Merge current work back to tree to minimize diffs and drift. Major highlights
for CellSPU modifications:

- SPUInstrInfo.td refactoring: "multiclass" really is _your_ friend.
- Other improvements based on refactoring effort in SPUISelLowering.cpp,
  esp. in SPUISelLowering::PerformDAGCombine(), where zero amount shifts and
  rotates are now eliminiated, other scalar-to-vector-to-scalar silliness
  is also eliminated.
- 64-bit operations are being implemented, _muldi3.c gcc runtime now
  compiles and generates the right code. More work still needs to be done.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@47532 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/CellSPU/CellSDKIntrinsics.td b/lib/Target/CellSPU/CellSDKIntrinsics.td
index 2f453b1..5d759a4 100644
--- a/lib/Target/CellSPU/CellSDKIntrinsics.td
+++ b/lib/Target/CellSPU/CellSDKIntrinsics.td
@@ -210,20 +210,20 @@
       (SHLIv4i32 VECREG:$rA, uimm7:$val)>;
 
 def CellSDKshlqbi:
-  Pat<(int_spu_si_shlqbi VECREG:$rA, VECREG:$rB),
-      (SHLQBIvec VECREG:$rA, VECREG:$rB)>;
+  Pat<(int_spu_si_shlqbi VECREG:$rA, R32C:$rB),
+      (SHLQBIv16i8 VECREG:$rA, R32C:$rB)>;
 
 def CellSDKshlqii:
   Pat<(int_spu_si_shlqbii VECREG:$rA, uimm7:$val),
-      (SHLQBIIvec VECREG:$rA, uimm7:$val)>;
+      (SHLQBIIv16i8 VECREG:$rA, uimm7:$val)>;
 
 def CellSDKshlqby:
-  Pat<(int_spu_si_shlqby VECREG:$rA, VECREG:$rB),
-      (SHLQBYvec VECREG:$rA, VECREG:$rB)>;
+  Pat<(int_spu_si_shlqby VECREG:$rA, R32C:$rB),
+      (SHLQBYv16i8 VECREG:$rA, R32C:$rB)>;
 
 def CellSDKshlqbyi:
   Pat<(int_spu_si_shlqbyi VECREG:$rA, uimm7:$val),
-      (SHLQBYIvec VECREG:$rA, uimm7:$val)>;
+      (SHLQBYIv16i8 VECREG:$rA, uimm7:$val)>;
           
 //===----------------------------------------------------------------------===//
 // Branch/compare intrinsics: