Re: hwpstate_intel hangs kernel

From: Conrad Meyer <cem_at_freebsd.org>
Date: Tue, 11 Feb 2020 12:05:19 -0800
Hi Andreas,

No, nothing new on this front.  I don't have any of this hardware nor
do I work for Intel, and nothing I see in the code or published spec
would result in hangs.  I don't have more cycles to spend on this
driver.  So I suggest using 'hint.hwpstate_intel.0.disabled="1"' as a
workaround indefinitely.

On a private email with Andrey I discussed making disabled="1" the
default, and allowing users to opt in with disabled="0" (or perhaps
enabled="1", for fewer double negatives).  It's not especially
satisfying, but would prevent this class of unknown issue with
hwpstate_intel(4).

Best,
Conrad

On Tue, Feb 11, 2020 at 11:49 AM Andreas Nilsson <andrnils_at_gmail.com> wrote:
>
> Hello,
>
> any new leads on this? As of what is in git on 2020-02-05, my computer still hangs on kernel boot on "hwpstate_intel0: <Intel Speed Shift> on cpu0"
>
> I guess it would have been easier to debug a panic. Setting debug.hwpstate_verbose=1 doesn't really give any difference in output.
>
> Best regards
> Andreas
>
> On Wed, Feb 5, 2020 at 2:45 PM Andreas Nilsson <andrnils_at_gmail.com> wrote:
>>
>> Hello,
>>
>> I upgraded to a newer version,  git 87d669d3863-c266265, and I do not experience the random hang anymore. The machine still hangs on boot on "hwpstate_intel0: <Intel Speed Shift> on cpu0" unless I set 'hint.hwpstate_intel.0.disabled="1"' in loader.conf.
>>
>> On a side note I cannot set or unset that hint from loader prompt;
>>
>> ok> set hint.hwpstate_intel.0.disabled="0"
>> ok> show
>> ....
>> hint.hwpstate=
>> ...
>>
>> Best regards
>> Andreas
>>
>> On Sat, Feb 1, 2020 at 11:26 PM Andreas Nilsson <andrnils_at_gmail.com> wrote:
>>>
>>> Hello Conrad,
>>>
>>> thank you Andrey for bisecting! I'll try with that hint and see how it works for me.
>>>
>>> Best regards
>>> Andreas
>>>
>>> On Sat, Feb 1, 2020, 18:18 Conrad Meyer <cem_at_freebsd.org> wrote:
>>>>
>>>> Hi Andrey,
>>>>
>>>> Please try 'hint.hwpstate_intel.0.disabled="1"' as a workaround for now.
>>>>
>>>> I think I have identified at least one problematic piece of code,
>>>> although I don't know if it's the root cause.  I will go ahead and fix
>>>> that, which may not fix the hang, and also add some debug printfs that
>>>> can be enabled to help identify the real issue.
>>>>
>>>> Thanks for the report and bisect.
>>>>
>>>> Best,
>>>> Conrad
>>>>
>>>> On Sat, Feb 1, 2020 at 6:06 AM Andrey V. Elsukov <bu7cher_at_yandex.ru> wrote:
>>>> >
>>>> > 31.01.2020 18:11, Andrey V. Elsukov пишет:
>>>> > > On 24.01.2020 19:52, Andreas Nilsson wrote:
>>>> > >> It hangs during kernel boot and the last message printed on console is:
>>>> > >> hwpstate_intel0: <Intel Speed Shift> on cpu0
>>>> > >
>>>> > > Hi,
>>>> > >
>>>> > > Did you find the cause of this hang?
>>>> > > I also tried to update today from r350816 to r357330. But my Lenovo X1
>>>> > > Carbon 4th hangs on the same message.
>>>> > >
>>>> >
>>>> > Hi,
>>>> >
>>>> > I have bisected the bad commit, it is r357002.
>>>> >
>>>> > --
>>>> > WBR, Andrey V. Elsukov
>>>> >
Received on Tue Feb 11 2020 - 19:05:33 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:23 UTC