Skip to contentSkip to conversation

High pitched noise from laptop: processor.c in linux 2.6

DISCUSSION Fraz 10 participants 16 messages Oct 7, 2004 – Oct 22, 2004

lore
Share & tools
lore

Fraz

author
#0 lore raw

I recently upgraded from linux 2.4.27 to 2.6.8.1 and noticed my laptop now makes a high pitch noise while idle. I traced it back to the processor module for acpi. 'rmmod processor' stops the noise.

Using speed step to turn it down to 733 Mhz makes it a
little quieter and doesn't change the tone.

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

The laptop is a Compaq Evo N160 from Dec 2001. The reference in google was to a Centrino laptop with the same problem. I wrote to Dominik Brodowski who suggested posting here.

fraz

#1 lore raw

Hi!

Quoted 11 lines

I recently upgraded from linux 2.4.27 to 2.6.8.1 and noticed my laptop now makes a high pitch noise while idle. I traced it back to the processor module for acpi. 'rmmod processor' stops the noise.

Using speed step to turn it down to 733 Mhz makes it a
little quieter and doesn't change the tone.

Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

Ouch and btw it is hardware problem -- too cheap capacitors.
Pavel

Signature · 1 line
Boycott Kodak -- for their patent abuse against Java.
#2 lore raw
Quoted 13 lines · Pavel Machek

Hi!

Quoted 11 lines

I recently upgraded from linux 2.4.27 to 2.6.8.1 and noticed my laptop now makes a high pitch noise while idle. I traced it back to the processor module for acpi. 'rmmod processor' stops the noise.

Using speed step to turn it down to 733 Mhz makes it a
little quieter and doesn't change the tone.

Quoted 6 lines
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000. What would happen if one were to set HZ to a higher value, like 10000?

Quoted 4 lines

Ouch and btw it is hardware problem -- too cheap capacitors.
Pavel

Signature · 1 line
Boycott Kodak -- for their patent abuse against Java.

Actually, I don't know which is worse, patent abuse or Java misuse.

Signature · 2 lines
Måns Rullgård
[email protected]
#3 lore raw

Hi!

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000.

What benefits? HZ=1000 takes 1W more on my system.

Quoted 2 lines

What would happen if one
were to set HZ to a higher value, like 10000?

Try it.

Quoted 3 lines
Quoted 1 line

Boycott Kodak -- for their patent abuse against Java.

Actually, I don't know which is worse, patent abuse or Java misuse.

Well, java is ugly but not dangerous.
Pavel

Signature · 1 line
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         
#4 lore raw
Quoted 7 lines · Pavel Machek

Hi!

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Quoted 5 lines
Quoted 3 lines
Quoted 1 line

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000.

What benefits? HZ=1000 takes 1W more on my system.

Isn't it supposed to give more accurate timing?

Quoted 13 lines
Quoted 2 lines

What would happen if one
were to set HZ to a higher value, like 10000?

Try it.

Quoted 3 lines
Quoted 1 line

Boycott Kodak -- for their patent abuse against Java.

Actually, I don't know which is worse, patent abuse or Java misuse.

Well, java is ugly but not dangerous.
Pavel

Signature · 1 line
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         

Has that really happened?

Signature · 2 lines
Måns Rullgård
[email protected]
#5 lore raw

Hi!

Quoted 9 lines
Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
Quoted 1 line

code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000.

What benefits? HZ=1000 takes 1W more on my system.

Isn't it supposed to give more accurate timing?

Well, yes, but if it beeps for you... Accurate timing is probably not that important.

Quoted 3 lines
Quoted 1 line

64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms

Has that really happened?

Yes. On gprs it is rather easy to reproduce, unfortunately...

Pavel

Signature · 2 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!
#6 lore raw
In reply to Pavel Machek
Quoted 7 lines · Pavel Machek

Hi!

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000.

What benefits? HZ=1000 takes 1W more on my system.

Better timer resolution?

Lee

#7 lore raw
Quoted 2 lines · Lee Revell
Quoted 1 line

What benefits? HZ=1000 takes 1W more on my system.

Better timer resolution?

And heavily reduced accuracy on a lot of laptops where 1000Hz is enough to make the clock slide every time the battery state is queried or an SMM event triggers.

Getting the best of both worlds depends on the stuff discussed at OLS being finished, then you can have 1Khz accurancy and battery life

#8 lore raw
Quoted 8 lines · Alan Cox
Quoted 2 lines · Lee Revell
Quoted 1 line

What benefits? HZ=1000 takes 1W more on my system.

Better timer resolution?

And heavily reduced accuracy on a lot of laptops where 1000Hz is enough to make the clock slide every time the battery state is queried or an SMM event triggers.

Wouldn't such a laptop be horribly broken? 1ms is a LONG time to disable interrupts. That's millions of CPU cycles...

Quoted 3 lines

Getting the best of both worlds depends on the stuff discussed at OLS being finished, then you can have 1Khz accurancy and battery life

I was not there but I imagine this involves a way to get 1khz accuracy with a 100Hz timer interrupt rate?

Lee

#9 lore raw
Quoted 2 lines · Lee Revell

I was not there but I imagine this involves a way to get 1khz accuracy with a 100Hz timer interrupt rate?

I think the idea is to (depending on the hardware) dynamically adjust the system timers to exactly the time necessary. If you don't have any important processes that will interrupt within the next 10ms then you can just go ahead and set the timer longer. If properly done you might even be able to set HZ to 20 or 50. Hardware interrupts would trigger immediate responses, but otherwise a non-interactive system could minimize the number of context switches and make the processor cache that much more useful. This would still be bad on an interactive/desktop system because it would mean that perceptive users could detect the switching if they have two computationally intensive jobs running simultaneously.

Cheers,
Kyle Moffett

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a17 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------

#10 lore raw
In reply to Lee Revell
Quoted 5 lines · Lee Revell
Quoted 3 lines

And heavily reduced accuracy on a lot of laptops where 1000Hz is enough to make the clock slide every time the battery state is queried or an SMM event triggers.

Wouldn't such a laptop be horribly broken? 1ms is a LONG time to disable interrupts. That's millions of CPU cycles...

Yes, and most laptops have this problem. They use SMM traps to talk to the battery including huge delay loops and during those SMM traps no interrupt code runs.

Quoted 4 lines
Quoted 2 lines

Getting the best of both worlds depends on the stuff discussed at OLS being finished, then you can have 1Khz accurancy and battery life

I was not there but I imagine this involves a way to get 1khz accuracy with a 100Hz timer interrupt rate?

Think about

add_timeout(timer, when, precision_desired)

#11 lore raw
Quoted 11 lines · Alan Cox
Quoted 5 lines · Lee Revell
Quoted 3 lines

And heavily reduced accuracy on a lot of laptops where 1000Hz is enough to make the clock slide every time the battery state is queried or an SMM event triggers.

Wouldn't such a laptop be horribly broken? 1ms is a LONG time to disable interrupts. That's millions of CPU cycles...

Yes, and most laptops have this problem. They use SMM traps to talk to the battery including huge delay loops and during those SMM traps no interrupt code runs.

Ugh! I was under the impression that mostly older machines had this problem and it was a minority of laptops. I could not find a lot of info on SMM - several of the links I found were DDJ "Undocumented Corner" articles.

Anyway this explains probably half the weird bug reports on the linux audio user list.

Lee

#12 lore raw
In reply to Alan Cox
Quoted 2 lines · Alan Cox

Yes, and most laptops have this problem. They use SMM traps to talk to the battery

Hmm... Since the SMM on this AMD64 notebook we've been talking about on the thread about USB handoff isn't exactly AMD64-compliant, could this be the reason why I can't get battery information? I'm yet to install a 32-bit-only system on this box to see whether it makes any difference.

Signature · 3 lines
Alexandre Oliva             http://www.ic.unicamp.br/~oliva/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}
#13 lore raw
In reply to Måns Rullgård

Hi!

Quoted 9 lines
Quoted 6 lines
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

... and lose all the benefits of HZ=1000. What would happen if one were to set HZ to a higher value, like 10000?

You'll probably get even uglier noise and 5%+ performance penalty.

Signature · 1 line
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         
#14 lore raw
In reply to Pavel Machek
Quoted 6 lines · Pavel Machek
Quoted 4 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

Change #define HZ 1000 to #define HZ 100...

Helped an Asus S1300N. Not sure it's gone completely, but it's more comfortable now.

To the original poster: look in include/asm/param.h

#15 lore raw
In reply to Fraz
Quoted 7 lines · Fraz

I recently upgraded from linux 2.4.27 to 2.6.8.1 and noticed my laptop now makes a high pitch noise while idle. I traced it back to the processor module for acpi. 'rmmod processor' stops the noise.

Using speed step to turn it down to 733 Mhz makes it a little quieter and doesn't change the tone.

Quoted 8 lines

Is there any way to stop this? I googled around and found it had something to do with idle frequency of 1000 Hz in 2.6 instead of 100Hz in the 2.4 kernel. I couldn't find much else on this. Hunting around the code didn't help much, I don't know C.

The laptop is a Compaq Evo N160 from Dec 2001. The reference in google was to a Centrino laptop with the same problem. I wrote to Dominik Brodowski who suggested posting here.

Take a look at this bug report:

http://bugme.osdl.org/show_bug.cgi?id=2478

Signature · 4 lines
Christian Hesse

geek by nature
linux by choice