> My guess is that the SIGBUS is probably something that happens because
> you've got it running under the debugger. The SIGSEGV more looks like
> the real problem here - I'm no expert at reading debugger output ;-) but
> it looks like the problem is in adding entries to the session cache, when
> adjusting the size of the table/buckets. I haven't looked too closely at
> the code yet, but a guess would be that you're filling up the session
> cache. What if you run with SSLSessionCache none - does it still segfault?
The sigbus is the real problem, I think the sigsegv is a result of having
all the children under the debugger. :-)
The reason being, is that after setting CoreDumpDirectory in my httpd.conf
and debugging the core dumps after the SIGBUS, I get the same stack traces
as the first ones I got. Here's a stack trace from the core file generated
from SIGBUS:
> dbx ../bin/httpd core
dbx version 7.2.1.2m patch3441 Oct 20 1998 14:00:38
Core from signal SIGBUS: Bus error
(dbx) where
> 0 memcpy(0x100fa634, 0x43cc000, 0x37, 0x8, 0x1, 0x37, 0x0, 0x37)
["/xlv47/6.5.8m/work/irix/lib/libc/libc_n32_M4/strings/bcopy.s":531,
0xfadeb30]
1 ap_bwrite(0x100bfac8, 0x43cc000, 0x37, 0x8, 0x1, 0x37, 0x0, 0x37)
["/local/share/downloads/Apache/apache_1.3.12/src/main/buff.c":1329,
0x10019018]
2 ap_send_mmap(0x43cc000, 0x1015ff50, 0x0, 0x37, 0x1, 0x37, 0x0, 0x37)
["/local/share/downloads/Apache/apache_1.3.12/src/main/http_protocol.c":2332
, 0x10039614]
3 default_handler(0x1015ff50, 0x7ffe8be1, 0x7ffe8be1, 0x8, 0x1, 0x37,
0x0, 0x37)
["/local/share/downloads/Apache/apache_1.3.12/src/main/http_core.c":3224,
0x10032d14]
4 ap_invoke_handler(0x1015ff50, 0x101020a8, 0x0, 0x0, 0x10160ad8, 0x0,
0x0, 0x37)
["/local/share/downloads/Apache/apache_1.3.12/src/main/http_config.c":520,
0x10025020]
5 ap_run_sub_req(0x1015ff50, 0x43cc000, 0x37, 0x8, 0x1, 0x37, 0x0, 0x37)
["/local/share/downloads/Apache/apache_1.3.12/src/main/http_request.c":929,
0x1001be74]
6 <Unknown>() [< unknown >, 0x5fee3c48]
(dbx)
As you can see it's the same as the first SIGBUSs I posted.
Since this is a SIGBUS in memcpy, this means that either the data we're
reading from or writing to is bad, but I'm not sure which at this point.
Hopefully Ralf can take a look, maybe he has a better idea of what's going
on.
-Dave
______________________________________________________________________
Apache Interface to OpenSSL (mod_ssl) www.modssl.org
User Support Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
