>Number:         2297
>Category:       port-i386
>Synopsis:       NFS server crashes on i386 system
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Apr  3 20:35:02 1996
>Last-Modified:
>Originator:     Matthias Scheler
>Organization:
Matthias Scheler
tron@colwyn.owl.de
>Release:        960402
>Environment:
System: NetBSD colwyn 1.1B NetBSD 1.1B (COLWYN) #0: Wed Apr 3 02:13:11 MET DST 1996 tron@colwyn:/usr/src/sys/arch/i386/compile/COLWYN i386


>Description:
When I entered "ls -la /homes/colwyn/src/lib/libc" on my NetBSD-amiga 1.1B
system my NetBSD-i386 1.1B system serving "/homes/colwyn" via NFS stopped
with this error message:
vm_fault(f82a5000, f87d6000, 1, 0) -> 1
kernel: page fault trap, code=0
Stopped at      _nfsrv_readdir+0xca4:   cmpl   $0,(0%ebx)

I rebooted and the problem occured again. This time I entered "cont"
at the debugger prompt and the NetBSD-i386 continued to run without
any problems.

The problem happens only with the NetBSD-i386 system as server. I wasn't
once able to crash the NetBSD-amiga system that way. So I suppose it's
an i386 specific problem.

>How-To-Repeat:
That's not easy. I can't reproduce it every time and the problem only
shows up if I access that particular directory with "ls -la". Maybe the
code offset will help you.

>Fix:
Sorry, no idea.
>Audit-Trail:
>Unformatted: