[PATCH 0/3] media: Fix CI warnings for 6.19

[PATCH 0/3] media: Fix CI warnings for 6.19

[Ricardo Ribalda]

New kernel version, new warnings.

This series only introduces a new patch:
media: iris: Document difference in size during allocation

The other two have been already sent to linux-media or linux-next ML,
but they have not found their way into the tree.

Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
Jacopo Mondi (1):
      media: uapi: c3-isp: Fix documentation warning

Ricardo Ribalda (2):
      media: iris: Document difference in size during allocation
      media: iris: Fix fps calculation

 drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c | 10 +++++++++-
 drivers/media/platform/qcom/iris/iris_venc.c             |  5 ++---
 include/uapi/linux/media/amlogic/c3-isp-config.h         |  2 +-
 3 files changed, 12 insertions(+), 5 deletions(-)
---
base-commit: 47b7b5e32bb7264b51b89186043e1ada4090b558
change-id: 20251202-warnings-6-19-960d9b686cff

Best regards,
-- 
Ricardo Ribalda <ribalda@chromium.org>

[PATCH 1/3] media: uapi: c3-isp: Fix documentation warning

[Ricardo Ribalda]

From: Jacopo Mondi <jacopo.mondi@ideasonboard.com>

Building htmldocs generates a warning:

WARNING: include/uapi/linux/media/amlogic/c3-isp-config.h:199
error: Cannot parse struct or union!

Which correctly highlights that the c3_isp_params_block_header symbol
is wrongly documented as a struct while it's a plain #define instead.

Fix this by removing the 'struct' identifier from the documentation of
the c3_isp_params_block_header symbol.

[ribalda: Add Closes:]

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/all/20251127131425.4b5b6644@canb.auug.org.au/
Fixes: 45662082855c ("media: uapi: Convert Amlogic C3 to V4L2 extensible params")
Cc: stable@vger.kernel.org
Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
 include/uapi/linux/media/amlogic/c3-isp-config.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/uapi/linux/media/amlogic/c3-isp-config.h b/include/uapi/linux/media/amlogic/c3-isp-config.h
index 0a3c1cc55ccbbad12f18037d65f32ec9ca1a4ec0..92db5dcdda181cb31665e230cc56b443fa37a0be 100644
--- a/include/uapi/linux/media/amlogic/c3-isp-config.h
+++ b/include/uapi/linux/media/amlogic/c3-isp-config.h
@@ -186,7 +186,7 @@ enum c3_isp_params_block_type {
 #define C3_ISP_PARAMS_BLOCK_FL_ENABLE	V4L2_ISP_PARAMS_FL_BLOCK_ENABLE
 
 /**
- * struct c3_isp_params_block_header - C3 ISP parameter block header
+ * c3_isp_params_block_header - C3 ISP parameter block header
  *
  * This structure represents the common part of all the ISP configuration
  * blocks and is identical to :c:type:`v4l2_isp_params_block_header`.

-- 
2.52.0.158.g65b55ccf14-goog

Re: [PATCH 1/3] media: uapi: c3-isp: Fix documentation warning

[Mauro Carvalho Chehab]

Em Wed, 03 Dec 2025 08:55:34 +0000
Ricardo Ribalda <ribalda@chromium.org> escreveu:

> From: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> 
> Building htmldocs generates a warning:
> 
> WARNING: include/uapi/linux/media/amlogic/c3-isp-config.h:199
> error: Cannot parse struct or union!
> 
> Which correctly highlights that the c3_isp_params_block_header symbol
> is wrongly documented as a struct while it's a plain #define instead.
> 
> Fix this by removing the 'struct' identifier from the documentation of
> the c3_isp_params_block_header symbol.
> 
> [ribalda: Add Closes:]
> 
> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> Closes: https://lore.kernel.org/all/20251127131425.4b5b6644@canb.auug.org.au/
> Fixes: 45662082855c ("media: uapi: Convert Amlogic C3 to V4L2 extensible params")
> Cc: stable@vger.kernel.org
> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  include/uapi/linux/media/amlogic/c3-isp-config.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/uapi/linux/media/amlogic/c3-isp-config.h b/include/uapi/linux/media/amlogic/c3-isp-config.h
> index 0a3c1cc55ccbbad12f18037d65f32ec9ca1a4ec0..92db5dcdda181cb31665e230cc56b443fa37a0be 100644
> --- a/include/uapi/linux/media/amlogic/c3-isp-config.h
> +++ b/include/uapi/linux/media/amlogic/c3-isp-config.h
> @@ -186,7 +186,7 @@ enum c3_isp_params_block_type {
>  #define C3_ISP_PARAMS_BLOCK_FL_ENABLE	V4L2_ISP_PARAMS_FL_BLOCK_ENABLE
>  
>  /**
> - * struct c3_isp_params_block_header - C3 ISP parameter block header
> + * c3_isp_params_block_header - C3 ISP parameter block header
>   *
>   * This structure represents the common part of all the ISP configuration
>   * blocks and is identical to :c:type:`v4l2_isp_params_block_header`.
> 

Just merged this one at media-committers next. 

My plan is to send later during the merge window, after the first
PR I sent upstream gets merged.

Regards,
Mauro

[PATCH 2/3] media: iris: Document difference in size during allocation

[Ricardo Ribalda]

As we get ready for kzalloc checking for invalid sizes, let's add
documentation for the cases where the size is different but valid.

This patch fixes this cocci warning:
./platform/qcom/iris/iris_hfi_gen2_command.c:1215:9-25: WARNING: casting value returned by memory allocation function to (struct iris_inst *) is useless.

Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
 drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
index f9129553209922fda548ca320494ae6ae797854c..ab91afd0597045bd876d0411b08b5a3421b12c70 100644
--- a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
+++ b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
@@ -1212,5 +1212,13 @@ void iris_hfi_gen2_command_ops_init(struct iris_core *core)
 
 struct iris_inst *iris_hfi_gen2_get_instance(void)
 {
-	return (struct iris_inst *)kzalloc(sizeof(struct iris_inst_hfi_gen2), GFP_KERNEL);
+	struct iris_inst_hfi_gen2 *out;
+
+	/*
+	 * The allocation is intentionally larger. The first member of
+	 * struct iris_hfi_gen2 is struct iris_inst.
+	 */
+	out = kzalloc(sizeof(*out), GFP_KERNEL);
+
+	return (struct iris_inst *)out;
 }

-- 
2.52.0.158.g65b55ccf14-goog

Re: [PATCH 2/3] media: iris: Document difference in size during allocation

[Laurent Pinchart]

On Wed, Dec 03, 2025 at 08:55:35AM +0000, Ricardo Ribalda wrote:
> As we get ready for kzalloc checking for invalid sizes, let's add
> documentation for the cases where the size is different but valid.
> 
> This patch fixes this cocci warning:
> ./platform/qcom/iris/iris_hfi_gen2_command.c:1215:9-25: WARNING: casting value returned by memory allocation function to (struct iris_inst *) is useless.
> 
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> index f9129553209922fda548ca320494ae6ae797854c..ab91afd0597045bd876d0411b08b5a3421b12c70 100644
> --- a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> +++ b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> @@ -1212,5 +1212,13 @@ void iris_hfi_gen2_command_ops_init(struct iris_core *core)
>  
>  struct iris_inst *iris_hfi_gen2_get_instance(void)
>  {
> -	return (struct iris_inst *)kzalloc(sizeof(struct iris_inst_hfi_gen2), GFP_KERNEL);
> +	struct iris_inst_hfi_gen2 *out;
> +
> +	/*
> +	 * The allocation is intentionally larger. The first member of
> +	 * struct iris_hfi_gen2 is struct iris_inst.
> +	 */
> +	out = kzalloc(sizeof(*out), GFP_KERNEL);
> +
> +	return (struct iris_inst *)out;

	return &out->inst;

would be more readable. You can then drop the comment.

>  }
> 

-- 
Regards,

Laurent Pinchart

Re: [PATCH 2/3] media: iris: Document difference in size during allocation

[Ricardo Ribalda]

Hi Laurent



On Wed, 3 Dec 2025 at 10:17, Laurent Pinchart
<laurent.pinchart@ideasonboard.com> wrote:
>
> On Wed, Dec 03, 2025 at 08:55:35AM +0000, Ricardo Ribalda wrote:
> > As we get ready for kzalloc checking for invalid sizes, let's add
> > documentation for the cases where the size is different but valid.
> >
> > This patch fixes this cocci warning:
> > ./platform/qcom/iris/iris_hfi_gen2_command.c:1215:9-25: WARNING: casting value returned by memory allocation function to (struct iris_inst *) is useless.
> >
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > ---
> >  drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c | 10 +++++++++-
> >  1 file changed, 9 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> > index f9129553209922fda548ca320494ae6ae797854c..ab91afd0597045bd876d0411b08b5a3421b12c70 100644
> > --- a/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> > +++ b/drivers/media/platform/qcom/iris/iris_hfi_gen2_command.c
> > @@ -1212,5 +1212,13 @@ void iris_hfi_gen2_command_ops_init(struct iris_core *core)
> >
> >  struct iris_inst *iris_hfi_gen2_get_instance(void)
> >  {
> > -     return (struct iris_inst *)kzalloc(sizeof(struct iris_inst_hfi_gen2), GFP_KERNEL);
> > +     struct iris_inst_hfi_gen2 *out;
> > +
> > +     /*
> > +      * The allocation is intentionally larger. The first member of
> > +      * struct iris_hfi_gen2 is struct iris_inst.
> > +      */
> > +     out = kzalloc(sizeof(*out), GFP_KERNEL);
> > +
> > +     return (struct iris_inst *)out;
>
>         return &out->inst;
>
> would be more readable. You can then drop the comment.

It looks better, and it is more robust. Thanks

I prefer to leave a small comment though:

+       struct iris_inst_hfi_gen2 *out;
+
+       /* The allocation is intentionally larger than struct iris_inst. */
+       out = kzalloc(sizeof(*out), GFP_KERNEL);
+
+       return &out->inst;


Regards

>
> >  }
> >
>
> --
> Regards,
>
> Laurent Pinchart



-- 
Ricardo Ribalda

[PATCH 3/3] media: iris: Fix fps calculation

[Ricardo Ribalda]

iris_venc_s_param() uses do_div to divide two 64 bits operators, this is
wrong. Luckily for us, both of the operators fit in 32 bits, so we can use
a normal division.

Now that we are at it, mark the fps smaller than 1 as invalid, the code
does not seem to handle them properly.

The following cocci warning is fixed with this patch:
./platform/qcom/iris/iris_venc.c:378:1-7: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead

Fixes: 4ff586ff28e3 ("media: iris: Add support for G/S_PARM for encoder video device")
Reviewed-by: Dikshita Agarwal <dikshita.agarwal@oss.qualcomm.com>
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
 drivers/media/platform/qcom/iris/iris_venc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/media/platform/qcom/iris/iris_venc.c b/drivers/media/platform/qcom/iris/iris_venc.c
index 5830eba93c68b27fa9db87bac63a691eaca338d2..730f2aa1a83b2f4f90227ab9a7e04aee7c2c3cfe 100644
--- a/drivers/media/platform/qcom/iris/iris_venc.c
+++ b/drivers/media/platform/qcom/iris/iris_venc.c
@@ -408,11 +408,10 @@ int iris_venc_s_param(struct iris_inst *inst, struct v4l2_streamparm *s_parm)
 	us_per_frame = timeperframe->numerator * (u64)USEC_PER_SEC;
 	do_div(us_per_frame, timeperframe->denominator);
 
-	if (!us_per_frame)
+	if (!us_per_frame || us_per_frame > USEC_PER_SEC)
 		return -EINVAL;
 
-	fps = (u64)USEC_PER_SEC;
-	do_div(fps, us_per_frame);
+	fps = USEC_PER_SEC / (u32)us_per_frame;
 	if (fps > max_rate) {
 		ret = -ENOMEM;
 		goto reset_rate;

-- 
2.52.0.158.g65b55ccf14-goog

Re: [PATCH 3/3] media: iris: Fix fps calculation

[Dmitry Baryshkov]

On Wed, Dec 03, 2025 at 08:55:36AM +0000, Ricardo Ribalda wrote:
> iris_venc_s_param() uses do_div to divide two 64 bits operators, this is
> wrong. Luckily for us, both of the operators fit in 32 bits, so we can use
> a normal division.
> 
> Now that we are at it, mark the fps smaller than 1 as invalid, the code
> does not seem to handle them properly.
> 
> The following cocci warning is fixed with this patch:
> ./platform/qcom/iris/iris_venc.c:378:1-7: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead
> 
> Fixes: 4ff586ff28e3 ("media: iris: Add support for G/S_PARM for encoder video device")
> Reviewed-by: Dikshita Agarwal <dikshita.agarwal@oss.qualcomm.com>
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/platform/qcom/iris/iris_venc.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/media/platform/qcom/iris/iris_venc.c b/drivers/media/platform/qcom/iris/iris_venc.c
> index 5830eba93c68b27fa9db87bac63a691eaca338d2..730f2aa1a83b2f4f90227ab9a7e04aee7c2c3cfe 100644
> --- a/drivers/media/platform/qcom/iris/iris_venc.c
> +++ b/drivers/media/platform/qcom/iris/iris_venc.c
> @@ -408,11 +408,10 @@ int iris_venc_s_param(struct iris_inst *inst, struct v4l2_streamparm *s_parm)
>  	us_per_frame = timeperframe->numerator * (u64)USEC_PER_SEC;
>  	do_div(us_per_frame, timeperframe->denominator);
>  
> -	if (!us_per_frame)
> +	if (!us_per_frame || us_per_frame > USEC_PER_SEC)
>  		return -EINVAL;
>  
> -	fps = (u64)USEC_PER_SEC;
> -	do_div(fps, us_per_frame);
> +	fps = USEC_PER_SEC / (u32)us_per_frame;

Can we replace this with:

fps = timeperframe->denominator / timeperframe->numerator ?

>  	if (fps > max_rate) {
>  		ret = -ENOMEM;
>  		goto reset_rate;
> 
> -- 
> 2.52.0.158.g65b55ccf14-goog
> 

-- 
With best wishes
Dmitry

Re: [PATCH 3/3] media: iris: Fix fps calculation

[Ricardo Ribalda]

On Sat, 6 Dec 2025 at 13:06, Dmitry Baryshkov
<dmitry.baryshkov@oss.qualcomm.com> wrote:
>
> On Wed, Dec 03, 2025 at 08:55:36AM +0000, Ricardo Ribalda wrote:
> > iris_venc_s_param() uses do_div to divide two 64 bits operators, this is
> > wrong. Luckily for us, both of the operators fit in 32 bits, so we can use
> > a normal division.
> >
> > Now that we are at it, mark the fps smaller than 1 as invalid, the code
> > does not seem to handle them properly.
> >
> > The following cocci warning is fixed with this patch:
> > ./platform/qcom/iris/iris_venc.c:378:1-7: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead
> >
> > Fixes: 4ff586ff28e3 ("media: iris: Add support for G/S_PARM for encoder video device")
> > Reviewed-by: Dikshita Agarwal <dikshita.agarwal@oss.qualcomm.com>
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > ---
> >  drivers/media/platform/qcom/iris/iris_venc.c | 5 ++---
> >  1 file changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/media/platform/qcom/iris/iris_venc.c b/drivers/media/platform/qcom/iris/iris_venc.c
> > index 5830eba93c68b27fa9db87bac63a691eaca338d2..730f2aa1a83b2f4f90227ab9a7e04aee7c2c3cfe 100644
> > --- a/drivers/media/platform/qcom/iris/iris_venc.c
> > +++ b/drivers/media/platform/qcom/iris/iris_venc.c
> > @@ -408,11 +408,10 @@ int iris_venc_s_param(struct iris_inst *inst, struct v4l2_streamparm *s_parm)
> >       us_per_frame = timeperframe->numerator * (u64)USEC_PER_SEC;
> >       do_div(us_per_frame, timeperframe->denominator);
> >
> > -     if (!us_per_frame)
> > +     if (!us_per_frame || us_per_frame > USEC_PER_SEC)
> >               return -EINVAL;
> >
> > -     fps = (u64)USEC_PER_SEC;
> > -     do_div(fps, us_per_frame);
> > +     fps = USEC_PER_SEC / (u32)us_per_frame;
>
> Can we replace this with:
>
> fps = timeperframe->denominator / timeperframe->numerator ?

Yep, I believe you are correct

I will send a new version soon.

>
> >       if (fps > max_rate) {
> >               ret = -ENOMEM;
> >               goto reset_rate;
> >
> > --
> > 2.52.0.158.g65b55ccf14-goog
> >
>
> --
> With best wishes
> Dmitry



-- 
Ricardo Ribalda