commit 7ba2725f5d3feb99b705de5e89b7c49b6687e5ed
author Craig Topper <craig.topper@gmail.com> Fri Dec 30 07:16:00 2011 +0000
committer Craig Topper <craig.topper@gmail.com> Fri Dec 30 07:16:00 2011 +0000
tree 6c08341fd2688b43b4223efb02cdf98412e9a45a
parent 06f554d06ab0f9390d04bcbaabb76f572d940249

Make FMA4 imply AVX so that YMM registers would be available. Necessitates removing from Bulldozer CPU types since it would enable AVX code generation implicitly. Also make SSE4A imply SSE3. Without some level of SSE implied, XMM registers wouldn't be legal.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147369 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/X86/X86.td

diff --git a/lib/Target/X86/X86.td b/lib/Target/X86/X86.td
index d053c76..6677936 100644
--- a/lib/Target/X86/X86.td
+++ b/lib/Target/X86/X86.td
@@ -77,7 +77,8 @@
                                         "IsUAMemFast", "true",
                                         "Fast unaligned memory access">;
 def FeatureSSE4A   : SubtargetFeature<"sse4a", "HasSSE4A", "true",
-                                      "Support SSE 4a instructions">;
+                                      "Support SSE 4a instructions",
+                                      [FeatureSSE3]>;
 
 def FeatureAVX     : SubtargetFeature<"avx", "HasAVX", "true",
                                       "Enable AVX instructions">;
@@ -90,8 +91,9 @@
                                       "Enable three-operand fused multiple-add",
                                       [FeatureAVX]>;
 def FeatureFMA4    : SubtargetFeature<"fma4", "HasFMA4", "true",
-                                      "Enable four-operand fused multiple-add">;
-def FeatureXOP    : SubtargetFeature<"xop", "HasXOP", "true",
+                                      "Enable four-operand fused multiple-add",
+                                      [FeatureAVX]>;
+def FeatureXOP     : SubtargetFeature<"xop", "HasXOP", "true",
                                       "Enable XOP instructions">;
 def FeatureVectorUAMem : SubtargetFeature<"vector-unaligned-mem",
                                           "HasVectorUAMem", "true",
@@ -201,12 +203,12 @@
 def : Proc<"amdfam10",        [FeatureSSE3,   FeatureSSE4A,
                                Feature3DNowA, FeatureCMPXCHG16B, FeatureLZCNT,
                                FeaturePOPCNT, FeatureSlowBTMem]>;
-// FIXME: Disabling AVX for now since it's not ready.
+// FIXME: Disabling AVX/FMA4 for now since it's not ready.
 def : Proc<"bdver1",          [FeatureSSE42, FeatureSSE4A, FeatureCMPXCHG16B,
-                               FeatureAES, FeatureCLMUL, FeatureFMA4,
+                               FeatureAES, FeatureCLMUL,
                                FeatureXOP, FeatureLZCNT, FeaturePOPCNT]>;
 def : Proc<"bdver2",          [FeatureSSE42, FeatureSSE4A, FeatureCMPXCHG16B,
-                               FeatureAES, FeatureCLMUL, FeatureFMA4,
+                               FeatureAES, FeatureCLMUL,
                                FeatureXOP, FeatureF16C, FeatureLZCNT,
                                FeaturePOPCNT, FeatureBMI]>;