Alan/Rui,
just built 2.2.18 (on a box that's been running 2.2.14 for a very very long
time, and loading an emu10k1 module from opensource.creative.com or wherever).
was pleased to note that emu10k1 finally made it in. Compiled and built. Dmesg
indicated that things were detecting nicely, but attempts to play sound result
in 'Cannot open /dev/dsp!' (not a rights problem). I had heard a bit earlier
this evening on #debian that someone was complaining of similar problems on a
2.2.17->2.2.18 upgrade, so on a hunch i pulled down a 2.2.17 kernel, and make
oldconfig'd it with my 2.2.18 config.
dmesg again fine, this time it works :) so there does appear to be a problem
there.
templestowe:~$ dmesg |less
Linux version 2.2.17 (root@templestowe) (gcc version 2.95.2 20000220 (Debian
GNU/Linux)) #1 Tue Dec 19 01:44:23 EST 2000
...
Creative EMU10K1 PCI Audio Driver, version 0.6, 01:45:48 Dec 19 2000
emu10k1: EMU10K1 rev 4 model 0x20 found, IO at 0xe400-0xe41f, IRQ 10
...
machine is running debian sid. i'll be happy to do anything else anyone
suggests to play with this. I did diff out the source file by file in
drivers/sound/emu10k1 (my *god* since when did this driver need to be 8k+ lines
of code? i just wrote a kernel and filesystem for my CMU OS class [15-412] in
that many lines :) ... and there's more than i'm prepared to deal with.
is this driver working for anyone else in 2.2.18? anyone recall the reasoning
behind this patch set (i may well have missed it going by on l-k)?
perhaps this is just a known problem by now but i haven't seen mention on l-k
so i'll risk the wrath of the appropriate gods.
Cheers,
Ari Heitner
All reports I've seen mention that the driver works fine if compiled as a
module. You may
want to try that. For this reason I also don't believe the problem is the
any of the emu10k1
source files (which you diff'ed), unfortunately I haven't had much
time/means to test this
since I'm without a home PC and a SBLive.
Rui
Ari Heitner
<aheitner@andre To: [email protected]
w.cmu.edu> cc: [email protected], [email protected],
[email protected]
19/12/00 08:43 Subject: emu10k1 broken in 2.2.18
Please respond
to Ari Heitner
Alan/Rui,
just built 2.2.18 (on a box that's been running 2.2.14 for a very very long
time, and loading an emu10k1 module from opensource.creative.com or
wherever).
was pleased to note that emu10k1 finally made it in. Compiled and built.
Dmesg
indicated that things were detecting nicely, but attempts to play sound
result
in 'Cannot open /dev/dsp!' (not a rights problem). I had heard a bit
earlier
this evening on #debian that someone was complaining of similar problems on
a
2.2.17->2.2.18 upgrade, so on a hunch i pulled down a 2.2.17 kernel, and
make
oldconfig'd it with my 2.2.18 config.
dmesg again fine, this time it works :) so there does appear to be a
problem
there.
templestowe:~$ dmesg |less
Linux version 2.2.17 (root@templestowe) (gcc version 2.95.2 20000220
(Debian
GNU/Linux)) #1 Tue Dec 19 01:44:23 EST 2000
...
Creative EMU10K1 PCI Audio Driver, version 0.6, 01:45:48 Dec 19 2000
emu10k1: EMU10K1 rev 4 model 0x20 found, IO at 0xe400-0xe41f, IRQ 10
...
machine is running debian sid. i'll be happy to do anything else anyone
suggests to play with this. I did diff out the source file by file in
drivers/sound/emu10k1 (my *god* since when did this driver need to be 8k+
lines
of code? i just wrote a kernel and filesystem for my CMU OS class [15-412]
in
that many lines :) ... and there's more than i'm prepared to deal with.
is this driver working for anyone else in 2.2.18? anyone recall the
reasoning
behind this patch set (i may well have missed it going by on l-k)?
perhaps this is just a known problem by now but i haven't seen mention on
l-k
so i'll risk the wrath of the appropriate gods.
Cheers,
Ari Heitner
[email protected] wrote:
>
> All reports I've seen mention that the driver works fine if compiled as a
> module. You may
> want to try that. For this reason I also don't believe the problem is the
> any of the emu10k1
> source files (which you diff'ed), unfortunately I haven't had much
> time/means to test this
> since I'm without a home PC and a SBLive.
Hi Rui,
this is unfortunately not true as I posted lately on emu10k1-devel.
If compiled as a module /dev/dsp and /dev/audio do _no_ work, but
/dev/dsp1 and /dev/audio1 do.
Someone posted a little patch that makes it work again. Here it is:
--- linux-2.2.18/drivers/sound/sound_core.c Sun Dec 17 11:11:46 2000
+++ linux-2.2.18-emubug/drivers/sound/sound_core.c Sun Dec 17
11:05:23 2000
@@ -436,7 +436,7 @@
cs4281_probe();
#endif
#ifdef CONFIG_SOUND_EMU10K1
- init_emu10k1();
+/* init_emu10k1();*/
#endif
#ifdef CONFIG_SOUND_YMFPCI
ymf_probe();
Greetings,
Juri
--
[email protected]
system engineer innominate AG
clustering & security the linux architects
tel: +49-30-308806-45 fax: -77 http://www.innominate.com
> want to try that. For this reason I also don't believe the problem is the
> any of the emu10k1
I dont think its the emu10k1 itself. Nor does it seem to be the init function
not being called (folks report seeing the init messages in both cases). I need
to trace the init order some time