Re: [PATCH v3 02/12] softmmu/physmem: Fix ram_block_discard_range() to handle shared anonymous memory

[David Hildenbrand]

On 11.03.21 22:37, Peter Xu wrote:
> On Mon, Mar 08, 2021 at 04:05:50PM +0100, David Hildenbrand wrote:
> > We can create shared anonymous memory via
> >      "-object memory-backend-ram,share=on,..."
> > which is, for example, required by PVRDMA for mremap() to work.
> >
> > Shared anonymous memory is weird, though. Instead of MADV_DONTNEED, we
> > have to use MADV_REMOVE. MADV_DONTNEED fails silently and does nothing.
> >
> > Fixes: 06329ccecfa0 ("mem: add share parameter to memory-backend-ram")
>
> I'm thinking whether we should keep this fixes - it's valid, however it could
> unveil issues if those remapped ranges didn't get unmapped in time.  After all
> "not releasing some memory existed" seems not a huge deal for stable.  No
> strong opinion, just raise it up as a pure question.

If someone would be using it along with postcopy (which should work 
apart from that issue) you could be in trouble. That's why i think at 
least the Fixes: tag is valid. CC: stable might be debatable indeed.

> > Signed-off-by: David Hildenbrand <david@redhat.com>
>
> Reviewed-by: Peter Xu <peterx@redhat.com>

Thanks!

--
Thanks,

David / dhildenb