[Letux-kernel] Strange things happening with latest kernels
H. Nikolaus Schaller
hns at goldelico.com
Thu Jun 13 19:16:56 CEST 2019
> Am 13.06.2019 um 18:46 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
>
>
>> Am 13.06.2019 um 18:33 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>
>> Hi,
>>
>> On Thu, 13 Jun 2019 18:08:12 +0200
>> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>>
>>>> Am 13.06.2019 um 17:51 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>>>
>>>> On Thu, 13 Jun 2019 15:45:42 +0200
>>>> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>>>>
>>>>>> Am 13.06.2019 um 11:08 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
>>>>>>
>>>>>>
>>>>>> Something which needs deeper inspection to understand the influencing
>>>>>> factors.
>>>>>
>>>>> As soon as I start X11, temperature goes up. Battery temp to 55.1°C and OMAP3
>>>>> reports 90°C during charging.
>>>>>
>>>> Strange thing, Does it depend on how X is accessing the fb (omapfb vs. fbdev vs. modesetting
>>>> driver)? SGX active vs. not active.
>>>
>>> Have not yet tested into that direction. SGX is not compiled into any of the
>>> kernels I have tried.
>>>
>>> Next thing is to test with LXDE or XfCE. And try a build with omap2plus_defconfig.
>>>
>> maybe there are multiple problems working together to heat up omap3.
>
> Yes, looks so. I just never had experienced that on OMAP3 before. Just
> on OMAP5...
>
>> I doubt cpufreq alone causes such problems. But I will test, if
>> everything is idle, things should be powered down anyways regardless of
>> cpufreq
>>
>> When I had a hot gta04 in the past, it was always at a bad time.
>> Touchscreen did not work. Last time was when I wanted to use the gta04
>> to show me the way from Brussels midi to my hotel for fosdem. That
>> was with 4.17 kernel.
>
> And strangely there aren't many recent omap3 specific changes in 4.19..4.19.49
>
> iMac:master hns$ git log v4.19..v4.19.49 --oneline arch/arm/mach-omap2/
> 30d503bae967 ARM: avoid Cortex-A9 livelock on tight dmb loops
> f49f7007de59 ARM: OMAP2+: fix lack of timer interrupts on CPU1 after hotplug
> f3f7a8b6faf8 ARM: OMAP2+: Variable "reg" in function omap4_dsi_mux_pads() could be uninitialized
> 1b1de8b95b97 ARM: OMAP5+: Fix inverted nirq pin interrupts with irq_set_type
> 8218fcf4666c ARM: OMAP2+: hwmod: Fix some section annotations
> 5a8fbba77bf8 ARM: OMAP2+: prm44xx: Fix section annotation on omap44xx_prm_enable_io_wakeup
> iMac:master hns$
>
> So it might be some (clock?) driver but there are just 2 patches in drivers/clk/ti.
>
> Maybe I just never had recognised this because the room temperature was lower?
Or it did degrade with every kernel upgrade since 4.19-rc1 and I rarely can do stress
tests with every new kernel version, just check basic functions (display, touch, USB,
charger, sound) or things explicitly backported/improved.
> On the other hand, the GTA04 is really getting hot (with risk for battery damage).
Ok, with LXDE it is initially (after boot) at ca. 70% 800MHz
but degrading to 20%@800MHz 5 minutes after boot and
It has a little less system load in "idle" than QuantumSTEP, but 10% vs. 2%
should not make such a big difference.
top - 00:03:25 up 3 min, 2 users, load average: 0.44, 0.60, 0.27
Tasks: 85 total, 1 running, 56 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.7 us, 1.3 sy, 0.0 ni, 97.1 id, 0.0 wa, 0.0 hi, 1.0 si, 0.0 st
KiB Mem: 507084 total, 160996 used, 346088 free, 10476 buffers
KiB Swap: 0 total, 0 used, 0 free. 93440 cached Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3612 root 20 0 3100 1804 1384 R 1.6 0.4 0:00.89 top
3606 root 20 0 9988 3992 3392 S 0.7 0.8 0:00.36 sshd
17 root 20 0 0 0 0 I 0.3 0.0 0:00.14 kworker/u2:1-ev
19 root 20 0 0 0 0 I 0.3 0.0 0:00.54 kworker/0:1-eve
3054 root 20 0 52060 15444 12588 S 0.3 3.0 0:07.88 lxpanel
1 root 20 0 2372 1224 1088 S 0.0 0.2 0:03.10 init
2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root 0 -20 0 0 0 I 0.0 0.0 0:00.00 rcu_gp
4 root 0 -20 0 0 0 I 0.0 0.0 0:00.00 rcu_par_gp
Temperature is also growing:
root at letux:~# ./temperatures
Sat Jan 1 00:06:58 UTC 2000 67° 33° 4158mV 800MHz
root at letux:~#
But: I did run the "high-load" script which tries to do a 100% load on each
ARM core (developed for heating up the Pyra) and there, the SoC temperature
slowly creeps up to 84°C and battery is at 41°.
And after stopping this process it goes down to the 75° and 45° which I
remember to be some thermal equilibrium with radiation to the surrounding
air.
So something in the OMAP3 SoC can be getting hotter than it IMHO should...
Maybe I should also compare to the 4.14 kernels. And try to look with the
thermo-image camera if it is really the OMAP3 and not e.g. the WLAN chip
which is close.
Much more work than planned for...
BR,
Nikolaus
More information about the Letux-kernel
mailing list