linux-kernel - Re: dealing with excessive includes

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, 23 Oct 2006 18:52:39 +0200 (CEST)
From:	Geert Uytterhoeven <geert@...ux-m68k.org>
To:	Linus Torvalds <torvalds@...l.org>
cc:	Andi Kleen <ak@...e.de>, Randy Dunlap <rdunlap@...otime.net>,
	Stefan Richter <stefanr@...6.in-berlin.de>,
	Al Viro <viro@....linux.org.uk>,
	Alexey Dobriyan <adobriyan@...il.com>,
	Linux Kernel Development <linux-kernel@...r.kernel.org>,
	linux-arch@...r.kernel.org
Subject: Re: dealing with excessive includes

On Mon, 23 Oct 2006, Linus Torvalds wrote:
> On Mon, 23 Oct 2006, Geert Uytterhoeven wrote:
> > > We have tons of issues that depend on config variables and architecture 
> > > details. 
> > 
> > Indeed, so the config variables and architecture details should be handled in
> > the include files, not in the (multiple) users of those include files.
> 
> The point is - _verifying_ that is actually hard.
> 
> If some inline function depends on a particular header, you'll have a hard 
> time checking for that if there's an #ifdef around it. Which is not 
> uncommon, we have things like:
> 
> 	#ifdef CONFIG_PROCFS
> 	.. number of inline functions ..
> 	#else
> 	#define function1(a,b,c) do { } while (0)
> 	...
> 	#endif
> 
> so I'm just saying that "just compile it" is _not_ a way of verifying that 
> the header file is complete - because it may well be complete for the 
> particular config you're testing, but not for some other.
> 
> So this is a hard problem. If it was easy, we'd not _have_ the problem in 
> the first place.

I agree _verifying_ this for all config and arch combinations is hard.
But my point is that right now we're `solving' this at the user (of the
include) level, which is an order of magnitude more work.
If the includes were (sufficiently) self-contained, the driver writers would
have to care less about config/arch dependencies.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds
-
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/