Re: -msoft-float for mipsel-linux?

["Kai Ruottu" <karuottu AT freenet.hut.fi>]

Tseng Chou Ming wrote:

> How can i enable "-msoft-float" for mipsel-linux-gcc?
> I tried to use "configure --nfp" but it doesn't work 

 The '-msoft-float' needs the 'fp-bit.c' / 'dp-bit.c' routines
for 'libgcc.a'. These are used now for the embedded 'mips-elf' and
'mips-ecoff' targets, but seemingly not for the 'system' targets.

 I would suggest you trying the following:

1. add a new line to the current 'mips*-linux' config entry (in gcc/configure),
   for a new 'tmake_file' of your own :

------------ mips*-linux entry -----------------------------
       mips*-*-linux*)                         # Linux MIPS, either endian.
+	    tmake_file=t-linux
	    xmake_file=x-linux
		xm_file="xm-siglist.h ${xm_file}"
		case $machine in
			mipsel-*)  tm_file="mips/elfl.h mips/linux.h" ;;
			*)         tm_file="mips/elf.h mips/linux.h" ;;
		esac
		extra_parts="crtbegin.o crtend.o"
		gnu_ld=yes
		gas=yes
		;;
------------ mips*-linux entry -----------------------------

 The text in the 'tmake_file' will be included to the resulted 'Makefile'
while running 'configure'...

 Then add the new 'gcc/config/t-linux', using the 'gcc/config/t-elf' as
a template (strongly stripped) :

------------ t-linux -----------------------------
# We have a premade insn-attrtab.c to save the hour it takes to run genattrtab.
# PREMADE_ATTRTAB = $(srcdir)/config/mips/mips-at.c
# PREMADE_ATTRTAB_MD = $(srcdir)/config/mips/mips-at.md

# Suppress building libgcc1.a, since the MIPS compiler port is complete
# and does not need anything from libgcc1.a.
LIBGCC1 =

# fp-bit and dp-bit are really part of libgcc1, but this will cause
# them to be built correctly, so... [taken from t-sparclite]
LIB2FUNCS_EXTRA = fp-bit.c dp-bit.c

dp-bit.c: $(srcdir)/config/fp-bit.c
	echo '#ifdef __MIPSEL__' > dp-bit.c
	echo '#define FLOAT_BIT_ORDER_MISMATCH' >> dp-bit.c
	echo '#endif' >> dp-bit.c
	echo '#define US_SOFTWARE_GOFAST' >> dp-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
	echo '#ifdef __MIPSEL__' >> fp-bit.c
	echo '#define FLOAT_BIT_ORDER_MISMATCH' >> fp-bit.c
	echo '#endif' >> fp-bit.c
	echo '#define US_SOFTWARE_GOFAST' >> fp-bit.c
	cat $(srcdir)/config/fp-bit.c >> fp-bit.c

# Build the libraries for both hard and soft floating point

MULTILIB_OPTIONS  = msoft-float
MULTILIB_DIRNAMES = soft-float
MULTILIB_MATCHES  = 

LIBGCC = stmp-multilib
INSTALL_LIBGCC = install-multilib
------------ t-linux -----------------------------

 Please compare this with the originally used, 'default', t-mips !!!
I didn't copy it and don't remember if the following lines :

# PREMADE_ATTRTAB = $(srcdir)/config/mips/mips-at.c
# PREMADE_ATTRTAB_MD = $(srcdir)/config/mips/mips-at.md

were with or without the comments there... The 't-linux' should
just be a superset of the 't-mips'...

 You can consider adding still more variations for the libs, like
the '-mips3', just look at the 't-elf'... The default is probably
still R3000...

 The extra libs will be installed into the subdirs mentioned in the
'MULTILIB_DIRNAMES', one name for each option in 'MULTILIB_OPTIONS'.
And 'gcc' will add the right '-L xxxx' for the used '-mxxxx' option.
(Here if you use '-msoft-float', the 'ld' command line will be put
to search first in the 'soft-float' subdir.)

Cheers, Kai

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com