[PATCH, applied] PR 70381, fix -mfloat128 setting

Michael Meissner meissner@linux.vnet.ibm.com
Thu Mar 24 00:04:00 GMT 2016


This patch undoes a change that I made that I did not intend to install in the
trunk.  On October 29th, 2015 I commited a major patch to enable the IEEE
128-bit floating point support (subversion id 229545).  I didn't realize I had
set OPTION_MASK_FLOAT128 in the ISA_2_6_MASKS_SERVER macro (which sets the
default options for power7).  I had only meant to set the mask in
POWERPC_MASKS, which would allow it to be set.

The upshot is -mfloat128 would be enabled for any power7, power8, or power9
system by default.  However, the IEEE 128-bit emulation functions are only
built on the Linux OS.  In addition, while the basic compiler and libgcc
support is there, more work needs to be done on libraries, and other
infrastructure before __float128 can be considered available in the GCC 7.0
time frame.

I built compilers with/without these patches on a Power8 system, and there were
no regressions.  Sorry about enabling it by default.

2016-03-23  Michael Meissner  <meissner@linux.vnet.ibm.com>

	PR target/70381
	* config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Do not set
	-mfloat128 here.

-- 
Michael Meissner, IBM
IBM, M/S 2506R, 550 King Street, Littleton, MA 01460-6245, USA
email: meissner@linux.vnet.ibm.com, phone: +1 (978) 899-4797
-------------- next part --------------
Index: gcc/config/rs6000/rs6000-cpus.def
===================================================================
--- gcc/config/rs6000/rs6000-cpus.def	(revision 234407)
+++ gcc/config/rs6000/rs6000-cpus.def	(working copy)
@@ -44,7 +44,6 @@
 #define ISA_2_6_MASKS_SERVER	(ISA_2_5_MASKS_SERVER			\
 				 | OPTION_MASK_POPCNTD			\
 				 | OPTION_MASK_ALTIVEC			\
-				 | OPTION_MASK_FLOAT128			\
 				 | OPTION_MASK_VSX			\
 				 | OPTION_MASK_UPPER_REGS_DF)
 


More information about the Gcc-patches mailing list