On Thu 10-12-20 06:21:57, Johannes Weiner wrote:
> On Wed, Dec 09, 2020 at 10:18:22AM +0100, Vlastimil Babka wrote:
> > On 12/9/20 2:24 AM, Yu Zhao wrote:
> > > We are capable of SetPageWorkingset based on refault distances after
> > > commit aae466b0052e ("mm/swap: implement workingset detection for anonymous LRU")
> > > This is done by workingset_refault(), which is right above the
> > > unconditional SetPageWorkingset deleted by this patch.
> > >
> > > The unconditional SetPageWorkingset miscategorizes pages that are
> > > read ahead or never belonged to the working set (e.g., tmpfs pages
> > > accessed by fd). When those pages are swapped in (after they were
> > > swapped out) for the first time, they skew PSI (when using
> > > async swap). When this happens again, depending on their refault
> > > distances, they might skew workingset_restore_anon counter in
> > > addition to PSI because their shadows say they were part of the
> > > working set.
> > >
> > > Signed-off-by: Yu Zhao <yuzhao@xxxxxxxxxx>
> >
> > Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
> >
> > Makes sense, especially now that we have anonymous LRU support. The flag setting
> > in this context seems to go back all the way to 1899ad18c607 ("mm: workingset:
> > tell cache transitions from workingset thrashing") where I'm not sure why it was
> > even used on the anonymous page, when workingset was only implemented for the
> > page cache. Maybe Johannes remembers?
>
> I just double checked that commit and the changelog is indeed
> incomplete and doesn't mention the swap aspect. :(
>
> That patch was part of the psi series. It was meant to mark incoming
> pages under IO with SetPageWorkingset when waiting for them
> constituted a memory stall.
>
> On the page cache side, because we HAVE workingset detection, this was
> specific to recently evicted pages that had been active in their
> previous life. On the anon side, the aging algorithm had no
> distinction between workingset and sporadically used pages. Given the
> choice between a) no swapin stalls are pressure and b) all swapin
> stalls are pressure, I went with the latter in order to detect swap
> storms. The false positive case - high rate of swapin without severe
> memory pressure - was relatively unlikely, because we tried to avoid
> swapping until everything was completely on fire in the first place.
>
> With the lru balancing rework, more prevalent use of proactive reclaim
> etc. the distinction between hot and cold swapins became more
> important. Thankfully, Joonsoo's patches made that possible.
This is a useful information, thanks! Yu Zhao can you make it into the
changelog so that we have it for a future reference please?
Feel free to add
Acked-by: Michal Hocko <mhocko@xxxxxxxx>
--
Michal Hocko
SUSE Labs
