On Fri, Dec 18, 2020 at 8:08 AM Michal Hocko mhocko@suse.com wrote:
On Fri 18-12-20 07:24:53, Pavel Tatashin wrote:
On Fri, Dec 18, 2020 at 4:43 AM Michal Hocko mhocko@suse.com wrote:
On Thu 17-12-20 13:52:38, Pavel Tatashin wrote:
* 1. Pinned pages: (long-term) pinning of movable pages is avoided* when pages are pinned and faulted, but it is still possible that* address space already has pages in ZONE_MOVABLE at the time when* pages are pinned (i.e. user has touches that memory before* pinning). In such case we try to migrate them to a different zone,* but if migration fails the pages can still end-up pinned in* ZONE_MOVABLE. In such case, memory offlining might retry a long* time and will only succeed once user application unpins pages.I still dislike this. Pinning can fail so there shouldn't be any reasons to break MOVABLE constrain for something that can be handled. If anything there should be a very good reasoning behind this decision documented.
This is basically current behaviour, after patch 8, we can never pin pages in the movable zone, so I will update this comment in that patch.
Then it would be much easier for review to state that the existing behavior is unchanged and do not update this comment just to remove it in a later patch. Because this patch should be straightforward change of the condition which pages to migrate (+some renaming which should be reasonably easy to follow).
Makes sense, I will update this comment correctly right away when the behaviour changes.