summaryrefslogtreecommitdiff
path: root/test/MC/PowerPC/ppc64-encoding-fp.s
diff options
context:
space:
mode:
authorHal Finkel <hfinkel@anl.gov>2013-08-19 05:01:02 +0000
committerHal Finkel <hfinkel@anl.gov>2013-08-19 05:01:02 +0000
commit953a78084b85ea88cd2b208153a72df70e27133f (patch)
tree0d675bc7cfc49f17f2da670f7fab40f064f6b84b /test/MC/PowerPC/ppc64-encoding-fp.s
parenta9b697975592b204e86496efc62a8b26fbfa1a03 (diff)
downloadllvm-953a78084b85ea88cd2b208153a72df70e27133f.tar.gz
llvm-953a78084b85ea88cd2b208153a72df70e27133f.tar.bz2
llvm-953a78084b85ea88cd2b208153a72df70e27133f.tar.xz
Add the PPC fcpsgn instruction
Modern PPC cores support a floating-point copysign instruction, and we can use this to lower the FCOPYSIGN node (which is created from calls to the libm copysign function). A couple of extra patterns are necessary because the operand types of FCOPYSIGN need not agree. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188653 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/MC/PowerPC/ppc64-encoding-fp.s')
-rw-r--r--test/MC/PowerPC/ppc64-encoding-fp.s6
1 files changed, 4 insertions, 2 deletions
diff --git a/test/MC/PowerPC/ppc64-encoding-fp.s b/test/MC/PowerPC/ppc64-encoding-fp.s
index ae0e2866a2..e288dea89f 100644
--- a/test/MC/PowerPC/ppc64-encoding-fp.s
+++ b/test/MC/PowerPC/ppc64-encoding-fp.s
@@ -65,8 +65,10 @@
fnabs 2, 3
# CHECK: fnabs. 2, 3 # encoding: [0xfc,0x40,0x19,0x11]
fnabs. 2, 3
-# FIXME: fcpsgn 2, 3
-# FIXME: fcpsgn. 2, 3
+# CHECK: fcpsgn 2, 3, 4 # encoding: [0xfc,0x43,0x20,0x10]
+ fcpsgn 2, 3, 4
+# CHECK: fcpsgn. 2, 3, 4 # encoding: [0xfc,0x43,0x20,0x11]
+ fcpsgn. 2, 3, 4
# Floating-point arithmetic instructions