[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <034f0b5e-3c61-4392-8360-ebd57749d53e@app.fastmail.com>
Date: Wed, 26 Jul 2023 19:02:55 +0200
From: "Arnd Bergmann" <arnd@...db.de>
To: "Nick Desaulniers" <ndesaulniers@...gle.com>,
"Arnd Bergmann" <arnd@...nel.org>
Cc: "Paul E. McKenney" <paulmck@...nel.org>,
"Andrew Morton" <akpm@...ux-foundation.org>,
"Kees Cook" <keescook@...omium.org>,
"Peter Zijlstra" <peterz@...radead.org>,
"Sami Tolvanen" <samitolvanen@...gle.com>,
"Alexey Dobriyan" <adobriyan@...il.com>,
"Masahiro Yamada" <masahiroy@...nel.org>,
"Josh Poimboeuf" <jpoimboe@...nel.org>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] init: keep boot_command_line after init
On Wed, Jul 26, 2023, at 18:25, Nick Desaulniers wrote:
> On Wed, Jul 26, 2023 at 7:33 AM Arnd Bergmann <arnd@...nel.org> wrote:
>>
>> From: Arnd Bergmann <arnd@...db.de>
>>
>> The boot command line is not available after the init section gets discarded,
>> so adding a permanent reference to it causes a link time warning:
>>
>> WARNING: modpost: vmlinux: section mismatch in reference: cmdline_load_proc_show+0x2 (section: .text) -> boot_command_line (section: .init.data)
>
> cmdline_load_proc_show is probably inlined, but should it also be
> marked __init? It's lone call site seems to be __init AFAICT.
>
No, that's not what it does: cmdline_load_proc_show() is called
when someone reads /proc/cmdline_load from userspace. It's only
the function that creates this procfs file that is __init, but the
call happens later.
Arnd
Powered by blists - more mailing lists