Software issue or (uggh) hardware?

[Bob Raymond]

Hi,

Details of situation:

Kernel 2.4.19-xfs won't boot after a power outage (that lasted only 5 secs!)
Kernel 2.5.42-ac1 locks up sometimes, and I can't get KDE, Gnome, etc., to 
start before it oopses, and now it, along with 2.5.41-ac2 (which always 
worked before) and 2.5.40-bk5, sometimes oops even before X can start.  
However, sometimes in the middle of booting any kernel, the system shuts 
itself off, and I get the "CPU overclock is failed" message.  I ran FreeBSD 
all through the night, because I can at least check e-mail, etc. there, and 
it locked up in the middle of the night as well.  I checked CPU/System temps 
in the BIOS and got 33C - 34C for the CPU, 29-34C for the system (it varied a 
lot).  I think my fan monitor actually is borken because it noted RPMS of 
9999, then 0, then back again.  Anyone have any ideas of what might be wrong?

BTW, none of the modules load in the 2.4.19-xfs kernel.   Unfortunately I lost 
the XFS patches for my 2.4.19 kernel and wvdial and my system aren't getting 
along at this point (KPPP won't work if I get an oops too early to get into 
KDE), so a kernel recompile is out of the question for anything but the 2.5 
series right now.  I also get an oops when GPM tries to load.  Funny what one 
little power outage can do to wreak havoc over an entire system :-(

oh, yeah- hardware specs:
EPoX 8KTA3Pro, AMD Athlon TBird 1.4 (not o/c'ed), 512MB PC133 SDRAM
ATI Radeon 8500, Antec SX1240 case with 400watt PS, TB Santa Cruz (right now 
Alsa modules for this won't load), Promise UDMA100 controller (FreeBSD is on 
this), IBM Deathstar 60GXP 41GB as /dev/hda on /mnt/winxp, Maxtor something 
or other 60GB as /dev/hdb on /, IBM Deathstar 120GXP 40GB as /dev/hde on 
/mnt/freebsd.

				TIA,

				Bob Raymond