Summary:	DAHLIN-00227: Kernel Panic when Hanging up call using FXS connected phone
Reporter:	Alec Davis (alecdavis)	Labels:
Date Opened:	2010-12-12 00:10:47.000-0600	Date Closed:	2011-01-01 16:07:11.000-0600
Priority:	Critical	Regression?	No
Status:	Closed/Complete	Components:	dahdi (the module)
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) 13122010001.jpg ( 1) chan_dahdi.conf ( 2) system.conf
Description:	Since dahdi-linux -r9524 panic occurs, and warning below. Since dahdi-linux -r9521, the warning below each time we make a call, and echo is there, very short delay, but is very noticeable. "[2010-12-12 19:02:15.600188] WARNING[19536]: chan_dahdi.c:4791 dahdi_train_ec: Unable to request echo training on channel 36: Invalid                argument ****** ADDITIONAL INFORMATION ****** Versions: Asterisk SVN-trunk-r298055M DAHDI Version: SVN-trunk-r9521 Echo Canceller: HWEC libpri version: SVN-branch-1.4-r2113 hardware: Wildcard TE122 Card 0 Wildcard TDM800P Board 1 + VPMADT032
Comments:	By: Alec Davis (alecdavis) 2010-12-12 00:12:51.000-0600 initially saw fault with dahdi -r9524, when using FXS port to call over an IAX trunk. Thus FXO port wasn't part of the equation. By: Michael Spiceland (mspiceland) 2010-12-12 22:02:22.000-0600 Thank you for the report.  Since you have Digium cards, you should call Digium technical support for a rapid response to this issue. Do you have a screenshot or picture of the kernel panic?  Do you expect that the kernel panic is as simple to reproduce as simply hanging up the call from the FXS port?  I would have assumed we would have caught that with automated testing.  Unless it is very easy to reproduce, we'll likely need the config files. By: Michael Spiceland (mspiceland) 2010-12-12 22:12:14.000-0600 alecdavis, The context of this issue just sank in.  Since you are testing trunk and not a release, I assume you aren't likely to need to call technical support but simply assisting us with this report.  We'll see if we can reproduce this.  If it isn't easy to reproduce, configs would surely help. Thanks! By: Alec Davis (alecdavis) 2010-12-13 03:39:40.000-0600 I guess it's possible there is a matching Asterisk team branch for these late versions of dahdi, needless to say surely it shouldn't panic though. I'll get a camera out next time, if unable to reproduce. attached  system.conf  chan_dahdi.conf dialplan: [globals] TRUNK=DAHDI/G0 .... exten => _1XX.,1,Dial(${TRUNK}/${EXTEN:1}) exten => _85XX.,1,Dial(IAX2/iaxtrunk/${EXTEN}) By: Alec Davis (alecdavis) 2010-12-13 04:11:20.000-0600 console output: Core debug is at least 3    -- Starting simple switch on 'DAHDI/33-1'    -- Executing [1083210@phones:1] Dial("DAHDI/33-1", "DAHDI/G0/083210") in new stack    -- Called G0/083210 [2010-12-13 23:01:44.897036] WARNING[8052]: chan_dahdi.c:4791 dahdi_train_ec: Unable to request echo training on channel 36: Invalid argument    -- DAHDI/36-1 answered DAHDI/33-1 asterix*CLI> Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] Process asterisk (pid: 8052, ti=f7ca4000 task=f68471c0 task.ti=f7ca4000) Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] Stack: f6cb7a80 00000000 f7ca5e94 f8ae0ef4 00000000 00000000 00200246 00000000 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]        f6f34dc0 00000000 f6c85790 c0176ed7 f6cb7a80 000000ff f6cb7a80 f6c85790 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]        00000000 c0176dbb c0173233 f741d200 f71c4be8 f7ca5f14 f6cb7a80 b6e4eed0 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] Call Trace: Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<f8ae0ef4>] dahdi_open+0x1bd/0x1da [dahdi] Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0176ed7>] chrdev_open+0x11c/0x159 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0176dbb>] chrdev_open+0x0/0x159 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0173233>] __dentry_open+0x10d/0x1fc Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c017333e>] nameidata_to_filp+0x1c/0x2c Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c017dab2>] do_filp_open+0x34f/0x684 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c013baca>] do_futex+0x303/0x70e Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<f8ae8079>] dahdi_ioctl+0x23e/0x245 [dahdi] Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0173050>] do_sys_open+0x40/0xb0 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0173104>] sys_open+0x1e/0x23 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  [<c0103857>] sysenter_past_esp+0x78/0xb1 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682]  ======================= Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] Code: 89 7c 24 04 eb 52 c7 44 24 04 fa ff ff ff eb 48 c7 44 24 04 f0 ff ff ff eb 3e 89 d8 e8 b1 92 7d c7 89 c5 8b 83 b4 00 00 00 85 c0 <8b> b0 60 01 00 00 0f 84 67 ff ff ff e9 69 ff ff ff 8b 14 24 89 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] EIP: [<f8ae0cfe>] dahdi_specchan_open+0x4fd/0x536 [dahdi] SS:ESP 0068:f7ca5e78 Message from syslogd@asterix at Mon Dec 13 23:01:58 2010 ... asterix kernel: [  238.838682] Oops: 0000 [#1] SMP By: Alec Davis (alecdavis) 2010-12-13 04:17:09.000-0600 hope you can make something of the picture, lights out, and I moved :( By: Digium Subversion (svnbot) 2010-12-13 08:57:14.000-0600 Repository: dahdi Revision: 9530 U   linux/trunk/drivers/dahdi/dahdi-base.c ------------------------------------------------------------------------ r9530 | sruffell | 2010-12-13 08:57:13 -0600 (Mon, 13 Dec 2010) | 11 lines dahdi: Do not dereference chan->span for pseudo channels. Fixes a regression introduced in r9510 which saves a pointer to the ops member of a channel's span before checking if the channel is a pseudo. Psuedo channels do not have spans. (issue DAHLIN-227) (issue DAHLIN-226) Reported by: alecdavis Signed-off-by: Shaun Ruffell <sruffell@digium.com> ------------------------------------------------------------------------ http://svn.digium.com/view/dahdi?view=rev&revision=9530 By: Shaun Ruffell (sruffell) 2010-12-13 08:59:50.000-0600 alecdavis:  Thanks...the text from the oops definitely helped.  The oops was actually due to an earlier commit that I had made last week.  Now...there is still the issue of the ECHOTRAIN to look at.... By: Shaun Ruffell (sruffell) 2010-12-13 15:22:41.000-0600 alecdavis:  When you next get a chance to try the new version of the driver, could you edit your /etc/dahdi/system.conf and add echocanceller=hwec,1-15,17-31 As of last week users will be able to switch between hardware and software echocan without reloading the drivers.  We still need to update dahdi_genconf to default to hwec if possible. By: Alec Davis (alecdavis) 2010-12-15 00:10:45.000-0600 unable to panic as previously, this problem seems to be resolved. also added as suggested in system.conf<pre> span=1,0,0,ccs,hdb3,crc4 #;esf,b8zs bchan=1-15,17-31 # set this to 1-15,17-31 for E1 dchan=16 # set this to 16 for E1 fxols=32-35 fxsls=36-37 #echocanceller=hwec,1-15,17-31 echocanceller=hwec,32-37 </pre> By: Shaun Ruffell (sruffell) 2011-01-01 16:07:11.000-0600 Seems like there isn't anything else to do here.  If I'm mistaken, please reopen.