Skip to contentSkip to conversation

New: still problems with k7 mobile frequency is lower than real

Bug 8720 [email protected] 1 participant 26 messages Jul 6, 2007 – Sep 30, 2007

lore
Share & tools
lore
#0 lore raw

http://bugzilla.kernel.org/show_bug.cgi?id=8720

Summary: still problems with k7 mobile frequency is lower than
real
Product: Power Management
Version: 2.5
KernelVersion: 2.6.22-rc5
Platform: All
OS/Version: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: cpufreq
AssignedTo: [email protected]
ReportedBy: [email protected]
CC: [email protected]

Most recent kernel where this bug did not occur: don't know
Distribution: Kubuntu 7.04 and 7.10 Tribe2
Hardware Environment: Athlon XP-M 2400+
Problem Description: This cpu has 1800 Mhz and is detected well during early startup but after by powernowdas 1473 Mhz, the last fix in bug #8255 didn't solve the problem for me, before and after that patch is the same.

I tried also Kubuntu 7.10 Tribe 2 that has a 2.6.22rc5 custom according to
https://lists.ubuntu.com/archives/gutsy-changes/2007-June/003601.html

I personally verified that has the patch posted in bug #8255 and it has, but same bug.

from dmesg attacched:

[    0.000000] Detected 1807.794 MHz processor.
[  118.852989] powernow: PowerNOW! Technology present. Can scale: frequency and
voltage.
[  118.862515] powernow: Trying ACPI perflib
[  118.864625] powernow: Minimum speed 1473 MHz. Maximum speed 1473 MHz.

The processor has the Barton core with 512 KB of cache

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#1 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #1 from [email protected] 2007-07-06 10:24 ------- Created an attachment (id=11968)
--> (http://bugzilla.kernel.org/attachment.cgi?id=11968&action=view)
dmesg from kubuntu 7.10 tribe1

this was with a previous kernel I think 2.6.22rc3, but the bug is the same with kubuntu 7.10 tribe2 that has 2.6.22rc5 and same is with kernel 2.6.20 from kubuntu 7.04

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#2 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #2 from [email protected] 2007-07-06 17:01 ------- Daniel if you put into a simple binary some debugging code I can run into the notebook and report here result.

Probably the best thing to understand the problem.

thanx
Marco

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#3 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #3 from [email protected] 2007-07-09 10:55 ------- seems all frequencies detected are the same one!

ale@portatile:~$ cat
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies
1472966 1472966 1472966 1472966 1472966 1472966
Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#4 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #4 from [email protected] 2007-07-22 17:13 ------- Created an attachment (id=12097)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12097&action=view)
dmesg from kubuntu 7.10 tribe3 + XP-M 2800+

this the dmesg with the same identical notebook, kubuntu 7.10 tribe3 but another Athlon XP Mobile, this time it's a 2800+ that should have 2,14 ghz instead it's always recognized as: 1472-1473 Mhz

[    0.000000] Detected 2142.556 MHz processor.
[  239.598408] powernow: PowerNOW! Technology present. Can scale: frequency and
voltage.
[  239.653057] powernow: Trying ACPI perflib
[  239.669325] powernow: Minimum speed 1472 MHz. Maximum speed 1472 MHz.
Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#5 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #5 from [email protected] 2007-07-22 17:14 ------- Created an attachment (id=12098)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12098&action=view)
scaling_available_frequencies from kubuntu 7.10 tribe3 + XP-M 2800+

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#6 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

[email protected] changed:

What |Removed |Added
----------------------------------------------------------------------------

      KernelVersion|2.6.22-rc5                  |2.6.22

------- Comment #6 from [email protected] 2007-07-28 07:45 ------- Tried a third cpu with the same notebook and this seems well recognized.

The cpu is an Athlon XP-M 2500+ at 1874 Mhz so probably there are still some problems with cpu scaling frequencies calculation.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#7 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #7 from [email protected] 2007-07-28 07:48 ------- Created an attachment (id=12188)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12188&action=view)
dmesg from kubuntu 7.10 tribe3 + XP-M 2500+

this cpu is correctly recognized by linux

[    0.000000] Detected 1874.742 MHz processor.
[  217.438431] powernow: PowerNOW! Technology present. Can scale: frequency and
voltage.
[  217.490428] powernow: SGTC: 13333
[  217.490431] powernow: Minimum speed 535 MHz. Maximum speed 1874 MHz.
Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#8 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #8 from [email protected] 2007-07-28 07:49 ------- Created an attachment (id=12189)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12189&action=view)
scaling_available_frequencies from kubuntu 7.10 tribe3 + XP-M 2500+

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#9 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #9 from [email protected] 2007-07-29 19:55 ------- Please boot an affected system with the "cpufreq.debug=3" parameter and attach full dmesg output here.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#10 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #10 from [email protected] 2007-07-30 11:27 ------- Daniel thanx so much for your answer.

There are other tests that I can do? Because the pc affected is not mine, so when I go to my friend I do as many tests ad I can.

thanx

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#11 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #11 from [email protected] 2007-07-30 11:59 ------- Theres nothing I can do without that info, sorry. You may also wish to dump the ACPI DSDT as detailed in the other bug as I suspect it will be needed later.

I think you also need to compile cpufreq debug support into your kernel for that debug parameter to have an effect.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#12 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #12 from [email protected] 2007-07-30 15:52 ------- this can be a problem.
Is there a live cd with a debugging kernel that I can just launch? All that I have is ubuntu tribe 3 for now.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#13 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #13 from [email protected] 2007-08-01 12:54 ------- well probably ubuntu 7.10 tribe3 is not built with debug level that I need

[    0.000000] Unknown boot option `cpufreq.debug=3': ignoring

but I have the ACPI dump that I obtained with acpidump utility, can be helpful?

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#14 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #14 from [email protected] 2007-08-01 12:56 ------- Created an attachment (id=12229)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12229&action=view)
acpidump with ubuntu 7.10 tribe3 and XP-M 2800+

obtained with:

sudo acpidump

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#15 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #15 from [email protected] 2007-09-01 05:41 ------- Created an attachment (id=12658)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12658&action=view)
dmesg with athlon xp-m 2400+ and kernel 2.6.22.5

this is a self-compiled 2.6.22.5 kernel with cpu freq debug support

dmesg attached and kernel started with cpufreq.debug=3 parameter under athlon-m
2400+ where the bug occurs

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#16 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #16 from [email protected] 2007-09-01 05:41 ------- Created an attachment (id=12659)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12659&action=view)
acpidump with athlon xp-m 2400+ and kernel 2.6.22.5

this is a self-compiled 2.6.22.5 kernel with cpu freq debug support

acpidump attached and kernel started with cpufreq.debug=3 parameter under athlon-m 2400+ where the bug occurs

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#17 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #17 from [email protected] 2007-09-01 05:43 ------- Created an attachment (id=12660)
--> (http://bugzilla.kernel.org/attachment.cgi?id=12660&action=view)
dmesg with athlon xp-m 2500+ and kernel 2.6.22.5

this is a self-compiled 2.6.22.5 kernel with cpu freq debug support

dmesg attached and kernel started with cpufreq.debug=3 parameter under athlon-m
2500+ where the bug DOESN'T occur

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#18 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #18 from [email protected] 2007-09-01 05:46 ------- sorry for so many attachments, but I think last 3 files are what count.

SUMMARY with kernel 2.6.22.5:

Athlon XP-M 2400+ BUGGY
Athlon XP-M 2500+ NON BUGGY
Athlon XP-M 2800+ BUGGY

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#19 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

[email protected] changed:

What |Removed |Added
----------------------------------------------------------------------------

  Attachment #12659|application/octet-stream    |text/plain
          mime type|                            |
Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#20 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

[email protected] changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |[email protected]

------- Comment #19 from [email protected] 2007-09-01 08:21 ------- Looking at dmesg in comment #15 and corresponding DSDT in comment #16:

powernow tables in this system are unusable so it falls back on ACPI. The powernow_acpi_control_t values in the control field of the _PSS entries always result in FID=0 VID=0 which looks wrong.

As there is _PCT, acpi-cpufreq may be a better cpufreq driver choice for this system. However the ACPI-supplied cpu frequencies in _PSS are wrong.

The DSL code for _PSS reads FID and VID values from system memory (region FSEG defined at 0x000FDF00 in system memory), and uses them as indexes into a local PS13 array which presumably contains a series of frequencies. However, again the FID values are always 0 so it always ends up looking at the first element, 0x5b7 (1463).

I'm not 100% sure if 0x000FDF00 is controlled by Linux or otherwise, however I am fairly sure the answer here is that your BIOS is double-broken. Dave, any comments?

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#21 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #20 from [email protected] 2007-09-06 07:33 ------- So it's a bios bug?

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#22 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #21 from [email protected] 2007-09-06 07:46 ------- I'm not certain, but yes I think so.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#23 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #22 from [email protected] 2007-09-23 02:35 ------- Daniel only a question: isn't possible a small hack when you have this kind of buggy bios then put the min and max frequency to the detected frequency that seems ok?

For example comment #7:

[    0.000000] Detected 1874.742 MHz processor.

If not this bug can be closed I think.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#24 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

------- Comment #23 from [email protected] 2007-09-23 04:19 ------- No, as we don't know what data to program into registers to ask the CPU to go to that frequency.

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
#25 lore raw

[Bug 8720] still problems with k7 mobile frequency is lower than real

In reply to [email protected]

http://bugzilla.kernel.org/show_bug.cgi?id=8720

[email protected] changed:

What |Removed |Added
----------------------------------------------------------------------------

             Status|NEW                         |REJECTED
         Resolution|                            |INVALID

------- Comment #24 from [email protected] 2007-09-30 08:07 ------- Ok tried the same cpu in another motherboard and it's correctly recognized, so I close this bug as "my bios is fucking buggy" :(

thanx Daniel for your time!

Signature · 3 lines
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.