linux-kernel - Re: [PATCH] ALSA: pcm: fix ELD constraints for (E)AC3, DTS(-HD) and MLP formats

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>] [day] [month] [year] [list]

Message-ID: <ZJhLdE7oXAvIi1Yi@camel3.lan>
Date:   Sun, 25 Jun 2023 16:13:08 +0200
From:   Matthias Reichl <hias@...us.com>
To:     Markus Elfring <Markus.Elfring@....de>
Cc:     alsa-devel@...a-project.org, kernel-janitors@...r.kernel.org,
        Dom Cobley <popcornmix@...il.com>,
        Jaroslav Kysela <perex@...ex.cz>,
        Takashi Iwai <tiwai@...e.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] ALSA: pcm: fix ELD constraints for (E)AC3, DTS(-HD) and
 MLP formats

On Sun, Jun 25, 2023 at 02:10:21PM +0200, Markus Elfring wrote:
> …
> > This patch fixes the constraints for the common AC3 and DTS formats,
> …
> 
> Please add an imperative change suggestion.

I assumed the motivation was pretty clear from the paragraph above which you
snipped off:

> The SADs of compressed formats contain the channel and sample rate
> info of the audio data inside the compressed stream, but when
> building constraints we must use the rates and channels used to
> transport the compressed streams.
>
> eg 48kHz 6ch EAC3 needs to be transmitted as a 2ch 192kHz stream.

The previous implementation added constraints that could be both too broad
and incomplete at the same time, leading to the audio device accepting
channel/rate combinations that are not supported by the sink while rejecting
combinations that are required to transmit the compressed bitstream.

Typical impact on users is eg "Dolby TrueHD passthrough does not work".

Consider this EDID audio block of a 2020 Sony TV which rejected Dolby TrueHD
passthrough:

    Linear PCM:
      Max channels: 6
      Supported sample rates (kHz): 192 176.4 96 88.2 48 44.1 32
      Supported sample sizes (bits): 24 20 16
    AC-3:
      Max channels: 6
      Supported sample rates (kHz): 48 44.1 32
      Maximum bit rate: 640 kb/s
    DTS:
      Max channels: 6
      Supported sample rates (kHz): 48 44.1 32
      Maximum bit rate: 1504 kb/s
    Enhanced AC-3 (DD+):
      Max channels: 8
      Supported sample rates (kHz): 48 44.1
      Supports Joint Object Coding
    MAT (MLP):
      Max channels: 8
      Supported sample rates (kHz): 48
      Supports Dolby TrueHD, object audio PCM and channel-based PCM
      Hash calculation not required for object audio PCM or channel-based PCM

The old implementation didn't add the 192kHz / 8ch combination that's required
to transport the MLP TrueHD bitstream, so opening the device in 8ch 192kHz mode
failed.

> How do you think about to add the tag “Fixes”?

I've thought about that but decided against it as adding exact constraints
has the chance of breaking existing applications that accidentally relied on
the (incorrect) previous behaviour of adding rather broad constraints.

Consider the following EDID of a 2009 Sony TV:

    Linear PCM:
      Max channels: 2
      Supported sample rates (kHz): 48 44.1 32
      Supported sample sizes (bits): 24 20 16
    AC-3:
      Max channels: 6
      Supported sample rates (kHz): 48 44.1 32
      Maximum bit rate: 640 kb/s
    Enhanced AC-3 (DD+):
      Max channels: 6
      Supported sample rates (kHz): 48 44.1 32

The old implementation would have constraints that allowed up to 6ch output at
32/44.1/48kHz while the correct setup would be to only allow max 2ch output
(both AC3 and EAC3 bitstreams are transmitted in 2ch mode).

So you could successfully output eg 6ch audio (which the sink likely wouldn't accept
and/or only output the first 2 channels) before but now this will get rejected.

so long,

Hias