linux-kernel - Re: [PATCH] kmemcheck: fix sparse warning

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, 8 Jul 2009 23:29:08 -0700
From:	Josh Triplett <josh@...htriplett.org>
To:	Vegard Nossum <vegard.nossum@...il.com>
Cc:	Linux Torvalds <torvalds@...ux-foundation.org>,
	Johannes Berg <johannes@...solutions.net>,
	Pekka J Enberg <penberg@...helsinki.fi>,
	linux-kernel <linux-kernel@...r.kernel.org>,
	linux-sparse@...r.kernel.org
Subject: Re: [PATCH] kmemcheck: fix sparse warning

[Adding Linus and Chris Li to CC; Linus for further background on
-Wdo-while, and Chris Li for Sparse.]

On Wed, Jul 08, 2009 at 09:28:24PM +0200, Vegard Nossum wrote:
> 2009/7/6 Johannes Berg <johannes@...solutions.net>:
> > Whether or not the sparse warning
> >
> > warning: do-while statement is not a compound statement
> >
> > is justified or not in this case, it is annoying and
> > trivial to fix.
[...]
> 
> I'll change the patch title to "kmemcheck: work around bogus sparse
> warning" and fix the indentation, sounds ok?
> 
> Meanwhile, I Cced sparse mailing list in case somebody else knows
> anything else about this warning (what it means, whether it's
> justified in this case, whether it should be fixed in sparse, etc.).

-Wdo-while gives a warning if you write:

do
    statement
while (...);

where "statement" does not consist of a compound statement surrounded by
braces.  As far as I know, this warning exists primarily because it
matched Linus's preference for readability.

However, note that Sparse does not have -Wdo-while enabled by default.
Sparse only issues that warning if you use -Wdo-while explicitly, or if
you pass -Wall.  (The latter often happens due to passing the same
warning flags to GCC and Sparse, without using cgcc which filters out
-Wall for Sparse for that reason.)

Much discussion has occurred previously suggesting that Sparse's -Wall
should match GCC's "all useful warnings" rather than "all possible
warnings", and some other option should exist for "all possible
warnings" (for instance, -Weverything).  This seems very reasonable.
Given that Sparse *exists* to give warnings, -Wall's "all useful
warnings" approach seems like it matches Sparse's default behavior;
thus, I think it would make the most sense to make Sparse just interpret
-Wall as "enable all the warnings enabled by default" (which would
normally act as a no-op, but would make a difference if you passed
"-Wno-foo -Wall" for some default warning foo).

This would improve Sparse's behavior for the case where you invoke
sparse $(CFLAGS) directly rather than setting CC=cgcc, if CFLAGS
contains -Wall.

- Josh Triplett
--
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/