Summary: | ASTERISK-10256: revision 81599 causes crash in gsm_read | ||
Reporter: | Dmitry Andrianov (dimas) | Labels: | |
Date Opened: | 2007-09-07 17:32:27 | Date Closed: | 2011-06-07 14:00:27 |
Priority: | Critical | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ||
Description: | revision 81599 introduces some error which crashes asterisk when VoiceMail is executed. I do not think that error is related to VoiceMail at all because it crashes during playback of .gsm file. Previous revision 81569 works just fine. 111 => { Answer; Voicemail(1011, u); }; ****** ADDITIONAL INFORMATION ****** (gdb) bt #0 0x481a711c in fread () from /lib/libc.so.6 #1 0x00c698b8 in gsm_read (s=0x9a469f8, whennext=0x157ea3c) at format_gsm.c:72 #2 0x0809b250 in ast_readaudio_callback (s=0x9a469f8) at file.c:635 #3 0x0809de1a in ast_streamfile (chan=0x9a4cbc0, filename=0x56af054 "vm-theperson", preflang=0x9a4cfdf "en") at file.c:725 #4 0x0809dfdc in ast_stream_and_wait (chan=0x9a4cbc0, file=0x40 <Address 0x40 out of bounds>, language=0x9a4cfdf "en", digits=0x1589ca0 "#") at file.c:1216 ASTERISK-1 0x056a3f9b in leave_voicemail (chan=0x9a4cbc0, ext=0x1589cf0 "1011", options=0x1589e3c) at app_voicemail.c:2080 ASTERISK-2 0x056a41c9 in vm_exec (chan=0x9a4cbc0, data=0x158df28) at app_voicemail.c:6975 ASTERISK-3 0x080c728e in pbx_extension_helper (c=0x9a4cbc0, con=0x0, context=0x9a4cd40 "ael-default", exten=0x9a4cd90 "111", priority=2, label=0x0, callerid=0x9a45a40 "1001", action=E_SPAWN) at pbx.c:532 ASTERISK-4 0x080c8f76 in __ast_pbx_run (c=0x9a4cbc0) at pbx.c:2288 ASTERISK-5 0x080c9f1e in pbx_thread (data=0x9a4cbc0) at pbx.c:2603 ASTERISK-6 0x080f5c4b in dummy_start (data=0x9a48a38) at utils.c:775 ASTERISK-7 0x482c42db in start_thread () from /lib/libpthread.so.0 ASTERISK-8 0x4821e12e in clone () from /lib/libc.so.6 | ||
Comments: | By: Russell Bryant (russell) 2007-09-07 22:27:44 Try running a "make clean", then rebuild and reinstall and see if that helps. By: Dmitry Andrianov (dimas) 2007-09-08 02:07:46 'make clean' helped, case closed. Why is 'make clean' required at all? doesn't it mean there are some dependencies not properly reflected in makefiles and something is not rebuilt when some .h file (or something) changes? By: Michiel van Baak (mvanbaak) 2007-09-08 04:01:49 make clean will remove all previously compiled objects. I'm not a makefile wizard but I found out most projects need a make clean from time to time. By: Digium Subversion (svnbot) 2007-09-08 08:24:26 Repository: asterisk Revision: 81952 ------------------------------------------------------------------------ r81952 | russell | 2007-09-08 08:24:26 -0500 (Sat, 08 Sep 2007) | 11 lines (closes issue ASTERISK-10256) Bump the cleancount so that a "make clean" will be forced. This is needed because my fix in revision 81599 made a change to a data structure in file.h, and since file dependency tracking is only on with dev-mode enabled, file format modules that don't get rebuilt may crash, as is the case with this issue. This makes me wonder - how much faster does the code build without the file dependency tracking enabled? If it doesn't make much of a difference, then it may be worth just keeping it on all of the time, or perhaps just not in release tarballs, so that this type of issue is avoided. ------------------------------------------------------------------------ By: Russell Bryant (russell) 2007-09-08 08:26:10 Your question on dependencies is correct. In fact, we *only* track file depencies if dev-mode is enabled (./configure --enable-dev-mode). However, this is making me think that the optimization isn't worth it. By: Digium Subversion (svnbot) 2007-09-08 08:27:10 Repository: asterisk Revision: 81953 ------------------------------------------------------------------------ r81953 | russell | 2007-09-08 08:27:09 -0500 (Sat, 08 Sep 2007) | 19 lines Merged revisions 81952 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81952 | russell | 2007-09-08 08:42:26 -0500 (Sat, 08 Sep 2007) | 11 lines (closes issue ASTERISK-10256) Bump the cleancount so that a "make clean" will be forced. This is needed because my fix in revision 81599 made a change to a data structure in file.h, and since file dependency tracking is only on with dev-mode enabled, file format modules that don't get rebuilt may crash, as is the case with this issue. This makes me wonder - how much faster does the code build without the file dependency tracking enabled? If it doesn't make much of a difference, then it may be worth just keeping it on all of the time, or perhaps just not in release tarballs, so that this type of issue is avoided. ........ ------------------------------------------------------------------------ By: Digium Subversion (svnbot) 2007-09-08 09:14:35 Repository: asterisk Revision: 81970 ------------------------------------------------------------------------ r81970 | murf | 2007-09-08 09:14:33 -0500 (Sat, 08 Sep 2007) | 260 lines Merged revisions 81683,81712,81730,81742,81744-81745,81777,81779,81781-81785,81813-81814,81827,81839,81849,81858,81873,81891,81910,81924,81953 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................ r81683 | mmichelson | 2007-09-06 09:21:45 -0600 (Thu, 06 Sep 2007) | 13 lines Merged revisions 81682 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81682 | mmichelson | 2007-09-06 10:20:36 -0500 (Thu, 06 Sep 2007) | 5 lines Fixes a memory leak (closes issue ASTERISK-10247, reported and patched by Ivan) ........ ................ r81712 | rizzo | 2007-09-06 09:43:49 -0600 (Thu, 06 Sep 2007) | 8 lines various changes to the documentation, and redefinition of ao2_hash_fn and ao2_callback_fn typedefs, in preparation to more cleanup of the _search_flags Please do not merge this change to 1.4 yet - there are no functional changes anyways. ................ r81730 | mmichelson | 2007-09-06 10:31:30 -0600 (Thu, 06 Sep 2007) | 14 lines Merged revisions 81713 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81713 | mmichelson | 2007-09-06 11:25:40 -0500 (Thu, 06 Sep 2007) | 6 lines Fixes an issue where valid DTMF had to be pressed twice to exit a queue if a member's phone was ringing. (closes issue ASTERISK-10245, reported by strider2k, patched by me) ........ ................ r81742 | mattf | 2007-09-06 10:38:54 -0600 (Thu, 06 Sep 2007) | 1 line Patch on 10575. Add support for unequipped CIC (UCIC) message as well as improve some of our CIC flags in chan_zap ................ r81744 | tilghman | 2007-09-06 10:57:20 -0600 (Thu, 06 Sep 2007) | 2 lines Incorporate the ability to log output of safe_asterisk to syslog (closes issue ASTERISK-9588) ................ r81745 | phsultan | 2007-09-06 11:00:58 -0600 (Thu, 06 Sep 2007) | 9 lines Merged revisions 81743 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81743 | phsultan | 2007-09-06 18:56:29 +0200 (Thu, 06 Sep 2007) | 1 line Various string length fixes. Removed an unused variable in aji_client structure (context) ........ ................ r81777 | file | 2007-09-06 13:43:20 -0600 (Thu, 06 Sep 2007) | 15 lines Merged revisions 81776 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81776 | file | 2007-09-06 16:40:37 -0300 (Thu, 06 Sep 2007) | 7 lines (closes issue ASTERISK-9809) Reported by: stevefeinstein Patches: meetme-unmute-manager.diff uploaded by qwell (license 4) Tested by: stevefeinstein After looking over the code I agree with Qwell. Setting the file descriptor to conference each time just causes a fight back and forth. ........ ................ r81779 | qwell | 2007-09-06 14:00:08 -0600 (Thu, 06 Sep 2007) | 10 lines Merged revisions 81778 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81778 | qwell | 2007-09-06 14:59:07 -0500 (Thu, 06 Sep 2007) | 2 lines This should fix a build issue that people building against uClibc were seeing with the addition of astobj2 ........ ................ r81781 | russell | 2007-09-06 14:05:50 -0600 (Thu, 06 Sep 2007) | 2 lines Fix the syntax of declaring a hint with a name to be compatible with trunk ................ r81782 | file | 2007-09-06 14:16:02 -0600 (Thu, 06 Sep 2007) | 6 lines (closes issue ASTERISK-10019) Reported by: mvanbaak Patches: chan_skinny_info.diff uploaded by mvanbaak (license 7) Add skinny show device, skinny show line, and skinny show settings CLI commands. ................ r81783 | russell | 2007-09-06 14:24:18 -0600 (Thu, 06 Sep 2007) | 4 lines Merge HINT() dialplan function from my sandbox branch into trunk. This function will let you retrieve the list of devices or name associated with a hint. (inspired by issue ASTERISK-10226) ................ r81784 | russell | 2007-09-06 14:27:26 -0600 (Thu, 06 Sep 2007) | 4 lines Rename the DEVSTATE() function to DEVICE_STATE() to better conform to how other functions are named. (inspired by issue ASTERISK-10226) ................ r81785 | russell | 2007-09-06 14:27:53 -0600 (Thu, 06 Sep 2007) | 2 lines s/DEVSTATE/DEVICE_STATE/ ................ r81813 | russell | 2007-09-06 14:54:07 -0600 (Thu, 06 Sep 2007) | 5 lines Add EXTENSION_STATE() function that can retrieve the state of an extension that has a hint. (closes issue ASTERISK-10226, adamgundy) ................ r81814 | file | 2007-09-06 15:01:10 -0600 (Thu, 06 Sep 2007) | 2 lines Initialize iax_frames variable to NULL, keeps valgrind happy. ................ r81827 | qwell | 2007-09-06 16:06:17 -0600 (Thu, 06 Sep 2007) | 9 lines Merged revisions 81826 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81826 | qwell | 2007-09-06 17:05:02 -0500 (Thu, 06 Sep 2007) | 1 line We added COPTS for ASTCFLAGS additions, but not LDOPTS for ASTLDFLAGS. This adds LDOPTS ........ ................ r81839 | russell | 2007-09-06 16:30:15 -0600 (Thu, 06 Sep 2007) | 24 lines Merged revisions 81832 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81832 | russell | 2007-09-06 17:28:57 -0500 (Thu, 06 Sep 2007) | 16 lines (closes issue ASTERISK-9438, closes issue ASTERISK-10017) Reported by: kenw Patches: 9724.txt uploaded by russell (license 2) Tested by: kenw, russell Resolve a deadlock that occurs when doing a SIP transfer to parking. I come across this type of deadlock fairly often it seems. It is very important to mind the boundary between the channel driver and the core in respect to the channel lock and the channel-pvt lock. Channel drivers lock to lock the pvt and then the channel once it calls into the core, while the core will do it in the opposite order. The way this is avoided is by having channel drivers either release their pvt lock while calling into the core, or such as in this case, unlocking the pvt just long enough to acquire the channel lock. ........ ................ r81849 | russell | 2007-09-06 16:32:03 -0600 (Thu, 06 Sep 2007) | 1 line fix the build ... oops ................ r81858 | file | 2007-09-06 16:34:44 -0600 (Thu, 06 Sep 2007) | 2 lines Fix memory issue that crept up with Russell's testing. It is *not* proper to free the frame we get in ast_write. ................ r81873 | file | 2007-09-07 06:32:04 -0600 (Fri, 07 Sep 2007) | 2 lines Don't check for epoll support when cross compiling. ................ r81891 | mmichelson | 2007-09-07 09:29:23 -0600 (Fri, 07 Sep 2007) | 11 lines Merged revisions 81886 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81886 | mmichelson | 2007-09-07 10:25:19 -0500 (Fri, 07 Sep 2007) | 3 lines Moving the explanation for joinempty to a more appropriate place ........ ................ r81910 | qwell | 2007-09-07 10:13:57 -0600 (Fri, 07 Sep 2007) | 4 lines Add an optional reason parameter to PauseQueueMember/UnpauseQueueMember applications and manager events. Issue 8738, patch by rgollent ................ r81924 | qwell | 2007-09-07 13:53:30 -0600 (Fri, 07 Sep 2007) | 14 lines Merged revisions 81923 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 (closes issue ASTERISK-10255) ........ r81923 | qwell | 2007-09-07 14:48:00 -0500 (Fri, 07 Sep 2007) | 5 lines Allow the MEMBERINTERFACE variable to be used as the mixmonitor filename. This moves the setting of the MEMBERINTERFACE variable to before mixmonitor. Issue 10671, patch by sim. ........ ................ r81953 | russell | 2007-09-08 07:45:40 -0600 (Sat, 08 Sep 2007) | 19 lines Merged revisions 81952 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.4 ........ r81952 | russell | 2007-09-08 08:42:26 -0500 (Sat, 08 Sep 2007) | 11 lines (closes issue ASTERISK-10256) Bump the cleancount so that a "make clean" will be forced. This is needed because my fix in revision 81599 made a change to a data structure in file.h, and since file dependency tracking is only on with dev-mode enabled, file format modules that don't get rebuilt may crash, as is the case with this issue. This makes me wonder - how much faster does the code build without the file dependency tracking enabled? If it doesn't make much of a difference, then it may be worth just keeping it on all of the time, or perhaps just not in release tarballs, so that this type of issue is avoided. ........ ................ ------------------------------------------------------------------------ |