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;
MIME-Version: 1.0
References: <20210612110014.GA1211@raspberrypi> <87eed7tk24.wl-maz@kernel.org>
In-Reply-To: <87eed7tk24.wl-maz@kernel.org>
From:   Austin Kim <austindh.kim@gmail.com>
Date:   Sat, 12 Jun 2021 22:32:54 +0900
Message-ID: <CADLLry7bYPYfqwQwWLtsXaAMG_deJ=TVVowAb5wtk_=AhTT64Q@mail.gmail.com>
Subject: Re: [PATCH] KVM: arm64: vgic: declear local variable 'flags' before
 for loop starts
To:     Marc Zyngier <maz@kernel.org>
Cc:     james.morse@arm.com, alexandru.elisei@arm.com,
        suzuki.poulose@arm.com, catalin.marinas@arm.com, will@kernel.org,
        linux-arm-kernel@lists.infradead.org, kvmarm@lists.cs.columbia.edu,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        =?UTF-8?B?6rmA64+Z7ZiE?= <austin.kim@lge.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

2021=EB=85=84 6=EC=9B=94 12=EC=9D=BC (=ED=86=A0) =EC=98=A4=ED=9B=84 8:10, M=
arc Zyngier <maz@kernel.org>=EB=8B=98=EC=9D=B4 =EC=9E=91=EC=84=B1:
>
> On Sat, 12 Jun 2021 12:00:14 +0100,
> Austin Kim <austindh.kim@gmail.com> wrote:
> >
> > From: Austin Kim <austin.kim@lge.com>
> >
> > Normally local variable 'flags' is defined out of for loop,
> > when 'flags' is used as the second parameter in a call to
> > spinlock_irq[save/restore] function.
> >
> > So it had better declear local variable 'flags' ahead of for loop.
>
> Why better? Reducing the scope of a variable is in general good
> practice. Do you see any material advantage in moving this variable
> out of the loop? Does the compiler generate better code?

First all of, thanks for feedback.

I checked how the compiler generate assembly code(before/after) using
objdump utility.
And then found out that compiler generates the same assembly code.

<compiler version: gcc-linaro-7.5.0-2019.12-x86_64_aarch64>

ffff80001005f5c8 <vgic_mmio_read_pending>:
ffff80001005f5c8:   d503233f    paciasp
...
ffff80001005f63c:   97ffe9af    bl  ffff800010059cf8 <vgic_get_irq>
ffff80001005f640:   aa0003f3    mov x19, x0
ffff80001005f644:   943886c3    bl  ffff800010e81150 <_raw_spin_lock_irqsav=
e>

Let me keep in mind how compiler generate assembly code with new patch,
which leads to material advantage.

BR,
Austin Kim

>
> Thanks,
>
>         M.
>
> --
> Without deviation from the norm, progress is not possible.