Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230803051401.710236-2-leobras@redhat.com> <CAJF2gTShfMzVZw5TVqBvLNAzEBMzhWxNGC9JS70euPcKKN85zQ@mail.gmail.com>
 <CAJ6HWG5mzxN=txnbdi-=c+=wLsDfSe7Me+1DQE79ZX0NN6U1_g@mail.gmail.com> <CAJF2gTTag3hR7ZgM8o+yk+0Ox3A+xm4jAVTikAA0HqXQFKayRA@mail.gmail.com>
In-Reply-To: <CAJF2gTTag3hR7ZgM8o+yk+0Ox3A+xm4jAVTikAA0HqXQFKayRA@mail.gmail.com>
From:   Leonardo Bras Soares Passos <leobras@redhat.com>
Date:   Fri, 4 Aug 2023 05:05:11 -0300
Message-ID: <CAJ6HWG6ofyiLRq-HZFmocxPFzQhM+HwBQVOqwp-XSfctFHLNTQ@mail.gmail.com>
Subject: Re: [RFC PATCH v2 0/3] Deduplicate RISCV cmpxchg.h and atomic.c macros
To:     Guo Ren <guoren@kernel.org>
Cc:     Will Deacon <will@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Boqun Feng <boqun.feng@gmail.com>,
        Mark Rutland <mark.rutland@arm.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Andrzej Hajda <andrzej.hajda@intel.com>,
        Arnd Bergmann <arnd@arndb.de>, Ingo Molnar <mingo@kernel.org>,
        Palmer Dabbelt <palmer@rivosinc.com>,
        linux-kernel@vger.kernel.org, linux-riscv@lists.infradead.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Aug 3, 2023 at 11:29=E2=80=AFPM Guo Ren <guoren@kernel.org> wrote:
>
> On Fri, Aug 4, 2023 at 10:20=E2=80=AFAM Leonardo Bras Soares Passos
> <leobras@redhat.com> wrote:
> >
> > On Thu, Aug 3, 2023 at 9:53=E2=80=AFPM Guo Ren <guoren@kernel.org> wrot=
e:
> > >
> > > On Thu, Aug 3, 2023 at 1:14=E2=80=AFPM Leonardo Bras <leobras@redhat.=
com> wrote:
> > > >
> > > > I unified previous patchsets into a single one, since the work is r=
elated.
> > > >
> > > > While studying riscv's cmpxchg.h file, I got really interested in
> > > > understanding how RISCV asm implemented the different versions of
> > > > {cmp,}xchg.
> > > >
> > > > When I understood the pattern, it made sense for me to remove the
> > > > duplications and create macros to make it easier to understand what=
 exactly
> > > > changes between the versions: Instruction sufixes & barriers.
> > > >
> > > > Also, did the same kind of work on atomic.c.
> > > >
> > > > Note to Guo Ren:
> > > > I did some further improvement after your previous reviews, so I en=
ded
> > > > up afraid including your Reviewed-by before cheching if the changes=
 are
> > > > ok for you. Please check it out again, I just removed some helper m=
acros
> > > > that were not being used elsewhere in the kernel.
> > > I found this optimization has conflicts with the below patches:
> > > https://lore.kernel.org/linux-riscv/20230802164701.192791-15-guoren@k=
ernel.org/
> > > https://lore.kernel.org/linux-riscv/20230802164701.192791-5-guoren@ke=
rnel.org/
> > >
> > > If yours merged, how do we support the inline cmpxchg/xchg_small
> > > function?
> >
> > Oh, I actually introduced my series so I could introduce new xchg and
> > cmpxchg for size 1 and 2. Is that what your patches are about, right?
> >
> > I was working on that yesterday, and decided to send the patchset
> > without them because I was still not sure enough.
> >
> > About implementation strategy, I was introducing a new macros for xchg
> > & cmpxchg with asm which would work for both for size 1 & size 2, and
> > use the switch-case to create the mask and and_value.
> >
> > You think that works enough?
> Good, go ahead.
>
> >
> > > It's very struggling to use macros to implement complex
> > > functions.
> >
> > I agree, but with this we can achieve more generic code, which makes
> > more clear what is the pattern for given function.
> >
> > > Could you consider a more relaxed optimization in which we could
> > > insert inline cmpxchg/xchg_small?
> >
> > What about this: I finish the patches I have been working with
> > (cmpxchg & xchg for sizes 1 and 2), and if they are fine we expand
> > this patchset with them.  If not, I try relaxing them a little so we
> > can merge with your set.
> >
> > Does that work for you?
> Great, you could provide cmpxchg & xchg for sizes 1 and 2, then my
> patch series would base on yours. After tested, I would give you
> Tested-by.

Awesome! Thanks!

I will send it shortly, just compile-testing the kernel.

> >
> > Best regards,
> > Leo
> >
> >
> > >
> > > >
> > > > Thanks!
> > > > Leo
> > > >
> > > >
> > > > Changes since squashed cmpxchg:
> > > > - Unified with atomic.c patchset
> > > > - Rebased on top of torvalds/master (thanks Andrea Parri!)
> > > > - Removed helper macros that were not being used elsewhere in the k=
ernel.
> > > >
> > > > Changes since (cmpxchg) RFCv3:
> > > > - Squashed the 6 original patches in 2: one for cmpxchg and one for=
 xchg
> > > > https://lore.kernel.org/all/20230404163741.2762165-1-leobras@redhat=
.com/
> > > >
> > > > Changes since (cmpxchg) RFCv2:
> > > > - Fixed  macros that depend on having a local variable with a magic=
 name
> > > > - Previous cast to (long) is now only applied on 4-bytes cmpxchg
> > > > https://lore.kernel.org/all/20230321074249.2221674-1-leobras@redhat=
.com/
> > > >
> > > > Changes since (cmpxchg) RFCv1:
> > > > - Fixed patch 4/6 suffix from 'w.aqrl' to '.w.aqrl', to avoid build=
 error
> > > > https://lore.kernel.org/all/20230318080059.1109286-1-leobras@redhat=
.com/
> > > >
> > > >
> > > > Leonardo Bras (3):
> > > >   riscv/cmpxchg: Deduplicate xchg() asm functions
> > > >   riscv/cmpxchg: Deduplicate cmpxchg() asm and macros
> > > >   riscv/atomic.h : Deduplicate arch_atomic.*
> > > >
> > > >  arch/riscv/include/asm/atomic.h  | 164 ++++++++--------
> > > >  arch/riscv/include/asm/cmpxchg.h | 318 +++++----------------------=
----
> > > >  2 files changed, 123 insertions(+), 359 deletions(-)
> > > >
> > > > --
> > > > 2.41.0
> > > >
> > >
> > >
> > > --
> > > Best Regards
> > >  Guo Ren
> > >
> >
>
>
> --
> Best Regards
>  Guo Ren
>