[gentoo-user] Abnormal processor temperature.

[Michael]

I have a system here running an Intel N97 processor, which is idling at
70-80C on Gentoo with all cores 99% idle. This is 40 degrees hotter than
it runs on Ubuntu or Windows 10.
Powertop confirms that the CPU is spending nearly all of its time in idle
mode. I have tried both the 6.1 and the 6.5 kernels, including the
pre-compiled gentoo-kernel-bin version just to be sure it's not something
in *my* config messing it up.
I'm not really sure where to look next. There doesn't seem to be anything
actually *using* the CPU or GPU to any significant degree. Certainly not
enough to explain the temperature difference. If anyone has advice on what
diagnostics to try next to find out what's going on I'd appreciate it.
Thanks,
LMP

The fact you get a report of 40° hotter temperature does not mean the CPU
temperature is actually higher. Do you get any other indications the PC is
running hot, e.g. the fans are spinning faster, the heatsink is hotter to the
touch, or when measured with an infrared thermometer?

Is the 'sensors' command output, after you have run sensors-detect, different
than Ubuntu's?

Is the content of /etc/sensors3.conf different between the two distros?

Are there different sensor modules loaded?

Some chip drivers report spurious results and need specifying a different
measurement method/sensor device. You can take a look here for you CPU/MoBo
chip:

https://www.kernel.org/doc/Documentation/hwmon/

[Michael]

-----Original Message-----
Sent: Tuesday, November 21, 2023 4:15 PM
Subject: Re: [gentoo-user] Abnormal processor temperature.

I have a system here running an Intel N97 processor, which is idling
at 70-80C on Gentoo with all cores 99% idle. This is 40 degrees
hotter than it runs on Ubuntu or Windows 10.
Powertop confirms that the CPU is spending nearly all of its time in
idle mode. I have tried both the 6.1 and the 6.5 kernels, including
the pre-compiled gentoo-kernel-bin version just to be sure it's not
something in *my* config messing it up.
I'm not really sure where to look next. There doesn't seem to be
anything actually *using* the CPU or GPU to any significant degree.
Certainly not enough to explain the temperature difference. If anyone
has advice on what diagnostics to try next to find out what's going on
I'd appreciate it.
Thanks,
LMP

The fact you get a report of 40° hotter temperature does not mean the CPU
temperature is actually higher. Do you get any other indications the PC
is running hot, e.g. the fans are spinning faster, the heatsink is hotter
to the touch, or when measured with an infrared thermometer?
Is the 'sensors' command output, after you have run sensors-detect,
different than Ubuntu's?
Is the content of /etc/sensors3.conf different between the two distros?
Are there different sensor modules loaded?
Some chip drivers report spurious results and need specifying a different
measurement method/sensor device. You can take a look here for you CPU/MoBo
https://www.kernel.org/doc/Documentation/hwmon/

Sorry, I forgot to mention that. The fans definitely run faster. Which is
fortunate because checking with an IR thermometer on the CPU heatsink
produces more varied results than the internal sensor, but still 30-40°
hotter.

OK, it's definitely not a sensor misreading the temperature then.

Driver modules loaded by each distro don't seem substantially different.
Ubuntu has intel-spi loaded while Gentoo does not, but that's it for
differences from lspci. I can look for other differences as well, but it
doesn't seem to be a difference in sensor calibration unfortunately.

In the first instance I suggest you add SPI module(s) to your kernel. It
controls comms with sensors and peripherals, so it could well play an
important role in controlling the MoBo temperature.

[Grant Edwards]

I have a system here running an Intel N97 processor, which is idling
at 70-80C on Gentoo with all cores 99% idle. This is 40 degrees
hotter than it runs on Ubuntu or Windows 10.
Powertop confirms that the CPU is spending nearly all of its time in
idle mode.

Are clock speeds being scaled down when idle? Or does the N97's "idle
mode" preclude the need to scale down clock speed when not busy to
avoid high temps?

[Michael]

-----Original Message-----
Sent: Wednesday, November 22, 2023 8:36 AM
Subject: [gentoo-user] Re: Abnormal processor temperature.

I have a system here running an Intel N97 processor, which is idling
at 70-80C on Gentoo with all cores 99% idle. This is 40 degrees
hotter than it runs on Ubuntu or Windows 10.
Powertop confirms that the CPU is spending nearly all of its time in
idle mode.

Are clock speeds being scaled down when idle? Or does the N97's "idle
mode" preclude the need to scale down clock speed when not busy to avoid
high temps?

That was part of my confusion, because even when I used cpupower to lock all
cores to 800MHz, it kept running hot.
Michael's suggestion of adding the SPI modules to the kernel appears to have
fixed it. I'm not sure why the gentoo-kernel-bin that I tried didn't work
at that rate, but I'm going to have to activate the
"somebody-else's-problem" field on that one.
LMP

I suggested enabling the SPI modules because they are used by the CPU to
communicate with various sensors, adjust clock frequency between components
and thereafter to receive signals a/synchronously to control temperatures.
Theoretically, a fail-safe system would thermally throttle, or over-cool the
CPU, when any of the critical SPI signals is lost. I can't say I understand
why without some temperature sensor feedback the CPU decided to overclock
itself. I would have thought these days oscillator crystals or equivalent
capacitor circuitry would be internal to the CPU die, so the CPU frequency
control would be self-contained. Either way, this seems to have been the
problem.

[Dr Rainer Woitok]

Michael,

Post by Michael
...
I suggested enabling the SPI modules because they are used by the CPU to
communicate with various sensors, adjust clock frequency between components
and thereafter to receive signals a/synchronously to control temperatures.

What exactly do you suggest? My kernel (6.1.57) configuration contains:

$ grep SPI /usr/src/linux/.config|grep -v SPIN
CONFIG_SCSI_SPI_ATTRS=y
# CONFIG_SPI is not set
# SPI RTC drivers
CONFIG_RTC_I2C_AND_SPI=y
# SPI and I2C RTC drivers
$

I played a little with the "/" and "?" commands of "make menuconfig",
activated "CONFIG_SPI" and found a lot of controllers and drivers, but
nothing obvious to enable additionally. However, apparently activating
"CONFIG_SPI" also activated

CONFIG_NET_VENDOR_ADI=y
CONFIG_SND_SPI=y
CONFIG_SPI_DYNAMIC=y
CONFIG_SPI_MASTER=y

Is this what you are suggesting?

Sincerely,
Rainer

[Michael]

Post by Dr Rainer Woitok
Michael,

Post by Michael
...
I suggested enabling the SPI modules because they are used by the CPU to
communicate with various sensors, adjust clock frequency between components
and thereafter to receive signals a/synchronously to control temperatures.

$ grep SPI /usr/src/linux/.config|grep -v SPIN
CONFIG_SCSI_SPI_ATTRS=y
# CONFIG_SPI is not set
# SPI RTC drivers
CONFIG_RTC_I2C_AND_SPI=y
# SPI and I2C RTC drivers
$
I played a little with the "/" and "?" commands of "make menuconfig",
activated "CONFIG_SPI" and found a lot of controllers and drivers, but
nothing obvious to enable additionally. However, apparently activating
"CONFIG_SPI" also activated
CONFIG_NET_VENDOR_ADI=y
CONFIG_SND_SPI=y
CONFIG_SPI_DYNAMIC=y
CONFIG_SPI_MASTER=y
Is this what you are suggesting?
Sincerely,
Rainer

It depends on the hardware, this is what I have enabled on an AMD MoBo:

~ $ grep SPI /usr/src/linux/.config
CONFIG_UNINLINE_SPIN_UNLOCK=y
CONFIG_MUTEX_SPIN_ON_OWNER=y
CONFIG_RWSEM_SPIN_ON_OWNER=y
CONFIG_LOCK_SPIN_ON_OWNER=y
CONFIG_ARCH_USE_QUEUED_SPINLOCKS=y
CONFIG_QUEUED_SPINLOCKS=y
# CONFIG_INET6_ESPINTCP is not set
CONFIG_SCSI_SPI_ATTRS=y
CONFIG_SPI=y
CONFIG_SPI_MASTER=y
# CONFIG_SPI_MEM is not set
# SPI Master Controller Drivers
# CONFIG_SPI_ALTERA is not set
# CONFIG_SPI_AXI_SPI_ENGINE is not set
# CONFIG_SPI_BITBANG is not set
# CONFIG_SPI_CADENCE is not set
# CONFIG_SPI_DESIGNWARE is not set
# CONFIG_SPI_NXP_FLEXSPI is not set
# CONFIG_SPI_GPIO is not set
# CONFIG_SPI_MICROCHIP_CORE is not set
# CONFIG_SPI_MICROCHIP_CORE_QSPI is not set
# CONFIG_SPI_LANTIQ_SSC is not set
# CONFIG_SPI_OC_TINY is not set
# CONFIG_SPI_PXA2XX is not set
# CONFIG_SPI_ROCKCHIP is not set
# CONFIG_SPI_SC18IS602 is not set
# CONFIG_SPI_SIFIVE is not set
# CONFIG_SPI_MXIC is not set
# CONFIG_SPI_XCOMM is not set
# CONFIG_SPI_XILINX is not set
# CONFIG_SPI_ZYNQMP_GQSPI is not set
CONFIG_SPI_AMD=m
# SPI Multiplexer support
# CONFIG_SPI_MUX is not set
# SPI Protocol Masters
# CONFIG_SPI_SPIDEV is not set
# CONFIG_SPI_LOOPBACK_TEST is not set
# CONFIG_SPI_TLE62X0 is not set
# CONFIG_SPI_SLAVE is not set
CONFIG_SPI_DYNAMIC=y
# SPI GPIO expanders
# end of SPI GPIO expanders
# CONFIG_SENSORS_LTC2947_SPI is not set
# CONFIG_MFD_DA9052_SPI is not set
# CONFIG_MFD_MC13XXX_SPI is not set
# CONFIG_MFD_TPS65912_SPI is not set
# CONFIG_MFD_ARIZONA_SPI is not set
# CONFIG_MFD_WM831X_SPI is not set
# CONFIG_SND_SPI is not set
# SPI RTC drivers
CONFIG_RTC_I2C_AND_SPI=y
# SPI and I2C RTC drivers
# CONFIG_HWSPINLOCK is not set

[Dr Rainer Woitok]

Michael,

Post by Michael
...
~ $ grep SPI /usr/src/linux/.config
...
CONFIG_SCSI_SPI_ATTRS=y
CONFIG_SPI=y
CONFIG_SPI_MASTER=y
...
CONFIG_SPI_AMD=m
...
CONFIG_SPI_DYNAMIC=y
...
CONFIG_RTC_I2C_AND_SPI=y

Thanks for the info. Apart from "CONFIG_SPI_AMD=m" this is what I've
now configured too.

Sincerely,
Rainer