This patchset first removes redundant checks on the acomp_ctx and its
"req" member in zswap_cpu_comp_dead().

Next, it persists the zswap pool's per-CPU acomp_ctx resources to
last until the pool is destroyed. It then simplifies the per-CPU
acomp_ctx mutex locking in zswap_compress()/zswap_decompress().

Code comments added after allocation and before checking to deallocate
the per-CPU acomp_ctx's members, based on expected crypto API return
values and zswap changes this patchset makes.

Patch 2 is an independent submission of patch 23 from [1], to
facilitate merging.

[1]: https://patchwork.kernel.org/project/linux-mm/list/?series=1046677

Changes since v2:
=================
1) Rebased to mm-unstable as of 3-31-2026 (commit 24e341657c28).
2) Collected Acks.

Changes since v1:
=================
1) Made the changes to eliminate redundant checks on
   acomp_ctx/acomp_ctx->req in zswap_cpu_comp_dead(), per Yosry.
2) Renamed acomp_ctx_dealloc() to acomp_ctx_free(), per Yosry.
3) Incorporated suggestions from Yosry and Sashiko to reset the
   acomp_ctx's members to NULL after freeing them, to prevent UAF and
   double free issues.
4) Replaced v1's patch 2 with v2's patch 1.


Kanchana P. Sridhar (2):
  mm: zswap: Remove redundant checks in zswap_cpu_comp_dead().
  mm: zswap: Tie per-CPU acomp_ctx lifetime to the pool.

 mm/zswap.c | 182 +++++++++++++++++++++++++----------------------------
 1 file changed, 85 insertions(+), 97 deletions(-)

-- 
2.39.5

On Tue, Mar 31, 2026 at 11:34 AM Kanchana P. Sridhar
<kanchanapsridhar2026@gmail.com> wrote:
>
> This patchset first removes redundant checks on the acomp_ctx and its
> "req" member in zswap_cpu_comp_dead().
>
> Next, it persists the zswap pool's per-CPU acomp_ctx resources to
> last until the pool is destroyed. It then simplifies the per-CPU
> acomp_ctx mutex locking in zswap_compress()/zswap_decompress().
>
> Code comments added after allocation and before checking to deallocate
> the per-CPU acomp_ctx's members, based on expected crypto API return
> values and zswap changes this patchset makes.
>
> Patch 2 is an independent submission of patch 23 from [1], to
> facilitate merging.
>
> [1]: https://patchwork.kernel.org/project/linux-mm/list/?series=1046677
>
> Changes since v2:
> =================
> 1) Rebased to mm-unstable as of 3-31-2026 (commit 24e341657c28).
> 2) Collected Acks.
>
> Changes since v1:
> =================
> 1) Made the changes to eliminate redundant checks on
>    acomp_ctx/acomp_ctx->req in zswap_cpu_comp_dead(), per Yosry.
> 2) Renamed acomp_ctx_dealloc() to acomp_ctx_free(), per Yosry.
> 3) Incorporated suggestions from Yosry and Sashiko to reset the
>    acomp_ctx's members to NULL after freeing them, to prevent UAF and
>    double free issues.
> 4) Replaced v1's patch 2 with v2's patch 1.
>
>
> Kanchana P. Sridhar (2):
>   mm: zswap: Remove redundant checks in zswap_cpu_comp_dead().
>   mm: zswap: Tie per-CPU acomp_ctx lifetime to the pool.

Sashiko was able to apply the patches and doesn't seem to find any
regressions: https://sashiko.dev/#/patchset/20260331183351.29844-1-kanchanapsridhar2026%40gmail.com.

On Tue, Mar 31, 2026 at 12:23 PM Yosry Ahmed <yosry@kernel.org> wrote:
>
> On Tue, Mar 31, 2026 at 11:34 AM Kanchana P. Sridhar
> <kanchanapsridhar2026@gmail.com> wrote:
> >
> > This patchset first removes redundant checks on the acomp_ctx and its
> > "req" member in zswap_cpu_comp_dead().
> >
> > Next, it persists the zswap pool's per-CPU acomp_ctx resources to
> > last until the pool is destroyed. It then simplifies the per-CPU
> > acomp_ctx mutex locking in zswap_compress()/zswap_decompress().
> >
> > Code comments added after allocation and before checking to deallocate
> > the per-CPU acomp_ctx's members, based on expected crypto API return
> > values and zswap changes this patchset makes.
> >
> > Patch 2 is an independent submission of patch 23 from [1], to
> > facilitate merging.
> >
> > [1]: https://patchwork.kernel.org/project/linux-mm/list/?series=1046677
> >
> > Changes since v2:
> > =================
> > 1) Rebased to mm-unstable as of 3-31-2026 (commit 24e341657c28).
> > 2) Collected Acks.
> >
> > Changes since v1:
> > =================
> > 1) Made the changes to eliminate redundant checks on
> >    acomp_ctx/acomp_ctx->req in zswap_cpu_comp_dead(), per Yosry.
> > 2) Renamed acomp_ctx_dealloc() to acomp_ctx_free(), per Yosry.
> > 3) Incorporated suggestions from Yosry and Sashiko to reset the
> >    acomp_ctx's members to NULL after freeing them, to prevent UAF and
> >    double free issues.
> > 4) Replaced v1's patch 2 with v2's patch 1.
> >
> >
> > Kanchana P. Sridhar (2):
> >   mm: zswap: Remove redundant checks in zswap_cpu_comp_dead().
> >   mm: zswap: Tie per-CPU acomp_ctx lifetime to the pool.
>
> Sashiko was able to apply the patches and doesn't seem to find any
> regressions: https://sashiko.dev/#/patchset/20260331183351.29844-1-kanchanapsridhar2026%40gmail.com.

That's good to know, thanks for confirming Yosry!

On Tue, 31 Mar 2026 13:58:37 -0700 "Kanchana P. Sridhar" <kanchanapsridhar2026@gmail.com> wrote:

> > > Kanchana P. Sridhar (2):
> > >   mm: zswap: Remove redundant checks in zswap_cpu_comp_dead().
> > >   mm: zswap: Tie per-CPU acomp_ctx lifetime to the pool.
> >
> > Sashiko was able to apply the patches and doesn't seem to find any
> > regressions: https://sashiko.dev/#/patchset/20260331183351.29844-1-kanchanapsridhar2026%40gmail.com.
> 
> That's good to know, thanks for confirming Yosry!

Thanks, both.

I'm surprised that a patchset of this signficance got through AI review
without comment, actually.  

I've confirmed that the code changes in v3 are identical to those in
v2, which I removed a couple of days ago for Sashiko's benefit.  v2 was
added to mm.git March 16. So all good, I've re-added this series.

On Tue, Mar 31, 2026 at 3:19 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Tue, 31 Mar 2026 13:58:37 -0700 "Kanchana P. Sridhar" <kanchanapsridhar2026@gmail.com> wrote:
>
> > > > Kanchana P. Sridhar (2):
> > > >   mm: zswap: Remove redundant checks in zswap_cpu_comp_dead().
> > > >   mm: zswap: Tie per-CPU acomp_ctx lifetime to the pool.
> > >
> > > Sashiko was able to apply the patches and doesn't seem to find any
> > > regressions: https://sashiko.dev/#/patchset/20260331183351.29844-1-kanchanapsridhar2026%40gmail.com.
> >
> > That's good to know, thanks for confirming Yosry!
>
> Thanks, both.
>
> I'm surprised that a patchset of this signficance got through AI review
> without comment, actually.

Thanks Andrew! I must thank Yosry because he pointed out end cases
that helped make the solution more fail-safe.
I am also very impressed by Sashiko's thorough analysis.

>
> I've confirmed that the code changes in v3 are identical to those in
> v2, which I removed a couple of days ago for Sashiko's benefit.  v2 was
> added to mm.git March 16. So all good, I've re-added this series.

Sounds great, thanks again!

Best regards,
Kanchana

>
>