Re: [PATCH] zram: revive swap_slot_free_notify

[Minchan Kim]

On Tue, Mar 22, 2016 at 05:20:08PM +0900, Joonsoo Kim wrote:
> 2016-03-22 17:00 GMT+09:00 Minchan Kim <minchan@xxxxxxxxxx>:
> > On Tue, Mar 22, 2016 at 02:08:59PM +0900, Joonsoo Kim wrote:
> >> On Fri, Mar 18, 2016 at 04:58:31PM +0900, Minchan Kim wrote:
> >> > <b430e9d1c6d4> "remove compressed copy from zram in-memory"
> >> > applied swap_slot_free_notify call in *end_swap_bio_read* to
> >> > remove duplicated memory between zram and memory.
> >> >
> >> > However, with introducing rw_page in zram <8c7f01025f7b>
> >> > "zram: implement rw_page operation of zram", it became void
> >> > because rw_page doesn't need bio.
> >> >
> >> > This patch restores the function for rw_page.
> >> >
> >> > Signed-off-by: Minchan Kim <minchan@xxxxxxxxxx>
> >> > ---
> >> >  mm/page_io.c | 93 ++++++++++++++++++++++++++++++++----------------------------
> >> >  1 file changed, 50 insertions(+), 43 deletions(-)
> >> >
> >> > diff --git a/mm/page_io.c b/mm/page_io.c
> >> > index ff74e512f029..18aac7819cc9 100644
> >> > --- a/mm/page_io.c
> >> > +++ b/mm/page_io.c
> >> > @@ -66,6 +66,54 @@ void end_swap_bio_write(struct bio *bio)
> >> >     bio_put(bio);
> >> >  }
> >> >
> >> > +static void swap_slot_free_notify(struct page *page)
> >> > +{
> >> > +   struct swap_info_struct *sis;
> >> > +   struct gendisk *disk;
> >> > +
> >> > +   /*
> >> > +    * There is no guarantee that the page is in swap cache - the software
> >> > +    * suspend code (at least) uses end_swap_bio_read() against a non-
> >> > +    * swapcache page.  So we must check PG_swapcache before proceeding with
> >> > +    * this optimization.
> >> > +    */
> >> > +   if (unlikely(!PageSwapCache(page)))
> >> > +           return;
> >> > +
> >> > +   sis = page_swap_info(page);
> >> > +   if (!(sis->flags & SWP_BLKDEV))
> >> > +           return;
> >> > +
> >> > +   /*
> >> > +    * The swap subsystem performs lazy swap slot freeing,
> >> > +    * expecting that the page will be swapped out again.
> >> > +    * So we can avoid an unnecessary write if the page
> >> > +    * isn't redirtied.
> >> > +    * This is good for real swap storage because we can
> >> > +    * reduce unnecessary I/O and enhance wear-leveling
> >> > +    * if an SSD is used as the as swap device.
> >> > +    * But if in-memory swap device (eg zram) is used,
> >> > +    * this causes a duplicated copy between uncompressed
> >> > +    * data in VM-owned memory and compressed data in
> >> > +    * zram-owned memory.  So let's free zram-owned memory
> >> > +    * and make the VM-owned decompressed page *dirty*,
> >> > +    * so the page should be swapped out somewhere again if
> >> > +    * we again wish to reclaim it.
> >> > +    */
> >> > +   disk = sis->bdev->bd_disk;
> >> > +   if (disk->fops->swap_slot_free_notify) {
> >> > +           swp_entry_t entry;
> >> > +           unsigned long offset;
> >> > +
> >> > +           entry.val = page_private(page);
> >> > +           offset = swp_offset(entry);
> >> > +
> >> > +           SetPageDirty(page);
> >> > +           disk->fops->swap_slot_free_notify(sis->bdev,
> >> > +                           offset);
> >> > +   }
> >> > +}
> >> > +
> >> >  static void end_swap_bio_read(struct bio *bio)
> >> >  {
> >> >     struct page *page = bio->bi_io_vec[0].bv_page;
> >> > @@ -81,49 +129,7 @@ static void end_swap_bio_read(struct bio *bio)
> >> >     }
> >> >
> >> >     SetPageUptodate(page);
> >> > -
> >> > -   /*
> >> > -    * There is no guarantee that the page is in swap cache - the software
> >> > -    * suspend code (at least) uses end_swap_bio_read() against a non-
> >> > -    * swapcache page.  So we must check PG_swapcache before proceeding with
> >> > -    * this optimization.
> >> > -    */
> >> > -   if (likely(PageSwapCache(page))) {
> >> > -           struct swap_info_struct *sis;
> >> > -
> >> > -           sis = page_swap_info(page);
> >> > -           if (sis->flags & SWP_BLKDEV) {
> >> > -                   /*
> >> > -                    * The swap subsystem performs lazy swap slot freeing,
> >> > -                    * expecting that the page will be swapped out again.
> >> > -                    * So we can avoid an unnecessary write if the page
> >> > -                    * isn't redirtied.
> >> > -                    * This is good for real swap storage because we can
> >> > -                    * reduce unnecessary I/O and enhance wear-leveling
> >> > -                    * if an SSD is used as the as swap device.
> >> > -                    * But if in-memory swap device (eg zram) is used,
> >> > -                    * this causes a duplicated copy between uncompressed
> >> > -                    * data in VM-owned memory and compressed data in
> >> > -                    * zram-owned memory.  So let's free zram-owned memory
> >> > -                    * and make the VM-owned decompressed page *dirty*,
> >> > -                    * so the page should be swapped out somewhere again if
> >> > -                    * we again wish to reclaim it.
> >> > -                    */
> >> > -                   struct gendisk *disk = sis->bdev->bd_disk;
> >> > -                   if (disk->fops->swap_slot_free_notify) {
> >> > -                           swp_entry_t entry;
> >> > -                           unsigned long offset;
> >> > -
> >> > -                           entry.val = page_private(page);
> >> > -                           offset = swp_offset(entry);
> >> > -
> >> > -                           SetPageDirty(page);
> >> > -                           disk->fops->swap_slot_free_notify(sis->bdev,
> >> > -                                           offset);
> >> > -                   }
> >> > -           }
> >> > -   }
> >> > -
> >> > +   swap_slot_free_notify(page);
> >> >  out:
> >> >     unlock_page(page);
> >> >     bio_put(bio);
> >> > @@ -347,6 +353,7 @@ int swap_readpage(struct page *page)
> >> >
> >> >     ret = bdev_read_page(sis->bdev, swap_page_sector(page), page);
> >> >     if (!ret) {
> >> > +           swap_slot_free_notify(page);
> >> >             count_vm_event(PSWPIN);
> >> >             return 0;
> >> >     }
> >>
> >> Hello,
> >
> > Hey Joonsoo,
> >
> >>
> >> You need to check PageUpdate() or something because bdev_read_page()
> >> can be asynchronous.
> >
> > I considered it but decided not to add the check :(.
> > Because I couldn't justify what benfit we can have with the check.
> > The swap_slot_free_notify is tightly coupled with zram for several
> > years and zram have been worked synchronously. So if bdev_read_page
> > returns 0, it means we already have read the page successfully.
> > Even, when I looked up other rw_page user, it seems there is no async
> > rw_page users at the moment.
> 
> Yes, I also looked up other rw_page users and found that
> there is no async rw_page now.
> 
> > If there is someone want to use *async* rw_page && *swap_slot_free_noity*
> > in future, we could add the check easily. But I hope anyone never use
> > swap_slot_free_notify any more which is mess. :(
> 
> But, I think that we should add the check. If someone want it, how does
> he/she know about it? Even, if someone makes zram to read/write
> asynchronously, we can miss it easily. This is error-prone practice.

Okay, I don't have strong against it.
If we really want to catch such case, let's add WARN_ON_ONCE.

diff --git a/mm/page_io.c b/mm/page_io.c
index 18aac7819cc9..6592893d16ca 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -353,6 +353,7 @@ int swap_readpage(struct page *page)
 
 	ret = bdev_read_page(sis->bdev, swap_page_sector(page), page);
 	if (!ret) {
+		WARN_ON_ONCE(!PageUptodate(page));
 		swap_slot_free_notify(page);
 		count_vm_event(PSWPIN);
 		return 0;

> 
> >>
> >> BTW, something like as swap_slot_free_notify() which invalidate
> >> backend of storage can also be possible for frontswap when
> >> frontswap_load() succeed. Isn't it?
> >
> > frontswap_tmem_exclusive_gets_enabled?
> 
> Wow... yes. that's what I try to find.
> Do you know the reason why zswap doesn't enable it?

Hmm, I couldn't remember. Maybe, it's not zswap stuff but frontswap stuff
so I guess zswap user can enable it via frontswap interface if he want.

> 
> Thanks.