I think my biggest problem is my trouble explaining things. I sent this to
the ALSA group and they say it could be USB drivers. This is copied from
the ALSA bugtrak archive.
When I plug to USB sound cards into my VIA (Avertec 3150H) based laptop or
(HP Pavillion 305w) desktop, and I begin using them (both streaming
audio), all applications begin core dumping. I am not able to log anything
that happens because the system stops logging and sometimes the system
will freeze totally. I have used every version of 2.6 kernel and Alsa
drivers 9.0+ up to 1.4.X. With one USB sound card everything is fine, with
2 all hell breaks loose. I copied all the stuff from the bug tracking from
ALSA, they said it could be a USB or hardware issue. Works fine in
winders, so it is hard to say. ;-). I have both EHCI and UHCI drivers
enabled, I have tried enabling and disabling various options including
expirimental ones.
I have tried just using UHCI drivers, which at first neither UHCI or EHCI
would work because of PCI IRQ assignment problems, which seem to be fixed
when I changed the power management scheme in linux from APM to ACPI,
something I (bug report which he let me know to send here from now on)
sent to Alan Stern earlier this year that I was able to resolve by
changing that scheme and that allowed my USB devices to be correctly
configured to the correct IRQs--so I don't know if this is related or not
to PCI irq assignment which now seems to be correctly working?
;-)
I hope this is clear enough:
Here are the ALSA corrispondance below.
USB sound works great in ALSA except when I try to run two USB sound
cards. I use it for portable sound recording for guitar fx processing. Any
application using two different DSPs will start causing applications to
crash (applications start core dumping.. such as gnome, logging, and other
such things), until the entire system is unusable. Tried different kernel
versions and alsa versions. regular OSS sound driver does not seem to work
with mp3+ so I can't even verify if it is the kernel or alsa or both. I
have been able to use the built in sound card (VIA dsp) and one USB sound
card with no problems. so it has nothing to do with 2 sound cards. but if
I compile in with just the USB driver or USB driver and VIA driver and
plug in two sound cards the system crashes.
Additional Information I am using UHCI and EHCI HCD. I have tried
disabling EHCI, enabling EHCI with high speed transactions patch, bandwith
alocation. The standard USB audio driver does not work with my soundcard
(sound blaster mp3+ USB) so I have not been able to test with that. This
problem is reproducable with any vanilla 2.6.X kernel with just EHCI and
UHCI enabled. I am using a avertec laptop with a VIA chipset and mobil
athlon. (model 3150H), it does the same on my hp desktop hp pavillion 304w
via chipset.
Distribution Mandrake 10.0
Kernel Version 2.6.5 rc2 mm5 (and 2.6 kernel version)
Attached Files config.gz (5,675 bytes) 03-30-04 07:17
Relationships
Upload File
Select File
(Max size: 5,000k)
Add Bugnote
Bugnote
Bug Notes
Clemens Ladisch
04-05-04 16:11
What kind of crash? Are there any interesting messages in the system log
/(var/log/messages)?
Does this happens with a 2.4.x kernel, too?
paperclip
04-08-04 22:24
[ Edit ][ Delete ] Nothing shows up in the /var/log/messages or
/var/log/syslog section when I have syslog daemon running when the
computer crashes. I have not tried it with 2.4 kernel, because I had
trouble getting alsa working well in 2.4 in the first place, so I moved to
2.6. This is something strange that I discovered when testing it, is if I
connect the USB sound cards after I have already started up and everything
is loaded, often it won't crash. but if I have them attached when I start
up either of my computers, the computer will initialize them and everthing
will seem fine and when I go to use them (use a sound related program) the
system will lock up, either totally freeze or it will become totally
unusable. When I look at the home directory of the user I was logged in
under a whole bunch of core files will be there. Here are some lines from
the /var/log/messages that I think might shed some light..
Apr 8 00:55:13 localhost kernel: ACPI: Interpreter enabled
Apr 8 00:55:13 localhost kernel: ACPI: Using PIC for interrupt routing Apr
8 00:55:13 localhost kernel: ACPI: PCI Root Bridge [PCI0] (00:00) Apr 8
00:55:13 localhost kernel: PCI: Probing PCI hardware (bus 00) Apr 8
00:55:13 localhost kernel: ACPI: Embedded Controller [ECD] (gpe 4) Apr 8
00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKA] (IRQs 4 *11 14
15)
Apr 8 00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKB] (IRQs 3 5
9 14 15)
Apr 8 00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKC] (IRQs *10
12)
Apr 8 00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKD] (IRQs 6 7)
Apr 8 00:55:13 localhost kernel: SCSI subsystem initialized
Apr 8 00:55:13 localhost kernel: drivers/usb/core/usb.c: registered new
driver
hub
Apr 8 00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKC] enabled
at IRQ 10
Apr 8 00:55:13 localhost kernel: ACPI: PCI Interrupt Link [LNKD] enabled
at IRQ 6
Apr 8 00:55:13 localhost kernel: ACPI: No IRQ known for interrupt pin A of
device 0000:00:11.1 - using IRQ 255
When I did an lspci -vvv and looked for that device 0000:00:11.1 I noticed
that it is one of my USB controlers. It is assigning an odd IRQ. That is
the only suspicous thing that I see off hand.
paperclip
04-08-04 22:31
[ Edit ][ Delete ] Ok.. I was wrong... I was looking at the wrong line..
here is the lspci output:
[EMAIL PROTECTED] log]# lspci
00:00.0 Host bridge: VIA Technologies, Inc. P/KN266 Host Bridge
00:01.0 PCI bridge: VIA Technologies, Inc. VT8633 [Apollo Pro266 AGP]
00:0a.0 CardBus bridge: O2 Micro, Inc. OZ6912 Cardbus Controller
00:10.0 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.1 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.2 USB Controller: VIA Technologies, Inc. USB (rev 80)
00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT8233/A/C/VT8235 PIPC Bus Master IDE (rev 06)
00:11.5 Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235
AC97 A udio Controller (rev 50)
00:11.6 Communication controller: VIA Technologies, Inc. Intel 537 [AC97
Modem](rev 80)
00:12.0 Ethernet controller: VIA Technologies, Inc. VT6102 [Rhine-II] (rev
74)01:00.0 VGA compatible controller: S3 Inc. VT8375 [ProSavage8
KM266/[EMAIL PROTECTED] log]#
That device is the IDE controller..
paperclip
04-20-04 00:15
[ Edit ][ Delete ] actually... the crashes still happen with two usb
sound cards plugged in after startup. (just does not happen as quickly --
both sound cards have to be streaming audio) I have not found anything in
the logs to show any kind of precursor thing. Infact I don't even see any
error messages before it happens. I don't know what else I should check.
Any ideas?
The type of crash experienced is either the system will totally freeze, or
all applications will start creating core dump files. The window manager
usually crashes first.. the gnome desktop bars, then X will crash leaving
core (the system does not become unstable until I stream audio with both
USB sound cards) files in my home directory. I am unable to shutdown if I
can get to the command line because every application I run core dumps. I
hope that is more clear. when I look in /var/log/messages, none of the
applications that crash show any errors in the log. Infact, the system
stops logging anything during the crash. I can reproduce this with any 2.6
kernel. I need to experiment with 2.4 and get sound working, but I would
not be suprised if any two usb sound cards will cause this same problem. I
don't have any others except for the two sound blaster mp3+ to test. You
have to be streaming audio on both cards; -- I use them full duplex with
guitar software, or with audacity--recording with one while playing the
other tracks out on the other--which does cause the crash.
Let me know if you need something else?
-Ron
Clemens Ladisch
04-22-04 15:39
With every application you run dumping core, this looks as if lots of
memory gets corrupted.
This may be a bug in the USB drivers, but I think this is more likely to
be a hardware bug (not with your MP3+es but somewhere in the computer).
VIA chipsets aren't known for reliability when there are high loads on the
PCI bus.
paperclip
04-22-04 22:06
[ Edit ][ Delete ] I don't have a non VIA chipset for testing this
either. Both my desktop and my laptop use VIA chipsets--I hope someone
else could test this at some point ;-) (PLEASE). I suspected USB first
actually, but have not had problems with other USB devices. I did have
problems getting the system to the detect USB bus in the first place (PCI
IRQ PROBLEMS) and I was able to get it to finally work by changing from
APM to ACPI and setting the PCI access to direct. I will work on getting
the OSS USB driver working and test it to see if that causes similar
problems. I suspected the ALSA driver only because I have not had problems
with other USB devices, including a USB Zip drive. Also, I use a USB
mouse--which I tried disabling and enabling in the kernel as well as Mass
storage to see if anything was causing a problem or conflict. It seems
like some sort of buffer overflow (Array out of bounds or screwy pointers,
when timing gets screwy during a lot of system load?), which would make
sense that it could be USB, but I would think we would see problems with
other devices? With a single USB sound card it works flawlessly, which
still makes me wonder. I am sure it is a headache on your end because you
don't control the USB (EHCI, OHCI, UHCI) driver end of things--still
without anyway to test other chipsets, I would hope someone reading these
would have two USB audio devices to try? Some people not loading their
system as much may not notice these crashes because it could be some sort
of concurrency issue (example using the same memory location and trashing
the contents). Card 1 caches something, Card 2 tries to use the same
memory location that Card 1 is using at the same time, trashing the
memory. I have seen this sort of thing with servers. In this instance I
could be totally wrong. I know it is just a binary stream out to a /dev
device, like a dsp, so it may not even work that way. I don't know.
Oh well =-),
-Ron
-------------------------------------------------------
This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
For a limited time only, get FREE Ground shipping on all orders of $35
or more. Hurry up and shop folks, this offer expires April 30th!
http://www.thinkgeek.com/freeshipping/?cpg=12297
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
