This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
PATCH: build libgcc dfp routines for x86-64 GNU/Linux
- From: Ben Elliston <bje at au1 dot ibm dot com>
- To: gcc-patches at gcc dot gnu dot org
- Cc: jh at suse dot cz
- Date: Wed, 8 Feb 2006 08:35:24 +1100
- Subject: PATCH: build libgcc dfp routines for x86-64 GNU/Linux
This patch simply adds the makefile targets required to build dfp
runtime routines in to libgcc on x86-64 GNU/Linux. These are slightly
different to the common case, hence they are added to t-linux64.
These routines are still not built by default and are only activated
with --enable-decimal-float.
Bootstrapped on x86_64-unknown-linux-gnu. Okay to install?
2006-02-08 Ben Elliston <bje@au.ibm.com>
* config/i386/t-linux64 (D32PBIT, D64PBIT, D128PBIT): Define.
(dfp32-bit.c, dfp64-bit.c, dfp128-bit.c): New targets.
Index: config/i386/t-linux64
===================================================================
--- config/i386/t-linux64 (revision 110712)
+++ config/i386/t-linux64 (working copy)
@@ -19,3 +19,24 @@ EXTRA_MULTILIB_PARTS=crtbegin.o crtend.o
# because then __FRAME_END__ might not be the last thing in .eh_frame
# section.
CRTSTUFF_T_CFLAGS = -fno-omit-frame-pointer -fno-asynchronous-unwind-tables
+
+# Build decimal floating point support routines.
+
+D32PBIT = dfp32-bit.c
+D64PBIT = dfp64-bit.c
+D128PBIT = dfp128-bit.c
+
+dfp32-bit.c: $(srcdir)/config/dfp-bit.c
+ echo '#define WIDTH 32' > $@
+ echo '#define CMPtype DItype' >> $@
+ cat $< >> $@
+
+dfp64-bit.c: $(srcdir)/config/dfp-bit.c
+ echo '#define WIDTH 64' > $@
+ echo '#define CMPtype DItype' >> $@
+ cat $< >> $@
+
+dfp128-bit.c: $(srcdir)/config/dfp-bit.c
+ echo '#define WIDTH 128' > $@
+ echo '#define CMPtype DItype' >> $@
+ cat $< >> $@