Re: [PATCH] mm: Optimize TLB flushes during page reclaim

[Rik van Riel]

On Thu, 2025-01-23 at 13:16 -0600, Vinay Banakar wrote:
> 
> We have two options:
> 1. Keep the current logic where TLB flush batching varies by caller
> 2. Enforce consistent 512-page batching in shrink_folio_list() and
> also convert to folio_batch as suggested by Matthew
> 

Given that your code is correct, and only the comment
was a little misleading, would you be open to resending
your patch without the comments about "PMD level"
batching?

It would be really nice to get this optimization into
the kernel.

If you don't have the time currently, I would also be
happy to clean up the patch and get it to Andrew.
Just let me know.

kind regards,

Rik
-- 
All Rights Reversed.