linux-kernel - Re: [PATCH 1/4] media: mediatek: vcodec: dec: Fix 4K frame size enumeration

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: <f5e68826df868ae5a3cd5737fd9d7f7683bbad73.camel@collabora.com>
Date:   Mon, 27 Jun 2022 11:32:09 -0400
From:   Nicolas Dufresne <nicolas.dufresne@...labora.com>
To:     Chen-Yu Tsai <wenst@...omium.org>,
        Tiffany Lin <tiffany.lin@...iatek.com>,
        Andrew-CT Chen <andrew-ct.chen@...iatek.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Hans Verkuil <hverkuil-cisco@...all.nl>
Cc:     linux-media@...r.kernel.org, linux-mediatek@...ts.infradead.org,
        linux-kernel@...r.kernel.org,
        AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>,
        Yunfei Dong <yunfei.dong@...iatek.com>
Subject: Re: [PATCH 1/4] media: mediatek: vcodec: dec: Fix 4K frame size
 enumeration

Hi Chen-Yu,

Le lundi 27 juin 2022 à 19:23 +0800, Chen-Yu Tsai a écrit :
> This partially reverts commit b018be06f3c7 ("media: mediatek: vcodec:
> Read max resolution from dec_capability"). In this commit, the maximum
> resolution ended up being a function of both the firmware capability and
> the current set format.
> 
> However, frame size enumeration for output (coded) formats should not
> depend on the format set, but should return supported resolutions for
> the format requested by userspace.

Good point. Though, I don't see any special casing for the CAPTURE case. As this
HW does not include a scaler, it must only return 1 resolution when being
enumerated for CAPTURE side (or not implement that enumeration, but its
complicated to half implement something in m2m). The return unique size should
match what G_FMT(CAPTURE) would return.

> 
> Fix this so that the driver returns the supported resolutions correctly,
> even if the instance only has default settings, or if the output format
> is currently set to VP8F, which does not support 4K.
> 
> Fixes: b018be06f3c7 ("media: mediatek: vcodec: Read max resolution from dec_capability")
> Signed-off-by: Chen-Yu Tsai <wenst@...omium.org>
> ---
>  drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec.c    | 2 --
>  .../platform/mediatek/vcodec/mtk_vcodec_dec_stateless.c    | 7 +++++++
>  2 files changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec.c b/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec.c
> index 5d6fdf18c3a6..fcb4b8131c49 100644
> --- a/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec.c
> +++ b/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec.c
> @@ -595,8 +595,6 @@ static int vidioc_enum_framesizes(struct file *file, void *priv,
>  		fsize->type = V4L2_FRMSIZE_TYPE_STEPWISE;
>  		fsize->stepwise = dec_pdata->vdec_framesizes[i].stepwise;
>  
> -		fsize->stepwise.max_width = ctx->max_width;
> -		fsize->stepwise.max_height = ctx->max_height;
>  		mtk_v4l2_debug(1, "%x, %d %d %d %d %d %d",
>  				ctx->dev->dec_capability,
>  				fsize->stepwise.min_width,
> diff --git a/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec_stateless.c b/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec_stateless.c
> index 16d55785d84b..9a4d3e3658aa 100644
> --- a/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec_stateless.c
> +++ b/drivers/media/platform/mediatek/vcodec/mtk_vcodec_dec_stateless.c
> @@ -360,6 +360,13 @@ static void mtk_vcodec_add_formats(unsigned int fourcc,
>  
>  		mtk_vdec_framesizes[count_framesizes].fourcc = fourcc;
>  		mtk_vdec_framesizes[count_framesizes].stepwise = stepwise_fhd;
> +		if (!(ctx->dev->dec_capability & VCODEC_CAPABILITY_4K_DISABLED) &&
> +		    fourcc != V4L2_PIX_FMT_VP8_FRAME) {
> +			mtk_vdec_framesizes[count_framesizes].stepwise.max_width =
> +				VCODEC_DEC_4K_CODED_WIDTH;
> +			mtk_vdec_framesizes[count_framesizes].stepwise.max_height =
> +				VCODEC_DEC_4K_CODED_HEIGHT;
> +		}

I don't particularly like to see this special cased check being added into
multiple places. Its also in your patch 2, and I think it exist in a third
place. Could it be possible to have an internal helper to ensure we don't
duplicate this logic ? Somehow, it seems there is something in common between
set_default, try_fmt and this code.

>  		num_framesizes++;
>  		break;
>  	case V4L2_PIX_FMT_MM21: