Received-SPF: pass (google.com: domain of linux-kernel+bounces-12391-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Wed, 27 Dec 2023 18:17:35 -0800
From: "H. Peter Anvin" <hpa@zytor.com>
To: Al Viro <viro@zeniv.linux.org.uk>
CC: Tanzir Hasan <tanzirh@google.com>, Andy Lutomirski <luto@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>,
        Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org,
        linux-kernel@vger.kernel.org, Nick Desaulniers <nnn@google.com>
Subject: Re: [PATCH] x86/syscalls: shrink entry/syscall_32.i via IWYU
User-Agent: K-9 Mail for Android
In-Reply-To: <20231228002614.GI1674809@ZenIV>
References: <20231227-syscall32-v1-1-9621140d33bd@google.com> <20231227233444.GH1674809@ZenIV> <8B198413-0C9B-40A2-8478-F22206AAE6FA@zytor.com> <20231228002614.GI1674809@ZenIV>
Message-ID: <7068F1FF-1266-447D-ACC8-2F1AD774F230@zytor.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

On December 27, 2023 4:26:14 PM PST, Al Viro <viro@zeniv=2Elinux=2Eorg=2Euk=
> wrote:
>On Wed, Dec 27, 2023 at 03:50:33PM -0800, H=2E Peter Anvin wrote:
>> > /*
>> >  * Only the low 32 bits of orig_ax are meaningful, so we return int=
=2E
>> >  * This importantly ignores the high bits on 64-bit, so comparisons
>>=20
>> __visible is for LTO, no?
>
>If we need it in cases when array defined in entry/syscall_32=2Ec and
>used in entry/common=2Ec, I would respectfully suggest that whatever
>we need it for is misguided garbage=2E  I don't think that LTO does
>need it, though=2E  How is arch/x86/entry/{syscall_32,common}=2Ec
>different from e=2Eg=2E fs/{namespace,d_path}=2Ec, where we have
>fs/namespace=2Ec:100:__cacheline_aligned_in_smp DEFINE_SEQLOCK(mount_lock=
);
>and
>fs/d_path=2Ec:166:        read_seqbegin_or_lock(&mount_lock, &m_seq);
>respectively?

You're correct of course; __visible for LTO is for functions called from a=
ssembly *at the point of definition*, not declaration and certainly not as =
a type=2E