go to bug id or search bugs for
I tried to run APC (downloaded from CVS 30minutes ago)
CFLAGS="-O3 -march=athlon64 -ffast-math -funroll-all-loops -funit-at-a-time -fpeel-loops -ftracer -funswitch-loops -fomit-frame-pointer -pipe"
always getting hunderts logs i erros_log:
[Sat Feb 25 12:17:39 2006] [notice] child pid 25298 exit signal Segmentation fault (11)
[Sat Feb 25 12:17:39 2006] [notice] child pid 25410 exit signal Segmentation fault (11)
[Sat Feb 25 12:17:39 2006] [notice] child pid 25440 exit signal Floating point exception (8)
[Sat Feb 25 12:17:39 2006] [notice] child pid 25442 exit signal Floating point exception (8)
[Sat Feb 25 12:17:39 2006] [notice] child pid 25496 exit signal Floating point exception (8)
[Sat Feb 25 12:17:40 2006] [notice] child pid 25526 exit signal Segmentation fault (11)
Add a Patch
Add a Pull Request
Can you get me a backtrace of one of these crashes?
Here is full backtrace made from coredump of crashed server with error: "child pid 25410 exit signal
Segmentation fault (11)"
I am ussing CVS(09.04.2006) version of APC
I am using apache 2.2.0 on SuSE 10.064bit with Athlon64
I was trying MPM: event and worker with the same result.
Does it happen with the prefork mpm? If so, please get a backtrace from that and compile without any optimization, not even -O2 to give me a readable backtrace.
Prefork gives me such errors:
#0 ZEND_UNSET_VAR_SPEC_CV_HANDLER (execute_data=0x7fffff819900)
#1 0x00002b8a8c36994b in execute (op_array=0x58d74a0) at zend_vm_execute.h:92
#2 0x00002b8a8c36eb99 in ZEND_INCLUDE_OR_EVAL_SPEC_CONST_HANDLER (
execute_data=0x7fffff81b460) at zend_vm_execute.h:1918
#3 0x00002b8a8c3698fb in execute (op_array=0x58c9f70) at zend_vm_execute.h:92
#4 0x00002b8a8c341386 in zend_execute_scripts (type=8,
retval=<value optimized out>, file_count=3)
#5 0x00002b8a8c2f0e43 in php_execute_script (primary_file=0x7fffff81da10)
#6 0x00002b8a8c3fc215 in php_handler (r=0x58a40c8)
#7 0x000000000043bb7d in ap_invoke_handler (r=0x58a40c8) at config.c:157
#8 0x000000000045d790 in ap_process_request (r=0x58a40c8)
#9 0x0000000000459a19 in ap_process_http_connection (c=0x589e288)
#10 0x000000000044280c in ap_process_connection (c=0x589e288, csd=0x589e098)
#11 0x00000000004805c7 in child_main (child_num_arg=<value optimized out>)
#12 0x0000000000480944 in make_child (s=0x5cafe8, slot=2) at prefork.c:736
---Type <return> to continue, or q <return> to quit---
#13 0x00000000004817d5 in ap_mpm_run (_pconf=<value optimized out>,
plog=<value optimized out>, s=<value optimized out>) at prefork.c:871
#14 0x0000000000421020 in main (argc=3, argv=0x7fffff81df18) at main.c:712
Full backtrace is available here:
Sorry, I used -O3. I am now recompiling without any optimization and will provide backtrace within 10 minutes.
It's quite funny, but httpd 2.2 prefork compiled witth CFLAGS="" + PHP5.1.2 with CFLAGS="" + APC(current cvs) works fine and have no crashes :)
How many strict aliasing errors do you get during the entire build ?
Also you might want to build it using -fno-strict-alias and -fno-gcse to confirm whether it is a hazard in instruction scheduling.
Closed as bogus (gcc optimization issue)