Re: GW Solo 5350 Laptop APIC trouble?

From: Mikael Pettersson (mikpe@csd.uu.se)
Date: Mon Apr 29 2002 - 15:37:00 EST


On Mon, 29 Apr 2002 12:02:05 -0600 (MDT), Benson Chow wrote:
>I'm running 2.4.18 on my GW Solo 5350, and if I compile APIC in for
>uniprocessors, it will fail pretty horribly - something like this:
>
>APIC + APM: System Freeze on idle, random time, instant Fn key freeze.
>APIC + ACPI: System Freeze when using Fn key with smm feature.
>APM w/o APIC: Works fine, but Fn key instant on-screen status info is
>corrupt
>ACPI w/o APIC: Haven't tried yet.
>Redhat 7.2 Kernel (2.4.9-RH I believe): No crashes, Fn-key onscreen works
>fine.
>...
>I'm still trying to figure out what's in the RH kernel that makes it work
>fine, that's when I found out compiling in APIC aggravated the failures.
>...
>Anyone have any ideas how to start debugging this? Chances are it's a
>hardware issue, I hope it can be worked around in software without
>outrightly disabling or not using the features.

RedHat's 7.2 UP kernels don't have CONFIG_X86_UP_APIC=y.
We've seen failures like these before. To date, they've always
been caused by BIOS problems, and chances are your machine also
has a local-APIC hostile BIOS.

The only known Linux workaround is to not enable the local APIC.

I maintain a local APIC blacklist, which the kernel uses to avoid
enabling the local APIC on buggy machines. If after testing you
are sure that keeping the local APIC disabled is a prerequisite
for stable operation on your machine, then please apply the patch
below and send me the DMI data from the kernel boot log so that
I can add your machine to the blacklist.

/Mikael

--- linux-2.4.18/arch/i386/kernel/dmi_scan.c.~1~ Tue Feb 26 13:26:56 2002
+++ linux-2.4.18/arch/i386/kernel/dmi_scan.c Mon Apr 29 22:24:41 2002
@@ -20,8 +20,8 @@
         u16 handle;
 };
 
-#define dmi_printk(x)
-//#define dmi_printk(x) printk x
+//#define dmi_printk(x)
+#define dmi_printk(x) printk x
 
 static char * __init dmi_string(struct dmi_header *dm, u8 s)
 {
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Apr 30 2002 - 22:00:18 EST