Can we emulate i386-pc-pe-coff with i386coff?

Can we emulate i386-pc-pe-coff with i386coff?

[H . J . Lu]

I don't think we can emulate i386-pc-pe-coff with i386coff. I don't
even know we can emulate i386-pc-pe-coff at all. The problem is
config/tc-i386.c has #ifdef TE_PE inside. We need to modify it to
check TE_PE at the runtime or do something else. Anyway, this patch
disable the i386coff emulation for i386-pc-pe-coff.


-- 
H.J. Lu (hjl@gnu.org)
----
2000-06-14  H.J. Lu <hjl@gnu.org>

	* configure.in: Don't emulate i386-pc-pe-coff with i386coff.
	* configure: Rebuild.

Index: configure.in
===================================================================
RCS file: /work/cvs/gnu/binutils/gas/configure.in,v
retrieving revision 1.24
diff -u -p -r1.24 configure.in
--- configure.in	2000/06/13 19:25:38	1.24
+++ configure.in	2000/06/14 18:00:46
@@ -557,6 +557,8 @@ changequote([,])dnl
 			big)	emulation="mipsbelf mipslelf mipself" ;;
 			*)	emulation="mipslelf mipsbelf mipself" ;;
 			esac ;;
+      # i386-pc-pe-coff != i386-pc-coff.
+      i386-*-pe-coff)	;;
       # Uncommenting the next line will turn on support for i386 AOUT
       # for the default linux configuration
       # i386-*-linux*-elf) emulation="i386elf i386aout" ;;

Re: Can we emulate i386-pc-pe-coff with i386coff?

[Alan Modra]

On Wed, 14 Jun 2000, H . J . Lu wrote:

> I don't think we can emulate i386-pc-pe-coff with i386coff. I don't
> even know we can emulate i386-pc-pe-coff at all. The problem is
> config/tc-i386.c has #ifdef TE_PE inside. We need to modify it to
> check TE_PE at the runtime or do something else. Anyway, this patch
> disable the i386coff emulation for i386-pc-pe-coff.

Agreed.  I dislike all those #ifdefs.  What's really needed is an overhaul
of the generic reloc handling code so that all the back-end fudges can be
removed.  A major undertaking.

-- 
Linuxcare.  Support for the Revolution.

New cross toolchain development utility...

[Steve Hill]

[-- Attachment #1: Type: text/plain, Size: 1394 bytes --]

Greetings.

It seems that every couple of weeks someone is asking about building a
toolchain, blah, blah, blah. I myself had a brain fart earlier this week
when I forgot to apply a patch. I also want to be able to keep my toolchain
up to date when a new kernel is released (especially since I am keeping up
with the development series). So, I decided to write a kind of smart build
utility that builds an entire ARM toolchain from start to finish,
automatically applies patches at the right time and provides logging and
informative messages. The fruits of my labor are attached below. I would be
interested in knowing how it works for everyone. I see no reason why it
cannot be used for other targets as well.

Also, I have included in the archive the stage1 and stage3 patches for GCC
that were made up of patches by Phil Blundell. If you do a diff between the
two patches you'll see the only change is the -Dinhibit hack. Let me know
how it works for everyone. I tested this with the following:

  binutils-2.10-rc3
  gcc-2.95.2
  glibc-2.1.3 (with crypt and linuxthread ad-ons)
  linux-2.4.0-test1 (with Russell King and personal patches)

I compiled working kernels and applications that ran perfectly. Cheers.

-Steve

-- 
 Steven J. Hill - Embedded SW Engineer
 Public Key: 'finger sjhill@mail.cotw.com'
 FPR1: E124 6E1C AF8E 7802 A815
 FPR2: 7D72 829C 3386 4C4A E17D
cxtool-0.1.tar.gz


[-- Attachment #2: cxtool-0.1.tar.gz --]
[-- Type: application/x-gzip, Size: 31227 bytes --]

My i386-pc-pe-coff pacth to gas/configure.in [Re: Can we emulate i386-pc-pe-coff with i386coff?]

[H . J . Lu]

On Thu, Jun 15, 2000 at 12:47:56PM +1000, Alan Modra wrote:
> On Wed, 14 Jun 2000, H . J . Lu wrote:
> 
> > I don't think we can emulate i386-pc-pe-coff with i386coff. I don't
> > even know we can emulate i386-pc-pe-coff at all. The problem is
> > config/tc-i386.c has #ifdef TE_PE inside. We need to modify it to
> > check TE_PE at the runtime or do something else. Anyway, this patch
> > disable the i386coff emulation for i386-pc-pe-coff.
> 
> Agreed.  I dislike all those #ifdefs.  What's really needed is an overhaul
> of the generic reloc handling code so that all the back-end fudges can be
> removed.  A major undertaking.
> 

I haven't heard any comments on my patch. Is that ok with everyone?

-- 
H.J. Lu (hjl@gnu.org)