Re: [PATCH] mm/huge_memory: consolidate order-related checks into folio_split_supported()

[Wei Yang <richard.weiyang@gmail.com>]

On Sat, Nov 15, 2025 at 05:07:45AM +0000, Matthew Wilcox wrote:
>On Sat, Nov 15, 2025 at 02:51:09AM +0000, Wei Yang wrote:
>> I am not 100% for sure. While if we trust it returns 0 if folio doesn't
>> support large folio, I am afraid we can trust it returns correct value it
>> supports. Or that would be the issue to related file system.
>> 
>> To be honest, I am lack of the capability to investigate all file system to
>> make sure this value is properly setup.
>
>Maybe you should just give up on this instead of asking so many
>questions, misunderstanding the answers, and sending more patches full
>of mistakes?

Hi, Matthew

I am glad to see your reply, but your comment makes me feel frustrated.

Well, I think we still talk about the fact. You pointed out three flaws:

  * too many questions
  * too many misunderstandings
  * full of mistakes

Per my understanding, there is only one question: whether
mapping_max_folio_order() works universally. I think it does and also Zi
suggested to use it. But since I am not sure about this, we can revert it to
keep the logic same as current.

For misunderstandings, I use the word to be polite and avoid conflict. I
went through the discussion with Zi and thought I did what he suggested. In
case I do miss something, Zi would correct me.

For full of mistakes, I am confused. I did some contributions in kernel, small
of course. Mostly correct and get merged. I know sometimes I made mistakes,
but full of mistakes, it seems not to be true.

Well, I still think this is a common review process, I am open to comment and
questions. And also glad to take suggestions from community. I still think it
is reasonable to combine the EINVAL handling, while the detail need to be
settled down. 

Have a nice weekend all.

-- 
Wei Yang
Help you, Help me