Subject: port-i386/112: /dev/sound with SBPRO hangs
To: None <gnats-admin>
From: None <>
List: netbsd-bugs
Date: 01/31/1994 19:20:33
>Number:         112
>Category:       port-i386
>Synopsis:       /dev/sound with SBPRO hangs
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jan 31 19:20:23 1994
>Originator:     Thomas Eberhardt
CeVis, University of Bremen, Germany
System: NetBSD ed209 0.9a ED209#0 i386

Machine: i386
	It looks that there are problems with the "high-speed" DMA mode
	in the Sound Blaster driver.

	Another thing: I've played with tracker 3.1 (a MOD player).
	I've previously used the Sound Blaster driver for 386bsd and
	everything was fine.  When I now try to make tracker do its thing
	based on the Sparc output module, I just get a bunch of short
	burst.  But when I modify tracker to write its samples to a file
	and then do a "cat" of this file to /dev/audio the output is ok.
	tracker does a lot of small write(2)'s to output its samples.
	Make kernel with "options SBPRO" and do "cat /usr/bin/tee >/dev/sound".
	This should interpret the data as linear encoded and 43478 samples/sec.
	But the process just hangs and is not interruptible.  "ps -axl" gives:

  100 13279   134   0  10  0    60   52 audio  IE+  p2    0:00.04 cat /usr/bin/

	Doing "cat /usr/bin/tee >/dev/audio" instead just makes the desired
	noise.  Here the data is interpreted as mu-law encoded and
	8000 samples/sec.

SUP 8.26 (4.3 BSD) for file NetBSD-current/supfile at Feb  1 03:48:07
Last update occurred at Mon Jan 31 11:41:57 1994 for collection src
Last update occurred at Mon Jan 31 11:39:56 1994 for collection ksrc
Last update occurred at Mon Jan 31 11:39:27 1994 for collection gamessrc
Last update occurred at Mon Jan 31 11:41:45 1994 for collection regress
Last update occurred at Mon Jan 31 11:41:27 1994 for collection othersrc