Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Sender: Arvind Sankar <niveditas98@gmail.com>
From:   Arvind Sankar <nivedita@alum.mit.edu>
Date:   Tue, 10 Nov 2020 17:54:21 -0500
To:     Nick Desaulniers <ndesaulniers@google.com>
Cc:     Arvind Sankar <nivedita@alum.mit.edu>,
        Adrian Ratiu <adrian.ratiu@collabora.com>,
        Nathan Chancellor <natechancellor@gmail.com>,
        Arnd Bergmann <arnd@arndb.de>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        clang-built-linux <clang-built-linux@googlegroups.com>,
        Russell King <linux@armlinux.org.uk>,
        LKML <linux-kernel@vger.kernel.org>,
        Collabora Kernel ML <kernel@collabora.com>,
        Ard Biesheuvel <ardb@kernel.org>
Subject: Re: [PATCH 2/2] arm: lib: xor-neon: disable clang vectorization
Message-ID: <20201110225421.GA29900@rani.riverdale.lan>
References: <20201106051436.2384842-3-adrian.ratiu@collabora.com>
 <20201106101419.GB3811063@ubuntu-m3-large-x86>
 <87wnyyvh56.fsf@collabora.com>
 <CAKwvOdkodob0M0r_AK_4nG3atLGMyNENMd6qVAHSPa92Zh7UZA@mail.gmail.com>
 <871rh2i9xg.fsf@iwork.i-did-not-set--mail-host-address--so-tickle-me>
 <CAKwvOdkm3u83TQDBB-fC0TwKZCFXGh5sAfahKXxA+mnzgDid_w@mail.gmail.com>
 <20201110221511.GA1373528@rani.riverdale.lan>
 <CAKwvOd=NQ==umC+N_Sgji5HCCFTRARh4jWiB3DaBfV6jDd5cRg@mail.gmail.com>
 <CAKwvOdmsCY8Hb6t8t9jW5DKCZKbja27bGufO=tw3+-UuTTFBOQ@mail.gmail.com>
 <CAKwvOd=6nPUvUY6XJULmkZywHJG2kfu46h7=Zm18j9O30ovdpw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <CAKwvOd=6nPUvUY6XJULmkZywHJG2kfu46h7=Zm18j9O30ovdpw@mail.gmail.com>
Precedence: bulk

On Tue, Nov 10, 2020 at 02:39:59PM -0800, Nick Desaulniers wrote:
> On Tue, Nov 10, 2020 at 2:39 PM Nick Desaulniers
> <ndesaulniers@google.com> wrote:
> >
> > On Tue, Nov 10, 2020 at 2:36 PM Nick Desaulniers
> > <ndesaulniers@google.com> wrote:
> > >
> > > On Tue, Nov 10, 2020 at 2:15 PM Arvind Sankar <nivedita@alum.mit.edu> wrote:
> > > >
> > > > On Tue, Nov 10, 2020 at 01:41:17PM -0800, Nick Desaulniers wrote:
> > > > > On Mon, Nov 9, 2020 at 11:51 AM Adrian Ratiu <adrian.ratiu@collabora.com> wrote:
> > > > > >
> > > > > > On Fri, 06 Nov 2020, Nick Desaulniers <ndesaulniers@google.com>
> > > > > > wrote:
> > > > > > > +#pragma clang loop vectorize(enable)
> > > > > > >         do {
> > > > > > >                 p1[0] ^= p2[0] ^ p3[0] ^ p4[0] ^ p5[0]; p1[1] ^=
> > > > > > >                 p2[1] ^ p3[1] ^ p4[1] ^ p5[1];
> > > > > > > ``` seems to generate the vectorized code.
> > > > > > >
> > > > > > > Why don't we find a way to make those pragma's more toolchain
> > > > > > > portable, rather than open coding them like I have above rather
> > > > > > > than this series?
> > > > > >
> > > > > > Hi again Nick,
> > > > > >
> > > > > > How did you verify the above pragmas generate correct vectorized
> > > > > > code?  Have you tested this specific use case?
> > > > >
> > > > > I read the disassembly before and after my suggested use of pragmas;
> > > > > look for vld/vstr.  You can also add -Rpass-missed=loop-vectorize to
> > > > > CFLAGS_xor-neon.o in arch/arm/lib/Makefile and rebuild
> > > > > arch/arm/lib/xor-neon.o with CONFIG_BTRFS enabled.
> > > > >
> > > >
> > > > https://godbolt.org/z/1oo9M6
> > > >
> > > > With the __restrict__ keywords added, clang seems to vectorize the loop,
> > > > but still reports that vectorization wasn't beneficial -- any idea
> > > > what's going on?
> >
> > Anyways, it's not safe to make that change in the kernel unless you
> > can guarantee that callers of these routines do not alias or overlap.
> 
> s/callers/parameters passed by callers/
> 

Yep, but that seems likely, it doesn't seem like the function would do
anything useful if the destination overlapped one of the sources. The
kernel just doesn't seem to make use of the restrict keyword.