On Tue, Dec 29, 1998, [EMAIL PROTECTED] wrote:
> sorry for mostly sleeping concerning mod_ssl over the last few weeks
> (I followed up to 2.1.0 but not afterwards).
>
> I just tested mod_ssl DSO support with Solaris 2.5.1, gcc-2.7.2:
> (tested with 2.1.0, 2.1.1 and 2.1.4)
>
> Everything works fine until the first restart (SIGHUP or SIGUSR1) to
> the server process:
>
> until then every SSL connection will get a SEGV directly after
> getting connected:
>
> 25563: write(19, 0xEFFFEA18, 95) = 95
> 25563: [ 2 9 / D e c / 1 9 9 8 1 1 : 3 1 : 5 1 ] [ i n f o ] C
> 25563: o n n e c t i o n t o c h i l d 1 e s t a b l i s h e d
> 25563: ( s e r v e r w e b d e v . d e . u u . n e t : 4 4 3 )\n
> 25563: brk(0x00294EC8) = 0
> 25563: brk(0x00296EC8) = 0
> 25563: Incurred fault #6, FLTBOUNDS %pc = 0xEF1EFD18
> 25563: siginfo: SIGSEGV SEGV_MAPERR addr=0xEF3DD37C
> 25563: Received signal #11, SIGSEGV [default]
> 25563: siginfo: SIGSEGV SEGV_MAPERR addr=0xEF3DD37C
>
> the problem for me is, that under any circumstances I did not get a
> core to look more closely at the source :-(
>
> any suggestions (to get a core or fix the problem) ?
Wasn't Solaris one of the platforms where getting a core isn't possible for
UID 0 based processes? Perhaps then you have to run it on ports 8080/8443
temporarily as a regular user to get the core. Without the coredump and a
backtrace it's hard to find the location of the SIGSEGV.
I'll try it out myself on my Solaris 2.6 box the next days.
In the meantime perhaps you can get a core on your platform, Jan?
Ralf S. Engelschall
[EMAIL PROTECTED]
www.engelschall.com
______________________________________________________________________
Apache Interface to SSLeay (mod_ssl) www.engelschall.com/sw/mod_ssl/
Official Support Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]