MIME-Version: 1.0
References: <CAHk-=wiPd8R8-zSqTOtJ9KYeZLBByHug7ny3rgP-ZqzpP_KELg@mail.gmail.com>
 <20230320180501.GA598084@dev-arch.thelio-3990X> <CAHk-=wgSqpdkeJBb92M37JNTdRQJRnRUApraHKE8uGHTqQuu2Q@mail.gmail.com>
In-Reply-To: <CAHk-=wgSqpdkeJBb92M37JNTdRQJRnRUApraHKE8uGHTqQuu2Q@mail.gmail.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Mon, 20 Mar 2023 11:49:55 -0700
Message-ID: <CAHk-=whXi1qgeDLHbaCyMpZ4XXYqS=fT2daYAmWntEK2rhG25g@mail.gmail.com>
Subject: Re: Linux 6.3-rc3
To:     Nathan Chancellor <nathan@kernel.org>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        David Airlie <airlied@gmail.com>,
        Daniel Vetter <daniel@ffwll.ch>,
        dri-devel@lists.freedesktop.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Mar 20, 2023 at 11:26=E2=80=AFAM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> Hmm. I do my arm64 allmodconfig builds with gcc, and I'm surprised
> that gcc doesn't warn about this.

Side note: I'm also wondering why that TEGRA_HOST1X config has that
ARM dependency in

        depends on ARCH_TEGRA || (ARM && COMPILE_TEST)

because it seems to build just fine at least on x86-64 if I change it to be=
 just

        depends on ARCH_TEGRA || COMPILE_TEST

ie there seems to be nothing ARM-specific in there.

Limiting it to just the tegra platform by default makes 100% sense,
but that "only do compile-testing on ARM" seems a bit bogus.

That limit goes back to forever (commit 6f44c2b5280f: "gpu: host1x:
Increase compile test coverage" back in Nov 2013), so maybe things
didn't use to work as well back in the dark ages?

None of this explains why gcc didn't catch it, but at least allowing
the build on x86-64 would likely have made it easier for people to see
clang catching this.

                   Linus