PR11867, bignum operations

Alan Modra amodra@gmail.com
Tue Aug 3 10:53:00 GMT 2010


On Tue, Aug 03, 2010 at 07:07:51PM +0930, Alan Modra wrote:
> 	* gas/all/octa.s, * gas/all/octa.d: New test.
> 	* gas/all/gas.exp: Run octa.
> 	* gas/elf/data-1.s, * gas/elf/data-1.d: Delete.
> 	* gas/elf/elf.exp: Don't run data-1.

Committed with # comment lines in octa.s removed (tic30 doesn't like
them), and address matching relaxed in octa.d (tic4x has 4 octet
bytes).

The test fails on alpha-dec-vms and hppa-hp-hpux10 due to missing
entries in gas-defs.exp:get_standard_section_names for these targets,
and on pdp11-dec-aout due to '~' being allowed to start a symbol on
that target.  In any case, pdp11 wouldn't match the expected output,
so I probably should not run the test for pdp11.

There were two other interesting failures due to bugs.

	* config/tc-d10v.h (TARGET_BYTES_BIG_ENDIAN): Define as 1.
	* config/tc-m32c.c (md_number_to_chars): Call bigendian
	form of number_to_chars, not littleendian.

Index: gas/config/tc-d10v.h
===================================================================
RCS file: /cvs/src/src/gas/config/tc-d10v.h,v
retrieving revision 1.19
diff -u -p -r1.19 tc-d10v.h
--- gas/config/tc-d10v.h	10 Sep 2009 14:31:23 -0000	1.19
+++ gas/config/tc-d10v.h	3 Aug 2010 09:38:51 -0000
@@ -22,7 +22,7 @@
 
 #define TC_D10V
 
-#define TARGET_BYTES_BIG_ENDIAN 0
+#define TARGET_BYTES_BIG_ENDIAN 1
 
 /* The target BFD architecture.  */
 #define TARGET_ARCH bfd_arch_d10v
Index: gas/config/tc-m32c.c
===================================================================
RCS file: /cvs/src/src/gas/config/tc-m32c.c,v
retrieving revision 1.23
diff -u -p -r1.23 tc-m32c.c
--- gas/config/tc-m32c.c	28 Jun 2010 14:06:57 -0000	1.23
+++ gas/config/tc-m32c.c	3 Aug 2010 09:38:51 -0000
@@ -1146,7 +1146,7 @@ m32c_force_relocation (fixS * fixp)
 void
 md_number_to_chars (char * buf, valueT val, int n)
 {
-  number_to_chars_littleendian (buf, val, n);
+  number_to_chars_bigendian (buf, val, n);
 }
 
 /* Turn a string in input_line_pointer into a floating point constant of type


-- 
Alan Modra
Australia Development Lab, IBM



More information about the Binutils mailing list