Re: [PATCH v2 1/2] zswap: make shrinking memcg-aware

[Yosry Ahmed]

On Wed, Sep 27, 2023 at 2:02 PM Johannes Weiner <hannes@xxxxxxxxxxx> wrote:
>
> On Wed, Sep 27, 2023 at 09:48:10PM +0200, Domenico Cerasuolo wrote:
> > > > @@ -485,6 +487,17 @@ struct page *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask,
> > > >         __folio_set_locked(folio);
> > > >         __folio_set_swapbacked(folio);
> > > >
> > > > +       /*
> > > > +        * Page fault might itself trigger reclaim, on a zswap object that
> > > > +        * corresponds to the same swap entry. However, as the swap entry has
> > > > +        * previously been pinned, the task will run into an infinite loop trying
> > > > +        * to pin the swap entry again.
> > > > +        *
> > > > +        * To prevent this from happening, we remove it from the zswap
> > > > +        * LRU to prevent its reclamation.
> > > > +        */
> > > > +       zswap_lru_removed = zswap_remove_swpentry_from_lru(entry);
> > > > +
> > >
> > > This will add a zswap lookup (and potentially an insertion below) in
> > > every single swap fault path, right?. Doesn't this introduce latency
> > > regressions? I am also not a fan of having zswap-specific details in
> > > this path.
> > >
> > > When you say "pinned", do you mean the call to swapcache_prepare()
> > > above (i.e. setting SWAP_HAS_CACHE)? IIUC, the scenario you are
> > > worried about is that the following call to charge the page may invoke
> > > reclaim, go into zswap, and try to writeback the same page we are
> > > swapping in here. The writeback call will recurse into
> > > __read_swap_cache_async(), call swapcache_prepare() and get EEXIST,
> > > and keep looping indefinitely. Is this correct?
>
> Yeah, exactly.
>
> > > If yes, can we handle this by adding a flag to
> > > __read_swap_cache_async() that basically says "don't wait for
> > > SWAP_HAS_CACHE and the swapcache to be consistent, if
> > > swapcache_prepare() returns EEXIST just fail and return"? The zswap
> > > writeback path can pass in this flag and skip such pages. We might
> > > want to modify the writeback code to put back those pages at the end
> > > of the lru instead of in the beginning.
> >
> > Thanks for the suggestion, this actually works and it seems cleaner so I think
> > we'll go for your solution.
>
> That sounds like a great idea.
>
> It should be pointed out that these aren't perfectly
> equivalent. Removing the entry from the LRU eliminates the lock
> recursion scenario on that very specific entry.
>
> Having writeback skip on -EEXIST will make it skip *any* pages that
> are concurrently entering the swapcache, even when it *could* wait for
> them to finish.
>
> However, pages that are concurrently read back into memory are a poor
> choice for writeback anyway, and likely to be removed from swap soon.
>
> So it happens to work out just fine in this case. I'd just add a
> comment that explains the recursion deadlock, as well as the
> implication of skipping any busy entry and why that's okay.

Good point, we will indeed skip even if the concurrent insertion from
the swapcache is coming from a different cpu.

As you said, it works out just fine in this case, as the page will be
removed from zswap momentarily anyway. A comment is indeed due.