[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250805142446.GN184255@nvidia.com>
Date: Tue, 5 Aug 2025 11:24:46 -0300
From: Jason Gunthorpe <jgg@...dia.com>
To: David Hildenbrand <david@...hat.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Alex Williamson <alex.williamson@...hat.com>,
"kvm@...r.kernel.org" <kvm@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"lizhe.67@...edance.com" <lizhe.67@...edance.com>
Subject: Re: [GIT PULL] VFIO updates for v6.17-rc1
On Tue, Aug 05, 2025 at 04:22:32PM +0200, David Hildenbrand wrote:
> On 05.08.25 16:20, Jason Gunthorpe wrote:
> > On Tue, Aug 05, 2025 at 04:10:45PM +0200, David Hildenbrand wrote:
> > > There are some weird scenarios where you hotplug memory after boot memory,
> > > and suddenly you can runtime-allocate a gigantic folio that spans both
> > > ranges etc.
> >
> > I was thinking we'd forbid this directly, but yes it is a another new
> > check.
> >
> > > So while related, the corner cases are all a bit nasty, and just forbidding
> > > folios to span a memory section on these problematic configs (sparse
> > > !vmemmap) sounds interesting.
> >
> > Indeed, this just sounds like forcing MAX_ORDER to be no larger than
> > the section size for this old mode?
>
> MAX_ORDER is always limited to the section size already.
>
> MAX_ORDER is only about buddy allocations. What hugetlb and dax do is
> independent of MAX_ORDER.
Oh I thought it limited folios too.
Still same idea is to have a MAX_FOLIO_ORDER for that case.
Jason
Powered by blists - more mailing lists