Re: [PATCH v2 2/3] mm: huge_memory: refactor anon_enabled_store() with change_anon_orders()

[Breno Leitao <leitao@debian.org>]

On Fri, Mar 06, 2026 at 11:32:44AM +0000, Lorenzo Stoakes (Oracle) wrote:
> On Thu, Mar 05, 2026 at 06:04:54AM -0800, Breno Leitao wrote:
> > Signed-off-by: Breno Leitao <leitao@debian.org>
>
> This is nice, you've improved my idea :)

Thanks! Your idea was a good starting point, and digging deeper led me to
sysfs_match_string(), which I wasn't familiar with before — it turned out to
be exactly what we needed here.

Having the modes in anon_enabled_mode_strings[] also opens the door to further
simplification. I haven't thought it through completely yet, but I suspect we
could consolidate quite a few of the _store()/_show() helpers as a follow-up.

> I know you change another instance of this in 3/3, but I wonder if there are
> other cases like this in the THP code we can fixup too?

Yes, there are several other candidates — defrag_store, the ones Baolin Wang
reported, shmem_enabled_store(), and thpsize_shmem_enabled_store() among them.

I'd prefer to address those in a separate patch series once this one lands,
if that works for everyone.