[CURRENT-20021016] innfeed dumps core

Russ Allbery rra at stanford.edu
Tue Nov 5 00:35:09 UTC 2002


Jeffrey M Vinocur <jeff at litech.org> writes:

> Haven't looked at the other stuff, but this is pretty funny looking.
> That's either severely data corruption of some sort, or we're making the
> incorrect assumption that malloc() zeros memory.

Okay, that bit of stupidity took me a while to find, and I read over it
twice without noticing the problem.  *sigh*  Mea culpa.

Committing this patch now, which will fix the problem.

The key observation that finally helped me figure out what was going on
was that the backtraces were showing that INN was trying to execute code
at 0x2, not just access that memory value.  So then I started thinking
about the message handling callbacks, and took a closer look at them.

Index: main.c
===================================================================
RCS file: /dist1/cvs/isc/inn/inn/innfeed/main.c,v
retrieving revision 1.43
diff -u -p -r1.43 main.c
--- main.c	2002/10/07 12:18:28	1.43
+++ main.c	2002/11/05 00:33:16
@@ -147,7 +147,7 @@ int main (int argc, char **argv)
 #endif
 
   message_handlers_die (2, error_log_stderr_date, message_log_syslog_err) ;
-  message_handlers_warn (2, message_log_syslog_warning);
+  message_handlers_warn (1, message_log_syslog_warning);
   message_handlers_notice (1, message_log_syslog_notice) ;
 
 #define OPT_STRING "a:b:c:Cd:e:hl:mMo:p:S:s:vxyz"

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.


More information about the inn-workers mailing list