Re: [PATCH v4 4/4] media: venus: hfi: add a check to handle OOB in sfr region

[Hans Verkuil]

On 2/7/25 09:24, Vikash Garodia wrote:
> sfr->buf_size is in shared memory and can be modified by malicious user.
> OOB write is possible when the size is made higher than actual sfr data
> buffer. Cap the size to allocated size for such cases.
> 
> Cc: stable@xxxxxxxxxxxxxxx
> Fixes: d96d3f30c0f2 ("[media] media: venus: hfi: add Venus HFI files")
> Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@xxxxxxxxxx>
> Signed-off-by: Vikash Garodia <quic_vgarodia@xxxxxxxxxxx>
> ---
>  drivers/media/platform/qcom/venus/hfi_venus.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/platform/qcom/venus/hfi_venus.c b/drivers/media/platform/qcom/venus/hfi_venus.c
> index 6b615270c5dae470c6fad408c9b5bc037883e56e..c3113420d266e61fcab44688580288d7408b50f4 100644
> --- a/drivers/media/platform/qcom/venus/hfi_venus.c
> +++ b/drivers/media/platform/qcom/venus/hfi_venus.c
> @@ -1041,18 +1041,23 @@ static void venus_sfr_print(struct venus_hfi_device *hdev)
>  {
>  	struct device *dev = hdev->core->dev;
>  	struct hfi_sfr *sfr = hdev->sfr.kva;
> +	u32 size;
>  	void *p;
>  
>  	if (!sfr)
>  		return;
>  
> -	p = memchr(sfr->data, '\0', sfr->buf_size);
> +	size = sfr->buf_size;

If this is ever 0...

> +	if (size > ALIGNED_SFR_SIZE)
> +		size = ALIGNED_SFR_SIZE;
> +
> +	p = memchr(sfr->data, '\0', size);
>  	/*
>  	 * SFR isn't guaranteed to be NULL terminated since SYS_ERROR indicates
>  	 * that Venus is in the process of crashing.
>  	 */
>  	if (!p)
> -		sfr->data[sfr->buf_size - 1] = '\0';
> +		sfr->data[size - 1] = '\0';

...then this will overwrite memory. It probably can't be 0, but a check or perhaps
just a comment might be good. It looks a bit scary.

Regards,

	Hans

>  
>  	dev_err_ratelimited(dev, "SFR message from FW: %s\n", sfr->data);
>  }
>