Message-ID: <1500525696.2055.5.camel@gmail.com>
Subject: Re: [lkp-robot] [include/linux/string.h] 6974f0c455:
 kernel_BUG_at_lib/string.c
From: Daniel Micay <danielmicay@gmail.com>
To: Linus Torvalds <torvalds@linux-foundation.org>,
        kernel test robot <xiaolong.ye@intel.com>,
        Ananth N Mavinakayanahalli <ananth@linux.vnet.ibm.com>,
        Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>,
        Masami Hiramatsu <mhiramat@kernel.org>
Cc: Kees Cook <keescook@chromium.org>, Arnd Bergmann <arnd@arndb.de>,
        Mark Rutland <mark.rutland@arm.com>, Daniel Axtens <dja@axtens.net>,
        Rasmus Villemoes <linux@rasmusvillemoes.dk>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Chris Metcalf <cmetcalf@ezchip.com>,
        Thomas Gleixner <tglx@linutronix.de>, "H. Peter Anvin" <hpa@zytor.com>,
        Ingo Molnar <mingo@elte.hu>, Andrew Morton <akpm@linux-foundation.org>,
        LKML <linux-kernel@vger.kernel.org>, LKP <lkp@01.org>
Date: Thu, 20 Jul 2017 00:41:36 -0400
In-Reply-To: <CA+55aFwuebKQ0-sAiRJWpMcaVJ8MjXvCg56DO5q8jMiPDXvc8A@mail.gmail.com>
References: <20170720014238.GH27396@yexl-desktop>
         <CA+55aFwuebKQ0-sAiRJWpMcaVJ8MjXvCg56DO5q8jMiPDXvc8A@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 925
Lines: 25

> So the fortify_string code has decided that only a single-byte (or
> empty) memcpy is ok.
> 
> And that, in turn, seems to be because we're copying from
> optprobe_template_entry, which is declared as
> 
>     extern __visible kprobe_opcode_t optprobe_template_entry;
> 
> so the fortify code decides it's a single character.
> 
> Does just changing all those things to be declared as arrays fix
> things?

Yeah, that fixes it because GCC will consider the size of 'char foo[]'
unknown (i.e. (size_t)-1 from __builtin_object_size).

GCC doesn't know this essentially constant value at compile-time so it
wasn't a compile-time error:

#define TMPL_END_IDX \
	((long)&optprobe_template_end - (long)&optprobe_template_entry)

-fsanitize=object-size works the same way for pointer dereferences so
replacing might fix some issues for CONFIG_UBSAN_SANITIZE_ALL. I guess
that's way too noisy at the moment thus the !COMPILE_TEST.