In servalan.mailinglist.fbsd-current Daniel M. Kurry writes:
>On Wed, Aug 15, 2001 at 07:01:46PM +0200, some SMTP stream spewed forth:
>>
>> One gets the first DMA buffer full, then the process hangs...
>Due to the lack of replies, I'll go ahead.
>I am seeing sound breakage also.
>My card is a
>Creative Labs SoundBlaster Live!.
>xmms will play a short (less than a second) spurt of audio and then stop
>responding. mpg123 will not play (any audio to the speakers) at all.
>I ran a buildworld today which apparently broke it.
>That puts the breakage between today and sometime less than 2 months
>ago.
>(I really cannot be more specific.)
I'm seeing much the same thing, on an SMP box with onboard sbc0 (Vibra16X)
sound chip. Attempting to play sound with madplay gets about 2 seconds of
sound and then silence, with the madplay process in an unkillable kernel
wait. Oddly enough, the sbc0 interrupt thread continues to occasionally gather
a tick of CPU time, but apparently not enough to do anything useful.
I'm busy doing binary-search on the CVS tree, checking out source from
different times and seeing if I can localize the commit that broke it.
My current results are that a kernel built from source as of
2001/08/10 00:00 CDT (i.e. 2001/08/09 22:00:00 PDT) works, one built
from source as of 2001/08/10 15:52 PDT does not, so the bug is
somewhere in between there. I'm now trying to narrow this down further,
to a specific commit somewhere in that region.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message