From:	<ams AT ludd DOT luth DOT se>
Message-Id:	<200305072141.h47LfSeB010670@speedy.ludd.luth.se>
Subject:	Re: (fwd) Re: SIGILL 386 (illegal opcode)
In-Reply-To:	<200305072355.16047.pavenis@latnet.lv> "from Andris Pavenis at May
	7, 2003 11:55:15 pm"
To:	djgpp-workers AT delorie DOT com
Date:	Wed, 7 May 2003 23:41:28 +0200 (CEST)
X-Mailer:	ELM [version 2.4ME+ PL78 (25)]
MIME-Version:	1.0
X-MailScanner:	Found to be clean
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

According to Andris Pavenis:
> On Wednesday 07 May 2003 20:59, Martin Str|mberg wrote:
> > Alexei K. <kzua AT kzua DOT udm DOT ru> wrote:
> > : I deeply use STL. Was libstdxx.a library compiled for 386 CPU?
> >
> > Your problem indicate it was not.
> >
> > It should work on 386s.
> >
> > I hope the GCC maintainer corrects this for the following releases of
> > GCC.
> 
> xaddl can be found in inline assembler in atomicity.h. Therefore specifying 
> -march=i386 or -m386 doesn't help much. It's also in gcc-3.2.3, and there 
> will be no more  releases from gcc-3.2 branch.

1. No wonder as it says "// Low-level functions for atomic operations:
x86, x >= 4 version". 
gnu/gcc-3.22/libstdc++-v3/config/cpu/i386/bits/atomicity.h should be
used, not gnu/gcc-3.22/libstdc++-v3/config/cpu/i486/bits/atomicity.h
(which says "atomic operations: x86, x < 4 version").

2. It won't help as xaddl is present in that one too. Nevertheless the
i486 version will most likely never do the right thing for i386.

3. Meanwhile what's wrong with
gnu/gcc-3.22/libstdc++-v3/config/cpu/generic/bits/atomicity.h?


Right,

						MartinS

- Raw text -