OpenWrt Forum Archive

Topic: Freeswitch Aborted

The content of this topic has been archived on 30 Apr 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

I have managed to upgrade (to the latest SVN trunk), compile, and install freeswitch + its packages on an external USB partition of my WGT634U. Every time when I try to run /etc/init.d/freeswitch start (once I have executed ulimit -s 240), freeswitch returns the Aborted message. So, I installed strace package and captured its output to a file. The file is about 27KB in size and I don't suppose it will be prudent to post it here unless I can attach it. However, after studying the output file from strace, I noticed the following two ioctl lines (towards the end of the file) before freeswitched terminated its process:

ioctl(0, TIOCNXCL, {B38400 opost isig icanon echo ...}) = 0
ioctl(1, TIOCNXCL, 0x7fa522c8)          = -1 ENOTTY (Inappropriate ioctl for device)
getpid()                                = 4768
getrlimit(RLIMIT_STACK, {rlim_cur=240*1024, rlim_max=240*1024}) = 0
rt_sigaction(SIGRT_0, {SIG_DFL, [], SA_STACK|SA_INTERRUPT|SA_SIGINFO|0x2c669f0}, NULL, 16)
rt_sigaction(SIGRT_1, {SIG_DFL, [], SA_STACK|SA_INTERRUPT|SA_SIGINFO|0x2c66910}, NULL, 16)
rt_sigaction(SIGRT_2, {SIG_DFL, [], SA_STACK|SA_INTERRUPT|0x2c65b24}, NULL, 16) = 0
rt_sigprocmask(SIG_BLOCK, [RT_0], NULL, 16) = 0
rt_sigprocmask(SIG_UNBLOCK, [RT_1], NULL, 16) = 0
brk(0)                                  = 0x416000
brk(0x417000)                           = 0x417000
brk(0x419000)                           = 0x419000
brk(0x41b000)                           = 0x41b000
rt_sigaction(SIGILL, {0x10000000, [], 0x401190 /* SA_??? */}, {SIG_DFL, [], 0}, 16) = 0
rt_sigaction(SIGTERM, {0x10000000, [], 0x401190 /* SA_??? */}, {SIG_DFL, [], 0}, 16) = 0
getrlimit(RLIMIT_STACK, {rlim_cur=240*1024, rlim_max=240*1024}) = 0
setrlimit(RLIMIT_STACK, {rlim_cur=240*1024, rlim_max=8192*1024}) = 0
setrlimit(RLIMIT_NOFILE, {rlim_cur=999999, rlim_max=999999}) = 0
setrlimit(RLIMIT_CPU, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
setrlimit(RLIMIT_DATA, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
setrlimit(RLIMIT_FSIZE, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
setrlimit(RLIMIT_NPROC, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
setrlimit(RLIMIT_RTPRIO, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
setrlimit(RLIMIT_AS, {rlim_cur=RLIM_INFINITY, rlim_max=RLIM_INFINITY}) = 0
brk(0x41c000)                           = 0x41c000
getpid()                                = 4768
brk(0x41e000)                           = 0x41e000
open("/tmp/freeswitch/log//freeswitch.pid", O_RDONLY) = 3
read(3, "4743"..., 32)                  = 4
close(3)                                = 0
open("/tmp/freeswitch/log//freeswitch.pid", O_WRONLY|O_CREAT|O_TRUNC, 0600) = 3
fcntl(3, F_SETLK, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0}) = 0
write(3, "4768"..., 4)                  = 4
brk(0x420000)                           = 0x420000
rt_sigprocmask(SIG_UNBLOCK, [IOT], NULL, 16) = 0
getpid()                                = 4768
kill(4768, SIGIOT)                      = 0
--- SIGIOT (Aborted) @ 0 (0) ---
+++ killed by SIGIOT +++
- /tmp/freeswitch.strace

From the above excerpt, the only thing I noticed is freeswitch had a problem with IOCTL. Being a non-programmer, I am hoping anyone out here can help me out to debug this problem.

To clarify the above post, I compiled OpenWRT SVN trunk + freeswitch with the following options:

Binutils Version (binutils 2.20)
GCC compiler Version (gcc 4.3.3 with CodeSourcery enhancements)
uClibc Version (uClibc 0.9.30.1)

Hi,

I have the lastest OpenWRT trunk compiled and running on a Router Station Pro.
Freeswitch is running correctly.
I do not find the LUA freeswitch module. Is missing in the Freeswitch OpenWRT packages ?

Regards
Satjuice

Thanks for your response. Two days ago, I just noticed the OpenWRT Makefile for freeswitch got updated. So, I did a clean compilation, but got stuck with compiling gd-2.0.35 package as shown here. I sure will give this a try once the problem has been fixed.

The discussion might have continued from here.