Re: libssh crash

[Aris Adamantiadis <aris@xxxxxxxxxxxx>]

Hi Vasuki,

This looks like an heap corruption :( The core in buffers was recently
inspected, we founds some bugs but they were unlikely to be triggable
like this. It is well possible that the crash in realloc is caused by an
heap corruption somewhere else in the program (including or not libssh).

Could you try to reproduce this problem within valgrind ? A valgrind
report would be of big help.

Aris
Le 7/02/13 19:22, Vasuki Gubbi a écrit :
> Hello,
> 
> I am using libssh 0.5.2. Sometimes I am seing a crash in my program.
> Coredump shows following stack:
> 
> #0  0x0000003ab2e32a45 in raise () from /lib64/libc.so.6
> #1  0x0000003ab2e34225 in abort () from /lib64/libc.so.6
> #2  0x0000003ab2e6fdfb in __libc_message () from /lib64/libc.so.6
> #3  0x0000003ab2e75716 in malloc_printerr () from /lib64/libc.so.6
> #4  0x0000003ab2e7abf7 in _int_realloc () from /lib64/libc.so.6
> #5  0x0000003ab2e7ad55 in realloc () from /lib64/libc.so.6
> #6  0x00007f050ad6026a in realloc_buffer () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> #7  0x00007f050ad603d7 in buffer_add_data () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> #8  0x00007f050ad81e3c in ssh_socket_pollcallback () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> #9  0x00007f050ad7e826 in ssh_poll_ctx_dopoll () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> #10 0x00007f050ad7f6a0 in ssh_handle_packets () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> #11 0x00007f050ad66941 in ssh_channel_poll () from
> /netsmart/netsmart8.0/lib/libssh.so.4
> 
> Is there any known bug in libssh related to ssh_channel_poll?
> 
> Regards,
> Vasuki