This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils project.

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

Re: PATCH for i386: --no-copy-relocs

To: Richard Henderson <rth at redhat dot com>
Subject: Re: PATCH for i386: --no-copy-relocs
From: Alan Modra <amodra at bigpond dot net dot au>
Date: Thu, 28 Jun 2001 23:46:32 +0930
Cc: binutils at sources dot redhat dot com
References: <002101c0f82c$935bf2a0$18671ed4@fr.mainsoft.com> <20010619233358.G27171@bubble.local> <sizob4nqia.fsf@daffy.airs.com> <20010620093554.C16699@bubble.local> <20010627112514.A15281@redhat.com>

On Wed, Jun 27, 2001 at 11:25:14AM -0700, Richard Henderson wrote:
> On Wed, Jun 20, 2001 at 09:35:54AM +0930, Alan Modra wrote:
> > As a side note, doing this does not provide any benefit at run-time for
> > data accesses, besides reducing dynamic loader time;  Data accesses still
> > go through the .got.
> 
> They wouldn't necessarily need to though.  You could implement
> a form of relaxation on x86 that transformed
> 
> 	movl    foo@GOT(%ebx), %edx
> into
> 	leal    foo@GOTOFF(%ebx), %edx
> 
> which is convenient in that the instructions are the same size.

Hmm, yes.  Or even
	leal	foo, %edx

Alan

References:
- PATCH for i386: --no-copy-relocs
  - From: Jose Luu
- Re: PATCH for i386: --no-copy-relocs
  - From: Alan Modra
- Re: PATCH for i386: --no-copy-relocs
  - From: Ian Lance Taylor
- Re: PATCH for i386: --no-copy-relocs
  - From: Alan Modra
- Re: PATCH for i386: --no-copy-relocs
  - From: Richard Henderson

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