httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug Bloebaum <>
Subject Re: Segmentation faults :(
Date Fri, 09 Oct 1998 08:31:41 GMT

I said (re: SEGV in httpd-1.3.3):

 > >mod_perl certainly appears to be involved...

Oops, sorry for the potentially false accusation.

Doug MacEachern wrote:
 > Except that there is no mod_perl code in the distribution that
 > calls mod_perl_set_opmask().  There's an xsub in Apache.xs so Perl
 > code can call $r->set_opmask(...), but it'll croak unless you've
 > enabled the experimental and undocumented opmask stuff.  I'd be
 > really surprised if you have code that's trying to use this stuff.
 > Can you compile with -g and get another backtrace so we can see
 > line numbers, etc.?

Here it is:

(Apache/1.3.3 (Solaris 2.5.1) mod_perl/1.16)

(gdb) where
#0  0x55168 in ap_proxy_send_fb (f=0x4a99a8, r=0x4a8840, c=0x4a98a8)
    at proxy_util.c:594
#1  0x537f4 in ap_proxy_http_handler (r=0x4a8840, c=0x4a98a8, 
    url=0x4a981e "",

    proxyhost=0x0, proxyport=0) at proxy_http.c:518
#2  0x48a24 in proxy_handler (r=0x4a8840) at mod_proxy.c:396
#3  0x8bb68 in ap_invoke_handler (r=0x4a8840) at http_config.c:505
#4  0xa615c in process_request_internal (r=0x4a8840) at http_request.c:1151
#5  0xa61dc in ap_process_request (r=0x4a8840) at http_request.c:1167
#6  0x9b110 in child_main (child_num_arg=47) at http_main.c:3826
#7  0x9b46c in make_child (s=0x164840, slot=47, now=907920306)
    at http_main.c:3945
#8  0x9b9a8 in perform_idle_server_maintenance () at http_main.c:4100
#9  0x9c134 in standalone_main (argc=3, argv=0xeffffd6c) at http_main.c:4331
#10 0x9c998 in main (argc=3, argv=0xeffffd6c) at http_main.c:4534

 > >By the way, to potentially save someone else the pain of figuring
 > >out how to get a core file from a setuid Apache httpd process in
 > >Solaris, without knowing which httpd child might be the one to
 > >die, here's what I ended up doing:
 > Can't you just use the CoreDumpDirectory directive?

Nope, it's not a file permissions thing.  From the Solaris core(4)

    A process with an effective user ID different from the real user
    ID will not produce a core image.

Doug Bloebaum                             interactive investor
Systems Engineer                          105-109 Strand, London WC2R 0AB                

View raw message