Re: [PATCH v2 1/3] arm64/mm: Refactor PMD_PRESENT_INVALID and PTE_PROT_NONE bits

[David Hildenbrand <david@redhat.com>]

On 30.04.24 14:53, Ryan Roberts wrote:
> On 30/04/2024 12:37, David Hildenbrand wrote:
>> On 30.04.24 13:11, Catalin Marinas wrote:
>>> On Mon, Apr 29, 2024 at 06:15:45PM +0100, Ryan Roberts wrote:
>>>> On 29/04/2024 17:20, Catalin Marinas wrote:
>>>>> On Mon, Apr 29, 2024 at 03:02:05PM +0100, Ryan Roberts wrote:
>>>>>> diff --git a/arch/arm64/include/asm/pgtable-prot.h
>>>>>> b/arch/arm64/include/asm/pgtable-prot.h
>>>>>> index dd9ee67d1d87..de62e6881154 100644
>>>>>> --- a/arch/arm64/include/asm/pgtable-prot.h
>>>>>> +++ b/arch/arm64/include/asm/pgtable-prot.h
>>>>>> @@ -18,14 +18,7 @@
>>>>>>    #define PTE_DIRTY        (_AT(pteval_t, 1) << 55)
>>>>>>    #define PTE_SPECIAL        (_AT(pteval_t, 1) << 56)
>>>>>>    #define PTE_DEVMAP        (_AT(pteval_t, 1) << 57)
>>>>>> -#define PTE_PROT_NONE        (_AT(pteval_t, 1) << 58) /* only when
>>>>>> !PTE_VALID */
>>>>>> -
>>>>>> -/*
>>>>>> - * This bit indicates that the entry is present i.e. pmd_page()
>>>>>> - * still points to a valid huge page in memory even if the pmd
>>>>>> - * has been invalidated.
>>>>>> - */
>>>>>> -#define PMD_PRESENT_INVALID    (_AT(pteval_t, 1) << 59) /* only when
>>>>>> !PMD_SECT_VALID */
>>>>>> +#define PTE_INVALID        (_AT(pteval_t, 1) << 59) /* only when
>>>>>> !PTE_VALID */
>>>>>
>>>>> Nitpick - I prefer the PTE_PRESENT_INVALID name as it makes it clearer
>>>>> it's a present pte. We already have PTE_VALID, calling it PTE_INVALID
>>>>> looks like a negation only.
>>>>
>>>> Meh, for me the pte can only be valid or invalid if it is present. So it's
>>>> implicit. And if you have PTE_PRESENT_INVALID you should also have
>>>> PTE_PRESENT_VALID.
>>>>
>>>> We also have pte_mkinvalid(), which is core-mm-defined. In your scheme, surely
>>>> it should be pte_mkpresent_invalid()?
>>>>
>>>> But you're the boss, I'll change this to PTE_PRESENT_INVALID. :-(
>>>
>>> TBH, I don't have a strong opinion but best to avoid the bikeshedding.
>>> I'll leave the decision to you ;). It would match the pmd_mkinvalid()
>>> core code. But if you drop 'present' make sure you add a comment above
>>> that it's meant for present ptes.
>>
>> FWIW, I was confused by
>>
>> present = valid | invalid
> 
> OK fair enough.
> 
>>
>> Something like
>>
>> present = present_valid | present_invalid
> 
> I don't want to change pte_valid() to pte_present_valid(); that would also be a
> fair bit of churn.

Yes.

> 
> I'll take Catalin's suggestion and make this PTE_PRESENT_INVALID and
> pte_present_invalid(). And obviously leave pmd_mkinvalid() as it is.
> (Conversation in the other thread has concluded that it's ok to invalidate a
> non-present pmd afterall).

Works for me.

-- 
Cheers,

David / dhildenb