Skip to contentSkip to conversation

cpu-freq causing kernel lock-up after upgrading to 1GB ram.

DISCUSSION chris stones 2 participants 2 messages May 15, 2005

lore
Share & tools
lore
#0 lore raw

This is my first attempt at a bug report, so forgive me for making a mess, I will try my best.

-------------------------
hardware:
-------------------------
mobo: Chaintech VNF3-250
chipset: nForce-3
CPU: amd64 Athlon 3400+
Ram: 2 * Kingston 512Meg DDR400 PC3200 2T timeing 3-3-8-3. (total 1GB) Swap : 1.3 Gb.

--------------------------
software:
--------------------------
Distro: Gentoo 2005.0 (comiped as 64bit with multilib)
GCC: gcc version 3.4.3 20041125 (Gentoo Linux 3.4.3-r1, ssp-3.4.3-0, pie-8.7.7) kernel versions: gentoo 2.6.11-r7; vanilla-2.6.10; vanilla-2.6.11.9

--------------------------
description:
--------------------------
I have been using the modules freq_table, cpufreq_ondemand, powernow_k8 to regulate my cpu clock for over 6 months with the hardware configureation shown above, minus one 512meg ram module.

Recently, i shutdown the box, and added the second memory module.
At the time, the cpu_fre drivers were compiled into the kernel, shortly after booting, the system would freese when cpu_freq lowered my cpu to its idle state of 1Ghz.

These conditions must all be met to cause the lock-up.
1) Both Ram moodules are used (1GB ram)
2) the modules freq_table, cpufreq_ondemand, powernow_k8 are loaded.
3) The lockup occurs when the cpu is slowing down from a period of heavy load.

2 hours running memtest86 shows that my memory is not the problem.

one one occasion, the kernel gave a memory dump before the freeze. i regret i didnt take note of its rull contents, but a function trace line mentioned a cpu_freq function call.

If i remove one of the 512meg memory modules, the cpu_freq performsperfectly, as it has for the past 6 months.

Thankyou for taking the time to read my post.
Is there anything i can do (other than go without ram or cpu_freq) to solve this ? is there any more information you require from me ?

I have googled, but found no other similar bug reports.

From [email protected].

#1 lore raw

Same problem here - VNF3-250 - if there's one RAM module C'n'Q works fine, and if I put two modules it freezes every time I try to change frequency, no matter linux or windows. So, it's not cpufreq related problem - it just buggy VNF3-250 C'n'Q inplementation.

Quoted 61 lines · chris stones

This is my first attempt at a bug report, so forgive me for making a mess, I will try my best.

-------------------------
hardware:
-------------------------
mobo: Chaintech VNF3-250
chipset: nForce-3
CPU: amd64 Athlon 3400+
Ram: 2 * Kingston 512Meg DDR400 PC3200 2T timeing 3-3-8-3. (total 1GB) Swap : 1.3 Gb.

--------------------------
software:
--------------------------
Distro: Gentoo 2005.0 (comiped as 64bit with multilib)
GCC: gcc version 3.4.3 20041125 (Gentoo Linux 3.4.3-r1, ssp-3.4.3-0, pie-8.7.7)
kernel versions: gentoo 2.6.11-r7; vanilla-2.6.10; vanilla-2.6.11.9

--------------------------
description:
--------------------------
I have been using the modules freq_table, cpufreq_ondemand, powernow_k8 to regulate my cpu clock for over 6 months with the hardware configureation shown above, minus one 512meg ram module.

Recently, i shutdown the box, and added the second memory module. At the time, the cpu_fre drivers were compiled into the kernel, shortly after booting, the system would freese when cpu_freq lowered my cpu to its idle state of 1Ghz.

These conditions must all be met to cause the lock-up.
1) Both Ram moodules are used (1GB ram)
2) the modules freq_table, cpufreq_ondemand, powernow_k8 are loaded.
3) The lockup occurs when the cpu is slowing down from a period of heavy load.

2 hours running memtest86 shows that my memory is not the problem.

one one occasion, the kernel gave a memory dump before the freeze. i regret i didnt take note of its rull contents, but a function trace line mentioned a cpu_freq function call.

If i remove one of the 512meg memory modules, the cpu_freq performsperfectly, as it has for the past 6 months.

Thankyou for taking the time to read my post.
Is there anything i can do (other than go without ram or cpu_freq) to solve this ?
is there any more information you require from me ?

I have googled, but found no other similar bug reports.

From [email protected].

_______________________________________________
Cpufreq mailing list
[email protected]
http://lists.linux.org.uk/mailman/listinfo/cpufreq