Date: Thu, 22 Jan 2004 07:27:52 -0800 (PST)
From: Linus Torvalds <torvalds@osdl.org>
To: Richard Henderson <rth@twiddle.net>
cc: Valdis.Kletnieks@vt.edu,
       Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: 2.6.1-mm5 versus gcc 3.5 snapshot
In-Reply-To: <20040122060253.GA18719@twiddle.net>
Message-ID: <Pine.LNX.4.58.0401220725180.2123@home.osdl.org>
References: <200401212236.i0LMaNuh020491@turing-police.cc.vt.edu>
 <Pine.LNX.4.58.0401212043200.2123@home.osdl.org> <20040122060253.GA18719@twiddle.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1137
Lines: 39


On Wed, 21 Jan 2004, Richard Henderson wrote:
> 
> You're reading that wrong way-round.  It's "+m" and "=m"/"0" that's
> disallowed.

Ok, but...

>	  I.e. if you have matching constraints (or read-write
> constrants, which are exactly short-hand for matching constraints),
> then you *must* have a register alternative.  I.e. you'll get this
> warning if you *only* allow memories.
> 
> The problem is partially conceptual -- what in the world does
> 
> 	"=m"(x) : "0"(y)

I agree about the latter one, but "+m" (which is what the kernel uses) has
well-defined meaning, and the compiler would be/is silly to complain about
it.

So your arguments fall down flat. If it was

	"=m" (x) : "0" (y)

I'd agree with you, but that's not the code the compiler complains about.

Shorthand or not, the "+m" usage is (a) totally logical and (b) 
historically allowed.

Please fix the compiler.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/