linux-kernel - Re: any value to "NORET

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:	Wed, 23 May 2007 19:16:24 +0530
From:	"Satyam Sharma" <satyam.sharma@...il.com>
To:	"Robert P. J. Day" <rpjday@...dspring.com>
Cc:	"Krzysztof Halasa" <khc@...waw.pl>, "Adrian Bunk" <bunk@...sta.de>,
	"Linux Kernel Mailing List" <linux-kernel@...r.kernel.org>
Subject: Re: any value to "NORET_TYPE" macro?

Hi Robert,

On 5/23/07, Robert P. J. Day <rpjday@...dspring.com> wrote:
> On Wed, 23 May 2007, Satyam Sharma wrote:
>
> > On 5/23/07, Krzysztof Halasa <khc@...waw.pl> wrote:
> > > "Robert P. J. Day" <rpjday@...dspring.com> writes:
> > >
> > > > that may be but, as i suggested earlier, that would get into
> > > > guessing what those developers were thinking, and i just didn't
> > > > want to go there.
> > >
> > > No guessing, I just checked it (though a second check wouldn't do
> > > any harm).
> > >
> > > > the simple version of the patch is now in andrew's tree, and
> > > > i'll worry about the harder stuff next time.
> > >
> > > The "next time" would be much harder as there would be no key for
> > > searching for these functions.
> >
> > Krzysztof's absolutely right ... we don't want to lose the
> > NORET_TYPE annotations on all these functions before we switch them
> > to ATTRIB_NORET. And yes, _all_ of these NORET_TYPE's do want to be
> > ATTRIB_NORET (except for those that are double-annotated, for those
> > we can just get rid of the NORET_TYPE macro).
>
> ok, i'll go back and take another look at this.

Actually there's another thing :-) The __attribute__((xxx)) must go with the
function _declarations_ (and not the implementations/definitions). I noticed
after my previous mail that most of the double annotations are actually in
the case of the _declarations_ of these non-returning functions, whereas
most of the single-occurrences of NORET_TYPE were in the function
definitions, which means your patch that simply got rid of NORET_TYPE
actually ended up doing exactly the right thing that we wanted :-)

The unfortunate / ugly part, however, is that there are few cases where
the above doesn't hold true ([1] NORET_TYPE without ATTRIB_NORET in
function declaration e.g. in include/asm-mips/ptrace.h -- this must
become ATTRIB_NORET, and [2] ATTRIB_NORET either alone or with
NORET_TYPE in function definitions e.g. arch/mips/kernel/traps.c -- these
can be eliminated entirely, and [3] NORET_AND occurrences where we
don't really want ATTRIB_NORET anymore, so just removing
NORET_TYPE would be right).

So this might not really be a simple patch anymore, and because the
benefit of ATTRIB_NORET is only miniscule, so you can ignore my
previous mail if you want :-)

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