Re: [PATCH] Fix the cpumask rewrite

From: Linus Torvalds <torvalds@osdl.org>
To: James Bottomley <James.Bottomley@SteelEye.com>
Cc: Vojtech Pavlik <vojtech@suse.cz>, Andrew Morton <akpm@osdl.org>,
	Paul Jackson <pj@sgi.com>,
	PARISC list <parisc-linux@lists.parisc-linux.org>,
	Linux Kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] Fix the cpumask rewrite
Date: Sat, 26 Jun 2004 12:13:51 -0700 (PDT)	[thread overview]
Message-ID: <Pine.LNX.4.58.0406261212190.16079@ppc970.osdl.org> (raw)
In-Reply-To: <1088276531.1750.113.camel@mulgrave>

On Sat, 26 Jun 2004, James Bottomley wrote:
> 
> Our test bit implementation would then become:
> 
> static __inline__ int test_bit(int nr, const volatile void *address)
> {
> 	return __test_bit(nr, (const void *)address);
> }
> 
> That would keep our implementation happy.

You just _want_ to be screwed over whenever your gcc bugs are fixed, don't 
you?

Are you going to complain to the gcc people when they fix their bugs? Or 
are you going to spend months to debug problems that only happen for other 
people, because they happen to have fixed compilers?

There's a real reason why there is a "volatile" there on other 
architectures. 

		Linus