Re: [PATCH 1/3] mm: Move arch_do_swap_page() call to before swap_free()

[Catalin Marinas]

On Fri, May 19, 2023 at 11:21:35AM +0200, David Hildenbrand wrote:
> > > Sorry, I meant actual anonymous memory pages, not shmem. Like, anonymous
> > > pages that are COW-shared due to fork() or KSM.
> > > 
> > > How does MTE, in general, interact with that? Assume one process ends up
> > > modifying the tags ... and the page is COW-shared with a different
> > > process that should not observe these tag modifications.
> > 
> > Tag modifications cause write faults if the page is read-only, so for
> > COW shared pages we would end up copying the page in the usual way,
> > which on arm64 would copy the tags as well via the copy_highpage hook
> > (see arch/arm64/mm/copypage.c).
> 
> Oh, that makes sense, thanks for pointing that out!
> 
> ... and I can spot that KSM also checks the tag when de-duplicating:
> pages_identical() ends up calling memcmp_pages(), which knows how to deal
> with tags.
> 
> Interestingly, calc_checksum() does not seem to care about tags. But that
> simply implies that pages with the same content have same checksum,
> independent of the tag. And pages_identical() is the single source of truth.

That was my assumption at the time, there would be a memcmp_pages() in
case of checksum collision.

-- 
Catalin