Re: x270 CPU temp / throttling and unhandled HKEY event when I close the lid

[neil k <host.crash@xxxxxxxxx>]

Sorry it has taken me so long to reply...  The 4.11 kernel fixed these MCE events and then broke my graphics so I've been spending weeks trying to figure that out. 

Here's an update: Lenovo got back to me after I posted on their forums about it and said this was by design, to prevent the laptop from overheating if it accidentally wakes up inside a bag...  And as far as 0x6032, I am not able to see anything in Windows that limits the GPU (at least from the stats being reported in HWInfo.)

https://forums.lenovo.com/t5/ThinkPad-X-Series-Laptops/x270-thermal-throttling-to-60C-when-lid-is-closed-on-battery/td-p/3680557

Thanks,

Neil

On Sun, May 21, 2017 at 12:18 PM, neil k <host.crash@xxxxxxxxx> wrote:

Do you think I would have any luck contacting Lenovo about it?  It seems unlikely that I'd be able to find anyone at their technical support who could give me a real answer about this...

I'll see if I can monitor the GPU on Windows and see what happens. 

Here's an annotated version of the acpi_listen events:

neil@totbox ~ $ sudo acpi_listen

--lid closed on battery power - first time after boot where 0x6032 is triggered
processor LNXCPU:00 00000080 00000001
processor LNXCPU:01 00000080 00000001
processor LNXCPU:02 00000080 00000001
processor LNXCPU:03 00000080 00000001
ibm/hotkey LEN0268:00 00000080 00006032
button/lid LID close

--lid opened
ibm/hotkey LEN0268:00 00000080 00006032
button/lid LID open
jack/lineout LINEOUT unplug
jack/videoout VIDEOOUT unplug
jack/lineout LINEOUT plug
jack/videoout VIDEOOUT plug

--A/C adapter plugged in. thinkpad_acpi reports 0x6030 as "EC reports that Thermal Table has changed."  I think that only happens if I have the "thermal" settings different for battery and A/C in the BIOS.
battery PNP0C0A:01 00000080 00000001
ac_adapter ACPI0003:00 00000080 00000001
ibm/hotkey LEN0268:00 00000080 00006030
thermal_zone LNXTHERM:00 00000081 00000000
battery PNP0C0A:01 00000080 00000001

--lid closed - no 0x6032 on A/C
button/lid LID close

--lid opened on A/C - triggers 0x6032
ibm/hotkey LEN0268:00 00000080 00006032
button/lid LID open
jack/lineout LINEOUT unplug
jack/videoout VIDEOOUT unplug
jack/lineout LINEOUT plug
jack/videoout VIDEOOUT plug

On Sun, May 21, 2017 at 10:06 AM, Henrique de Moraes Holschuh <hmh@xxxxxxxxxx> wrote:

On Sun, 21 May 2017, neil k wrote:
> attachment.)  If I boot up the laptop with the lid closed, it won't start
> thermal throttling until after I trigger the event for the first time.  So

Some people broke the LID handling hideously for dubious reasons on
later kernel versions, and I don't know exactly when.  And the breakage
is exactly that they screwed up LID state at power up if the LID is
closed.  I won't go into the details because I am very likely to insult
people if I do.

So, in order to avoid a possible Linux bug, it is best to test it by
closing the LID *later*, so that Linux can see the LID getting closed.

The LID mess is apparently being handled upstream right now.

> that makes me think the throttling is 1. caused by the firmware and 2.
> definitely related to this event somehow.

Looks so.

> Considering that 60 degrees is nowhere near the critical temperature, and
> the fact that I can bypass the throttling by closing the lid on A/C then
> unplugging, booting up with the lid closed, etc. It seems to me like this
> is either a bug in the firmware or a poorly implemented "feature."  Is that

It likely has a few Linux bugs at play on how easily you can bypass the
"feature", on top of whatever bugs/corner cases Lenovo left unhandled.

However, it is very very clear that your laptop *requires* a much more
conservative thermal envelope while the LID is closed to safely operate.

> what it seems like to you too?  If there was something dangerous about
> exceeding 60 degrees with the lid closed, I'd expect it to throttle to that
> temperature regardless of whether it's plugged in at the time the lid is
> closed...

Yeah, it should be doing that throttling *always*.  The fact that it
isn't is a bug (likely caused by more than one underlying bug).

> Is there anything else I can do on Windows to help investigate what the lid
> helper event means?  I have it dual booting now, so I can test things in
> Windows or Linux.

Can you somehow test the GPU performance state?

It clearly uses standard ACPI tricks to limit the processor (which are
not exactly working right, but still...), so why would they have added
that extra private event?  It must be doing something else as well...

> neil@totbox ~ $ sudo acpi_listen
> processor LNXCPU:00 00000080 00000001
> processor LNXCPU:01 00000080 00000001
> processor LNXCPU:02 00000080 00000001
> processor LNXCPU:03 00000080 00000001
> ibm/hotkey LEN0268:00 00000080 00006032
> button/lid LID close
> ibm/hotkey LEN0268:00 00000080 00006032
> button/lid LID open
> jack/lineout LINEOUT unplug
> jack/videoout VIDEOOUT unplug
> jack/lineout LINEOUT plug
> jack/videoout VIDEOOUT plug
> battery PNP0C0A:01 00000080 00000001
> ac_adapter ACPI0003:00 00000080 00000001
> ibm/hotkey LEN0268:00 00000080 00006030

This 6030 seems related to the 6032...

> thermal_zone LNXTHERM:00 00000081 00000000
> battery PNP0C0A:01 00000080 00000001
> button/lid LID close
> ibm/hotkey LEN0268:00 00000080 00006032
> button/lid LID open

can you annotate this with your actions, e.g.:

powered up LID closed on A/C:
<events>
opened LID:
<events>
etc...

?

--
  Henrique Holschuh

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

_______________________________________________
ibm-acpi-devel mailing list
ibm-acpi-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel