RE: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

RE: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

[Brown, Len]

Eric,
Do you hear this noise for both AC and battery modes?

Do you still get this noise when you run with HZ=100
or HZ=1000?

Are there any T43 users who do not hear this noise?

Turning off C3 is a pretty big hammer...

thanks,
-Len


>-----Original Message-----
>From: Eric Shattow [mailto:shadow@serverart.org] 
>Sent: Thursday, September 21, 2006 1:37 PM
>To: Brown, Len
>Subject: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy 
>thinkpad t43/p models
>
>IBM T43/p laptops exhibit a high pitched noise at higher ACPI 
>C-states. Making
>C3 the max state reduces this noise greatly.
>
>diff -uprN linux-2.6.18/drivers/acpi/processor_idle.c
>linux-2.6.18~ibm_t43p_limit_max_cstate/drivers/acpi/processor_idle.c
>--- linux-2.6.18/drivers/acpi/processor_idle.c    2006-09-19
>22:42:06.000000000 -0500
>+++ 
>linux-2.6.18~ibm_t43p_limit_max_cstate/drivers/acpi/processor_idle.c   
>2006-09-21 12:20:59.000000000 -0500
>@@ -78,6 +78,9 @@ module_param(bm_history, uint, 0644);
>  * IBM ThinkPad R40e crashes mysteriously when going into C2 or C3.
>  * For now disable this. Probably a bug somewhere else.
>  *
>+ * IBM ThinkPad T43/p exhibits strange CPU noises at higher wait
>+ * states. Forcing a wait state of C3 or less keeps things quiet.
>+ *
>  * To skip this limit, boot/load with a large max_cstate limit.
>  */
> static int set_max_cstate(struct dmi_system_id *id)
>@@ -145,6 +148,48 @@ static struct dmi_system_id __cpuinitdat
>     { set_max_cstate, "IBM ThinkPad R40e", {
>       DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>       DMI_MATCH(DMI_BIOS_VERSION,"1SET68WW") }, (void*)1},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET37WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET38WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET39WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET41WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET42WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET43WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET45WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET46WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET47WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET56WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET57WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET58WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET59WW") }, (void*)3},
>+    { set_max_cstate, "IBM ThinkPad T43/p", {
>+      DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
>+      DMI_MATCH(DMI_BIOS_VERSION,"1YET60WW") }, (void*)3},
>     { set_max_cstate, "Medion 41700", {
>       DMI_MATCH(DMI_BIOS_VENDOR,"Phoenix Technologies LTD"),
>       DMI_MATCH(DMI_BIOS_VERSION,"R01-A1J")}, (void *)1},
>
>Signed-off by: Eric J Shattow <shadow@serverart.org>
>
>
>
>------------------------------------------------------------------
>  This email was checked by AMaViS anti-virus system !
>  Get yourself a free email address at http://mail.serverart.org
>

Re: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

[Sanjoy Mahajan]

Interesting!

> Are there any T43 users who do not hear this noise?

Not sure about a T43 but I hear a high-pitched noise on my T60.

> Do you hear this noise for both AC and battery modes?

It happens as soon as I pull out the AC power.  Well, more like half a
second after.  It comes from the lower left corner of the laptop as
you are looking at the keyboard, underneath the "Intel centrino duo"
sticker.  For a while I thought it was the wireless card, but
switching it off with the radio kill switch didn't stop the noise.

But if I pull out the AC power while running this shell script:

  while true ; do true ; done

then the noise doesn't start.

Kernel: 2.6.15-25-386 (Ubuntu), UP version (SMP fails on S3)
Hardware: T60, model 1951-C2U (T2400, Intel graphics and wireless)

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

RE: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

[Timo Hoenig]

Hi,

On Thu, 2006-09-21 at 18:05 -0400, Brown, Len wrote:

> Turning off C3 is a pretty big hammer...

Indeed.

I think it is a more general problem.  The systems with Intel Core Duo I
got my hands on all show the symptom:  Once the system is idle, the
second core begins to whine.  It's annoying in quiet environments.

      * Turning off the second core makes the whining stop.
      * Booting with processor.max_cstate=2 does also make the whining
        go away; but it's a bad idea on laptops if once wants to gain a
        long run on batteries.

Tested with HZ=250; I'll check whether HZ=100/1000 helps.

   Timo

Re: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

[Len Brown]

On Sunday 24 September 2006 07:38, Timo Hoenig wrote:

> I think it is a more general problem.  The systems with Intel Core Duo I
> got my hands on all show the symptom:  Once the system is idle, the
> second core begins to whine.  It's annoying in quiet environments.
> 
>       * Turning off the second core makes the whining stop.
>       * Booting with processor.max_cstate=2 does also make the whining
>         go away; but it's a bad idea on laptops if once wants to gain a
>         long run on batteries.
> 
> Tested with HZ=250; I'll check whether HZ=100/1000 helps.

Curious, I have not noticed this on the Core Duo systems I have.

In the past this has been traced to the voltage switching on the motherboard
causing the capacitors to vibrate and the motherboard to resonate.
That is why changing HZ will likely have an effect on the problem.

Exactly what model system do you have that is making noise?

-Len

Re: [PATCH 2.6.18 1/1] ACPI: limit cstate on noisy thinkpad t43/p models

[Timo Hoenig]

Hi,

On Mon, 2006-09-25 at 03:06 -0400, Len Brown wrote:

> On Sunday 24 September 2006 07:38, Timo Hoenig wrote:
> 
> > I think it is a more general problem.  The systems with Intel Core Duo I
> > got my hands on all show the symptom:  Once the system is idle, the
> > second core begins to whine.  It's annoying in quiet environments.
> > 
> >       * Turning off the second core makes the whining stop.
> >       * Booting with processor.max_cstate=2 does also make the whining
> >         go away; but it's a bad idea on laptops if once wants to gain a
> >         long run on batteries.
> > 
> > Tested with HZ=250; I'll check whether HZ=100/1000 helps.
> 
> Curious, I have not noticed this on the Core Duo systems I have.
> 
> In the past this has been traced to the voltage switching on the motherboard
> causing the capacitors to vibrate and the motherboard to resonate.
> That is why changing HZ will likely have an effect on the problem.

Yes, that's what I have found to be the culprit, too.  Sorry that I do
not have yet been able to check with different HZ than 250 so far.

> Exactly what model system do you have that is making noise?

* Lenovo ThinkPad T60, several different models
* Fujitsu Siemens e8110

Also, I just got my hands on a Lenovo ThinkPad X60 and ran the same test
[1] as on the other machines but did not hear any noise.

   Timo

[1] Test with Linux 2.6.18-rc6 (with pwd /sys/devices/system/cpu/cpu1)
    while (true); do echo 0 > online; sleep 1; echo 1 > online; sleep 1; done;