Linux kernel and device drivers for NXP i.MX platforms
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: Frank.Li@oss.nxp.com
Cc: Frank.Li@kernel.org, imx@lists.linux.dev, conor+dt@kernel.org,
	devicetree@vger.kernel.org, robh@kernel.org
Subject: Re: [PATCH v6 1/9] media: mc-entity: Store parsed V4L2 fwnode endpoint in media_pad
Date: Wed, 24 Jun 2026 20:56:59 +0000	[thread overview]
Message-ID: <20260624205659.C63B01F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260624-imx8qxp_pcam-v6-1-4b3f45920d2f@nxp.com>

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Critical] Captured fwnode endpoint contains a dangling pointer to `local_fwnode` if the caller drops its reference after parsing.
- [Medium] Embedding a large, V4L2-specific `struct v4l2_fwnode_endpoint` into the generic Media Controller `struct media_pad` breaks subsystem abstraction and unconditionally bloats memory.
--

commit 7acfb56349845d6638c963825f0f6b087ebaa54d
Author: Frank Li <Frank.Li@nxp.com>

media: mc-entity: Store parsed V4L2 fwnode endpoint in media_pad

This commit captures the parsed v4l2 fwnode endpoint information directly
within struct media_pad to allow reuse by consumers without reparsing.

> diff --git a/include/media/media-entity.h b/include/media/media-entity.h
> index d9b72cd87d524..4a3785cd9f370 100644
> --- a/include/media/media-entity.h
> +++ b/include/media/media-entity.h
> @@ -20,6 +20,8 @@
>  #include <linux/minmax.h>
>  #include <linux/types.h>
>  
> +#include <media/v4l2-fwnode.h>
> +

[Severity: Medium]
Does this include break subsystem isolation? struct media_pad is a generic
structure used across several subsystems like alsa, dvb, drm, and iio.
Including a v4l2-specific header here appears to tie the generic media
controller to v4l2.

[ ... ]

> @@ -240,7 +243,7 @@ struct media_pad {
>  	u16 num_links;
>  	enum media_pad_signal_type sig_type;
>  	unsigned long flags;
> -
> +	struct v4l2_fwnode_endpoint vep;

[Severity: Medium]
Is it expected to embed this v4l2-specific structure into every generic
media_pad? The v4l2_fwnode_endpoint struct is quite large, and
unconditionally embedding it adds memory overhead for all non-v4l2
entities that do not use fwnode endpoints.

[Severity: Critical]
Can this cause a use-after-free regression with the fwnode pointers?

When the endpoint is parsed, v4l2_fwnode_endpoint_parse() assigns the
fwnode to vep->base.local_fwnode without taking a reference count. This
was safe when vep was a transient stack variable, but now that it is
stored persistently in the media_pad, the caller might drop the
reference.

For example, if the caller iterates using a scoped iterator:

drivers/media/v4l2-core/v4l2-fwnode.c:v4l2_async_nf_parse_fwnode() {
    ...
    fwnode_graph_for_each_endpoint_scoped(dev_fwnode(dev), ep) {
        ret = v4l2_fwnode_endpoint_parse(ep, &pad->vep);
    ...
}

The scoped iterator automatically drops the reference to ep at the end
of the iteration, leaving pad->vep.base.local_fwnode as a dangling
pointer.

Later access or pointer comparisons, like in v4l2_common_notifier_bound(),
might dereference freed memory if the fwnode is deleted.

>  	/*
>  	 * The fields below are private, and should only be accessed via
>  	 * appropriate functions.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260624-imx8qxp_pcam-v6-0-4b3f45920d2f@nxp.com?part=1

  reply	other threads:[~2026-06-24 20:57 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-24 20:37 [PATCH v6 0/9] media: add new API simple 1to1 subdev register and add imx parallel camera support Frank.Li
2026-06-24 20:37 ` [PATCH v6 1/9] media: mc-entity: Store parsed V4L2 fwnode endpoint in media_pad Frank.Li
2026-06-24 20:56   ` sashiko-bot [this message]
2026-06-24 20:37 ` [PATCH v6 2/9] media: subdev: Add set_pad_by_ep() callback to internal ops Frank.Li
2026-06-24 20:50   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 3/9] media: subdev: Add media_async_register_subdev() helper Frank.Li
2026-06-24 21:00   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 4/9] media: synopsys: Use v4l2_subdev_get_frame_desc_passthrough() Frank.Li
2026-06-24 20:37 ` [PATCH v6 5/9] media: synopsys: Use media_async_register_subdev() to simplify code Frank.Li
2026-06-24 21:07   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 6/9] dt-bindings: media: add i.MX parallel CPI support Frank.Li
2026-06-24 20:57   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 7/9] media: nxp: add V4L2 subdev driver for camera parallel interface (CPI) Frank.Li
2026-06-24 21:03   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 8/9] arm64: dts: imx8: add camera parallel interface (CPI) node Frank.Li
2026-06-24 21:00   ` sashiko-bot
2026-06-24 20:37 ` [PATCH v6 9/9] arm64: dts: imx8qxp-mek: add parallel ov5640 camera support Frank.Li

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=20260624205659.C63B01F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=Frank.Li@kernel.org \
    --cc=Frank.Li@oss.nxp.com \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=imx@lists.linux.dev \
    --cc=robh@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /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