< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index Next in Thread >

Re: [reSIProcate] Ares crash


Hi.
I'm not expert in ares - but this calls seems outdated a bit. Maybe in old versions it was useful...
I add 
  if (!read_fds)
    return;
on read_tcp_data() begin. Will see how it changes processing (but i develop client applications).



On Thu, Dec 27, 2012 at 6:47 PM, palladin <p-aladin@xxxxxxxxx> wrote:
Hi, developer's team,
I have met crash when resip runs dns query.
Stack of crash looks like this

#0  0x00000031e6805d81 in read_tcp_data (channel=0x128feb0, server_idx=0, read_fds=<value optimized out>, now=1356626029) at ares_process.c:274
#1  0x00000031e6806147 in ares_process_poll (channel=0x128feb0, server_idx=0, rdFd=40, wrFd=<value optimized out>, now=1356626029) at ares_process.c:102
#2  0x00000031e584c344 in resip::AresDnsPollItem::processPollEvent(unsigned short) () from /usr/lib64/librutil-1.8.so
#3  0x00000031e584252d in resip::FdPollGrp::processItem(resip::FdPollItemIf*, unsigned short) () from /usr/lib64/librutil-1.8.so
#4  0x00000031e5843159 in resip::FdPollImplEpoll::epollWait(int) () from /usr/lib64/librutil-1.8.so
#5  0x00000031e58434da in resip::FdPollImplEpoll::waitAndProcess(int) () from /usr/lib64/librutil-1.8.so

>From provided core dump I could understand 2 things:
1) there is no verification of fd_set on NULL, when calling FD_ISSET.
2) we get here if there is no answer from dns server in 25 ms. Am I wrong?

Please advise with this issue. It appears occasionally and I can't understand scenario, which causes crash.

Best regards,
Taras.
_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@resiprocate.org
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel