On 5/10/2011 8:54 AM, Jeff Rickman wrote:
On 5/9/2011 2:42 PM, Jean Delvare wrote:
Hi Jeff,
On Sun, 08 May 2011 22:32:16 -0500, Jeff Rickman wrote:
What Harry's comments tell me is the original "i8k" code lacks
appropriate code to operate in a SMP environment. I think the Console
messages are trying to say that for "non-programmers" like myself.
Note these differences:
- Harry has a single CPU; I have 2 CPU.
This is a valuable observation. I presume that these CPUs are too old
to be multicore, best they could have it hyperthreading. Harry, can you
please share the contents of /proc/cpuinfo?
Jeff, could you please try booting with parameter max_cpus=1? I wonder
if it would make a difference. If it does, that would confirm your
theory.
- We both run the same Dell BIOS A07 code.
- Harry's Linux kernel is slightly older than my 2.6.35.12-90 Fedora
code.
- I think we would all agree that Jean's changes do not cause my crash
since that code works on Harry's machine but not on mine.
Time permitting, I will start doing the research to submit a properly
documented bug against Fedora Core 14 regarding this code. I suspect
this issue will also reside in FC15 so that is another point to check if
time allows. jean raises the real question: does anyone in the Fedora
team want to fix this?
I've seen this nug hit an openSUSE user (on a laptop with one
multi-core CPU), so it's definitely not specific to Fedora. It may be
related to the version of gcc used by the distribution though. If
Fedora people aren't interested, we may open a bug on
bugzilla.kernel.org instead.
In closing, I agree with Harry: it is very nice to have some hardware
monitoring in a Dell platform.
I'm glad you like it :)
I checked for any packages that could manipulate the CPU frequency. I
only found "cpuspeed". I removed the "cpuspeed" package and rebooted to
ensure a known state. Attempting to install the "stock i8k" module
causes a crash. Reboot to return to known state.
I changed the "maxcpus=4" in "grub.conf" to "maxcpus=1". Reboot.
Attempt to load "stock i8k" module. Crash. "/proc/cpuinfo" shows 1 CPU.
Return system to previous configuration with "cpuspeed" and reboot.
My only other choice would be to physically pull CPU2, leaving CPU1 in
the system, but we know that config should work based on Harry's testing.
Forgot to add....
The crashdump info looked the same in all cases. The Console messages
looked the same in all cases...complaint about SMP.
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
