linux-kernel - Re: [PATCH v5 5/8] media: mediatek: vcodec: fix vp9 4096x2176 fail for profile2

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: <899fb4c2ce292bcce4caeb6cf6f38744c449e8e9.camel@collabora.com>
Date: Thu, 06 Nov 2025 09:46:22 -0500
From: Nicolas Dufresne <nicolas.dufresne@...labora.com>
To: Kyrie Wu <kyrie.wu@...iatek.com>, Tiffany Lin
 <tiffany.lin@...iatek.com>,  Andrew-CT Chen <andrew-ct.chen@...iatek.com>,
 Yunfei Dong <yunfei.dong@...iatek.com>, Mauro Carvalho Chehab	
 <mchehab@...nel.org>, Rob Herring <robh@...nel.org>, Krzysztof Kozlowski	
 <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>, Matthias Brugger	
 <matthias.bgg@...il.com>, AngeloGioacchino Del Regno	
 <angelogioacchino.delregno@...labora.com>, Hans Verkuil
 <hverkuil@...all.nl>,  Nathan Hebert <nhebert@...omium.org>, Arnd Bergmann
 <arnd@...db.de>, Irui Wang <irui.wang@...iatek.com>,  George Sun
 <george.sun@...iatek.com>, linux-media@...r.kernel.org,
 devicetree@...r.kernel.org, 	linux-kernel@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, 	linux-mediatek@...ts.infradead.org,
 Yilong Zhou <yilong.zhou@...iatek.com>
Cc: Neil Armstrong <neil.armstrong@...aro.org>, Andrzej Pietrasiewicz
	 <andrzejtp2010@...il.com>
Subject: Re: [PATCH v5 5/8] media: mediatek: vcodec: fix vp9 4096x2176 fail
 for profile2

Hi,

Le jeudi 06 novembre 2025 à 14:13 +0800, Kyrie Wu a écrit :
> The dram addr of vp9 decoder tile number, which use dram mode
> to set tile information, may reach to 36bits for 4096x2176.
> It needs to get the highest 4bit of tile buffer address to
> set tile buffer address.

Please rework that commit message. Perhaps:

	The DRAM address of the VP9 decoder mode info (MI) buffers may require
	as much as 36bits for 4096x2176 resolution. Fold the 4 most significant
	bits into the lower (padding) four bits of address.

> 
> Fixes: 5d418351ca8f1 ("media: mediatek: vcodec: support stateless VP9 decoding")
> 
> Signed-off-by: Kyrie Wu <kyrie.wu@...iatek.com>
> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
> ---
>  .../mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c       | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c
> index d966914db4b9..91c563c049bd 100644
> --- a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c
> +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c
> @@ -1156,7 +1156,10 @@ static int vdec_vp9_slice_setup_tile_buffer:q(struct vdec_vp9_slice_instance *inst
>  			tiles->size[i][j] = size;
>  			if (tiles->mi_rows[i]) {
>  				*tb++ = (size << 3) + ((offset << 3) & 0x7f);
> -				*tb++ = pa & ~0xf;
> +				*tb = pa & ~0xf;
> +				if (IS_ENABLED(CONFIG_ARCH_DMA_ADDR_T_64BIT))
> +					*tb |= (pa >> 32) & 0xf;
> +				tb++;

You are following the style here, so no complaints, though the macros
GENMASK(3, 0) instead of 0xf and GENMASK(31, 4) instead of ~0xf is another way
to write this code.

Since this is not always obvious when the alignment padding of an address is
being reused to extend the number of bits of an address, it would be nice if you
add a comment explaining in words what his is doing.

cheers,
Nicolas

>  				*tb++ = (pa << 3) & 0x7f;
>  				mi_row = (tiles->mi_rows[i] - 1) & 0x1ff;
>  				mi_col = (tiles->mi_cols[j] - 1) & 0x3f;

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)