On Tue, Jan 25, 2005 at 12:18:15AM -0600, David Young wrote:
> Just when I thought everything was working....  Earlier today I saw an
> MCHK trap in uvm_analloc.  Now I've seen a new trap:
> 
> panic+0x19c
> trap+0xfc
> kernel MCHK trap by splraise+0x20: srr1=0x49030
>     r1=0xd5433c80 cr=0x88004042 xer=0 ctr=0x29d2bc
> bio_doread+0x5c
> ADBDevTable+0xffa017e8
> bread+0x20
> ffs_blkatoff+0x80
> VOP_BLKATOFF+0x48
> ufs_dirrewrite+0x60
> ufs_rename+0x524
> VOP_RENAME+0x50
> rename_files+0x2c4
> syscall_plain+0xe0
> user SC trap #128 by 0xeff0775c: srr1=0xf032
>    r1=0xffffda70 cr=0x44004084 xer=0 ctr=0xeff07754
> saved LR(0xfffffffb) is invalid
> 
> I was extracting man.tgz with pax in single-user mode when this panic
> occurred.  This is with 1 day-old -current; I'm going to update to this
> minute's sources and try again.

Same problem with new sources.  The offending instruction is this:

lwz r0, r1, 0x14

'show registers' indicates:

r0	0x31d2d0	panic+0x1a0
r1	0xd5433aa0

I can reproduce this easily by extracting a man.tgz or misc.tgz in the
root directory.

Dave

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Urbana, IL * (217) 278-3933