Re: [PATCH v1 2/4] mm: smaller folio_pte_batch() improvements

[David Hildenbrand <david@redhat.com>]

On 02.07.25 11:08, Lorenzo Stoakes wrote:
> On Wed, Jul 02, 2025 at 11:00:48AM +0200, David Hildenbrand wrote:
>> On 02.07.25 10:51, Lorenzo Stoakes wrote:
>>> On Wed, Jul 02, 2025 at 10:48:20AM +0200, David Hildenbrand wrote:
>>>> On 02.07.25 10:42, Oscar Salvador wrote:
>>>>> On Fri, Jun 27, 2025 at 01:55:08PM +0200, David Hildenbrand wrote:
>>>>>> Let's clean up a bit:
>>>>>>
>>>>>> (1) No need for start_ptep vs. ptep anymore, we can simply use ptep
>>>>>>
>>>>>> (2) Let's switch to "unsigned int" for everything
>>>>>>
>>>>>> (3) We can simplify the code by leaving the pte unchanged after the
>>>>>>        pte_same() check.
>>>>>>
>>>>>> (4) Clarify that we should never exceed a single VMA; it indicates a
>>>>>>        problem in the caller.
>>>>>>
>>>>>> No functional change intended.
>>>>>>
>>>>>> Signed-off-by: David Hildenbrand <david@redhat.com>
>>>>>
>>>>> Hi David :-),
>>>>>
>>>>> I have to confess that I fell in the same trap as Lorenzo wrt.
>>>>> __pte_batch_clear_ignored changing the pte value.
>>>>> So I'm not sure if it would be nice to place a little comment in
>>>>> __pte_batch_clear_ignored claryfing that pte's value remains unchanged ?
>>>>
>>>> I mean, that's how all our pte modification functions work, really? :)
>>>>
>>>> Thanks!
>>>
>>> I mean, it might be that me and Oscar are similarly 'challenged' in this
>>> respect :P (high 5 Oscar!) but I think the issue here is that it's sort of
>>> a compounded use, and in fact some functions do modify stuff, which is why
>>> we end up with all the ptep ptent etc. fun.
>>>
>>> Up to you re: comment, but I think maybe in cases where it's a reallly
>>> compounded set of stuff it's potentially useful.
>>>
>>> But obviously we still do do this all over the place elsewhere with no
>>> comment...
>>
>> Well, if you are not passing in a *value* and not a pointer to a function,
>> you would not expect for that *value* to change? :)
>>
>> Yes, once we pass pointers it's different. Or when we're using weird macros.
>>
>> Adding a comment that a function will not modify a value that is ...
>> passed-by-value? Maybe it's just me that doesn't get why that should be
>> particularly helpful :)
> 
> I think the issue is that we've passed around 'pte' as value and pointer (and of
> course, via macros...)  previously so that's the cause of the confusion, often.
> 
> This is why I really am a fan of us consistently saying ptep when passing a
> pointer.

100%: pte for pointers is absolutely nasty.

-- 
Cheers,

David / dhildenb