For a fresh hd install to all reiserfs of yesterday's cooker, i.e. 8.2beta 
w/contribs.  This new Ali M1647 motherboard with Radeon 32ddr agp card, Athlon 
xp1700 will perform a sloppy-looking reboot successfully with kernel-2.4.8-26mdk 
and newer cooker kernels on 8.1 but hangs on halt, with disconcerting fields of 
red and shutdown messages overwritten on what looks like the boot messages. 
With 8.2, the halt process looks quite normal until:

Unmounting file system:            [OK]
Halting system...
md: stopping all md devices.
flushing ide devices: hda hdc hdd
Power down.
general protection fault: fdf8
CPU:0
EIP: 0050:[<00008887>]  Tainted: P
EFLAGS: 00010086
eax:, ebx:, etc. (available)

Process halt (pid:3333, stackpage=dca07000)

Stack: (available)

Call Trace: (available)
Code: Bad EIP value.
/etc/rc0.d/S01halt: line1: 3333 Segmentation fault   halt -i -d -p


After Alt-SysRq-S, U, O,

<0>Kernel panic: Aiee, killing interrupt handler!
In interrupt handler-not syncing

I must reset to proceed and attempt to boot either any kernel of 8.1 on hda6 or 
8.2 on hda11 from grub/menu.lst on hda6 results in "inconsistent filesystem". 
W98 will boot from grub.  Either Mandrake will boot from a floppy.  reiserfsck 
-x on hda11 and hda6 makes no difference.  reiserfsck --rebuild-tree /dev/hda11 
and either Mandrake now boots from grub.

One more troubling message from boot messages of all kernels:

Ali15x3: IDE controller on PCI bus 00 dev 20
PCI: no IRQ known for interrupt pin A of dev 00:04:0.
Please try using pci=biosirq

Adding pci=biosirq to the grub entry causes the 'Please try' message to go away 
but nothing else.

Also, sensors-detect wants to insmod i2c-ali15x3 but fails, saying 
/lib/modules/(uname)/kernel/drivers/i2c/i2c-ali15x3.o.gz does not exist, which 
it does.  Trying to modprobe this by hand fails the same, while modprobe of 
random other driver is OK.

This bios is 'maturing' rapidly and there might be some improvement there. 
Many, like myself, seem to be attracted to this board to escape the problems of 
VIA but it now appears the Athlon and AGP are posing new problems.  Thought 
there might be some utility to getting this report out before I am swallowed up 
completely:p

Thanks.


Reply via email to