From mboxrd@z Thu Jan  1 00:00:00 1970
From: jamie@shareable.org (Jamie Lokier)
Date: Sun, 28 Mar 2010 00:14:29 +0000
Subject: [PATCH 1/1] [RFCv2] arm: add half-word __xchg
In-Reply-To: <20100327225240.GC13203@n2100.arm.linux.org.uk>
References: <20100318135008.GA11800@Krystal>
	<1269535366-25175-1-git-send-email-virtuoso@slind.org>
	<20100327225240.GC13203@n2100.arm.linux.org.uk>
Message-ID: <20100328001429.GA16697@shareable.org>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Russell King - ARM Linux wrote:
> I wonder if we should be using __alignof__ here.
> 
> 	unsigned long *ptrbig = (unsigned long *)((unsigned long)ptr &
> 		(__alignof__(unsigned long) - 1));

Are there ARM targets with a smaller value from __alignof__()?
I think you meant:

	unsigned long *ptrbig = (unsigned long *)((unsigned long)ptr &
		~(unsigned long)(__alignof__(unsigned long) - 1));

Perhaps in asm-generic that has a place.  It would simplify the asm if
the alignment is 1 on some machine.

But I don't think it'd happen anyway.  There are machines which don't
require full alignment of long, but insufficiently aligned *atomic*
accesses like cmpxchg are *not atomic*.  x86 is one such machine.
Fortunately GCC aligns the types sufficiently well - and we rely on
that all over the kernel.

I'm not sure about ARM, when doing a 64-bit cmpxchg, if the doubleword
must be 64-bit aligned to get atomicity.

Note that the posted code doesn't work as is for 64-bit longs: the
"mask" calculation overflows if called with size >= 4.

But seeing as this is for ARM only at present, I'd just change the
types to u32 and be done with it.  It does seem like a good thing to
go to asm-generic eventually though.

-- Jamie