On Tue, Jan 24, 2023 at 7:54 AM Ricardo Koller <ricarkol@xxxxxxxxxx> wrote:
>
> On Mon, Jan 23, 2023 at 10:41:18AM -0800, David Matlack wrote:
> >
> > Ricardo, if you're interested in adding page size stats to KVM/ARM ahead
> > of the Common MMU, e.g. to test eager page splitting, let me know.
>
> Sure, I can do that. Sounds pretty useful too.
>
> > I
> > want to make sure we align on the userspace-visible stat names to avoid
> > churn down the road. Specifically, do we want to expose neutral names
> > like pages_{pte,pmd,pud} or expand the KVM/x86 list to include all of
> > ARM's possible pages sizes like pages_{4k,16k,64k,...} (or both)?
>
> I would prefer the latter, mainly to match the x86 names:
>
> + stats->pages_4k = vm_get_stat(vm, "pages_4k");
> + stats->pages_2m = vm_get_stat(vm, "pages_2m");
> + stats->pages_1g = vm_get_stat(vm, "pages_1g");
> (from this patch)
We can always add pages_{pte,pmd,pud} to x86 as aliases of the
existing stats. The series I recently sent out to allow custom names
for stats [1] would make adding the aliases trivial actually.
[1] https://lore.kernel.org/kvm/20230118175300.790835-1-dmatlack@xxxxxxxxxx/
>
> but pages_{pte,pmd,pud} would certainly make this test simpler
> as it would handle all guest page sizes:
>
> + stats->pages_pte = vm_get_stat(vm, "pages_pte");
>
Yeah pages_{pte,pmd,pud} would certainly make the test simpler.
At this point I'm leaning toward pages_{pte,pmd,pud} to unblock this
testing in an architecture neutral way, and we can add
pages_{4k,16k,...} to ARM in the future using aliases.
