Re: [PATCH 2.43] malloc: Show hugetlb tunable default in --list-tunables

Hi Wilco,

On Fri, 2026-08-14 at 14:57 +0000, Wilco Dijkstra wrote:
> > > The backport looks fine - I can do the commit if you don't have write access.
> > 
> > sure, thank you for the quick reaction! (no, I don't have write access)
> 
> OK, committed as 7d26579873791748c20d874adb66e877412175dd
> 
> > I believe the pages are getting accessed, but the side effect comes on the systems
> > which always had
> > 
> > /sys/kernel/mm/transparent_hugepage/enabled: [always] madvise never
> > 
> > so madvise() syscall plays no role for them, but padding does.
> 
> If you set THP to [always], you explicitly state that you want hugepages to be used as
> much as possible. You will always get hugepages for large allocations or as a result of
> background page coalescing. Other allocators are far more aggressive than GLIBC -
> this work is reducing these obvious gaps.

which is expected and doesn't cause any wasted RAM as long as no extra padding on hugepage
boundary happens in the allocator.  

-- 
Alexander Sverdlin
Siemens AG
www.siemens.com