Define patterns for shld and shrd that match immediate

shift counts, and patterns that match dynamic shift counts
when the subtract is obscured by a truncate node.

Add DAGCombiner support for recognizing rotate patterns
when the shift counts are defined by truncate nodes.

Fix and simplify the code for commuting shld and shrd
instructions to work even when the given instruction doesn't
have a parent, and when the caller needs a new instruction.

These changes allow LLVM to use the shld, shrd, rol, and ror
instructions on x86 to replace equivalent code using two
shifts and an or in many more cases.

llvm-svn: 57662

1 files changed, 73 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/rot32.ll b/llvm/test/CodeGen/X86/rot32.ll
new file mode 100644
index 00000000000..7cebcb86ce1
--- /dev/null
+++ b/llvm/test/CodeGen/X86/rot32.ll
@@ -0,0 +1,73 @@
+; RUN: llvm-as < %s | llc -march=x86 > %t
+; RUN: grep rol %t | count 3
+; RUN: grep ror %t | count 1
+; RUN: grep shld %t | count 2
+; RUN: grep shrd %t | count 2
+
+define i32 @foo(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = shl i32 %x, %z
+	%1 = sub i32 32, %z
+	%2 = lshr i32 %x, %1
+	%3 = or i32 %2, %0
+	ret i32 %3
+}
+
+define i32 @bar(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = shl i32 %y, %z
+	%1 = sub i32 32, %z
+	%2 = lshr i32 %x, %1
+	%3 = or i32 %2, %0
+	ret i32 %3
+}
+
+define i32 @un(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = lshr i32 %x, %z
+	%1 = sub i32 32, %z
+	%2 = shl i32 %x, %1
+	%3 = or i32 %2, %0
+	ret i32 %3
+}
+
+define i32 @bu(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = lshr i32 %y, %z
+	%1 = sub i32 32, %z
+	%2 = shl i32 %x, %1
+	%3 = or i32 %2, %0
+	ret i32 %3
+}
+
+define i32 @xfoo(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = lshr i32 %x, 25
+	%1 = shl i32 %x, 7
+	%2 = or i32 %0, %1
+	ret i32 %2
+}
+
+define i32 @xbar(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = shl i32 %y, 7
+	%1 = lshr i32 %x, 25
+	%2 = or i32 %0, %1
+	ret i32 %2
+}
+
+define i32 @xun(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = lshr i32 %x, 7
+	%1 = shl i32 %x, 25
+	%2 = or i32 %0, %1
+	ret i32 %2
+}
+
+define i32 @xbu(i32 %x, i32 %y, i32 %z) nounwind readnone {
+entry:
+	%0 = lshr i32 %y, 7
+	%1 = shl i32 %x, 25
+	%2 = or i32 %0, %1
+	ret i32 %2
+}