Re: [PATCH RFC 0/4] KVM: MIPS: Provide arch-specific kvm_flush_remote_tlbs()

[Paolo Bonzini]

On 12/02/20 17:30, Paul Burton wrote:
> Hi Paolo,
> 
> On Wed, Feb 12, 2020 at 01:25:30PM +0100, Paolo Bonzini wrote:
>> MIPS folks, I see that arch/mips/kvm/mmu.c uses pud_index, so it's not
>> clear to me if it's meant to only work if CONFIG_PGTABLE_LEVELS=4 or
>> it's just bit rot.  Should I add a "depends on PGTABLE_LEVEL=4" to
>> arch/mips/Kconfig?
> 
> I'm no expert on this bit of code, but I'm pretty sure the systems
> KVM/VZ has been used on the most internally had PGTABLE_LEVEL=3.
> 
> I suspect this is actually a regression from commit 31168f033e37 ("mips:
> drop __pXd_offset() macros that duplicate pXd_index() ones"). Whilst
> that commit is correct that pud_index() & __pud_offset() are the same
> when pud_index() is actually provided, it doesn't take into account the
> __PAGETABLE_PUD_FOLDED case. There __pud_offset() was available but
> would always evaluate to zero, whereas pud_index() isn't defined...

Ok, I'll try to whip out a patch that handles __PAGETABLE_PUD_FOLDED.
On the other hand this makes me worry about how much KVM is being tested
by people that care about MIPS (even just compile-tested).

Paolo