Re: [RFA:] Fix bug with #APP/#NO_APP when using macros.

[Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>]

> From: Zack Weinberg <zack@codesourcery.com>
> Date: Thu, 08 May 2003 21:01:14 -0700

> Ian Lance Taylor <ian@airs.com> writes:
> 
> > It's true that #NO_APP needs to be printed in ASM_FILE_START, but it's
> > not true that the CRIS target is the only target which does it.
> > m68k.h does it also.  I actually thought that the m68k was the only
> > target which used #NO_APP, but I see that you used it for the CRIS.

Right, I see now and should have looked closer at m68k.h (lots
of ASM_FILE_START in config/m68k/*) particularly since you
pointed out that port.  Sorry for the misinformation.  (Luckily
I used weasel phrasing.  Bother, it seems to prove I can't tell
very far! :-)

> > I take your point, though, which is that though targets other than
> > CRIS and m68k emit #APP/#NO_APP around assembler statements, it is
> > ineffective and pointless.
> 
> And since no one has noticed this up to now, does that mean that there
> is no real performance benefit to be had from #NO_APP, and we can blow
> away all the code that emits it from gcc?  Or should gcc be changed to
> emit #NO_APP at file start?

The latter, per port, with caution and after fixing glibc, since
Alan's measurements (thanks!) prove it's good enough (IMHO) to
use it.  Rare insn output templates in particular need
inspection for strict formatting.  Check the emitted trampoline
code too, or emit #APP/#NO_APP around it.

brgds, H-P