Hi.  I'm running -current (with Manuel's pciide/dma patches currently,
but I was seeing odd behavior before, too) on an AMD K6/233 with 128M
of RAM and 128M of swap.

It's now consistently crashing when I run amanda's amdump command, so I
compiled DEBUG and have a core file.  I seem to recall some discussion
of crashes with 'panic: ptdi' in the past.  If anyone has any
suggestions on where I can go with this to get to the bottom of this
particular panic, I'd appreciate it.

I suspect that there might be a hardware bug, too, based on some of the
other behavior that I've seen.  I have been running the rc5des client
from distributed.net.  If I try to compile a kernel (or anything else),
I'll get a segfault and core dump in cc1.  If I suspend rc5des, the
compile proceeds OK.  I have also seen panics in vmfault (uvmfault when
I was running UVM/PMAP_NEW) and getblk().

The traceback from gdb matches the one I saw in ddb (thankfully).  It is:

#3  0xf012ca35 in panic ()
#4  0xf0207f63 in pmap_enter ()
#5  0xf01f5fae in vm_fault ()
#6  0xf01f60e5 in vm_fault_wire ()
#7  0xf01f828d in vm_map_pageable ()
#8  0xf01f64b1 in vslock ()
#9  0xf0124077 in physio ()
#10 0xf02159b8 in stwrite ()
#11 0xf014f14e in spec_write ()
#12 0xf01f29c9 in ufsspec_write ()
#13 0xf014809e in vn_write ()
#14 0xf012e594 in sys_write ()
#15 0xf020a174 in syscall ()

I've fiddled with the BIOS settings to try to choose the safest
possible settings to no avail.

Any thoughts or suggestions?

-allen

-- 
                    Allen Briggs - briggs@ninthwonder.com