longhaul: Unknown VIA CPU

Hi,

I tried to get the longhaul module to work on my VIA EPIA M10000 motherboard (C3-2 cpu), but I did not succeed yet.

I use cpufreq 2.4.21-2, and I see the following message in the syslog when I "modprobe longhaul":

longhaul: Unknown VIA CPU. Contact [email protected]

I'd be pleased to help with what seems to be a "new" cpu, but I do not know what to begin with.

Here is my /proc/cpuinfo; don't know if this is of any use to anyone (:

---
processor : 0
vendor_id : CentaurHauls
cpu family : 6
model : 9
model name : VIA Nehemiah
stepping : 1
cpu MHz : 999.531
cache size : 64 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu de tsc msr mtrr pge cmov mmx fxsr sse
bogomips : 1992.29
---

Regards,

Vincent Stehlé

Nehemiah is still unsupported as I've not seen a production one thats actually capable of working properly yet. Still waiting to hear back from the folks at Centaur.

Dave

 Dave Jones     http://www.codemonkey.org.uk

sorry, I thought I replied..

Already fixed in 2.6, not sure about the 2.4/2.4-ac backports.

I have the relevant documentation, and a Nehemiah that should support power scaling, however when I try to use it according to how the spec says I should I get nonsense returned from certain MSRs, and no speed change. I'm still waiting to hear back from Centaur.

I pinged them again yesterday, hopefully I can get an answer on this pretty soon.

Dave

 Dave Jones     http://www.codemonkey.org.uk

ducrot@neptune:~$ cvs rdiff -u -r LINUX_2_4 -rLINUX_2_5 cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c
Index: cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c
diff -u cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c:1.1.1.1 cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c:1.93
--- cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c:1.1.1.1   Mon Aug 25 15:58:46 2003
+++ cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c   Tue Aug 19 14:38:32 2003
@@ -597,6 +597,7 @@
        .target         = longhaul_target,
        .init           = longhaul_cpu_init,
        .name           = "longhaul",
+       .owner          = THIS_MODULE,
 };

static int __init longhaul_init (void)

ducrot@neptune:~$ cvs rdiff -u -r LINUX_2_4 -rLINUX_2_5 cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.h

Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.

I'm not sure if I pushed the changes back to CVS after merging to 2.5 mainline.

Dave

 Dave Jones     http://www.codemonkey.org.uk

ducrot@neptune:~$ diff -u ~/tmp/cpufreq/25/cpufreq/linux/arch/i386/kernel/cpufreq/longhaul.c \
	/var/share/cvs/bitkeeper/cpufreq-2.5/arch/i386/kernel/cpu/cpufreq/longhaul.c
@@ -584,7 +584,7 @@
        if (ret != 0)
                return ret;

- policy->governor = CPUFREQ_DEFAULT_GOVERNOR;

+       policy->policy = CPUFREQ_POLICY_PERFORMANCE;
        policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL;
 
        policy->cur = (unsigned int) (longhaul_get_cpu_fsb() * longhaul_get_cpu_mult() * 100);

(cvs don't have some core update).

Cheers,

Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.

I sent DaveJones a patch that at least says that the Nehemiah is not implemented, but i have not heard from him since.. also, the email address will need to be changed to the codemonkey one :)

However, VIA have not yet made that info available to the public (I'm hoping Dave has a copy)

sven