Hi,
The CPU on my EeePC laptop has been slowly overheating since yesterday (24 April), shortly after performing an update. The last update I did was on Tuesday (21 April), but I do not think the problem was present after that.
The sensors seem to report flawed information, at least regarding the fan speed: it is reported to be around 4000 rpm. The fan does spin (I can hear it if I hold my ear really close to the chassis), but definitely not at 4000 rpm (I remember from past issues that I shouldn't need any effort to hear it at that speed). The reported speed also sometimes drops to zero without any continuity.
I mention that the overheating is "slow", I mean that I can work around 15 minutes before the computer shuts down, and that, idle, it takes about 1 hour to reach critical temperature. If the fan was not spinning at all, it would be 2 minutes.
Also, I could not get any form of manual control over the fan with pwm1_enable set to 1.
Booting on a previous kernel did not help. If I boot on a different distribution or on windows, the OS does not see the fan speed sensor, and the computer overheats slowly all the same. But as I said, I already had similar problems with this computer (the fan was not spinning at all, then), and I remember that it showed across distributions in the same way. If I remember, I concluded back then that something had set a bios flag that prevented the bios from handling the fan properly for other OSes.
I have uploaded some information about my hardware on https://gist.github.com/ThibautVerron/27da7b4da3940f9894c0 .
Does anyone have any clue as to how to investigate (and hopefully fix) the issue?
Thanks,
T. Verron
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
On Sat, 25 Apr 2015 10:47:58 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
Does anyone have any clue as to how to investigate (and hopefully fix) the issue?
Your laptop is probably clogged with dust. You need to take it apart and clean it. I need to clean almost all my machines every year or two.
- -- Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B
N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩ�
2015-04-27 12:44 GMT+02:00 Jan Matejka JMatejka@suse.cz:
On Sat, 25 Apr 2015 10:47:58 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
Does anyone have any clue as to how to investigate (and hopefully fix) the issue?
Your laptop is probably clogged with dust. You need to take it apart and clean it. I need to clean almost all my machines every year or two.
Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B
Hi,
Thank you for your answer.
Although I cannot rule this possibility out with 100% certainty, I find it unlikely that the problem would appear all of a sudden. And it would not explain the inaccurate fan speed reported by the sensors, either.
Sadly, I don't have the surgical skills required for taking apart and back together this kind of laptops.
T. Verron
On Mon, Apr 27, 2015 at 8:50 AM, Thibaut Verron thibaut.verron@gmail.com wrote:
2015-04-27 12:44 GMT+02:00 Jan Matejka JMatejka@suse.cz:
On Sat, 25 Apr 2015 10:47:58 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
Does anyone have any clue as to how to investigate (and hopefully fix) the issue?
Your laptop is probably clogged with dust. You need to take it apart and clean it. I need to clean almost all my machines every year or two.
Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B
Hi,
Thank you for your answer.
Although I cannot rule this possibility out with 100% certainty, I find it unlikely that the problem would appear all of a sudden. And it would not explain the inaccurate fan speed reported by the sensors, either.
It's quite common that it happens all of a sudden.
Either you notice all of a sudden, or it crosses the threshold from not-a-problem into overheating-problem all of a sudden.
The inaccuracies are quite common too, I can usually see them even in the BIOS of pristine machines.
Sadly, I don't have the surgical skills required for taking apart and back together this kind of laptops.
Just get a compressed air can (make sure it's for electronics, otherwise it might have incompatible solvents), make sure your laptop is turned off, and blow some air through the cooler vent. Make sure to hold the can correctly (read the instructions) to avoid propellent to go out as well. That should move dust around enough to make things better. And if it does make things better, then you've got confirmation that dust is to blame.
2015-04-27 18:18 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Mon, Apr 27, 2015 at 8:50 AM, Thibaut Verron thibaut.verron@gmail.com wrote:
2015-04-27 12:44 GMT+02:00 Jan Matejka JMatejka@suse.cz:
On Sat, 25 Apr 2015 10:47:58 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
Does anyone have any clue as to how to investigate (and hopefully fix) the issue?
Your laptop is probably clogged with dust. You need to take it apart and clean it. I need to clean almost all my machines every year or two.
Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B
Hi,
Thank you for your answer.
Although I cannot rule this possibility out with 100% certainty, I find it unlikely that the problem would appear all of a sudden. And it would not explain the inaccurate fan speed reported by the sensors, either.
It's quite common that it happens all of a sudden.
Either you notice all of a sudden, or it crosses the threshold from not-a-problem into overheating-problem all of a sudden.
In a matter of hours (during which the computer was off), it went from "able to perform heavy-duty computations without becoming abnormally hot under my hands" to "reaching critical temperature while idle". Even taking into account my hands' inaccuracy, that still sounds like a gap.
The inaccuracies are quite common too, I can usually see them even in the BIOS of pristine machines.
But this sensor used to be accurate. I usually don't pay attention to it of course, but since I already had fan issues with this computer, I know that 4000ish rpm is the maximal speed of this fan, and that at this (reported) speed, I should not need any particular effort to hear it spin.
Also, when I say that I can't get any manual control on the fan, this includes *stopping* it... How can dust prevent a fan from stopping on-demand?
Sadly, I don't have the surgical skills required for taking apart and back together this kind of laptops.
Just get a compressed air can (make sure it's for electronics, otherwise it might have incompatible solvents), make sure your laptop is turned off, and blow some air through the cooler vent. Make sure to hold the can correctly (read the instructions) to avoid propellent to go out as well. That should move dust around enough to make things better. And if it does make things better, then you've got confirmation that dust is to blame.
I will try it, though, but I don't know when I will have time to drop by the electronics store, so it may take some time.
Thank you!
T. Verron
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
You could try your machine with the 13.1 or 13.2 live. If it doesn't heat up, it is not dust.
On 2015-04-28 00:08, Thibaut Verron wrote:
Just get a compressed air can (make sure it's for electronics, otherwise it might have incompatible solvents), make sure your laptop is turned off, and blow some air through the cooler vent. Make sure to hold the can correctly (read the instructions) to avoid propellent to go out as well. That should move dust around enough to make things better. And if it does make things better, then you've got confirmation that dust is to blame.
I will try it, though, but I don't know when I will have time to drop by the electronics store, so it may take some time.
Around here (Spain), I can get the can at some convenience stores, that we call here Chinese stores. The cans are cheaper and contains less gas, but it suffices.
The gas is typically butane or propane, so don't use it in a closed room or near fires. No solvents. It is impossible to compress enough air in a light metal can; it has to be a gas that when compressed at room temperature liquifies. And butane is just it, and cheap.
Previously freon was optimal, till it was found how damaging it was to the ozone layer... so it was forbidden.
- -- Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 "Bottle" (Minas Tirith))
On 04/28/2015 02:20 AM, Carlos E. R. wrote:
And butane is just it, and cheap. Previously freon was optimal, till it was found how damaging it was to the ozone layer... so it was forbidden.
__________
- just a remark :: Electrolux swedish vacuum cleaners used to have a plug , under the air-filter , into which plug one could insert a tube to "blow" { instead of suck}
............
regards
2015-04-28 9:37 GMT+02:00 ellanios82 ellanios82@gmail.com:
On 04/28/2015 02:20 AM, Carlos E. R. wrote:
And butane is just it, and cheap. Previously freon was optimal, till it was found how damaging it was to the ozone layer... so it was forbidden.
- just a remark :: Electrolux swedish vacuum cleaners used to have a plug ,
under the air-filter , into which plug one could insert a tube to "blow" { instead of suck}
Don't worry, I know where I can get a can of compressed air, I will just need some time. ;)
I found some more information on the issue, while randomly googling. It turns out that not all archives on my past issue are on a now-deleted distribution forum, I also asked a question on superuser : http://superuser.com/q/446610/119172
So apparently I remembered things wrong, the fan sensor has always been misreporting low speeds as 4090-ish rpm. It also appears that the problem I had 3 years ago was really close to the one I have today, except that back then, pwm1_enable was working as expected (and so I could turn the fan to max speed at start, so that I could investigate the problem without the computer shutting down).
I should have kept a trace of what I did to fix the issue back then... All that I remember was that it involved booting with different acpi_osi settings, maybe tweaking some values in the /sys/class/hwmon folders, maybe also loading and unloading some modules (there were several competing drivers for eeepc acpi)...
Ok, no panic, easy things first, first the dust.
Thibaut
2015-04-28 10:07 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-28 9:37 GMT+02:00 ellanios82 ellanios82@gmail.com:
On 04/28/2015 02:20 AM, Carlos E. R. wrote:
And butane is just it, and cheap. Previously freon was optimal, till it was found how damaging it was to the ozone layer... so it was forbidden.
- just a remark :: Electrolux swedish vacuum cleaners used to have a plug ,
under the air-filter , into which plug one could insert a tube to "blow" { instead of suck}
Don't worry, I know where I can get a can of compressed air, I will just need some time. ;)
I found some more information on the issue, while randomly googling. It turns out that not all archives on my past issue are on a now-deleted distribution forum, I also asked a question on superuser : http://superuser.com/q/446610/119172
So apparently I remembered things wrong, the fan sensor has always been misreporting low speeds as 4090-ish rpm. It also appears that the problem I had 3 years ago was really close to the one I have today, except that back then, pwm1_enable was working as expected (and so I could turn the fan to max speed at start, so that I could investigate the problem without the computer shutting down).
I should have kept a trace of what I did to fix the issue back then... All that I remember was that it involved booting with different acpi_osi settings, maybe tweaking some values in the /sys/class/hwmon folders, maybe also loading and unloading some modules (there were several competing drivers for eeepc acpi)...
Ok, no panic, easy things first, first the dust.
Thibaut
Just one more thing I noticed : the last update before the system stopped working properly included an update to grub2, which is (to the best of my knowledge) one of the packages with the power to break this kind of features.
I couldn't manage to revert to 2.02~beta2.35.1, so instead I reinstalled grub2 from stable (2.02~beta2.20.1.5). It didn't help. But since I have already seen non-trivially reversible changes of behavior of this bios, I must ask : was there any change in the package that could affect acpi features? For reference, the update to grub2 was in snapshot 20150421.
Thibaut
2015-04-28 13:46 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-28 10:07 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-28 9:37 GMT+02:00 ellanios82 ellanios82@gmail.com:
On 04/28/2015 02:20 AM, Carlos E. R. wrote:
And butane is just it, and cheap. Previously freon was optimal, till it was found how damaging it was to the ozone layer... so it was forbidden.
- just a remark :: Electrolux swedish vacuum cleaners used to have a plug ,
under the air-filter , into which plug one could insert a tube to "blow" { instead of suck}
Don't worry, I know where I can get a can of compressed air, I will just need some time. ;)
I found some more information on the issue, while randomly googling. It turns out that not all archives on my past issue are on a now-deleted distribution forum, I also asked a question on superuser : http://superuser.com/q/446610/119172
So apparently I remembered things wrong, the fan sensor has always been misreporting low speeds as 4090-ish rpm. It also appears that the problem I had 3 years ago was really close to the one I have today, except that back then, pwm1_enable was working as expected (and so I could turn the fan to max speed at start, so that I could investigate the problem without the computer shutting down).
I should have kept a trace of what I did to fix the issue back then... All that I remember was that it involved booting with different acpi_osi settings, maybe tweaking some values in the /sys/class/hwmon folders, maybe also loading and unloading some modules (there were several competing drivers for eeepc acpi)...
Ok, no panic, easy things first, first the dust.
Thibaut
Just one more thing I noticed : the last update before the system stopped working properly included an update to grub2, which is (to the best of my knowledge) one of the packages with the power to break this kind of features.
I couldn't manage to revert to 2.02~beta2.35.1, so instead I reinstalled grub2 from stable (2.02~beta2.20.1.5). It didn't help. But since I have already seen non-trivially reversible changes of behavior of this bios, I must ask : was there any change in the package that could affect acpi features? For reference, the update to grub2 was in snapshot 20150421.
Thibaut
So. I bought a can of compressed air, I blowed thoroughly in the input vent, it definitely made some dust go out the output vent (= between the keyboard keys), but the cpu still overheats even when idle.
The fan appears to be able to spin at reasonable speeds, since it reports 3900 RPM at some times. Based on previous experience, if the fan sensor is as accurate as before, this report should be taken as exact. However, this fan speed is only reported for a very short amount of time, as if the cooling policy specified to turn the fan to full speed only between 60 and 62°.
Thibaut
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
On Wed, 29 Apr 2015 13:27:55 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
So. I bought a can of compressed air, I blowed thoroughly in the input vent, it definitely made some dust go out the output vent (= between the keyboard keys), but the cpu still overheats even when idle.
The fan appears to be able to spin at reasonable speeds, since it reports 3900 RPM at some times. Based on previous experience, if the fan sensor is as accurate as before, this report should be taken as exact. However, this fan speed is only reported for a very short amount of time, as if the cooling policy specified to turn the fan to full speed only between 60 and 62°.
Question is not so much if it doesn't overheat anymore but if it takes longer. I'm not using the compressed air as I find it's pretty much impossible to really clean the parts properly with it. It just doesn't generate enough force. Especially in spaces that are enclosed a lot.
It's just easier to take it apart. You can find service manuals that guide your step by step through the disassembly on the notebook's manufacturer's web pages.
- -- Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B
N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩ�
2015-04-29 14:10 GMT+02:00 Jan Matejka JMatejka@suse.cz:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
On Wed, 29 Apr 2015 13:27:55 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
So. I bought a can of compressed air, I blowed thoroughly in the input vent, it definitely made some dust go out the output vent (= between the keyboard keys), but the cpu still overheats even when idle.
The fan appears to be able to spin at reasonable speeds, since it reports 3900 RPM at some times. Based on previous experience, if the fan sensor is as accurate as before, this report should be taken as exact. However, this fan speed is only reported for a very short amount of time, as if the cooling policy specified to turn the fan to full speed only between 60 and 62°.
Question is not so much if it doesn't overheat anymore but if it takes longer. I'm not using the compressed air as I find it's pretty much impossible to really clean the parts properly with it. It just doesn't generate enough force. Especially in spaces that are enclosed a lot.
It's just easier to take it apart. You can find service manuals that guide your step by step through the disassembly on the notebook's manufacturer's web pages.
There is no sign that it overheats any slower. To be 100% sure, I should have measured the time before and after "cleaning", but my rough estimations will have to do.
As for taking the netbook apart, this is a step I will not take. The compressed can thing was easy to test, and I should have had one of these at hand for a long time anyway. Taking the computer apart will take an afternoon, and I have no certainty that I will be able to recreate a thermal contact as good as the current, or even that I will be able to reassemble the netbook at all. The computer is not entirely functional with this thermal issue, but it is still way more useful than a brick.
I know how this whole story looks like, but you have to believe me, I had this kind of issues twice already, and each time it was caused by a software issue (combination of drivers, grub options and bios black magic). Here the problem appeared with a system update, and it has at least one purely software symptom (the fact that I cannot get manual control of the fan, I am not even able to manually stop it). Could we please try to investigate this?
Thibaut
I'm really exited to see this coming "into the light". I'll be awaiting the video when it arrives.
On Wed, Apr 29, 2015 at 7:48 AM, Thibaut Verron thibaut.verron@gmail.com wrote:
2015-04-29 14:10 GMT+02:00 Jan Matejka JMatejka@suse.cz:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
On Wed, 29 Apr 2015 13:27:55 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
So. I bought a can of compressed air, I blowed thoroughly in the input vent, it definitely made some dust go out the output vent (= between the keyboard keys), but the cpu still overheats even when idle.
The fan appears to be able to spin at reasonable speeds, since it reports 3900 RPM at some times. Based on previous experience, if the fan sensor is as accurate as before, this report should be taken as exact. However, this fan speed is only reported for a very short amount of time, as if the cooling policy specified to turn the fan to full speed only between 60 and 62°.
Question is not so much if it doesn't overheat anymore but if it takes longer. I'm not using the compressed air as I find it's pretty much impossible to really clean the parts properly with it. It just doesn't generate enough force. Especially in spaces that are enclosed a lot.
It's just easier to take it apart. You can find service manuals that guide your step by step through the disassembly on the notebook's manufacturer's web pages.
There is no sign that it overheats any slower. To be 100% sure, I should have measured the time before and after "cleaning", but my rough estimations will have to do.
As for taking the netbook apart, this is a step I will not take. The compressed can thing was easy to test, and I should have had one of these at hand for a long time anyway. Taking the computer apart will take an afternoon, and I have no certainty that I will be able to recreate a thermal contact as good as the current, or even that I will be able to reassemble the netbook at all. The computer is not entirely functional with this thermal issue, but it is still way more useful than a brick.
I know how this whole story looks like, but you have to believe me, I had this kind of issues twice already, and each time it was caused by a software issue (combination of drivers, grub options and bios black magic). Here the problem appeared with a system update, and it has at least one purely software symptom (the fact that I cannot get manual control of the fan, I am not even able to manually stop it). Could we please try to investigate this?
Thibaut
To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
On 2015-04-29 14:48, Thibaut Verron wrote:
I know how this whole story looks like, but you have to believe me, I had this kind of issues twice already, and each time it was caused by a software issue (combination of drivers, grub options and bios black magic). Here the problem appeared with a system update, and it has at least one purely software symptom (the fact that I cannot get manual control of the fan, I am not even able to manually stop it). Could we please try to investigate this?
Well, I suggested you try running the live of 13.1 or 13.2 to find out.
- -- Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 "Bottle" (Minas Tirith))
On Wed, Apr 29, 2015 at 9:10 AM, Jan Matejka JMatejka@suse.cz wrote:
On Wed, 29 Apr 2015 13:27:55 +0200 Thibaut Verron thibaut.verron@gmail.com wrote:
So. I bought a can of compressed air, I blowed thoroughly in the input vent, it definitely made some dust go out the output vent (= between the keyboard keys), but the cpu still overheats even when idle.
The fan appears to be able to spin at reasonable speeds, since it reports 3900 RPM at some times. Based on previous experience, if the fan sensor is as accurate as before, this report should be taken as exact. However, this fan speed is only reported for a very short amount of time, as if the cooling policy specified to turn the fan to full speed only between 60 and 62°.
Question is not so much if it doesn't overheat anymore but if it takes longer. I'm not using the compressed air as I find it's pretty much impossible to really clean the parts properly with it. It just doesn't generate enough force. Especially in spaces that are enclosed a lot.
I've had issues like the one described in the OP and blowing compressed air helps a lot.
Of course it does depend on how tightly packed the components are inside the laptop, in my case not very tight, but it should have made a bigger difference.
I'd guess the reported experience does put the software (BIOS probably) in a very suspicious light.
Thing is, if you can't get it to work right in Windows even, there's little chance of finding a software fix for it. Most BIOS firmwares are quite tuned for windows, and if windows is also failing it does show a very pervasive bug.
If I were you, I'd report that with your vendor. Seems like a hardware or firmware issue that they should be able or at least try to resolve. Even if it's just dust that the compressed air didn't unclog, they'd be able to disassemble and properly clean it.
Am 29.04.2015 um 17:46 schrieb Claudio Freire:
If I were you, I'd report that with your vendor. Seems like a hardware or firmware issue that they should be able or at least try to resolve. Even if it's just dust that the compressed air didn't unclog, they'd be able to disassemble and properly clean it.
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
The machine was working before. After a software update (I'd guess it is a kernel problem), it is not working anymore.
The heatsink has been cleaned (at least partially), and it did not help at all.
Can we now please start investigating the software issue and stop blaming Thibaut?
Thibaut: the 13.2-livemedium test would really be useful, just to convince the infidels that it *really* is a software issue :-)
I believe you, I have seen such things before.
And no, claiming "but the BIOS is broken" will not help. I don't think ASUS is going to fix their crap.
Am 29.04.2015 um 22:01 schrieb Stefan Seyfried:
And no, claiming "but the BIOS is broken" will not help. I don't think ASUS is going to fix their crap.
I forgot: Thibaut, once you confirmed that it works with 13.2 but fails with factory/tumbleweed, proceed directly to the opensuse-kernel list, because that's the place where the people are that are more likely to actually help you.
Can we now please start investigating the software issue and stop
blaming Thibaut?
No offense taken, it was really the first step to take. But I was somewhat afraid that past the mandatory "have you turned it off and on again" (oops, sorry, "please take out the dust") the thread would die a quick death. Thank you very much for your support.
Thibaut: the 13.2-livemedium test would really be useful, just to
convince the infidels that it *really* is a software issue :-)
I am downloading it right now, for some reason I missed it in the first mail from Carlos, sorry.
I forgot: Thibaut, once you confirmed that it works with 13.2 but fails with factory/tumbleweed, proceed directly to the opensuse-kernel list, because that's the place where the people are that are more likely to actually help you.
I like your confidence. :) But thanks for the pointers!
Thibaut
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
- -- Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 "Bottle" (Minas Tirith))
On Thu, 2015-04-30 at 00:47 +0200, Carlos E. R. wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's what I got too, problem is global, but the magnitude changed.
-Mike
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
On Thu, Apr 30, 2015 at 1:32 AM, Claudio Freire klaussfreire@gmail.com wrote:
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
But sure, testing a live media can't hurt.
2015-04-30 6:35 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Thu, Apr 30, 2015 at 1:32 AM, Claudio Freire klaussfreire@gmail.com wrote:
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
But sure, testing a live media can't hurt.
I share your doubts on that, but the live medium comes with its own kernel and its own bootloader, so I figure that there is a very slight chance that it shuffles more parameters than other kernels sharing the same bootloader.
I guess that the goal for now is precisely to get out of this situation where we can not test fixes: if the live USB can "fix" whatever is saved wrong in the BIOS, then even if the fan still does not work properly with tumbleweed, there will be a good chance that it will work for other distros and for windows. And then we will have something to test against. I understand that it is wishful thinking for now, but it is this kind of black magic that has already mysteriously solved this issue twice, so... why not?
As for contacting ASUS, I already did that 3 years ago when I had this issue for the first time. They didn't try to burn me or get me to reinstall windows, but (as could be predicted) they could not give any better advice than to bring the computer back to a repair shop. The initial cost (of evaluating the issue) would have been 70€, for a (then) 1-year-old 200€ computer. It was not worth it 3 years ago, and it's even less worth it now.
And yes, I could buy a similar computer in the same price range today, but this piece of hardware has proved reliable (yes...) and useful over the years, way beyond what its pricetag could suggest, and I'm not sure that this is a reproducible experience. ;-)
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
I'll keep you guys updated.
Thanks again all for your advice.
Thibaut
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
On 2015-04-30 10:43, Thibaut Verron wrote:
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
I place my laptop on a... "thing" with a largish fan inside, which can plug to one of the USB sockets of the laptop, or to a mains-usb converter (I do the later). It keeps the base of the laptop cool.
The internal fan cools well only the CPU. Ram and HD can get quite hot.
If I can't do this, I place the laptop on 1 centimeter pegs, not just on a table, so that air can circulate.
- -- Cheers / Saludos,
Carlos E. R.
(from 13.1 x86_64 "Bottle" (Minas Tirith))
On Thu, 2015-04-30 at 14:56 +0200, Carlos E. R. wrote:
On 2015-04-30 10:43, Thibaut Verron wrote:
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
I place my laptop on a... "thing" with a largish fan inside, which can plug to one of the USB sockets of the laptop, or to a mains-usb converter (I do the later). It keeps the base of the laptop cool.
The internal fan cools well only the CPU. Ram and HD can get quite hot.
If I can't do this, I place the laptop on 1 centimeter pegs, not just on a table, so that air can circulate.
Hm, I didn't make my desktop box, I bought it :) Much more powerful than lappy on stilts, and easier to get at the dust bunnies inside it.
-Mike
2015-04-30 10:43 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-30 6:35 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Thu, Apr 30, 2015 at 1:32 AM, Claudio Freire klaussfreire@gmail.com wrote:
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
But sure, testing a live media can't hurt.
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
Good news, I think! The live USB did not seem to work: it still could not get it to boot, and even if I start it with a hot (>80°C) CPU, the fan doesn't seem to step up.
But. I landed on a list of kernel flags, and I tried a couple of them out... just to see? Anyway, it turns out that if I boot my tumbleweed system with "thermal.off=1", I lose the main temperature sensor, and I *can* force the fan to full speed with pwm1_enable=1. The fan speed reading is not any useful, but I can hear that the fan is notably faster. And the remaining temperature readings are stabilized.
So, the fan is working at least.
After seeing all that, I rebooted (with the default boot flags this time), pwm1_enable was still set to 1 (before that, it was always 2 after a reboot). Again, I could control the fan (stop it or set it to full speed), and this time the fan speed readings seem to be accurate (at least for high speeds). I have now set pwm1_enable to 2 (automatic), I will see how the temperature is in 30-60 minutes.
Thibaut
On Thu, Apr 30, 2015 at 5:23 PM, Thibaut Verron thibaut.verron@gmail.com wrote:
2015-04-30 10:43 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-30 6:35 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Thu, Apr 30, 2015 at 1:32 AM, Claudio Freire klaussfreire@gmail.com wrote:
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
This is not useful.
No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
But sure, testing a live media can't hurt.
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
Good news, I think! The live USB did not seem to work: it still could not get it to boot, and even if I start it with a hot (>80°C) CPU, the fan doesn't seem to step up.
But. I landed on a list of kernel flags, and I tried a couple of them out... just to see? Anyway, it turns out that if I boot my tumbleweed system with "thermal.off=1", I lose the main temperature sensor, and I *can* force the fan to full speed with pwm1_enable=1. The fan speed reading is not any useful, but I can hear that the fan is notably faster. And the remaining temperature readings are stabilized.
So, the fan is working at least.
After seeing all that, I rebooted (with the default boot flags this time), pwm1_enable was still set to 1 (before that, it was always 2 after a reboot). Again, I could control the fan (stop it or set it to full speed), and this time the fan speed readings seem to be accurate (at least for high speeds). I have now set pwm1_enable to 2 (automatic), I will see how the temperature is in 30-60 minutes.
Well, you've proven it's a kernel thing. Or a kernel-BIOS interaction.
So I'd suggest trying to get help from the kernel ML mentioned upthread, where the wise guys are.
Ok, thanks for the help guys!
Thibaut
2015-04-30 22:52 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Thu, Apr 30, 2015 at 5:23 PM, Thibaut Verron thibaut.verron@gmail.com wrote:
2015-04-30 10:43 GMT+02:00 Thibaut Verron thibaut.verron@gmail.com:
2015-04-30 6:35 GMT+02:00 Claudio Freire klaussfreire@gmail.com:
On Thu, Apr 30, 2015 at 1:32 AM, Claudio Freire klaussfreire@gmail.com wrote:
On Wed, Apr 29, 2015 at 7:47 PM, Carlos E. R. robin.listas@telefonica.net wrote:
On 2015-04-29 22:01, Stefan Seyfried wrote:
> This is not useful. > > No fscking way is ASUS going to fix their BIOS for Linux users.
I understand he said that Windows had the same problem.
That's why I think it unlikely treating it as a software bug will help.
Even if it is.
There's no way to test fixes, because supposedly working kernels (windows) still exhibit the issue.
But sure, testing a live media can't hurt.
Just an update on the tests you asked me to run: until now I could not get the live USB to boot (it hangs at some point), and I can't let it heat idle to see what happens, because it reboots after a while. The fan does sound a bit louder, but this is only my inaccurate ear measuring. I had to go to work, I will try troubleshooting why the boot fails this evening, and if I still cannot boot, I'll try to start the live USB with a hot CPU, to see if the fan goes berserk as it should.
Good news, I think! The live USB did not seem to work: it still could not get it to boot, and even if I start it with a hot (>80°C) CPU, the fan doesn't seem to step up.
But. I landed on a list of kernel flags, and I tried a couple of them out... just to see? Anyway, it turns out that if I boot my tumbleweed system with "thermal.off=1", I lose the main temperature sensor, and I *can* force the fan to full speed with pwm1_enable=1. The fan speed reading is not any useful, but I can hear that the fan is notably faster. And the remaining temperature readings are stabilized.
So, the fan is working at least.
After seeing all that, I rebooted (with the default boot flags this time), pwm1_enable was still set to 1 (before that, it was always 2 after a reboot). Again, I could control the fan (stop it or set it to full speed), and this time the fan speed readings seem to be accurate (at least for high speeds). I have now set pwm1_enable to 2 (automatic), I will see how the temperature is in 30-60 minutes.
Well, you've proven it's a kernel thing. Or a kernel-BIOS interaction.
So I'd suggest trying to get help from the kernel ML mentioned upthread, where the wise guys are.
On Sat, 2015-04-25 at 10:47 +0200, Thibaut Verron wrote:
Booting on a previous kernel did not help. If I boot on a different distribution or on windows, the OS does not see the fan speed sensor, and the computer overheats slowly all the same. But as I said, I already had similar problems with this computer (the fan was not spinning at all, then), and I remember that it showed across distributions in the same way. If I remember, I concluded back then that something had set a bios flag that prevented the bios from handling the fan properly for other OSes.
Hm. On top of what others have said about crud maybe plugging things up, this paragraph suggests to me that your laptop may have crap for bios, that you should perhaps look for an update.
-Mike
2015-04-28 9:19 GMT+02:00 Mike Galbraith umgwanakikbuti@gmail.com:
On Sat, 2015-04-25 at 10:47 +0200, Thibaut Verron wrote:
Booting on a previous kernel did not help. If I boot on a different distribution or on windows, the OS does not see the fan speed sensor, and the computer overheats slowly all the same. But as I said, I already had similar problems with this computer (the fan was not spinning at all, then), and I remember that it showed across distributions in the same way. If I remember, I concluded back then that something had set a bios flag that prevented the bios from handling the fan properly for other OSes.
Hm. On top of what others have said about crud maybe plugging things up, this paragraph suggests to me that your laptop may have crap for bios, that you should perhaps look for an update.
-Mike
Indeed that's something I hadn't thought of. But I updated that bios the first time I had an issue with the fans (back then, it didn't help, or at least not directly), and there has been no new version since then.
Thibaut
-
Carlos E. R. -
Claudio Freire -
ellanios82 -
Jan Matejka -
Mike Galbraith -
Mike Galbraith -
Robert Martens -
Stefan Seyfried -
Thibaut Verron