no speedstep for PIII coppermine

Bug #10813 reported by Jean Privat
22
This bug affects 1 person
Affects Status Importance Assigned to Milestone
powernowd (Ubuntu)
Invalid
Medium
Paul Sladen

Bug Description

http://www.bay-wolf.com/speedstep.htm says the latitude C800 with PIII has
SpeedStep.
But the file /proc/cpufrec and the directory /sys/devices/system/cpu/cpu0/ are
empty and the cpufreq-detect.sh script from
http://www.ubuntulinux.org/wiki/CPUFreqModuleLoading only says "This processor
does not support any power-saving reductions."

$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 6
cpu MHz : 848.323
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat
pse36 mmx fxsr sse
bogomips : 1679.36

Revision history for this message
Paul Sladen (sladen) wrote :

Bonjour Jean!

Thanks for your report, could you grab a new version of the script and see what
it recommends now.

Depending on your chipset is and what it suggests, could you try:

  sudo modprobe speedstep-something.o coppermine=1

and let me know if that does/doesn't work. I think your chip should be able to
switch between 850Mhz and either 500Mhz or 400Mhz depending on which CPU
revision it is.

Many Thanks,

Paul

Revision history for this message
Jean Privat (privat) wrote :

$ bash cpufreq-detect.sh
speedstep-smi.o coppermine=1
$ sudo modprobe speedstep-smi.o coppermine=1
FATAL: Module speedstep_smi.o not found.
--> Bash completion says me the .o is not needed (and I have heard of .ko for
2.6 kernels)
$ sudo modprobe speedstep-smi coppermine=1
--> The screen became blank durring 2 seconds and the command response was
FATAL: Error inserting speedstep_smi (/lib/modules/2.6.9-1-386/kernel/arch/i386/
kernel/cpu/cpufreq/speedstep-smi.ko): No such device
--> Here what was append to dmegs :
speedstep_smi: Unknown parameter `coppermine'
cpufreq: change failed with new_state 2 and result 0
cpufreq: change failed with new_state 2 and result 0

--> Ok, try an other (I know it is not a centrino but why not)
$ sudo modprobe speedstep-centrino coppermine=1
FATAL: Error inserting speedstep_centrino
(/lib/modules/2.6.9-1-386/kernel/arch/i386/kernel/cpu/cpufreq/speedstep-centrino.ko):
No such device
--> New in dmesg :
speedstep_centrino: Unknown parameter `coppermine'

--> Try the last one (speedstep_lib was already loaded by previous module tries)
$ sudo modprobe speedstep-ich coppermine=1
FATAL: Error inserting speedstep_ich
(/lib/modules/2.6.9-1-386/kernel/arch/i386/kernel/cpu/cpufreq/speedstep-ich.ko):
No such device
--> And in dmesg :
speedstep_ich: Unknown parameter `coppermine'
cpufreq: Intel(R) SpeedStep(TM) for this chipset not (yet) available.
--> To be honest, I don't like the last line

Revision history for this message
Chris Ladd (caladd) wrote :

My Sony Vaio PCG-SR17 also has a PIII Coppermine that fails to install the
speedstep-smi module under the Hoary Hedgehog release. Here is my /proc/cpuinfo:

 processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 6
cpu MHz : 699.637
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat
pse36 mmx fxsr sse
bogomips : 1388.54

Doing a modprobe speedstep-smi fails with:

FATAL: Error inserting speedstep_smi
(/lib/modules/2.6.10-5-686/kernel/arch/i386/kernel/cpu/cpufreq/speedstep-smi.ko):
No such device

And the dmesg gets:

cpufreq: change failed with new_state 2 and result 0
cpufreq: change failed with new_state 2 and result 0

Revision history for this message
l33tmonkey (l33tmonkey) wrote :

http://thinkwiki.org/wiki/How_to_get_SpeedStep_working_on_Coppermine-piix4-smi_based_Thinkpads

I think this link may be helpful.
It appears that these processors do not report as speedstep capable.

Revision history for this message
Jakob Malm (malmjakob) wrote :

I added the following to /etc/modules, as suggested on a site somewhere:

speedstep-lib
speedstep-smi smi_port=0xb2 smi_cmd=0x82 smi_sig=1

Speedstep is now working fine for me on Hoary. I have a Pentium III 800 Mhz (650
Mhz in low power mode) with
Host bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX Host bridge (rev 03).

Revision history for this message
Paul Sladen (sladen) wrote :

This needs to be special-cased for this particular laptop.

Windows includes a massive overrides/bug fixes file---about 75% which are just
overrides to use the 'speedstep-smi' fallback on broken machines and/or on alter
smi ports.

Needs using 'dmi-decode' to figure out the laptop.

Revision history for this message
Nico (nico-rdo) wrote :

I have a Dell Lattitude C600 with a P III Coppermine 1 GHz.

Powernowd determines the CPU and loads the speedstep-smi module.

From then on I have a 1 second system freeze everytime the cpu frequency is
changed (correlated using the Gnome Applet).

In fact, the Dell has a weird BIOS, which takes care of frequency changes
itself, thus the clash with the kernel Speedstep.

Without Powernowd, I can still witness frequency changes between 20 and 1002 MHz
using the x86info module from gkrellm.

Here are the outputs that powernowd uses on my machine in order to determine the
CPU type. It may be useful in order to change the
/usr/share/powernowd/cpufreq-detect.sh script so that it detects machines like
mine properly.

/proc/cpuinfo
=============

processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 10
cpu MHz : 1002.511
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat
pse36 mmx fxsr sse
bogomips : 1990.65

/proc/ioports
=============

0000-001f : dma1
0020-0021 : pic1
0022-0022 : PM2_CNT_BLK
0040-0043 : timer0
0050-0053 : timer1
0060-006f : keyboard
0070-0077 : rtc
0080-008f : dma page reg
00a0-00a1 : pic2
00c0-00df : dma2
00f0-00ff : fpu
0170-0177 : ide1
01f0-01f7 : ide0
02f8-02ff : serial
0376-0376 : ide1
0378-037a : parport0
03c0-03df : vesafb
03f0-03f1 : pnp 00:08
03f6-03f6 : ide0
03f8-03ff : serial
04d0-04d1 : pnp 00:01
0600-067f : pnp 00:02
0680-06ff : pnp 00:02
0778-077a : parport0
0800-083f : 0000:00:07.3
  0800-0803 : PM1a_EVT_BLK
  0806-0807 : pnp 00:02
  0808-080b : PM_TMR
  080c-080f : GPE0_BLK
  0810-083f : pnp 00:02
    0810-0815 : ACPI CPU throttle
0840-085f : 0000:00:07.3
  0840-084f : pnp 00:02
    0840-0847 : piix4-smbus
  0850-0853 : pnp 00:02
  0854-0855 : PM1a_CNT_BLK
  0856-085f : pnp 00:02
0860-086f : 0000:00:07.1
  0860-0867 : ide0
  0868-086f : ide1
0cf8-0cff : PCI conf1
4000-40ff : PCI CardBus #02
4400-44ff : PCI CardBus #02
4800-48ff : PCI CardBus #06
4c00-4cff : PCI CardBus #06
d000-d0ff : 0000:00:10.1
d400-d4ff : 0000:00:10.0
  d400-d47f : 0000:00:10.0
d800-d8ff : 0000:00:08.0
  d800-d8ff : Maestro3
dce0-dcff : 0000:00:07.2
  dce0-dcff : uhci_hcd
e000-efff : PCI Bus #01
  ec00-ecff : 0000:01:00.0
f000-f0fe : motherboard
  f000-f0fe : pnp 00:03
f100-f1fe : motherboard
  f100-f1fe : pnp 00:03
f200-f2fe : motherboard
  f200-f2fe : pnp 00:03
f400-f4fe : motherboard
  f400-f4fe : pnp 00:03
f500-f5fe : motherboard
  f500-f5fe : pnp 00:03
f600-f6fe : motherboard
  f600-f6fe : pnp 00:03
f800-f8fe : motherboard
  f800-f8fe : pnp 00:03
f900-f9fe : motherboard
  f900-f9fe : pnp 00:03
fa00-fafe : motherboard
fc00-fcfe : motherboard
fd00-fdfe : motherboard
fe00-fefe : motherboard

Revision history for this message
agd5f (agd5f) wrote :

I'm having the same problem on my T20 Thinkpad. speedstep-smi fails the same regardless of the options and whether or not I pass relaxed_check=1 to speedstep-lib

modprobe speedstep-smi smi_port=0xb2 smi_cmd=0x82 smi_sig=1
FATAL: Error inserting speedstep_smi (/lib/modules/2.6.15-17-386/kernel/arch/i386/kernel/cpu/cpufreq/speedstep-smi.ko): No such device

processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 6
cpu MHz : 697.126
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr sse
bogomips : 1394.74

Any ideas?

Revision history for this message
agd5f (agd5f) wrote :

Dapper BTW.

Revision history for this message
Paul Sladen (sladen) wrote :

Hi everyone, I think I'd like to close this bug. It has been around for a very long time and has now collected several different bug-reports.

Please can you test against the lastest Dapper. If anyone still has an issue below, please can you file them again in a new bug, specifically mentioning the model of your laptop in the title of the bug report and then we can tackle each one separately. Please attach they DMI decode data for your CPU, as found on the bottom of:

  https://wiki.ubuntu.com/LaptopTestingTeam/HotkeyResearch

Jakob: we should be able to do a specific over-ride for your laptop (please include the dmidecode data so that it can be identified).

Chris: Speedstep may not work on your laptop, if you can google and find anyone else with your model who *has* found a solution, please post that so that it can be implemented.

Jean: do you still have problems running under Dapper?

Nico: we may just have to blacklist scaling for your machine.

Changed in powernowd:
status: Unconfirmed → Rejected
Revision history for this message
Paul Sladen (sladen) wrote :

Jakob Malm, Nico, agd5f; Please see the last comment in this bug report and resubmit it under a new bug specific to your laptop model if you still are having problems.

Revision history for this message
Nico (nico-rdo) wrote :

Paul,

Until a couple of upgrades ago, the powernowd daemon still initiated those 1s freezes regularly, and I had to stop it manually.

Now it is all fine and CPU throttling is still working, either form the daemon, or from the Dell BIOS (probably this, if you blacklisted my chip), but without the 1s freezes.

All of that in Dapper.

Thanks for this !

Nico

Revision history for this message
agd5f (agd5f) wrote :
Revision history for this message
Jean Privat (privat) wrote :
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.