[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Bug 62552] Loading mod_md segfaults when using DBDriver from mod_authn_dbd

--- Comment #13 from Stefan Eissing <stefan@xxxxxxxxxxx> ---
I agree that something seems strange, which normally indicates that there is a
hidden variable which we are not aware of yet.

In your backtrace attached here, I read:

[Wed Jul 18 11:23:52.903327 2018] [so:debug] [pid 10467:tid 140737354008448]
mod_so.c(266): AH01575: loaded module ssl_module from 
Program received signal SIGSEGV, Segmentation fault.
0x00007fffef1c2610 in ?? ()

(gdb) thread apply all bt

Thread 1 (Thread 0x7ffff7fe0780 (LWP 10467)):
#8  0x00005555555c0638 in load_module (cmd=0x7fffffffdaa0, dummy=<optimized
out>, modname=0x555555853a68 "md_module", filename=0x555555853a78
"modules/") at mod_so.c:302
#9  0x00005555555ac410 in invoke_cmd (cmd=0x5555557f35e0 <so_cmds>,
parms=parms@entry=0x7fffffffdaa0, mconfig=mconfig@entry=0x7fffffffd5b0,
args=<optimized out>, args@entry=0x55555582dc53 "md_module modules/")
at config.c:948

That is why I thought mod_ssl has been loaded before mod_md. How am I reading
that wrong?

Once we know for sure exactly which ordering is problematic, maybe we can find
the difference between mod_ssl's initalization of openssl and the way mod_md
does it.

The other thing that worries me is: if you say that only the configuration

DBDriver mysql

triggers the segfault then this can only happen if that config line comes
_before_ the LoadModule of mod_md. But in your minimal example, this is not the
case. So, I think we do not have a clear model of what happens when...

You are receiving this mail because:
You are the assignee for the bug.
To unsubscribe, e-mail: bugs-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: bugs-help@xxxxxxxxxxxxxxxx