From 3870b750e6d8af533926138e670f4643a5953e42 Mon Sep 17 00:00:00 2001
From: Jim Grosbach <grosbach@apple.com>
Date: Fri, 22 Oct 2010 18:35:16 +0000
Subject: [PATCH] More ARM multiply instuction binary encodings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117121 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Target/ARM/ARMInstrFormats.td |  29 ++++-
 lib/Target/ARM/ARMInstrInfo.td    | 182 ++++++++++++++----------------
 test/MC/ARM/simple-encoding.ll    |  30 +++++
 3 files changed, 141 insertions(+), 100 deletions(-)

diff --git a/lib/Target/ARM/ARMInstrFormats.td b/lib/Target/ARM/ARMInstrFormats.td
index 867571cbf4f..0194e2de363 100644
--- a/lib/Target/ARM/ARMInstrFormats.td
+++ b/lib/Target/ARM/ARMInstrFormats.td
@@ -934,15 +934,42 @@ class AMul2Ia<bits<7> opcod, bits<4> opc7_4, dag oops, dag iops,
 }
 
 // SMUL<x><y> / SMULW<y> / SMLA<x><y> / SMLAW<x><y>
-class AMulxyI<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops,
+class AMulxyIbase<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops,
               InstrItinClass itin, string opc, string asm, list<dag> pattern>
   : I<oops, iops, AddrModeNone, Size4Bytes, IndexModeNone, MulFrm, itin,
       opc, asm, "", pattern> {
+  bits<4> Rn;
+  bits<4> Rm;
   let Inst{4}     = 0;
   let Inst{7}     = 1;
   let Inst{20}    = 0;
   let Inst{27-21} = opcod;
   let Inst{6-5}   = bit6_5;
+  let Inst{11-8}  = Rm;
+  let Inst{3-0}   = Rn;
+}
+class AMulxyI<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops,
+              InstrItinClass itin, string opc, string asm, list<dag> pattern>
+  : AMulxyIbase<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> {
+  bits<4> Rd;
+  let Inst{19-16} = Rd;
+}
+
+// AMulxyI with Ra operand
+class AMulxyIa<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops,
+              InstrItinClass itin, string opc, string asm, list<dag> pattern>
+  : AMulxyI<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> {
+  bits<4> Ra;
+  let Inst{15-12} = Ra;
+}
+// SMLAL*
+class AMulxyI64<bits<7> opcod, bits<2> bit6_5, dag oops, dag iops,
+              InstrItinClass itin, string opc, string asm, list<dag> pattern>
+  : AMulxyIbase<opcod, bit6_5, oops, iops, itin, opc, asm, pattern> {
+  bits<4> RdLo;
+  bits<4> RdHi;
+  let Inst{19-16} = RdHi;
+  let Inst{15-12} = RdLo;
 }
 
 // Extend instructions.
diff --git a/lib/Target/ARM/ARMInstrInfo.td b/lib/Target/ARM/ARMInstrInfo.td
index ee4fee3bd0e..4fc2adda046 100644
--- a/lib/Target/ARM/ARMInstrInfo.td
+++ b/lib/Target/ARM/ARMInstrInfo.td
@@ -2372,132 +2372,116 @@ def SMMLSR : AMul2Ia <0b0111010, 0b1111, (outs GPR:$Rd),
             Requires<[IsARM, HasV6]>;
 
 multiclass AI_smul<string opc, PatFrag opnode> {
-  def BB : AMulxyI<0b0001011, 0b00, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "bb"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (opnode (sext_inreg GPR:$a, i16),
-                                      (sext_inreg GPR:$b, i16)))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def BB : AMulxyI<0b0001011, 0b00, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "bb"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (opnode (sext_inreg GPR:$Rn, i16),
+                                      (sext_inreg GPR:$Rm, i16)))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def BT : AMulxyI<0b0001011, 0b10, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "bt"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (opnode (sext_inreg GPR:$a, i16),
-                                      (sra GPR:$b, (i32 16))))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def BT : AMulxyI<0b0001011, 0b10, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "bt"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (opnode (sext_inreg GPR:$Rn, i16),
+                                      (sra GPR:$Rm, (i32 16))))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def TB : AMulxyI<0b0001011, 0b01, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "tb"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (opnode (sra GPR:$a, (i32 16)),
-                                      (sext_inreg GPR:$b, i16)))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def TB : AMulxyI<0b0001011, 0b01, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "tb"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (opnode (sra GPR:$Rn, (i32 16)),
+                                      (sext_inreg GPR:$Rm, i16)))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def TT : AMulxyI<0b0001011, 0b11, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "tt"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (opnode (sra GPR:$a, (i32 16)),
-                                      (sra GPR:$b, (i32 16))))]>,
-            Requires<[IsARM, HasV5TE]> {
-           }
+  def TT : AMulxyI<0b0001011, 0b11, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "tt"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (opnode (sra GPR:$Rn, (i32 16)),
+                                      (sra GPR:$Rm, (i32 16))))]>,
+            Requires<[IsARM, HasV5TE]>;
 
-  def WB : AMulxyI<0b0001001, 0b01, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "wb"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (sra (opnode GPR:$a,
-                                    (sext_inreg GPR:$b, i16)), (i32 16)))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def WB : AMulxyI<0b0001001, 0b01, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "wb"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (sra (opnode GPR:$Rn,
+                                    (sext_inreg GPR:$Rm, i16)), (i32 16)))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def WT : AMulxyI<0b0001001, 0b11, (outs GPR:$dst), (ins GPR:$a, GPR:$b),
-              IIC_iMUL16, !strconcat(opc, "wt"), "\t$dst, $a, $b",
-              [(set GPR:$dst, (sra (opnode GPR:$a,
-                                    (sra GPR:$b, (i32 16))), (i32 16)))]>,
-            Requires<[IsARM, HasV5TE]> {
-           }
+  def WT : AMulxyI<0b0001001, 0b11, (outs GPR:$Rd), (ins GPR:$Rn, GPR:$Rm),
+              IIC_iMUL16, !strconcat(opc, "wt"), "\t$Rd, $Rn, $Rm",
+              [(set GPR:$Rd, (sra (opnode GPR:$Rn,
+                                    (sra GPR:$Rm, (i32 16))), (i32 16)))]>,
+            Requires<[IsARM, HasV5TE]>;
 }
 
 
 multiclass AI_smla<string opc, PatFrag opnode> {
-  def BB : AMulxyI<0b0001000, 0b00, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "bb"), "\t$dst, $a, $b, $acc",
-              [(set GPR:$dst, (add GPR:$acc,
-                               (opnode (sext_inreg GPR:$a, i16),
-                                       (sext_inreg GPR:$b, i16))))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def BB : AMulxyI<0b0001000, 0b00, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "bb"), "\t$Rd, $Rn, $Rm, $Ra",
+              [(set GPR:$Rd, (add GPR:$Ra,
+                               (opnode (sext_inreg GPR:$Rn, i16),
+                                       (sext_inreg GPR:$Rm, i16))))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def BT : AMulxyI<0b0001000, 0b10, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "bt"), "\t$dst, $a, $b, $acc",
-              [(set GPR:$dst, (add GPR:$acc, (opnode (sext_inreg GPR:$a, i16),
-                                                    (sra GPR:$b, (i32 16)))))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def BT : AMulxyI<0b0001000, 0b10, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "bt"), "\t$Rd, $Rn, $Rm, $Ra",
+              [(set GPR:$Rd, (add GPR:$Ra, (opnode (sext_inreg GPR:$Rn, i16),
+                                                   (sra GPR:$Rm, (i32 16)))))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def TB : AMulxyI<0b0001000, 0b01, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "tb"), "\t$dst, $a, $b, $acc",
-              [(set GPR:$dst, (add GPR:$acc, (opnode (sra GPR:$a, (i32 16)),
-                                                 (sext_inreg GPR:$b, i16))))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def TB : AMulxyI<0b0001000, 0b01, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "tb"), "\t$Rd, $Rn, $Rm, $Ra",
+              [(set GPR:$Rd, (add GPR:$Ra, (opnode (sra GPR:$Rn, (i32 16)),
+                                                (sext_inreg GPR:$Rm, i16))))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def TT : AMulxyI<0b0001000, 0b11, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "tt"), "\t$dst, $a, $b, $acc",
-             [(set GPR:$dst, (add GPR:$acc, (opnode (sra GPR:$a, (i32 16)),
-                                                    (sra GPR:$b, (i32 16)))))]>,
-            Requires<[IsARM, HasV5TE]> {
-           }
+  def TT : AMulxyI<0b0001000, 0b11, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "tt"), "\t$Rd, $Rn, $Rm, $Ra",
+             [(set GPR:$Rd, (add GPR:$Ra, (opnode (sra GPR:$Rn, (i32 16)),
+                                                   (sra GPR:$Rm, (i32 16)))))]>,
+            Requires<[IsARM, HasV5TE]>;
 
-  def WB : AMulxyI<0b0001001, 0b00, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "wb"), "\t$dst, $a, $b, $acc",
-              [(set GPR:$dst, (add GPR:$acc, (sra (opnode GPR:$a,
-                                       (sext_inreg GPR:$b, i16)), (i32 16))))]>,
-           Requires<[IsARM, HasV5TE]> {
-           }
+  def WB : AMulxyI<0b0001001, 0b00, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "wb"), "\t$Rd, $Rn, $Rm, $Ra",
+              [(set GPR:$Rd, (add GPR:$Ra, (sra (opnode GPR:$Rn,
+                                      (sext_inreg GPR:$Rm, i16)), (i32 16))))]>,
+           Requires<[IsARM, HasV5TE]>;
 
-  def WT : AMulxyI<0b0001001, 0b10, (outs GPR:$dst),
-              (ins GPR:$a, GPR:$b, GPR:$acc),
-              IIC_iMAC16, !strconcat(opc, "wt"), "\t$dst, $a, $b, $acc",
-              [(set GPR:$dst, (add GPR:$acc, (sra (opnode GPR:$a,
-                                         (sra GPR:$b, (i32 16))), (i32 16))))]>,
-            Requires<[IsARM, HasV5TE]> {
-           }
+  def WT : AMulxyI<0b0001001, 0b10, (outs GPR:$Rd),
+              (ins GPR:$Rn, GPR:$Rm, GPR:$Ra),
+              IIC_iMAC16, !strconcat(opc, "wt"), "\t$Rd, $Rn, $Rm, $Ra",
+              [(set GPR:$Rd, (add GPR:$Ra, (sra (opnode GPR:$Rn,
+                                        (sra GPR:$Rm, (i32 16))), (i32 16))))]>,
+            Requires<[IsARM, HasV5TE]>;
 }
 
 defm SMUL : AI_smul<"smul", BinOpFrag<(mul node:$LHS, node:$RHS)>>;
 defm SMLA : AI_smla<"smla", BinOpFrag<(mul node:$LHS, node:$RHS)>>;
 
 // Halfword multiply accumulate long: SMLAL<x><y> -- for disassembly only
-def SMLALBB : AMulxyI<0b0001010, 0b00, (outs GPR:$ldst, GPR:$hdst),
-                      (ins GPR:$a, GPR:$b),
-                      IIC_iMAC64, "smlalbb", "\t$ldst, $hdst, $a, $b",
+def SMLALBB : AMulxyI64<0b0001010, 0b00, (outs GPR:$RdLo, GPR:$RdHi),
+                      (ins GPR:$Rn, GPR:$Rm),
+                      IIC_iMAC64, "smlalbb", "\t$RdLo, $RdHi, $Rn, $Rm",
                       [/* For disassembly only; pattern left blank */]>,
-              Requires<[IsARM, HasV5TE]> {
-}
+              Requires<[IsARM, HasV5TE]>;
 
-def SMLALBT : AMulxyI<0b0001010, 0b10, (outs GPR:$ldst, GPR:$hdst),
-                      (ins GPR:$a, GPR:$b),
-                      IIC_iMAC64, "smlalbt", "\t$ldst, $hdst, $a, $b",
+def SMLALBT : AMulxyI64<0b0001010, 0b10, (outs GPR:$RdLo, GPR:$RdHi),
+                      (ins GPR:$Rn, GPR:$Rm),
+                      IIC_iMAC64, "smlalbt", "\t$RdLo, $RdHi, $Rn, $Rm",
                       [/* For disassembly only; pattern left blank */]>,
-              Requires<[IsARM, HasV5TE]> {
-}
+              Requires<[IsARM, HasV5TE]>;
 
-def SMLALTB : AMulxyI<0b0001010, 0b01, (outs GPR:$ldst, GPR:$hdst),
-                      (ins GPR:$a, GPR:$b),
-                      IIC_iMAC64, "smlaltb", "\t$ldst, $hdst, $a, $b",
+def SMLALTB : AMulxyI64<0b0001010, 0b01, (outs GPR:$RdLo, GPR:$RdHi),
+                      (ins GPR:$Rn, GPR:$Rm),
+                      IIC_iMAC64, "smlaltb", "\t$RdLo, $RdHi, $Rn, $Rm",
                       [/* For disassembly only; pattern left blank */]>,
-              Requires<[IsARM, HasV5TE]> {
-}
+              Requires<[IsARM, HasV5TE]>;
 
-def SMLALTT : AMulxyI<0b0001010, 0b11, (outs GPR:$ldst, GPR:$hdst),
-                      (ins GPR:$a, GPR:$b),
-                      IIC_iMAC64, "smlaltt", "\t$ldst, $hdst, $a, $b",
+def SMLALTT : AMulxyI64<0b0001010, 0b11, (outs GPR:$RdLo, GPR:$RdHi),
+                      (ins GPR:$Rn, GPR:$Rm),
+                      IIC_iMAC64, "smlaltt", "\t$RdLo, $RdHi, $Rn, $Rm",
                       [/* For disassembly only; pattern left blank */]>,
-              Requires<[IsARM, HasV5TE]> {
-}
+              Requires<[IsARM, HasV5TE]>;
 
 // Helper class for AI_smld -- for disassembly only
 class AMulDualI<bit long, bit sub, bit swap, dag oops, dag iops,
diff --git a/test/MC/ARM/simple-encoding.ll b/test/MC/ARM/simple-encoding.ll
index 970e0e616c4..f2c0c7c2bd9 100644
--- a/test/MC/ARM/simple-encoding.ll
+++ b/test/MC/ARM/simple-encoding.ll
@@ -150,4 +150,34 @@ define i32 @f15(i32 %x, i32 %y) {
         %tmp3.upgrd.2 = trunc i64 %tmp3 to i32
         ret i32 %tmp3.upgrd.2
 }
+
+define i32 @f16(i16 %x, i32 %y) {
+; CHECK: f16:
+; CHECK: smulbt r0, r0, r1            @ encoding: [0xc0,0x01,0x60,0xe1]
+        %tmp1 = add i16 %x, 2
+        %tmp2 = sext i16 %tmp1 to i32
+        %tmp3 = ashr i32 %y, 16
+        %tmp4 = mul i32 %tmp2, %tmp3
+        ret i32 %tmp4
+}
+
+define i32 @f17(i32 %x, i32 %y) {
+; CHECK: f17:
+; CHECK: smultt r0, r1, r0            @ encoding: [0xe1,0x00,0x60,0xe1]
+        %tmp1 = ashr i32 %x, 16
+        %tmp3 = ashr i32 %y, 16
+        %tmp4 = mul i32 %tmp3, %tmp1
+        ret i32 %tmp4
+}
+
+define i32 @f18(i32 %a, i16 %x, i32 %y) {
+; CHECK: f18:
+; CHECK: smlabt r0, r1, r2, r0        @ encoding: [0xc1,0x02,0x00,0xe1]
+        %tmp = sext i16 %x to i32
+        %tmp2 = ashr i32 %y, 16
+        %tmp3 = mul i32 %tmp2, %tmp
+        %tmp5 = add i32 %tmp3, %a
+        ret i32 %tmp5
+}
+
 declare void @llvm.trap() nounwind