x86-assembly w/ binutils 2.10+, intel-syntax

[h-peter recktenwald address@hidden]

Alan Modra wrote:
> 
> On Mon, Oct 01, 2001 at 08:38:24PM +0000, h-peter recktenwald address@hidden 
> wrote:
> > problems (i.e. bugs, faults,...) w/ <as> of binutils 2.11.91 and w/ xgcc
> > 3.1.
> 
> Please try the latest binutils CVS version.  You problems may be fixed.
> The bug reporting address for binutils, and instructions on how to obtain
> the source can be found at http://sources.redhat.com/binutils/
>
> See http://gcc.gnu.org/bugs.html for instructions on reporting gcc bugs.

I tried almost any currently available version. no success.

many faults, i.e. defined constants compiled as memory ref's, no way to
get them in a register other than w/ <lea>, binary pseudo-ops won't work
at all, standard intel ops missing, & more such silly stuff (which all
works perfectly well in t&t-mode). I'd recommend the intel-"syntax"
implementor(s) to simply check their works against facts - where I'd offer
my assistance, though I won't do any "C"-ish "bug"-fixes, &c.

btw, did anyone ever compile a x586(k6/2)-pc-linux-gnu kernel with any of
those recent "as" versions (2.10.xx; 2.11.xx, 2.11.xx-cross32/64)?
pse, give me a notice on how you'd manage...

and, probably how to get rid of those incredibly stupid warnings about
what might be "deprecated" which produce loads of garbage, wasting time
(life!) and efficiently hiding any truely relevant messages - I'd ask the
program to #work#, not to #educate#. maybe "gcc" related, but also,
evidently, relate to most "GNU" stuff...

and, did anyone else ever manage to actually #useing# as in
.intel-"syntax" mode?
I'd be really grateful for any hint on how.


best,
        hp

-- 
             ==> mailto:address@hidden <==
        .hpr - h.-peter recktenwald, berlin - t: +49 30 85967858
Linux,Assembly,Forth: http://home.snafu.de/phpr/index-lx.shtml (en,de)
======================================================================