kern/135556: [snd_hda] Sound (mostly) broken w/ snd_hda on
Joseph S. Atkinson
jsa.bsd at gmail.com
Mon Jun 22 08:20:05 UTC 2009
The following reply was made to PR kern/135556; it has been noted by GNATS.
From: "Joseph S. Atkinson" <jsa.bsd at gmail.com>
To: bug-followup at FreeBSD.org, Ariff Abdullag <ariff at FreeBSD.org>
Subject: Re: kern/135556: [snd_hda] Sound (mostly) broken w/ snd_hda on -CURRENT
Date: Mon, 22 Jun 2009 04:17:07 -0400
What I am suspecting is going on here is that sound will only pipe through pcm0
and that hw.snd.default_unit may be honored, but meaningless.
I'm doing as much testing on this as I can with what I have.
So far, sound only comes up when I manually kldload snd_hda still. Doing so
makes the AL888 come up as hdac0 and pcm0 through pcm2 and the R6xx HDMI port as
I can switch hw.snd.default_unit around however, and sound will stop on pcm0 if
I point it at another pcm device. There may be sound going there, but I don't
have speakers on those ports yet (plan to test that tomorrow).
I am thinking that if sound isn't on pcm0, then it goes nowhere. So booting with
snd_hda_load="YES" forces detection of the HDMI port first, making it pcm0, then
the sysctl sets it to pcm1. But via my suspected bug, the signal isn't making it
to pcm1, because it is only working for pcm0.
Tomorrow, I will test this theory by hooking up speakers to another AL888 sound
jack. If I get no sound from the pcm1 device on AL888 as hdac0, it should be an
indicator that the problem is just some misdirection for devices other than pcm0.
If you know a way of disabling the HDMI port, preventing it from even being
probed, it might help me. Doing so would force snd_hda to detect AL888 only and
other factors affecting sound-on-boot could be eliminated.
I will email again tomorrow when I test my pcm1 device, and am willing to try
any and all patches you may have here.
More information about the freebsd-multimedia