httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <br...@hyperreal.org>
Subject Re: Another runaway process on hyperreal
Date Tue, 04 Aug 1998 06:28:43 GMT

Compiling with 

-Wall -DSUEXEC_BIN=\"/usr/local/bin/suexec\" -Wshadow -Wmissing-prototypes
-Wmissi
ng-declarations -Wpointer-arith -pipe -g -O2

I've recompiled without -O2, after syncing with the current tree.  I'll
post again if we get another core.

	Brian

At 12:57 PM 8/3/98 -0700, Dean Gaudet wrote:
>Perhaps Brian is compiling with -fomit-frame-pointer or with a -O level
>that's high enough to lose the frame pointer... and that is messing up the
>gdb stack trace.  I doubt it's in the middle-of-nowhere.
>
>Dean
>
>On Sun, 2 Aug 1998, Roy T. Fielding wrote:
>
>> >I'm not Dean, but no way. This is not shared memory!
>> 
>> I was thnking it might be a performance optimization on freebsd,
>> since the children never write to that memory.  In any case, I haven't
>> found any other way for the handlers array to get screwed.
>> 
>> ....Roy
>> 
> 
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
"Common sense is the collection of prejudices  |     brian@apache.org
acquired by the age of eighteen." - Einstein   |  brian@hyperreal.org

Mime
View raw message