Linux kernel development

Comrade DOS | 29 May 2012 22:20

Picon

Re: PROBLEM: Performance drop

> Unfortunately you have used acpi=debug instead of apic=debug.  So I
> can't compare I/O APIC configurations between the different test
> scenarios.

Sorry me for this mistake.

> So what I'd need next is a log (with apic=debug) from
> - booting 3.3.5

I use linux 3.3.7 on this machine now.
Log from 3.3.7:

Initializing cgroup subsys cpuset
Initializing cgroup subsys cpu
Linux version 3.3.7-gentoo (root <at> dos) (gcc version 4.7.0 (Gentoo 4.7.0
p1.0, pie-0.5.3) ) #3 SMP PREEMPT Mon May 28 10:59:42 OMST 2012
Command line: root=/dev/sda3 apic=debug
BIOS-provided physical RAM map:
 BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
 BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
 BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
 BIOS-e820: 0000000000100000 - 00000000b7fb0000 (usable)
 BIOS-e820: 00000000b7fb0000 - 00000000b7fc8000 (reserved)
 BIOS-e820: 00000000b7fc8000 - 00000000b7fe7fb8 (ACPI NVS)
 BIOS-e820: 00000000b7fe7fb8 - 00000000c0000000 (reserved)
 BIOS-e820: 00000000e0000000 - 00000000f0000000 (reserved)
 BIOS-e820: 00000000fec00000 - 00000000fec02000 (reserved)
 BIOS-e820: 00000000ffbc0000 - 00000000ffcc0000 (reserved)
 BIOS-e820: 00000000fff00000 - 0000000100000000 (reserved)
 BIOS-e820: 0000000100000000 - 0000000140000000 (usable)

(Continue reading)

Andreas Herrmann | 30 May 2012 14:43

Re: PROBLEM: Performance drop

On Wed, May 30, 2012 at 03:20:27AM +0700, Comrade DOS wrote:
> > Unfortunately you have used acpi=debug instead of apic=debug.  So I
> > can't compare I/O APIC configurations between the different test
> > scenarios.
> 
> Sorry me for this mistake.

No problem.

The logs show no difference in IO-APIC pin usage.
So it's not the old problem ...

Comparing both logs I found following differences:

(Most other stuff seems just to be changed formatting.)

-ACPI: Thermal Zone [TZ1] (67 C)
+ACPI: Thermal Zone [TZ1] (62 C)

I think what's shown is the temperature value which just differed
between the boots. But that made me look at acpi/thermal.c where the
messages came from. The only change between 3.3 and 3.4 is this
commit:

  commit 9bcb8118965ab4631a65ee0726e6518f75cda6c5
  Author: Matthew Garrett <mjg <at> redhat.com>
  Date:   Wed Feb 1 10:26:54 2012 -0500

    ACPI: Evaluate thermal trip points before reading temperature

(Continue reading)

Jason Vas Dias | 7 Jul 2012 16:40

Picon

Re: PROBLEM: Performance drop

I can confirm that the AMD Turion X2 2.2Ghz  HP Compaq 6715b
"business" x86_64 k8 dual-core laptops circa 2007
DO get stuck in 800Khz mode  and cannot switch out of it after booting
the "stable" "v3.4.4" tagged kernel.

I followed the containing post and reverted commit
ff74ae50f01ee67764564815c023c362c87ce18b :

Commit d51cdad33bb5bb370c05129f7c7f3a16a55eff40
Author: root <root <at> jvdspc.localdomain>
Date:   Fri Jul 6 18:57:03 2012 +0000

    Revert "ACPI: Evaluate thermal trip points before reading temperature"

    This reverts commit 9bcb8118965ab4631a65ee0726e6518f75cda6c5.

commit ff74ae50f01ee67764564815c023c362c87ce18b
Author: Greg Kroah-Hartman <gregkh <at> linuxfoundation.org>
Date:   Fri Jun 22 11:37:50 2012 -0700

And wow !  what a difference - back to a circa 2007 machine versus  a
circa 1987 machine.

Not too many of us  left around trying to run the latest version of
linux on nearly 5-year-old hardware I guess, but still -
please can you restore correct Linux cpufreq & thermal operation on
old-style AMD k8 CPUs ?
They do seem to depend on the temperature being set BEFORE 1st entry .

Thanks & Regards,

(Continue reading)

Jason Vas Dias | 7 Jul 2012 16:45

Picon

Re: PROBLEM: Performance drop

Sorry, of course the commit I backed out was :
9bcb8118965ab4631a65ee0726e6518f75cda6c5.

On Sat, Jul 7, 2012 at9bcb8118965ab4631a65ee0726e6518f75cda6c5. 3:40
PM, Jason Vas Dias <jason.vas.dias <at> gmail.com> wrote:
> I can confirm that the AMD Turion X2 2.2Ghz  HP Compaq 6715b
> "business" x86_64 k8 dual-core laptops circa 2007
> DO get stuck in 800Khz mode  and cannot switch out of it after booting
> the "stable" "v3.4.4" tagged kernel.
>
> I followed the containing post and reverted commit
> ff74ae50f01ee67764564815c023c362c87ce18b :
>
> Commit d51cdad33bb5bb370c05129f7c7f3a16a55eff40
> Author: root <root <at> jvdspc.localdomain>
> Date:   Fri Jul 6 18:57:03 2012 +0000
>
>     Revert "ACPI: Evaluate thermal trip points before reading temperature"
>
>     This reverts commit 9bcb8118965ab4631a65ee0726e6518f75cda6c5.
>
> commit ff74ae50f01ee67764564815c023c362c87ce18b
> Author: Greg Kroah-Hartman <gregkh <at> linuxfoundation.org>
> Date:   Fri Jun 22 11:37:50 2012 -0700
>
> And wow !  what a difference - back to a circa 2007 machine versus  a
> circa 1987 machine.
>
> Not too many of us  left around trying to run the latest version of
> linux on nearly 5-year-old hardware I guess, but still -

(Continue reading)

Greg Kroah-Hartman | 12 Jul 2012 00:06

Re: PROBLEM: Performance drop

On Sat, Jul 07, 2012 at 03:40:58PM +0100, Jason Vas Dias wrote:
> I can confirm that the AMD Turion X2 2.2Ghz  HP Compaq 6715b
> "business" x86_64 k8 dual-core laptops circa 2007
> DO get stuck in 800Khz mode  and cannot switch out of it after booting
> the "stable" "v3.4.4" tagged kernel.
> 
> I followed the containing post and reverted commit
> ff74ae50f01ee67764564815c023c362c87ce18b :
> 
> Commit d51cdad33bb5bb370c05129f7c7f3a16a55eff40
> Author: root <root <at> jvdspc.localdomain>
> Date:   Fri Jul 6 18:57:03 2012 +0000
> 
>     Revert "ACPI: Evaluate thermal trip points before reading temperature"
> 
>     This reverts commit 9bcb8118965ab4631a65ee0726e6518f75cda6c5.
> 
> commit ff74ae50f01ee67764564815c023c362c87ce18b
> Author: Greg Kroah-Hartman <gregkh <at> linuxfoundation.org>
> Date:   Fri Jun 22 11:37:50 2012 -0700
> 
> And wow !  what a difference - back to a circa 2007 machine versus  a
> circa 1987 machine.
> 
> Not too many of us  left around trying to run the latest version of
> linux on nearly 5-year-old hardware I guess, but still -
> please can you restore correct Linux cpufreq & thermal operation on
> old-style AMD k8 CPUs ?
> They do seem to depend on the temperature being set BEFORE 1st entry .
>

(Continue reading)

Jason Vas Dias | 14 Jul 2012 13:11

Picon

Re: PROBLEM: Performance drop

Hi - any progress on this or on the patch I submitted for it ? -
please see enclosed - apologies for my being forced to use gmail which
has mandatory line wrap -
Please do something about restoring correct thermal operation on
x86_64 K8's with HP BIOS !
Thanks & Regards,
Jason

Re: [PATCH: 1/1] ACPI: make evaluation of thermal trip points before
temperature or vice versa dependant on new "temp_b4_trip" module
parameter to support older AMD x86_64s
Kernel
	x
Jason Vas Dias
	
Jul 9 (5 days ago)
		
Reply
to Rusty, linux-kernel, Andreas, Matthew, Len, Comrade
Thanks Rusty - sorry I didn't see your email until now - revised patch
addressing your comments attached -
BTW,  sorry about the word wrap on the initial posting - should I
attach a '.patch' file or inline ?  Trying both .

The Revised Patch (against :
commit bd0a521e88aa7a06ae7aabaed7ae196ed4ad867a
Author: Linus Torvalds <torvalds <at> linux-foundation.org>
Date:   Sat Jul 7 17:23:56 2012 -0700

    Linux 3.5-rc6

(Continue reading)

Greg Kroah-Hartman | 27 Sep 2012 19:34

Re: PROBLEM: Performance drop

On Sat, Jul 14, 2012 at 12:11:41PM +0100, Jason Vas Dias wrote:
> Hi - any progress on this or on the patch I submitted for it ? -
> please see enclosed - apologies for my being forced to use gmail which
> has mandatory line wrap -
> Please do something about restoring correct thermal operation on
> x86_64 K8's with HP BIOS !
> Thanks & Regards,
> Jason
> 
> Re: [PATCH: 1/1] ACPI: make evaluation of thermal trip points before
> temperature or vice versa dependant on new "temp_b4_trip" module
> parameter to support older AMD x86_64s

I don't see this patch in Linus's tree, so I can't do anything about it
in a stable kernel release.  Please work with the ACPI developers to get
it pushed to Linus's tree, and after that, let me know what the git
commit id is of it so I can add it to a stable release.

thanks,

greg k-h

Borislav Petkov | 27 Sep 2012 19:48

Picon

Re: PROBLEM: Performance drop

On Thu, Sep 27, 2012 at 10:34:56AM -0700, Greg Kroah-Hartman wrote:
> On Sat, Jul 14, 2012 at 12:11:41PM +0100, Jason Vas Dias wrote:
> > Hi - any progress on this or on the patch I submitted for it ? -
> > please see enclosed - apologies for my being forced to use gmail which
> > has mandatory line wrap -
> > Please do something about restoring correct thermal operation on
> > x86_64 K8's with HP BIOS !
> > Thanks & Regards,
> > Jason
> > 
> > Re: [PATCH: 1/1] ACPI: make evaluation of thermal trip points before
> > temperature or vice versa dependant on new "temp_b4_trip" module
> > parameter to support older AMD x86_64s
> 
> I don't see this patch in Linus's tree, so I can't do anything about it
> in a stable kernel release.  Please work with the ACPI developers to get
> it pushed to Linus's tree, and after that, let me know what the git
> commit id is of it so I can add it to a stable release.

Yeah,

this patch adds a dmi quirk *and* a module parameter which looks kinda
too much to me. In any case, Andreas needs to look at this since he's
been dealing with those old laptops.

Jason, can you clean up your patch, drop the module parameter, write a
proper commit message, run it through checkpatch.pl and send it as a
reply to this thread with a mail client which doesn't mangle it so that
we can look at it further?

(Continue reading)

katabami | 10 Oct 2012 00:19

Re: PROBLEM: Performance drop

Hi, just for recording. Bug 43284[1] is the duplicate of this bug report. It
stops by reporting the "bad" commit, 9bcb8118965ab4631a65ee0726e6518f75cda6c5.

[1]https://bugzilla.kernel.org/show_bug.cgi?id=43284

katabami | 18 Oct 2012 03:21

Re: PROBLEM: Performance drop

Sent a patch to the bugzilla,[1] following Boris' and Rusty Russell's advice.

[1]
https://bugzilla.kernel.org/show_bug.cgi?id=43284

Thanks a lot.

Return

Return to gmane.linux.kernel.

Search Archive

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane