Re: [GIT PULL] KVM changes for Linux 6.9 merge window

[Paolo Bonzini]

Retrying without HTML.

Paolo

Il 17 marzo 2024 14:34:02 CET, Paolo Bonzini <pbonzini@xxxxxxxxxx> ha scritto:
>[first time writing to lkml from phone so I hope the formatting isn't too bad]
>
>Il 17 marzo 2024 11:36:37 CET, Marc Zyngier <maz@xxxxxxxxxx> ha scritto:
>>On Sat, 16 Mar 2024 16:01:47 +0000,
>>Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>>> > You can also make CONFIG_KVM_ARM64_RES_BITS_PARANOIA depend on !COMPILE_TEST.
>>> 
>>> No.
>>> 
>>> WTF is wrong with you?
>>> 
>>> You're saying "let's turn off this compile-time sanity check when
>>> we're doing compile testing".
>>>     https://lore.kernel.org/linux-next/20240222220349.1889c728@xxxxxxxxxxxxxxxx/
>>> 
>>> and you're still trying to just *HIDE* this garbage?
>>> 
>>> Stop it.
>>
>>Well, if you really need to shout at someone, it should be me, as I
>>was the one who didn't get Stephen's hint last time.
>
>No problem with being shouted at, but "depends on !COMPILE_TEST" is actually something that *is* used for "maintainers will look at it, it shouldn't matter for linux-next compile testing". Most notably it's used for -Werror.
>
>When Stephen reported the failure, I should have noticed that the bandaid doesn't do anything to fix allyesconfig/allmodconfig. If there's anything I can blame you for, I thought/understood that you would be able to fix the failure between the report and the beginning of the merge window, so there's that small miscommunication but that's it.
>
>>I'll try to resurrect it as a selftest, or maybe just keep it out of
>>tree for my own use.
>
>I still believe that "depends on !COMPILE_TEST" is what you want here, but yeah keeping out of tree or even under a special make target is an option if Linus disagrees.
>
>Selftests have the advantage that they can be marked XFAIL, but I am not sure they're a good match here (also because the flip side is that I think XPASS fails the run).
>
>Paolo
Paolo