Any better like that ? (Should be better in fact ;)
Bastien Nocera wrote:
Bastien Nocera wrote:
<snip>
This bug is already worked-around for the Pismo. As I have the
hardware, I'll try to look at it soon, and submit the fix to BenH and
on linuxppc-dev. In the end, this is definitely a kernel problem.
You might want to try this out (not tested at all):
around line 2543 of dmasound_awacs.c, try adding the lines with a + in
front:
if (hw_can_byteswap)
dmasound.mach.hardware_afmts = (AFMT_S16_BE | AFMT_S16_LE) ;
else
dmasound.mach.hardware_afmts = AFMT_S16_BE ;
+/* if are using a Screamer, device id 11 (iMac 2000)don't claim we
+ * support LE data when we don't */
+if ((sound_device_id == 0xb) && (awacs_revision == AWACS_SCREAMER))
+{
+
hw_can_byteswap = 0;
+ dmasound.mach.hardware_afmts = AFMT_S16_BE ;
+}
/* shut out chips that do output only.
may need to extend this to machines which have no inputs - even tho'
they use screamer - IIRC one of the powerbooks is like this.
