Re: [RFCv2 1/6] mm: introduce MADV_COLD

[Minchan Kim]

On Mon, Jun 03, 2019 at 09:16:07AM +0200, Michal Hocko wrote:
> On Fri 31-05-19 23:34:07, Minchan Kim wrote:
> > On Fri, May 31, 2019 at 04:03:32PM +0200, Michal Hocko wrote:
> > > On Fri 31-05-19 22:39:04, Minchan Kim wrote:
> > > > On Fri, May 31, 2019 at 10:47:52AM +0200, Michal Hocko wrote:
> > > > > On Fri 31-05-19 15:43:08, Minchan Kim wrote:
> > > > > > When a process expects no accesses to a certain memory range, it could
> > > > > > give a hint to kernel that the pages can be reclaimed when memory pressure
> > > > > > happens but data should be preserved for future use.  This could reduce
> > > > > > workingset eviction so it ends up increasing performance.
> > > > > > 
> > > > > > This patch introduces the new MADV_COLD hint to madvise(2) syscall.
> > > > > > MADV_COLD can be used by a process to mark a memory range as not expected
> > > > > > to be used in the near future. The hint can help kernel in deciding which
> > > > > > pages to evict early during memory pressure.
> > > > > > 
> > > > > > Internally, it works via deactivating pages from active list to inactive's
> > > > > > head if the page is private because inactive list could be full of
> > > > > > used-once pages which are first candidate for the reclaiming and that's a
> > > > > > reason why MADV_FREE move pages to head of inactive LRU list. Therefore,
> > > > > > if the memory pressure happens, they will be reclaimed earlier than other
> > > > > > active pages unless there is no access until the time.
> > > > > 
> > > > > [I am intentionally not looking at the implementation because below
> > > > > points should be clear from the changelog - sorry about nagging ;)]
> > > > > 
> > > > > What kind of pages can be deactivated? Anonymous/File backed.
> > > > > Private/shared? If shared, are there any restrictions?
> > > > 
> > > > Both file and private pages could be deactived from each active LRU
> > > > to each inactive LRU if the page has one map_count. In other words,
> > > > 
> > > >     if (page_mapcount(page) <= 1)
> > > >         deactivate_page(page);
> > > 
> > > Why do we restrict to pages that are single mapped?
> > 
> > Because page table in one of process shared the page would have access bit
> > so finally we couldn't reclaim the page. The more process it is shared,
> > the more fail to reclaim.
> 
> So what? In other words why should it be restricted solely based on the
> map count. I can see a reason to restrict based on the access
> permissions because we do not want to simplify all sorts of side channel
> attacks but memory reclaim is capable of reclaiming shared pages and so
> far I haven't heard any sound argument why madvise should skip those.
> Again if there are any reasons, then document them in the changelog.

I will go with removing the part so that defer to decision to the VM reclaim
based on the review.

>  
> [...]
> 
> > > Please document this, if this is really a desirable semantic because
> > > then you have the same set of problems as we've had with the early
> > > MADV_FREE implementation mentioned above.
> > 
> > IIRC, the problem of MADV_FREE was that we couldn't discard freeable
> > pages because VM never scan anonymous LRU with swapless system.
> > However, it's not the our case because we should reclaim them, not
> > discarding.
> 
> Right. But there is still the page cache reclaim. Is it expected that
> an explicitly cold memory doesn't get reclaimed because we have a
> sufficient amount of page cache (a very common case) and we never age
> anonymous memory because of that?

If there are lots of used-once pages in file-LRU, I think there is no
need to reclaim anonymous pages because it needs bigger overhead due to
IO. It has been true for a long time in current VM policy.

Reclaim preference model based on hints is as following based on cost:

MADV_DONTNEED >> MADV_PAGEOUT > used-once pages > MADV_FREE >= MADV_COLD

It is desirable for the new hints to be placed in the reclaiming preference
order such that a) they don't overlap functionally with existing hints and
b) we have a balanced ordering of disruptive and non-disruptive hints.