gdb dying while debugging asterisk. The problem seems to be related
with a race in the handling of module_list, which in turn is triggeded
by calling dlopen() on a system which uses initializers to create
locks.
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@95772 f38db490-d61c-443f-a65b-d21fe96a405b
This is only to complete the build, clearly the linker
behaviour will be completely different and likely to
cause trouble in those cases.
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@89367 f38db490-d61c-443f-a65b-d21fe96a405b
build times - tested, there is no measureable difference before and
after this commit.
In this change:
use asterisk/compat.h to include a small set of system headers:
inttypes.h, unistd.h, stddef.h, stddint.h, sys/types.h, stdarg.h,
stdlib.h, alloca.h, stdio.h
Where available, the inclusion is conditional on HAVE_FOO_H as determined
by autoconf.
Normally, source files should not include any of the above system headers,
and instead use either "asterisk.h" or "asterisk/compat.h" which does it
better.
For the time being I have left alone second-level directories
(main/db1-ast, etc.).
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@89333 f38db490-d61c-443f-a65b-d21fe96a405b
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r89325 | kpfleming | 2007-11-16 10:47:46 -0600 (Fri, 16 Nov 2007) | 4 lines
To help combat problems where people build external modules (asterisk-addons or others) and then change the build options of the Asterisk build in a way that makes the incompatible without warning, this commit introduces an MD5 signature of the important build-time options and includes that signature into modules when they are built. When the loader loads one of these modules and notices the problem, it will emit a warning to console and refuse to initialize the module, as doing so could cause the system to be unstable or even crash.
If you upgrade to this version of Asterisk, you must rebuild *all* of your modules that came from other sources before trying to run this version. If you are using Digium's G.729 binary codec module, you will need v33 or newer.
........
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@89326 f38db490-d61c-443f-a65b-d21fe96a405b
- the *_CURRENT macros no longer need the list head pointer argument
- add AST_LIST_MOVE_CURRENT to encapsulate the remove/add operation when moving entries between lists
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@89106 f38db490-d61c-443f-a65b-d21fe96a405b
the constructor for the list of modules was run
after the constructors for the embedded modules
(which appended entries to the list).
As a result, the list appeared empty when it was
time to use it.
On linux the order of execution of constructor
was evidently different (it may depend on the
ordering of modules in the ELF file).
This is only a workaround - there may be other
situations where the execution of constructors
causes problems, so if we manage to find a more
general solution this workaround can go away.
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@89031 f38db490-d61c-443f-a65b-d21fe96a405b
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r67308 | russell | 2007-06-05 10:51:53 -0500 (Tue, 05 Jun 2007) | 5 lines
When shutting down "gracefully", go through and run the unload() callbacks for
all of the modules. "stop now" is considered a non-graceful shutdown and will
not go through this process.
(issue #9804, reported by chrisost, patch by me)
........
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@67310 f38db490-d61c-443f-a65b-d21fe96a405b
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r62986 | kpfleming | 2007-05-03 11:38:56 -0500 (Thu, 03 May 2007) | 2 lines
improve loader a bit, by avoiding trying to initialize embedded modules twice and avoiding trying to load modules from disk when they have been loaded already during the 'preload' pass (reported by blitzrage on IRC, patch by me)
........
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@62988 f38db490-d61c-443f-a65b-d21fe96a405b
Might need better locking support, giving up if we can't get the lock. Right now,
using existing locking in find_resource
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@46582 f38db490-d61c-443f-a65b-d21fe96a405b
various PBX installations and checks if a module is loaded before using
it.
example IFMODULE(chan_sip3.so)
issue #6671 in the bug tracker, finally gone. Thanks to mithraen for keeping
it updated.
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@46513 f38db490-d61c-443f-a65b-d21fe96a405b
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r45817 | file | 2006-10-21 14:48:58 -0400 (Sat, 21 Oct 2006) | 2 lines
Don't use promotion on Darwin because it doesn't seem to work quite right in all cases, this should solve the unresolved symbol issue people have been seeing.
........
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@45820 f38db490-d61c-443f-a65b-d21fe96a405b
also eliminate a warning from dlopen() because some flags were missing on the second dlopen() call
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@41087 f38db490-d61c-443f-a65b-d21fe96a405b
also, keep trying to dlclose() a module until it actually goes away, since it may have other modules it brought in when it was loaded (thanks PCadach for pointing this problem out to me)
git-svn-id: http://svn.digium.com/svn/asterisk/trunk@40949 f38db490-d61c-443f-a65b-d21fe96a405b
qwell