Subject: re: netbsd32 vs. alpha
To: Dave Huang <khym@bga.com>
From: matthew green <mrg@eterna.com.au>
List: port-alpha
Date: 01/05/2000 17:23:56
Is anyone else having trouble with a recent -current kernel on the
alpha? If I grep for something in a few files, things are okay, but if I
grep in a bunch of files, grep dies with SIGABRT and no core dump. I
don't know what the cutoff is, but it seems to have something to do with
the length of the command line. The last thing ktrace sees is the NAMI
on the executable. Also, OSF1 binaries die the same way, regardless of
the length of the commandline.
I grepped for SIGABRT in /usr/src/sys/kern (which I couldn't do all at
once :), and the only mention of it is in kern_exec.c, which was
recently changed for the EXEC_32 stuff. Whereas before, it'd always use
the sizeof(char *) version, now it uses sizeof(int) unless EXEC_32 is
set, which I'm guessing it's not :) So it sounds like it's no longer
allocating enough room on the stack for the commandline or something
like that...
i see the same problem on my alpha. it is annoying. i believe that
the logic in this code is backwards and it wants to have these then
and else clauses swapped.
eeh?