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

[Nadia Chambers]

On Thu, 19 Feb 2026 15:08:51 +0000 Kiryl Shutsemau <kas@xxxxxxxxxx> wrote:
>> No, there's no new hardware (that I know of). I want to explore what
>> page size means.
>> The kernel uses the same value - PAGE_SIZE - for two things:
>>   - the order-0 buddy allocation size;
>>   - the granularity of virtual address space mapping;

Am Fr, Feb 20, 2026 um 09:04:09 +0000, David Laight schrieb:
> Also the 'random' buffers that are PAGE_SIZE rather than 4k.
> I also wonder how is affects mmap of kernel memory and the alignement
> of PCIe windows (etc).

It doesn't. Any adequate implementation has to sweep drivers and update
them to use MMUPAGE_SIZE or other address calculations that are still
correct for device memory, buffers need sizing so they work usw.

Yes, those things are there. Sweeping them is integral to implementing
page clustering, whether it was Hugh, me, Andrea or Kiryl.


-- nyc