[PATCH] media: vimc: add RGB/YUV input entity implementation

[PATCH] media: vimc: add RGB/YUV input entity implementation

[Pavan Bobba]

Introduce a new vimc-input entity to simulate a color frame source in the
Virtual Media Controller (VIMC) pipeline. This entity outputs RGB888 frames
and allows testing of pipelines that start from a pre-processed RGB/YUV
source instead of a raw Bayer sensor.

The patch adds vimc-input.c with basic pad operations for format
enumeration, get/set, and stream enable/disable handlers. The entity is
registered in the VIMC core configuration, replacing the previous temporary
use of vimc-sensor. Frame generation is not yet implemented and remains a
TODO for future work.

This change enables link validation and format negotiation for the
RGB/YUV input path, paving the way for software frame injection and
test-pattern generation.

Signed-off-by: Pavan Bobba <opensource206@gmail.com>
---
 drivers/media/test-drivers/vimc/Makefile      |   3 +-
 drivers/media/test-drivers/vimc/vimc-common.h |   1 +
 drivers/media/test-drivers/vimc/vimc-core.c   |   3 +-
 drivers/media/test-drivers/vimc/vimc-input.c  | 210 ++++++++++++++++++
 4 files changed, 214 insertions(+), 3 deletions(-)
 create mode 100644 drivers/media/test-drivers/vimc/vimc-input.c

diff --git a/drivers/media/test-drivers/vimc/Makefile b/drivers/media/test-drivers/vimc/Makefile
index 9b9631562473..7e1fdb2f2a78 100644
--- a/drivers/media/test-drivers/vimc/Makefile
+++ b/drivers/media/test-drivers/vimc/Makefile
@@ -1,6 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0
 vimc-y := vimc-core.o vimc-common.o vimc-streamer.o vimc-capture.o \
-		vimc-debayer.o vimc-scaler.o vimc-sensor.o vimc-lens.o
+		vimc-debayer.o vimc-scaler.o vimc-sensor.o vimc-lens.o \
+		vimc-input.o
 
 obj-$(CONFIG_VIDEO_VIMC) += vimc.o
 
diff --git a/drivers/media/test-drivers/vimc/vimc-common.h b/drivers/media/test-drivers/vimc/vimc-common.h
index 7a45a2117748..6c94b1635fa8 100644
--- a/drivers/media/test-drivers/vimc/vimc-common.h
+++ b/drivers/media/test-drivers/vimc/vimc-common.h
@@ -172,6 +172,7 @@ extern const struct vimc_ent_type vimc_debayer_type;
 extern const struct vimc_ent_type vimc_scaler_type;
 extern const struct vimc_ent_type vimc_capture_type;
 extern const struct vimc_ent_type vimc_lens_type;
+extern const struct vimc_ent_type vimc_input_type;
 
 /**
  * vimc_pix_map_by_index - get vimc_pix_map struct by its index
diff --git a/drivers/media/test-drivers/vimc/vimc-core.c b/drivers/media/test-drivers/vimc/vimc-core.c
index f632c77e52f5..2f6846facb23 100644
--- a/drivers/media/test-drivers/vimc/vimc-core.c
+++ b/drivers/media/test-drivers/vimc/vimc-core.c
@@ -107,9 +107,8 @@ static const struct vimc_ent_config ent_config[] = {
 		.type = &vimc_capture_type
 	},
 	[RGB_YUV_INPUT] = {
-		/* TODO: change this to vimc-input when it is implemented */
 		.name = "RGB/YUV Input",
-		.type = &vimc_sensor_type
+		.type = &vimc_input_type
 	},
 	[SCALER] = {
 		.name = "Scaler",
diff --git a/drivers/media/test-drivers/vimc/vimc-input.c b/drivers/media/test-drivers/vimc/vimc-input.c
new file mode 100644
index 000000000000..cedcc450d59e
--- /dev/null
+++ b/drivers/media/test-drivers/vimc/vimc-input.c
@@ -0,0 +1,210 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * vimc-input.c Virtual Media Controller Driver
+ *
+ * Copyright (C) 2025 Virtual Input Entity Implementation
+ */
+
+#include <linux/v4l2-mediabus.h>
+#include <media/v4l2-subdev.h>
+
+#include "vimc-common.h"
+
+struct vimc_input_device {
+	struct vimc_ent_device ved;
+	struct v4l2_subdev sd;
+	struct media_pad pad;
+};
+
+static const struct v4l2_mbus_framefmt fmt_default = {
+	.width = 640,
+	.height = 480,
+	.code = MEDIA_BUS_FMT_RGB888_1X24,
+	.field = V4L2_FIELD_NONE,
+	.colorspace = V4L2_COLORSPACE_SRGB,
+};
+
+static int vimc_input_init_state(struct v4l2_subdev *sd,
+				 struct v4l2_subdev_state *sd_state)
+{
+	struct v4l2_mbus_framefmt *mf;
+	unsigned int i;
+
+	for (i = 0; i < sd->entity.num_pads; i++) {
+		mf = v4l2_subdev_state_get_format(sd_state, i);
+		*mf = fmt_default;
+	}
+
+	return 0;
+}
+
+static int vimc_input_enum_mbus_code(struct v4l2_subdev *sd,
+				     struct v4l2_subdev_state *sd_state,
+				     struct v4l2_subdev_mbus_code_enum *code)
+{
+	if (code->index > 0)
+		return -EINVAL;
+
+	code->code = MEDIA_BUS_FMT_RGB888_1X24;
+
+	return 0;
+}
+
+static int vimc_input_enum_frame_size(struct v4l2_subdev *sd,
+				      struct v4l2_subdev_state *sd_state,
+				      struct v4l2_subdev_frame_size_enum *fse)
+{
+	const struct vimc_pix_map *vpix;
+
+	if (fse->index)
+		return -EINVAL;
+
+	/* Only accept code in the pix map table */
+	vpix = vimc_pix_map_by_code(fse->code);
+	if (!vpix)
+		return -EINVAL;
+
+	fse->min_width = VIMC_FRAME_MIN_WIDTH;
+	fse->max_width = VIMC_FRAME_MAX_WIDTH;
+	fse->min_height = VIMC_FRAME_MIN_HEIGHT;
+	fse->max_height = VIMC_FRAME_MAX_HEIGHT;
+
+	return 0;
+}
+
+static int vimc_input_get_fmt(struct v4l2_subdev *sd,
+			      struct v4l2_subdev_state *sd_state,
+			      struct v4l2_subdev_format *fmt)
+{
+	struct v4l2_mbus_framefmt *mf;
+
+	mf = v4l2_subdev_state_get_format(sd_state, fmt->pad);
+
+	fmt->format = *mf;
+
+	return 0;
+}
+
+static int vimc_input_set_fmt(struct v4l2_subdev *sd,
+			      struct v4l2_subdev_state *sd_state,
+			      struct v4l2_subdev_format *fmt)
+{
+	struct v4l2_mbus_framefmt *mf;
+
+	mf = v4l2_subdev_state_get_format(sd_state, fmt->pad);
+
+	/* Set the new format */
+	*mf = fmt->format;
+
+	return 0;
+}
+
+static int vimc_input_enable_streams(struct v4l2_subdev *sd,
+				     struct v4l2_subdev_state *state,
+				     u32 pad, u64 streams_mask)
+{
+	/* For input entity, we don't allocate frames since we expect
+	 * external frame injection. Just mark that streaming is active.
+	 *
+	 * TODO: For future enhancement, consider implementing frame generation
+	 * or userspace frame injection mechanism. This would require:
+	 * - Frame buffer allocation (similar to vimc-sensor.c)
+	 * - Interface for userspace to inject frames (e.g., via sysfs/debugfs)
+	 * - Frame rate control for generated test patterns
+	 * - Integration with VIMC's streaming infrastructure
+	 * This would make the input entity suitable for more testing scenarios.
+	 */
+	return 0;
+}
+
+static int vimc_input_disable_streams(struct v4l2_subdev *sd,
+				      struct v4l2_subdev_state *state,
+				      u32 pad, u64 streams_mask)
+{
+	/* Streaming stopped - no cleanup needed for input entity */
+	return 0;
+}
+
+static const struct v4l2_subdev_pad_ops vimc_input_pad_ops = {
+	.enum_mbus_code		= vimc_input_enum_mbus_code,
+	.enum_frame_size	= vimc_input_enum_frame_size,
+	.get_fmt		= vimc_input_get_fmt,
+	.set_fmt		= vimc_input_set_fmt,
+	.enable_streams		= vimc_input_enable_streams,
+	.disable_streams	= vimc_input_disable_streams,
+};
+
+static const struct v4l2_subdev_ops vimc_input_ops = {
+	.pad = &vimc_input_pad_ops,
+};
+
+static const struct v4l2_subdev_internal_ops vimc_input_internal_ops = {
+	.init_state = vimc_input_init_state,
+};
+
+static void vimc_input_release(struct vimc_ent_device *ved)
+{
+	struct vimc_input_device *vinput =
+		container_of(ved, struct vimc_input_device, ved);
+
+	v4l2_subdev_cleanup(&vinput->sd);
+	media_entity_cleanup(vinput->ved.ent);
+	kfree(vinput);
+}
+
+/*
+ * Input process frame function
+ * For an input entity, just return the received frame unchanged
+ */
+static void *vimc_input_process_frame(struct vimc_ent_device *ved,
+				      const void *frame)
+{
+	/* For an input entity, just return the received frame unchanged.
+	 *
+	 * TODO: Future enhancement could implement:
+	 * - Frame validation and format checking
+	 * - Frame transformation or processing
+	 * - Frame injection from userspace buffers
+	 * - Frame rate limiting or buffering
+	 * Currently, this is a simple pass-through for external frame sources.
+	 */
+	return (void *)frame;
+}
+
+static struct vimc_ent_device *vimc_input_add(struct vimc_device *vimc,
+					      const char *vcfg_name)
+{
+	struct v4l2_device *v4l2_dev = &vimc->v4l2_dev;
+	struct vimc_input_device *vinput;
+	int ret;
+
+	/* Allocate the vinput struct */
+	vinput = kzalloc(sizeof(*vinput), GFP_KERNEL);
+	if (!vinput)
+		return ERR_PTR(-ENOMEM);
+
+	/* Initialize the media pad */
+	vinput->pad.flags = MEDIA_PAD_FL_SOURCE;
+
+	ret = vimc_ent_sd_register(&vinput->ved, &vinput->sd, v4l2_dev,
+				   vcfg_name,
+				   MEDIA_ENT_F_IO_V4L, 1, &vinput->pad,
+				   &vimc_input_internal_ops, &vimc_input_ops);
+	if (ret)
+		goto err_free_vinput;
+
+	vinput->ved.process_frame = vimc_input_process_frame;
+	vinput->ved.dev = vimc->mdev.dev;
+
+	return &vinput->ved;
+
+err_free_vinput:
+	kfree(vinput);
+
+	return ERR_PTR(ret);
+}
+
+const struct vimc_ent_type vimc_input_type = {
+	.add = vimc_input_add,
+	.release = vimc_input_release
+};
-- 
2.43.0

Re: [PATCH] media: vimc: add RGB/YUV input entity implementation

[Shuah Khan]

On 10/31/25 23:26, Pavan Bobba wrote:
> Introduce a new vimc-input entity to simulate a color frame source in the
> Virtual Media Controller (VIMC) pipeline. This entity outputs RGB888 frames
> and allows testing of pipelines that start from a pre-processed RGB/YUV
> source instead of a raw Bayer sensor.
> 
> The patch adds vimc-input.c with basic pad operations for format
> enumeration, get/set, and stream enable/disable handlers. The entity is
> registered in the VIMC core configuration, replacing the previous temporary
> use of vimc-sensor. Frame generation is not yet implemented and remains a
> TODO for future work.
> 
> This change enables link validation and format negotiation for the
> RGB/YUV input path, paving the way for software frame injection and
> test-pattern generation.
> 
> Signed-off-by: Pavan Bobba <opensource206@gmail.com>

I see 4 patches - are they dependent then gerenrate a sries
with cover letter explaining the changes you are making.

thanks,
-- Shuah

Re: [PATCH] media: vimc: add RGB/YUV input entity implementation

[opensource india]

Hi Shuah,

On Wed, Nov 5, 2025 at 10:43 PM Shuah Khan <skhan@linuxfoundation.org> wrote:
>
> On 10/31/25 23:26, Pavan Bobba wrote:
> > Introduce a new vimc-input entity to simulate a color frame source in the
> > Virtual Media Controller (VIMC) pipeline. This entity outputs RGB888 frames
> > and allows testing of pipelines that start from a pre-processed RGB/YUV
> > source instead of a raw Bayer sensor.
> >
> > The patch adds vimc-input.c with basic pad operations for format
> > enumeration, get/set, and stream enable/disable handlers. The entity is
> > registered in the VIMC core configuration, replacing the previous temporary
> > use of vimc-sensor. Frame generation is not yet implemented and remains a
> > TODO for future work.
> >
> > This change enables link validation and format negotiation for the
> > RGB/YUV input path, paving the way for software frame injection and
> > test-pattern generation.
> >
> > Signed-off-by: Pavan Bobba <opensource206@gmail.com>
>
> I see 4 patches - are they dependent then gerenrate a sries
> with cover letter explaining the changes you are making.
>
> thanks,
> -- Shuah

even though all 4 patches are part of vimc driver,
these are independent patches

Re: [PATCH] media: vimc: add RGB/YUV input entity implementation

[Shuah Khan]

On 11/5/25 22:03, opensource india wrote:
> Hi Shuah,
> 
> On Wed, Nov 5, 2025 at 10:43 PM Shuah Khan <skhan@linuxfoundation.org> wrote:
>>
>> On 10/31/25 23:26, Pavan Bobba wrote:
>>> Introduce a new vimc-input entity to simulate a color frame source in the
>>> Virtual Media Controller (VIMC) pipeline. This entity outputs RGB888 frames
>>> and allows testing of pipelines that start from a pre-processed RGB/YUV
>>> source instead of a raw Bayer sensor.
>>>
>>> The patch adds vimc-input.c with basic pad operations for format
>>> enumeration, get/set, and stream enable/disable handlers. The entity is
>>> registered in the VIMC core configuration, replacing the previous temporary
>>> use of vimc-sensor. Frame generation is not yet implemented and remains a
>>> TODO for future work.
>>>
>>> This change enables link validation and format negotiation for the
>>> RGB/YUV input path, paving the way for software frame injection and
>>> test-pattern generation.
>>>
>>> Signed-off-by: Pavan Bobba <opensource206@gmail.com>
>>
>> I see 4 patches - are they dependent then gerenrate a sries
>> with cover letter explaining the changes you are making.
>>
>> thanks,
>> -- Shuah
> 
> even though all 4 patches are part of vimc driver,
> these are independent patches

Independent in the sense that they can be applied and tested
independently?

thanks,
-- Shuah

Re: [PATCH] media: vimc: add RGB/YUV input entity implementation

[opensource india]

On Fri, Nov 7, 2025 at 5:41 AM Shuah Khan <skhan@linuxfoundation.org> wrote:
>
> On 11/5/25 22:03, opensource india wrote:
> > Hi Shuah,
> >
> > On Wed, Nov 5, 2025 at 10:43 PM Shuah Khan <skhan@linuxfoundation.org> wrote:
> >>
> >> On 10/31/25 23:26, Pavan Bobba wrote:
> >>> Introduce a new vimc-input entity to simulate a color frame source in the
> >>> Virtual Media Controller (VIMC) pipeline. This entity outputs RGB888 frames
> >>> and allows testing of pipelines that start from a pre-processed RGB/YUV
> >>> source instead of a raw Bayer sensor.
> >>>
> >>> The patch adds vimc-input.c with basic pad operations for format
> >>> enumeration, get/set, and stream enable/disable handlers. The entity is
> >>> registered in the VIMC core configuration, replacing the previous temporary
> >>> use of vimc-sensor. Frame generation is not yet implemented and remains a
> >>> TODO for future work.
> >>>
> >>> This change enables link validation and format negotiation for the
> >>> RGB/YUV input path, paving the way for software frame injection and
> >>> test-pattern generation.
> >>>
> >>> Signed-off-by: Pavan Bobba <opensource206@gmail.com>
> >>
> >> I see 4 patches - are they dependent then gerenrate a sries
> >> with cover letter explaining the changes you are making.
> >>
> >> thanks,
> >> -- Shuah
> >
> > even though all 4 patches are part of vimc driver,
> > these are independent patches
>
> Independent in the sense that they can be applied and tested
> independently?

Yes. all 4 patches can be applied and test independently

>
> thanks,
> -- Shuah
>