RTAI 3.2 and new binutils

[NZG]

> You might be right that binutils is not the culprit.
Supposedly it's GAS that's the ultimate culprit, which I thought was tied to 
binutils, but how Hannes is not seeing it I don't know, I will have to keep 
investigating.

NZG

On Wednesday 25 January 2006 2:51 pm, J.D. Yamokoski wrote:
> You might be right that binutils is not the culprit. But is the fix just
> a matter of doing a few find/replace's on "mov" in the RTAI code? That's
> what seemed to fix the kernel problems. For instance, before applying
> linux-2.6-seg-5.patch
> <http://www.kernel.org/pub/linux/devel/binutils/linux-2.6-seg-5.patch> I
> got the following errors:
>
> CC arch/i386/kernel/process.o
> {standard input}: Assembler messages:
> {standard input}:880: Error: suffix or operands invalid for `mov'
> ....
> {standard input}:1324: Error: suffix or operands invalid for `mov'
> make[1]: *** [arch/i386/kernel/process.o] Error 1
> make: *** [arch/i386/kernel] Error 2
>
> After applying the patch, no more problems. 99% of that patch is the
> following:
>
> -    asm volatile("movl %%fs,%0":"=m" (*(int *)&prev->fs));
> -    asm volatile("movl %%gs,%0":"=m" (*(int *)&prev->gs));
> +    asm volatile("mov %%fs,%0":"=m" (prev->fs));
> +    asm volatile("mov %%gs,%0":"=m" (prev->gs));
>
> Now however, I get the same error signature in the RTAI code:
>
> CC adeos/x86.o
> {standard input}: Assembler messages:
> {standard input}:401: Error: suffix or operands invalid for `mov'
> ....
> {standard input}:1418: Error: suffix or operands invalid for `mov'
> make[1]: *** [adeos/x86.o] Error 1
> make: *** [adeos] Error 2
>
>
> J.D. Yamokoski
>
> Hannes Mayer wrote:
> > J.D. Yamokoski wrote:
> >> I posted a message to the list just a few days ago regarding this
> >> same problem. I was getting mov compile errors in RTAI code. I was
> >> getting them in the linux kernel as well until found the same patches
> >> you reference. Unfortunately, the responses I got to my post were
> >> simply, "RTAI does not support GCC4, try recompiling with GCC3.2" -
> >> which does not address the problem whatsoever.
> >
> > Ok, it does not address your problem, but compiling with GCC3.3.x most
> > likely avoids other issues.
> >
> > I just checked the latest binutils version:
> > It is 2.16.1, right ?
> > I checked my development machine (debian sarge):
> > # ld --version
> > GNU ld version 2.16.1 Debian GNU/Linux
> >
> > Since I haven't seen such errors like you do, I'm guessing, but maybe
> > binutils isn't the problem here.
> >
> > Best regards,
> > Hannes.
>
> _______________________________________________
> RTAI mailing list
> RTAI at rtai.org
> https://mail.rtai.org/cgi-bin/mailman/listinfo/rtai