FS#18014 - [oss] crashes repeatedly
Attached to Project:
Community Packages
Opened by Rene (hit) - Tuesday, 26 January 2010, 13:45 GMT
Last edited by Dan Griffiths (Ghost1227) - Monday, 01 March 2010, 20:35 GMT
Opened by Rene (hit) - Tuesday, 26 January 2010, 13:45 GMT
Last edited by Dan Griffiths (Ghost1227) - Monday, 01 March 2010, 20:35 GMT
|
Details
Description:
OSS crashes repeatedly for some reason. Additional info: * dmesg irq 17: nobody cared (try booting with the "irqpoll" option) Pid: 0, comm: swapper Not tainted 2.6.32-ARCH #1 Call Trace: [<c108a134>] ? __report_bad_irq+0x24/0x90 [<c108a2f0>] ? note_interrupt+0x150/0x190 [<c108aa4b>] ? handle_fasteoi_irq+0xab/0xd0 [<c10064e5>] ? handle_irq+0x15/0x30 [<c1005a57>] ? do_IRQ+0x47/0xc0 [<c1047002>] ? irq_exit+0x52/0x70 [<c1005a60>] ? do_IRQ+0x50/0xc0 [<c1035463>] ? select_nohz_load_balancer+0x53/0x140 [<c10040d0>] ? common_interrupt+0x30/0x38 [<c100b00d>] ? mwait_idle+0x4d/0xc0 [<c1002904>] ? cpu_idle+0x94/0xe0 [<c13e285d>] ? start_kernel+0x335/0x33a [<c13e2366>] ? unknown_bootoption+0x0/0x190 handlers: [<f5f7eb00>] (osscore_intr+0x0/0x60 [osscore]) Disabling IRQ #17 osscore: Output timed out (sync) on audio engine 1 198831 fifo errors were detected Restarted OSS and another one kicks in dmesg: osscore: Output timed out on audio engine 1/'Intel ICH5 (24D5) (vmix)' (count=42274816) irq 17: nobody cared (try booting with the "irqpoll" option) Pid: 13541, comm: deluged Not tainted 2.6.32-ARCH #1 Call Trace: [<c108a134>] ? __report_bad_irq+0x24/0x90 [<f63516a8>] ? ichintr+0x2e8/0x8f0 [oss_ich] [<c108a2f0>] ? note_interrupt+0x150/0x190 [<c108aa4b>] ? handle_fasteoi_irq+0xab/0xd0 [<c10064e5>] ? handle_irq+0x15/0x30 [<c1005a57>] ? do_IRQ+0x47/0xc0 [<f6295b28>] ? osscore_intr+0x28/0x60 [osscore] [<c10040d0>] ? common_interrupt+0x30/0x38 [<c1221d60>] ? __copy_skb_header+0x0/0x170 [<c1221eef>] ? __skb_clone+0x1f/0xd0 [<c1267869>] ? tcp_transmit_skb+0x69/0x6d0 [<c10040d0>] ? common_interrupt+0x30/0x38 [<c1269d40>] ? tcp_write_xmit+0x190/0x980 [<c126a691>] ? tcp_send_fin+0x91/0x180 [<c126a59e>] ? __tcp_push_pending_frames+0x2e/0x90 [<c125bcff>] ? tcp_close+0x32f/0x410 [<c127b2e3>] ? inet_release+0x33/0x60 [<c121d2aa>] ? sock_release+0x1a/0x80 [<c121d31f>] ? sock_close+0xf/0x30 [<c10e4e67>] ? __fput+0xc7/0x1d0 [<c10e1847>] ? filp_close+0x47/0x80 [<c10e18f0>] ? sys_close+0x70/0xc0 [<c10039f3>] ? sysenter_do_call+0x12/0x28 handlers: [<f6295b00>] (osscore_intr+0x0/0x60 [osscore]) Disabling IRQ #17 157819 fifo errors were detected After these errors sound start stuttering and again, restarting OSS fixes it for a couple of minutes (or less). * package version(s) firefox 3.6-2 jre 6u17-1 oss 4.2_2002-2 kernel26 2.6.32.6-1 Steps to reproduce: 1. Play sounds with OSS Yes, [testing] is enabled, but only because this also happened before, so I enabled testing to see if it's fixed. Booting with "irqpoll" didn't help. Is there any other information I can provide to help investigation? |
This task depends upon
Closed by Dan Griffiths (Ghost1227)
Monday, 01 March 2010, 20:35 GMT
Reason for closing: Works for me
Additional comments about closing: OP requested close.
Monday, 01 March 2010, 20:35 GMT
Reason for closing: Works for me
Additional comments about closing: OP requested close.
Also, which was the last kernel version where everything used to work correctly?
Rebuilt OSS now, along with alsa-oss and alsa-plugins, and problem seems to be only with Java apps in browser now (games and so). I can hear the sounds, but the browser hangs heavily and system mem usage is massive. Also dmesg is filled with repeating messages which I attached now.
if (n + 1 >= MAX_CLIENTS) /* Cannot create more client engines */
return OSS_EBUSY;
As 16 is EBUSY, and that's the only place where it returns OSS_EBUSY that could result in such message.
Because of this, the only way I can see this message would show up is that more than MAX_CLIENTS (in our build, 9) instances of /dev/dsp are being open at the same time. If this is the case, you should see a lot processes in the per-application mixer controls at ossxmix when this errors occurs.
In this case it would be a good measure to see which applications are that and looking if some of them are opening /dev/dsp crazily, causing this.
About the "page allocation failure", I can't see why it happens, nor if it could be a cause for the issues that happens after that (including the vmix engine creation failure). I tried looking for cesium at IRC but couldn't find him. You may want to post this bug at the OSS forum (http://www.opensound.com/forum/) to see if he or someone else experienced on OSS can help debugging that.
I have no other ideas by now :(
However... I can't replicate this issue (and I have been using OSS for a long time now...)
[hit@sahver ~]$ dmesg
...
irq 17: nobody cared (try booting with the "irqpoll" option)
Pid: 0, comm: swapper Not tainted 2.6.32-ARCH #1
Call Trace:
[<c108a2b4>] ? __report_bad_irq+0x24/0x90
[<c108a470>] ? note_interrupt+0x150/0x190
[<c108abcb>] ? handle_fasteoi_irq+0xab/0xd0
[<c10064e5>] ? handle_irq+0x15/0x30
[<c1005a57>] ? do_IRQ+0x47/0xc0
[<c10354f3>] ? select_nohz_load_balancer+0x53/0x140
[<c10040d0>] ? common_interrupt+0x30/0x38
[<c100b01d>] ? mwait_idle+0x4d/0xc0
[<c1002904>] ? cpu_idle+0x94/0xe0
[<c13e485d>] ? start_kernel+0x335/0x33a
[<c13e4366>] ? unknown_bootoption+0x0/0x190
handlers:
[<f1382b00>] (osscore_intr+0x0/0x60 [osscore])
Disabling IRQ #17
osscore: Output timed out on audio engine 1/'Intel ICH5 (24D5) (vmix)' (count=7659520)
109278 fifo errors were detected
[hit@sahver ~]$ sudo /etc/rc.d/oss restart
:: Stopping Open Sound System [DONE]
:: Starting Open Sound System [DONE]
[hit@sahver ~]$
After restart everything is OK again. It just drives me crazy and makes me want to make a big fire and thorw everything into.
So I guess we can close it...
[1] udisks-git, gvfs-git, etc
[2] IRQ addresses set to manual instead of automatic