llvm-6502/lib/Target/PowerPC/PPCInstr64Bit.td
Hal Finkel 4e703f82f2 [PowerPC] Add a DAGToDAG peephole to remove unnecessary zero-exts
On PPC64, we end up with lots of i32 -> i64 zero extensions, not only from all
of the usual places, but also from the ABI, which specifies that values passed
are zero extended. Almost all 32-bit PPC instructions in PPC64 mode are defined
to do *something* to the higher-order bits, and for some instructions, that
action clears those bits (thus providing a zero-extended result). This is
especially common after rotate-and-mask instructions. Adding an additional
instruction to zero-extend the results of these instructions is unnecessary.

This PPCISelDAGToDAG peephole optimization examines these zero-extensions, and
looks back through their operands to see if all instructions will implicitly
zero extend their results. If so, we convert these instructions to their 64-bit
variants (which is an internal change only, the actual encoding of these
instructions is the same as the original 32-bit ones) and remove the
unnecessary zero-extension (changing where the INSERT_SUBREG instructions are
to make everything internally consistent).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224169 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-12 23:59:36 +00:00

1149 lines
52 KiB
TableGen

//===-- PPCInstr64Bit.td - The PowerPC 64-bit Support ------*- tablegen -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file describes the PowerPC 64-bit instructions. These patterns are used
// both when in ppc64 mode and when in "use 64-bit extensions in 32-bit" mode.
//
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// 64-bit operands.
//
def s16imm64 : Operand<i64> {
let PrintMethod = "printS16ImmOperand";
let EncoderMethod = "getImm16Encoding";
let ParserMatchClass = PPCS16ImmAsmOperand;
let DecoderMethod = "decodeSImmOperand<16>";
}
def u16imm64 : Operand<i64> {
let PrintMethod = "printU16ImmOperand";
let EncoderMethod = "getImm16Encoding";
let ParserMatchClass = PPCU16ImmAsmOperand;
let DecoderMethod = "decodeUImmOperand<16>";
}
def s17imm64 : Operand<i64> {
// This operand type is used for addis/lis to allow the assembler parser
// to accept immediates in the range -65536..65535 for compatibility with
// the GNU assembler. The operand is treated as 16-bit otherwise.
let PrintMethod = "printS16ImmOperand";
let EncoderMethod = "getImm16Encoding";
let ParserMatchClass = PPCS17ImmAsmOperand;
let DecoderMethod = "decodeSImmOperand<16>";
}
def tocentry : Operand<iPTR> {
let MIOperandInfo = (ops i64imm:$imm);
}
def tlsreg : Operand<i64> {
let EncoderMethod = "getTLSRegEncoding";
let ParserMatchClass = PPCTLSRegOperand;
}
def tlsgd : Operand<i64> {}
def tlscall : Operand<i64> {
let PrintMethod = "printTLSCall";
let MIOperandInfo = (ops calltarget:$func, tlsgd:$sym);
let EncoderMethod = "getTLSCallEncoding";
}
//===----------------------------------------------------------------------===//
// 64-bit transformation functions.
//
def SHL64 : SDNodeXForm<imm, [{
// Transformation function: 63 - imm
return getI32Imm(63 - N->getZExtValue());
}]>;
def SRL64 : SDNodeXForm<imm, [{
// Transformation function: 64 - imm
return N->getZExtValue() ? getI32Imm(64 - N->getZExtValue()) : getI32Imm(0);
}]>;
def HI32_48 : SDNodeXForm<imm, [{
// Transformation function: shift the immediate value down into the low bits.
return getI32Imm((unsigned short)(N->getZExtValue() >> 32));
}]>;
def HI48_64 : SDNodeXForm<imm, [{
// Transformation function: shift the immediate value down into the low bits.
return getI32Imm((unsigned short)(N->getZExtValue() >> 48));
}]>;
//===----------------------------------------------------------------------===//
// Calls.
//
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
let isTerminator = 1, isBarrier = 1, PPC970_Unit = 7 in {
let isBranch = 1, isIndirectBranch = 1, Uses = [CTR8] in {
def BCTR8 : XLForm_2_ext<19, 528, 20, 0, 0, (outs), (ins), "bctr", IIC_BrB,
[]>,
Requires<[In64BitMode]>;
def BCCCTR8 : XLForm_2_br<19, 528, 0, (outs), (ins pred:$cond),
"b${cond:cc}ctr${cond:pm} ${cond:reg}", IIC_BrB,
[]>,
Requires<[In64BitMode]>;
def BCCTR8 : XLForm_2_br2<19, 528, 12, 0, (outs), (ins crbitrc:$bi),
"bcctr 12, $bi, 0", IIC_BrB, []>,
Requires<[In64BitMode]>;
def BCCTR8n : XLForm_2_br2<19, 528, 4, 0, (outs), (ins crbitrc:$bi),
"bcctr 4, $bi, 0", IIC_BrB, []>,
Requires<[In64BitMode]>;
}
}
let Defs = [LR8] in
def MovePCtoLR8 : Pseudo<(outs), (ins), "#MovePCtoLR8", []>,
PPC970_Unit_BRU;
let isBranch = 1, isTerminator = 1, hasCtrlDep = 1, PPC970_Unit = 7 in {
let Defs = [CTR8], Uses = [CTR8] in {
def BDZ8 : BForm_1<16, 18, 0, 0, (outs), (ins condbrtarget:$dst),
"bdz $dst">;
def BDNZ8 : BForm_1<16, 16, 0, 0, (outs), (ins condbrtarget:$dst),
"bdnz $dst">;
}
let isReturn = 1, Defs = [CTR8], Uses = [CTR8, LR8, RM] in {
def BDZLR8 : XLForm_2_ext<19, 16, 18, 0, 0, (outs), (ins),
"bdzlr", IIC_BrB, []>;
def BDNZLR8 : XLForm_2_ext<19, 16, 16, 0, 0, (outs), (ins),
"bdnzlr", IIC_BrB, []>;
}
}
let isCall = 1, PPC970_Unit = 7, Defs = [LR8] in {
// Convenient aliases for call instructions
let Uses = [RM] in {
def BL8 : IForm<18, 0, 1, (outs), (ins calltarget:$func),
"bl $func", IIC_BrB, []>; // See Pat patterns below.
def BL8_TLS : IForm<18, 0, 1, (outs), (ins tlscall:$func),
"bl $func", IIC_BrB, []>;
def BLA8 : IForm<18, 1, 1, (outs), (ins abscalltarget:$func),
"bla $func", IIC_BrB, [(PPCcall (i64 imm:$func))]>;
}
let Uses = [RM], isCodeGenOnly = 1 in {
def BL8_NOP : IForm_and_DForm_4_zero<18, 0, 1, 24,
(outs), (ins calltarget:$func),
"bl $func\n\tnop", IIC_BrB, []>;
def BL8_NOP_TLS : IForm_and_DForm_4_zero<18, 0, 1, 24,
(outs), (ins tlscall:$func),
"bl $func\n\tnop", IIC_BrB, []>;
def BLA8_NOP : IForm_and_DForm_4_zero<18, 1, 1, 24,
(outs), (ins abscalltarget:$func),
"bla $func\n\tnop", IIC_BrB,
[(PPCcall_nop (i64 imm:$func))]>;
}
let Uses = [CTR8, RM] in {
def BCTRL8 : XLForm_2_ext<19, 528, 20, 0, 1, (outs), (ins),
"bctrl", IIC_BrB, [(PPCbctrl)]>,
Requires<[In64BitMode]>;
let isCodeGenOnly = 1 in {
def BCCCTRL8 : XLForm_2_br<19, 528, 1, (outs), (ins pred:$cond),
"b${cond:cc}ctrl${cond:pm} ${cond:reg}", IIC_BrB,
[]>,
Requires<[In64BitMode]>;
def BCCTRL8 : XLForm_2_br2<19, 528, 12, 1, (outs), (ins crbitrc:$bi),
"bcctrl 12, $bi, 0", IIC_BrB, []>,
Requires<[In64BitMode]>;
def BCCTRL8n : XLForm_2_br2<19, 528, 4, 1, (outs), (ins crbitrc:$bi),
"bcctrl 4, $bi, 0", IIC_BrB, []>,
Requires<[In64BitMode]>;
}
}
}
} // Interpretation64Bit
// FIXME: Duplicating this for the asm parser should be unnecessary, but the
// previous definition must be marked as CodeGen only to prevent decoding
// conflicts.
let Interpretation64Bit = 1, isAsmParserOnly = 1 in
let isCall = 1, PPC970_Unit = 7, Defs = [LR8], Uses = [RM] in
def BL8_TLS_ : IForm<18, 0, 1, (outs), (ins tlscall:$func),
"bl $func", IIC_BrB, []>;
// Calls
def : Pat<(PPCcall (i64 tglobaladdr:$dst)),
(BL8 tglobaladdr:$dst)>;
def : Pat<(PPCcall_nop (i64 tglobaladdr:$dst)),
(BL8_NOP tglobaladdr:$dst)>;
def : Pat<(PPCcall (i64 texternalsym:$dst)),
(BL8 texternalsym:$dst)>;
def : Pat<(PPCcall_nop (i64 texternalsym:$dst)),
(BL8_NOP texternalsym:$dst)>;
def : Pat<(PPCcall_nop_tls texternalsym:$func, tglobaltlsaddr:$sym),
(BL8_NOP_TLS texternalsym:$func, tglobaltlsaddr:$sym)>;
// Atomic operations
let usesCustomInserter = 1 in {
let Defs = [CR0] in {
def ATOMIC_LOAD_ADD_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_ADD_I64",
[(set i64:$dst, (atomic_load_add_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_LOAD_SUB_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_SUB_I64",
[(set i64:$dst, (atomic_load_sub_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_LOAD_OR_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_OR_I64",
[(set i64:$dst, (atomic_load_or_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_LOAD_XOR_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_XOR_I64",
[(set i64:$dst, (atomic_load_xor_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_LOAD_AND_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_AND_i64",
[(set i64:$dst, (atomic_load_and_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_LOAD_NAND_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$incr), "#ATOMIC_LOAD_NAND_I64",
[(set i64:$dst, (atomic_load_nand_64 xoaddr:$ptr, i64:$incr))]>;
def ATOMIC_CMP_SWAP_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$old, g8rc:$new), "#ATOMIC_CMP_SWAP_I64",
[(set i64:$dst, (atomic_cmp_swap_64 xoaddr:$ptr, i64:$old, i64:$new))]>;
def ATOMIC_SWAP_I64 : Pseudo<
(outs g8rc:$dst), (ins memrr:$ptr, g8rc:$new), "#ATOMIC_SWAP_I64",
[(set i64:$dst, (atomic_swap_64 xoaddr:$ptr, i64:$new))]>;
}
}
// Instructions to support atomic operations
def LDARX : XForm_1<31, 84, (outs g8rc:$rD), (ins memrr:$ptr),
"ldarx $rD, $ptr", IIC_LdStLDARX,
[(set i64:$rD, (PPClarx xoaddr:$ptr))]>;
let Defs = [CR0] in
def STDCX : XForm_1<31, 214, (outs), (ins g8rc:$rS, memrr:$dst),
"stdcx. $rS, $dst", IIC_LdStSTDCX,
[(PPCstcx i64:$rS, xoaddr:$dst)]>,
isDOT;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [RM] in
def TCRETURNdi8 :Pseudo< (outs),
(ins calltarget:$dst, i32imm:$offset),
"#TC_RETURNd8 $dst $offset",
[]>;
let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [RM] in
def TCRETURNai8 :Pseudo<(outs), (ins abscalltarget:$func, i32imm:$offset),
"#TC_RETURNa8 $func $offset",
[(PPCtc_return (i64 imm:$func), imm:$offset)]>;
let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1, Uses = [RM] in
def TCRETURNri8 : Pseudo<(outs), (ins CTRRC8:$dst, i32imm:$offset),
"#TC_RETURNr8 $dst $offset",
[]>;
let isTerminator = 1, isBarrier = 1, PPC970_Unit = 7, isBranch = 1,
isIndirectBranch = 1, isCall = 1, isReturn = 1, Uses = [CTR8, RM] in
def TAILBCTR8 : XLForm_2_ext<19, 528, 20, 0, 0, (outs), (ins), "bctr", IIC_BrB,
[]>,
Requires<[In64BitMode]>;
let isBranch = 1, isTerminator = 1, hasCtrlDep = 1, PPC970_Unit = 7,
isBarrier = 1, isCall = 1, isReturn = 1, Uses = [RM] in
def TAILB8 : IForm<18, 0, 0, (outs), (ins calltarget:$dst),
"b $dst", IIC_BrB,
[]>;
let isBranch = 1, isTerminator = 1, hasCtrlDep = 1, PPC970_Unit = 7,
isBarrier = 1, isCall = 1, isReturn = 1, Uses = [RM] in
def TAILBA8 : IForm<18, 0, 0, (outs), (ins abscalltarget:$dst),
"ba $dst", IIC_BrB,
[]>;
} // Interpretation64Bit
def : Pat<(PPCtc_return (i64 tglobaladdr:$dst), imm:$imm),
(TCRETURNdi8 tglobaladdr:$dst, imm:$imm)>;
def : Pat<(PPCtc_return (i64 texternalsym:$dst), imm:$imm),
(TCRETURNdi8 texternalsym:$dst, imm:$imm)>;
def : Pat<(PPCtc_return CTRRC8:$dst, imm:$imm),
(TCRETURNri8 CTRRC8:$dst, imm:$imm)>;
// 64-bit CR instructions
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
let hasSideEffects = 0 in {
def MTOCRF8: XFXForm_5a<31, 144, (outs crbitm:$FXM), (ins g8rc:$ST),
"mtocrf $FXM, $ST", IIC_BrMCRX>,
PPC970_DGroup_First, PPC970_Unit_CRU;
def MTCRF8 : XFXForm_5<31, 144, (outs), (ins i32imm:$FXM, g8rc:$rS),
"mtcrf $FXM, $rS", IIC_BrMCRX>,
PPC970_MicroCode, PPC970_Unit_CRU;
let hasExtraSrcRegAllocReq = 1 in // to enable post-ra anti-dep breaking.
def MFOCRF8: XFXForm_5a<31, 19, (outs g8rc:$rT), (ins crbitm:$FXM),
"mfocrf $rT, $FXM", IIC_SprMFCRF>,
PPC970_DGroup_First, PPC970_Unit_CRU;
def MFCR8 : XFXForm_3<31, 19, (outs g8rc:$rT), (ins),
"mfcr $rT", IIC_SprMFCR>,
PPC970_MicroCode, PPC970_Unit_CRU;
} // hasSideEffects = 0
let hasSideEffects = 1, isBarrier = 1, usesCustomInserter = 1 in {
let Defs = [CTR8] in
def EH_SjLj_SetJmp64 : Pseudo<(outs gprc:$dst), (ins memr:$buf),
"#EH_SJLJ_SETJMP64",
[(set i32:$dst, (PPCeh_sjlj_setjmp addr:$buf))]>,
Requires<[In64BitMode]>;
let isTerminator = 1 in
def EH_SjLj_LongJmp64 : Pseudo<(outs), (ins memr:$buf),
"#EH_SJLJ_LONGJMP64",
[(PPCeh_sjlj_longjmp addr:$buf)]>,
Requires<[In64BitMode]>;
}
//===----------------------------------------------------------------------===//
// 64-bit SPR manipulation instrs.
let Uses = [CTR8] in {
def MFCTR8 : XFXForm_1_ext<31, 339, 9, (outs g8rc:$rT), (ins),
"mfctr $rT", IIC_SprMFSPR>,
PPC970_DGroup_First, PPC970_Unit_FXU;
}
let Pattern = [(PPCmtctr i64:$rS)], Defs = [CTR8] in {
def MTCTR8 : XFXForm_7_ext<31, 467, 9, (outs), (ins g8rc:$rS),
"mtctr $rS", IIC_SprMTSPR>,
PPC970_DGroup_First, PPC970_Unit_FXU;
}
let hasSideEffects = 1, Defs = [CTR8] in {
let Pattern = [(int_ppc_mtctr i64:$rS)] in
def MTCTR8loop : XFXForm_7_ext<31, 467, 9, (outs), (ins g8rc:$rS),
"mtctr $rS", IIC_SprMTSPR>,
PPC970_DGroup_First, PPC970_Unit_FXU;
}
let Pattern = [(set i64:$rT, readcyclecounter)] in
def MFTB8 : XFXForm_1_ext<31, 339, 268, (outs g8rc:$rT), (ins),
"mfspr $rT, 268", IIC_SprMFTB>,
PPC970_DGroup_First, PPC970_Unit_FXU;
// Note that encoding mftb using mfspr is now the preferred form,
// and has been since at least ISA v2.03. The mftb instruction has
// now been phased out. Using mfspr, however, is known not to work on
// the POWER3.
let Defs = [X1], Uses = [X1] in
def DYNALLOC8 : Pseudo<(outs g8rc:$result), (ins g8rc:$negsize, memri:$fpsi),"#DYNALLOC8",
[(set i64:$result,
(PPCdynalloc i64:$negsize, iaddr:$fpsi))]>;
let Defs = [LR8] in {
def MTLR8 : XFXForm_7_ext<31, 467, 8, (outs), (ins g8rc:$rS),
"mtlr $rS", IIC_SprMTSPR>,
PPC970_DGroup_First, PPC970_Unit_FXU;
}
let Uses = [LR8] in {
def MFLR8 : XFXForm_1_ext<31, 339, 8, (outs g8rc:$rT), (ins),
"mflr $rT", IIC_SprMFSPR>,
PPC970_DGroup_First, PPC970_Unit_FXU;
}
} // Interpretation64Bit
//===----------------------------------------------------------------------===//
// Fixed point instructions.
//
let PPC970_Unit = 1 in { // FXU Operations.
let Interpretation64Bit = 1 in {
let hasSideEffects = 0 in {
let isCodeGenOnly = 1 in {
let isReMaterializable = 1, isAsCheapAsAMove = 1, isMoveImm = 1 in {
def LI8 : DForm_2_r0<14, (outs g8rc:$rD), (ins s16imm64:$imm),
"li $rD, $imm", IIC_IntSimple,
[(set i64:$rD, imm64SExt16:$imm)]>;
def LIS8 : DForm_2_r0<15, (outs g8rc:$rD), (ins s17imm64:$imm),
"lis $rD, $imm", IIC_IntSimple,
[(set i64:$rD, imm16ShiftedSExt:$imm)]>;
}
// Logical ops.
let isCommutable = 1 in {
defm NAND8: XForm_6r<31, 476, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"nand", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (not (and i64:$rS, i64:$rB)))]>;
defm AND8 : XForm_6r<31, 28, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"and", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (and i64:$rS, i64:$rB))]>;
} // isCommutable
defm ANDC8: XForm_6r<31, 60, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"andc", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (and i64:$rS, (not i64:$rB)))]>;
let isCommutable = 1 in {
defm OR8 : XForm_6r<31, 444, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"or", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (or i64:$rS, i64:$rB))]>;
defm NOR8 : XForm_6r<31, 124, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"nor", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (not (or i64:$rS, i64:$rB)))]>;
} // isCommutable
defm ORC8 : XForm_6r<31, 412, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"orc", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (or i64:$rS, (not i64:$rB)))]>;
let isCommutable = 1 in {
defm EQV8 : XForm_6r<31, 284, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"eqv", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (not (xor i64:$rS, i64:$rB)))]>;
defm XOR8 : XForm_6r<31, 316, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"xor", "$rA, $rS, $rB", IIC_IntSimple,
[(set i64:$rA, (xor i64:$rS, i64:$rB))]>;
} // let isCommutable = 1
// Logical ops with immediate.
let Defs = [CR0] in {
def ANDIo8 : DForm_4<28, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"andi. $dst, $src1, $src2", IIC_IntGeneral,
[(set i64:$dst, (and i64:$src1, immZExt16:$src2))]>,
isDOT;
def ANDISo8 : DForm_4<29, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"andis. $dst, $src1, $src2", IIC_IntGeneral,
[(set i64:$dst, (and i64:$src1, imm16ShiftedZExt:$src2))]>,
isDOT;
}
def ORI8 : DForm_4<24, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"ori $dst, $src1, $src2", IIC_IntSimple,
[(set i64:$dst, (or i64:$src1, immZExt16:$src2))]>;
def ORIS8 : DForm_4<25, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"oris $dst, $src1, $src2", IIC_IntSimple,
[(set i64:$dst, (or i64:$src1, imm16ShiftedZExt:$src2))]>;
def XORI8 : DForm_4<26, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"xori $dst, $src1, $src2", IIC_IntSimple,
[(set i64:$dst, (xor i64:$src1, immZExt16:$src2))]>;
def XORIS8 : DForm_4<27, (outs g8rc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"xoris $dst, $src1, $src2", IIC_IntSimple,
[(set i64:$dst, (xor i64:$src1, imm16ShiftedZExt:$src2))]>;
let isCommutable = 1 in
defm ADD8 : XOForm_1r<31, 266, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"add", "$rT, $rA, $rB", IIC_IntSimple,
[(set i64:$rT, (add i64:$rA, i64:$rB))]>;
// ADD8 has a special form: reg = ADD8(reg, sym@tls) for use by the
// initial-exec thread-local storage model.
def ADD8TLS : XOForm_1<31, 266, 0, (outs g8rc:$rT), (ins g8rc:$rA, tlsreg:$rB),
"add $rT, $rA, $rB", IIC_IntSimple,
[(set i64:$rT, (add i64:$rA, tglobaltlsaddr:$rB))]>;
let isCommutable = 1 in
defm ADDC8 : XOForm_1rc<31, 10, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"addc", "$rT, $rA, $rB", IIC_IntGeneral,
[(set i64:$rT, (addc i64:$rA, i64:$rB))]>,
PPC970_DGroup_Cracked;
let Defs = [CARRY] in
def ADDIC8 : DForm_2<12, (outs g8rc:$rD), (ins g8rc:$rA, s16imm64:$imm),
"addic $rD, $rA, $imm", IIC_IntGeneral,
[(set i64:$rD, (addc i64:$rA, imm64SExt16:$imm))]>;
def ADDI8 : DForm_2<14, (outs g8rc:$rD), (ins g8rc_nox0:$rA, s16imm64:$imm),
"addi $rD, $rA, $imm", IIC_IntSimple,
[(set i64:$rD, (add i64:$rA, imm64SExt16:$imm))]>;
def ADDIS8 : DForm_2<15, (outs g8rc:$rD), (ins g8rc_nox0:$rA, s17imm64:$imm),
"addis $rD, $rA, $imm", IIC_IntSimple,
[(set i64:$rD, (add i64:$rA, imm16ShiftedSExt:$imm))]>;
let Defs = [CARRY] in {
def SUBFIC8: DForm_2< 8, (outs g8rc:$rD), (ins g8rc:$rA, s16imm64:$imm),
"subfic $rD, $rA, $imm", IIC_IntGeneral,
[(set i64:$rD, (subc imm64SExt16:$imm, i64:$rA))]>;
defm SUBFC8 : XOForm_1r<31, 8, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"subfc", "$rT, $rA, $rB", IIC_IntGeneral,
[(set i64:$rT, (subc i64:$rB, i64:$rA))]>,
PPC970_DGroup_Cracked;
}
defm SUBF8 : XOForm_1r<31, 40, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"subf", "$rT, $rA, $rB", IIC_IntGeneral,
[(set i64:$rT, (sub i64:$rB, i64:$rA))]>;
defm NEG8 : XOForm_3r<31, 104, 0, (outs g8rc:$rT), (ins g8rc:$rA),
"neg", "$rT, $rA", IIC_IntSimple,
[(set i64:$rT, (ineg i64:$rA))]>;
let Uses = [CARRY] in {
let isCommutable = 1 in
defm ADDE8 : XOForm_1rc<31, 138, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"adde", "$rT, $rA, $rB", IIC_IntGeneral,
[(set i64:$rT, (adde i64:$rA, i64:$rB))]>;
defm ADDME8 : XOForm_3rc<31, 234, 0, (outs g8rc:$rT), (ins g8rc:$rA),
"addme", "$rT, $rA", IIC_IntGeneral,
[(set i64:$rT, (adde i64:$rA, -1))]>;
defm ADDZE8 : XOForm_3rc<31, 202, 0, (outs g8rc:$rT), (ins g8rc:$rA),
"addze", "$rT, $rA", IIC_IntGeneral,
[(set i64:$rT, (adde i64:$rA, 0))]>;
defm SUBFE8 : XOForm_1rc<31, 136, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"subfe", "$rT, $rA, $rB", IIC_IntGeneral,
[(set i64:$rT, (sube i64:$rB, i64:$rA))]>;
defm SUBFME8 : XOForm_3rc<31, 232, 0, (outs g8rc:$rT), (ins g8rc:$rA),
"subfme", "$rT, $rA", IIC_IntGeneral,
[(set i64:$rT, (sube -1, i64:$rA))]>;
defm SUBFZE8 : XOForm_3rc<31, 200, 0, (outs g8rc:$rT), (ins g8rc:$rA),
"subfze", "$rT, $rA", IIC_IntGeneral,
[(set i64:$rT, (sube 0, i64:$rA))]>;
}
} // isCodeGenOnly
// FIXME: Duplicating this for the asm parser should be unnecessary, but the
// previous definition must be marked as CodeGen only to prevent decoding
// conflicts.
let isAsmParserOnly = 1 in
def ADD8TLS_ : XOForm_1<31, 266, 0, (outs g8rc:$rT), (ins g8rc:$rA, tlsreg:$rB),
"add $rT, $rA, $rB", IIC_IntSimple, []>;
let isCommutable = 1 in {
defm MULHD : XOForm_1r<31, 73, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"mulhd", "$rT, $rA, $rB", IIC_IntMulHW,
[(set i64:$rT, (mulhs i64:$rA, i64:$rB))]>;
defm MULHDU : XOForm_1r<31, 9, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"mulhdu", "$rT, $rA, $rB", IIC_IntMulHWU,
[(set i64:$rT, (mulhu i64:$rA, i64:$rB))]>;
} // isCommutable
}
} // Interpretation64Bit
let isCompare = 1, hasSideEffects = 0 in {
def CMPD : XForm_16_ext<31, 0, (outs crrc:$crD), (ins g8rc:$rA, g8rc:$rB),
"cmpd $crD, $rA, $rB", IIC_IntCompare>, isPPC64;
def CMPLD : XForm_16_ext<31, 32, (outs crrc:$crD), (ins g8rc:$rA, g8rc:$rB),
"cmpld $crD, $rA, $rB", IIC_IntCompare>, isPPC64;
def CMPDI : DForm_5_ext<11, (outs crrc:$crD), (ins g8rc:$rA, s16imm64:$imm),
"cmpdi $crD, $rA, $imm", IIC_IntCompare>, isPPC64;
def CMPLDI : DForm_6_ext<10, (outs crrc:$dst), (ins g8rc:$src1, u16imm64:$src2),
"cmpldi $dst, $src1, $src2",
IIC_IntCompare>, isPPC64;
}
let hasSideEffects = 0 in {
defm SLD : XForm_6r<31, 27, (outs g8rc:$rA), (ins g8rc:$rS, gprc:$rB),
"sld", "$rA, $rS, $rB", IIC_IntRotateD,
[(set i64:$rA, (PPCshl i64:$rS, i32:$rB))]>, isPPC64;
defm SRD : XForm_6r<31, 539, (outs g8rc:$rA), (ins g8rc:$rS, gprc:$rB),
"srd", "$rA, $rS, $rB", IIC_IntRotateD,
[(set i64:$rA, (PPCsrl i64:$rS, i32:$rB))]>, isPPC64;
defm SRAD : XForm_6rc<31, 794, (outs g8rc:$rA), (ins g8rc:$rS, gprc:$rB),
"srad", "$rA, $rS, $rB", IIC_IntRotateD,
[(set i64:$rA, (PPCsra i64:$rS, i32:$rB))]>, isPPC64;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
defm EXTSB8 : XForm_11r<31, 954, (outs g8rc:$rA), (ins g8rc:$rS),
"extsb", "$rA, $rS", IIC_IntSimple,
[(set i64:$rA, (sext_inreg i64:$rS, i8))]>;
defm EXTSH8 : XForm_11r<31, 922, (outs g8rc:$rA), (ins g8rc:$rS),
"extsh", "$rA, $rS", IIC_IntSimple,
[(set i64:$rA, (sext_inreg i64:$rS, i16))]>;
defm SLW8 : XForm_6r<31, 24, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"slw", "$rA, $rS, $rB", IIC_IntGeneral, []>;
defm SRW8 : XForm_6r<31, 536, (outs g8rc:$rA), (ins g8rc:$rS, g8rc:$rB),
"srw", "$rA, $rS, $rB", IIC_IntGeneral, []>;
} // Interpretation64Bit
// For fast-isel:
let isCodeGenOnly = 1 in {
def EXTSB8_32_64 : XForm_11<31, 954, (outs g8rc:$rA), (ins gprc:$rS),
"extsb $rA, $rS", IIC_IntSimple, []>, isPPC64;
def EXTSH8_32_64 : XForm_11<31, 922, (outs g8rc:$rA), (ins gprc:$rS),
"extsh $rA, $rS", IIC_IntSimple, []>, isPPC64;
} // isCodeGenOnly for fast-isel
defm EXTSW : XForm_11r<31, 986, (outs g8rc:$rA), (ins g8rc:$rS),
"extsw", "$rA, $rS", IIC_IntSimple,
[(set i64:$rA, (sext_inreg i64:$rS, i32))]>, isPPC64;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
defm EXTSW_32_64 : XForm_11r<31, 986, (outs g8rc:$rA), (ins gprc:$rS),
"extsw", "$rA, $rS", IIC_IntSimple,
[(set i64:$rA, (sext i32:$rS))]>, isPPC64;
defm SRADI : XSForm_1rc<31, 413, (outs g8rc:$rA), (ins g8rc:$rS, u6imm:$SH),
"sradi", "$rA, $rS, $SH", IIC_IntRotateDI,
[(set i64:$rA, (sra i64:$rS, (i32 imm:$SH)))]>, isPPC64;
defm CNTLZD : XForm_11r<31, 58, (outs g8rc:$rA), (ins g8rc:$rS),
"cntlzd", "$rA, $rS", IIC_IntGeneral,
[(set i64:$rA, (ctlz i64:$rS))]>;
def POPCNTD : XForm_11<31, 506, (outs g8rc:$rA), (ins g8rc:$rS),
"popcntd $rA, $rS", IIC_IntGeneral,
[(set i64:$rA, (ctpop i64:$rS))]>;
// popcntw also does a population count on the high 32 bits (storing the
// results in the high 32-bits of the output). We'll ignore that here (which is
// safe because we never separately use the high part of the 64-bit registers).
def POPCNTW : XForm_11<31, 378, (outs gprc:$rA), (ins gprc:$rS),
"popcntw $rA, $rS", IIC_IntGeneral,
[(set i32:$rA, (ctpop i32:$rS))]>;
defm DIVD : XOForm_1r<31, 489, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"divd", "$rT, $rA, $rB", IIC_IntDivD,
[(set i64:$rT, (sdiv i64:$rA, i64:$rB))]>, isPPC64,
PPC970_DGroup_First, PPC970_DGroup_Cracked;
defm DIVDU : XOForm_1r<31, 457, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"divdu", "$rT, $rA, $rB", IIC_IntDivD,
[(set i64:$rT, (udiv i64:$rA, i64:$rB))]>, isPPC64,
PPC970_DGroup_First, PPC970_DGroup_Cracked;
let isCommutable = 1 in
defm MULLD : XOForm_1r<31, 233, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
"mulld", "$rT, $rA, $rB", IIC_IntMulHD,
[(set i64:$rT, (mul i64:$rA, i64:$rB))]>, isPPC64;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
def MULLI8 : DForm_2<7, (outs g8rc:$rD), (ins g8rc:$rA, s16imm64:$imm),
"mulli $rD, $rA, $imm", IIC_IntMulLI,
[(set i64:$rD, (mul i64:$rA, imm64SExt16:$imm))]>;
}
let hasSideEffects = 0 in {
let isCommutable = 1 in {
defm RLDIMI : MDForm_1r<30, 3, (outs g8rc:$rA),
(ins g8rc:$rSi, g8rc:$rS, u6imm:$SH, u6imm:$MBE),
"rldimi", "$rA, $rS, $SH, $MBE", IIC_IntRotateDI,
[]>, isPPC64, RegConstraint<"$rSi = $rA">,
NoEncode<"$rSi">;
}
// Rotate instructions.
defm RLDCL : MDSForm_1r<30, 8,
(outs g8rc:$rA), (ins g8rc:$rS, gprc:$rB, u6imm:$MBE),
"rldcl", "$rA, $rS, $rB, $MBE", IIC_IntRotateD,
[]>, isPPC64;
defm RLDCR : MDSForm_1r<30, 9,
(outs g8rc:$rA), (ins g8rc:$rS, gprc:$rB, u6imm:$MBE),
"rldcr", "$rA, $rS, $rB, $MBE", IIC_IntRotateD,
[]>, isPPC64;
defm RLDICL : MDForm_1r<30, 0,
(outs g8rc:$rA), (ins g8rc:$rS, u6imm:$SH, u6imm:$MBE),
"rldicl", "$rA, $rS, $SH, $MBE", IIC_IntRotateDI,
[]>, isPPC64;
// For fast-isel:
let isCodeGenOnly = 1 in
def RLDICL_32_64 : MDForm_1<30, 0,
(outs g8rc:$rA),
(ins gprc:$rS, u6imm:$SH, u6imm:$MBE),
"rldicl $rA, $rS, $SH, $MBE", IIC_IntRotateDI,
[]>, isPPC64;
// End fast-isel.
defm RLDICR : MDForm_1r<30, 1,
(outs g8rc:$rA), (ins g8rc:$rS, u6imm:$SH, u6imm:$MBE),
"rldicr", "$rA, $rS, $SH, $MBE", IIC_IntRotateDI,
[]>, isPPC64;
defm RLDIC : MDForm_1r<30, 2,
(outs g8rc:$rA), (ins g8rc:$rS, u6imm:$SH, u6imm:$MBE),
"rldic", "$rA, $rS, $SH, $MBE", IIC_IntRotateDI,
[]>, isPPC64;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
defm RLWINM8 : MForm_2r<21, (outs g8rc:$rA),
(ins g8rc:$rS, u5imm:$SH, u5imm:$MB, u5imm:$ME),
"rlwinm", "$rA, $rS, $SH, $MB, $ME", IIC_IntGeneral,
[]>;
defm RLWNM8 : MForm_2r<23, (outs g8rc:$rA),
(ins g8rc:$rS, g8rc:$rB, u5imm:$MB, u5imm:$ME),
"rlwnm", "$rA, $rS, $rB, $MB, $ME", IIC_IntGeneral,
[]>;
// RLWIMI can be commuted if the rotate amount is zero.
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
defm RLWIMI8 : MForm_2r<20, (outs g8rc:$rA),
(ins g8rc:$rSi, g8rc:$rS, u5imm:$SH, u5imm:$MB,
u5imm:$ME), "rlwimi", "$rA, $rS, $SH, $MB, $ME",
IIC_IntRotate, []>, PPC970_DGroup_Cracked,
RegConstraint<"$rSi = $rA">, NoEncode<"$rSi">;
let isSelect = 1 in
def ISEL8 : AForm_4<31, 15,
(outs g8rc:$rT), (ins g8rc_nox0:$rA, g8rc:$rB, crbitrc:$cond),
"isel $rT, $rA, $rB, $cond", IIC_IntGeneral,
[]>;
} // Interpretation64Bit
} // hasSideEffects = 0
} // End FXU Operations.
//===----------------------------------------------------------------------===//
// Load/Store instructions.
//
// Sign extending loads.
let canFoldAsLoad = 1, PPC970_Unit = 2 in {
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
def LHA8: DForm_1<42, (outs g8rc:$rD), (ins memri:$src),
"lha $rD, $src", IIC_LdStLHA,
[(set i64:$rD, (sextloadi16 iaddr:$src))]>,
PPC970_DGroup_Cracked;
def LWA : DSForm_1<58, 2, (outs g8rc:$rD), (ins memrix:$src),
"lwa $rD, $src", IIC_LdStLWA,
[(set i64:$rD,
(aligned4sextloadi32 ixaddr:$src))]>, isPPC64,
PPC970_DGroup_Cracked;
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
def LHAX8: XForm_1<31, 343, (outs g8rc:$rD), (ins memrr:$src),
"lhax $rD, $src", IIC_LdStLHA,
[(set i64:$rD, (sextloadi16 xaddr:$src))]>,
PPC970_DGroup_Cracked;
def LWAX : XForm_1<31, 341, (outs g8rc:$rD), (ins memrr:$src),
"lwax $rD, $src", IIC_LdStLHA,
[(set i64:$rD, (sextloadi32 xaddr:$src))]>, isPPC64,
PPC970_DGroup_Cracked;
// For fast-isel:
let isCodeGenOnly = 1, mayLoad = 1 in {
def LWA_32 : DSForm_1<58, 2, (outs gprc:$rD), (ins memrix:$src),
"lwa $rD, $src", IIC_LdStLWA, []>, isPPC64,
PPC970_DGroup_Cracked;
def LWAX_32 : XForm_1<31, 341, (outs gprc:$rD), (ins memrr:$src),
"lwax $rD, $src", IIC_LdStLHA, []>, isPPC64,
PPC970_DGroup_Cracked;
} // end fast-isel isCodeGenOnly
// Update forms.
let mayLoad = 1, hasSideEffects = 0 in {
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
def LHAU8 : DForm_1<43, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memri:$addr),
"lhau $rD, $addr", IIC_LdStLHAU,
[]>, RegConstraint<"$addr.reg = $ea_result">,
NoEncode<"$ea_result">;
// NO LWAU!
let Interpretation64Bit = 1, isCodeGenOnly = 1 in
def LHAUX8 : XForm_1<31, 375, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"lhaux $rD, $addr", IIC_LdStLHAUX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">;
def LWAUX : XForm_1<31, 373, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"lwaux $rD, $addr", IIC_LdStLHAUX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">, isPPC64;
}
}
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
// Zero extending loads.
let canFoldAsLoad = 1, PPC970_Unit = 2 in {
def LBZ8 : DForm_1<34, (outs g8rc:$rD), (ins memri:$src),
"lbz $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi8 iaddr:$src))]>;
def LHZ8 : DForm_1<40, (outs g8rc:$rD), (ins memri:$src),
"lhz $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi16 iaddr:$src))]>;
def LWZ8 : DForm_1<32, (outs g8rc:$rD), (ins memri:$src),
"lwz $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi32 iaddr:$src))]>, isPPC64;
def LBZX8 : XForm_1<31, 87, (outs g8rc:$rD), (ins memrr:$src),
"lbzx $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi8 xaddr:$src))]>;
def LHZX8 : XForm_1<31, 279, (outs g8rc:$rD), (ins memrr:$src),
"lhzx $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi16 xaddr:$src))]>;
def LWZX8 : XForm_1<31, 23, (outs g8rc:$rD), (ins memrr:$src),
"lwzx $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (zextloadi32 xaddr:$src))]>;
// Update forms.
let mayLoad = 1, hasSideEffects = 0 in {
def LBZU8 : DForm_1<35, (outs g8rc:$rD, ptr_rc_nor0:$ea_result), (ins memri:$addr),
"lbzu $rD, $addr", IIC_LdStLoadUpd,
[]>, RegConstraint<"$addr.reg = $ea_result">,
NoEncode<"$ea_result">;
def LHZU8 : DForm_1<41, (outs g8rc:$rD, ptr_rc_nor0:$ea_result), (ins memri:$addr),
"lhzu $rD, $addr", IIC_LdStLoadUpd,
[]>, RegConstraint<"$addr.reg = $ea_result">,
NoEncode<"$ea_result">;
def LWZU8 : DForm_1<33, (outs g8rc:$rD, ptr_rc_nor0:$ea_result), (ins memri:$addr),
"lwzu $rD, $addr", IIC_LdStLoadUpd,
[]>, RegConstraint<"$addr.reg = $ea_result">,
NoEncode<"$ea_result">;
def LBZUX8 : XForm_1<31, 119, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"lbzux $rD, $addr", IIC_LdStLoadUpdX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">;
def LHZUX8 : XForm_1<31, 311, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"lhzux $rD, $addr", IIC_LdStLoadUpdX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">;
def LWZUX8 : XForm_1<31, 55, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"lwzux $rD, $addr", IIC_LdStLoadUpdX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">;
}
}
} // Interpretation64Bit
// Full 8-byte loads.
let canFoldAsLoad = 1, PPC970_Unit = 2 in {
def LD : DSForm_1<58, 0, (outs g8rc:$rD), (ins memrix:$src),
"ld $rD, $src", IIC_LdStLD,
[(set i64:$rD, (aligned4load ixaddr:$src))]>, isPPC64;
// The following four definitions are selected for small code model only.
// Otherwise, we need to create two instructions to form a 32-bit offset,
// so we have a custom matcher for TOC_ENTRY in PPCDAGToDAGIsel::Select().
def LDtoc: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg),
"#LDtoc",
[(set i64:$rD,
(PPCtoc_entry tglobaladdr:$disp, i64:$reg))]>, isPPC64;
def LDtocJTI: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg),
"#LDtocJTI",
[(set i64:$rD,
(PPCtoc_entry tjumptable:$disp, i64:$reg))]>, isPPC64;
def LDtocCPT: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg),
"#LDtocCPT",
[(set i64:$rD,
(PPCtoc_entry tconstpool:$disp, i64:$reg))]>, isPPC64;
def LDtocBA: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc:$reg),
"#LDtocCPT",
[(set i64:$rD,
(PPCtoc_entry tblockaddress:$disp, i64:$reg))]>, isPPC64;
let hasSideEffects = 1, isCodeGenOnly = 1, RST = 2, Defs = [X2] in
def LDinto_toc: DSForm_1<58, 0, (outs), (ins memrix:$src),
"ld 2, $src", IIC_LdStLD,
[(PPCload_toc ixaddr:$src)]>, isPPC64;
def LDX : XForm_1<31, 21, (outs g8rc:$rD), (ins memrr:$src),
"ldx $rD, $src", IIC_LdStLD,
[(set i64:$rD, (load xaddr:$src))]>, isPPC64;
def LDBRX : XForm_1<31, 532, (outs g8rc:$rD), (ins memrr:$src),
"ldbrx $rD, $src", IIC_LdStLoad,
[(set i64:$rD, (PPClbrx xoaddr:$src, i64))]>, isPPC64;
let mayLoad = 1, hasSideEffects = 0 in {
def LDU : DSForm_1<58, 1, (outs g8rc:$rD, ptr_rc_nor0:$ea_result), (ins memrix:$addr),
"ldu $rD, $addr", IIC_LdStLDU,
[]>, RegConstraint<"$addr.reg = $ea_result">, isPPC64,
NoEncode<"$ea_result">;
def LDUX : XForm_1<31, 53, (outs g8rc:$rD, ptr_rc_nor0:$ea_result),
(ins memrr:$addr),
"ldux $rD, $addr", IIC_LdStLDUX,
[]>, RegConstraint<"$addr.ptrreg = $ea_result">,
NoEncode<"$ea_result">, isPPC64;
}
}
def : Pat<(PPCload ixaddr:$src),
(LD ixaddr:$src)>;
def : Pat<(PPCload xaddr:$src),
(LDX xaddr:$src)>;
// Support for medium and large code model.
def ADDIStocHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, tocentry:$disp),
"#ADDIStocHA",
[(set i64:$rD,
(PPCaddisTocHA i64:$reg, tglobaladdr:$disp))]>,
isPPC64;
def LDtocL: Pseudo<(outs g8rc:$rD), (ins tocentry:$disp, g8rc_nox0:$reg),
"#LDtocL",
[(set i64:$rD,
(PPCldTocL tglobaladdr:$disp, i64:$reg))]>, isPPC64;
def ADDItocL: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, tocentry:$disp),
"#ADDItocL",
[(set i64:$rD,
(PPCaddiTocL i64:$reg, tglobaladdr:$disp))]>, isPPC64;
// Support for thread-local storage.
def ADDISgotTprelHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDISgotTprelHA",
[(set i64:$rD,
(PPCaddisGotTprelHA i64:$reg,
tglobaltlsaddr:$disp))]>,
isPPC64;
def LDgotTprelL: Pseudo<(outs g8rc:$rD), (ins s16imm64:$disp, g8rc_nox0:$reg),
"#LDgotTprelL",
[(set i64:$rD,
(PPCldGotTprelL tglobaltlsaddr:$disp, i64:$reg))]>,
isPPC64;
def : Pat<(PPCaddTls i64:$in, tglobaltlsaddr:$g),
(ADD8TLS $in, tglobaltlsaddr:$g)>;
def ADDIStlsgdHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDIStlsgdHA",
[(set i64:$rD,
(PPCaddisTlsgdHA i64:$reg, tglobaltlsaddr:$disp))]>,
isPPC64;
def ADDItlsgdL : Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDItlsgdL",
[(set i64:$rD,
(PPCaddiTlsgdL i64:$reg, tglobaltlsaddr:$disp))]>,
isPPC64;
def ADDIStlsldHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDIStlsldHA",
[(set i64:$rD,
(PPCaddisTlsldHA i64:$reg, tglobaltlsaddr:$disp))]>,
isPPC64;
def ADDItlsldL : Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDItlsldL",
[(set i64:$rD,
(PPCaddiTlsldL i64:$reg, tglobaltlsaddr:$disp))]>,
isPPC64;
def ADDISdtprelHA: Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDISdtprelHA",
[(set i64:$rD,
(PPCaddisDtprelHA i64:$reg,
tglobaltlsaddr:$disp))]>,
isPPC64;
def ADDIdtprelL : Pseudo<(outs g8rc:$rD), (ins g8rc_nox0:$reg, s16imm64:$disp),
"#ADDIdtprelL",
[(set i64:$rD,
(PPCaddiDtprelL i64:$reg, tglobaltlsaddr:$disp))]>,
isPPC64;
let PPC970_Unit = 2 in {
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
// Truncating stores.
def STB8 : DForm_1<38, (outs), (ins g8rc:$rS, memri:$src),
"stb $rS, $src", IIC_LdStStore,
[(truncstorei8 i64:$rS, iaddr:$src)]>;
def STH8 : DForm_1<44, (outs), (ins g8rc:$rS, memri:$src),
"sth $rS, $src", IIC_LdStStore,
[(truncstorei16 i64:$rS, iaddr:$src)]>;
def STW8 : DForm_1<36, (outs), (ins g8rc:$rS, memri:$src),
"stw $rS, $src", IIC_LdStStore,
[(truncstorei32 i64:$rS, iaddr:$src)]>;
def STBX8 : XForm_8<31, 215, (outs), (ins g8rc:$rS, memrr:$dst),
"stbx $rS, $dst", IIC_LdStStore,
[(truncstorei8 i64:$rS, xaddr:$dst)]>,
PPC970_DGroup_Cracked;
def STHX8 : XForm_8<31, 407, (outs), (ins g8rc:$rS, memrr:$dst),
"sthx $rS, $dst", IIC_LdStStore,
[(truncstorei16 i64:$rS, xaddr:$dst)]>,
PPC970_DGroup_Cracked;
def STWX8 : XForm_8<31, 151, (outs), (ins g8rc:$rS, memrr:$dst),
"stwx $rS, $dst", IIC_LdStStore,
[(truncstorei32 i64:$rS, xaddr:$dst)]>,
PPC970_DGroup_Cracked;
} // Interpretation64Bit
// Normal 8-byte stores.
def STD : DSForm_1<62, 0, (outs), (ins g8rc:$rS, memrix:$dst),
"std $rS, $dst", IIC_LdStSTD,
[(aligned4store i64:$rS, ixaddr:$dst)]>, isPPC64;
def STDX : XForm_8<31, 149, (outs), (ins g8rc:$rS, memrr:$dst),
"stdx $rS, $dst", IIC_LdStSTD,
[(store i64:$rS, xaddr:$dst)]>, isPPC64,
PPC970_DGroup_Cracked;
def STDBRX: XForm_8<31, 660, (outs), (ins g8rc:$rS, memrr:$dst),
"stdbrx $rS, $dst", IIC_LdStStore,
[(PPCstbrx i64:$rS, xoaddr:$dst, i64)]>, isPPC64,
PPC970_DGroup_Cracked;
}
// Stores with Update (pre-inc).
let PPC970_Unit = 2, mayStore = 1 in {
let Interpretation64Bit = 1, isCodeGenOnly = 1 in {
def STBU8 : DForm_1<39, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memri:$dst),
"stbu $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.reg = $ea_res">, NoEncode<"$ea_res">;
def STHU8 : DForm_1<45, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memri:$dst),
"sthu $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.reg = $ea_res">, NoEncode<"$ea_res">;
def STWU8 : DForm_1<37, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memri:$dst),
"stwu $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.reg = $ea_res">, NoEncode<"$ea_res">;
def STBUX8: XForm_8<31, 247, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memrr:$dst),
"stbux $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.ptrreg = $ea_res">, NoEncode<"$ea_res">,
PPC970_DGroup_Cracked;
def STHUX8: XForm_8<31, 439, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memrr:$dst),
"sthux $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.ptrreg = $ea_res">, NoEncode<"$ea_res">,
PPC970_DGroup_Cracked;
def STWUX8: XForm_8<31, 183, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memrr:$dst),
"stwux $rS, $dst", IIC_LdStStoreUpd, []>,
RegConstraint<"$dst.ptrreg = $ea_res">, NoEncode<"$ea_res">,
PPC970_DGroup_Cracked;
} // Interpretation64Bit
def STDU : DSForm_1<62, 1, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memrix:$dst),
"stdu $rS, $dst", IIC_LdStSTDU, []>,
RegConstraint<"$dst.reg = $ea_res">, NoEncode<"$ea_res">,
isPPC64;
def STDUX : XForm_8<31, 181, (outs ptr_rc_nor0:$ea_res), (ins g8rc:$rS, memrr:$dst),
"stdux $rS, $dst", IIC_LdStSTDUX, []>,
RegConstraint<"$dst.ptrreg = $ea_res">, NoEncode<"$ea_res">,
PPC970_DGroup_Cracked, isPPC64;
}
// Patterns to match the pre-inc stores. We can't put the patterns on
// the instruction definitions directly as ISel wants the address base
// and offset to be separate operands, not a single complex operand.
def : Pat<(pre_truncsti8 i64:$rS, iPTR:$ptrreg, iaddroff:$ptroff),
(STBU8 $rS, iaddroff:$ptroff, $ptrreg)>;
def : Pat<(pre_truncsti16 i64:$rS, iPTR:$ptrreg, iaddroff:$ptroff),
(STHU8 $rS, iaddroff:$ptroff, $ptrreg)>;
def : Pat<(pre_truncsti32 i64:$rS, iPTR:$ptrreg, iaddroff:$ptroff),
(STWU8 $rS, iaddroff:$ptroff, $ptrreg)>;
def : Pat<(aligned4pre_store i64:$rS, iPTR:$ptrreg, iaddroff:$ptroff),
(STDU $rS, iaddroff:$ptroff, $ptrreg)>;
def : Pat<(pre_truncsti8 i64:$rS, iPTR:$ptrreg, iPTR:$ptroff),
(STBUX8 $rS, $ptrreg, $ptroff)>;
def : Pat<(pre_truncsti16 i64:$rS, iPTR:$ptrreg, iPTR:$ptroff),
(STHUX8 $rS, $ptrreg, $ptroff)>;
def : Pat<(pre_truncsti32 i64:$rS, iPTR:$ptrreg, iPTR:$ptroff),
(STWUX8 $rS, $ptrreg, $ptroff)>;
def : Pat<(pre_store i64:$rS, iPTR:$ptrreg, iPTR:$ptroff),
(STDUX $rS, $ptrreg, $ptroff)>;
//===----------------------------------------------------------------------===//
// Floating point instructions.
//
let PPC970_Unit = 3, hasSideEffects = 0,
Uses = [RM] in { // FPU Operations.
defm FCFID : XForm_26r<63, 846, (outs f8rc:$frD), (ins f8rc:$frB),
"fcfid", "$frD, $frB", IIC_FPGeneral,
[(set f64:$frD, (PPCfcfid f64:$frB))]>, isPPC64;
defm FCTID : XForm_26r<63, 814, (outs f8rc:$frD), (ins f8rc:$frB),
"fctid", "$frD, $frB", IIC_FPGeneral,
[]>, isPPC64;
defm FCTIDZ : XForm_26r<63, 815, (outs f8rc:$frD), (ins f8rc:$frB),
"fctidz", "$frD, $frB", IIC_FPGeneral,
[(set f64:$frD, (PPCfctidz f64:$frB))]>, isPPC64;
defm FCFIDU : XForm_26r<63, 974, (outs f8rc:$frD), (ins f8rc:$frB),
"fcfidu", "$frD, $frB", IIC_FPGeneral,
[(set f64:$frD, (PPCfcfidu f64:$frB))]>, isPPC64;
defm FCFIDS : XForm_26r<59, 846, (outs f4rc:$frD), (ins f8rc:$frB),
"fcfids", "$frD, $frB", IIC_FPGeneral,
[(set f32:$frD, (PPCfcfids f64:$frB))]>, isPPC64;
defm FCFIDUS : XForm_26r<59, 974, (outs f4rc:$frD), (ins f8rc:$frB),
"fcfidus", "$frD, $frB", IIC_FPGeneral,
[(set f32:$frD, (PPCfcfidus f64:$frB))]>, isPPC64;
defm FCTIDUZ : XForm_26r<63, 943, (outs f8rc:$frD), (ins f8rc:$frB),
"fctiduz", "$frD, $frB", IIC_FPGeneral,
[(set f64:$frD, (PPCfctiduz f64:$frB))]>, isPPC64;
defm FCTIWUZ : XForm_26r<63, 143, (outs f8rc:$frD), (ins f8rc:$frB),
"fctiwuz", "$frD, $frB", IIC_FPGeneral,
[(set f64:$frD, (PPCfctiwuz f64:$frB))]>, isPPC64;
}
//===----------------------------------------------------------------------===//
// Instruction Patterns
//
// Extensions and truncates to/from 32-bit regs.
def : Pat<(i64 (zext i32:$in)),
(RLDICL (INSERT_SUBREG (i64 (IMPLICIT_DEF)), $in, sub_32),
0, 32)>;
def : Pat<(i64 (anyext i32:$in)),
(INSERT_SUBREG (i64 (IMPLICIT_DEF)), $in, sub_32)>;
def : Pat<(i32 (trunc i64:$in)),
(EXTRACT_SUBREG $in, sub_32)>;
// Implement the 'not' operation with the NOR instruction.
// (we could use the default xori pattern, but nor has lower latency on some
// cores (such as the A2)).
def i64not : OutPatFrag<(ops node:$in),
(NOR8 $in, $in)>;
def : Pat<(not i64:$in),
(i64not $in)>;
// Extending loads with i64 targets.
def : Pat<(zextloadi1 iaddr:$src),
(LBZ8 iaddr:$src)>;
def : Pat<(zextloadi1 xaddr:$src),
(LBZX8 xaddr:$src)>;
def : Pat<(extloadi1 iaddr:$src),
(LBZ8 iaddr:$src)>;
def : Pat<(extloadi1 xaddr:$src),
(LBZX8 xaddr:$src)>;
def : Pat<(extloadi8 iaddr:$src),
(LBZ8 iaddr:$src)>;
def : Pat<(extloadi8 xaddr:$src),
(LBZX8 xaddr:$src)>;
def : Pat<(extloadi16 iaddr:$src),
(LHZ8 iaddr:$src)>;
def : Pat<(extloadi16 xaddr:$src),
(LHZX8 xaddr:$src)>;
def : Pat<(extloadi32 iaddr:$src),
(LWZ8 iaddr:$src)>;
def : Pat<(extloadi32 xaddr:$src),
(LWZX8 xaddr:$src)>;
// Standard shifts. These are represented separately from the real shifts above
// so that we can distinguish between shifts that allow 6-bit and 7-bit shift
// amounts.
def : Pat<(sra i64:$rS, i32:$rB),
(SRAD $rS, $rB)>;
def : Pat<(srl i64:$rS, i32:$rB),
(SRD $rS, $rB)>;
def : Pat<(shl i64:$rS, i32:$rB),
(SLD $rS, $rB)>;
// SHL/SRL
def : Pat<(shl i64:$in, (i32 imm:$imm)),
(RLDICR $in, imm:$imm, (SHL64 imm:$imm))>;
def : Pat<(srl i64:$in, (i32 imm:$imm)),
(RLDICL $in, (SRL64 imm:$imm), imm:$imm)>;
// ROTL
def : Pat<(rotl i64:$in, i32:$sh),
(RLDCL $in, $sh, 0)>;
def : Pat<(rotl i64:$in, (i32 imm:$imm)),
(RLDICL $in, imm:$imm, 0)>;
// Hi and Lo for Darwin Global Addresses.
def : Pat<(PPChi tglobaladdr:$in, 0), (LIS8 tglobaladdr:$in)>;
def : Pat<(PPClo tglobaladdr:$in, 0), (LI8 tglobaladdr:$in)>;
def : Pat<(PPChi tconstpool:$in , 0), (LIS8 tconstpool:$in)>;
def : Pat<(PPClo tconstpool:$in , 0), (LI8 tconstpool:$in)>;
def : Pat<(PPChi tjumptable:$in , 0), (LIS8 tjumptable:$in)>;
def : Pat<(PPClo tjumptable:$in , 0), (LI8 tjumptable:$in)>;
def : Pat<(PPChi tblockaddress:$in, 0), (LIS8 tblockaddress:$in)>;
def : Pat<(PPClo tblockaddress:$in, 0), (LI8 tblockaddress:$in)>;
def : Pat<(PPChi tglobaltlsaddr:$g, i64:$in),
(ADDIS8 $in, tglobaltlsaddr:$g)>;
def : Pat<(PPClo tglobaltlsaddr:$g, i64:$in),
(ADDI8 $in, tglobaltlsaddr:$g)>;
def : Pat<(add i64:$in, (PPChi tglobaladdr:$g, 0)),
(ADDIS8 $in, tglobaladdr:$g)>;
def : Pat<(add i64:$in, (PPChi tconstpool:$g, 0)),
(ADDIS8 $in, tconstpool:$g)>;
def : Pat<(add i64:$in, (PPChi tjumptable:$g, 0)),
(ADDIS8 $in, tjumptable:$g)>;
def : Pat<(add i64:$in, (PPChi tblockaddress:$g, 0)),
(ADDIS8 $in, tblockaddress:$g)>;
// Patterns to match r+r indexed loads and stores for
// addresses without at least 4-byte alignment.
def : Pat<(i64 (unaligned4sextloadi32 xoaddr:$src)),
(LWAX xoaddr:$src)>;
def : Pat<(i64 (unaligned4load xoaddr:$src)),
(LDX xoaddr:$src)>;
def : Pat<(unaligned4store i64:$rS, xoaddr:$dst),
(STDX $rS, xoaddr:$dst)>;
// 64-bits atomic loads and stores
def : Pat<(atomic_load_64 ixaddr:$src), (LD memrix:$src)>;
def : Pat<(atomic_load_64 xaddr:$src), (LDX memrr:$src)>;
def : Pat<(atomic_store_64 ixaddr:$ptr, i64:$val), (STD g8rc:$val, memrix:$ptr)>;
def : Pat<(atomic_store_64 xaddr:$ptr, i64:$val), (STDX g8rc:$val, memrr:$ptr)>;