linux-kernel - Re: [tip:core/locking] locking, x86: Slightly shorten __ticket_spin

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, 02 Dec 2009 09:23:49 -0800
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
CC:	Jan Beulich <JBeulich@...ell.com>, a.p.zijlstra@...llo.nl,
	mingo@...e.hu, tglx@...utronix.de, mingo@...hat.com,
	npiggin@...e.de, linux-kernel@...r.kernel.org
Subject: Re: [tip:core/locking] locking, x86: Slightly shorten  __ticket_spin_trylock()

On 12/02/2009 09:05 AM, Linus Torvalds wrote:
> 
> Btw, even if gcc just treats 'bool' as 'char' (which is the sane thing to 
> do on x86 anyway - I don't see why it should ever do anything else), we 
> actually mess up in the kernel and make that type confusion even worse.
> 

For what it's worth, the gcc ABI for i386-Linux treats _Bool (bool) as
follows:

When in memory, except stack slots:

	sizeof(_Bool) = 1
	0 is false, 1 is true, any other value is *undefined behavior*.

When in registers, or in a stack slot:

	Registers, and stack slots, are always 4 bytes
	0 is false, 1 is true, any other value is *undefined behavior*.
	
All of which is well-defined.  However, it also only applies to values
in memory, or values passed across function boundaries in registers[1]
-- anything else is *by definition outside the scope of the ABI* and
therefore the compiler can legitimately do whatever is appropriate at
any point in time.

As such, I would agree with Linus in that using an u8 is the right thing
to be handed through the inline asm boundary -- if the compiler needs to
extend it, it will, and if it doesn't, there is no penalty.  Similarly,
a bool can be cast to u8 without penalty.

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.

--
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/