linux-kernel - Re: [PATCH v2] media: davinci: vpif: Fix memory leaks in probe error path

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <Z_9KKrbKFDekGLOw@hovoldconsulting.com>
Date: Wed, 16 Apr 2025 08:11:54 +0200
From: Johan Hovold <johan@...nel.org>
To: Dmitry Nikiforov <Dm1tryNk@...dex.ru>
Cc: "Lad, Prabhakar" <prabhakar.csengg@...il.com>,
	Mauro Carvalho Chehab <mchehab@...nel.org>,
	Hans Verkuil <hverkuil@...all.nl>, linux-media@...r.kernel.org,
	linux-kernel@...r.kernel.org, lvc-project@...uxtesting.org
Subject: Re: [PATCH v2] media: davinci: vpif: Fix memory leaks in probe error
 path

On Wed, Apr 16, 2025 at 01:38:20AM +0300, Dmitry Nikiforov wrote:
> If `of_graph_get_endpoint_by_regs()` fails, the probe function currently
> returns 0, skipping cleanup. 
> 
> Also, if an error occurs during the initialization of `pdev_display`,
> the allocated platform device `pdev_capture` is not released properly,
> leading to a memory leak.
> 
> Adjust error path handling to fix the leaks.
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> Fixes: 43acb728bbc4 ("media: davinci: vpif: fix use-after-free on driver unbind")
> Signed-off-by: Dmitry Nikiforov <Dm1tryNk@...dex.ru>
> ---
> v2: also fix of_graph_get_endpoint_by_regs() error path (Johan Hovold).
>  drivers/media/platform/ti/davinci/vpif.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/platform/ti/davinci/vpif.c b/drivers/media/platform/ti/davinci/vpif.c
> index a81719702a22..4839e34e5d29 100644
> --- a/drivers/media/platform/ti/davinci/vpif.c
> +++ b/drivers/media/platform/ti/davinci/vpif.c
> @@ -467,7 +467,8 @@ static int vpif_probe(struct platform_device *pdev)
>  	 */
>  	endpoint = of_graph_get_endpoint_by_regs(pdev->dev.of_node, 0, -1);
>  	if (!endpoint)
> -		return 0;
> +		ret = -ENODEV;
> +		goto err_put_rpm;

This looks wrong, since you're changing an early success path into a
probe failure.

Either way, that would need to go into a separate patch.

I was referring to the error handling for the pdev_display allocation
earlier, which also needs to deregister the capture device on errors
(and that can be done as part of this patch).

>  	of_node_put(endpoint);
>  
>  	/*
> @@ -527,6 +528,7 @@ static int vpif_probe(struct platform_device *pdev)
>  
>  err_put_pdev_display:
>  	platform_device_put(pdev_display);
> +	platform_device_del(pdev_capture);
>  err_put_pdev_capture:
>  	platform_device_put(pdev_capture);
>  err_put_rpm:

Johan