Re: [GIT PULL] thermal for v5.14-rc1

[Dmitry Osipenko]

04.07.2021 16:19, Daniel Lezcano пишет:
> On 04/07/2021 14:04, Dmitry Osipenko wrote:
>> 04.07.2021 12:51, Daniel Lezcano пишет:
>>> On 03/07/2021 21:56, Dmitry Osipenko wrote:
>>>> 03.07.2021 22:34, Linus Torvalds пишет:
>>>>> On Sat, Jul 3, 2021 at 9:17 AM Daniel Lezcano <daniel.lezcano@xxxxxxxxxx> wrote:
>>>>>>
>>>>>> ssh://git@xxxxxxxxxxxxxxxxxxx/pub/scm/linux/kernel/git/thermal/linux.git
>>>>>> tags/thermal-v5.14-rc1
>>>>>
>>>>> This does not build for me. And I suspect it never built in linux-next either.
>>>>>
>>>>> I get
>>>>>
>>>>>   ERROR: modpost: "tegra_fuse_readl"
>>>>> [drivers/thermal/tegra/tegra30-tsensor.ko] undefined!
>>>>>   ERROR: modpost: "tegra_fuse_readl"
>>>>> [drivers/thermal/tegra/tegra-soctherm.ko] undefined!
>>>>>   make[1]: *** [scripts/Makefile.modpost:150: modules-only.symvers] Error 1
>>>>>   make[1]: *** Deleting file 'modules-only.symvers'
>>>>>   make: *** [Makefile:1762: modules] Error 2
>>>>>
>>>>> and I think it's due to that commit 1f9c5936b10c
>>>>> ("thermal/drivers/tegra: Correct compile-testing of drivers") which
>>>>> quite sensibly tries to extend build coverage for the tegra thermal
>>>>> drivers, but that build coverage doesn't actually *work* outside the
>>>>> tegra world.
>>>>>
>>>>> That commit says "All Tegra thermal drivers support compile-testing",
>>>>> but clearly they stumble at the last hurdle.
>>>>>
>>>>> I made the decision to just unpull this, not because I couldn't fix
>>>>> it, but because if it was this untested, I don't want to worry about
>>>>> all the *other* code in there too.
>>>>
>>>> It was tested in linux-next for about two weeks and it was a known
>>>> problem already [1], sorry again. The compile-testing depends on the
>>>> patch [2] that should come with the ARM32 PR. The linux-next was okay
>>>> because [2] got first into -next and I forgot about it. It will be fine
>>>> to defer these Tegra patches till the next release if this is the
>>>> easiest solution.
>>>
>>> My bad, I did a wrong decision assuming the ARM32 branch would be
>>> already there when pulling the thermal branch.
>>>
>>> I'll remove the patches.
>>>
>>> Sorry for the inconvenience.
>>
>> Daniel, please let me know if I will need to resend the thermal patches
>> for 5.15.
> 
> I'll drop:
> 
> "thermal/drivers/tegra: Add driver for Tegra30 thermal sensor"
> "thermal/drivers/tegra: Correct compile-testing of drivers"
> 
> but keep the bindings.
> 
> It won't be necessary to resend them.

Great, thank you!