Changing temperature trip points

Changing temperature trip points

[Edmund Rhudy]

On my laptop, a Gateway Solo 1400 with one of those super-integrated VIA
Apollo Pro133A chipsets that has basically every system function
integrated into it (video, sound, modem, networking, and so forth), I
note that the temperature trip points are excruciatingly high; so high,
in fact, that the processor (a Celeron 900) would be close to thermal
death by the time the fan would purportedly turn itself on.

cat /proc/acpi/thermal_zone/TZN0/trip_points returns the following
output:

critical (S5):		105 C
passive:			90 C: tc1=4 tc2=3 tsp=100
devices=0xc129f620
active[0]:			65 C: devices=0xc12a9fe0

As it stands, I can currently turn the fan on manually with echo on >
/proc/acpi/fan/FAN0/state (though not turn it off with echo off, for
whatever reason), but I can't be constantly monitoring the temperature,
and I'd like the fan to just turn itself on and off only when necessary,
as it's rather loud.

This is with the latest version of ACPI (20021101) on 2.4.20-rc1 under
Red Hat 8.0.  If you require any further information, let me know and
I'll dig it out for you.  For the record, here's everything that dmesg
returned with regard to ACPI (any lines I skipped are indicated with
ellipses):

ACPI: have wakeup address 0xc0001000
On node 0 totalpages: 61424
zone(0): 4096 pages.
zone(1): 57328 pages.
zone(2): 0 pages.
ACPI: RSDP (v000 OID_00                     ) @ 0x000e5010
ACPI: RSDT (v001 GATEWA SOLO1400 08193.04647) @ 0x0efffbc0
ACPI: FADT (v001 GATEWA SOLO1400 08193.04647) @ 0x0efffac0
ACPI: BOOT (v001 GATEWA SOLO1400 08193.04647) @ 0x0efffb50
ACPI: DBGP (v001 GATEWA SOLO1400 08193.04647) @ 0x0efffb80
ACPI: DSDT (v001 GATEWA SOLO1400 08193.04647) @ 0x00000000
ACPI: BIOS passes blacklist
ACPI: Subsystem revision 20021101
PCI: PCI BIOS revision 2.10 entry at 0xe8a24, last bus=1
PCI: Using configuration type 1
    ACPI-0508: *** Info: GPE Block0 defined as GPE0 to GPE15
    ACPI-0508: *** Info: GPE Block1 defined as GPE16 to GPE31
ACPI: Interpreter enabled
ACPI: Using PIC for interrupt routing
ACPI: System [ACPI] (supports S0 S1 S3 S4 S5)
ACPI: PCI Root Bridge [PCI0] (00:00)
PCI: Probing PCI hardware (bus 00)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: Embedded Controller [EC0] (gpe 1)
ACPI: PCI Interrupt Link [LNKA] (IRQs 5 *7 10 11)
ACPI: PCI Interrupt Link [LNKB] (IRQs 5 7 10 *11)
ACPI: PCI Interrupt Link [LNKC] (IRQs 5 7 *10 11)
ACPI: PCI Interrupt Link [LNKD] (IRQs 5 7 10 *11)
ACPI: Power Resource [PFAN] (off)
...
ACPI: AC Adapter [ADP0] (off-line)
ACPI: Battery Slot [BAT0] (battery present)
ACPI: Power Button (FF) [PWRF]
ACPI: Lid Switch [LID]
ACPI: Sleep Button (CM) [SLPB]
ACPI: Fan [FAN0] (off)
ACPI: Processor [CPU0] (supports C1)
ACPI: Thermal Zone [TZN0] (39 C)

If you need anything else, let me know and I'll do my best to provide
it.  I apologise in advance for the length of this post and any
irrelevant information I may have included.




-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> On my laptop, a Gateway Solo 1400 with one of those super-integrated VIA
> Apollo Pro133A chipsets that has basically every system function
> integrated into it (video, sound, modem, networking, and so forth), I
> note that the temperature trip points are excruciatingly high; so high,
> in fact, that the processor (a Celeron 900) would be close to thermal
> death by the time the fan would purportedly turn itself on.

I have similar problem here. HP Omnibook turns itself off long before
passive trip point is reached.

Try echo "1:2:3:4:5" > /proc/acpi/thermal_zone/THRM/trip_points. Should work in 2.5.44; dunno
if 2.4. backport supports it.
								Pavel 
-- 
When do you have heart between your knees?


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Edmund Rhudy]

I did in fact try that, and while it lets me change the values, they
don't actually seem to affect the kernel's idea of when the fan should
come on, if at all.  I set the active trip point to 45 C and the passive
to 50 C, then ran a compile to see if the fan triggered.  While the
temperature got above 50 C, the fan never came on.  In fact, I don't
know if the fan would come on even if the originally-specified trip
points were reached.  Either way, this isn't exactly the most comforting
situation to observe.

On Sat, 2002-11-02 at 12:17, Pavel Machek wrote:
> Hi!
> 
> > On my laptop, a Gateway Solo 1400 with one of those super-integrated VIA
> > Apollo Pro133A chipsets that has basically every system function
> > integrated into it (video, sound, modem, networking, and so forth), I
> > note that the temperature trip points are excruciatingly high; so high,
> > in fact, that the processor (a Celeron 900) would be close to thermal
> > death by the time the fan would purportedly turn itself on.
> 
> I have similar problem here. HP Omnibook turns itself off long before
> passive trip point is reached.
> 
> Try echo "1:2:3:4:5" > /proc/acpi/thermal_zone/THRM/trip_points. Should work in 2.5.44; dunno
> if 2.4. backport supports it.
> 								Pavel 
> -- 
> When do you have heart between your knees?
> 





-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> I did in fact try that, and while it lets me change the values, they
> don't actually seem to affect the kernel's idea of when the fan should
> come on, if at all.  I set the active trip point to 45 C and the passive
> to 50 C, then ran a compile to see if the fan triggered.  While the
> temperature got above 50 C, the fan never came on.  In fact, I don't
> know if the fan would come on even if the originally-specified trip
> points were reached.  Either way, this isn't exactly the most comforting
> situation to observe.

That's because your fan is driven by hardware, not ACPI. I see it
here, too. Passive cooling should work, through.
								Pavel
-- 
Worst form of spam? Adding advertisment signatures ala sourceforge.net.
What goes next? Inserting advertisment *into* email?


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 08:56:16PM +0100, Pavel Machek wrote:
> Hi!
> 
> > I did in fact try that, and while it lets me change the values, they
> > don't actually seem to affect the kernel's idea of when the fan should
> > come on, if at all.  I set the active trip point to 45 C and the passive
> > to 50 C, then ran a compile to see if the fan triggered.  While the
> > temperature got above 50 C, the fan never came on.  In fact, I don't
> > know if the fan would come on even if the originally-specified trip
> > points were reached.  Either way, this isn't exactly the most comforting
> > situation to observe.
> 
> That's because your fan is driven by hardware, not ACPI. I see it
> here, too. Passive cooling should work, through.

Actually, I think the problem is something else: The BIOS is supposed to
generate ACPI events whenever these trip points are reached. So if you
change these values (by patching the DSDT, by "hardwiring" it into the
kernel) you need to use the polling method. Then ACPI will compare the
current temperature to the specified temperatures with a specified
frequency. Any chance you can try whether my assuption is correct?

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > > I did in fact try that, and while it lets me change the values, they
> > > don't actually seem to affect the kernel's idea of when the fan should
> > > come on, if at all.  I set the active trip point to 45 C and the passive
> > > to 50 C, then ran a compile to see if the fan triggered.  While the
> > > temperature got above 50 C, the fan never came on.  In fact, I don't
> > > know if the fan would come on even if the originally-specified trip
> > > points were reached.  Either way, this isn't exactly the most comforting
> > > situation to observe.
> > 
> > That's because your fan is driven by hardware, not ACPI. I see it
> > here, too. Passive cooling should work, through.
> 
> Actually, I think the problem is something else: The BIOS is supposed to
> generate ACPI events whenever these trip points are reached. So if you
> change these values (by patching the DSDT, by "hardwiring" it into the
> kernel) you need to use the polling method. Then ACPI will compare the
> current temperature to the specified temperatures with a specified
> frequency. Any chance you can try whether my assuption is correct?


YOU are right that I need to turn the polling on.

But I'm sure fan is done by bios on my machine.

Here is what my cool-it-down script looks like:

root@amd:~# cat watchtemp
#!/bin/bash
echo "83:83:60:70:0" > /proc/acpi/thermal_zone/THRM/trip_points
echo "5" > /proc/acpi/thermal_zone/THRM/polling_frequency
watch '( cat /proc/acpi/processor/CPU0/limit ; cat
/proc/acpi/thermal_zone/THRM/{temperature,state,trip_points} )'
root@amd:~#

						Pavel

-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

Hi!

On Sun, Nov 03, 2002 at 09:21:30PM +0100, Pavel Machek wrote:
> 
> Hi!
> 
> > Actually, I think the problem is something else: The BIOS is supposed to
> > generate ACPI events whenever these trip points are reached. So if you
> > change these values (by patching the DSDT, by "hardwiring" it into the
> > kernel) you need to use the polling method. Then ACPI will compare the
> > current temperature to the specified temperatures with a specified
> > frequency. Any chance you can try whether my assuption is correct?
> 
> 
> YOU are right that I need to turn the polling on.
> 
> But I'm sure fan is done by bios on my machine.

Yes, that might well be - the fan on my machine turns on and off both when
it thinks it should, and when ACPI tells it should *weird*

Actually, I think any write into trip_points should automatically enable
the polling mode. Will submit a patch which does that later. Do 3 seconds
sound sane?

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > > Actually, I think the problem is something else: The BIOS is supposed to
> > > generate ACPI events whenever these trip points are reached. So if you
> > > change these values (by patching the DSDT, by "hardwiring" it into the
> > > kernel) you need to use the polling method. Then ACPI will compare the
> > > current temperature to the specified temperatures with a specified
> > > frequency. Any chance you can try whether my assuption is correct?
> > 
> > 
> > YOU are right that I need to turn the polling on.
> > 
> > But I'm sure fan is done by bios on my machine.
> 
> Yes, that might well be - the fan on my machine turns on and off both when
> it thinks it should, and when ACPI tells it should *weird*
> 
> Actually, I think any write into trip_points should automatically enable
> the polling mode. Will submit a patch which does that later. Do 3 seconds
> sound sane?

I believe user should be clever enough to set polling rate manually.

Perhaps documenting it is enough?
				Pavel
PS: Polling rate has strong effect  on how passive works. If you set
it too fast it will basically only use "max performance" and "max
throttle", which is not what you want.

-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 09:39:49PM +0100, Pavel Machek wrote:
> Hi!
> 
> > > YOU are right that I need to turn the polling on.
> > > 
> > > But I'm sure fan is done by bios on my machine.
> > 
> > Yes, that might well be - the fan on my machine turns on and off both when
> > it thinks it should, and when ACPI tells it should *weird*
> > 
> > Actually, I think any write into trip_points should automatically enable
> > the polling mode. Will submit a patch which does that later. Do 3 seconds
> > sound sane?
> 
> I believe user should be clever enough to set polling rate manually.
> 
> Perhaps documenting it is enough?
> 				Pavel

> PS: Polling rate has strong effect  on how passive works. If you set
> it too fast it will basically only use "max performance" and "max
> throttle", which is not what you want.

That's true, yes. But it really should be documented. Well, maybe I'll find
the time to do the same for /thermal_zone what I've already done for
/processor at http://www.brodo.de/english/pub/acpi/proc/processor.html

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!


> > > > YOU are right that I need to turn the polling on.
> > > > 
> > > > But I'm sure fan is done by bios on my machine.
> > > 
> > > Yes, that might well be - the fan on my machine turns on and off both when
> > > it thinks it should, and when ACPI tells it should *weird*
> > > 
> > > Actually, I think any write into trip_points should automatically enable
> > > the polling mode. Will submit a patch which does that later. Do 3 seconds
> > > sound sane?
> > 
> > I believe user should be clever enough to set polling rate manually.
> > 
> > Perhaps documenting it is enough?
> > 				Pavel
> 
> > PS: Polling rate has strong effect  on how passive works. If you set
> > it too fast it will basically only use "max performance" and "max
> > throttle", which is not what you want.
> 
> That's true, yes. But it really should be documented. Well, maybe I'll find
> the time to do the same for /thermal_zone what I've already done for
> /processor at
> > > > http://www.brodo.de/english/pub/acpi/proc/processor.html

Actually, Documentation/acpi/processor.txt seems like right
place. "Polling" warning should be in Documentation/, really.

				Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

> Actually, Documentation/acpi/processor.txt seems like right
> place. "Polling" warning should be in Documentation/, really.

Which does not exist ==> extremely dangerous situation. Even though you
should know what you're doing when you meddle with the affairs of kernels,
kernels should try at least to some extent to be safe to use. 

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Overheating is *not* dangerous Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > Actually, Documentation/acpi/processor.txt seems like right
> > place. "Polling" warning should be in Documentation/, really.
> 
> Which does not exist ==> extremely dangerous situation. Even though you
> should know what you're doing when you meddle with the affairs of kernels,
> kernels should try at least to some extent to be safe to use. 

So perhaps its good time to create Documentation/acpi? ;-).

Anyway it is not *dangerous* situation. ACPI machines power down on
overtemp.
								Pavel,
who forced his toshiba satellite to 95Celsius CPU temp to see how it
behaves, and who has hp omnibook xe3 with broken cpu fan and 900mhz
athlon. Yes I have seen emergency shutdowns many time already.

[BTW toshiba is designed properly, and even without fan is able to
passively cool itself by going from 300MHz to 40Mhz. Omnibook is
broken and its thermal throttling is not capable of keeping CPU cool
enough. What is worse, in some cases omnibook cpu makes harddisk hot
(while cpu is in limit and machine does not do emergency poweroff),
leading to bad things (tm).]

-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Overheating is *not* dangerous Re: Changing temperature trip points

[Edmund Rhudy]

The point that I, and it would seem Dominik too, am trying to make is
that you can't always rely on this sort of stuff.  My laptop didn't shut
down when it was in danger, and it doesn't have any sort of throttling
safeguard, nor is there anything that can be changed in the BIOS.  My
desktop allows you to set a shutdown temperature, but by default it's
left blank, so the system would happily slag itself unless there is
direct user intervention.  User intervention is not guaranteed, nor is a
nicely behaved BIOS that watches for this sort of thing.

-----Original Message-----
From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org] 
Sent: Sunday, 3 November 2002 5:20 PM
To: Dominik Brodowski
Cc: Edmund Rhudy; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Overheating is *not* dangerous Re: [ACPI] Changing temperature
trip points


Hi!

> > Actually, Documentation/acpi/processor.txt seems like right place. 
> > "Polling" warning should be in Documentation/, really.
> 
> Which does not exist ==> extremely dangerous situation. Even though 
> you should know what you're doing when you meddle with the affairs of 
> kernels, kernels should try at least to some extent to be safe to use.

So perhaps its good time to create Documentation/acpi? ;-).

Anyway it is not *dangerous* situation. ACPI machines power down on
overtemp.
								Pavel,
who forced his toshiba satellite to 95Celsius CPU temp to see how it
behaves, and who has hp omnibook xe3 with broken cpu fan and 900mhz
athlon. Yes I have seen emergency shutdowns many time already.

[BTW toshiba is designed properly, and even without fan is able to
passively cool itself by going from 300MHz to 40Mhz. Omnibook is broken
and its thermal throttling is not capable of keeping CPU cool enough.
What is worse, in some cases omnibook cpu makes harddisk hot (while cpu
is in limit and machine does not do emergency poweroff), leading to bad
things (tm).]

-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Overheating is *not* dangerous Re: Changing temperature trip points

[Pavel Machek]

Hi!

> The point that I, and it would seem Dominik too, am trying to make is
> that you can't always rely on this sort of stuff.  My laptop didn't shut
> down when it was in danger, and it doesn't have any sort of
> throttling

Your laptop is still alive, so I do not think it was in
danger. [Athlon/900 seems to work stable at 83Celsius. Celeron/300
worked stable up to 95Celsius. I do not know why you think 70Celsius
is too much for celeron].

> safeguard, nor is there anything that can be changed in the BIOS.  My
> desktop allows you to set a shutdown temperature, but by default it's
> left blank, so the system would happily slag itself unless there is
> direct user intervention.  User intervention is not guaranteed, nor is a
> nicely behaved BIOS that watches for this sort of thing.

*If* fan on your desktop dies (1st hardware bug), and
your ACPI bios is broken (2nd hardware bug), and
you configure trip points but forget polling frequency (2 hw bugs and
user error), and 
your CPU does not have thermal protection (it has to according to ACPI
specs -- 3 hw bugs and user error)
you can cook your cpu (given its older athlon).

Okay but even if we compensated for your user error, still 3 hw bugs
and one software bug would cook your machine (kernel crash).

If hw vendor made 3 hardware errors... I guess he deserves to have
that machine cooked.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Overheating is *not* dangerous Re: Changing temperature trip points

[Edmund Rhudy]

So you're saying that if the hardware vendor screwed up, the end user
should suffer?  Are you entirely certain that this is an appropriate
view to take?

As for the Celeron, Intel lists the maximum junction temperature of the
4 different varieties of 900 MHz Celerons (the speed of the one in my
laptop) at either 72 C (in the case of the SL633) or 77 C (the SL5WY,
SL5MQ, SL5LX).

-----Original Message-----
From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org] 
Sent: Sunday, 3 November 2002 5:38 PM
To: Edmund Rhudy
Cc: 'Dominik Brodowski'; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: Overheating is *not* dangerous Re: [ACPI] Changing
temperature trip points


Hi!

> The point that I, and it would seem Dominik too, am trying to make is 
> that you can't always rely on this sort of stuff.  My laptop didn't 
> shut down when it was in danger, and it doesn't have any sort of 
> throttling

Your laptop is still alive, so I do not think it was in
danger. [Athlon/900 seems to work stable at 83Celsius. Celeron/300
worked stable up to 95Celsius. I do not know why you think 70Celsius is
too much for celeron].

> safeguard, nor is there anything that can be changed in the BIOS.  My 
> desktop allows you to set a shutdown temperature, but by default it's 
> left blank, so the system would happily slag itself unless there is 
> direct user intervention.  User intervention is not guaranteed, nor is

> a nicely behaved BIOS that watches for this sort of thing.

*If* fan on your desktop dies (1st hardware bug), and
your ACPI bios is broken (2nd hardware bug), and
you configure trip points but forget polling frequency (2 hw bugs and
user error), and 
your CPU does not have thermal protection (it has to according to ACPI
specs -- 3 hw bugs and user error) you can cook your cpu (given its
older athlon).

Okay but even if we compensated for your user error, still 3 hw bugs and
one software bug would cook your machine (kernel crash).

If hw vendor made 3 hardware errors... I guess he deserves to have that
machine cooked.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Overheating is *not* dangerous Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 11:19:45PM +0100, Pavel Machek wrote:
> Hi!
> 
> > > Actually, Documentation/acpi/processor.txt seems like right
> > > place. "Polling" warning should be in Documentation/, really.
> > 
> > Which does not exist ==> extremely dangerous situation. Even though you
> > should know what you're doing when you meddle with the affairs of kernels,
> > kernels should try at least to some extent to be safe to use. 
> 
> So perhaps its good time to create Documentation/acpi? ;-).
> 
> Anyway it is not *dangerous* situation. ACPI machines power down on
> overtemp.
ACPI machines power down on overtemp. What is overtemp? The temperature
specified in CRITICAL in the DSDT. And if this value is wrong, then you need
to rely on the sanity of BIOS programmers. And a log of good luck.

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Overheating is *not* dangerous Re: Changing temperature trip points

[Pavel Machek]

Hi!
> > > > Actually, Documentation/acpi/processor.txt seems like right
> > > > place. "Polling" warning should be in Documentation/, really.
> > > 
> > > Which does not exist ==> extremely dangerous situation. Even though you
> > > should know what you're doing when you meddle with the affairs of kernels,
> > > kernels should try at least to some extent to be safe to use. 
> > 
> > So perhaps its good time to create Documentation/acpi? ;-).
> > 
> > Anyway it is not *dangerous* situation. ACPI machines power down on
> > overtemp.
> ACPI machines power down on overtemp. What is overtemp? The temperature
> specified in CRITICAL in the DSDT. And if this value is wrong, then you need
> to rely on the sanity of BIOS programmers. And a log of good luck.

Actually, I believe its sanity of HW designers. I'd not trust BIOS
programmers to get it right.

There's one more point above _CRT, that's what I call overtemp:

[from specs:]

This says that if you can cook your machine, it was not acpi
compliant.
							Pavel

When the heat reaches the temperature indicated by _CRT, the OS must
immediately shutdown the system.
The system must disable the power either after the temperature reaches
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
some hardware-determined level above
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
_CRT or after a predetermined time has passed. Before disabling power,
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
platform designers should incorporate
some time that allows the OS to run its critical shutdown
operation. There is no requirement for a minimum
shutdown operation window that commences immediately after the
temperature reaches _CRT.  This is
because
*  Heat might rise rapidly in some systems and slower on others,
depending on casing design and
    environmental factors.
*  Shutdown can take several minutes on a server and only a few short
seconds on a hand-held device.
Because of this indistinct discrepancy and the fact that a critical
heat situation is a remarkably rare occurrence,
ACPI does not specify a target window for a safe shutdown. It is
entirely up to the OEM to build in a safe
buffer that it sees fit for the target platform.


-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Overheating is *not* dangerous Re: Changing temperature trip points

[Diego Zuccato]

Pavel Machek wrote:

> [BTW toshiba is designed properly, and even without fan is able to
> passively cool itself by going from 300MHz to 40Mhz. Omnibook is
> broken and its thermal throttling is not capable of keeping CPU cool
Acer (at least my TM525TE) seems badly designed: sometimes (even if not
overheating...) it locks the CPU to 200MHz and won't allow it to raise
till a reboot. :-(

BYtE,
 Diego.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 09:39:49PM +0100, Pavel Machek wrote:
> PS: Polling rate has strong effect  on how passive works. If you set
> it too fast it will basically only use "max performance" and "max
> throttle", which is not what you want.
No, it doesn't: the _trend_ is deciding, and not the fact that the temp is
above the limit. additionally, polling is used automatically when passive
cooling a.k.a. "throttle || performance" is used. Patch will follow soon.

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > PS: Polling rate has strong effect  on how passive works. If you set
> > it too fast it will basically only use "max performance" and "max
> > throttle", which is not what you want.
> No, it doesn't: the _trend_ is deciding, and not the fact that the temp is
> above the limit. additionally, polling is used automatically when passive
> cooling a.k.a. "throttle || performance" is used. Patch will follow soon.

AFAICR, its something like a * temperature + b * trend.

Assume your polling frequency is 1msec, and passive limit 50C.

t: Temp is 50C, machine is cool at P0:T0.
t+1 msec: Temp is 51C. Trend is +, temperature is over, machine goes P0:T1.
t+2 msec: Temp is still 51C (1 msec is too little time for temp to
change), trend is no change, but temperature is still over, machine
goes P0:T2.
t+3 msec: Temp is still 51C (2 msec is too little time for temp to
change), trend is no change, but temperature is still over, machine
goes P0:T3.
at t+7 msec machine reaches P0:T7.
at t+1000msec cpu cools down, and you reach P0:T0 in 7 msec.

Not good.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 11:15:32PM +0100, Pavel Machek wrote:
> Hi!
> 
> > > PS: Polling rate has strong effect  on how passive works. If you set
> > > it too fast it will basically only use "max performance" and "max
> > > throttle", which is not what you want.
> > No, it doesn't: the _trend_ is deciding, and not the fact that the temp is
> > above the limit. additionally, polling is used automatically when passive
> > cooling a.k.a. "throttle || performance" is used. Patch will follow soon.
> 
> AFAICR, its something like a * temperature + b * trend.
> 
> Assume your polling frequency is 1msec, and passive limit 50C.
> 
> t: Temp is 50C, machine is cool at P0:T0.
> t+1 msec: Temp is 51C. Trend is +, temperature is over, machine goes P0:T1.
> t+2 msec: Temp is still 51C (1 msec is too little time for temp to
> change), trend is no change, but temperature is still over, machine
> goes P0:T2.
> t+3 msec: Temp is still 51C (2 msec is too little time for temp to
> change), trend is no change, but temperature is still over, machine
> goes P0:T3.
> at t+7 msec machine reaches P0:T7.
> at t+1000msec cpu cools down, and you reach P0:T0 in 7 msec.
> 
> Not good.

No, that's wrong: 

1.) If passive cooling is active, _TSP is used _in all cases_
as polling frequency (_TSP is an integer given in 1/10th seconds) - see
acpi_thremal_check:
*
	if (tz->state.passive)
		sleep_time = tz->trips.passive.tsp * 100;
	else if (tz->polling_frequency > 0)
		sleep_time = tz->polling_frequency * 100;
*
2.) This means that in acpi_thermal_passive:
		trend = (passive->tc1 * (tz->temperature -
tz->last_temperature)) + (passive->tc2 * (tz->temperature -
passive->temperature));
    is only calculated each _TSP frequency. And only when this trend is 
   positive, throttling or perf. state is increased.
*
		if (trend > 0)
			for (i=0; i<passive->devices.count; i++)
				acpi_processor_set_thermal_limit(
					passive->devices.handles[i], 
					ACPI_PROCESSOR_LIMIT_INCREMENT);
*

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > goes P0:T3.
> > at t+7 msec machine reaches P0:T7.
> > at t+1000msec cpu cools down, and you reach P0:T0 in 7 msec.
> > 
> > Not good.
> 
> No, that's wrong: 
> 
> 1.) If passive cooling is active, _TSP is used _in all cases_
> as polling frequency (_TSP is an integer given in 1/10th seconds) - see
> acpi_thremal_check:

Okay, you were right and I was wrong.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Changing temperature trip points

[Edmund Rhudy]

Unfortunately not, because I'm not a programmer of any sort.  If I get
anything done, it's by the grace of the kernel compiling gods.  My modus
operandi consists of blindly plugging away and hoping that something
works, because I am barely acquainted with C and know nothing whatsoever
of regular expressions.  The most I can do is really test something that
somebody else comes up with and report back.

-----Original Message-----
From: Dominik Brodowski [mailto:linux-JhLEnvuH02M@public.gmane.org] 
Sent: Sunday, 3 November 2002 3:16 PM
To: Pavel Machek
Cc: Edmund Rhudy; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [ACPI] Changing temperature trip points


On Sun, Nov 03, 2002 at 08:56:16PM +0100, Pavel Machek wrote:
> Hi!
> 
> > I did in fact try that, and while it lets me change the values, they

> > don't actually seem to affect the kernel's idea of when the fan 
> > should come on, if at all.  I set the active trip point to 45 C and 
> > the passive to 50 C, then ran a compile to see if the fan triggered.

> > While the temperature got above 50 C, the fan never came on.  In 
> > fact, I don't know if the fan would come on even if the 
> > originally-specified trip points were reached.  Either way, this 
> > isn't exactly the most comforting situation to observe.
> 
> That's because your fan is driven by hardware, not ACPI. I see it 
> here, too. Passive cooling should work, through.

Actually, I think the problem is something else: The BIOS is supposed to
generate ACPI events whenever these trip points are reached. So if you
change these values (by patching the DSDT, by "hardwiring" it into the
kernel) you need to use the polling method. Then ACPI will compare the
current temperature to the specified temperatures with a specified
frequency. Any chance you can try whether my assuption is correct?

	Dominik





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

Hi,

On Sun, Nov 03, 2002 at 04:08:43PM -0500, Edmund Rhudy wrote:
> Unfortunately not, because I'm not a programmer of any sort.  If I get
> anything done, it's by the grace of the kernel compiling gods.  My modus
> operandi consists of blindly plugging away and hoping that something
> works, because I am barely acquainted with C and know nothing whatsoever
> of regular expressions.  The most I can do is really test something that
> somebody else comes up with and report back.

Actually, I only wanted to know whether you set a "polling frequency" (e.g.
echo 1 > polling_frequency after setting these temperature trip points.

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Changing temperature trip points

[Edmund Rhudy]

See, now that I can understand.  :)

Anyway, let's see...having set temperatures to 60, 50 and 45 C
respectively, and running a compile to move the core temperature up, the
fan does in fact trigger at 45 C.  Outstanding.  I'll script it into
changing them automatically at startup.

Pavel also wrote the following:

> I believe user should be clever enough to set polling rate manually.

I wouldn't leave anything to the user unless it's extremely clear cut,
particularly with the movement towards a more Windows-like Linux (see
Bluecurve) that tries to soften the reality of Linux not being that
friendly to the average schmuck.

-----Original Message-----
From: Dominik Brodowski [mailto:linux-JhLEnvuH02M@public.gmane.org] 
Sent: Sunday, 3 November 2002 4:15 PM
To: Edmund Rhudy
Cc: 'Pavel Machek'; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [ACPI] Changing temperature trip points


Hi,

On Sun, Nov 03, 2002 at 04:08:43PM -0500, Edmund Rhudy wrote:
> Unfortunately not, because I'm not a programmer of any sort.  If I get

> anything done, it's by the grace of the kernel compiling gods.  My 
> modus operandi consists of blindly plugging away and hoping that 
> something works, because I am barely acquainted with C and know 
> nothing whatsoever of regular expressions.  The most I can do is 
> really test something that somebody else comes up with and report 
> back.

Actually, I only wanted to know whether you set a "polling frequency"
(e.g. echo 1 > polling_frequency after setting these temperature trip
points.

	Dominik





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

Hi,

On Sun, Nov 03, 2002 at 04:24:51PM -0500, Edmund Rhudy wrote:
> See, now that I can understand.  :)

Sorry for being unclear before.

> Anyway, let's see...having set temperatures to 60, 50 and 45 C
> respectively, and running a compile to move the core temperature up, the
> fan does in fact trigger at 45 C.  Outstanding.  I'll script it into
> changing them automatically at startup.

:-)
 
> Pavel also wrote the following:
> 
> > I believe user should be clever enough to set polling rate manually.
> 
> I wouldn't leave anything to the user unless it's extremely clear cut,
> particularly with the movement towards a more Windows-like Linux (see
> Bluecurve) that tries to soften the reality of Linux not being that
> friendly to the average schmuck.

OK, I'll change my opinion once again -- upon further looking at the code.
Patch will follow soon.

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> See, now that I can understand.  :)
> 
> Anyway, let's see...having set temperatures to 60, 50 and 45 C
> respectively, and running a compile to move the core temperature up, the
> fan does in fact trigger at 45 C.  Outstanding.  I'll script it into
> changing them automatically at startup.
> 
> Pavel also wrote the following:
> 
> > I believe user should be clever enough to set polling rate manually.
> 
> I wouldn't leave anything to the user unless it's extremely clear cut,
> particularly with the movement towards a more Windows-like Linux (see
> Bluecurve) that tries to soften the reality of Linux not being that
> friendly to the average schmuck.

No.

You only ever need to touch polling_frequency in case your hardware is
broken. If you are working around broken hardware,you'd better know
what you are doing.

Kernel is *not* meant to be user friendly. You can have your "broken
hardware setup wizard" in your distribution if you want, but not in
kernel.

				Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Changing temperature trip points

[Edmund Rhudy]

My point is more that it might be better to assume that things won't
work.  This isn't a sound driver.  This is something that the life of
the computer depends on, and if somebody's processor melts down because
they didn't read some piece of kernel documentation that they didn't
even know existed, they're not likely to be particularly forgiving and
think "Oh, it was my fault for not realising that my hardware was
broken, I'll be more careful next time".

-----Original Message-----
From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org] 
Sent: Sunday, 3 November 2002 5:03 PM
To: Edmund Rhudy
Cc: 'Dominik Brodowski'; 'Pavel Machek';
acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [ACPI] Changing temperature trip points


Hi!

> See, now that I can understand.  :)
> 
> Anyway, let's see...having set temperatures to 60, 50 and 45 C 
> respectively, and running a compile to move the core temperature up, 
> the fan does in fact trigger at 45 C.  Outstanding.  I'll script it 
> into changing them automatically at startup.
> 
> Pavel also wrote the following:
> 
> > I believe user should be clever enough to set polling rate manually.
> 
> I wouldn't leave anything to the user unless it's extremely clear cut,

> particularly with the movement towards a more Windows-like Linux (see
> Bluecurve) that tries to soften the reality of Linux not being that 
> friendly to the average schmuck.

No.

You only ever need to touch polling_frequency in case your hardware is
broken. If you are working around broken hardware,you'd better know what
you are doing.

Kernel is *not* meant to be user friendly. You can have your "broken
hardware setup wizard" in your distribution if you want, but not in
kernel.

				Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

> My point is more that it might be better to assume that things won't
> work.  This isn't a sound driver.  This is something that the life of
> the computer depends on, and if somebody's processor melts down because
> they didn't read some piece of kernel documentation that they didn't
> even know existed, they're not likely to be particularly forgiving and
> think "Oh, it was my fault for not realising that my hardware was
> broken, I'll be more careful next time".

No. Worst that can happen is that your machine powers off. And thats
after machine already has at least one hardware bug.
								Pavel


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

Hi!

On Sun, Nov 03, 2002 at 11:11:43PM +0100, Pavel Machek wrote:
> > My point is more that it might be better to assume that things won't
> > work.  This isn't a sound driver.  This is something that the life of
> > the computer depends on, and if somebody's processor melts down because
> > they didn't read some piece of kernel documentation that they didn't
> > even know existed, they're not likely to be particularly forgiving and
> > think "Oh, it was my fault for not realising that my hardware was
> > broken, I'll be more careful next time".
> 
> No. Worst that can happen is that your machine powers off. And thats
> after machine already has at least one hardware bug.

Actually, I doubt that it will power off. Assume your BIOS vendor set
critical to 110°C (my vendor did so). And the CPU burns through at 70°C.
Let's assume no thermal management objects like fans exist. Without custom
polling, a trip point of 65°C for critical has no, really no effect. The
CPU melts down. Or am I missing something?

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> > > My point is more that it might be better to assume that things won't
> > > work.  This isn't a sound driver.  This is something that the life of
> > > the computer depends on, and if somebody's processor melts down because
> > > they didn't read some piece of kernel documentation that they didn't
> > > even know existed, they're not likely to be particularly forgiving and
> > > think "Oh, it was my fault for not realising that my hardware was
> > > broken, I'll be more careful next time".
> > 
> > No. Worst that can happen is that your machine powers off. And thats
> > after machine already has at least one hardware bug.
> 
> Actually, I doubt that it will power off. Assume your BIOS vendor set
> critical to 110°C (my vendor did so). And the CPU burns through at 70°C.
> Let's assume no thermal management objects like fans exist. Without custom
> polling, a trip point of 65°C for critical has no, really no effect. The
> CPU melts down. Or am I missing something?

Yes, you are missing reality.

Try stopping your fan and see what happens. I believe you'll see your
notebook powering off itself at <90C, low enough to keep cpu alive.

									Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 11:22:22PM +0100, Pavel Machek wrote:
> Hi!
> 
> > Actually, I doubt that it will power off. Assume your BIOS vendor set
> > critical to 110°C (my vendor did so). And the CPU burns through at 70°C.
> > Let's assume no thermal management objects like fans exist. Without custom
> > polling, a trip point of 65°C for critical has no, really no effect. The
> > CPU melts down. Or am I missing something?
> 
> Yes, you are missing reality.
> 
> Try stopping your fan and see what happens. I believe you'll see your
> notebook powering off itself at <90C, low enough to keep cpu alive.

I hope that such a safeguard is present. But I don't know it. And I won't
test it. Better be safe than sorry. 

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

RE: Changing temperature trip points

[Edmund Rhudy]

Before I discovered the problem with my computer, I'd run a kernel
compile that brought the core temperature up to over 75 C.  A Celeron
reaches critical temperature with risk of heat death at 70 C.  You can't
assume that all computers are going to have the same safeguards.

-----Original Message-----
From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org] 
Sent: Sunday, 3 November 2002 5:12 PM
To: Edmund Rhudy
Cc: 'Pavel Machek'; 'Dominik Brodowski';
acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [ACPI] Changing temperature trip points


> My point is more that it might be better to assume that things won't 
> work.  This isn't a sound driver.  This is something that the life of 
> the computer depends on, and if somebody's processor melts down 
> because they didn't read some piece of kernel documentation that they 
> didn't even know existed, they're not likely to be particularly 
> forgiving and think "Oh, it was my fault for not realising that my 
> hardware was broken, I'll be more careful next time".

No. Worst that can happen is that your machine powers off. And thats
after machine already has at least one hardware bug.
								Pavel





-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> Before I discovered the problem with my computer, I'd run a kernel
> compile that brought the core temperature up to over 75 C.  A Celeron
> reaches critical temperature with risk of heat death at 70 C.  You can't
> assume that all computers are going to have the same safeguards.

They have to.

ACPI is software, and in case of kernel crash there will be no
cooling... So if you don't have your safeguard (I believe you *do*
have it), it will kill itself during kernelcrash when you are not
looking.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Dominik Brodowski]

On Sun, Nov 03, 2002 at 05:09:43PM -0500, Edmund Rhudy wrote:
> My point is more that it might be better to assume that things won't
> work.  This isn't a sound driver.  This is something that the life of
> the computer depends on, and if somebody's processor melts down because
> they didn't read some piece of kernel documentation that they didn't
> even know existed,
... or which even does not exist yet ...
> they're not likely to be particularly forgiving and
> think "Oh, it was my fault for not realising that my hardware was
> broken, I'll be more careful next time".
While we're protected by the GPL so that nobody can sue us because of this,
we should try to be a bit "friendly". And it doesn't lose us anything.

	Dominik


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com

Re: Changing temperature trip points

[Arndt Schoenewald]

On Sat, Nov 02, 2002 at 12:35:56AM -0500, Edmund Rhudy wrote:
> [...]
> As it stands, I can currently turn the fan on manually with echo on >
> /proc/acpi/fan/FAN0/state (though not turn it off with echo off, for
> whatever reason), but I can't be constantly monitoring the temperature,
> [...]

Just FYI, "on" and "off" are not the right keywords here:

    echo >/proc/acpi/fan/FAN0/state 0		turns the fan on
    echo >/proc/acpi/fan/FAN0/state 3		turns the fan off

(On my machine, though, 3 doesn't turn the fan off, but makes it run
slower and thus more quiet.)

Arndt


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Maier Gerfried]

Zitiere Arndt Schoenewald <abs-SA7OhAOe25xnNxvc45mVi0K323yFvGpRdefyYXQ/eNw@public.gmane.org>:

[...]
> Just FYI, "on" and "off" are not the right keywords here:
> 
>     echo >/proc/acpi/fan/FAN0/state 0		turns the fan on
>     echo >/proc/acpi/fan/FAN0/state 3		turns the fan off
> 
> (On my machine, though, 3 doesn't turn the fan off, but makes it run
> slower and thus more quiet.)

Does this mean, that there shold be /proc/acpi/fan/... present in the latest 
acpi (kernel 2.4.20-rc1)?

Well, unfortunately I do not have it :-(
(fan-module seems to be loaded without any problems - meaning: error messages; 
acpi-debug enabled)

Maier Gerfried


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Pavel Machek]

Hi!

> [...]
> > Just FYI, "on" and "off" are not the right keywords here:
> > 
> >     echo >/proc/acpi/fan/FAN0/state 0		turns the fan on
> >     echo >/proc/acpi/fan/FAN0/state 3		turns the fan off
> > 
> > (On my machine, though, 3 doesn't turn the fan off, but makes it run
> > slower and thus more quiet.)
> 
> Does this mean, that there shold be /proc/acpi/fan/... present in the latest 
> acpi (kernel 2.4.20-rc1)?
> 
> Well, unfortunately I do not have it :-(
> (fan-module seems to be loaded without any problems - meaning: error messages; 
> acpi-debug enabled)

We are talking 2.5.X here.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Arndt Schoenewald]

On Fri, Nov 08, 2002 at 08:45:06AM +0100, Maier Gerfried wrote:
> Zitiere Arndt Schoenewald <abs-SA7OhAOe25xnNxvc45mVi0K323yFvGpRdefyYXQ/eNw@public.gmane.org>:
> 
> [...]
> > Just FYI, "on" and "off" are not the right keywords here:
> > 
> >     echo >/proc/acpi/fan/FAN0/state 0		turns the fan on
> >     echo >/proc/acpi/fan/FAN0/state 3		turns the fan off
> > 
> > (On my machine, though, 3 doesn't turn the fan off, but makes it run
> > slower and thus more quiet.)
> 
> Does this mean, that there shold be /proc/acpi/fan/... present in the latest 
> acpi (kernel 2.4.20-rc1)?
> 
> Well, unfortunately I do not have it :-(
> (fan-module seems to be loaded without any problems - meaning: error messages; 
> acpi-debug enabled)

Yes, /proc/acpi/fan/... should be present, also in 2.4.x kernels; I had
it all the time since 2.4.18+acpi-20020329 (the first Linux+ACPI kernel
I put on my then brandnew Gericom M6-T laptop). I haven't tried kernel
2.4.20-rc1 yet, though.

Are you using the ACPI patch from http://sf.net/projects/acpi?

Of course the fan can only be controlled through ACPI if there is a fan
and it is listed correctly in the ACPI description tables. This is
typically the case for laptops and notebooks (but not for all of them),
and typically not the case for desktop machines.

Arndt


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Jörg Sonnenberger]

Dominik wrote:
>While we're protected by the GPL so that nobody can sue us because of
>this,
>we should try to be a bit "friendly". And it doesn't lose us anything.

At least for the German law, you are definitivly wrong. You can be sued
for a known and not clearly stated fault in your "product".

Joerg


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Alan Cox]

On Mon, 2002-11-04 at 21:20, Jörg Sonnenberger wrote:
> Dominik wrote:
> >While we're protected by the GPL so that nobody can sue us because of
> >this,
> >we should try to be a bit "friendly". And it doesn't lose us anything.
> 
> At least for the German law, you are definitivly wrong. You can be sued
> for a known and not clearly stated fault in your "product".

Open source code isnt a product, nor is it sold, nor do you have any
other contract option than the GPL or refusal. Its just like any other
scientific publication, a refereed journal if you like. In fact the
contribution statistics actually even match that of a refereed journal

If you bought it in a box then the people who sold you the box may well
have duties but thats their problem. In the case of a PC which self
destructs when given perfectly valid code for a PC compatible to execute
(which may or may not be the case - I suspect HPaq people who actually
know the answer will indeed say the box is thermal protected) I suspect
the PC vendor would be the one at fault anyway, especially if they
described it as a "PC" or "IBM PC compatible". If they didn't well why
are you running software sold for IBM PC compatibles on it 8)

Alan



-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en

Re: Changing temperature trip points

[Maier Gerfried]

Zitiere Arndt Schoenewald <abs-SA7OhAOe25xnNxvc45mVi0K323yFvGpRdefyYXQ/eNw@public.gmane.org>:

> Yes, /proc/acpi/fan/... should be present, also in 2.4.x kernels; I had
> it all the time since 2.4.18+acpi-20020329 (the first Linux+ACPI kernel
> I put on my then brandnew Gericom M6-T laptop). I haven't tried kernel
> 2.4.20-rc1 yet, though.
> 
> Are you using the ACPI patch from http://sf.net/projects/acpi?

Yes. With the suggested debugging-patch _HIGH -> _LOW and with acpi-debugging 
enabled.

> Of course the fan can only be controlled through ACPI if there is a fan
> and it is listed correctly in the ACPI description tables. This is
> typically the case for laptops and notebooks (but not for all of them),
> and typically not the case for desktop machines.

It's a Acer Travelmate Notebook and it has got a fan. I'm afraid, that I am not 
able to have a look to the ACPI-description table with my recent knowlege. How 
could I check this?

(According to lsmod the fan-module is loaded.)

Gerfried


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en