kms hdmi audio xine

[forum1793]

Sound from Xine over hdmi using kms plays itself three times about 1/2 to 1 seconds apart. Very odd.

Using mplayer, everything else the same, sound over hdmi with kms is fine.

Prior to kms, sound over hdmi using radeonhd worked fine with xine.

Using 780G/3200hd with Slackware64-current.
libdrm, mesa, radeon are current to master. I'm not using Airlie's drm-next. Takes too much time using git for kernel and drm-next. So I'm using what comes with 2.6.33-rc2.

edit: by the way, changing xine to output to the front speakers works fine and there is no stuttering. The problem occurs only when sound is over hdmi.

I've tried Catalyst on the 64-bit system in the past and it worked fine with both programs. On a Slackware-13 partition, Catalyst plays sound over hdmi fine in both xine and mplayer.

Does anyone else use sound over hdmi? Can you confirm the issue with xine and kms or any fix (other than use mplayer).

[forum1793]

I just tried kernel 2.6.33-rc3 and the audio stuttering over HDMI still occurs with xine but not with mplayer.

I then booted 2.6.29.6 kernel and loaded catalyst 9-12. No audio stuttering in xine or mplayer over HDMI. They were both fine. This is with same xine and mplayer packages as above.

The observation is the same using the xine and mplayer packages distributed with slackware-current, as well as from building the newest gits.

So, somehow the changes in KMS cause audio stuttering over hdmi in xine but not mplayer. How weird is that?

[forum1793]

Just built 2.6.32.3. Made no changes to drm/mesa/radeon (ati)/or radeonhd. Built kernel without staging and without much of anything regarding graphics but agpgart is module.

Using radeon, no sound over hdmi, as expected.
Using radeonhd, perfect sound over hdmi with both xine and mplayer.

So stuttering problem for hdmi audio in xine is related to that functionality in kernel 2.6.33-rc3. Why it doesn't also impact mplayer is odd.

PS. Can't try radeonhd with 2.6.33-rc3 with staging as it hardlocks the pc requiring manually turning off power.

[forum1793]

Just built 2.6.33-rc5

Same stuttering problem occurs with audio over hdmi in xine.

[forum1793]

With build of 2.6.34-rc3, xine now plays audio OK over hdmi again. mplayer continues to work. Didn't work with rc1. I didn't see much in changelog for -rc2 and didn't try it.

Only radeon part of changelog for rc3 is:

commit a5ee4eb75413c145334c30e43f1af9875dad6fd7
Author: Clemens Ladisch <[email protected]>
Date:   Mon Mar 22 09:52:16 2010 +0100

    PCI quirk: RS780/RS880: work around missing MSI initialization
    
    AMD says in section 2.5.4 (GFX MSI Enable) of #43291 (AMD 780G Family
    Register Programming Requirements):
    
      The SBIOS must enable internal graphics MSI capability in GCCFG by
      setting the following: NBCFG.NB_CNTL.STRAP_MSI_ENABLE='1'
    
    Quite a few BIOS writers misinterpret this sentence and think that
    enabling MSI is an optional feature.  However, clearing that bit just
    prevents delivery of MSI messages but does not remove the MSI PCI
    capabilities registers, and so leaves these devices unusable for any
    driver that attempts to use MSI.
    
    Setting that bit is not possible after the BIOS has locked down the
    configuration registers, so we have to manually disable MSI for the
    affected devices.
    
    This fixes the codec communication errors in the HDA driver when
    accessing the HDMI audio device, and allows us to get rid of the
    overcautious quirk in radeon_irq_kms.c.
    
    Signed-off-by: Clemens Ladisch <[email protected]>
    Tested-by: Alex Deucher <[email protected]>
    Cc: <[email protected]>
    Signed-off-by: Jesse Barnes <[email protected]>

This must have done. Thanks devs.