linux-kernel - Re: Fix empty macros in acpi.

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:	Tue, 3 Jul 2007 01:22:47 -0400
From:	Len Brown <lenb@...nel.org>
To:	Al Viro <viro@....linux.org.uk>, Dave Jones <davej@...hat.com>
Cc:	Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: Fix empty macros in acpi.

On Tuesday 12 June 2007 20:26, Al Viro wrote:
> On Tue, Jun 12, 2007 at 08:21:15PM -0400, Dave Jones wrote:
> > On Wed, Jun 13, 2007 at 01:00:29AM +0100, Al Viro wrote:
> >  > On Tue, Jun 12, 2007 at 07:33:09PM -0400, Dave Jones wrote:
> >  > > +#define DBG(x...) do { } while(0)
> >  > 
> >  > Eh...  Please, stop it - if you want a function-call-like no-op returning void,
> >  > use ((void)0).  At least that way one can say DBG(....),foo(), etc.
> > 
> > They both end up compiled to nothing anyway, so I'm not bothered
> > either way..   I'm not sure I follow why the syntax of that last part
> > is a good thing.  It looks like something we'd want to avoid rather
> > than promote?
> 
> If on one side of ifdef it's a void-valued expression, so it should be
> on another; the reason is that we don't get surprise differences between
> the builds...

true, DBG() in this case would expand to printk(), which returns int.

But in practice, DBG isn't used in any expressions -- and the other zillion
places in the kernel where it is used, it is used as in dave's patch.

Indeed, I don't see printk used in expressions either...

While I know it is common Linux style, and by default it is okay with gcc,
it seems somewhat half-baked to call functions and not check their return
value by default.  IMHO, if they return something, it should be checked,
or explicitly ignored -- or it shouldn't return anything in the first place...

> IOW, if it doesn't build in some context, it should consistently fail to
> build in that context.

whelp, it seems that the reason for this patch is this:

#define DBG()

if(...)
	DBG();
next_c_statement

which turns into
if(...) ;
next_c_statement

But since there is an intervening ';', this code is still functionally correct
and a decent compiler will delete the test altogether, yes?

So is there some real problem here that I missed,
or is this to make some code-checking tool that I don't have happy?

thanks,
-Len
-
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/