dbf6192d34f6b7fa1f0c.newsdawg.com
From: "Dariusz Piatkowski"
Hi Dave!
On Sun, 22 Jan 2012 03:18:04 UTC, Dave Yeo wrote:
> Dariusz Piatkowski wrote:
> > Hi Trevor!
> >
> > On Sat, 21 Jan 2012 20:40:03 UTC, "Trevor Hemsley"
> > wrote:
> >
> >> On Sat, 21 Jan 2012 17:03:14 UTC in comp.os.os2.misc, "Dariusz Piatkowski"
> >> wrote:
> >>
> >>> I am able to run my 14.105 SMP based stuff in a 4-core configuration on
my new
> >>> MSI board...however attempting to even boot when the full 6-core setup is
> >>> enabled in BIOS causes the machine to hand HARD in RESOURCE.SYS, even
though I
> >>> am still attempting to boot with just a 4-core enabled OS2 configuration
through
> >>> OS2APCI.PSD driver. Subsequently my machine currently has the 2 of the 6
cores
> >>> physically disabled in BIOS.
> >>>
> >>> So...I'm thinking the driver is running into some problems. I have never
seen
> >>> this before, and just about every mention of this problem I'm finding on
the net
> >>> is dealing with eCS ACPI boots...which I'm not certain I can relate to
in any
> >>> way.
> >>>
> >>> Any hints as to what may be at play here? What/where to look for the
issue?
> >>
> >> The ECS acpi stuff requires that you install their version of
resource.sys. No
> >> idea why that is as they do not say. Perhaps it is related to your issue,
> >> perhaps it has nothing to do with it.
> >
> >
> > I'm not running any eCS SMP stuff on my machine however. It's all IBM
stuff, and
> > the 4-core SMP configuration is courtesy of OS2APIC.PSD driver.
> >
> > All I can tell is that I get a HARD hang when the boot logo displays
> > ..RESOURCE.SYS...at the bottom of the screen (using ALT-F2 during boot
start).
> > Both (main and maintenance) my partitions hang, that means 14.104&
14.105. I
> > suppose that might mean the driver which JUST loaded prior to RESOURCE.SYS
hung
> > the system...however, I'm not sure of how to determine that.
>
> Have you got hardware detection turned on? You could try removing
> everything from snoop.lst except ibmkbd.snp and pcibus.snp in that order.
> Also the IBM supplied ibmkbd.sys is buggy when it comes to SMP,
> unluckily you have to purchase eCS to get the updated one.
> You could also experiment with underclocking your system, at least that
> shouldn't fry anything.
I tried several different approaches:
1) the maintenance partition is a bare-bones configuration, snoop.lst doesn't
even exist there, yet I see the same hang...so I'm thinking it may be
something
outside of snoop.lst, not quite sure what though
2) I did try a bare-bones snoop.lst in my regular partition, that did not have
any effects on the hang
3) I tried FULL hardware detection as well, no difference
4) finally, I under-clocked the machine, in several stages, down to a 2Gig
setting from a default of 3.4, no difference there
Relying on some of the other system tools I see that 'dmidecode' fails to
produce output, what I get is:
=== START ===
[G:\test\dmidecode]dmidecode
# dmidecode 2.10
SMBIOS 2.5 present.
54 structures occupying 1934 bytes.
Table at 0x0009F800.
Invalid entry length (0). DMI table is broken! Stop.
=== STOP ===
In comparison, the previous MSI board showed the following result:
=== START ===
# dmidecode 2.10
SMBIOS 2.6 present.
53 structures occupying 2033 bytes.
Table at 0x000FCEB0.
...details of the dump....
=== STOP ===
So at the very least, the NEW board is using an older version of SMBIOS...and
it
appears to be incorrect...I can always try to revert back to an older version
of
BIOS itself and see if that fixes the issue.
PCI scan produces a good result however, no IRQ conflicts, etc.
--- Internet Rex 2.31
* Origin: NewsGuy - Unlimited Usenet $19.95 (1:261/20.999)
|
