[phpBB Debug] PHP Warning: in file [ROOT]/phpbb/session.php on line 583: sizeof(): Parameter must be an array or an object that implements Countable
[phpBB Debug] PHP Warning: in file [ROOT]/phpbb/session.php on line 639: sizeof(): Parameter must be an array or an object that implements Countable
[phpBB Debug] PHP Warning: in file [ROOT]/includes/functions.php on line 4516: Cannot modify header information - headers already sent by (output started at [ROOT]/includes/functions.php:3262)
[phpBB Debug] PHP Warning: in file [ROOT]/includes/functions.php on line 4516: Cannot modify header information - headers already sent by (output started at [ROOT]/includes/functions.php:3262)
[phpBB Debug] PHP Warning: in file [ROOT]/includes/functions.php on line 4516: Cannot modify header information - headers already sent by (output started at [ROOT]/includes/functions.php:3262)
linux-PHC • Patch for 2.6.34 appears to be broken
Page 1 of 2

Patch for 2.6.34 appears to be broken

Posted: Mon 17. May 2010, 18:43
by The Vault Dweller
The problem that I have described in this post http://www.linux-phc.org/forum/viewtopic.php?f=8&t=180 still persists with 2.6.34 and phc-intel-0.3.2-11-offtree.

Attempt of executing this command

Code: Select all

echo 28 28 27 23 0 > /sys/devices/system/cpu/cpu0/cpufreq/phc_vids
fails with

Code: Select all

echo: write error: no such device
Patch for 2.6.33 works fine.

HW configuration: Pentium-M Dothan.

How do I debug this failure?

Re: Patch for 2.6.34 appears to be broken

Posted: Sat 22. May 2010, 08:20
by the-fallen
I apologize for the late answear.

The message "no such device" normally appears when the cpufreq driver decides (by whatever) that your CPU is not capable of setting control data (vids, fids, etc) through ACPI.

It is strange that this happen to you only with the new kernel. Within the PHC-patch we only check the Flag if the CPU is capable of setting data. The decision itself falls somewhere in the ACPI subsystem.

There is nothing I an do for you now but maybe this will help you to find the problem.

Re: Patch for 2.6.34 appears to be broken

Posted: Sun 23. May 2010, 13:46
by mullet
Hi,

I'm experiencing the exact same problem. Wih 2.6.33 it works just fine but the -11 patch for 2.6.34 shows that error message.

My CPU is also a Pentium M Dothan (IBM Thinkpad X41).

Re: Patch for 2.6.34 appears to be broken

Posted: Fri 28. May 2010, 12:42
by FieserKiller
Same problem here
CPU: Intel(R) Pentium(R) M processor 2.00GHz stepping 08

I used phc-intel-0.3.2-10 with kernel 2.6.33 without any problems, now with phc-intel-0.3.2-11 and kernel 2.6.34 I get that "no such device" error on reading and writing

EDIT:
could the new kernel option
Processor Clocking Control interface driver (X86_PCC_CPUFREQ)
cause this error?
I'll try to check that asap

EDIT2:
Nope, wrong idea

Re: Patch for 2.6.34 appears to be broken

Posted: Fri 28. May 2010, 14:52
by mullet
toggling X86_PCC_CPUFREQ doesn't work for me

Re: Patch for 2.6.34 appears to be broken

Posted: Sat 29. May 2010, 18:53
by chrisF
I did some debugging on my Samsung P35 with Dothan CPU and found out the following:
For proper work, linux-phc checks a flag "SYSTEM_INTEL_MSR_CAPABLE", set somewhere in the ACPI sub-system for the detected CPU.
Kernel 2.6.34 does not set the "SYSTEM_INTEL_MSR_CAPABLE" any more, it sets the "SYSTEM_IO_CAPABLE" instead.
I have no idea, why the linux-phc does not see the MSR capability of the CPU. On the command line, the "msr-tools" can work on the MSRs, CPU voltage can be read and set.

ChrisF

Re: Patch for 2.6.34 appears to be broken

Posted: Thu 17. Jun 2010, 22:29
by edishtron
the patch which can be found at the following url should fix the problem.

http://marc.info/?l=linux-acpi&m=127678734705556&w=2

cheers
edishtron

Re: Patch for 2.6.34 appears to be broken

Posted: Fri 18. Jun 2010, 16:27
by the-fallen
PHC can not set values if the processor driver decides to be "SYSTEM_IO_CAPABLE" instead of "SYSTEM_MSR_CAPABLE".
It is telling the system how to control the CPU.

Reading/Writing to the MSR's still works because it does not depends on that capability.



The patch edishtron linked above should fix that problem.
Please report if that is now working for you so I can sleep well :) Thanks.

Re: Patch for 2.6.34 appears to be broken

Posted: Sun 27. Jun 2010, 17:28
by chrisF
Hi, the patch edishtron mentioned, works fine!
The voltages can be adjusted on my Pentium M Dothan with the patched kernel 2.6.34.

Regards,
ChrisF

Re: Patch for 2.6.34 appears to be broken

Posted: Tue 29. Jun 2010, 19:12
by the-fallen
Great news, and thanks for your reply.