Fan always on with 3.18.2-2

marcoc · 2015-02-02 12:47:34

Kernel 3.18.5-1 fixed the issue on my Compaq 6910p.

No. Sorry, I was wrong.
After a reboot, the fan is always on again.

I downgrade to 3.17.6....

cardiel · 2015-02-03 21:03:33

I'm having the same problem. The CPU will also get 7 celsius hotter than before, probably why the fan is spinning more than it used to before.
I use LTS kernel instead, low temp and no fan noise

Last edited by cardiel (2015-02-03 21:18:09)

xtian · 2015-02-05 15:02:35

This is the first time I've needed to downgrade any packages using Arch. (>_<)

The Downgrading Package wiki page suggests to downgrade `linux-headers` and `*module*` packages at the same time. Is this necessary?

Then I should do this?
linux-3.18.2-2-x86_64.pkg.tar.xz --> linux-3.17.6-1-x86_64.pkg.tar.xz
linux-api-headers-3.18.1-1-x86_64.pkg.tar.xz --> linux-api-headers-3.16.2-1-x86_64.pkg.tar.xz

prash · 2015-02-05 15:24:25

@xtian

A simpler option is to not downgrade anything at all. Just install the package called "linux-lts". Installing both, the latest kernel, and the LTS kernel has a couple of advantages. First, you won't have to worry about "pacman -Syu" clobbering your preferred kernel, and second, "linux" and "linux-lts" are updated independent of each other, with separate options during grub, it's easy to test the latest kernel because you have it installed all the time.

You just need to update grub.cfg to either select your preferred kernel by default, or to select the previously booted kernel. IMO, the latter option makes things simplest.

RobyWanKenobi · 2015-02-05 16:25:18

I confirm that on HP-620 laptop with 3.18.5-1 the issue (fan always on at full speed) is still there.
I also suggest what prash already said: install both "linux" and "linux-lts", choosing the latter one as the "default" kernel in grub.cfg.
It's a time saver, compared the upgrade / test / downgrade solution...

sluckxz · 2015-02-07 02:19:08

For me the fan is full speed on linux-3.18.1-1-x86_64 as well!
Reverted to 3.17.6.-1 troubleshooting (just found this thread) and will probably go LTS soon.
Probook 4510s

skanky · 2015-02-09 09:11:23

Just for info: 3.18.6-1 - no change in behaviour.

Last edited by skanky (2015-02-09 09:11:41)

prash · 2015-02-09 19:15:31

I had the time to test 3.18.1. The fan blasts at full speed. I saw that 3.19 was released on kernel.org, so I installed linux-git from AUR. The fans still blasts so:

% acpi -tc
Thermal 0: ok, 90.0 degrees C
Thermal 1: ok, 23.3 degrees C
Thermal 2: ok, 36.0 degrees C
Thermal 3: ok, 34.0 degrees C
Thermal 4: ok, 37.0 degrees C
Thermal 5: ok, 16.0 degrees C
Cooling 0: LCD 0 of 24
Cooling 1: Fan 1 of 1
Cooling 2: Fan 1 of 1
Cooling 3: Fan 1 of 1
Cooling 4: Fan 1 of 1
Cooling 5: Fan 1 of 1
Cooling 6: Fan 1 of 1
Cooling 7: Fan 1 of 1
Cooling 8: Fan 1 of 1
Cooling 9: Fan 1 of 1
Cooling 10: Fan 1 of 1
Cooling 11: Fan 1 of 1
Cooling 12: Fan 1 of 1
Cooling 13: Fan 0 of 1
Cooling 14: Processor 0 of 10
Cooling 15: Processor 0 of 10

On my system, Thermal 0, which corresponds to FDTZ and thermal_zone5, refers to the fan speed.

clfarron4 · 2015-02-09 20:45:20

Well. that slashes any hope I had for 3.19 being a fix for this problem. It'd be great if you post the output and I'll edit the affected versions to indicate that 3.19 is still affected.

As a reminder to newcomers to this thread, this is the upstream report and temperature information (from commands such as acpi -tc ) should be posted there.

Cooleech · 2015-02-10 07:56:24

On bugzilla.kernel.org it says problem started on 3.18.2, but i think it begins with 3.18.0. Can I get a link to that kernel version so that I can test it on my HP 6730s? It seems only Probooks have the same problem as we do with 6730s series.

amish · 2015-02-10 10:55:59

You can download from here:
http://seblu.net/a/arm/packages/l/linux/

3.18.0 was never released for Arch linux. It begins with 3.18.1

clfarron4 · 2015-02-10 12:21:41

Cooleech wrote:

On bugzilla.kernel.org it says problem started on 3.18.2, but i think it begins with 3.18.0. Can I get a link to that kernel version so that I can test it on my HP 6730s? It seems only Probooks have the same problem as we do with 6730s series.

Well fudge... Does anyone want me to build 3.18.0 with the same kernel config as the [core] kernel?

clfarron4 · 2015-02-10 20:08:32

OK, so I built 3.18 with the official Arch config. This is because upstream wants to know if it's all of 3.18 that is affected or starting from a particular version of 3.18

clfarron4 wrote:

OK, to see whether 3.18 solves this problem without having to enable [testing] (in all it's glory), I've setup the following repository:
[kernel-testing]
Server = http://dl.dropbox.com/u/298301785/arch/kernel-testing/$arch
To use it, you'll need to put it ABOVE [core] in /etc/pacman.conf and import my PGP key for verification of the repository and the packages in it.
NOTE: This is just the kernel. You will have to build any modules and other drivers yourself.

Last edited by clfarron4 (2015-02-10 20:15:39)

rpodgorny · 2015-02-10 20:19:20

clfarron4 wrote:

OK, so I built 3.18 with the official Arch config. This is because upstream wants to know if it's all of 3.18 that is affected or starting from a particular version of 3.18
clfarron4 wrote:
OK, to see whether 3.18 solves this problem without having to enable [testing] (in all it's glory), I've setup the following repository:
[kernel-testing]
Server = http://dl.dropbox.com/u/298301785/arch/kernel-testing/$arch
To use it, you'll need to put it ABOVE [core] in /etc/pacman.conf and import my PGP key for verification of the repository and the packages in it.
NOTE: This is just the kernel. You will have to build any modules and other drivers yourself.

...and what's the result for you? i've built it as well but i am unable to get consistent results. each boot is different! :-(

to be more precise, it seems it works like this: if the cpu temp is low during the boot, the fan is quiet and stays that way even when i heat up the cpu. when the cpu is hot during the boot, the fan is full on and stays that way even after the cpu cools down. this holds true for 3.18.0. 3.18.1 seems to be full on even when the cpu is cool.

not certain, thou. it's so hard to repeat the results...

prash · 2015-02-10 20:41:12

@clfarron4: Thank you for making 3.18.0 available so quickly for testing. The fan goes full-blast in this one too, as I have reported on the kernel bugzilla.

Apart from the issue with the fan, there is another problem on 3.18.0 and 3.18.1. The video appears torn, but only in the top quarter of the screen.

I mention this only because it may be relevant sometime in the future, if the problem were to show up again. Anyway, 3.18.2 fixed the video problem.

amish · 2015-02-11 00:42:55

Yes, I had already reported that 3.18.1 had weird graphical issue. Dont know if it is related. But it was fixed in 3.18.2

clfarron4 · 2015-02-11 09:13:33

prash wrote:

@clfarron4: Thank you for making 3.18.0 available so quickly for testing. The fan goes full-blast in this one too, as I have reported on the kernel bugzilla.

Thanks!

prash wrote:

Apart from the issue with the fan, there is another problem on 3.18.0 and 3.18.1. The video appears torn, but only in the top quarter of the screen.
[snip]
I mention this only because it may be relevant sometime in the future, if the problem were to show up again. Anyway, 3.18.2 fixed the video problem.

That will have been why we had to use a patch for drm/i915: Prevent negative relocation deltas from wrapping in 3.18.2 and 3

skanky · 2015-02-11 09:35:43

rpodgorny wrote:

...and what's the result for you? i've built it as well but i am unable to get consistent results. each boot is different! :-(
to be more precise, it seems it works like this: if the cpu temp is low during the boot, the fan is quiet and stays that way even when i heat up the cpu. when the cpu is hot during the boot, the fan is full on and stays that way even after the cpu cools down. this holds true for 3.18.0. 3.18.1 seems to be full on even when the cpu is cool.
not certain, thou. it's so hard to repeat the results...

This sounds like what I get - and may explain why I don't get the issue every boot (though with most).

It normally seems to coincide with the systemd message about flushing the journal to persistent storage. The operation gets the Cylon/KNightirder eyes treatment for a second or few. My guess would be that that's possibly what's causing the initial heat up and is nothing to do with the actual issue.

amish · 2015-02-13 05:45:39

One more similar looking bug which also reports that issue occurs after 3.18.x kernel but does not occur with 3.17

https://bugzilla.kernel.org/show_bug.cgi?id=91411

Someone has identified bad commit from git bisect (I have no idea what it is!)

@clfarron4 - is it possible for you to UNDO that commit and release kernel for testing? Thanks,

clfarron4 · 2015-02-13 16:30:35

amish wrote:

One more similar looking bug which also reports that issue occurs after 3.18.x kernel but does not occur with 3.17
https://bugzilla.kernel.org/show_bug.cgi?id=91411
Someone has identified bad commit from git bisect (I have no idea what it is!)
@clfarron4 - is it possible for you to UNDO that commit and release kernel for testing? Thanks,

Hi, it should be possible for me to undo it. Whether anything works is a different matter, but there's no harm trying.

Also, I have no experience using Git or anything like that, so I'm going to struggle right from the off.

Last edited by clfarron4 (2015-02-13 16:44:41)

amish · 2015-02-13 16:37:33

Thanks. I can try as long as its not going to cause any harm to hardware OR file system!

clfarron4 · 2015-02-13 17:32:23

clfarron4 wrote:

Hi, it should be possible for me to undo it. Whether anything works is a different matter, but there's no harm trying.
Also, I have no experience using Git or anything like that, so I'm going to struggle right from the off.

I just noticed the guy is doing the git bisect from 3.17 to 3.18 ¬¬. So essentially, you're running 3.17 with a few of the patches which would take you to 3.18; This is explains why un-doing the commit from isn't working for me. I'll get onto this tomorrow, as I've been battling with git for quite a bit now and I need to attend to other stuff IRL.

amish · 2015-02-14 03:28:16

Thanks. I was studying git bisect, it appears to be very tedious job.

Pick a commit, compile a kernel, test it. Declare it good or bad.

Then again pick next one ... so on.

Since you dont have the issue, for you, git bisect will not help. Unless you compile a kernel, put it on testing repo and then we test and report back.

So its going to be very lengthy process (may take days and days for 15-20 commits).

Probably its not worth?

My only hope is that some kernel developer looks at it.

I requested them politely to look at it urgently but one guy did not reply positively. (dont know if he is kernel developer or not)

Lets see - hoping for the best.

amish · 2015-02-15 14:52:06

Just wanted to correct my previous report that bug did not exist in 3.18.1 for me.

I tried again with 3.18.1 (from rollback machine) as well as 3.18.0 (provided by clfarron4).

In both cases FAN ran at full speed after booting.

After GRUB load and for 10-20 seconds FAN was normal and then suddenly goes to full speed.

JJK · 2015-02-18 14:59:16

I have something to add to this thread, I have exactly the same problem on my machine (HP probook 4510s) which I think is the same one that you're using @amish.

I can confirm that the problem doesn't exist in the LTS kernel.
However, I tried running the latest arch-linux live cd this morning and it doesn't have the problem even though it is running linux 3.18.4.

Can anyone confirm this?

JJK

Last edited by JJK (2015-02-18 14:59:27)

Arch Linux

#76 2015-02-02 12:47:34

Re: Fan always on with 3.18.2-2

#77 2015-02-03 21:03:33

Re: Fan always on with 3.18.2-2

#78 2015-02-05 15:02:35

Re: Fan always on with 3.18.2-2

#79 2015-02-05 15:24:25

Re: Fan always on with 3.18.2-2

#80 2015-02-05 16:25:18

Re: Fan always on with 3.18.2-2

#81 2015-02-07 02:19:08

Re: Fan always on with 3.18.2-2

#82 2015-02-09 09:11:23

Re: Fan always on with 3.18.2-2

#83 2015-02-09 19:15:31

Re: Fan always on with 3.18.2-2

#84 2015-02-09 20:45:20

Re: Fan always on with 3.18.2-2

#85 2015-02-10 07:56:24

Re: Fan always on with 3.18.2-2

#86 2015-02-10 10:55:59

Re: Fan always on with 3.18.2-2

#87 2015-02-10 12:21:41

Re: Fan always on with 3.18.2-2

#88 2015-02-10 20:08:32

Re: Fan always on with 3.18.2-2

#89 2015-02-10 20:19:20

Re: Fan always on with 3.18.2-2

#90 2015-02-10 20:41:12

Re: Fan always on with 3.18.2-2

#91 2015-02-11 00:42:55

Re: Fan always on with 3.18.2-2

#92 2015-02-11 09:13:33

Re: Fan always on with 3.18.2-2

#93 2015-02-11 09:35:43

Re: Fan always on with 3.18.2-2

#94 2015-02-13 05:45:39

Re: Fan always on with 3.18.2-2

#95 2015-02-13 16:30:35

Re: Fan always on with 3.18.2-2

#96 2015-02-13 16:37:33

Re: Fan always on with 3.18.2-2

#97 2015-02-13 17:32:23

Re: Fan always on with 3.18.2-2

#98 2015-02-14 03:28:16

Re: Fan always on with 3.18.2-2

#99 2015-02-15 14:52:06

Re: Fan always on with 3.18.2-2

#100 2015-02-18 14:59:16

Re: Fan always on with 3.18.2-2

Board footer