[Sparc]: correct the 'set' synthetic instruction

[dougk]

changes as suggested, add moar tests.

llvm/trunk/lib/Target/Sparc/AsmParser/SparcAsmParser.cpp

void SparcAsmParser::expandSET(MCInst &Inst, SMLoc IDLoc,
                                SmallVectorImpl<MCInst> &Instructions) {
   MCOperand MCRegOp = Inst.getOperand(0);
   MCOperand MCValOp = Inst.getOperand(1);
   assert(MCRegOp.isReg());
   assert(MCValOp.isImm() || MCValOp.isExpr());
 
   // the imm operand can be either an expression or an immediate.
   bool IsImm = Inst.getOperand(1).isImm();
-  uint64_t ImmValue = IsImm ? MCValOp.getImm() : 0;
+  int64_t RawImmValue = IsImm ? MCValOp.getImm() : 0;
+
+  // Allow either a signed or unsigned 32-bit immediate.
+  if (RawImmValue < -2147483648 || RawImmValue > 4294967295) {
+    Error(IDLoc, "set: argument must be between -2147483648 and 4294967295");
+    return;
+  }
+
+  // If the value was expressed as a large unsigned number, that's ok.
+  // We want to see if it "looks like" a small signed number.
+  int32_t ImmValue = RawImmValue;
+  // For 'set' you can't use 'or' with a negative operand on V9 because
+  // that would splat the sign bit across the upper half of the destination
+  // register, whereas 'set' is defined to zero the high 32 bits.
+  bool IsEffectivelyImm13 =
+      IsImm && ((is64Bit() ? 0 : -4096) <= ImmValue && ImmValue < 4096);
   const MCExpr *ValExpr;
   if (IsImm)
     ValExpr = MCConstantExpr::create(ImmValue, getContext());
   else
     ValExpr = MCValOp.getExpr();
 
   MCOperand PrevReg = MCOperand::createReg(Sparc::G0);
 
-  if (!IsImm || (ImmValue & ~0x1fff)) {
+  // If not just a signed imm13 value, then either we use a 'sethi' with a
+  // following 'or', or a 'sethi' by itself if there are no more 1 bits.
+  // In either case, start with the 'sethi'.
+  if (!IsEffectivelyImm13) {
     MCInst TmpInst;
     const MCExpr *Expr =
         SparcMCExpr::create(SparcMCExpr::VK_Sparc_HI, ValExpr, getContext());
     TmpInst.setLoc(IDLoc);
     TmpInst.setOpcode(SP::SETHIi);
     TmpInst.addOperand(MCRegOp);
     TmpInst.addOperand(MCOperand::createExpr(Expr));
     Instructions.push_back(TmpInst);
     PrevReg = MCRegOp;
   }
 
-  if (!IsImm || ((ImmValue & 0x1fff) != 0 || ImmValue == 0)) {
+  // The low bits require touching in 3 cases:
+  // * A non-immediate value will always require both instructions.
+  // * An effectively imm13 value needs only an 'or' instruction.
+  // * Otherwise, an immediate that is not effectively imm13 requires the
+  //   'or' only if bits remain after clearing the 22 bits that 'sethi' set.
+  // If the low bits are known zeros, there's nothing to do.
+  // In the second case, and only in that case, must we NOT clear
+  // bits of the immediate value via the %lo() assembler function.
+  // Note also, the 'or' instruction doesn't mind a large value in the case
+  // where the operand to 'set' was 0xFFFFFzzz - it does exactly what you mean.
+  if (!IsImm || IsEffectivelyImm13 || (ImmValue & 0x3ff)) {
     MCInst TmpInst;
-    const MCExpr *Expr =
+    const MCExpr *Expr;
+    if (IsEffectivelyImm13)
+      Expr = ValExpr;
+    else
+      Expr =
           SparcMCExpr::create(SparcMCExpr::VK_Sparc_LO, ValExpr, getContext());
     TmpInst.setLoc(IDLoc);
     TmpInst.setOpcode(SP::ORri);
     TmpInst.addOperand(MCRegOp);
     TmpInst.addOperand(PrevReg);
     TmpInst.addOperand(MCOperand::createExpr(Expr));
     Instructions.push_back(TmpInst);
   }
 }

llvm/trunk/test/MC/Sparc/sparc-asm-errors.s

+! RUN: not llvm-mc %s -arch=sparc   -show-encoding 2>&1 | FileCheck %s
+! RUN: not llvm-mc %s -arch=sparcv9 -show-encoding 2>&1 | FileCheck %s
+
+! Test the lower and upper bounds of 'set'
+        ! CHECK: argument must be between
+        set -2147483649, %o1
+        ! CHECK: argument must be between
+        set 4294967296, %o1

llvm/trunk/test/MC/Sparc/sparc-synthetic-instructions.s

! Section A.3 Synthetic Instructions
         ! CHECK: save                             ! encoding: [0x81,0xe0,0x00,0x00]
         save
 
         ! CHECK: sethi %hi(40000), %g1            ! encoding: [0x03,0b00AAAAAA,A,A]
         ! CHECK:                                  !   fixup A - offset: 0, value: %hi(40000), kind: fixup_sparc_hi22
         ! CHECK: or %g1, %lo(40000), %g1          ! encoding: [0x82,0x10,0b011000AA,A]
         ! CHECK:                                  !   fixup A - offset: 0, value: %lo(40000), kind: fixup_sparc_lo10
         set 40000, %g1
-        ! CHECK: mov      %lo(1), %g1             ! encoding: [0x82,0x10,0b001000AA,A]
-        ! CHECK:                                  !   fixup A - offset: 0, value: %lo(1), kind: fixup_sparc_lo10
+        ! CHECK: mov 1, %g1 ! encoding: [0x82,0x10,0x20,0x01]
         set 1, %g1
         ! CHECK: sethi %hi(32768), %g1            ! encoding: [0x03,0b00AAAAAA,A,A]
         ! CHECK:                                  !   fixup A - offset: 0, value: %hi(32768), kind: fixup_sparc_hi22
         set 32768, %g1
 
+        ! Expect a 'sethi' without an 'or'.
+        ! CHECK: sethi %hi(268431360), %o1       ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! CHECK:                                 !   fixup A - offset: 0, value: %hi(268431360), kind: fixup_sparc_hi22
+        set 0x0ffff000, %o1
+
+        ! CHECK: sethi %hi(268433408), %o1       ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! CHECK:                                 !   fixup A - offset: 0, value: %hi(268433408), kind: fixup_sparc_hi22
+        set 0x0ffff800, %o1
+
+        ! This is the boundary case that uses the lowest of the 22 bits in sethi.
+        ! CHECK: sethi %hi(268434432), %o1       ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! CHECK:                                 !   fixup A - offset: 0, value: %hi(268434432), kind: fixup_sparc_hi22
+        set 0x0ffffc00, %o1
+
+        ! Now the synthetic instruction becomes two instructions.
+        ! CHECK: sethi %hi(2147483647), %o1      ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! CHECK:                                 !   fixup A - offset: 0, value: %hi(2147483647), kind: fixup_sparc_hi22
+        ! CHECK: or %o1, %lo(2147483647), %o1    ! encoding: [0x92,0x12,0b011000AA,A]
+        ! CHECK:                                 !   fixup A - offset: 0, value: %lo(2147483647), kind: fixup_sparc_lo10
+        set 2147483647, %o1
+
         ! CHECK: xnor %g1, %g0, %g2               ! encoding: [0x84,0x38,0x40,0x00]
         not %g1, %g2
         ! CHECK: xnor %g1, %g0, %g1               ! encoding: [0x82,0x38,0x40,0x00]
         not %g1
 
         ! CHECK: sub %g0, %g1, %g2                ! encoding: [0x84,0x20,0x00,0x01]
         neg %g1, %g2
         ! CHECK: sub %g0, %g1, %g1                ! encoding: [0x82,0x20,0x00,0x01]
! Other aliases
         ! CHECK: wr %g0, %i0, %wim                ! encoding: [0x81,0x90,0x00,0x18]
         wr %i0, %wim
         ! CHECK: wr %g0, 5, %wim                  ! encoding: [0x81,0x90,0x20,0x05]
         wr 5, %wim
         ! CHECK: wr %g0, %i0, %tbr                ! encoding: [0x81,0x98,0x00,0x18]
         wr %i0, %tbr
         ! CHECK: wr %g0, 5, %tbr                  ! encoding: [0x81,0x98,0x20,0x05]
         wr 5, %tbr
+
+! The following tests exercise 'set' in such a way that its output differs
+! depending on whether targeting V8 or V9.
+!
+! RUN: llvm-mc %s -arch=sparc   -show-encoding | FileCheck %s --check-prefix=V8
+! RUN: llvm-mc %s -arch=sparcv9 -show-encoding | FileCheck %s --check-prefix=V9
+
+        ! V8: mov        -1, %o1              ! encoding: [0x92,0x10,0x3f,0xff]
+        ! V9: sethi %hi(-1), %o1              ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-1), kind: fixup_sparc_hi22
+        ! V9: or %o1, %lo(-1), %o1            ! encoding: [0x92,0x12,0b011000AA,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %lo(-1), kind: fixup_sparc_lo10
+        set 0xffffffff, %o1
+
+        ! V8: mov        -2, %o1              ! encoding: [0x92,0x10,0x3f,0xfe]
+        ! V9: sethi %hi(-2), %o1              ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-2), kind: fixup_sparc_hi22
+        ! V9: or %o1, %lo(-2), %o1            ! encoding: [0x92,0x12,0b011000AA,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %lo(-2), kind: fixup_sparc_lo10
+        set 0xfffffffe, %o1
+
+        ! V8: mov        -16, %o1             ! encoding: [0x92,0x10,0x3f,0xf0]
+        ! V9: sethi %hi(-16), %o1             ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-16), kind: fixup_sparc_hi22
+        ! V9: or %o1, %lo(-16), %o1           ! encoding: [0x92,0x12,0b011000AA,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %lo(-16), kind: fixup_sparc_lo10
+        set 0xfffffff0, %o1
+
+        ! V8: mov        -256, %o1            ! encoding: [0x92,0x10,0x3f,0x00]
+        ! V9: sethi %hi(-256), %o1            ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-256), kind: fixup_sparc_hi22
+        ! V9: or %o1, %lo(-256), %o1          ! encoding: [0x92,0x12,0b011000AA,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %lo(-256), kind: fixup_sparc_lo10
+        set 0xffffff00, %o1
+
+        ! V8: mov        -4096, %o1           ! encoding: [0x92,0x10,0x30,0x00]
+        ! V9: sethi %hi(-4096), %o1           ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-4096), kind: fixup_sparc_hi22
+        set 0xfffff000, %o1
+
+        ! These results are the same for V8 and V9, so this test could have
+        ! been with the others that weren't segregated by architecture,
+        ! but logically it belongs here as a boundary case.
+        ! V8: sethi %hi(-8192), %o1           ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V8:                                 !   fixup A - offset: 0, value: %hi(-8192), kind: fixup_sparc_hi22
+        ! V9: sethi %hi(-8192), %o1           ! encoding: [0x13,0b00AAAAAA,A,A]
+        ! V9:                                 !   fixup A - offset: 0, value: %hi(-8192), kind: fixup_sparc_hi22
+        set 0xffffe000, %o1