>> +	JMSK_PLN2_RD_VINIT_INT,
>> +
>> +	/* write engine (WE) configuration fields */
>> +	JMSK_WE_CFG_BYTE_ORDERING,
>> +	JMSK_WE_CFG_BURST_LENGTH_MAX,
>> +	JMSK_WE_CFG_MEMORY_FORMAT,
>> +	JMSK_WE_CFG_CBCR_ORDER,
>> +	JMSK_WE_CFG_PLN0_EN,
>> +	JMSK_WE_CFG_PLN1_EN,
>> +	JMSK_WE_CFG_PLN2_EN,
>> +	JMSK_WE_CFG_MAL_BOUNDARY,
>> +	JMSK_WE_CFG_MAL_EN,
>> +	JMSK_WE_CFG_POP_BUFF_ON_EOS,
>> +
>> +	/* write engine plane address/geometry fields */
>> +	JMSK_PLNS_WR_BUF_SIZE_WIDTH,
>> +	JMSK_PLNS_WR_BUF_SIZE_HEIGHT,
>> +	JMSK_PLNS_WR_STRIDE,
>> +	JMSK_PLNS_WR_HINIT,
>> +	JMSK_PLNS_WR_VINIT,
>> +	JMSK_PLNS_WR_HSTEP,
>> +	JMSK_PLNS_WR_VSTEP,
>> +	JMSK_PLNS_WR_BLOCK_CFG_PER_COL,
>> +	JMSK_PLNS_WR_BLOCK_CFG_PER_RAW,
>> +
>> +	/* encoder configuration fields */
>> +	JMSK_ENC_CFG_IMAGE_FORMAT,
>> +	JMSK_ENC_CFG_APPLY_EOI,
>> +	JMSK_ENC_CFG_HUFFMAN_SEL,
>> +	JMSK_ENC_CFG_FSC_ENABLE,
>> +	JMSK_ENC_CFG_OUTPUT_DISABLE,
>> +	JMSK_ENC_CFG_RST_MARKER_PERIOD,
>> +	JMSK_ENC_IMAGE_SIZE_WIDTH,
>> +	JMSK_ENC_IMAGE_SIZE_HEIGHT,
>> +
>> +	/* scaler configuration fields */
>> +	JMSK_SCALE_CFG_HSCALE_ENABLE,
>> +	JMSK_SCALE_CFG_VSCALE_ENABLE,
>> +	JMSK_SCALE_CFG_UPSAMPLE_EN,
>> +	JMSK_SCALE_CFG_SUBSAMPLE_EN,
>> +	JMSK_SCALE_CFG_HSCALE_ALGO,
>> +	JMSK_SCALE_CFG_VSCALE_ALGO,
>> +	JMSK_SCALE_CFG_H_SCALE_FIR_ALGO,
>> +	JMSK_SCALE_CFG_V_SCALE_FIR_ALGO,
>> +	JMSK_SCALE_PLNS_OUT_CFG_BLK_WIDTH,
>> +	JMSK_SCALE_PLNS_OUT_CFG_BLK_HEIGHT,
>> +	JMSK_SCALE_PLNS_HSTEP_FRACTIONAL,
>> +	JMSK_SCALE_PLNS_HSTEP_INTEGER,
>> +	JMSK_SCALE_PLNS_VSTEP_FRACTIONAL,
>> +	JMSK_SCALE_PLNS_VSTEP_INTEGER,
>> +
>> +	/* dMI table configuration and write fields */
>> +	JMSK_DMI_CFG,
>> +	JMSK_DMI_ADDR,
>> +	JMSK_DMI_DATA,
>> +
>> +	JMSK_ID_MAX
>> +};
>> +
>> +struct qcom_jpeg_reg_offs {
>> +	u32 hw_version;
>> +	u32 hw_capability;
>> +	u32 reset_cmd;
>> +	u32 core_cfg;
>> +	u32 int_mask;
>> +	u32 int_clr;
>> +	u32 int_status;
>> +	u32 hw_cmd;
>> +	u32 enc_core_state;
>> +
>> +	struct {
>> +		u32 pntr[QCOM_JPEG_MAX_PLANES];
>> +		u32 offs[QCOM_JPEG_MAX_PLANES];
>> +		u32 bsize[QCOM_JPEG_MAX_PLANES];
>> +		u32 stride[QCOM_JPEG_MAX_PLANES];
>> +		u32 hinit[QCOM_JPEG_MAX_PLANES];
>> +		u32 vinit[QCOM_JPEG_MAX_PLANES];
>> +		u32 pntr_cnt;
>> +		u32 vbpad_cfg;
>> +	} fe;
>> +	u32 fe_cfg;
>> +
>> +	struct {
>> +		u32 pntr[QCOM_JPEG_MAX_PLANES];
>> +		u32 bsize[QCOM_JPEG_MAX_PLANES];
>> +		u32 stride[QCOM_JPEG_MAX_PLANES];
>> +		u32 hinit[QCOM_JPEG_MAX_PLANES];
>> +		u32 hstep[QCOM_JPEG_MAX_PLANES];
>> +		u32 vinit[QCOM_JPEG_MAX_PLANES];
>> +		u32 vstep[QCOM_JPEG_MAX_PLANES];
>> +		u32 blocks[QCOM_JPEG_MAX_PLANES];
>> +		u32 pntr_cnt;
>> +	} we;
>> +	u32 we_cfg;
>> +
>> +	struct {
>> +		u32 hstep[QCOM_JPEG_MAX_PLANES];
>> +		u32 vstep[QCOM_JPEG_MAX_PLANES];
>> +	} scale;
>> +	u32 scale_cfg;
>> +	u32 scale_out_cfg[QCOM_JPEG_MAX_PLANES];
>> +
>> +	u32 enc_cfg;
>> +	u32 enc_img_size;
>> +	u32 enc_out_size;
>> +
>> +	u32 dmi_cfg;
>> +	u32 dmi_data;
>> +	u32 dmi_addr;
> 
> Ugh. How many platforms do you support? How many of those have
> significantly different registers here?
> 
>> +};
>> +
>> +#endif /* QCOM_JENC_DEFS_H */
>> diff --git a/drivers/media/platform/qcom/jpeg/qcom_jenc_dev.c b/drivers/media/platform/qcom/jpeg/qcom_jenc_dev.c
>> new file mode 100644
>> index 000000000000..c3573853f24a
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/jpeg/qcom_jenc_dev.c
>> @@ -0,0 +1,337 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
>> + */
>> +
>> +#include <linux/clk.h>
>> +#include <linux/interconnect.h>
>> +#include <linux/interrupt.h>
>> +#include <linux/mod_devicetable.h>
>> +#include <linux/module.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/pm_opp.h>
>> +#include <linux/pm_runtime.h>
>> +#include <linux/slab.h>
>> +
>> +#include <media/v4l2-mem2mem.h>
>> +
>> +#include "qcom_jenc_dev.h"
>> +
>> +#include "qcom_jenc_defs.h"
>> +#include "qcom_jenc_ops.h"
>> +#include "qcom_jenc_res.h"
>> +#include "qcom_jenc_v4l2.h"
>> +
>> +static int qcom_jpeg_clk_init(struct qcom_jenc_dev *jenc)
>> +{
>> +	int c_idx;
>> +
>> +	jenc->num_clks = devm_clk_bulk_get_all(jenc->dev, &jenc->clks);
>> +	if (jenc->num_clks < 0)
>> +		return jenc->num_clks;
>> +
>> +	for (c_idx = 0; c_idx < jenc->num_clks; c_idx++) {
>> +		if (!strcmp(jenc->clks[c_idx].id, "jpeg")) {
>> +			jenc->core_clk = jenc->clks[c_idx].clk;
> 
> Why?
> 
>> +			return 0;
>> +		}
>> +	}
>> +
>> +	return -ENODEV;
>> +}
>> +
>> +static int qcom_jpeg_clk_on(struct qcom_jenc_dev *jenc)
>> +{
>> +	struct dev_pm_opp *opp;
>> +	unsigned long freq;
>> +	int rc;
>> +
>> +	/* look up the OPP entry by level to obtain the core clock frequency */
>> +	opp = dev_pm_opp_find_level_exact(jenc->dev, jenc->perf);
>> +	if (IS_ERR(opp)) {
>> +		dev_err(jenc->dev, "OPP not found for perf level %u: %pe\n", jenc->perf, opp);
>> +		return PTR_ERR(opp);
>> +	}
>> +	freq = dev_pm_opp_get_freq(opp);
>> +	dev_pm_opp_put(opp);
>> +
>> +	rc = clk_set_rate(jenc->core_clk, freq);
> 
> Oh, I see why. It's because you didn't figure out dev_pm_opp_set_rate().
> Please do.
> 
>> +	if (rc) {
>> +		dev_err(jenc->dev, "core clock set rate failed: %d\n", rc);
>> +		return rc;
>> +	}
>> +
>> +	rc = clk_bulk_prepare_enable(jenc->num_clks, jenc->clks);
>> +	if (rc) {
>> +		clk_set_rate(jenc->core_clk, 0);
>> +		return rc;
>> +	}
>> +
>> +	dev_dbg(jenc->dev, "JPEG core clock: %ld\n", clk_get_rate(jenc->core_clk));
>> +
>> +	return 0;
>> +}
>> +
>> +static void qcom_jpeg_clk_off(struct qcom_jenc_dev *jenc)
>> +{
>> +	clk_set_rate(jenc->core_clk, 0);
>> +	clk_bulk_disable_unprepare(jenc->num_clks, jenc->clks);
> 
> Should it not be other way around?
> 
>> +}
>> +
>> +static int qcom_jpeg_icc_on(struct qcom_jenc_dev *jenc)
>> +{
>> +	const struct qcom_dev_resources	*res = jenc->res;
>> +	int p_idx;
>> +	int rc;
>> +
>> +	for (p_idx = 0; p_idx < res->num_of_icc; p_idx++) {
>> +		rc = icc_set_bw(jenc->icc_paths[p_idx], res->icc_res[p_idx].pair.aggr,
>> +				res->icc_res[p_idx].pair.peak);
>> +		if (rc) {
>> +			dev_err(jenc->dev, "icc_set_bw failed for path %s: %d\n",
>> +				res->icc_res[p_idx].icc_id, rc);
>> +			goto err_icc_set_bw;
>> +		}
>> +	}
> 
> Be a nice person, add bulk API to ICC?
> 
>> +
>> +	return 0;
>> +
>> +err_icc_set_bw:
>> +	while (--p_idx >= 0)
>> +		icc_set_bw(jenc->icc_paths[p_idx], 0, 0);
>> +
>> +	return rc;
>> +}
>> +
>> +static void qcom_jpeg_icc_off(struct qcom_jenc_dev *jenc)
>> +{
>> +	const struct qcom_dev_resources	*res = jenc->res;
>> +	int p_idx;
>> +
>> +	for (p_idx = 0; p_idx < res->num_of_icc; p_idx++)
>> +		icc_set_bw(jenc->icc_paths[p_idx], 0, 0);
>> +}
>> +
>> +static int qcom_jpeg_icc_init(struct qcom_jenc_dev *jenc)
>> +{
>> +	const struct qcom_dev_resources	*res = jenc->res;
>> +	int p_idx;
>> +
>> +	jenc->icc_paths = devm_kcalloc(jenc->dev, res->num_of_icc, sizeof(*jenc->icc_paths),
>> +				       GFP_KERNEL);
>> +	if (!jenc->icc_paths)
>> +		return -ENOMEM;
>> +
>> +	for (p_idx = 0; p_idx < res->num_of_icc; p_idx++) {
>> +		jenc->icc_paths[p_idx] = devm_of_icc_get(jenc->dev, res->icc_res[p_idx].icc_id);
>> +		if (IS_ERR(jenc->icc_paths[p_idx])) {
>> +			return dev_err_probe(jenc->dev, PTR_ERR(jenc->icc_paths[p_idx]),
>> +					     "failed to get ICC path: %pe\n",
>> +					     jenc->icc_paths[p_idx]);
>> +		}
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static int qcom_jpeg_pm_suspend(struct device *dev)
>> +{
>> +	struct qcom_jenc_dev *jenc = dev_get_drvdata(dev);
>> +
>> +	qcom_jpeg_clk_off(jenc);
>> +
>> +	qcom_jpeg_icc_off(jenc);
>> +
>> +	return 0;
>> +}
>> +
>> +static int qcom_jpeg_pm_resume(struct device *dev)
>> +{
>> +	struct qcom_jenc_dev *jenc = dev_get_drvdata(dev);
>> +	int rc;
>> +
>> +	rc = qcom_jpeg_icc_on(jenc);
>> +	if (rc)
>> +		return rc;
>> +
>> +	rc = qcom_jpeg_clk_on(jenc);
>> +	if (rc) {
>> +		qcom_jpeg_icc_off(jenc);
>> +		return rc;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static int qcom_jpeg_suspend(struct device *dev)
>> +{
>> +	struct qcom_jenc_dev *jenc = dev_get_drvdata(dev);
>> +	int rc;
>> +
>> +	v4l2_m2m_suspend(jenc->m2m_dev);
>> +
>> +	rc = pm_runtime_force_suspend(dev);
>> +	if (rc)
>> +		v4l2_m2m_resume(jenc->m2m_dev);
>> +
>> +	return rc;
>> +}
>> +
>> +static int qcom_jpeg_resume(struct device *dev)
>> +{
>> +	struct qcom_jenc_dev *jenc = dev_get_drvdata(dev);
>> +	int rc;
>> +
>> +	rc = pm_runtime_force_resume(dev);
>> +	if (rc)
>> +		return rc;
>> +
>> +	v4l2_m2m_resume(jenc->m2m_dev);
>> +
>> +	return 0;
>> +}
>> +
>> +static const struct dev_pm_ops qcom_jpeg_pm_ops = {
>> +	SYSTEM_SLEEP_PM_OPS(qcom_jpeg_suspend, qcom_jpeg_resume)
>> +	RUNTIME_PM_OPS(qcom_jpeg_pm_suspend, qcom_jpeg_pm_resume, NULL)
>> +};
> 
> UNIVERSAL_DEV_PM_OPS? _DEFINE_DEV_PM_OPS?
> 
>> +
>> +static int qcom_jpeg_probe(struct platform_device *pdev)
>> +{
>> +	const struct qcom_dev_resources *res;
>> +	struct qcom_jenc_dev *jenc;
>> +	int rc;
>> +
>> +	jenc = kzalloc_obj(*jenc, GFP_KERNEL);
> 
> devm?
> 
>> +	if (!jenc)
>> +		return -ENOMEM;
>> +
>> +	jenc->dev = &pdev->dev;
>> +	platform_set_drvdata(pdev, jenc);
>> +	rc = devm_mutex_init(&pdev->dev, &jenc->dev_mutex);
>> +	if (rc)
>> +		goto err_free_jenc;
>> +	spin_lock_init(&jenc->hw_lock);
>> +	init_completion(&jenc->reset_complete);
>> +	init_completion(&jenc->stop_complete);
>> +
>> +	res = device_get_match_data(jenc->dev);
>> +	if (!res) {
>> +		rc = dev_err_probe(jenc->dev, -ENODEV, "unsupported SoC\n");
>> +		goto err_free_jenc;
>> +	}
>> +	jenc->res = res;
>> +
>> +	if (!jenc->res->hw_offs || !jenc->res->hw_ops) {
>> +		rc = dev_err_probe(jenc->dev, -EINVAL, "missing hw resources\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	rc = dma_set_mask_and_coherent(jenc->dev, DMA_BIT_MASK(32));
>> +	if (rc) {
>> +		dev_err_probe(jenc->dev, rc, "failed to set DMA mask\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	jenc->jpeg_base = devm_platform_ioremap_resource(pdev, 0);
>> +	if (IS_ERR(jenc->jpeg_base)) {
>> +		rc = dev_err_probe(jenc->dev, PTR_ERR(jenc->jpeg_base),
>> +				   "failed to map JPEG resource\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	rc = devm_pm_opp_of_add_table(jenc->dev);
>> +	if (rc && rc != -ENODEV) {
>> +		dev_err_probe(jenc->dev, rc, "failed to add OPP table\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	rc = qcom_jpeg_clk_init(jenc);
>> +	if (rc) {
>> +		dev_err_probe(jenc->dev, rc, "failed to get clocks\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	jenc->irq = platform_get_irq(pdev, 0);
>> +	if (jenc->irq < 0) {
>> +		rc = dev_err_probe(jenc->dev, jenc->irq, "failed to get IRQ\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	rc = devm_request_threaded_irq(jenc->dev, jenc->irq,
>> +				       jenc->res->hw_ops->hw_irq_top,
>> +				       jenc->res->hw_ops->hw_irq_bot,
>> +				       IRQF_ONESHOT, dev_name(jenc->dev), jenc);
> 
> Are you ready to serve the IRQ here? If not, either request it later,
> when everything is ready, or use IRQF_NO_AUTOEN
> 
>> +	if (rc) {
>> +		dev_err_probe(jenc->dev, rc, "failed to request IRQ\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	rc = qcom_jpeg_icc_init(jenc);
>> +	if (rc)
>> +		goto err_free_jenc;
>> +
>> +	rc = v4l2_device_register(jenc->dev, &jenc->v4l2_dev);
>> +	if (rc) {
>> +		dev_err_probe(jenc->dev, rc, "failed to register V4L2 device\n");
>> +		goto err_free_jenc;
>> +	}
>> +
>> +	jenc->perf = QCOM_SOC_PERF_NOMINAL;
>> +
>> +	rc = devm_pm_runtime_enable(jenc->dev);
>> +	if (rc)
>> +		goto err_v4l2_device_unregister;
>> +
>> +	rc = qcom_jpeg_v4l2_register(jenc);
>> +	if (rc) {
>> +		dev_err_probe(jenc->dev, rc, "failed to register video device\n");
>> +		goto err_v4l2_device_unregister;
>> +	}
> 
> rc = devm_add_action_or_reset(dev, unregister_action)
> 
>> +
>> +	dev_dbg(jenc->dev, "Qualcomm JPEG encoder registered\n");
>> +
>> +	return 0;
>> +
>> +err_v4l2_device_unregister:
>> +	v4l2_device_unregister(&jenc->v4l2_dev);
>> +	kfree(jenc);
>> +	return rc;
>> +err_free_jenc:
>> +	kfree(jenc);
>> +	return rc;
>> +}
>> +
>> +static void qcom_jpeg_remove(struct platform_device *pdev)
>> +{
>> +	struct qcom_jenc_dev *jenc = platform_get_drvdata(pdev);
>> +
>> +	/* v4l2_device_unregister() is called from the release callback. */
>> +	qcom_jpeg_v4l2_unregister(jenc);
>> +
>> +	dev_dbg(jenc->dev, "Qualcomm JPEG encoder deregistered\n");
>> +}
>> +
> 
> With the reset action in place, the remove can be removed.
> 
> 
>> diff --git a/drivers/media/platform/qcom/jpeg/qcom_jenc_hdr.c b/drivers/media/platform/qcom/jpeg/qcom_jenc_hdr.c
>> new file mode 100644
>> index 000000000000..cce15cc18186
>> --- /dev/null
>> +++ b/drivers/media/platform/qcom/jpeg/qcom_jenc_hdr.c
>> @@ -0,0 +1,354 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
>> + */
>> +
>> +#include <linux/errno.h>
>> +#include <linux/string.h>
>> +
>> +#include <media/jpeg.h>
>> +#include <media/v4l2-jpeg.h>
>> +
>> +#include "qcom_jenc_dev.h"
>> +#include "qcom_jenc_hdr.h"
>> +
>> +/*
>> + * The elements defined in this header are specified
>> + * in the ITU-T T.81 / JPEG specification.
>> + *
>> + * https://www.w3.org/Graphics/JPEG/itu-t81.pdf
>> + */
>> +
>> +#define JFIF_HEADER_WIDTH_OFFS		0x07
>> +#define JFIF_HEADER_HEIGHT_OFFS		0x05
>> +
>> +#define JPEG_MARKER_PREFIX		0xff
> 
> None of these definitions are specific to your driver. Move all of them
> to the shared headers. Reuse and extend v4l2-jpeg as much as possible.
> Ideally, only the hardware specifics should be in the driver. Check if
> there is anything to be leveraged from or shared with other existing
> drivers. Another option might be to define a new pixel format,
> header-less JPEG, then make it a userspace job to write all the headers.
> 
>> +
>> +static const u32 qcom_jpeg_v420_hw_reg_mask[] = {
>> +	[JMSK_HW_VER_STEP]	= JPEG_V420_HW_VER_STEP_MSK,
> 
> I think I have written this feedback already. Drop all this indirection.
> Use registers directly, as is.
> 
>> +	[JMSK_HW_VER_MINOR]	= JPEG_V420_HW_VER_MINOR_MSK,
>> +	[JMSK_HW_VER_MAJOR]	= JPEG_V420_HW_VER_MAJOR_MSK,
>> +
>> +};
>> +
>> +static const struct qcom_jpeg_reg_offs qcom_jpeg_v420_hw_reg_offs = {
>> +	.hw_version	= 0x000,
> 
> #define all these values, drop indirection.
> 
>> +	.hw_capability	= 0x004,
>> +	.reset_cmd	= 0x008,
>> +	.core_cfg	= 0x00c,
>> +	.hw_cmd		= 0x010,
>> +	.int_mask	= 0x018,
>> +	.int_clr	= 0x01c,
>> +	.int_status	= 0x020,
>> +	.enc_core_state = 0x014,
>> +
> 


  reply	other threads:[~2026-06-30 14:14 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-29 12:17 [PATCH v3 0/4] Add Qualcomm JPEG V4L2 encoder for SM8250 Atanas Filipov
2026-06-29 12:17 ` [PATCH v3 1/4] media: qcom: camss: populate child platform devices Atanas Filipov
2026-06-29 12:30   ` sashiko-bot
2026-06-29 13:41   ` Vladimir Zapolskiy
2026-06-29 12:17 ` [PATCH v3 2/4] dt-bindings: media: qcom: Add JPEG encoder binding Atanas Filipov
2026-06-29 12:25   ` sashiko-bot
2026-06-29 12:58   ` Dmitry Baryshkov
2026-06-29 13:38   ` Vladimir Zapolskiy
2026-06-30 13:19     ` Bryan O'Donoghue
2026-06-30 13:32       ` Vladimir Zapolskiy
2026-06-30 20:24         ` Bryan O'Donoghue
2026-06-30 21:02           ` Vladimir Zapolskiy
2026-06-29 13:41   ` Rob Herring (Arm)
2026-06-29 12:17 ` [PATCH v3 3/4] arm64: dts: qcom: sm8250: Add JPEG encoder node Atanas Filipov
2026-06-29 12:34   ` sashiko-bot
2026-06-29 13:02   ` Dmitry Baryshkov
2026-06-29 12:17 ` [PATCH v3 4/4] media: qcom: jpeg: Add Qualcomm JPEG V4L2 encoder Atanas Filipov
2026-06-29 12:46   ` sashiko-bot
2026-06-29 13:48   ` Bryan O'Donoghue
2026-06-29 13:50     ` Bryan O'Donoghue
2026-06-30  9:13   ` Konrad Dybcio
2026-06-30 13:06   ` Dmitry Baryshkov
2026-06-30 14:14     ` Atanas Filipov [this message]
2026-07-01  0:13       ` Dmitry Baryshkov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=2a7cd484-ee0f-40bd-bba7-3faa86bb8247@oss.qualcomm.com \
    --to=atanas.filipov@oss.qualcomm.com \
    --cc=andersson@kernel.org \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=dmitry.baryshkov@oss.qualcomm.com \
    --cc=krzysztof.kozlowski+dt@linaro.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=quic_jesszhan@quicinc.com \
    --cc=quic_vgarodia@quicinc.com \
    --cc=robh@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox