Subject: Re: port-i386/6337: fatal page fault in supervisor mode
To: None <navas@bayarea.net>
From: Chris G. Demetriou <cgd@netbsd.org>
List: netbsd-bugs
Date: 10/20/1998 16:54:05
George Navas <navas@bayarea.net> writes:
> >Number:         6337
> >Category:       port-i386
> >Synopsis:       trap type 6 code 0 eip f0142a6c cs 8 eflags 10206 cr2 14 cpl 0
> >Confidential:   no
> [ ... ]
> #3  0xf02148fa in trap (frame={tf_es = -251592688, tf_ds = -258408432, 
>       tf_edi = -252049152, tf_esi = -251997696, tf_ebp = -37675252, 
>       tf_ebx = -252536064, tf_edx = 0, tf_ecx = 0, tf_eax = 2147483647, 
>       tf_trapno = 6, tf_err = 0, tf_eip = -267113876, tf_cs = 8, 
>       tf_eflags = 66054, tf_esp = 11, tf_ss = -251997696, tf_vm86_es = 11, 
>       tf_vm86_ds = -266314792, tf_vm86_fs = -264904704, 
>       tf_vm86_gs = -37675512}) at ../../../../arch/i386/i386/trap.c:238

Alas, this trace doesn't say where the bug is occurring, only the
trace after the trap.

Do you have a kernel cmpiled with debugging symbols?  Can you gdb that
kernel, then say "list *0xf0142a6c" (the eip where the fault
happened)?

If no kernel with debug symbols, can you "disass 0xf0142a6c" and reply
with the result?

please reply to this message (with all cc's) so that it gets appended
to the PR.



> >Fix:
> 	Don't have one, but how I wish I had one.

Well, we'll see if we can fix that.  8-)



cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/Pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.