Re: [BUG mm-unstable] "kernel BUG at mm/swap.c:393!" on commit b9c91c43412f2e

[Yosry Ahmed]

On Wed, Jun 21, 2023 at 12:01 AM Hyeonggon Yoo <42.hyeyoo@xxxxxxxxx> wrote:
>
> On Wed, Jun 07, 2023 at 07:51:43PM +0000, Yosry Ahmed wrote:
> > Commit 71024cb4a0bf ("frontswap: remove frontswap_tmem_exclusive_gets")
> > removed support for exclusive loads from frontswap as it was not used.
> > Bring back exclusive loads support to frontswap by adding an "exclusive"
> > output parameter to frontswap_ops->load.
> >
> > On the zswap side, add a module parameter to enable/disable exclusive
> > loads, and a config option to control the boot default value.
> > Refactor zswap entry invalidation in zswap_frontswap_invalidate_page()
> > into zswap_invalidate_entry() to reuse it in zswap_frontswap_load() if
> > exclusive loads are enabled.
> >
> > With exclusive loads, we avoid having two copies of the same page in
> > memory (compressed & uncompressed) after faulting it in from zswap. On
> > the other hand, if the page is to be reclaimed again without being
> > dirtied, it will be re-compressed. Compression is not usually slow, and
> > a page that was just faulted in is less likely to be reclaimed again
> > soon.
> >
> > Suggested-by: Yu Zhao <yuzhao@xxxxxxxxxx>
> > Signed-off-by: Yosry Ahmed <yosryahmed@xxxxxxxxxx>
> > ---
> >
> > v1 -> v2:
> > - Add a module parameter to control whether exclusive loads are enabled
> >   or not, the config option now controls the default boot value instead.
> >   Replaced frontswap_ops->exclusive_loads by an output parameter to
> >   frontswap_ops->load() (Johannes Weiner).
> > ---
>
> Hi Yosry, I was testing the latest mm-unstable and encountered a bug.
> It was bisectable and this is the first bad commit.
>
>
> Attached config file and bisect log.
> The oops message is available at:
>
> https://social.kernel.org/media/eace06d71655b3cc76411366573e4a8ce240ad65b8fd20977d7c73eec9dc2253.jpg
>
> (the head commit is b9c91c43412f2e07 "mm: zswap: support exclusive loads")
> (it's an image because I tested it on real machine)
>
>
> This is what I have as swap space:
>
> $ cat /proc/swaps
> Filename                                Type            Size            Used            Priority
> /var/swap                               file            134217724       0               -2
> /dev/zram0                              partition       8388604         0               100


Hi Hyeonggon,

Thanks for reporting this! I think I know what went wrong. Could you
please verify if the below fix works if possible?

Domenico, I believe the below fix would also fix a problem with the
recent writeback series. If the entry is invalidated before we grab the
lock to put the local ref in zswap_frontswap_load(), then the entry
will be freed once we call zswap_entry_put(), and the movement to the
beginning LRU will be operating on a freed entry. It also modifies
your recently added commit 418fd29d9de5 ("mm: zswap: invaldiate entry
after writeback"). I would appreciate it if you also take a look.

If this works as intended, I can send a formal patch (applies on top
of fd247f029cd0 ("mm/gup: do not return 0 from pin_user_pages_fast()
for bad args")):