linux-kernel - Re: [PATCH 2/2] media: vsp1: vsp1

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

Message-ID: <20250616133134.GC10542@pendragon.ideasonboard.com>
Date: Mon, 16 Jun 2025 16:31:34 +0300
From: Laurent Pinchart <laurent.pinchart@...asonboard.com>
To: Jacopo Mondi <jacopo.mondi+renesas@...asonboard.com>
Cc: Kieran Bingham <kieran.bingham+renesas@...asonboard.com>,
	Mauro Carvalho Chehab <mchehab@...nel.org>,
	linux-media@...r.kernel.org, linux-renesas-soc@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] media: vsp1: vsp1_dl: Count display lists

Hi Jacopo,

Thank you for the patch.

On Thu, May 29, 2025 at 06:36:31PM +0200, Jacopo Mondi wrote:
> To detect invalid usage patterns of the display list helpers, store

We can be more precise:

"To detect leaks of display lists, ..."

> in the display list manager the number of available display lists

s/available/allocated/

> when the manager is created and verify that when the display manager
> is reset the same number of lists is available.
> 
> Signed-off-by: Jacopo Mondi <jacopo.mondi+renesas@...asonboard.com>
> ---
>  drivers/media/platform/renesas/vsp1/vsp1_dl.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/drivers/media/platform/renesas/vsp1/vsp1_dl.c b/drivers/media/platform/renesas/vsp1/vsp1_dl.c
> index 8a3c0274a163..5c4eeb65216f 100644
> --- a/drivers/media/platform/renesas/vsp1/vsp1_dl.c
> +++ b/drivers/media/platform/renesas/vsp1/vsp1_dl.c
> @@ -214,6 +214,7 @@ struct vsp1_dl_list {
>   * @pending: list waiting to be queued to the hardware
>   * @pool: body pool for the display list bodies
>   * @cmdpool: commands pool for extended display list
> + * @list_count: display list counter

"number of allocated display lists"

>   */
>  struct vsp1_dl_manager {
>  	unsigned int index;
> @@ -228,6 +229,8 @@ struct vsp1_dl_manager {
>  
>  	struct vsp1_dl_body_pool *pool;
>  	struct vsp1_dl_cmd_pool *cmdpool;
> +
> +	size_t list_count;
>  };
>  
>  /* -----------------------------------------------------------------------------
> @@ -1073,7 +1076,9 @@ void vsp1_dlm_setup(struct vsp1_device *vsp1)
>  
>  void vsp1_dlm_reset(struct vsp1_dl_manager *dlm)
>  {
> +	size_t dlm_list_count;
>  	unsigned long flags;
> +	size_t list_count;
>  
>  	spin_lock_irqsave(&dlm->lock, flags);
>  
> @@ -1081,8 +1086,13 @@ void vsp1_dlm_reset(struct vsp1_dl_manager *dlm)
>  	__vsp1_dl_list_put(dlm->queued);
>  	__vsp1_dl_list_put(dlm->pending);
>  
> +	list_count = list_count_nodes(&dlm->free);
> +	dlm_list_count = dlm->list_count;

dlm->list_count is not documented as protected by the lock. I don't
think that's an oversight, as it can only be set when the dlm is
created. You can drop the dlm_list_count variable and use
dlm->list_count below.

> +
>  	spin_unlock_irqrestore(&dlm->lock, flags);
>  
> +	WARN_ON_ONCE(list_count != dlm_list_count);
> +
>  	dlm->active = NULL;
>  	dlm->queued = NULL;
>  	dlm->pending = NULL;
> @@ -1150,6 +1160,7 @@ struct vsp1_dl_manager *vsp1_dlm_create(struct vsp1_device *vsp1,
>  				      + sizeof(*dl->header);
>  
>  		list_add_tail(&dl->list, &dlm->free);
> +		dlm->list_count = list_count_nodes(&dlm->free);

Does this need to be done inside the loop, can't you just write

	dlm->list_count = prealloc;

after the loop ?

>  	}
>  
>  	if (vsp1_feature(vsp1, VSP1_HAS_EXT_DL)) {

-- 
Regards,

Laurent Pinchart