On Sun, Jul 01, 2001 at 09:53:24PM +0200, Manuel Bouyer wrote:
> FYI, I have the same problem on my alpha. I've also failures from more.
> I've also seen this on my vax a few weeks ago and I didn't investigate further.
> I've a new make build running on this machine, I'll see if it fails too.

OK, my vax has the problem too: the old awk binary dumps core "bad system call"
with the new libc, in whith 'make includes' in sys/arch
#0  0x1897b6c in __syscall (0, 0, 0)
(gdb) where
#0  0x1897b6c in __syscall (0, 0, 0)
#1  0x1897b1e in lseek (0, 0, 0, 0)
#2  0x10936 in optimal_bufsize (0, 2147482052)
#3  0x13397 in pathopen (0, 71025, 158820)
#4  0x12ce7 in devopen (71025, 71023, 158820)
#5  0x11751 in do_nextfile (0)
#6  0x119a8 in do_input ()
#7  0x141ac in main (2, 2147482536, 2147482548)
#8  0x109e in __start (2147482532)
#9  0x1028 in start

Hum, I've already seen this the last time. It dumps core on lseek
when input is a pipe.
I'm running a kernel without NEW_PIPE, and I've already seen this long
before NEW_PIPE was integrated.

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--