Menu

Freezes

Dan Mayr
2010-04-08
2013-06-13
  • Dan Mayr

    Dan Mayr - 2010-04-08

    Hi there!
    I received my pod xt today and was very excited to try it and test the line6usb driver.
    So far I got everything working fine: alsa, pulse, jackd, ardour.
    I use the pod as input and output device and even the sample rate issue seems to be of no problem.

    However, I already got 3 freezes today which is something that should not happen, especially on a system considered stable.
    Can you guide me to tracking down the reason for those freezes?
    What is your experience with the stability of the line6usb driver?
    It appears that the freezes occurred while I was playing the guitar and using the (connected) floorboard.
    I did not record while this was happening nor was I playing back sound from the computer.

    Any help is much appreciated!
    Thanks,
    Dan

    P.S. Keep up the good work!

     
  • Dan Mayr

    Dan Mayr - 2010-04-08

    Oh btw., I am using the currect svn version (Rev 601) of the driver, compiled from sources on a Ubuntu 9.10 machine with stock 2.6.31 kernel (no rt) and alsa 1.0.20 and jack 0.116.1

     
  • Dan Mayr

    Dan Mayr - 2010-04-11

    Hi again,

    I did some more testing and
    1. the good news is, everything seems to work fine with my thinkpad running Debian lenny and line6usb 0.8.1 !
    2. On my other machine deinstalling pulseaudio seems to have improved stability of the driver, too.
    3. However, hibernation doesn't work anymore using either 0.8.1 or 0.9 (R601). The screen gets black but from there nothing happens so I have to reboot. Only workaround I found is to unload the line6usb driver before hibernation (which requires killing all audio applications beforehand). Interesting is that I can hibernate on the Thinkpad without unloading the driver.
    4. Debian squueze and Ubuntu 9.10 often crash when I try to unload line6usb 0.9 (R601). 0.8.1 unloads fine.
    5. The kernel gives me some warnings. These repeat periodically every 2 seconds as long as line6usb (0.8.1 or 0.9) are loaded:

    [  950.377095] BUG: scheduling while atomic: swapper/0/0x10000100
    [  950.377099] Modules linked in: line6usb fuse snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_midi snd_rawmidi snd_seq_midi_event snd_seq snd_timer snd_seq_device nvidia(P) i2c_nforce2 snd agpgart i2c_core soundcore applesmc led_class btusb input_polldev shpchp bluetooth snd_page_alloc pci_hotplug rfkill pcspkr evdev processor ext3 jbd mbcache hid_apple usbhid hid sg sr_mod sd_mod cdrom crc_t10dif ide_pci_generic ide_core ata_generic ohci_hcd ahci ssb libata mmc_core pcmcia firewire_ohci firewire_core crc_itu_t button ehci_hcd forcedeth pcmcia_core scsi_mod usbcore nls_base thermal fan thermal_sys [last unloaded: line6usb]
    [  950.377170]
    [  950.377174] Pid: 0, comm: swapper Tainted: P           (2.6.32-trunk-686 #1) Macmini3,1
    [  950.377178] EIP: 0060:[<f80de0d7>] EFLAGS: 00000296 CPU: 1
    [  950.377185] EIP is at acpi_idle_enter_bm+0x246/0x281 [processor]
    [  950.377188] EAX: c13ef9a8 EBX: 0000001c ECX: 000000dd EDX: 01214000
    [  950.377191] ESI: 00000000 EDI: 00000008 EBP: 00000000 ESP: f6c71f78
    [  950.377194]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
    [  950.377202] CR0: 8005003b CR2: b66a89a0 CR3: 013f9000 CR4: 000406d0
    [  950.377204] DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
    [  950.377206] DR6: ffff0ff0 DR7: 00000400
    [  950.377208] Call Trace:
    [  950.377215]  [<c11b5421>] ? cpuidle_idle_call+0x68/0xbb
    [  950.377219]  [<c1002388>] ? cpu_idle+0x89/0xa5
    

    If there is anything I can do to help you debugging this just let me know!

     
  • Markus Grabner

    Markus Grabner - 2010-04-12

    Could you please tell me exactly which device you are using (i.e., the output of "lsusb | grep Line6")? And did you try the most recent driver version (R604)? I made some changes which might be related to the periodically repeating messages. BTW, the snippet from your log file is cut off at the point where it becomes interesting :-), is there more information available below " Call Trace:"?

    Kind regards,
    Markus

     
  • Dan Mayr

    Dan Mayr - 2010-04-15

    Hi Markus!

    Thanks for your reply.
    I do have a PodXT (no pro, no live).
    The output of lsusb  is:
    Bus 003 Device 005: ID 0e41:5044 Line6, Inc. PODxt

    I tried the R605 version of the driver however the problem persits.
    I also noted a system freeze, probably caused by an un-killable jackd running with the pod as audio interface and consuming nearly 99% cpu time.

    There were no more lines in that particular call trace. However I found a line which is often followed by the previously posted trace in one of the system logs:
    [ ? start_kernel+0x30c/0x311

    Any ideas?
    Best regards,
    Dan

     
  • Markus Grabner

    Markus Grabner - 2010-04-15

    Do the periodic warnings (your issue #5) still appear with R605? If they do, you can try the debug version at "https://line6linux.svn.sourceforge.net/svnroot/line6linux/driver/branches/debug". It is identical to the "trunk" driver, but outputs a (somewhat) detailled trace in the syslog, which might indicate where the warning is triggered.

    It would be extremely helpful if you could find out the exact driver revision at which problem #4 appears for the first time. This is straightforward, although quite tedious: pick any revision somewhere in the middle between 547 (0.8.1 release) and the current revision 605. If that revision has the same problem, try again with a lower revision, otherwise with a higher one, and so.

    Kind regards,
    Markus

     
  • Dan Mayr

    Dan Mayr - 2010-04-17

    The periodic warnings are always there, from R547 to R606.
    However, R547 seems much more stable on my system.
    I am using the debug version of R606 right now and the call trace remains the same for most of the time.
    However, I also encountered the following trace:

       ? __scsi_put_command+0x38/0x51
       ? scsi_next_command+0x1e/0x2f
       ? scsi_io_completion+0x373/0x394
       ? scsi_finish_command+0xaa/0xc2
       ? blk_done_softirq+0x53/0x5f
       ? __do_softirq+0xaa/0x151
       ? do_softirq+0x31/0x3c
       ? irq_exit+0x26/0x58
       ? do_IRQ+0x78/0x89
       ? common_interrupt+0x30/0x38
       ? acpi_idle_enter_bm+0x246/0x281
       ? cpuidle_idle_call+0x68/0xbb
       ? cpu_idle+0x89/0xa5
       ? start_kernel+0x30c/0x311

    Although it doesn't look related to line6usb imo, I have only seen those error messages when line6usb is loaded.

    Here is another one (likely more related):
    Apr 17 15:41:44 mini kernel:  BUG: scheduling while atomic: ksoftirqd/0/4/0x10000100
    Apr 17 15:41:44 mini kernel:  Modules linked in: line6usb snd_seq_dummy fuse snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_seq_midi snd_pcm_oss snd_rawmidi snd_mixer_oss snd_seq_midi_event snd_seq snd_pcm btusb bluetooth nvidia(P) snd_timer rfkill agpgart snd_seq_device i2c_nforce2 snd i2c_core shpchp soundcore snd_page_alloc pci_hotplug applesmc led_class input_polldev evdev pcspkr processor hid_apple ext3 jbd mbcache usbhid hid sg sr_mod sd_mod cdrom crc_t10dif ide_pci_generic ide_core ata_generic ssb mmc_core ahci pcmcia ohci_hcd firewire_ohci firewire_core crc_itu_t pcmcia_core forcedeth libata button scsi_mod ehci_hcd usbcore nls_base thermal fan thermal_sys
    Apr 17 15:41:44 mini kernel:  Pid: 4, comm: ksoftirqd/0 Tainted: P        W  2.6.32-trunk-686 #1
    Apr 17 15:41:44 mini kernel:  Call Trace:
    Apr 17 15:41:44 mini kernel:    ? schedule+0x7e/0x7dc
    Apr 17 15:41:44 mini kernel:    ? update_curr+0x106/0x1b3
    Apr 17 15:41:44 mini kernel:    ? update_curr+0x106/0x1b3
    Apr 17 15:41:44 mini kernel:    ? _cond_resched+0x25/0x3c
    Apr 17 15:41:44 mini kernel:    ? __kmalloc+0x8e/0x128
    Apr 17 15:41:44 mini kernel:    ? line6_alloc_sysex_buffer+0x1b/0x6c
    Apr 17 15:41:44 mini kernel:    ? line6_alloc_sysex_buffer+0x1b/0x6c
    Apr 17 15:41:44 mini kernel:    ? pod_startup_timeout+0x0/0x8a
    Apr 17 15:41:44 mini kernel:    ? pod_startup_timeout+0x53/0x8a
    Apr 17 15:41:44 mini kernel:    ? run_timer_softirq+0x16a/0x1eb
    Apr 17 15:41:44 mini kernel:    ? __do_softirq+0xaa/0x151
    Apr 17 15:41:44 mini kernel:    ? do_softirq+0x31/0x3c
    Apr 17 15:41:44 mini kernel:    ? ksoftirqd+0x44/0xa5
    Apr 17 15:41:44 mini kernel:    ? ksoftirqd+0x0/0xa5
    Apr 17 15:41:44 mini kernel:    ? kthread+0x61/0x66
    Apr 17 15:41:44 mini kernel:    ? kthread+0x0/0x66
    Apr 17 15:41:44 mini kernel:    ? kernel_thread_helper+0x7/0x10

    I first thought it might be an irq issue so I tried another port usb port but without success.

     
  • Markus Grabner

    Markus Grabner - 2010-04-18

    >   ? __scsi_put_command+0x38/0x51
    These are not (obviously) related to the Line6 driver - no idea where they come from…

    > Apr 17 15:41:44 mini kernel:  BUG: scheduling while atomic: ksoftirqd/0/4/0x10000100
    This one is very interesting, and fortunately the syslog tells us where it happens:

    > Apr 17 15:41:44 mini kernel:   ? line6_alloc_sysex_buffer+0x1b/0x6c
    It looks like you are running out of physical memory on your machine while the Line6 device is being used. I never encountered this bug since I have 8GB of RAM in my development machine :-) I tried to blindly fix this, though I can't confirm it since - as I said - I didn't have the problem before. Could you please try the most recent driver version (the proposed fix is both in the trunk and in the debug branch)?

    Thanks & kind regards,
    Markus

     
  • Markus Grabner

    Markus Grabner - 2010-04-18

    BTW, I just noticed another thing in your syslog:
    > Apr 17 15:41:44 mini kernel:   ? pod_startup_timeout+0x0/0x8a

    There is no "pod_startup_timeout" in the current driver version, so you are not using the most recent version. Can you please run the command

    grep "line6usb driver version" /var/log/messages

    to identify the exact version of the Line6 driver loaded on your system?

    Thanks & kind regards,
    Markus

     
  • Dan Mayr

    Dan Mayr - 2010-04-20

    Hi Markus!

    A first remark on the current driver version:
    The error message did not go away. I start thinking this is rather caused by the kernel than line6usb, although it's weird that it only occurs if the driver is loaded. But here is something similar: http://www.mail-archive.com/debian-kernel@lists.debian.org/msg51554.html. A lot of people seem to receive similar call traces wit 2.6.30+ kernels.

    What I noticed is that with the current driver versions sometimes alsa fail to detect the pod as soundcard. So it is not listed in /proc/asound/cards although line6usb is loaded. Reloading (rmmod + modprobe) the driver doesn't help but switching back to 0.8.1 resolves this issue.

    I also noticed that sometimes the sound from the pod starts lagging including the sound from the guitar. And it even happend to me that all of a sudden the current channel of the pod got changed. Don't know if this is caused by the driver or just a malfunctioning pod..

    Thanks!
    Dan

     
  • Dan Mayr

    Dan Mayr - 2010-04-20

    Just reproduced the channel switching thing a few times with loaded line6usb but couldnt reproduce it with unloaded line6usb.
    To reproduce it selected a high gain amp model and produced some pinch harmonics and the pod would always switch the channel 3A.  Could this be an issue with clipping detection/handling?
    Driver version was 0.8.1

     
  • Markus Grabner

    Markus Grabner - 2010-04-21

    > What I noticed is that with the current driver versions sometimes alsa fail to detect the pod as soundcard.
    I modified the startup procedure to only create ALSA entries after the device is fully initialized, which takes a couple of seconds. So if you check ALSA soundcards immediately after inserting the line6usb module, your Line6 device won't show up. Make sure to wait 5-10 seconds. Another reason could be a different startup behaviour of the PODxt. In any case, it would be useful to see the USB messages travelling over the bus. The driver contains debugging code to dump these messages to the syslog. Please use the debug branch (see message #6 above) at revision >= 616 and proceed as follows:
    *) type "make xconfig"
    *) search for "Line6" (menu "Edit->Find")
    *) check the box "dump control messages"
    *) save the configuration and exit the configuration tool
    *) recompile and reinsert the driver
    Please post your syslog from inserting the driver until ~10 seconds later. Moreover, if the driver is responsible for switching the channel, the corresponding messages should also show up in the syslog, which is best observed in a "live view" of the log file (i.e., "tail -f /var/log/messages"). If you find these codes in the syslog, please post them as well.

    BTW, in which timezone are you located? It's probably more efficient to have a skype chat.

    Kind regards,
    Markus

     
  • Dan Mayr

    Dan Mayr - 2010-04-22

    Hi!

    I guess that would be really easier :)
    My timezone is GMT+1, so we should be able to find a common spot :)
    Maybe Friday after 18.00 or Saturday afternoon? Unfortunately I dont have a headset (but my soundcard's mic-in is not working anyways) so I will have text chat.

    I ran into some troubles following your xconfig guide.
    First, the directory  $(KDIR)/source/$(SUBDIR) does not exist, so I created it.
    However make xconfig still fails with this error message:
    make: Entering directory `/usr/src/linux-headers-2.6.32-trunk-686'
    /usr/src/linux-headers-2.6.32-trunk-common/scripts/Makefile.build:44: /usr/src/linux-headers-2.6.32-trunk-common/scripts/basic/Makefile: No such file or directory
    seems like it is looking for a makefile in all subfolders of scripts/basic however there aren't any..

    Best regards,
    Dan

     
  • Dan Mayr

    Dan Mayr - 2010-04-22

    Oh and btw. I waited for almost 5 minutes and the device did (and right now does) not show up in aplay -l or /proc/asound/cards.
    This is what dmesg shows:
    line6usb driver version 0.9.0 (revision 616M)
    line6usb 2-7:1.0: Line6 PODxt found
    line6usb 2-7:1.0: Line6 PODxt now attached
    Line6 device 0: PODxt:0
    Line6 device 1: (not used)
    Line6 device 2: (not used)
    Line6 device 3: (not used)
    Line6 device 4: (not used)
    Line6 device 5: (not used)
    Line6 device 6: (not used)
    Line6 device 7: (not used)
    usbcore: registered new interface driver line6usb

    (why is there no edit function for posts ? )

     
  • Markus Grabner

    Markus Grabner - 2010-04-22

    > I ran into some troubles following your xconfig guide.
    It seems your Linux distro has other conventions for the location of Linux kernel headers and sources :-( I'm using openSUSE-11.2, which distro are you using?

    You should at least have the directory "/lib/modules/$(uname -r)" containing the modules for the running kernel. Can you locate the corresponding sources and objects on your system? For me, these are "/usr/src/linux-2.6.31.12-0.2" and "/usr/src/linux-2.6.31.12-0.2-obj/x86_64/desktop", respectively.

    > This is what dmesg shows:
    Are you sure that this is the "debug" driver (from "https://line6linux.svn.sourceforge.net/svnroot/line6linux/driver/branches/debug")? It should be much more verbose.

    Kind regards,
    Markus

     
  • Dan Mayr

    Dan Mayr - 2010-04-23

    Do you have a skype account? Would make it a lot easier…

    I am using debian squeeze.
    There is a lib/modules/2.6.32-trunk-686/ and link called sources pointing to /usr/src/linux-headers-2.6.32-trunk-common
    The kernel sources a located in /usr/src/linux-source-2.6.32.
    The headers in /usr/src/linux-headers-2.6.32-trunk-686 (arch dependent) and /usr/src/linux-headers-2.6.32-trunk-common (arch independent).
    The kernel *.ko files are located in /lib/modules/2.6.32-trunk-686/kernel.

     
  • Markus Grabner

    Markus Grabner - 2010-04-23

    > Do you have a skype account?
    I sent it to your sourceforge email address, did you get it?

    Kind regards,
    Markus

     

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.