[PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Barry Song]

From: Barry Song <v-songbaohua@oppo.com>

If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
zswap_never_enabled() should return true.

Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
Signed-off-by: Barry Song <v-songbaohua@oppo.com>
Cc: Yosry Ahmed <yosryahmed@google.com>
Cc: Nhat Pham <nphamcs@gmail.com>
Cc: Chengming Zhou <chengming.zhou@linux.dev>
Cc: Chris Li <chrisl@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/zswap.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/zswap.h b/include/linux/zswap.h
index bf83ae5e285d..6cecb4a4f68b 100644
--- a/include/linux/zswap.h
+++ b/include/linux/zswap.h
@@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
 
 static inline bool zswap_never_enabled(void)
 {
-	return false;
+	return true;
 }
 
 #endif
-- 
2.34.1

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Chengming Zhou]

On 2024/6/30 07:22, Barry Song wrote:
> From: Barry Song <v-songbaohua@oppo.com>
> 
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
> 
> Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
> Signed-off-by: Barry Song <v-songbaohua@oppo.com>
> Cc: Yosry Ahmed <yosryahmed@google.com>
> Cc: Nhat Pham <nphamcs@gmail.com>
> Cc: Chengming Zhou <chengming.zhou@linux.dev>
> Cc: Chris Li <chrisl@kernel.org>
> Cc: David Hildenbrand <david@redhat.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

>   include/linux/zswap.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/zswap.h b/include/linux/zswap.h
> index bf83ae5e285d..6cecb4a4f68b 100644
> --- a/include/linux/zswap.h
> +++ b/include/linux/zswap.h
> @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
>   
>   static inline bool zswap_never_enabled(void)
>   {
> -	return false;
> +	return true;
>   }
>   
>   #endif

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Yosry Ahmed]

On Sat, Jun 29, 2024 at 4:22 PM Barry Song <21cnbao@gmail.com> wrote:
>
> From: Barry Song <v-songbaohua@oppo.com>
>
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
>
> Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
> Signed-off-by: Barry Song <v-songbaohua@oppo.com>

Acked-by: Yosry Ahmed <yosryahmed@google.com>

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Chris Li]

Acked-by: Chris Li <chrisl@kernel.org>

Chris

On Sat, Jun 29, 2024 at 4:22 PM Barry Song <21cnbao@gmail.com> wrote:
>
> From: Barry Song <v-songbaohua@oppo.com>
>
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
>
> Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
> Signed-off-by: Barry Song <v-songbaohua@oppo.com>
> Cc: Yosry Ahmed <yosryahmed@google.com>
> Cc: Nhat Pham <nphamcs@gmail.com>
> Cc: Chengming Zhou <chengming.zhou@linux.dev>
> Cc: Chris Li <chrisl@kernel.org>
> Cc: David Hildenbrand <david@redhat.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---
>  include/linux/zswap.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/linux/zswap.h b/include/linux/zswap.h
> index bf83ae5e285d..6cecb4a4f68b 100644
> --- a/include/linux/zswap.h
> +++ b/include/linux/zswap.h
> @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
>
>  static inline bool zswap_never_enabled(void)
>  {
> -       return false;
> +       return true;
>  }
>
>  #endif
> --
> 2.34.1
>

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Andrew Morton]

On Sun, 30 Jun 2024 11:22:31 +1200 Barry Song <21cnbao@gmail.com> wrote:

> From: Barry Song <v-songbaohua@oppo.com>
> 
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
> 
> ...
>
> --- a/include/linux/zswap.h
> +++ b/include/linux/zswap.h
> @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
>  
>  static inline bool zswap_never_enabled(void)
>  {
> -	return false;
> +	return true;
>  }

Well, that code was as wrong as it's possible to get.

But what effect does this have?  Seems "not much"?  Perhaps we'll
attempt a zswap_load() which later fails for other reasons?

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[David Hildenbrand]

On 30.06.24 01:22, Barry Song wrote:
> From: Barry Song <v-songbaohua@oppo.com>
> 
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
> 
> Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
> Signed-off-by: Barry Song <v-songbaohua@oppo.com>
> Cc: Yosry Ahmed <yosryahmed@google.com>
> Cc: Nhat Pham <nphamcs@gmail.com>
> Cc: Chengming Zhou <chengming.zhou@linux.dev>
> Cc: Chris Li <chrisl@kernel.org>
> Cc: David Hildenbrand <david@redhat.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---
>   include/linux/zswap.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/zswap.h b/include/linux/zswap.h
> index bf83ae5e285d..6cecb4a4f68b 100644
> --- a/include/linux/zswap.h
> +++ b/include/linux/zswap.h
> @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
>   
>   static inline bool zswap_never_enabled(void)
>   {
> -	return false;
> +	return true;
>   }
>   
>   #endif

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers,

David / dhildenb

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Barry Song]

On Tue, Jul 2, 2024 at 6:52 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Sun, 30 Jun 2024 11:22:31 +1200 Barry Song <21cnbao@gmail.com> wrote:
>
> > From: Barry Song <v-songbaohua@oppo.com>
> >
> > If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> > zswap_never_enabled() should return true.
> >
> > ...
> >
> > --- a/include/linux/zswap.h
> > +++ b/include/linux/zswap.h
> > @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
> >
> >  static inline bool zswap_never_enabled(void)
> >  {
> > -     return false;
> > +     return true;
> >  }
>
> Well, that code was as wrong as it's possible to get.
>
> But what effect does this have?  Seems "not much"?  Perhaps we'll
> attempt a zswap_load() which later fails for other reasons?

Yes, but the API was created to inform the mm core that zswap has never been
enabled, allowing the mm core to perform mTHP swap-in. This is a transitional
solution until zswap supports mTHP. If zswap has been enabled, performing
mTHP swap-in will result in corrupted data. You may find the answer in the
mTHP swap-in series:

https://lore.kernel.org/linux-mm/CAJD7tkZ4FQr6HZpduOdvmqgg_-whuZYE-Bz5O2t6yzw6Yg+v1A@mail.gmail.com/

Thanks
Barry

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Yosry Ahmed]

On Mon, Jul 1, 2024 at 11:52 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Sun, 30 Jun 2024 11:22:31 +1200 Barry Song <21cnbao@gmail.com> wrote:
>
> > From: Barry Song <v-songbaohua@oppo.com>
> >
> > If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> > zswap_never_enabled() should return true.
> >
> > ...
> >
> > --- a/include/linux/zswap.h
> > +++ b/include/linux/zswap.h
> > @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
> >
> >  static inline bool zswap_never_enabled(void)
> >  {
> > -     return false;
> > +     return true;
> >  }
>
> Well, that code was as wrong as it's possible to get.
>
> But what effect does this have?  Seems "not much"?  Perhaps we'll
> attempt a zswap_load() which later fails for other reasons?

Actually zswap_load() is a noop with !CONFIG_ZSWAP, so it doesn't have
an effect there. The only effect is that with Barry's latest large
folio swapin patches for zram, we will always fallback to order-0
swapin, even mistakenly when !CONFIG_ZSWAP.

Basically the bug just makes Barry's in progress patches not work at all.

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Andrew Morton]

On Tue, 2 Jul 2024 04:58:05 -0700 Yosry Ahmed <yosryahmed@google.com> wrote:

> > > --- a/include/linux/zswap.h
> > > +++ b/include/linux/zswap.h
> > > @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
> > >
> > >  static inline bool zswap_never_enabled(void)
> > >  {
> > > -     return false;
> > > +     return true;
> > >  }
> >
> > Well, that code was as wrong as it's possible to get.
> >
> > But what effect does this have?  Seems "not much"?  Perhaps we'll
> > attempt a zswap_load() which later fails for other reasons?
> 
> Actually zswap_load() is a noop with !CONFIG_ZSWAP, so it doesn't have
> an effect there. The only effect is that with Barry's latest large
> folio swapin patches for zram, we will always fallback to order-0
> swapin, even mistakenly when !CONFIG_ZSWAP.
> 
> Basically the bug just makes Barry's in progress patches not work at all.

OK, thanks, I added this to the changelog:

The only effect of this issue is that with Barry's latest large folio
swapin patches for zram ("mm: support mTHP swap-in for zRAM-like
swapfile"), we will always fallback to order-0 swapin, even mistakenly
when !CONFIG_ZSWAP.

Basically this bug makes Barry's in progress patches not work at all.

Re: [PATCH] mm: zswap: fix zswap_never_enabled() for CONFIG_ZSWAP==N

[Nhat Pham]

On Sat, Jun 29, 2024 at 4:22 PM Barry Song <21cnbao@gmail.com> wrote:
>
> From: Barry Song <v-songbaohua@oppo.com>
>
> If CONFIG_ZSWAP is set to N, it means zswap cannot be enabled.
> zswap_never_enabled() should return true.
>
> Fixes: 0300e17d67c3 ("mm: zswap: add zswap_never_enabled()")
> Signed-off-by: Barry Song <v-songbaohua@oppo.com>
> Cc: Yosry Ahmed <yosryahmed@google.com>
> Cc: Nhat Pham <nphamcs@gmail.com>
> Cc: Chengming Zhou <chengming.zhou@linux.dev>
> Cc: Chris Li <chrisl@kernel.org>
> Cc: David Hildenbrand <david@redhat.com>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---
>  include/linux/zswap.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/linux/zswap.h b/include/linux/zswap.h
> index bf83ae5e285d..6cecb4a4f68b 100644
> --- a/include/linux/zswap.h
> +++ b/include/linux/zswap.h
> @@ -68,7 +68,7 @@ static inline bool zswap_is_enabled(void)
>
>  static inline bool zswap_never_enabled(void)
>  {
> -       return false;
> +       return true;
>  }
>
>  #endif
> --
> 2.34.1
>

That's some pretty yikesy bug :) Thankfully this seems unused thus far
(which is probably why it is not caught until now).
Thanks for fixing this, Barry!

Reviewed-by: Nhat Pham <nphamcs@gmail.com>