1
# DP: Fix --with-long-double-128 for sparc32 when defaulting to 64-bit.
3
On sparc, the --with-long-double-128 option doesn't change anything for
4
a 64-bit compiler, as it always default to 128-bit long doubles. For
5
a 32/64-bit compiler defaulting to 32-bit this correctly control the
6
size of long double of the 32-bit compiler, however for a 32/64-bit
7
compiler defaulting to 64-bit, the built-in specs force the
8
-mlong-double-64 option. This makes the option useless in this case.
10
The patch below fixes that by removing the -mlong-double-64 from the
11
built-in spec, using the default instead.
15
2013-12-04 Aurelien Jarno <aurelien@aurel32.net>
17
* config/sparc/linux64.h (CC1_SPEC): When defaulting to 64-bit,
18
don't force -mlong-double-64 when -m32 or -mv8plus is given.
20
Index: b/src/gcc/config/sparc/linux64.h
21
===================================================================
22
--- a/src/gcc/config/sparc/linux64.h
23
+++ b/src/gcc/config/sparc/linux64.h
24
@@ -154,9 +154,9 @@ extern const char *host_detect_local_cpu
26
#define CC1_SPEC "%{profile:-p} \
27
%{m32:%{m64:%emay not use both -m32 and -m64}} \
28
-%{m32:-mptr32 -mno-stack-bias %{!mlong-double-128:-mlong-double-64} \
29
+%{m32:-mptr32 -mno-stack-bias \
30
%{!mcpu*:-mcpu=cypress}} \
31
-%{mv8plus:-mptr32 -mno-stack-bias %{!mlong-double-128:-mlong-double-64} \
32
+%{mv8plus:-mptr32 -mno-stack-bias \
34
%{!m32:%{!mcpu*:-mcpu=ultrasparc}} \
35
%{!mno-vis:%{!m32:%{!mcpu=v9:-mvis}}} \