Re: Intermittent Seg Fault with PHP4.3.8 + FTP
| From: | Steve Brown | Date: | Wed, 08 Sep 2004 15:14:18 +0000 |
| Subject: | Re: Intermittent Seg Fault with PHP4.3.8 + FTP | ||
| References: | 1 2 3 4 5 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-196134@lists.php.net to get a copy of this message | ||
> OK. Sounds like a genuine segfault to me. The CLI version will simply print
> "Segmentation fault" to the screen or "Segmentation fault (core dumped)"
> when configured with debugging enabled.
Thanks for the help Jim. I had to install gdb and recompile php, but
I was able to get consistant backtraces from both systems. When the
fault occured, the backtrace was identical every time. It appears to
be an issue with memchr() in ftp_get():
=================
Program received signal SIGSEGV, Segmentation fault.
0x4207bae0 in memchr () from /lib/tls/libc.so.6
(gdb) bt
#0 0x4207bae0 in memchr () from /lib/tls/libc.so.6
#1 0x0807ebb0 in ftp_get (ftp=0x8366c4c, outstream=0x83a2204,
path=0x839868c "/x-stuff/mir_libraries/lib-nusoap-066.php",
type=FTPTYPE_ASCII, resumepos=0)
at /usr/local/src/php-4.3.8/ext/ftp/ftp.c:730
#2 0x0807bf69 in zif_ftp_get (ht=4, return_value=0x83a0f9c,
this_ptr=0x0, return_value_used=1) at
/usr/local/src/php-4.3.8/ext/ftp/php_ftp.c:637
#3 0x081ecfb0 in execute (op_array=0x836c920) at
/usr/local/src/php-4.3.8/Zend/zend_execute.c:1635
#4 0x081ed22b in execute (op_array=0x836d648) at
/usr/local/src/php-4.3.8/Zend/zend_execute.c:1679
#5 0x081ed22b in execute (op_array=0x8366b74) at
/usr/local/src/php-4.3.8/Zend/zend_execute.c:1679
#6 0x081d9783 in zend_execute_scripts (type=8, retval=0x0,
file_count=3) at /usr/local/src/php-4.3.8/Zend/zend.c:891
#7 0x0819e9b7 in php_execute_script (primary_file=0xbffffad0) at
/usr/local/src/php-4.3.8/main/main.c:1734
#8 0x081f3e3d in main (argc=2, argv=0xbffffb64) at
/usr/local/src/php-4.3.8/sapi/cli/php_cli.c:822
=======================
I did some searching through the bug list and didn't find anything
that appeared to be related to this issue. I've created bug #30027,
so hopefully this will be addressed soon. Thanks for your help!
Steve