linux-kernel - Re: [PATCH v4 6/6] media: venus: vdec: Make the range of us_per

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: <9f089225-9c4a-4510-8b0c-da5ba9812a3d@xs4all.nl>
Date: Tue, 7 Jan 2025 10:35:04 +0100
From: Hans Verkuil <hverkuil@...all.nl>
To: Ricardo Ribalda <ribalda@...omium.org>,
 Mauro Carvalho Chehab <mchehab@...nel.org>,
 Stanimir Varbanov <stanimir.k.varbanov@...il.com>,
 Vikash Garodia <quic_vgarodia@...cinc.com>,
 Bryan O'Donoghue <bryan.odonoghue@...aro.org>,
 Hans Verkuil <hans.verkuil@...co.com>
Cc: linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
 Stanimir Varbanov <stanimir.varbanov@...aro.org>,
 linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH v4 6/6] media: venus: vdec: Make the range of us_per_frame
 explicit

On 06/01/2025 14:40, Ricardo Ribalda wrote:
> Fps bigger than 0.000232829 fps, this fits in a 32 bit us_per_frame.
> There is no need to do a 64 bit division here.
> Also, the driver only works with whole fps.
> 
> Found by cocci:
> drivers/media/platform/qcom/venus/vdec.c:488:1-7: WARNING: do_div() does a 64-by-32 division, please consider using div64_u64 instead.
> 
> Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@...aro.org>
> Signed-off-by: Ricardo Ribalda <ribalda@...omium.org>
> ---
>  drivers/media/platform/qcom/venus/vdec.c | 7 ++-----
>  1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/media/platform/qcom/venus/vdec.c b/drivers/media/platform/qcom/venus/vdec.c
> index 6b8906ced6bc..88f6b5a3a4fe 100644
> --- a/drivers/media/platform/qcom/venus/vdec.c
> +++ b/drivers/media/platform/qcom/venus/vdec.c
> @@ -464,7 +464,7 @@ static int vdec_s_parm(struct file *file, void *fh, struct v4l2_streamparm *a)
>  	struct venus_inst *inst = to_inst(file);
>  	struct v4l2_captureparm *cap = &a->parm.capture;
>  	struct v4l2_fract *timeperframe = &cap->timeperframe;
> -	u64 us_per_frame, fps;
> +	u64 us_per_frame;
>  
>  	if (a->type != V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE &&
>  	    a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE)
> @@ -486,10 +486,7 @@ static int vdec_s_parm(struct file *file, void *fh, struct v4l2_streamparm *a)
>  	if (!us_per_frame || us_per_frame > USEC_PER_SEC)
>  		return -EINVAL;
>  
> -	fps = (u64)USEC_PER_SEC;
> -	do_div(fps, us_per_frame);
> -
> -	inst->fps = fps;
> +	inst->fps = USEC_PER_SEC / (u32)us_per_frame;

This still allows for an fps value of USEC_PER_SEC if us_per_frame is 1.

Looking at where inst->fps is used I see:

drivers/media/platform/qcom/venus/pm_helpers.c: return mbs * inst->fps;
drivers/media/platform/qcom/venus/venc.c:       frate.framerate = inst->fps * (1 << 16);

(mbs is at most 512x512)

So if fps is USEC_PER_SEC those calculations will wrap around.

What is the real maximum fps that the HW can handle?

Stan? Bryan? It would be nice if there is a proper sanity check here.

Regards,

	Hans


>  	timeperframe->numerator = 1;
>  	timeperframe->denominator = inst->fps;
>  
>