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;
Date:   Wed, 28 Jun 2023 18:24:40 +0100
From:   Conor Dooley <conor@kernel.org>
To:     Evan Green <evan@rivosinc.com>
Cc:     Conor Dooley <conor.dooley@microchip.com>,
        Samuel Ortiz <sameo@rivosinc.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Palmer Dabbelt <palmer@dabbelt.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        linux-riscv@lists.infradead.org,
        "Hongren (Zenithal) Zheng" <i@zenithal.me>, linux@rivosinc.com,
        Andrew Jones <ajones@ventanamicro.com>,
        Heiko Stuebner <heiko.stuebner@vrull.eu>,
        Anup Patel <apatel@ventanamicro.com>,
        linux-kernel@vger.kernel.org, Guo Ren <guoren@kernel.org>,
        Atish Patra <atishp@rivosinc.com>,
        =?iso-8859-1?Q?Bj=F6rn_T=F6pel?= <bjorn@rivosinc.com>,
        Jiatai He <jiatai2021@iscas.ac.cn>
Subject: Re: [PATCH 1/3] RISC-V: add Bitmanip/Scalar Crypto parsing from DT
Message-ID: <20230628-dragonish-lullaby-b44d2df09d66@spud>
References: <20230627143747.1599218-1-sameo@rivosinc.com>
 <20230627143747.1599218-2-sameo@rivosinc.com>
 <CALs-HssMkVikspnEi-Ek2t=ABvFvgptAhsBjk1+aLuVjiP7P7w@mail.gmail.com>
 <20230627-debating-twelve-da2c1ed60948@spud>
 <ZJwE5wRVkoND3Z6P@vermeer>
 <20230628-unfeeling-tavern-edd4f58396fa@wendy>
 <CALs-HstZe+bw_fY--4FQXfFoA67tvMSaEjKhZ2pRstNT07xBxA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
        protocol="application/pgp-signature"; boundary="Nk2liGfu8KoG29TY"
Content-Disposition: inline
In-Reply-To: <CALs-HstZe+bw_fY--4FQXfFoA67tvMSaEjKhZ2pRstNT07xBxA@mail.gmail.com>
Precedence: bulk


--Nk2liGfu8KoG29TY
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Jun 28, 2023 at 10:18:34AM -0700, Evan Green wrote:
> On Wed, Jun 28, 2023 at 4:10=E2=80=AFAM Conor Dooley <conor.dooley@microc=
hip.com> wrote:
> >
> > On Wed, Jun 28, 2023 at 12:01:11PM +0200, Samuel Ortiz wrote:
> > > On Tue, Jun 27, 2023 at 07:48:15PM +0100, Conor Dooley wrote:
> > > > On Tue, Jun 27, 2023 at 11:14:30AM -0700, Evan Green wrote:
> > > > > On Tue, Jun 27, 2023 at 7:38=E2=80=AFAM Samuel Ortiz <sameo@rivos=
inc.com> wrote:
> >
> > > > > It would be nice to consolidate the ones together that search for=
 a
> > > > > single string and set multiple bits, though I don't have any super
> > > > > elegant ideas for how off the top of my head.
> > > >
> > > > I've got a refactor of this code in progress, dropping all of these
> > > > copy-paste in place of a loop. It certainly looks more elegant than
> > > > this, but it will fall over a bit for these "one string matches many
> > > > extensions" cases. See here:
> > > > https://patchwork.kernel.org/project/linux-riscv/patch/20230626-thi=
eving-jockstrap-d35d20b535c5@wendy/
> > > > My immediate thought is to add another element to riscv_isa_ext_dat=
a,
> > > > that contains "parent" extensions to check for. Should be fairly do=
able,
> > > > I'll whip something up on top of that...
> > >
> > > Nice, and thanks for the review.
> >
> > > Should I wait for your refactor to be merged before pushing this one?
> >
> > I don't know. I think that you should continue on with your series here,
> > and whichever goes in second gets rebased on top of the other.
> > I don't think it makes material difference to review of this patchset as
> > to whether you rebase on top of what I'm working on, so I wouldn't
> > bother until it gets merged.
> >
> > Rather hacky, had less time than expected this morning:
> > https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux.git/commit/=
?h=3Driscv-extensions-strings-supersets
> > Clearly there's issues with looping to RISCV_ISA_MAX_SUPERSETS & I just
> > repurposed Zicsr for the sake of testing something in the time I had.
> >
> > Evan, at a high level, does that look more elegant to you, or have I ma=
de
> > things worse?
> >
>=20
> I see what you're going for at least. It's unfortunate that when
> someone bumps up RISCV_ISA_MAX_SUPERSETS it squares the whole array.
> Another way to go might be to define the elements in a separate array,
> like:
>=20
> unsigned int riscv_zks_exts[] =3D {
>        RISCV_ISA_EXT_ZBKB,
>        RISCV_ISA_EXT_ZBKC,
>        ....
> };
>=20
> then the macro entry looks like:
>=20
> SET_ISA_EXT_MAP_MULTI("zks", riscv_zks_exts),
>=20
> where the SET_ISA_EXT_MAP_MULTI() could use ARRAY_SIZE() to stash both
> the pointer to the array and the number of elements.

Yup, I like the sound of that. I like the variadic stuff as it'd not
require defining a bunch of sub-arrays of supersets. I guess if it grows
too badly, we can just dump it off into another file or w/e.

--Nk2liGfu8KoG29TY
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----

iHUEABYIAB0WIQRh246EGq/8RLhDjO14tDGHoIJi0gUCZJxs2AAKCRB4tDGHoIJi
0hyeAPoCJou4iW+SwgKBQCp5o/rvu8bFFnG2SdAC26nA3X5ifAD8Cev8fwX4Y1Qg
5T7g/LvQlXnVSz10udi5To5hLiisfQM=
=dznn
-----END PGP SIGNATURE-----

--Nk2liGfu8KoG29TY--