linux-kernel - Re: gcc 4.5.1 / as 2.20.51.0.11 miscompiling drivers/char/i8k.c ?

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

Date:	Mon, 15 Nov 2010 20:12:49 +0100
From:	Jakub Jelinek <jakub@...hat.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Jim Bos <jim876@...all.nl>, Andi Kleen <andi@...stfloor.org>,
	James Cloos <cloos@...loos.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Andreas Schwab <schwab@...hat.com>,
	Michael Matz <matz@...e.de>,
	Dave Korn <dave.korn.cygwin@...il.com>,
	Richard Guenther <richard.guenther@...il.com>, gcc@....gnu.org
Subject: Re: gcc 4.5.1 / as 2.20.51.0.11 miscompiling drivers/char/i8k.c ?

On Mon, Nov 15, 2010 at 07:58:48PM +0100, Jakub Jelinek wrote:
> Now, not sure why this happens, as there is
>     case GIMPLE_ASM:
>       for (i = 0; i < gimple_asm_nclobbers (stmt); i++)
>         {
>           tree op = gimple_asm_clobber_op (stmt, i);
>           if (simple_cst_equal(TREE_VALUE (op), memory_identifier_string) == 1)
>             {
>               if (dump_file)
>                 fprintf (dump_file, "    memory asm clobber is not const/pure");
>               /* Abandon all hope, ye who enter here. */
>               local->pure_const_state = IPA_NEITHER;
>             }
>         }
> Debugging...

Ah, the problem is that memory_identifier_string is only initialized in
ipa-reference.c's initialization, so it can be (and is in this case) NULL in
ipa-pure-const.c.

Two possible fixes (the latter is apparently what is used in
tree-ssa-operands.c, so is probably sufficient).  Guess ipa-reference.c
should be changed to do the same and just drop memory_identifier_string.

	Jakub

View attachment "1" of type "text/plain" (706 bytes)

View attachment "2" of type "text/plain" (562 bytes)