linux-kernel - Re: [RFC PATCH] powerpc: use __builtin

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20190818122036.GW31406@gate.crashing.org>
Date:   Sun, 18 Aug 2019 07:20:36 -0500
From:   Segher Boessenkool <segher@...nel.crashing.org>
To:     Christophe Leroy <christophe.leroy@....fr>
Cc:     Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        linux-kernel@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [RFC PATCH] powerpc: use __builtin_trap() in BUG/WARN macros.

Hi Christophe,

On Sat, Aug 17, 2019 at 06:37:50PM +0000, Christophe Leroy wrote:
>  #define BUG() do {						\
> +	__builtin_trap();					\

GCC will optimise away all code after this, it knows it is unreachable.
But you want to keep that BUG_ENTRY stuff I think?

The same will happen with a BUG_ON if the compiler can prove your
condition is always true.

>  	__asm__ __volatile__(					\
> -		"1:	twi 31,0,0\n"				\
> +		"1:\n"						\
>  		_EMIT_BUG_ENTRY					\
>  		: : "i" (__FILE__), "i" (__LINE__),		\
>  		    "i" (0), "i"  (sizeof(struct bug_entry)));	\

(GCC wil generate a different trap btw; what is called "trap" as extended
mnemonic, that is, "tw 31,0,0", not the same thing as "twi", if that
matters for the exception handler).

Segher