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]
Date: Thu, 1 Feb 2024 18:39:35 +0100
From: Louis Chauvet <louis.chauvet@...tlin.com>
To: Arthur Grillo <arthurgrillo@...eup.net>
Cc: Daniel Vetter <daniel@...ll.ch>, David Airlie <airlied@...il.com>,
	Haneen Mohammed <hamohammed.sa@...il.com>,
	Harry Wentland <harry.wentland@....com>,
	Jonathan Corbet <corbet@....net>,
	Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
	Maxime Ripard <mripard@...nel.org>,
	MaĆ­ra Canal <mairacanal@...eup.net>,
	Melissa Wen <melissa.srw@...il.com>,
	Rodrigo Siqueira <rodrigosiqueiramelo@...il.com>,
	Thomas Zimmermann <tzimmermann@...e.de>,
	linux-kernel@...r.kernel.org, dri-devel@...ts.freedesktop.org,
	linux-doc@...r.kernel.org, nicolejadeyee@...gle.com,
	seanpaul@...gle.com, thomas.petazzoni@...tlin.com,
	miquel.raynal@...tlin.com
Subject: Re: [PATCH v2 4/7] drm/vkms: Add chroma subsampling

[...]

> @@ -146,18 +149,23 @@ void vkms_compose_row(struct line_buffer *stage_buffer, struct vkms_plane_state
>  	for (size_t x = 0; x < limit; x++) {
>  		int x_pos = get_x_position(frame_info, limit, x);
>  
> +		bool shoud_inc = !((x + 1) % frame_format->num_planes);

I think this line will break if the subsampling is not the same as the 
plane count. For NV12 it works only because there are two planes and 
hsub=2/vsub=2, but I believe NV24 will not work because of plane 2, as 
we need to increment x at the same speed on all planes.

I have a proposal to solve this issue (see my patchset applying on top of 
yours). You probably at least need to use .hsub/vsub to 
increment/decrement properly src_pixels pointer.

Currently the tests pass for it because it only use "horizontal 
lines" and "full color" pictures. 

In the series [1] I proposed to change the pattern to detect this kind of 
issue.

[...]

[1]: https://lore.kernel.org/dri-devel/20240201-yuv-v1-0-3ca376f27632@bootlin.com/T/#t

-- 
Louis Chauvet, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ