Re: -msoft-float for 386?

• Date: Mon, 18 Nov 1996 09:34:43 -0800
• From:
  Doug Evans <dje@cygnus.com>
• Reply-To:
  Doug Evans <dje@cygnus.com>
• CC: crossgcc@cygnus.com
• In-reply-to: <56pvi9$ma0@halon.vggas.com> (JYoungman@vggas.com)

   From: JYoungman@vggas.com (James Youngman)
   Date: 18 Nov 1996 15:32:25 GMT

	   I just compiled a trivial program under 9386 Linux using -msoft-float
   and read FP instructions in the (-static) linked result because the C library 
   wasn't compiled -msoft-float.  This isn't a problem under Linux itself but it 
   is for my embedded target.  Is there a libgcc.a for the 386 that doesn't use FP 
   instructions?  What ./configure command should I use to get the various 
   combinations?  Are these availabel in RPM form?

What configuration are you using for your embedded target?

There are several ways to proceed, but the easiest to use an
embedded i386 target (i386-aout or some such), and work from there
by, say, adding this to config/i386/t-i386bare.
grep for fp-bit in config/*/t-* for more examples.

Note that this doesn't support long doubles, just floats and doubles.

--- add to t-i386bare ---
# These are really part of libgcc1, but this will cause them to be
# built correctly, so...

LIB2FUNCS_EXTRA = fp-bit.c dp-bit.c

dp-bit.c: $(srcdir)/config/fp-bit.c
	cat $(srcdir)/config/fp-bit.c > dp-bit.c

fp-bit.c: $(srcdir)/config/fp-bit.c
	echo '#define FLOAT' > fp-bit.c
	cat $(srcdir)/config/fp-bit.c >> fp-bit.c

MULTILIB_OPTIONS = msoft-float
MULTILIB_DIRNAMES = soft

LIBGCC = stmp-multilib
INSTALL_LIBGCC = install-multilib

• -msoft-float for 386?
• From: JYoungman@vggas.com (James Youngman)

• Prev: -msoft-float for 386?
• Next: [vaysbor@charlie.cns.iit.edu: Need mips cross compiler]
• Index: Cross gcc Mailing list
• Thread: Cross gcc Mailing list