On 17/05/11 09:15, Harald Welte wrote:
Hi Pablo!
I am sometimes getting osmo-nitb segfaults, and sometimes random bytes as output of log messages.
I've quickly started it in valgrind, and even during connection of the BTS I get the following issues:
<000d> input/ipaccess.c:693 accept()ed new OML link from 192.168.200.117 ==21897== Conditional jump or move depends on uninitialised value(s) ==21897== at 0x60D4BD4: vfprintf (vfprintf.c:1620) ==21897== by 0x60F8F01: vsnprintf (vsnprintf.c:120) ==21897== by 0x4E33670: _output (logging.c:177) ==21897== by 0x4E33878: _logp (logging.c:227) ==21897== by 0x4E33BCB: logp2 (logging.c:246) ==21897== by 0x435E5C: listen_fd_cb (ipaccess.c:692) ==21897== by 0x4E32419: osmo_select_main (select.c:132) ==21897== by 0x40555B: main (bsc_hack.c:306) ==21897== ==21897== Use of uninitialised value of size 8 ==21897== at 0x60D5008: vfprintf (vfprintf.c:1620) ==21897== by 0x60F8F01: vsnprintf (vsnprintf.c:120) ==21897== by 0x4E33670: _output (logging.c:177) ==21897== by 0x4E33878: _logp (logging.c:227) ==21897== by 0x4E33BCB: logp2 (logging.c:246) ==21897== by 0x435E5C: listen_fd_cb (ipaccess.c:692) ==21897== by 0x4E32419: osmo_select_main (select.c:132) ==21897== by 0x40555B: main (bsc_hack.c:306) ==21897== ==21897== Use of uninitialised value of size 8 ==21897== at 0x6101CE0: _IO_default_xsputn (genops.c:480) ==21897== by 0x60D4ED3: vfprintf (vfprintf.c:1620) ==21897== by 0x60F8F01: vsnprintf (vsnprintf.c:120) ==21897== by 0x4E33670: _output (logging.c:177) ==21897== by 0x4E33878: _logp (logging.c:227) ==21897== by 0x4E33BCB: logp2 (logging.c:246) ==21897== by 0x435E5C: listen_fd_cb (ipaccess.c:692) ==21897== by 0x4E32419: osmo_select_main (select.c:132) ==21897== by 0x40555B: main (bsc_hack.c:306) ==21897== ==21897== Conditional jump or move depends on uninitialised value(s) ==21897== at 0x6101CF2: _IO_default_xsputn (genops.c:479) ==21897== by 0x60D4ED3: vfprintf (vfprintf.c:1620) ==21897== by 0x60F8F01: vsnprintf (vsnprintf.c:120) ==21897== by 0x4E33670: _output (logging.c:177) ==21897== by 0x4E33878: _logp (logging.c:227) ==21897== by 0x4E33BCB: logp2 (logging.c:246) ==21897== by 0x435E5C: listen_fd_cb (ipaccess.c:692) ==21897== by 0x4E32419: osmo_select_main (select.c:132) ==21897== by 0x40555B: main (bsc_hack.c:306) ==21897== <000d> input/ipaccess.c:755 accept()ed new RSL link from 192.168.200.117 <0004> bsc_init.c:251 bootstrapping RSL for BTS/TRX (0/0) on ARFCN 871 using MCC=602 MNC=42 LAC=1 CID=0 BSIC=63 TSC=7 ==21897== Use of uninitialised value of size 8 ==21897== at 0x60D2EDB: _itoa_word (_itoa.c:195) ==21897== by 0x60D5386: vfprintf (vfprintf.c:1619) ==21897== by 0x60F8F01: vsnprintf (vsnprintf.c:120) ==21897== by 0x4E33670: _output (logging.c:177) ==21897== by 0x4E33878: _logp (logging.c:227) ==21897== by 0x4E33BCB: logp2 (logging.c:246) ==21897== by 0x410141: inp_sig_cb (bsc_init.c:247)
Can you please investigate? I suspect it to be related to your recent _output() related changes in libosmocore.
Not sure what's wrong yet, I don't hit those problems here, probabling we're using a different config file?
You may want to revert this and give it a try:
http://cgit.osmocom.org/cgit/libosmocore/commit/?id=81e9636454294ae10ef9bc8b...