On 08.03.23 17:51, Peter Xu wrote:
On Wed, Mar 08, 2023 at 04:57:55PM +0100, David Hildenbrand wrote:
Oh, my thinking was that with markers you postpone placing the shared
zeropage. So the next read access will require a pagefault to map the shared
zeropage. Your v1 would have performed best in that case I guess.
Ah, so I misunderstood because the sentence being quote was in another
context I think..
Sorry, I might have messed up the quoting.
Pagemap may bring overhead constantly though, not sure when vs page faults.
May relevant to num of missing pages.
Right, one idea (raised by Dave G. IIRC) was to to the pagemap scanning
concurrently from another thread and avoid most of the shared-zeropage
dance.
--
Thanks,
David / dhildenb
