linux-kernel - select validation (was: Re: [2.6 patch] HID: fix default building of all quirky devices)

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:	Thu, 16 Oct 2008 10:06:41 +0200 (CEST)
From:	Geert Uytterhoeven <geert@...ux-m68k.org>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
cc:	Jiri Kosina <jkosina@...e.cz>, Adrian Bunk <bunk@...nel.org>,
	zippel@...ux-m68k.org, sam@...nborg.org,
	linux-input@...r.kernel.org, linux-kernel@...r.kernel.org,
	Jiri Slaby <jirislaby@...il.com>,
	Mauro Carvalho Chehab <mchehab@...hat.com>
Subject: select validation (was: Re: [2.6 patch] HID: fix default building
 of all quirky devices)

On Wed, 15 Oct 2008, Linus Torvalds wrote:
> On Wed, 15 Oct 2008, Jiri Kosina wrote:
> > Hmm ... I am not really sure, isn't this a bug in Kconfig?
> 
> Not really, more of a misfeature.
> 
> That said, even if we were to consider it a bug, Adrian's patch is 
> obviously the better way to do it. You shouldn't do non-local dependencies 
> (have one config option select fifty other ones), when the local ones are 
> clearer and more readable (have one config option just describe its *own* 
> dependencies).
> 
> Don't get me wrong - select is very useful, but not for "should I ask this 
> question or not". The point to use select is when you have some library or 
> other common infrastructure that isn't worth a question of its own (eg 
> "Why the hell would the user want to care whether he needs CRC32 
> routines?"), and some code says "I will need this infrastructure" by just 
> saying "select CRC32" to let the build system know that it needs that 
> particular piece of functionality.
> 
> In other words, "select" is kind of a "depends on", but for things that it 
> is insane to ask. It would be totally _idiotic_ to ask a user "do you want 
> to have CRC32 routines in the kernel?" and then based on that say "ok, you 
> didn't ask for CRC32, so now you cannot use the AX88796 network driver".
> 
> See? THAT is what "select" is for. And when you use select, the way we do 
> things now, you have to select everything you need. You cannot assume that 
> it will recursively select whatever it needs.
> 
> And notice how you mis-use select. That's not how to disable a question. A 
> question gets disabled by just doing an "if xyz" on the question itself, 
> like Adrian did.

Would it be possible for kconfig to check for invalid usage of select?

Like you should not use select to enable something that has
dependencies. However, that would fail in case both the selector option
and the selected option depend on the same.

And doing it dynamically (you should not use select to enable something that
has dependencies that are (currently) not fulfilled) wouldn't find all
incorrect usages.

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/