6 Apr
2013
6 Apr
'13
12:54 a.m.
Hello everyone!
So, I've got an E-MU 1212m PCIE (PCI express) soundcard now. And it doesn't
work. so there are a few questions.
First it's mebeing blid and not familiar with PCIE slots. There are two
cards. Youcan see the product from alsa-project.org or I can give you a direct
link to the manual off-list.
So in my PC I have one full PCI express slot (a long one, almost like a PCI
slot) and I've got a short one. A friend supposed PCIE 1.
there there are two cards a big one, the one with the digital I/O, which I
slotted into the small slot and a small card (the analogue I/O and MIDI I/O),
which I slotted into the long slot. So far so good.
My lspci gives me:
[...]
03:00.0 PCI bridge: Pericom Semiconductor Device e111 (rev 02)
04:04.0 Multimedia audio controller: Creative Labs SB0400 Audigy2 Value
I think these two are new, I didn't recognise them and they are right at the
end of the list.
I have compile ALSA modules for everything E-MU/creative labs (10K1, 10K1X,
the dell bit and 20K2). My ALSA version is 1.0.23 and kernel version is 2.6.38
(self-compiled).
Still ALSA doesn't give any new soundcard.
The two cards are coorectly connected with the IDE-like cable, but not
screwed in. Still they look firm enough.
Any ideas? Would that be the wrong slot for the big card (the main board of
the two, I think)? Or is the ALSA driver not detecting it correctly? Should
lspci show more?
Remember, I can't see myself and the sighted people around me are so NOT
into computers, that they aren't of any help really.
Thanks for a ny good advise!
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
6 Apr
6 Apr
3:11 a.m.
I guess you can't connect them to wrong slots, but it's likely that they
don't sit correctly in the slots. We sighted people often need to mount
the cards without good view too, but we at least see, if the angle does
look good, something that can be hard to feel out.
Ask the most self-confident "noop" just to size up, if the angle of the
card does look right-angled enough and if there's any latch that is not
correctly locked in place.
Perhaps you need to loosen the screws for other cards, to ensure that
there's less tension. I had to do this and I also had to bend the plate
where the cards are screwed.
11:14 a.m.
Julien Claassen wrote:
04:04.0 Multimedia audio controller: Creative Labs
SB0400 Audigy2 Value
This proves that the hardware is installed correctly (at least the main card).
Still ALSA doesn't give any new soundcard.
The snd-emu10k1 driver should have been loaded automatically.
Does it appear in "/proc/asound/cards"?
If not, what is the output of "rmmod snd-emu10k1; modprobe snd-emu10k1;
dmesg|tail"?
Regards,
Clemens
11:34 a.m.
Hello Clemens!
No it does not appear in /proc/asound/cards . and modprobe and rmmod won't
show anything, I've built the driver into the kernel. I have a monolithic
kernel. Sorry.
What else could I check or try?
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
11:38 a.m.
Julien Claassen wrote:
Hello Clemens!
No it does not appear in /proc/asound/cards . and modprobe and
rmmod won't show anything, I've built the driver into the kernel. I
have a monolithic kernel. Sorry.
What else could I check or try?
Any reason not to use a standard distro kernel for testing
purposes?
Warm regards
Julien
--
Joel Roth
6:03 p.m.
Hi Joel!
Well there are a couple of reasons for not using one of the Debian kernels.
I never configured anything to do with module loading. I think I acutally had
to change a few things years ago to suite my kernels. In those days standard
kernels weren't that good for low latency.
I think I might try the Arch Linux though. It has given me some trouble. But
my Delta was correctly detected. So I have the hope, that this will work with
the second one as well, if there's a chance at all.
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
6 p.m.
Hello Clemens!
The output of lspci -s 4:4 -v is:
04:04.0 Multimedia audio controller: Creative Labs SB0400 Audigy2 Value
Subsystem: Creative Labs Device 4007
Flags: medium devsel, IRQ 19
I/O ports at ec00 [size=64]
Capabilities: [dc] Power Management version 2
Should there be more, seeing, that there are actually two cards?
Warmly yours and thanks
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
6:14 p.m.
On Sat, April 6, 2013 9:00 am, Julien Claassen wrote:
Hello Clemens!
The output of lspci -s 4:4 -v is:
04:04.0 Multimedia audio controller: Creative Labs SB0400 Audigy2 Value
Subsystem: Creative Labs Device 4007
Flags: medium devsel, IRQ 19
I/O ports at ec00 [size=64]
Capabilities: [dc] Power Management version 2
Should there be more, seeing, that there are actually two cards?
Looking at the manual, it looks like one card doesn't really connect to
the MB, It is just extra board/connector space.
--
Len Ovens
www.OvenWerks.net
6:43 p.m.
Julien Claassen wrote:
The output of lspci -s 4:4 -v is:
04:04.0 Multimedia audio controller: Creative Labs SB0400 Audigy2 Value
Subsystem: Creative Labs Device 4007
Flags: medium devsel, IRQ 19
I/O ports at ec00 [size=64]
Capabilities: [dc] Power Management version 2
I had hoped for the kernel driver, but I guess lspci shows this only for
modules.
Is there a symlink "0000:04:04.0" in the
"/sys/bus/pci/drivers/snd_emu10k1"
directory?
Should there be more, seeing, that there are actually
two cards?
The add-on card is not a separate PCI device.
Regards,
Clemens
7:18 p.m.
Hello Clemens1
No, no 0000:04:04.0 . But te directory was:
/sys/bus/pci/drivers/EMU10K1_Audigy/
Same goes for the 10K1X directory.
I've also booted my ARch Linux, which is a bit of a pain, but it has a
normal 3.8.3 kernel and candetect hardware and load modules. Same there, I get
the info on lspci, but no entry in /proc/asound/cards . ALSA version on that
system is 1.0.26 .
I've also cleaned up the motherboard, just to be sure and checked, if the
card slotted correctly. It feels that way. If one of them seemed a little
doubtful it was the main board, not the daughter board.
Anything else, that I might test?
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
8:07 p.m.
On Sat, April 6, 2013 10:18 am, Julien Claassen wrote:
Hello Clemens1
No, no 0000:04:04.0 . But te directory was:
/sys/bus/pci/drivers/EMU10K1_Audigy/
Same goes for the 10K1X directory.
I've also booted my ARch Linux, which is a bit of a pain, but it has a
normal 3.8.3 kernel and candetect hardware and load modules. Same there, I
get
the info on lspci, but no entry in /proc/asound/cards . ALSA version on
that
system is 1.0.26 .
I've also cleaned up the motherboard, just to be sure and checked, if
the
card slotted correctly. It feels that way. If one of them seemed a little
doubtful it was the main board, not the daughter board.
Anything else, that I might test?
There seems to be some indication that it matters what other pci(e) cards
are installed. Some MB have only so many PCIe channels. For example, A MB
might have 4 PCIe channels and then have a 4 channel slot a 2 channel slot
and 1 or 2 1 channel slots. Normally, the video card goes in the widest
slot, but you may like to try putting in a narrower slot to see if that
frees up some bandwidth. This will slow down the video card some, but in
this case the audio is important. Just as a test try swapping the slots
where you video and audio cards are.
--
Len Ovens
www.OvenWerks.net
8:36 p.m.
Hello Len!
If the number of PCIE channels on my board is a problem, then I won't
overcome it. I only have a built-in graphics card (no need for it :-) ), my
PCI MAudio Delta 1010LT and then there is one more PCI slot and those two PCIE
slots. The two cards only fit in one constallation. One PCIE slot is long,
theother is short. The mainboard (with the digital I/Os) fits into the very
short one and the daughter board fits into the long one (alghough it only has
a small piece right at the end, which fits into the slot.
I'll try to find more info on my board and see, if that might be the case.
Kind regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
9:08 p.m.
On Sat, April 6, 2013 11:36 am, Julien Claassen wrote:
Hello Len!
If the number of PCIE channels on my board is a problem, then I won't
overcome it. I only have a built-in graphics card (no need for it :-) ),
my
PCI MAudio Delta 1010LT and then there is one more PCI slot and those two
PCIE
Ok that should not be the case then I had just heard of video card
problems. Have you tried without the delta installed? (my guess is yes)
--
Len Ovens
www.OvenWerks.net
10:11 p.m.
Hi Len1
Your guess is wrong. I haven't uninstalled my Delta. It's an old card and I
have the feeling, that it's getting a little picky. It's practically a
grandfather with its eleven or maybe twelve years now. :-)
Best wishes
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
11:34 p.m.
Can you access the BIOS configuartion with braille? Perhaps there's some voodoo by the
BIOS. I don't have a clue what it could be, but since you now can't get the card
working for hours, taking a look at the BIOS settings couldn't harm. Or did you
already check the BIOS?
7 Apr
7 Apr
12:08 a.m.
Hello Ralf!
It MIGHT be an idea. Although, I don't see why. Well and I won't see, so
will nobody else. No screen... :-( I dispensed with a screen long ago. Too
much space taken up, that could better be taken by a tape deck. Anyway: I got
this board new and I can't see, why anyone in their right mind would like to
disable anything about PCI or PCIE slots. Nothying, that would prevent people
from putting hardware into them. I would have been more suspicious of the PCI
confiruation, since I got this board only two years ago.
Kind regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
1:58 a.m.
New subject: PCIe: E-MU 1212m soundcard installation help for the blind
On Sun, 2013-04-07 at 00:08 +0200, Julien Claassen wrote:
No screen... :-( I dispensed with a screen long ago.
From time to time it can be important to access the
BIOS configuration.
Latest drawback of my RME PCIe card:
Jack only does start with the sample rates 22050, 32000, 44100, 48000.
It doesn't run with the sample rates 88200, 96000, 192000.
:D
The other issues still are xruns, there's no setting without xruns and
only 2 of 8 ADAT channels are available. HDSP conf doesn't work and ALSA
mixer doesn't work correctly, just HDSP mixer is ok, excepted of the
ADAT issue.
I assume you made inquiries about your card?
That's a serious issue for Linux, I made inquiries but got to much wrong
information, not only for my RME card.
I had a modem that was listed for all board revisions as usable with
Linux and when I got a board revision that didn't work and sent it as
information, I got the reply, that the board version I got, was known as
not working with Linux, it just wasn't listed, because this board
revision is very seldom.
I shouldn't have wait that long with my card, but instead I should have
sent it back to the dealer and bought another card.
If you can't get your card working, don't wait to long, within 14 days
you can give it back.
12:10 p.m.
On Sun, 2013-04-07 at 11:43 +0200, Clemens Ladisch wrote:
Ralf Mardorf wrote:
I know and IMO your troubleshooting is the best method to find the cause
for the issue, but since I don't use braille, it's easy for me to take a
look at the BIOS configuration. Just to be absolutely safe, I would take
a look for "voodoo" settings, e.g. an option, "disable/enable E-MU audio
card support for PCIe" ;), I also would consider to update the BIOS.
That Linux can "see" hardware, that is known to work with Linux, doesn't
mean that it will work with each mobo and/or BIOS version. If a flat
ribbon cable is involved or something similar, I even would replace a
brand new cable if something shouldn't work. It's quiet possible to get
unstable connectivity, I for example hot glued SATA connectors to get
rid of issues, because the connectors loosed after a while.
That for Julien it's not that easy to take a look at the BIOS
configuration is another issue, but everybody else in this situation
should spend a few seconds, just to ensure, that nothing "capricious" is
caused by the BIOS.
Perhaps there's some voodoo by the BIOS.
PCIe accesses work, as proven by the lspci output.
The problem probably is with the driver.
11:46 a.m.
Julien Claassen wrote:
No, no 0000:04:04.0 .
Any error message is logged when booting.
Try "dmesg | grep -i emu" or "grep -i emu /var/log/messages".
I've also booted my ARch Linux, which is a bit of
a pain, but it has a normal 3.8.3 kernel and candetect hardware and load modules.
Okay, here "dmesg | tail" should show something after "rmmod snd-emu10k1;
modprobe snd-emu10k1".
Regards,
Clemens
11:50 a.m.
Hello Clemens!
OK, we're getting there, this is the output from my Debian (2.6.38
self-compiled) dmesg:
[ 2.726552] EMU10K1_Audigy 0000:04:04.0: PCI INT A -> Link[LNEB] -> GSI 19
(level, low) -> IRQ 19
[ 3.739893] EMU10K1_Audigy 0000:04:04.0: PCI INT A disabled
[ 3.739900] EMU10K1_Audigy: probe of 0000:04:04.0 failed with error -5
does that look, as if Ralf might have a point there and there's something
disabled or badly configured in my BIOS afterall?
Kind regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
12:45 p.m.
On Sun, 2013-04-07 at 11:51 +0200, Julien Claassen wrote:
does that look, as if Ralf might have a point there
and there's something
disabled or badly configured in my BIOS afterall?
If somebody will help out with a monitor within the next hours, than no,
if nobody can assist with a monitor for the next three month, than yes.
6:47 p.m.
Julien Claassen wrote:
EMU10K1_Audigy 0000:04:04.0: PCI INT A ->
Link[LNEB] -> GSI 19 (level, low) -> IRQ 19
EMU10K1_Audigy 0000:04:04.0: PCI INT A disabled
EMU10K1_Audigy: probe of 0000:04:04.0 failed with error -5
does that look, as if Ralf might have a point there and there's
something disabled or badly configured in my BIOS afterall?
No, this has nothing to do with the BIOS. (Once a PCE-E device shows
up correctly in lspci, it's almost impossible for the BIOS to breaks
things.)
What this shows is that the driver aborts during its initialization.
The only place where it does this with error code -5 but without
outputting any specific error message is in emufx.c in the functions
snd_emu10k1_*init_efx(), when a consistency check for the number of
allocated registers/RAM locations fails.
I'd guess that some change to the DSP program was never tested with the
particular configuration of the PCIe 1212m.
I don't know how the Emu10k1's DSP is programmed, or how the program
could be changed to get these pointers correct. Mikael? James?
Regards,
Clemens
8 Apr
8 Apr
1:58 a.m.
Hello Clemens!
Thank you for your very valued and valuable guidance here.
Just to be sure: could it also be due to a contact not really connecting? I
have a feeling, the card might not be a 100% tight. The chassis of my pc is
quite old and thus there might be a little doubt due to the screws. When you
said, that my board showing up, meant, that it was correctly slotted, I didn't
bet on that.
I suppose beyond waiting there's no more I can do right now, to help, short
of sending my card to a developer, so they can have some "fun" in testing
code. :-)
Best wishes
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
11:08 a.m.
Julien Claassen wrote:
Just to be sure: could it also be due to a contact not
really connecting?
No, the contacts work fine. PCI Express is a quite complex protocol,
but all communication uses the same packet format over a single serial
link, so when lspci shows correct information, all other things like
register accesses, bus-master DMA, and interrupts will work.
The problem is that the driver does not program the DSP correctly
(according to its internal consistency checks); it appears different
parts of the driver have different ideas of how many channels there can
be.
I suppose beyond waiting there's no more I can do
right now, to help,
short of sending my card to a developer, so they can have some "fun"
in testing code. :-)
As far as I can tell, the PCI and PCIe versions of the 1212m are
identical as far as the driver is concerned, so we just have to find
a developer with some 1212m.
Regards,
Clemens
11:30 a.m.
On Mon, 2013-04-08 at 11:08 +0200, Clemens Ladisch wrote:
PCI Express is a quite complex protocol,but all
communication uses the
same packet format over a single serial link, so when lspci shows
correct information, all other things like register accesses,
bus-master DMA, and interrupts will work.
Thank you for the information.
11:17 a.m.
(let's try a different e-mail for Pavel)
Julien Claassen wrote:
Just to be sure: could it also be due to a contact not
really connecting?
No, the contacts work fine. PCI Express is a quite complex protocol,
but all communication uses the same packet format over a single serial
link, so when lspci shows correct information, all other things like
register accesses, bus-master DMA, and interrupts will work.
The problem is that the driver does not program the DSP correctly
(according to its internal consistency checks); it appears different
parts of the driver have different ideas of how many channels there can
be.
I suppose beyond waiting there's no more I can do
right now, to help,
short of sending my card to a developer, so they can have some "fun"
in testing code. :-)
As far as I can tell, the PCI and PCIe versions of the 1212m are
identical as far as the driver is concerned, so we just have to find
a developer with some 1212m.
Regards,
Clemens
1:58 p.m.
I didn't read it completely myself, since I've got my own issues to
solve, but perhaps it's related, even if it's an USB device here:
-------- Forwarded Message --------
From: Torstein Hegge <hegge(a)resisty.net>
To: huwshimi <huwshimi(a)gmail.com>
Cc: alsa-user(a)lists.sourceforge.net
Subject: Re: [Alsa-user] AudioBox 44VSL bad descriptors patch
Date: Mon, 8 Apr 2013 13:39:05 +0200
On Mon, Apr 08, 2013 at 21:06:39 +1000, huwshimi wrote:
Hi,
I have a Presonus AudioBox 44VSL that is failing with bad descriptors:
[ 246.649802] usbaudio: unit 51: unexpected type 0x09
[ 246.649859] snd-usb-audio: probe of 2-1.2:1.0 failed with error -5
This seems to be a problem people have had with other devices and with
some
help in #alsa I have made the attached patch that
fixes the problem.
More details of the device and errors in this thread:
http://linuxmusicians.com/viewtopic.php?f=6&t=7738
This is fixed in 3.9-rc4 and newer kernels, and the fix will also be
included in the next Ubuntu Quantal kernel update.
See the recent "ALSA and SCARLETT FOCUSRITE 8i6" thread for details.
Torstein
9 Apr
9 Apr
5:41 p.m.
Hello!
I have just compiled a 3.9.0-rc6 kernel to see what happens.
As far as I could read from /var/log/messages afterwards I got an error -2.
I couldn't do much more in the system itself, for I had miscompiled something
and no braille support at the time. this will take a while to fix. It's been a
while, since I did a kernel. :-)
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
12 Apr
12 Apr
8:06 p.m.
Hello again!
I was just wondering, if anything new had come to light? I have to keep
track of time. Since if this soundcard turns out unusable, I have to send it
back. As you might imagine, there is a limited time period.
I know, that theoretically it shouldn't be a problem, given that some work
will be put into it, but most of the developers do this in their sparetime and
sparetime is limited as well.
So could you give me a guestimate? Do you think, that you will have the
time, motivation and resorces to get the E-MU 1212m PCIE working in the near
future or not.
Thanks and sorry for the lengthy mail.
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
10:39 p.m.
Julien Claassen wrote:
I was just wondering, if anything new had come to
light?
I don't know why 3.9 behaves differently, but one reason for the "-2"
error might be that the alsa-firmware package is not installed.
(There should be files in the directory /lib/firmware/emu/.)
Regards,
Clemens
10:59 p.m.
Hello Clemens!
There is no emu directory in /lib/firmware? Is it a special kernel option,
that I need? Or is it an extra package, that is required?
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
13 Apr
13 Apr
12:14 a.m.
Hello Clemens!
I'm sorry. I didn't even thik about the firmware before. It is installed
now, but doesn't make a difference to the 2.6.38 kernel. the error is still -5
and I don't get any other messages.
I may try the 3.9 kernel, but not tonight, too late and it's a hassle, since
the braille display still isn't working. So I just have to log a bit and
reboot.
Thanks and goodnight
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
16 Apr
16 Apr
6:31 p.m.
New subject: [solved] E-MU 1212m soundcard installation help for th blind
Hello!
OK, the card is now working under Linux 3.9.0-rc6. There were a few oddities
along the way.
The driver works under 3.9.0-rc6, but not under 2.6.38, using exactly the
same configuration.
The trick: It seems, that the emu10k1 driver should definitely not be built
into the kernel. I tried the same configurations only with the module built-in
and modularised. Modprobe from the commandline now works.
Then there re some questions about its workings and configuration in
alsamixer, but I'll ask those seperately.
Thanks very much for all the help and support.
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
9:04 p.m.
New subject: [solved] E-MU 1212m soundcard installation help for th blind
Julien Claassen wrote:
It seems, that the emu10k1 driver should definitely
not be built into the kernel.
Now I remember: When the driver is not built as a module, the firmware
must also be compiled into the kernel.
Regards,
Clemens
6 Apr
6 Apr
11:39 a.m.
On Sat, 2013-04-06 at 11:14 +0200, Clemens Ladisch wrote:
Julien Claassen wrote:
Yesno, on my machine there is clearance, it's possible that not all
contacts are ok, regarding to the angle. If the slots were unused for
month even dust could be an issue for some contacts. It's less likely
and more likely that it is installed correctly, but it's not ensured
just by this information.
04:04.0 Multimedia audio controller: Creative
Labs SB0400 Audigy2
Value
This proves that the hardware is installed correctly (at least the
main card).
4049
days inactive
4060
days old
linux-audio-user@lists.linuxaudio.org
37 comments
5 participants
participants (5)
-
Clemens Ladisch -
Joel Roth -
Julien Claassen -
Len Ovens -
Ralf Mardorf