On Thu, Mar 06, 2025 at 07:50:37AM +1100, Dave Chinner wrote:
> This may be one of the reasons why you don't see any change in real
> performance with 64kB directory blocks - we spend more time in
> folio allocation because of compaction overhead than we gain back
> from avoiding the use of vmapped buffers....
>
> i.e.
> if (size > PAGE_SIZE) {
> if (!is_power_of_2(size))
> goto fallback;
> gfp_mask ~= __GFP_DIRECT_RECLAIM;
> gfp_mask |= __GFP_NORETRY;
> }
> folio = folio_alloc(gfp_mask, get_order(size));
I'll give it a try.
