Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1515073 > unrolled thread
| Started by | Pavel Machek <pavel@ucw.cz> |
|---|---|
| First post | 2016-11-04 09:40 +0100 |
| Last post | 2016-11-05 13:30 +0100 |
| Articles | 20 on this page of 30 — 5 participants |
Back to article view | Back to linux.kernel
v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-04 09:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-04 10:00 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Viresh Kumar <viresh.kumar@linaro.org> - 2016-11-04 10:20 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-04 10:30 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Viresh Kumar <viresh.kumar@linaro.org> - 2016-11-04 10:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Henrique de Moraes Holschuh <hmh@hmh.eng.br> - 2016-11-05 19:30 +0100
Thinkpad power management (was Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build) Pavel Machek <pavel@ucw.cz> - 2016-11-05 20:30 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-04 15:10 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-04 21:50 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-04 22:20 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 09:50 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Henrique de Moraes Holschuh <hmh@hmh.eng.br> - 2016-11-05 18:50 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 20:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Henrique de Moraes Holschuh <hmh@hmh.eng.br> - 2016-11-06 04:50 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-04 23:20 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-05 00:30 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 14:30 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-05 14:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 15:00 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 15:10 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-05 15:20 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 16:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> - 2016-11-05 15:00 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 21:40 +0100
thinkpad x60, T40p: overheat with v4.9-rc4 (was Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build) Pavel Machek <pavel@ucw.cz> - 2016-11-09 12:40 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Henrique de Moraes Holschuh <hmh@hmh.eng.br> - 2016-11-05 19:10 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 21:00 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Zhang Rui <rui.zhang@intel.com> - 2016-11-05 12:30 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 14:20 +0100
Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build Pavel Machek <pavel@ucw.cz> - 2016-11-05 13:30 +0100
Page 1 of 2 [1] 2 Next page →
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-04 09:40 +0100 |
| Subject | v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szIcx-7IN-1@gated-at.bofh.it> |
[Multipart message — attachments visible in raw view] — view raw
Hi! I'm debugging overheats on v4.9-rc1... which did not seem to happen in v4.8-rc1. I'm running basically "nice make -j 3" on kernel... cpus are fully loaded. %Cpu(s): 7.5 us, 18.5 sy, 72.6 ni, 0.0 id, 0.0 wa, 0.0 hi, 1.5 si, 0.0 st KiB Mem: 3087096 total, 2993076 used, 94020 free, 52900 buffers KiB Swap: 1681428 total, 60900 used, 1620528 free. 1183664 cached Mem Still, cpus don't stay on maximum frequency on v4.8-rc1. (I suspect that may be why machine does not overheat). pavel@duo:/data/l/linux$ sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq 1833000 1833000 1000000 1000000 1000000 1000000 1833000 1833000 1833000 1833000 1000000 1000000 pavel@duo:/data/l/linux pavel@duo:/data/l/linux$ grep -i . /sys/devices/system/cpu/cpu0/cpufreq/* /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0 /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1833000 grep: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq: Permission denied /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:1833000 /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:1000000 /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000 /sys/devices/system/cpu/cpu0/cpufreq/freqdomain_cpus:0 1 /sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0 /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:1833000 1333000 1000000 /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:conservative powersave schedutil ondemand performance /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1833000 /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:ondemand /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:1833000 /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:1000000 /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported> grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory Is that expected behaviour? Any ideas? Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-04 10:00 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szIvU-7Ph-15@gated-at.bofh.it> |
| In reply to | #1515073 |
[Multipart message — attachments visible in raw view] — view raw
On Fri 2016-11-04 09:38:49, Pavel Machek wrote:
> Hi!
>
> I'm debugging overheats on v4.9-rc1... which did not seem to happen in
> v4.8-rc1. I'm running basically "nice make -j 3" on kernel... cpus are
> fully loaded.
>
> %Cpu(s): 7.5 us, 18.5 sy, 72.6 ni, 0.0 id, 0.0 wa, 0.0 hi, 1.5
> si, 0.0 st
> KiB Mem: 3087096 total, 2993076 used, 94020 free, 52900
> buffers
> KiB Swap: 1681428 total, 60900 used, 1620528 free. 1183664
> cached Mem
>
> Still, cpus don't stay on maximum frequency on v4.8-rc1. (I suspect
> that may be why machine does not overheat).
What is worse, they go to low frequency even with "performance"
governor on v4.8-rc1?!
pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq
1000000
1000000
1000000
1000000
1833000
1833000
1000000
1000000
1833000
1833000
1000000
1000000
pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ grep -i
. /sys/devices/system/cpu/cpu0/cpufreq/*
/sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000
grep: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq:
Permission denied
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:1833000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:1000000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000
/sys/devices/system/cpu/cpu0/cpufreq/freqdomain_cpus:0 1
/sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:1833000
1333000 1000000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:conservative
powersave schedutil ondemand performance
/sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1000000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:performance
/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:1000000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:1000000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory
pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$
Let me try v4.9-rc2... that works ok (cpus at the high frequency
during the kernel build). Unfortunately that sends my cpus to 99C
temperature range (and eventually forces emergency shutdown).
v4.9-rc2, current policy changes without me touching it. Notice the
1.47GHz below? I did not do that, it oscilates itself. Is that thermal
protection?
analyzing CPU 1:
driver: acpi-cpufreq
CPUs which run at the same hardware frequency: 1
CPUs which need to have their frequency coordinated by software: 1
maximum transition latency: 10.0 us.
hardware limits: 1000 MHz - 1.83 GHz
available frequency steps: 1.83 GHz, 1.33 GHz, 1000 MHz
available cpufreq governors: conservative, powersave, ondemand, performance, schedutil
current policy: frequency should be within 1000 MHz and 1.47 GHz.
The governor "ondemand" may decide which speed to use
within this range.
current CPU frequency is 1.33 GHz.
cpufreq stats: 1.83 GHz:38.24%, 1.33 GHz:34.83%, 1000 MHz:26.94% (9929)
dmesg is full of
[ 716.196096] thinkpad_acpi: THERMAL EMERGENCY: a sensor reports
something is extremely hot!
[ 716.198181] thinkpad_acpi: temperatures (Celsius): 97 52 N/A 87 37
N/A 37 N/A 53 62 N/A N/A N/A N/A N/A N/A
but no information about thermal throttling...
Any ideas?
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Viresh Kumar <viresh.kumar@linaro.org> |
|---|---|
| Date | 2016-11-04 10:20 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szIPg-8bj-23@gated-at.bofh.it> |
| In reply to | #1515078 |
Hi Pavel,
I am really confused about where the problem is. 4.8 or 4.9 ? :)
On 04-11-16, 09:58, Pavel Machek wrote:
> On Fri 2016-11-04 09:38:49, Pavel Machek wrote:
> > Hi!
> >
> > I'm debugging overheats on v4.9-rc1... which did not seem to happen in
> > v4.8-rc1. I'm running basically "nice make -j 3" on kernel... cpus are
> > fully loaded.
> >
> > %Cpu(s): 7.5 us, 18.5 sy, 72.6 ni, 0.0 id, 0.0 wa, 0.0 hi, 1.5
> > si, 0.0 st
> > KiB Mem: 3087096 total, 2993076 used, 94020 free, 52900
> > buffers
> > KiB Swap: 1681428 total, 60900 used, 1620528 free. 1183664
> > cached Mem
> >
> > Still, cpus don't stay on maximum frequency on v4.8-rc1. (I suspect
> > that may be why machine does not overheat).
>
> What is worse, they go to low frequency even with "performance"
> governor on v4.8-rc1?!
You sure about it? How did you check it?
Also why are you testing on 4.8-rc1? And not a 4.8 stable kernel? What if the
core is already fixed upstream ?
There is one core fix in 4.8:
commit 899bb6642f2a ("cpufreq: skip invalid entries when searching the
frequency")
> pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat
> /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq
> 1000000
> 1000000
> 1000000
> 1000000
> 1833000
> 1833000
> 1000000
> 1000000
> 1833000
> 1833000
> 1000000
> 1000000
Is this happening because of thermal capping ? That is the only reason that I
could think of where freq can change with performance governor.
> pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ grep -i
> . /sys/devices/system/cpu/cpu0/cpufreq/*
> /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0
> /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000
> grep: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq:
> Permission denied
> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:1833000
> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:1000000
> /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000
> /sys/devices/system/cpu/cpu0/cpufreq/freqdomain_cpus:0 1
> /sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:1833000
> 1333000 1000000
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:conservative
> powersave schedutil ondemand performance
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1000000
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:performance
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:1000000
And this value sort of confirms it.
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:1000000
> /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
> grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory
> pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$
>
> Let me try v4.9-rc2... that works ok (cpus at the high frequency
> during the kernel build). Unfortunately that sends my cpus to 99C
> temperature range (and eventually forces emergency shutdown).
Unbelievable.
> v4.9-rc2, current policy changes without me touching it. Notice the
> 1.47GHz below? I did not do that, it oscilates itself. Is that thermal
> protection?
Looks like to me.
Can we verify somehow about what's the situation should look like? Perhaps with
some older stable kernel? And then see if 4.8.X works fine or 4.9-rc.
--
viresh
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-04 10:30 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szIYV-8eL-17@gated-at.bofh.it> |
| In reply to | #1515085 |
[Multipart message — attachments visible in raw view] — view raw
Hi!
> I am really confused about where the problem is. 4.8 or 4.9 ? :)
Well, v4.8 runs at too low frequency without explanation, and v4.9
overheats. Both are a problem :-). But it starts to look like v4.9 is
the one where the real problem is.
> On 04-11-16, 09:58, Pavel Machek wrote:
> > On Fri 2016-11-04 09:38:49, Pavel Machek wrote:
> > > Hi!
> > >
> > > I'm debugging overheats on v4.9-rc1... which did not seem to happen in
> > > v4.8-rc1. I'm running basically "nice make -j 3" on kernel... cpus are
> > > fully loaded.
> > >
> > > %Cpu(s): 7.5 us, 18.5 sy, 72.6 ni, 0.0 id, 0.0 wa, 0.0 hi, 1.5
> > > si, 0.0 st
> > > KiB Mem: 3087096 total, 2993076 used, 94020 free, 52900
> > > buffers
> > > KiB Swap: 1681428 total, 60900 used, 1620528 free. 1183664
> > > cached Mem
> > >
> > > Still, cpus don't stay on maximum frequency on v4.8-rc1. (I suspect
> > > that may be why machine does not overheat).
> >
> > What is worse, they go to low frequency even with "performance"
> > governor on v4.8-rc1?!
>
> You sure about it? How did you check it?
See the dumps below. cpuinfo_cur_freq shows 1GHz while compilation is running.
> Also why are you testing on 4.8-rc1? And not a 4.8 stable kernel? What if the
> core is already fixed upstream ?
>
> There is one core fix in 4.8:
>
> commit 899bb6642f2a ("cpufreq: skip invalid entries when searching the
> frequency")
Ok, I guess that's not it.
> > 1000000
> > 1000000
>
> Is this happening because of thermal capping ? That is the only reason that I
> could think of where freq can change with performance governor.
How would I know if it is thermal capping? There's nothing in dmesg.
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1000000
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:performance
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:1000000
>
> And this value sort of confirms it.
>
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:1000000
> > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
> > grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory
> > pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$
> >
> > Let me try v4.9-rc2... that works ok (cpus at the high frequency
> > during the kernel build). Unfortunately that sends my cpus to 99C
> > temperature range (and eventually forces emergency shutdown).
>
> Unbelievable.
>
> > v4.9-rc2, current policy changes without me touching it. Notice the
> > 1.47GHz below? I did not do that, it oscilates itself. Is that thermal
> > protection?
>
> Looks like to me.
>
> Can we verify somehow about what's the situation should look like? Perhaps with
> some older stable kernel? And then see if 4.8.X works fine or 4.9-rc.
I can try older kernel from Debian distribution, I guess.
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Viresh Kumar <viresh.kumar@linaro.org> |
|---|---|
| Date | 2016-11-04 10:40 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szJ8C-8ib-15@gated-at.bofh.it> |
| In reply to | #1515090 |
On 04-11-16, 10:26, Pavel Machek wrote: > How would I know if it is thermal capping? There's nothing in dmesg. I am not sure what code is responsible for doing that in case of x86, maybe Rafael and Rui can explain it that better. But surely it involves userspace in this case as scaling_max_freq is getting changed. -- viresh
[toc] | [prev] | [next] | [standalone]
| From | Henrique de Moraes Holschuh <hmh@hmh.eng.br> |
|---|---|
| Date | 2016-11-05 19:30 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sAdT4-2WL-23@gated-at.bofh.it> |
| In reply to | #1515096 |
On Fri, 04 Nov 2016, Viresh Kumar wrote: > On 04-11-16, 10:26, Pavel Machek wrote: > > How would I know if it is thermal capping? There's nothing in dmesg. > > I am not sure what code is responsible for doing that in case of x86, maybe > Rafael and Rui can explain it that better. > > But surely it involves userspace in this case as scaling_max_freq is getting > changed. Note that a X60 might limit the maximum frequency through ACPI when on AC *and* its battery is not installed or not in working order. The EC tries to detect the power brick type (65W, 90W), which often doesn't work well on non-Lenovo power bricks, and signals the BIOS to limit the maximum frequency if it thinks its only power source is a 65W power brick. This behavior would be readly visible by the X60 refusing to reach the maximum clock frequency *at all* -- it is a static limit, only removed when a sufficiently charged battery becomes available. This indirect power draw capping should never engage if your battery is installed and working, regardless of type and brand of power adapter, AFAIK. I am only mentioning it so that you are aware of this possible behavior while debugging. Anyway, this overheating thinkpad very, very likely needs a hardware repair (on top of the kernel bug fixes ;-) ). Regardless of any kernel/userspace bugs, the X60 was engineered to not require thermal capping on reasonable indoor environmental temperature ranges. If it needs the kernel to do something at all to avoid overheating just because it has been running at full CPU power, and ambient temperature is below 28°C, it very likely has a hardware issue (lint on the heatsink/fan, and/or cracked thermal interface between heatsink and processor). Every old-style ThinkPad (like the X60) that I have ever seen overheating when running at full CPU power (after fan lint was removed) in a cool place (say, 26°C), stopped overheating after a heatsink reseat and thermal compound replacement (and if you're going to do it, use non-shelf-aged Arctic Silver 5 or something better, and "cure" it properly -- something rather easy to do on a ThinkPad X60). -- Henrique Holschuh
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 20:30 +0100 |
| Subject | Thinkpad power management (was Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build) |
| Message-ID | <sAeP7-3vm-17@gated-at.bofh.it> |
| In reply to | #1515604 |
[Multipart message — attachments visible in raw view] — view raw
Hi! On Sat 2016-11-05 16:21:21, Henrique de Moraes Holschuh wrote: > On Fri, 04 Nov 2016, Viresh Kumar wrote: > > On 04-11-16, 10:26, Pavel Machek wrote: > > > How would I know if it is thermal capping? There's nothing in dmesg. > > > > I am not sure what code is responsible for doing that in case of x86, maybe > > Rafael and Rui can explain it that better. > > > > But surely it involves userspace in this case as scaling_max_freq is getting > > changed. > > Note that a X60 might limit the maximum frequency through ACPI when on > AC *and* its battery is not installed or not in working order. > > The EC tries to detect the power brick type (65W, 90W), which often > doesn't work well on non-Lenovo power bricks, and signals the BIOS to > limit the maximum frequency if it thinks its only power source is a 65W > power brick. > This behavior would be readly visible by the X60 refusing to reach the > maximum clock frequency *at all* -- it is a static limit, only removed > when a sufficiently charged battery becomes available. Umm. I don't think it is this -- maximum frequency is reached when machine is cold. It does have reasonably good battery. I guess I should also mention that it is currently running off 19V, 50.5W power brick. [If EC can figure it out, I guess it will hate me :-)] > Anyway, this overheating thinkpad very, very likely needs a hardware > repair (on top of the kernel bug fixes ;-) ). Ok, lets do kernel fixes, first :-). > Regardless of any kernel/userspace bugs, the X60 was engineered to not > require thermal capping on reasonable indoor environmental temperature > ranges. If it needs the kernel to do something at all to avoid > overheating just because it has been running at full CPU power, and > ambient temperature is below 28°C, it very likely has a hardware issue > (lint on the heatsink/fan, and/or cracked thermal interface between > heatsink and processor). > > Every old-style ThinkPad (like the X60) that I have ever seen > overheating when running at full CPU power (after fan lint was removed) > in a cool place (say, 26°C), stopped overheating after a heatsink reseat > and thermal compound replacement (and if you're going to do it, use > non-shelf-aged Arctic Silver 5 or something better, and "cure" it > properly -- something rather easy to do on a ThinkPad X60). Aha, thanks, good to know. Temperature is something like 19C, so it is not hot here, but a) I'm using it on soft surface and b) the legs are long gone. I was actually looking at newer thinkpad, but anything newer than X220 seems to have horrible keyboard... X220 looks reasonably nice, but people warned me that X60s are built from better materials, and I realize that X220 may have rowhammer bug... So I guess X60 is going to stay a bit longer :-). Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> |
|---|---|
| Date | 2016-11-04 15:10 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szNlT-2Hd-9@gated-at.bofh.it> |
| In reply to | #1515078 |
On Fri, 2016-11-04 at 09:58 +0100, Pavel Machek wrote: > On Fri 2016-11-04 09:38:49, Pavel Machek wrote: > > > > [...] > > What is worse, they go to low frequency even with "performance" > governor on v4.8-rc1?! > > pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq ; sudo cat > /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq > 1000000 > 1000000 > 1000000 > 1000000 > 1833000 > 1833000 > 1000000 > 1000000 > 1833000 > 1833000 > 1000000 > 1000000 > pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ grep -i > . /sys/devices/system/cpu/cpu0/cpufreq/* > /sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0 > /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000 You are having thermal issue, so the BIOS is forcing to lowest possible freq. Just for a test only, boot by adding to your kernel command line during boot processor.ignore_ppc=1 > grep: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq: > Permission denied > /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:1833000 > /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:1000000 > /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000 > /sys/devices/system/cpu/cpu0/cpufreq/freqdomain_cpus:0 1 > /sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:18 > 33000 > 1333000 1000000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:cons > ervative > powersave schedutil ondemand performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1000000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:1000000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:1000000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported> > grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory > pavel@duo:/sys/devices/system/cpu/cpu0/cpufreq$ > > Let me try v4.9-rc2... that works ok (cpus at the high frequency > during the kernel build). Unfortunately that sends my cpus to 99C > temperature range (and eventually forces emergency shutdown). This we have to debug. Do you see same line like " /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000 " If not we need to find out why. I suggest enter kernel bugzilla. I will add couple of more suggestion to debug. Thanks, Srinivas
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-04 21:50 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szTAZ-6Ct-13@gated-at.bofh.it> |
| In reply to | #1515206 |
[Multipart message — attachments visible in raw view] — view raw
Hi! > > Let me try v4.9-rc2... that works ok (cpus at the high frequency > > during the kernel build). Unfortunately that sends my cpus to 99C > > temperature range (and eventually forces emergency shutdown). > > This we have to debug. Do you see same line like > " > /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000 > " > If not we need > to find out why. I'd prefer mails over bugzilla for now... 4.9-rc2 has bios_limit: pavel@duo:~$ cat /sys/devices/system/cpu/cpu0/cpufreq/bios_limit 1833000 and it has thermal zones: /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_temp 127000 /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_type critical /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_temp 97000 /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_type critical /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_temp 92500 /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_type passive ..so it should slow down CPU at 92C. So lets push the temperature up a bit... sudo watch cat /proc/acpi/ibm/thermal /sys/devices/system/cpu/cpu0/cpufreq/bios_limit /sys/devices/virtual/thermal/thermal_zone1/temp /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq temperatures: 98 49 -128 85 28 -128 28 -128 49 58 -128 -128 -128 -128 -128 -128 1833000 95000 1833000 Hmm. bios_limit does not seem to change, even when the temperature is clearly above the trip point. (It is also interestng that acpi/ibm reports bigger temperatures than /sys/devices/virtual/thermal/thermal_zone1/temp . I have seen 103C there.) Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> |
|---|---|
| Date | 2016-11-04 22:20 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szU42-71A-57@gated-at.bofh.it> |
| In reply to | #1515396 |
On Fri, 2016-11-04 at 21:44 +0100, Pavel Machek wrote: > Hi! > > > > > > > > > Let me try v4.9-rc2... that works ok (cpus at the high frequency > > > during the kernel build). Unfortunately that sends my cpus to 99C > > > temperature range (and eventually forces emergency shutdown). > > > > This we have to debug. Do you see same line like > > " > > /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:1000000 > > " > > If not we need > > to find out why. > > I'd prefer mails over bugzilla for now... > > 4.9-rc2 has bios_limit: > > pavel@duo:~$ cat /sys/devices/system/cpu/cpu0/cpufreq/bios_limit > 1833000 > > and it has thermal zones: > > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_temp 127000 > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_type critical > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_temp 97000 > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_type critical > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_temp 92500 > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_type passive > It will not act if there is no binding information. Do you have more files in this folder? grep -r . * in /sys/class/thermal will be helpful. > ..so it should slow down CPU at 92C. > > So lets push the temperature up a bit... > > sudo watch cat /proc/acpi/ibm/thermal > /sys/devices/system/cpu/cpu0/cpufreq/bios_limit > /sys/devices/virtual/thermal/thermal_zone1/temp /sys/devices/system/ > cpu/cpu0/cpufreq/cpuinfo_cur_freq > > temperatures: 98 49 -128 85 28 -128 28 -128 49 58 -128 -128 -128 > -128 -128 -128 > 1833000 > 95000 > 1833000 > > Hmm. bios_limit does not seem to change, even when the temperature is > clearly above the trip point. (It is also interestng that acpi/ibm > reports bigger temperatures than > /sys/devices/virtual/thermal/thermal_zone1/temp . I have seen 103C > there.) Probably they are showing package and core temperature or have a different sampling interval. Try enabling thermald service in Debian. it has access to more knobs to control thermals. Thanks, Srinivas
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 09:50 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sA4PL-5CQ-3@gated-at.bofh.it> |
| In reply to | #1515430 |
[Multipart message — attachments visible in raw view] — view raw
Hi! > > 4.9-rc2 has bios_limit: > > > > pavel@duo:~$ cat /sys/devices/system/cpu/cpu0/cpufreq/bios_limit > > 1833000 > > > > and it has thermal zones: > > > > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_temp 127000 > > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_type critical > > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_temp 97000 > > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_type critical > > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_temp 92500 > > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_type passive > > > It will not act if there is no binding information. Do you have more > files in this folder? > > grep -r . * in /sys/class/thermal will be helpful. Yes, I do. Here you go: pavel@duo:/data/l/linux$ sudo grep -ri . /sys/class/thermal/* /sys/class/thermal/cooling_device0/cur_state:0 /sys/class/thermal/cooling_device0/power/runtime_active_kids:0 /sys/class/thermal/cooling_device0/power/runtime_suspended_time:0 grep: /sys/class/thermal/cooling_device0/power/autosuspend_delay_ms: Input/output error /sys/class/thermal/cooling_device0/power/runtime_enabled:disabled /sys/class/thermal/cooling_device0/power/runtime_active_time:0 /sys/class/thermal/cooling_device0/power/control:auto /sys/class/thermal/cooling_device0/power/async:disabled /sys/class/thermal/cooling_device0/power/runtime_usage:0 /sys/class/thermal/cooling_device0/power/runtime_status:unsupported /sys/class/thermal/cooling_device0/type:LCD /sys/class/thermal/cooling_device0/max_state:7 /sys/class/thermal/cooling_device1/cur_state:0 /sys/class/thermal/cooling_device1/power/runtime_active_kids:0 /sys/class/thermal/cooling_device1/power/runtime_suspended_time:0 grep: /sys/class/thermal/cooling_device1/power/autosuspend_delay_ms: Input/output error /sys/class/thermal/cooling_device1/power/runtime_enabled:disabled /sys/class/thermal/cooling_device1/power/runtime_active_time:0 /sys/class/thermal/cooling_device1/power/control:auto /sys/class/thermal/cooling_device1/power/async:disabled /sys/class/thermal/cooling_device1/power/runtime_usage:0 /sys/class/thermal/cooling_device1/power/runtime_status:unsupported /sys/class/thermal/cooling_device1/type:Processor /sys/class/thermal/cooling_device1/max_state:10 /sys/class/thermal/cooling_device2/cur_state:0 /sys/class/thermal/cooling_device2/power/runtime_active_kids:0 /sys/class/thermal/cooling_device2/power/runtime_suspended_time:0 grep: /sys/class/thermal/cooling_device2/power/autosuspend_delay_ms: Input/output error /sys/class/thermal/cooling_device2/power/runtime_enabled:disabled /sys/class/thermal/cooling_device2/power/runtime_active_time:0 /sys/class/thermal/cooling_device2/power/control:auto /sys/class/thermal/cooling_device2/power/async:disabled /sys/class/thermal/cooling_device2/power/runtime_usage:0 /sys/class/thermal/cooling_device2/power/runtime_status:unsupported /sys/class/thermal/cooling_device2/type:Processor /sys/class/thermal/cooling_device2/max_state:10 /sys/class/thermal/thermal_zone0/passive:0 /sys/class/thermal/thermal_zone0/available_policies:step_wise /sys/class/thermal/thermal_zone0/policy:step_wise /sys/class/thermal/thermal_zone0/mode:enabled grep: /sys/class/thermal/thermal_zone0/offset: Input/output error grep: /sys/class/thermal/thermal_zone0/integral_cutoff: Input/output error /sys/class/thermal/thermal_zone0/power/runtime_active_kids:0 /sys/class/thermal/thermal_zone0/power/runtime_suspended_time:0 grep: /sys/class/thermal/thermal_zone0/power/autosuspend_delay_ms: Input/output error /sys/class/thermal/thermal_zone0/power/runtime_enabled:disabled /sys/class/thermal/thermal_zone0/power/runtime_active_time:0 /sys/class/thermal/thermal_zone0/power/control:auto /sys/class/thermal/thermal_zone0/power/async:disabled /sys/class/thermal/thermal_zone0/power/runtime_usage:0 /sys/class/thermal/thermal_zone0/power/runtime_status:unsupported grep: /sys/class/thermal/thermal_zone0/k_d: Input/output error /sys/class/thermal/thermal_zone0/trip_point_0_type:critical /sys/class/thermal/thermal_zone0/type:acpitz grep: /sys/class/thermal/thermal_zone0/slope: Input/output error grep: /sys/class/thermal/thermal_zone0/k_po: Input/output error grep: /sys/class/thermal/thermal_zone0/k_i: Input/output error grep: /sys/class/thermal/thermal_zone0/sustainable_power: Input/output error /sys/class/thermal/thermal_zone0/trip_point_0_temp:127000 grep: /sys/class/thermal/thermal_zone0/k_pu: Input/output error /sys/class/thermal/thermal_zone0/temp:58000 /sys/class/thermal/thermal_zone1/available_policies:step_wise /sys/class/thermal/thermal_zone1/policy:step_wise /sys/class/thermal/thermal_zone1/mode:enabled /sys/class/thermal/thermal_zone1/trip_point_1_type:passive grep: /sys/class/thermal/thermal_zone1/offset: Input/output error grep: /sys/class/thermal/thermal_zone1/integral_cutoff: Input/output error /sys/class/thermal/thermal_zone1/power/runtime_active_kids:0 /sys/class/thermal/thermal_zone1/power/runtime_suspended_time:0 grep: /sys/class/thermal/thermal_zone1/power/autosuspend_delay_ms: Input/output error /sys/class/thermal/thermal_zone1/power/runtime_enabled:disabled /sys/class/thermal/thermal_zone1/power/runtime_active_time:0 /sys/class/thermal/thermal_zone1/power/control:auto /sys/class/thermal/thermal_zone1/power/async:disabled /sys/class/thermal/thermal_zone1/power/runtime_usage:0 /sys/class/thermal/thermal_zone1/power/runtime_status:unsupported grep: /sys/class/thermal/thermal_zone1/k_d: Input/output error /sys/class/thermal/thermal_zone1/trip_point_0_type:critical /sys/class/thermal/thermal_zone1/type:acpitz /sys/class/thermal/thermal_zone1/cdev1_weight:0 /sys/class/thermal/thermal_zone1/trip_point_1_temp:92500 grep: /sys/class/thermal/thermal_zone1/slope: Input/output error grep: /sys/class/thermal/thermal_zone1/k_po: Input/output error grep: /sys/class/thermal/thermal_zone1/k_i: Input/output error /sys/class/thermal/thermal_zone1/cdev0_weight:0 grep: /sys/class/thermal/thermal_zone1/sustainable_power: Input/output error /sys/class/thermal/thermal_zone1/trip_point_0_temp:97000 /sys/class/thermal/thermal_zone1/cdev0_trip_point:1 grep: /sys/class/thermal/thermal_zone1/k_pu: Input/output error /sys/class/thermal/thermal_zone1/temp:80000 /sys/class/thermal/thermal_zone1/cdev1_trip_point:1 > > Hmm. bios_limit does not seem to change, even when the temperature is > > clearly above the trip point. (It is also interestng that acpi/ibm > > reports bigger temperatures than > > /sys/devices/virtual/thermal/thermal_zone1/temp . I have seen 103C > > there.) > Probably they are showing package and core temperature or have a > different sampling interval. > > Try enabling thermald service in Debian. it has access to more knobs to > control thermals. Hmm, thanks for the pointer. But it seems like I'll have to build my own, as /proc/acpi/ibm does not follow the usual infrastructure... Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Henrique de Moraes Holschuh <hmh@hmh.eng.br> |
|---|---|
| Date | 2016-11-05 18:50 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sAdgm-2qi-15@gated-at.bofh.it> |
| In reply to | #1515524 |
On Sat, 05 Nov 2016, Pavel Machek wrote: > Hmm, thanks for the pointer. But it seems like I'll have to build my > own, as /proc/acpi/ibm does not follow the usual infrastructure... /proc/acpi/ibm has been deprecated for years. 99% of the functionality is available through more modern, standard interfaces. thinkpad-acpi is supposed to export standard hwmon temperature sensors as well. Try them instead, please. Also, thinkpad-acpi will report the EC thermal sensors. They don't need to (and most often won't) match whatever you read from the processor core(s). -- Henrique Holschuh
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 20:40 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sAeYN-3yw-5@gated-at.bofh.it> |
| In reply to | #1515596 |
[Multipart message — attachments visible in raw view] — view raw
On Sat 2016-11-05 15:46:12, Henrique de Moraes Holschuh wrote: > On Sat, 05 Nov 2016, Pavel Machek wrote: > > Hmm, thanks for the pointer. But it seems like I'll have to build my > > own, as /proc/acpi/ibm does not follow the usual infrastructure... > > /proc/acpi/ibm has been deprecated for years. 99% of the functionality > is available through more modern, standard interfaces. Right, I see sensors can do it these days. Would it be good to expose them as /sys/class/thermal/thermal_zone*, too? Is it known what various fields in /proc/acpi/ibm/thermal measure? Basically... 100C is okay for semiconductors, but I'd prefer not to kill the hard drive.... > thinkpad-acpi is supposed to export standard hwmon temperature sensors > as well. Try them instead, please. Heh, I just finished python to work with /proc/acpi/ibm. Oh well. Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Henrique de Moraes Holschuh <hmh@hmh.eng.br> |
|---|---|
| Date | 2016-11-06 04:50 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sAmD0-8Y-9@gated-at.bofh.it> |
| In reply to | #1515616 |
On Sat, 05 Nov 2016, Pavel Machek wrote: > On Sat 2016-11-05 15:46:12, Henrique de Moraes Holschuh wrote: > > On Sat, 05 Nov 2016, Pavel Machek wrote: > > > Hmm, thanks for the pointer. But it seems like I'll have to build my > > > own, as /proc/acpi/ibm does not follow the usual infrastructure... > > > > /proc/acpi/ibm has been deprecated for years. 99% of the functionality > > is available through more modern, standard interfaces. > > Right, I see sensors can do it these days. Would it be good to expose > them as /sys/class/thermal/thermal_zone*, too? I don't like the idea of touching vendor-screwup-land like thermal zones, especially when thinkpads *already* have thermal zones and they must come from the very same sensors... This would need a lot of careful studying and planning. > Is it known what various fields in /proc/acpi/ibm/thermal measure? It varies with each model. You can map it out with better precision by using a cold spray while doing an "open-heart" surgery on the thinkpad. I kid you not, that's how they were mapped for some models. Look for information about this in thinkwiki: http://www.thinkwiki.org/wiki/Thermal_Sensors#ThinkPad_X60 The battery-pack-related sensor that is stuck at 50°C is something the Lenovo Yamato labs guys wouldn't be clear about. They told me to look at what the Windows drivers do, but that would require (1) Windows in the first place, and (2) clean room reverse engineering protocols. > Basically... 100C is okay for semiconductors, but I'd prefer not to > kill the hard drive.... Well, that first sensor getting to 100°C is your CPU for sure. > > thinkpad-acpi is supposed to export standard hwmon temperature sensors > > as well. Try them instead, please. > > Heh, I just finished python to work with /proc/acpi/ibm. Oh well. Hey, I *did* properly document that driver, and the documentation is even mostly up-to-date... Please refer to Documentation/laptops/thinkpad-acpi.txt in the kernel tree. -- Henrique Holschuh
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-04 23:20 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szV06-7FO-9@gated-at.bofh.it> |
| In reply to | #1515396 |
[Multipart message — attachments visible in raw view] — view raw
Hi! > I'd prefer mails over bugzilla for now... > > 4.9-rc2 has bios_limit: > > pavel@duo:~$ cat /sys/devices/system/cpu/cpu0/cpufreq/bios_limit > 1833000 > > and it has thermal zones: > > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_temp 127000 > /sys/devices/virtual/thermal/thermal_zone0/trip_point_0_type critical > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_temp 97000 > /sys/devices/virtual/thermal/thermal_zone1/trip_point_0_type critical > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_temp 92500 > /sys/devices/virtual/thermal/thermal_zone1/trip_point_1_type passive > > ..so it should slow down CPU at 92C. > > So lets push the temperature up a bit... > > sudo watch cat /proc/acpi/ibm/thermal > /sys/devices/system/cpu/cpu0/cpufreq/bios_limit > /sys/devices/virtual/thermal/thermal_zone1/temp /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq > > temperatures: 98 49 -128 85 28 -128 28 -128 49 58 -128 -128 -128 > -128 -128 -128 > 1833000 > 95000 > 1833000 > > Hmm. bios_limit does not seem to change, even when the temperature is > clearly above the trip point. (It is also interestng that acpi/ibm > reports bigger temperatures than > /sys/devices/virtual/thermal/thermal_zone1/temp . I have seen 103C > there.) Under v4.8-rc, behaviour is different: bios_limit goes to 1GHz there when temperature is around 84C at the thermal zone. That keeps ibm/thermal temperatures under 90C, and no "thermal emergency" messages in syslog. So we seem to have thermal or ACPI regression in v4.9-rc3. Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> |
|---|---|
| Date | 2016-11-05 00:30 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <szW5Q-8kN-5@gated-at.bofh.it> |
| In reply to | #1515454 |
On Fri, 2016-11-04 at 23:16 +0100, Pavel Machek wrote: > Hi! > [...] > So we seem to have thermal or ACPI regression in v4.9-rc3. > It is possible. Can you add either add printk in acpi_processor_ppc_has_changed() or use ftrace and see do you get to these functions acpi_processor_ppc_init() acpi_processor_ppc_has_changed() acpi_processor_ppc_notifier() ? When temperature limit is reached acpi_processor_ppc_notifier() should be called. Thanks, Srinivas
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 14:30 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sA9cK-8oH-1@gated-at.bofh.it> |
| In reply to | #1515476 |
[Multipart message — attachments visible in raw view] — view raw
On Fri 2016-11-04 23:20:53, Pandruvada, Srinivas wrote: > On Fri, 2016-11-04 at 23:16 +0100, Pavel Machek wrote: > > Hi! > > > > [...] > > > So we seem to have thermal or ACPI regression in v4.9-rc3. > > > It is possible. Can you add either add printk > in acpi_processor_ppc_has_changed() or use ftrace and see do you get to > these functions > > acpi_processor_ppc_init() > acpi_processor_ppc_has_changed() > acpi_processor_ppc_notifier() > > ? Ok, can do, let me recompile and reboot. > When temperature limit is reached acpi_processor_ppc_notifier() should > be called. No, that's not correct for ACPI passive trip points, is it? If I recall correctly, those should be monitored even when temperature is below them so that it does not reach them...? Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | "Pandruvada, Srinivas" <srinivas.pandruvada@intel.com> |
|---|---|
| Date | 2016-11-05 14:40 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sA9mq-8rO-25@gated-at.bofh.it> |
| In reply to | #1515545 |
On Sat, 2016-11-05 at 14:20 +0100, Pavel Machek wrote: > On Fri 2016-11-04 23:20:53, Pandruvada, Srinivas wrote: > > > > On Fri, 2016-11-04 at 23:16 +0100, Pavel Machek wrote: > > > > > > Hi! > > > > > > > [...] > > > > > > > > So we seem to have thermal or ACPI regression in v4.9-rc3. > > > > > It is possible. Can you add either add printk > > in acpi_processor_ppc_has_changed() or use ftrace and see do you > > get to > > these functions > > > > acpi_processor_ppc_init() > > acpi_processor_ppc_has_changed() > > acpi_processor_ppc_notifier() > > > > ? > > Ok, can do, let me recompile and reboot. > > > > > When temperature limit is reached acpi_processor_ppc_notifier() > > should > > be called. > > No, that's not correct for ACPI passive trip points, is it? If I > recall correctly, those should be monitored even when temperature is > below them so that it does not reach them...? No if BIOS is sending PPC, it will be called. You can try first in 4.8. Also try Don't run workload, just on an idle system. # echo 1 > /sys/class/thermal/cooling_device2/cur_state monitor the scaling_max_freq, it should reduce # echo 2 > /sys/class/thermal/cooling_device2/cur_state monitor the scaling_max_freq, it should reduce Thanks, Srinivas
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 15:00 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sA9FL-6I-11@gated-at.bofh.it> |
| In reply to | #1515547 |
[Multipart message — attachments visible in raw view] — view raw
Hi! > > Ok, can do, let me recompile and reboot. > > > > > > > > When temperature limit is reached acpi_processor_ppc_notifier() > > > should > > > be called. > > > > No, that's not correct for ACPI passive trip points, is it? If I > > recall correctly, those should be monitored even when temperature is > > below them so that it does not reach them...? > No if BIOS is sending PPC, it will be called. You can try first in 4.8. > > Also try > > Don't run workload, just on an idle system. > > # echo 1 > /sys/class/thermal/cooling_device2/cur_state > monitor the scaling_max_freq, it should reduce > > # echo 2 > /sys/class/thermal/cooling_device2/cur_state > monitor the scaling_max_freq, it should reduce Yes, this seems to work. scaling_max goes to 1.5, then 1.1. Also under load, scaling_max_freq changes. But at that point, we are already around 98C... and bios_limit stays the same all the time in v4.9. Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-11-05 15:10 +0100 |
| Subject | Re: v4.8-rc1: thinkpad x60: running at low frequency even during kernel build |
| Message-ID | <sA9Pr-q7-3@gated-at.bofh.it> |
| In reply to | #1515554 |
[Multipart message — attachments visible in raw view] — view raw
On Sat 2016-11-05 14:53:13, Pavel Machek wrote: > Hi! > > > > Ok, can do, let me recompile and reboot. > > > > > > > > > > > When temperature limit is reached acpi_processor_ppc_notifier() > > > > should > > > > be called. > > > > > > No, that's not correct for ACPI passive trip points, is it? If I > > > recall correctly, those should be monitored even when temperature is > > > below them so that it does not reach them...? > > No if BIOS is sending PPC, it will be called. You can try first in 4.8. > > > > Also try > > > > Don't run workload, just on an idle system. > > > > # echo 1 > /sys/class/thermal/cooling_device2/cur_state > > monitor the scaling_max_freq, it should reduce > > > > # echo 2 > /sys/class/thermal/cooling_device2/cur_state > > monitor the scaling_max_freq, it should reduce > > Yes, this seems to work. scaling_max goes to 1.5, then 1.1. Also under > load, scaling_max_freq changes. But at that point, we are already > around 98C... and bios_limit stays the same all the time in v4.9. ...while in v4.8-rc1, bios limit goes to 1.0 GHz at 90C, and temperature just doesn't go above that. Does it make sense to try with v4.8-final? Best regards, Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
Page 1 of 2 [1] 2 Next page →
Back to top | Article view | linux.kernel
csiph-web