This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: strange compiling problem

From: Jamie Lokier <jamie at shareable dot org>
To: Andrew Pinski <pinskia at physics dot uc dot edu>
Cc: Kyle Centers <kcent at programmer dot net>, gcc at gcc dot gnu dot org
Date: Sun, 7 Sep 2003 16:44:19 +0100
Subject: Re: strange compiling problem
References: <20030907152221.GJ19977@mail.jlokier.co.uk> <B835CB4E-E147-11D7-A011-000393A6D2F2@physics.uc.edu>

Andrew Pinski wrote:
> >	:"=m"(*&ra)
> >	:"0"(*&ra), "m"(*&rb)
> 
> 
> Even better, I think this is where the problem is any way:
> 	:"+m"(*&ra)
> 	: "m"(*&rb)
> 
> And change all %2 to %1 and %1 to %0.
> Note the + means that ra is being read from and written to.
> Almost every one wants to write it as the way you had it
> though for some reason.

It's because older GCC didn't support the "+" syntax in asms, and
current GCC supports either and automatically converts the first form
into the second one, in an asm.

-- Jamie

References:
- Re: strange compiling problem
  - From: Jamie Lokier
- Re: strange compiling problem
  - From: Andrew Pinski

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]