Re: [PATCH v2] mm/memory: Don't require head page for do_set_pmd()
From: Hugh Dickins
Date: Tue Aug 20 2024 - 16:34:25 EST
On Mon, 19 Aug 2024, Vincent Donnefort wrote:
> On Tue, Jun 11, 2024 at 07:22:03PM +0100, Matthew Wilcox wrote:
> > On Tue, Jun 11, 2024 at 11:06:22AM -0700, Andrew Morton wrote:
...
> > > You know what I'm going to ask ;) I'm assuming that the runtime effects
> > > are "small performance optimization" and that "should we backport the
> > > fix" is "no".
> >
> > We're going to stop using PMDs to map large folios unless the fault is
> > within the first 4KiB of the PMD. No idea how many workloads that
> > affects, but it only needs to be backported as far as v6.8, so we
> > may as well backport it.
>
> Hi, I am reviving this thread after noticing this comment attached
> to the fix.
>
> If you intend to install PTE level mappings for faults that happen outside of
> the first 4KiB, I believe this will make THP support for KVM ineffective.
You can relax, it's okay: where Matthew wrote "We're going to stop...",
he was describing the runtime effects of the bug (now fixed) to Andrew,
not proposing to make a change to mess up THP support.
The fix was backported to v6.9.N, but was too late for v6.8.N EOL.
Hugh