From: Johannes Weiner <hannes@saeurebad.de>
To: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mike Travis <travis@sgi.com>, linux-kernel@vger.kernel.org,
       "H. Anvin" <hpa@zytor.com>, Christoph Lameter <cl@linux-foundation.org>,
       Ingo Molnar <mingo@elte.hu>
Subject: Re: Dangerous code in cpumask_of_cpu?
References: <200807081816.40623.rusty@rustcorp.com.au>
	<87myksn587.fsf@saeurebad.de> <87iqvgn4c1.fsf@saeurebad.de>
	<87ej64n3xt.fsf@saeurebad.de>
Date: Tue, 08 Jul 2008 11:28:32 +0200
In-Reply-To: <87ej64n3xt.fsf@saeurebad.de> (Johannes Weiner's message of "Tue,
	08 Jul 2008 11:03:10 +0200")
Message-ID: <877ibwn2rj.fsf@saeurebad.de>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.0.60 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1076
Lines: 32


[ fixed christoph's address in cc]

Johannes Weiner <hannes@saeurebad.de> writes:

>> I guess because we dereference it immediately again, the location is not
>> clobbered yet.  At least in my test case, gcc assembled it to code that
>> puts the address in eax and derefences it immediately, before eax is
>> reused:
>
> Gee, just ignore this bs.  The address is in eax, not the value.

My theory was half-right.  Since the code is a macro, there is no call
and hence no stack clean-up.  And although it is UB, it works correctly
as the value is not yet clobbered when we access it again.  Converting
foo to a macro yields this:

        movl    $42, -8(%ebp)
        leal    -8(%ebp), %eax
        movl    (%eax), %eax
	...
	ret

gcc only emits a warning if the scope we leak a local address from is
that of a function.

	Hannes
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/