| Summary: | DAHLIN-00063: kernel oopses with similar stack traces | ||
| Reporter: | Alexander 'Leo' Bergolth (bergolth) | Labels: | |
| Date Opened: | 2008-11-25 03:48:52.000-0600 | Date Closed: | 2008-12-29 07:40:33.000-0600 |
| Priority: | Critical | Regression? | No |
| Status: | Closed/Complete | Components: | wct4xxp |
| Versions: | 2.0.0 | Frequency of Occurrence | |
| Related Issues: | |||
| Environment: | Attachments: | ( 0) asterisk-crash-dahdi.txt ( 1) asterisk-crash-dmesg-excerpt.txt ( 2) asterisk-crash-zaptel.txt ( 3) asterisk-crash-zaptel2.txt | |
| Description: | My asterisk-box running asterisk-1.4.22 and fedora kernel 2.6.26.5-28.fc8PAE and 2.6.26.6-49.fc8PAE constantly crashes after several days, leaving the same stack-trace every time. (Attached.) Zaptel versions 1.4.11 and 1.4.12 also show the same problem. The crashes appear after a few days, I couldn't trigger them with phone calls. kernel-PAE-2.6.24.4-64.fc8 together with a zaptel SVN version from 2008-08-05 worked fine. (An earlier version had stack overflow problems with this kernel.) Unfortunately the stack traces seem to be subsequent errors following a memory corruption so I suppose the stack traces won't be very informative. ****** ADDITIONAL INFORMATION ****** Machine: Dell DL380G5, 8 GB RAM Wildcard TE420 (4th Gen) | ||
| Comments: | By: Leif Madsen (lmadsen) 2008-11-25 07:34:03.000-0600 I have a feeling that you may need to provide some valgrind output if it is indeed memory corruption. I've assigned this issue to sruffell for now, but please reassign if this is not correct. Thanks! By: Shaun Ruffell (sruffell) 2008-11-25 10:17:11.000-0600 Couple of questions. Are you seeing any errors from vsnprintf in your dmesg logs before the crashes occur? Could you try the dahdi-linux trunk? Specifically, I'm interested in the behavior when http://svn.digium.com/view/dahdi?view=rev&revision=5355 is in play. That commit isn't in any release candidates yet but it will be in 2.1.0-rc5 as soon as it's released. By: Alexander 'Leo' Bergolth (bergolth) 2008-12-29 07:12:51.000-0600 The snprintf patch seems to fix the problem. Using 2.1.0-rc5, the box runs without a crash for 20 days now. The only messages, I get are: --------------------------- Unassigning channel 0/1! Unassigning channel 0/1! Unassigning channel 0/19! Unassigning channel 0/19! --------------------------- I'm using "echo 1 > /sys/module/dahdi/parameters/debug" and "echo 1 > /sys/module/wct4xxp/parameters/debug". Thanks, --leo By: Leif Madsen (lmadsen) 2008-12-29 07:40:33.000-0600 Closed as reporter indicates this is fixed in a future version. | ||