Hexagon: Add encoding bits to the mpy instructions.

llvm-svn: 209132

2 files changed, 582 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/Hexagon/intrinsics-mpy.ll b/llvm/test/CodeGen/Hexagon/intrinsics-mpy.ll
new file mode 100644
index 00000000000..74d18131930
--- /dev/null
+++ b/llvm/test/CodeGen/Hexagon/intrinsics-mpy.ll
@@ -0,0 +1,456 @@
+; RUN: llc -march=hexagon -mcpu=hexagonv5 < %s | FileCheck %s
+
+; Verify that the mpy intrinsics are lowered into the right instructions.
+
+@c = external global i32
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l)
+
+define void @test1(i32 %a1, i32 %b1) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.ll.s0(i32 %a1, i32 %b1)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.ll.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h)
+
+define void @test2(i32 %a2, i32 %b2) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.lh.s0(i32 %a2, i32 %b2)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.lh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l)
+
+define void @test3(i32 %a3, i32 %b3) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.hl.s0(i32 %a3, i32 %b3)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.hl.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h)
+
+define void @test4(i32 %a4, i32 %b4) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.hh.s0(i32 %a4, i32 %b4)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.hh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):sat
+
+define void @test5(i32 %a5, i32 %b5) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.ll.s0(i32 %a5, i32 %b5)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.ll.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):sat
+
+define void @test6(i32 %a6, i32 %b6) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.lh.s0(i32 %a6, i32 %b6)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.lh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):sat
+
+define void @test7(i32 %a7, i32 %b7) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.hl.s0(i32 %a7, i32 %b7)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.hl.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):sat
+
+define void @test8(i32 %a8, i32 %b8) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.hh.s0(i32 %a8, i32 %b8)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.hh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):rnd
+
+define void @test9(i32 %a9, i32 %b9) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.ll.s0(i32 %a9, i32 %b9)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.ll.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):rnd
+
+define void @test10(i32 %a10, i32 %b10) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.lh.s0(i32 %a10, i32 %b10)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.lh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):rnd
+
+define void @test11(i32 %a11, i32 %b11) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.hl.s0(i32 %a11, i32 %b11)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.hl.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):rnd
+
+define void @test12(i32 %a12, i32 %b12) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.hh.s0(i32 %a12, i32 %b12)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.hh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):rnd:sat
+
+define void @test13(i32 %a13, i32 %b13) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.ll.s0(i32 %a13, i32 %b13)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.ll.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):rnd:sat
+
+define void @test14(i32 %a14, i32 %b14) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.lh.s0(i32 %a14, i32 %b14)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.lh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):rnd:sat
+
+define void @test15(i32 %a15, i32 %b15) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.hl.s0(i32 %a15, i32 %b15)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.hl.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):rnd:sat
+
+define void @test16(i32 %a16, i32 %b16) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.hh.s0(i32 %a16, i32 %b16)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.hh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l)
+
+define void @test17(i32 %a17, i32 %b17) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.ll.s0(i32 %a17, i32 %b17)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.ll.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h)
+
+define void @test18(i32 %a18, i32 %b18) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.lh.s0(i32 %a18, i32 %b18)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.lh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l)
+
+define void @test19(i32 %a19, i32 %b19) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.hl.s0(i32 %a19, i32 %b19)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.hl.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h)
+
+define void @test20(i32 %a20, i32 %b20) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.hh.s0(i32 %a20, i32 %b20)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.hh.s0(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):<<1
+
+define void @test21(i32 %a21, i32 %b21) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.ll.s1(i32 %a21, i32 %b21)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.ll.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):<<1
+
+define void @test22(i32 %a22, i32 %b22) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.lh.s1(i32 %a22, i32 %b22)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.lh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):<<1
+
+define void @test23(i32 %a23, i32 %b23) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.hl.s1(i32 %a23, i32 %b23)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.hl.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):<<1
+
+define void @test24(i32 %a24, i32 %b24) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.hh.s1(i32 %a24, i32 %b24)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.hh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:sat
+
+define void @test25(i32 %a25, i32 %b25) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.ll.s1(i32 %a25, i32 %b25)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.ll.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:sat
+
+define void @test26(i32 %a26, i32 %b26) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.lh.s1(i32 %a26, i32 %b26)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.lh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:sat
+
+define void @test27(i32 %a27, i32 %b27) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.hl.s1(i32 %a27, i32 %b27)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.hl.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:sat
+
+define void @test28(i32 %a28, i32 %b28) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.hh.s1(i32 %a28, i32 %b28)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.hh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:rnd
+
+define void @test29(i32 %a29, i32 %b29) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.ll.s1(i32 %a29, i32 %b29)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.ll.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:rnd
+
+define void @test30(i32 %a30, i32 %b30) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.lh.s1(i32 %a30, i32 %b30)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.lh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:rnd
+
+define void @test31(i32 %a31, i32 %b31) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.hl.s1(i32 %a31, i32 %b31)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.hl.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:rnd
+
+define void @test32(i32 %a32, i32 %b32) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.rnd.hh.s1(i32 %a32, i32 %b32)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.rnd.hh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:rnd:sat
+
+define void @test33(i32 %a33, i32 %b33) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.ll.s1(i32 %a33, i32 %b33)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.ll.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:rnd:sat
+
+define void @test34(i32 %a34, i32 %b34) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.lh.s1(i32 %a34, i32 %b34)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.lh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):<<1:rnd:sat
+
+define void @test35(i32 %a35, i32 %b35) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.hl.s1(i32 %a35, i32 %b35)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.hl.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpy(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):<<1:rnd:sat
+
+define void @test36(i32 %a36, i32 %b36) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpy.sat.rnd.hh.s1(i32 %a36, i32 %b36)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpy.sat.rnd.hh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.l):<<1
+
+define void @test37(i32 %a37, i32 %b37) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.ll.s1(i32 %a37, i32 %b37)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.ll.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.l{{ *}},{{ *}}r{{[0-9]+}}.h):<<1
+
+define void @test38(i32 %a38, i32 %b38) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.lh.s1(i32 %a38, i32 %b38)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.lh.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.l):<<1
+
+define void @test39(i32 %a39, i32 %b39) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.hl.s1(i32 %a39, i32 %b39)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.hl.s1(i32, i32) #1
+
+; CHECK: r{{[0-9]+}}{{ *}}={{ *}}mpyu(r{{[0-9]+}}.h{{ *}},{{ *}}r{{[0-9]+}}.h):<<1
+
+define void @test40(i32 %a40, i32 %b40) #0 {
+entry:
+  %0 = tail call i32 @llvm.hexagon.M2.mpyu.hh.s1(i32 %a40, i32 %b40)
+  store i32 %0, i32* @c, align 4, !tbaa !1
+  ret void
+}
+
+declare i32 @llvm.hexagon.M2.mpyu.hh.s1(i32, i32) #1
+
+attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { nounwind readnone }
+
+!llvm.ident = !{!0}
+
+!0 = metadata !{metadata !"QuIC LLVM Hexagon Clang version 7.1-internal"}
+!1 = metadata !{metadata !2, metadata !2, i64 0}
+!2 = metadata !{metadata !"int", metadata !3, i64 0}
+!3 = metadata !{metadata !"omnipotent char", metadata !4, i64 0}
+!4 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/llvm/test/MC/Hexagon/mpy.s b/llvm/test/MC/Hexagon/mpy.s
new file mode 100644
index 00000000000..9e9706fc4c7
--- /dev/null
+++ b/llvm/test/MC/Hexagon/mpy.s
@@ -0,0 +1,126 @@
+#REQUIRES: object-emission
+#This test will be enabled when assembler support has been added.
+
+#RUN: llvm-mc -filetype=obj %s | llvm-objdump -d - | FileCheck %s
+
+# Check encoding bits for half-word multiply instructions.
+
+r7=mpy(r28.l,r20.h):<<1:rnd
+#CHECK: ecbcd427 { r7 = mpy(r28.l, r20.h):<<1:rnd }
+
+r18=mpy(r9.l,r21.h):rnd
+#CHECK: ec29d532 { r18 = mpy(r9.l, r21.h):rnd }
+
+r19=mpyu(r23.l,r20.l)
+#CHECK: ec57d413 { r19 = mpyu(r23.l, r20.l) }
+
+r22=mpyu(r19.l,r30.l):<<1
+#CHECK: ecd3de16 { r22 = mpyu(r19.l, r30.l):<<1 }
+
+r19=mpy(r16.h,r19.h)
+#CHECK: ec10d373 { r19 = mpy(r16.h, r19.h) }
+
+r30=mpy(r0.h,r16.h):<<1
+#CHECK: ec80d07e { r30 = mpy(r0.h, r16.h):<<1 }
+
+lr=mpy(r15.h,r25.l)
+#CHECK: ec0fd95f { r31 = mpy(r15.h, r25.l) }
+
+r29=mpy(r28.h,r15.l):<<1
+#CHECK: ec9ccf5d { r29 = mpy(r28.h, r15.l):<<1 }
+
+r20=mpy(r31.l,r19.h)
+#CHECK: ec1fd334 { r20 = mpy(r31.l, r19.h) }
+
+r24=mpy(r19.l,r15.h):<<1
+#CHECK: ec93cf38 { r24 = mpy(r19.l, r15.h):<<1 }
+
+r30=mpy(r10.l,sp.l)
+#CHECK: ec0add1e { r30 = mpy(r10.l, r29.l) }
+
+r7=mpy(r3.l,r4.l):<<1
+#CHECK: ec83c407 { r7 = mpy(r3.l, r4.l):<<1 }
+
+r30=mpy(r23.h,r2.h):rnd:sat
+#CHECK: ec37c2fe { r30 = mpy(r23.h, r2.h):rnd:sat }
+
+r5=mpy(r28.h,r27.h):<<1:rnd:sat
+#CHECK: ecbcdbe5 { r5 = mpy(r28.h, r27.h):<<1:rnd:sat }
+
+r26=mpy(r21.l,r23.l):rnd
+#CHECK: ec35d71a { r26 = mpy(r21.l, r23.l):rnd }
+
+sp=mpy(r25.h,r12.h):<<1:rnd
+#CHECK: ecb9cc7d { r29 = mpy(r25.h, r12.h):<<1:rnd }
+
+r1=mpy(r27.h,r29.h):rnd
+#CHECK: ec3bdd61 { r1 = mpy(r27.h, r29.h):rnd }
+
+r0=mpy(r2.h,r11.h):<<1:sat
+#CHECK: ec82cbe0 { r0 = mpy(r2.h, r11.h):<<1:sat }
+
+r3=mpy(r20.l,r30.l):rnd:sat
+#CHECK: ec34de83 { r3 = mpy(r20.l, r30.l):rnd:sat }
+
+r4=mpy(r21.h,r5.l):<<1:sat
+#CHECK: ec95c5c4 { r4 = mpy(r21.h, r5.l):<<1:sat }
+
+fp=mpy(r20.l,r12.h):rnd:sat
+#CHECK: ec34ccbe { r30 = mpy(r20.l, r12.h):rnd:sat }
+
+r12=mpy(sp.l,r30.h):<<1:rnd:sat
+#CHECK: ecbddeac { r12 = mpy(r29.l, r30.h):<<1:rnd:sat }
+
+r6=mpy(r10.h,fp.l):rnd:sat
+#CHECK: ec2adec6 { r6 = mpy(r10.h, r30.l):rnd:sat }
+
+r24=mpy(r12.h,r1.h):sat
+#CHECK: ec0cc1f8 { r24 = mpy(r12.h, r1.h):sat }
+
+r29=mpyu(r25.h,sp.l)
+#CHECK: ec59dd5d { r29 = mpyu(r25.h, r29.l) }
+
+r24=mpyu(lr.h,r29.l):<<1
+#CHECK: ecdfdd58 { r24 = mpyu(r31.h, r29.l):<<1 }
+
+r26=mpyu(r21.l,r18.h)
+#CHECK: ec55d23a { r26 = mpyu(r21.l, r18.h) }
+
+r29=mpyu(r4.l,r26.h):<<1
+#CHECK: ecc4da3d { r29 = mpyu(r4.l, r26.h):<<1 }
+
+fp=mpy(r8.l,r0.l):sat
+#CHECK: ec08c09e { r30 = mpy(r8.l, r0.l):sat }
+
+r1=mpy(r26.l,r16.l):<<1:sat
+#CHECK: ec9ad081 { r1 = mpy(r26.l, r16.l):<<1:sat }
+
+r16=mpyu(r26.h,r6.h)
+#CHECK: ec5ac670 { r16 = mpyu(r26.h, r6.h) }
+
+lr=mpyu(r23.h,r13.h):<<1
+#CHECK: ecd7cd7f { r31 = mpyu(r23.h, r13.h):<<1 }
+
+r14=mpy(r2.l,r7.h):sat
+#CHECK: ec02c7ae { r14 = mpy(r2.l, r7.h):sat }
+
+r9=mpy(r1.l,r9.h):<<1:sat
+#CHECK: ec81c9a9 { r9 = mpy(r1.l, r9.h):<<1:sat }
+
+r9=mpy(r30.l,r4.l):<<1:rnd:sat
+#CHECK: ecbec489 { r9 = mpy(r30.l, r4.l):<<1:rnd:sat }
+
+r9=mpy(r15.h,r27.l):<<1:rnd
+#CHECK: ecafdb49 { r9 = mpy(r15.h, r27.l):<<1:rnd }
+
+r16=mpy(r6.h,r16.l):rnd
+#CHECK: ec26d050 { r16 = mpy(r6.h, r16.l):rnd }
+
+r1=mpy(r10.l,r29.l):<<1:rnd
+#CHECK: ecaadd01 { r1 = mpy(r10.l, r29.l):<<1:rnd }
+
+r7=mpy(r4.h,r23.l):sat
+#CHECK: ec04d7c7 { r7 = mpy(r4.h, r23.l):sat }
+
+r17=mpy(r12.h,r26.l):<<1:rnd:sat
+#CHECK: ecacdad1 { r17 = mpy(r12.h, r26.l):<<1:rnd:sat }