Re: [LSF/MM/BPF TOPIC] 64k (or 16k) base page size on x86

[Kiryl Shutsemau]

On Mon, Feb 23, 2026 at 12:27:33PM +0100, David Hildenbrand (Arm) wrote:
> On 2/23/26 12:13, Kiryl Shutsemau wrote:
> > On Mon, Feb 23, 2026 at 12:04:10PM +0100, David Hildenbrand (Arm) wrote:
> > > On 2/20/26 20:33, Kalesh Singh wrote:
> > > > On Fri, Feb 20, 2026 at 8:30 AM David Hildenbrand (Arm)
> > > > <david@xxxxxxxxxx> wrote:
> > > > 
> > > > I think most issues will stem from linkers setting the default ELF
> > > > segment alignment (max-page-size) for x86 to 4096. So those ELFs will
> > > > not load correctly or at all on the larger emulated granularity.
> > > 
> > > Right, I assume that they will have to be thought about that, and possibly,
> > > some binaries/libraries recompiled.
> > 
> > I think backward compatibility is important and I believe we can get
> > there without ABI break. And optimize from there.
> > 
> > BTW, x86-64 SysV ABI allows for 64k page size:
> > 
> > 	Systems are permitted to use any power-of-two page size between
> > 	4KB and 64KB, inclusive.
> > 
> > But it doesn't work in practice.
> 
> Even in well controlled environments you would run in a hyperscaler?

I have not invested much time into investigating this.

I intentionally targeted compatible version assuming it will be better
received by upstream. I want it to be usable outside specially cured
userspace. 64k might not be good fit for a desktop, but 16k can be a
different story.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov