def ADJCALLSTACKUP : Pseudo<(outs), (ins i64imm:$amt1, i64imm:$amt2),
[(callseq_end timm:$amt1, timm:$amt2)]>;
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
// Takes as input the value of the stack pointer after a dynamic allocation
// has been made. Sets the output to the address of the dynamically-
// allocated area itself, skipping the outgoing arguments.
def BRCL : InstRIL<0xC04, (outs), (ins cond4:$valid, cond4:$R1,
brtarget32:$I2), "jg$R1\t$I2", []>;
}
- def AsmBRC : InstRI<0xA74, (outs), (ins uimm8zx4:$R1, brtarget16:$I2),
+ def AsmBRC : InstRI<0xA74, (outs), (ins imm32zx4:$R1, brtarget16:$I2),
"brc\t$R1, $I2", []>;
- def AsmBRCL : InstRIL<0xC04, (outs), (ins uimm8zx4:$R1, brtarget32:$I2),
+ def AsmBRCL : InstRIL<0xC04, (outs), (ins imm32zx4:$R1, brtarget32:$I2),
"brcl\t$R1, $I2", []>;
+ def AsmBCR : InstRR<0x07, (outs), (ins imm32zx4:$R1, GR64:$R2),
+ "bcr\t$R1, $R2", []>;
}
// Fused compare-and-branch instructions. As for normal branches,
}
let isCodeGenOnly = 1 in
defm C : CompareBranches<cond4, "$M3", "">;
-defm AsmC : CompareBranches<uimm8zx4, "", "$M3, ">;
+defm AsmC : CompareBranches<imm32zx4, "", "$M3, ">;
// Define AsmParser mnemonics for each general condition-code mask
// (integer or floating-point)
"j"##name##"\t$I2", []>;
def JG : InstRIL<0xC04, (outs), (ins brtarget32:$I2),
"jg"##name##"\t$I2", []>;
+ def BR : InstRR<0x07, (outs), (ins ADDR64:$R2), "b"##name##"r\t$R2", []>;
}
def LOCR : FixedCondUnaryRRF<"locr"##name, 0xB9F2, GR32, GR32, ccmask>;
def LOCGR : FixedCondUnaryRRF<"locgr"##name, 0xB9E2, GR64, GR64, ccmask>;
// Select instructions
//===----------------------------------------------------------------------===//
-def Select32 : SelectWrapper<GR32>;
-def Select64 : SelectWrapper<GR64>;
-
-defm CondStore8 : CondStores<GR32, nonvolatile_truncstorei8,
- nonvolatile_anyextloadi8, bdxaddr20only>;
-defm CondStore16 : CondStores<GR32, nonvolatile_truncstorei16,
- nonvolatile_anyextloadi16, bdxaddr20only>;
-defm CondStore32 : CondStores<GR32, nonvolatile_store,
- nonvolatile_load, bdxaddr20only>;
+def Select32Mux : SelectWrapper<GRX32>, Requires<[FeatureHighWord]>;
+def Select32 : SelectWrapper<GR32>;
+def Select64 : SelectWrapper<GR64>;
+
+// We don't define 32-bit Mux stores because the low-only STOC should
+// always be used if possible.
+defm CondStore8Mux : CondStores<GRX32, nonvolatile_truncstorei8,
+ nonvolatile_anyextloadi8, bdxaddr20only>,
+ Requires<[FeatureHighWord]>;
+defm CondStore16Mux : CondStores<GRX32, nonvolatile_truncstorei16,
+ nonvolatile_anyextloadi16, bdxaddr20only>,
+ Requires<[FeatureHighWord]>;
+defm CondStore8 : CondStores<GR32, nonvolatile_truncstorei8,
+ nonvolatile_anyextloadi8, bdxaddr20only>;
+defm CondStore16 : CondStores<GR32, nonvolatile_truncstorei16,
+ nonvolatile_anyextloadi16, bdxaddr20only>;
+defm CondStore32 : CondStores<GR32, nonvolatile_store,
+ nonvolatile_load, bdxaddr20only>;
defm : CondStores64<CondStore8, CondStore8Inv, nonvolatile_truncstorei8,
nonvolatile_anyextloadi8, bdxaddr20only>;
// Call instructions
//===----------------------------------------------------------------------===//
-// The definitions here are for the call-clobbered registers.
-let isCall = 1, Defs = [R0D, R1D, R2D, R3D, R4D, R5D, R14D,
- F0D, F1D, F2D, F3D, F4D, F5D, F6D, F7D, CC] in {
+let isCall = 1, Defs = [R14D, CC] in {
def CallBRASL : Alias<6, (outs), (ins pcrel32:$I2, variable_ops),
[(z_call pcrel32:$I2)]>;
def CallBASR : Alias<2, (outs), (ins ADDR64:$R2, variable_ops),
def CallBR : Alias<2, (outs), (ins), [(z_sibcall R1D)]>;
}
+// TLS calls. These will be lowered into a call to __tls_get_offset,
+// with an extra relocation specifying the TLS symbol.
+let isCall = 1, Defs = [R14D, CC] in {
+ def TLS_GDCALL : Alias<6, (outs), (ins tlssym:$I2, variable_ops),
+ [(z_tls_gdcall tglobaltlsaddr:$I2)]>;
+ def TLS_LDCALL : Alias<6, (outs), (ins tlssym:$I2, variable_ops),
+ [(z_tls_ldcall tglobaltlsaddr:$I2)]>;
+}
+
// Define the general form of the call instructions for the asm parser.
// These instructions don't hard-code %r14 as the return address register.
-def BRAS : InstRI<0xA75, (outs), (ins GR64:$R1, brtarget16:$I2),
+// Allow an optional TLS marker symbol to generate TLS call relocations.
+def BRAS : InstRI<0xA75, (outs), (ins GR64:$R1, brtarget16tls:$I2),
"bras\t$R1, $I2", []>;
-def BRASL : InstRIL<0xC05, (outs), (ins GR64:$R1, brtarget32:$I2),
+def BRASL : InstRIL<0xC05, (outs), (ins GR64:$R1, brtarget32tls:$I2),
"brasl\t$R1, $I2", []>;
def BASR : InstRR<0x0D, (outs), (ins GR64:$R1, ADDR64:$R2),
"basr\t$R1, $R2", []>;
//===----------------------------------------------------------------------===//
// Register moves.
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
+ // Expands to LR, RISBHG or RISBLG, depending on the choice of registers.
+ def LRMux : UnaryRRPseudo<"l", null_frag, GRX32, GRX32>,
+ Requires<[FeatureHighWord]>;
def LR : UnaryRR <"l", 0x18, null_frag, GR32, GR32>;
def LGR : UnaryRRE<"lg", 0xB904, null_frag, GR64, GR64>;
}
}
// Immediate moves.
-let neverHasSideEffects = 1, isAsCheapAsAMove = 1, isMoveImm = 1,
+let hasSideEffects = 0, isAsCheapAsAMove = 1, isMoveImm = 1,
isReMaterializable = 1 in {
- // 16-bit sign-extended immediates.
+ // 16-bit sign-extended immediates. LHIMux expands to LHI or IIHF,
+ // deopending on the choice of register.
+ def LHIMux : UnaryRIPseudo<bitconvert, GRX32, imm32sx16>,
+ Requires<[FeatureHighWord]>;
def LHI : UnaryRI<"lhi", 0xA78, bitconvert, GR32, imm32sx16>;
def LGHI : UnaryRI<"lghi", 0xA79, bitconvert, GR64, imm64sx16>;
// Register loads.
let canFoldAsLoad = 1, SimpleBDXLoad = 1 in {
+ // Expands to L, LY or LFH, depending on the choice of register.
+ def LMux : UnaryRXYPseudo<"l", load, GRX32, 4>,
+ Requires<[FeatureHighWord]>;
defm L : UnaryRXPair<"l", 0x58, 0xE358, load, GR32, 4>;
def LFH : UnaryRXY<"lfh", 0xE3CA, load, GRH32, 4>,
Requires<[FeatureHighWord]>;
// Register stores.
let SimpleBDXStore = 1 in {
+ // Expands to ST, STY or STFH, depending on the choice of register.
+ def STMux : StoreRXYPseudo<store, GRX32, 4>,
+ Requires<[FeatureHighWord]>;
defm ST : StoreRXPair<"st", 0x50, 0xE350, store, GR32, 4>;
def STFH : StoreRXY<"stfh", 0xE3CB, store, GRH32, 4>,
Requires<[FeatureHighWord]>;
defm MVC : MemorySS<"mvc", 0xD2, z_mvc, z_mvc_loop>;
// String moves.
-let mayLoad = 1, mayStore = 1, Defs = [CC], Uses = [R0L] in
+let mayLoad = 1, mayStore = 1, Defs = [CC] in
defm MVST : StringRRE<"mvst", 0xB255, z_stpcpy>;
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// 32-bit extensions from registers.
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
def LBR : UnaryRRE<"lb", 0xB926, sext8, GR32, GR32>;
def LHR : UnaryRRE<"lh", 0xB927, sext16, GR32, GR32>;
}
// 64-bit extensions from registers.
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
def LGBR : UnaryRRE<"lgb", 0xB906, sext8, GR64, GR64>;
def LGHR : UnaryRRE<"lgh", 0xB907, sext16, GR64, GR64>;
def LGFR : UnaryRRE<"lgf", 0xB914, sext32, GR64, GR32>;
}
let Defs = [CC], CCValues = 0xE, CompareZeroCCMask = 0xE in
- def LTGFR : UnaryRRE<"ltgf", 0xB912, null_frag, GR64, GR64>;
+ def LTGFR : UnaryRRE<"ltgf", 0xB912, null_frag, GR64, GR32>;
// Match 32-to-64-bit sign extensions in which the source is already
// in a 64-bit register.
def : Pat<(sext_inreg GR64:$src, i32),
(LGFR (EXTRACT_SUBREG GR64:$src, subreg_l32))>;
-// 32-bit extensions from memory.
-def LB : UnaryRXY<"lb", 0xE376, asextloadi8, GR32, 1>;
+// 32-bit extensions from 8-bit memory. LBMux expands to LB or LBH,
+// depending on the choice of register.
+def LBMux : UnaryRXYPseudo<"lb", asextloadi8, GRX32, 1>,
+ Requires<[FeatureHighWord]>;
+def LB : UnaryRXY<"lb", 0xE376, asextloadi8, GR32, 1>;
+def LBH : UnaryRXY<"lbh", 0xE3C0, asextloadi8, GRH32, 1>,
+ Requires<[FeatureHighWord]>;
+
+// 32-bit extensions from 16-bit memory. LHMux expands to LH or LHH,
+// depending on the choice of register.
+def LHMux : UnaryRXYPseudo<"lh", asextloadi16, GRX32, 2>,
+ Requires<[FeatureHighWord]>;
defm LH : UnaryRXPair<"lh", 0x48, 0xE378, asextloadi16, GR32, 2>;
+def LHH : UnaryRXY<"lhh", 0xE3C4, asextloadi16, GRH32, 2>,
+ Requires<[FeatureHighWord]>;
def LHRL : UnaryRILPC<"lhrl", 0xC45, aligned_asextloadi16, GR32>;
// 64-bit extensions from memory.
//===----------------------------------------------------------------------===//
// 32-bit extensions from registers.
-let neverHasSideEffects = 1 in {
- def LLCR : UnaryRRE<"llc", 0xB994, zext8, GR32, GR32>;
- def LLHR : UnaryRRE<"llh", 0xB995, zext16, GR32, GR32>;
+let hasSideEffects = 0 in {
+ // Expands to LLCR or RISB[LH]G, depending on the choice of registers.
+ def LLCRMux : UnaryRRPseudo<"llc", zext8, GRX32, GRX32>,
+ Requires<[FeatureHighWord]>;
+ def LLCR : UnaryRRE<"llc", 0xB994, zext8, GR32, GR32>;
+ // Expands to LLHR or RISB[LH]G, depending on the choice of registers.
+ def LLHRMux : UnaryRRPseudo<"llh", zext16, GRX32, GRX32>,
+ Requires<[FeatureHighWord]>;
+ def LLHR : UnaryRRE<"llh", 0xB995, zext16, GR32, GR32>;
}
// 64-bit extensions from registers.
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
def LLGCR : UnaryRRE<"llgc", 0xB984, zext8, GR64, GR64>;
def LLGHR : UnaryRRE<"llgh", 0xB985, zext16, GR64, GR64>;
def LLGFR : UnaryRRE<"llgf", 0xB916, zext32, GR64, GR32>;
def : Pat<(and GR64:$src, 0xffffffff),
(LLGFR (EXTRACT_SUBREG GR64:$src, subreg_l32))>;
-// 32-bit extensions from memory.
-def LLC : UnaryRXY<"llc", 0xE394, azextloadi8, GR32, 1>;
+// 32-bit extensions from 8-bit memory. LLCMux expands to LLC or LLCH,
+// depending on the choice of register.
+def LLCMux : UnaryRXYPseudo<"llc", azextloadi8, GRX32, 1>,
+ Requires<[FeatureHighWord]>;
+def LLC : UnaryRXY<"llc", 0xE394, azextloadi8, GR32, 1>;
+def LLCH : UnaryRXY<"llch", 0xE3C2, azextloadi8, GRH32, 1>,
+ Requires<[FeatureHighWord]>;
+
+// 32-bit extensions from 16-bit memory. LLHMux expands to LLH or LLHH,
+// depending on the choice of register.
+def LLHMux : UnaryRXYPseudo<"llh", azextloadi16, GRX32, 2>,
+ Requires<[FeatureHighWord]>;
def LLH : UnaryRXY<"llh", 0xE395, azextloadi16, GR32, 2>;
+def LLHH : UnaryRXY<"llhh", 0xE3C6, azextloadi16, GRH32, 2>,
+ Requires<[FeatureHighWord]>;
def LLHRL : UnaryRILPC<"llhrl", 0xC42, aligned_azextloadi16, GR32>;
// 64-bit extensions from memory.
def : Pat<(i32 (trunc GR64:$src)),
(EXTRACT_SUBREG GR64:$src, subreg_l32)>;
-// Truncations of 32-bit registers to memory.
-defm STC : StoreRXPair<"stc", 0x42, 0xE372, truncstorei8, GR32, 1>;
-defm STH : StoreRXPair<"sth", 0x40, 0xE370, truncstorei16, GR32, 2>;
-def STHRL : StoreRILPC<"sthrl", 0xC47, aligned_truncstorei16, GR32>;
+// Truncations of 32-bit registers to 8-bit memory. STCMux expands to
+// STC, STCY or STCH, depending on the choice of register.
+def STCMux : StoreRXYPseudo<truncstorei8, GRX32, 1>,
+ Requires<[FeatureHighWord]>;
+defm STC : StoreRXPair<"stc", 0x42, 0xE372, truncstorei8, GR32, 1>;
+def STCH : StoreRXY<"stch", 0xE3C3, truncstorei8, GRH32, 1>,
+ Requires<[FeatureHighWord]>;
+
+// Truncations of 32-bit registers to 16-bit memory. STHMux expands to
+// STH, STHY or STHH, depending on the choice of register.
+def STHMux : StoreRXYPseudo<truncstorei16, GRX32, 1>,
+ Requires<[FeatureHighWord]>;
+defm STH : StoreRXPair<"sth", 0x40, 0xE370, truncstorei16, GR32, 2>;
+def STHH : StoreRXY<"sthh", 0xE3C7, truncstorei16, GRH32, 2>,
+ Requires<[FeatureHighWord]>;
+def STHRL : StoreRILPC<"sthrl", 0xC47, aligned_truncstorei16, GR32>;
// Truncations of 64-bit registers to memory.
defm : StoreGR64Pair<STC, STCY, truncstorei8>;
//===----------------------------------------------------------------------===//
// Byte-swapping register moves.
-let neverHasSideEffects = 1 in {
+let hasSideEffects = 0 in {
def LRVR : UnaryRRE<"lrv", 0xB91F, bswap, GR32, GR32>;
def LRVGR : UnaryRRE<"lrvg", 0xB90F, bswap, GR64, GR64>;
}
//===----------------------------------------------------------------------===//
// Load BDX-style addresses.
-let neverHasSideEffects = 1, isAsCheapAsAMove = 1, isReMaterializable = 1,
+let hasSideEffects = 0, isAsCheapAsAMove = 1, isReMaterializable = 1,
DispKey = "la" in {
let DispSize = "12" in
def LA : InstRX<0x41, (outs GR64:$R1), (ins laaddr12pair:$XBD2),
// Load a PC-relative address. There's no version of this instruction
// with a 16-bit offset, so there's no relaxation.
-let neverHasSideEffects = 1, isAsCheapAsAMove = 1, isMoveImm = 1,
+let hasSideEffects = 0, isAsCheapAsAMove = 1, isMoveImm = 1,
isReMaterializable = 1 in {
def LARL : InstRIL<0xC00, (outs GR64:$R1), (ins pcrel32:$I2),
"larl\t$R1, $I2",
[(set GR64:$R1, pcrel32:$I2)]>;
}
+// Load the Global Offset Table address. This will be lowered into a
+// larl $R1, _GLOBAL_OFFSET_TABLE_
+// instruction.
+def GOT : Alias<6, (outs GR64:$R1), (ins),
+ [(set GR64:$R1, (global_offset_table))]>;
+
//===----------------------------------------------------------------------===//
// Absolute and Negation
//===----------------------------------------------------------------------===//
let Defs = [CC] in {
let CCValues = 0xF, CompareZeroCCMask = 0x8 in {
- def LPR : UnaryRR <"lp", 0x10, z_iabs32, GR32, GR32>;
- def LPGR : UnaryRRE<"lpg", 0xB900, z_iabs64, GR64, GR64>;
+ def LPR : UnaryRR <"lp", 0x10, z_iabs, GR32, GR32>;
+ def LPGR : UnaryRRE<"lpg", 0xB900, z_iabs, GR64, GR64>;
}
let CCValues = 0xE, CompareZeroCCMask = 0xE in
def LPGFR : UnaryRRE<"lpgf", 0xB910, null_frag, GR64, GR32>;
}
+def : Pat<(z_iabs32 GR32:$src), (LPR GR32:$src)>;
+def : Pat<(z_iabs64 GR64:$src), (LPGR GR64:$src)>;
+defm : SXU<z_iabs, LPGFR>;
defm : SXU<z_iabs64, LPGFR>;
let Defs = [CC] in {
let CCValues = 0xF, CompareZeroCCMask = 0x8 in {
- def LNR : UnaryRR <"ln", 0x11, z_inegabs32, GR32, GR32>;
- def LNGR : UnaryRRE<"lng", 0xB901, z_inegabs64, GR64, GR64>;
+ def LNR : UnaryRR <"ln", 0x11, z_inegabs, GR32, GR32>;
+ def LNGR : UnaryRRE<"lng", 0xB901, z_inegabs, GR64, GR64>;
}
let CCValues = 0xE, CompareZeroCCMask = 0xE in
def LNGFR : UnaryRRE<"lngf", 0xB911, null_frag, GR64, GR32>;
}
+def : Pat<(z_inegabs32 GR32:$src), (LNR GR32:$src)>;
+def : Pat<(z_inegabs64 GR64:$src), (LNGR GR64:$src)>;
+defm : SXU<z_inegabs, LNGFR>;
defm : SXU<z_inegabs64, LNGFR>;
let Defs = [CC] in {
// Insertions of a 16-bit immediate, leaving other bits unaffected.
// We don't have or_as_insert equivalents of these operations because
// OI is available instead.
+//
+// IIxMux expands to II[LH]x, depending on the choice of register.
+def IILMux : BinaryRIPseudo<insertll, GRX32, imm32ll16>,
+ Requires<[FeatureHighWord]>;
+def IIHMux : BinaryRIPseudo<insertlh, GRX32, imm32lh16>,
+ Requires<[FeatureHighWord]>;
def IILL : BinaryRI<"iill", 0xA53, insertll, GR32, imm32ll16>;
def IILH : BinaryRI<"iilh", 0xA52, insertlh, GR32, imm32lh16>;
+def IIHL : BinaryRI<"iihl", 0xA51, insertll, GRH32, imm32ll16>;
+def IIHH : BinaryRI<"iihh", 0xA50, insertlh, GRH32, imm32lh16>;
def IILL64 : BinaryAliasRI<insertll, GR64, imm64ll16>;
def IILH64 : BinaryAliasRI<insertlh, GR64, imm64lh16>;
-def IIHL : BinaryRI<"iihl", 0xA51, inserthl, GR64, imm64hl16>;
-def IIHH : BinaryRI<"iihh", 0xA50, inserthh, GR64, imm64hh16>;
+def IIHL64 : BinaryAliasRI<inserthl, GR64, imm64hl16>;
+def IIHH64 : BinaryAliasRI<inserthh, GR64, imm64hh16>;
// ...likewise for 32-bit immediates. For GR32s this is a general
// full-width move. (We use IILF rather than something like LLILF
// for 32-bit moves because IILF leaves the upper 32 bits of the
// GR64 unchanged.)
-let isAsCheapAsAMove = 1, isMoveImm = 1, isReMaterializable = 1 in
+let isAsCheapAsAMove = 1, isMoveImm = 1, isReMaterializable = 1 in {
+ def IIFMux : UnaryRIPseudo<bitconvert, GRX32, uimm32>,
+ Requires<[FeatureHighWord]>;
def IILF : UnaryRIL<"iilf", 0xC09, bitconvert, GR32, uimm32>;
+ def IIHF : UnaryRIL<"iihf", 0xC08, bitconvert, GRH32, uimm32>;
+}
def IILF64 : BinaryAliasRIL<insertlf, GR64, imm64lf32>;
-def IIHF : BinaryRIL<"iihf", 0xC08, inserthf, GR64, imm64hf32>;
+def IIHF64 : BinaryAliasRIL<inserthf, GR64, imm64hf32>;
// An alternative model of inserthf, with the first operand being
// a zero-extended value.
def : Pat<(or (zext32 GR32:$src), imm64hf32:$imm),
- (IIHF (INSERT_SUBREG (i64 (IMPLICIT_DEF)), GR32:$src, subreg_l32),
- imm64hf32:$imm)>;
+ (IIHF64 (INSERT_SUBREG (i64 (IMPLICIT_DEF)), GR32:$src, subreg_l32),
+ imm64hf32:$imm)>;
//===----------------------------------------------------------------------===//
// Addition
def AGFR : BinaryRRE<"agf", 0xB918, null_frag, GR64, GR32>;
// Addition of signed 16-bit immediates.
+ defm AHIMux : BinaryRIAndKPseudo<"ahimux", add, GRX32, imm32sx16>;
defm AHI : BinaryRIAndK<"ahi", 0xA7A, 0xECD8, add, GR32, imm32sx16>;
defm AGHI : BinaryRIAndK<"aghi", 0xA7B, 0xECD9, add, GR64, imm64sx16>;
// Addition of signed 32-bit immediates.
+ def AFIMux : BinaryRIPseudo<add, GRX32, simm32>,
+ Requires<[FeatureHighWord]>;
def AFI : BinaryRIL<"afi", 0xC29, add, GR32, simm32>;
+ def AIH : BinaryRIL<"aih", 0xCC8, add, GRH32, simm32>,
+ Requires<[FeatureHighWord]>;
def AGFI : BinaryRIL<"agfi", 0xC28, add, GR64, imm64sx32>;
// Addition of memory.
// Subtraction
//===----------------------------------------------------------------------===//
-// Plain substraction. Although immediate forms exist, we use the
+// Plain subtraction. Although immediate forms exist, we use the
// add-immediate instruction instead.
let Defs = [CC], CCValues = 0xF, CompareZeroCCMask = 0x8 in {
// Subtraction of a register.
let isConvertibleToThreeAddress = 1 in {
// ANDs of a 16-bit immediate, leaving other bits unaffected.
// The CC result only reflects the 16-bit field, not the full register.
+ //
+ // NIxMux expands to NI[LH]x, depending on the choice of register.
+ def NILMux : BinaryRIPseudo<and, GRX32, imm32ll16c>,
+ Requires<[FeatureHighWord]>;
+ def NIHMux : BinaryRIPseudo<and, GRX32, imm32lh16c>,
+ Requires<[FeatureHighWord]>;
def NILL : BinaryRI<"nill", 0xA57, and, GR32, imm32ll16c>;
def NILH : BinaryRI<"nilh", 0xA56, and, GR32, imm32lh16c>;
+ def NIHL : BinaryRI<"nihl", 0xA55, and, GRH32, imm32ll16c>;
+ def NIHH : BinaryRI<"nihh", 0xA54, and, GRH32, imm32lh16c>;
def NILL64 : BinaryAliasRI<and, GR64, imm64ll16c>;
def NILH64 : BinaryAliasRI<and, GR64, imm64lh16c>;
- def NIHL : BinaryRI<"nihl", 0xA55, and, GR64, imm64hl16c>;
- def NIHH : BinaryRI<"nihh", 0xA54, and, GR64, imm64hh16c>;
+ def NIHL64 : BinaryAliasRI<and, GR64, imm64hl16c>;
+ def NIHH64 : BinaryAliasRI<and, GR64, imm64hh16c>;
// ANDs of a 32-bit immediate, leaving other bits unaffected.
// The CC result only reflects the 32-bit field, which means we can
// use it as a zero indicator for i32 operations but not otherwise.
- let CCValues = 0xC, CompareZeroCCMask = 0x8 in
+ let CCValues = 0xC, CompareZeroCCMask = 0x8 in {
+ // Expands to NILF or NIHF, depending on the choice of register.
+ def NIFMux : BinaryRIPseudo<and, GRX32, uimm32>,
+ Requires<[FeatureHighWord]>;
def NILF : BinaryRIL<"nilf", 0xC0B, and, GR32, uimm32>;
+ def NIHF : BinaryRIL<"nihf", 0xC0A, and, GRH32, uimm32>;
+ }
def NILF64 : BinaryAliasRIL<and, GR64, imm64lf32c>;
- def NIHF : BinaryRIL<"nihf", 0xC0A, and, GR64, imm64hf32c>;
+ def NIHF64 : BinaryAliasRIL<and, GR64, imm64hf32c>;
}
// ANDs of memory.
}
// AND to memory
- defm NI : BinarySIPair<"ni", 0x94, 0xEB54, null_frag, uimm8>;
+ defm NI : BinarySIPair<"ni", 0x94, 0xEB54, null_frag, imm32zx8>;
// Block AND.
let mayLoad = 1, mayStore = 1 in
// ORs of a 16-bit immediate, leaving other bits unaffected.
// The CC result only reflects the 16-bit field, not the full register.
+ //
+ // OIxMux expands to OI[LH]x, depending on the choice of register.
+ def OILMux : BinaryRIPseudo<or, GRX32, imm32ll16>,
+ Requires<[FeatureHighWord]>;
+ def OIHMux : BinaryRIPseudo<or, GRX32, imm32lh16>,
+ Requires<[FeatureHighWord]>;
def OILL : BinaryRI<"oill", 0xA5B, or, GR32, imm32ll16>;
def OILH : BinaryRI<"oilh", 0xA5A, or, GR32, imm32lh16>;
+ def OIHL : BinaryRI<"oihl", 0xA59, or, GRH32, imm32ll16>;
+ def OIHH : BinaryRI<"oihh", 0xA58, or, GRH32, imm32lh16>;
def OILL64 : BinaryAliasRI<or, GR64, imm64ll16>;
def OILH64 : BinaryAliasRI<or, GR64, imm64lh16>;
- def OIHL : BinaryRI<"oihl", 0xA59, or, GR64, imm64hl16>;
- def OIHH : BinaryRI<"oihh", 0xA58, or, GR64, imm64hh16>;
+ def OIHL64 : BinaryAliasRI<or, GR64, imm64hl16>;
+ def OIHH64 : BinaryAliasRI<or, GR64, imm64hh16>;
// ORs of a 32-bit immediate, leaving other bits unaffected.
// The CC result only reflects the 32-bit field, which means we can
// use it as a zero indicator for i32 operations but not otherwise.
- let CCValues = 0xC, CompareZeroCCMask = 0x8 in
+ let CCValues = 0xC, CompareZeroCCMask = 0x8 in {
+ // Expands to OILF or OIHF, depending on the choice of register.
+ def OIFMux : BinaryRIPseudo<or, GRX32, uimm32>,
+ Requires<[FeatureHighWord]>;
def OILF : BinaryRIL<"oilf", 0xC0D, or, GR32, uimm32>;
+ def OIHF : BinaryRIL<"oihf", 0xC0C, or, GRH32, uimm32>;
+ }
def OILF64 : BinaryAliasRIL<or, GR64, imm64lf32>;
- def OIHF : BinaryRIL<"oihf", 0xC0C, or, GR64, imm64hf32>;
+ def OIHF64 : BinaryAliasRIL<or, GR64, imm64hf32>;
// ORs of memory.
let CCValues = 0xC, CompareZeroCCMask = 0x8 in {
}
// OR to memory
- defm OI : BinarySIPair<"oi", 0x96, 0xEB56, null_frag, uimm8>;
+ defm OI : BinarySIPair<"oi", 0x96, 0xEB56, null_frag, imm32zx8>;
// Block OR.
let mayLoad = 1, mayStore = 1 in
// XORs of a 32-bit immediate, leaving other bits unaffected.
// The CC result only reflects the 32-bit field, which means we can
// use it as a zero indicator for i32 operations but not otherwise.
- let CCValues = 0xC, CompareZeroCCMask = 0x8 in
+ let CCValues = 0xC, CompareZeroCCMask = 0x8 in {
+ // Expands to XILF or XIHF, depending on the choice of register.
+ def XIFMux : BinaryRIPseudo<xor, GRX32, uimm32>,
+ Requires<[FeatureHighWord]>;
def XILF : BinaryRIL<"xilf", 0xC07, xor, GR32, uimm32>;
+ def XIHF : BinaryRIL<"xihf", 0xC06, xor, GRH32, uimm32>;
+ }
def XILF64 : BinaryAliasRIL<xor, GR64, imm64lf32>;
- def XIHF : BinaryRIL<"xihf", 0xC06, xor, GR64, imm64hf32>;
+ def XIHF64 : BinaryAliasRIL<xor, GR64, imm64hf32>;
// XORs of memory.
let CCValues = 0xC, CompareZeroCCMask = 0x8 in {
}
// XOR to memory
- defm XI : BinarySIPair<"xi", 0x97, 0xEB57, null_frag, uimm8>;
+ defm XI : BinarySIPair<"xi", 0x97, 0xEB57, null_frag, imm32zx8>;
// Block XOR.
let mayLoad = 1, mayStore = 1 in
//===----------------------------------------------------------------------===//
// Shift left.
-let neverHasSideEffects = 1 in {
- defm SLL : ShiftRSAndK<"sll", 0x89, 0xEBDF, shl, GR32>;
- def SLLG : ShiftRSY<"sllg", 0xEB0D, shl, GR64>;
+let hasSideEffects = 0 in {
+ defm SLL : BinaryRSAndK<"sll", 0x89, 0xEBDF, shl, GR32>;
+ def SLLG : BinaryRSY<"sllg", 0xEB0D, shl, GR64>;
}
// Logical shift right.
-let neverHasSideEffects = 1 in {
- defm SRL : ShiftRSAndK<"srl", 0x88, 0xEBDE, srl, GR32>;
- def SRLG : ShiftRSY<"srlg", 0xEB0C, srl, GR64>;
+let hasSideEffects = 0 in {
+ defm SRL : BinaryRSAndK<"srl", 0x88, 0xEBDE, srl, GR32>;
+ def SRLG : BinaryRSY<"srlg", 0xEB0C, srl, GR64>;
}
// Arithmetic shift right.
let Defs = [CC], CCValues = 0xE, CompareZeroCCMask = 0xE in {
- defm SRA : ShiftRSAndK<"sra", 0x8A, 0xEBDC, sra, GR32>;
- def SRAG : ShiftRSY<"srag", 0xEB0A, sra, GR64>;
+ defm SRA : BinaryRSAndK<"sra", 0x8A, 0xEBDC, sra, GR32>;
+ def SRAG : BinaryRSY<"srag", 0xEB0A, sra, GR64>;
}
// Rotate left.
-let neverHasSideEffects = 1 in {
- def RLL : ShiftRSY<"rll", 0xEB1D, rotl, GR32>;
- def RLLG : ShiftRSY<"rllg", 0xEB1C, rotl, GR64>;
+let hasSideEffects = 0 in {
+ def RLL : BinaryRSY<"rll", 0xEB1D, rotl, GR32>;
+ def RLLG : BinaryRSY<"rllg", 0xEB1C, rotl, GR64>;
}
// Rotate second operand left and inserted selected bits into first operand.
def RISBG : RotateSelectRIEf<"risbg", 0xEC55, GR64, GR64>;
}
+// On zEC12 we have a variant of RISBG that does not set CC.
+let Predicates = [FeatureMiscellaneousExtensions] in
+ def RISBGN : RotateSelectRIEf<"risbgn", 0xEC59, GR64, GR64>;
+
// Forms of RISBG that only affect one word of the destination register.
// They do not set CC.
-let isCodeGenOnly = 1 in
- def RISBLG32 : RotateSelectRIEf<"risblg", 0xEC51, GR32, GR32>,
- Requires<[FeatureHighWord]>;
-def RISBHG : RotateSelectRIEf<"risbhg", 0xEC5D, GRH32, GR64>,
- Requires<[FeatureHighWord]>;
-def RISBLG : RotateSelectRIEf<"risblg", 0xEC51, GR32, GR64>,
- Requires<[FeatureHighWord]>;
+let Predicates = [FeatureHighWord] in {
+ def RISBMux : RotateSelectRIEfPseudo<GRX32, GRX32>;
+ def RISBLL : RotateSelectAliasRIEf<GR32, GR32>;
+ def RISBLH : RotateSelectAliasRIEf<GR32, GRH32>;
+ def RISBHL : RotateSelectAliasRIEf<GRH32, GR32>;
+ def RISBHH : RotateSelectAliasRIEf<GRH32, GRH32>;
+ def RISBLG : RotateSelectRIEf<"risblg", 0xEC51, GR32, GR64>;
+ def RISBHG : RotateSelectRIEf<"risbhg", 0xEC5D, GRH32, GR64>;
+}
// Rotate second operand left and perform a logical operation with selected
// bits of the first operand. The CC result only describes the selected bits,
def CHI : CompareRI<"chi", 0xA7E, z_scmp, GR32, imm32sx16>;
def CGHI : CompareRI<"cghi", 0xA7F, z_scmp, GR64, imm64sx16>;
- // Comparison with a signed 32-bit immediate.
+ // Comparison with a signed 32-bit immediate. CFIMux expands to CFI or CIH,
+ // depending on the choice of register.
+ def CFIMux : CompareRIPseudo<z_scmp, GRX32, simm32>,
+ Requires<[FeatureHighWord]>;
def CFI : CompareRIL<"cfi", 0xC2D, z_scmp, GR32, simm32>;
+ def CIH : CompareRIL<"cih", 0xCCD, z_scmp, GRH32, simm32>,
+ Requires<[FeatureHighWord]>;
def CGFI : CompareRIL<"cgfi", 0xC2C, z_scmp, GR64, imm64sx32>;
// Comparison with memory.
defm CH : CompareRXPair<"ch", 0x49, 0xE379, z_scmp, GR32, asextloadi16, 2>;
+ def CMux : CompareRXYPseudo<z_scmp, GRX32, load, 4>,
+ Requires<[FeatureHighWord]>;
defm C : CompareRXPair<"c", 0x59, 0xE359, z_scmp, GR32, load, 4>;
+ def CHF : CompareRXY<"chf", 0xE3CD, z_scmp, GRH32, load, 4>,
+ Requires<[FeatureHighWord]>;
def CGH : CompareRXY<"cgh", 0xE334, z_scmp, GR64, asextloadi16, 2>;
def CGF : CompareRXY<"cgf", 0xE330, z_scmp, GR64, asextloadi32, 4>;
def CG : CompareRXY<"cg", 0xE320, z_scmp, GR64, load, 8>;
def CLGFR : CompareRRE<"clgf", 0xB931, null_frag, GR64, GR32>;
def CLGR : CompareRRE<"clg", 0xB921, z_ucmp, GR64, GR64>;
- // Comparison with a signed 32-bit immediate.
+ // Comparison with an unsigned 32-bit immediate. CLFIMux expands to CLFI
+ // or CLIH, depending on the choice of register.
+ def CLFIMux : CompareRIPseudo<z_ucmp, GRX32, uimm32>,
+ Requires<[FeatureHighWord]>;
def CLFI : CompareRIL<"clfi", 0xC2F, z_ucmp, GR32, uimm32>;
+ def CLIH : CompareRIL<"clih", 0xCCF, z_ucmp, GRH32, uimm32>,
+ Requires<[FeatureHighWord]>;
def CLGFI : CompareRIL<"clgfi", 0xC2E, z_ucmp, GR64, imm64zx32>;
// Comparison with memory.
+ def CLMux : CompareRXYPseudo<z_ucmp, GRX32, load, 4>,
+ Requires<[FeatureHighWord]>;
defm CL : CompareRXPair<"cl", 0x55, 0xE355, z_ucmp, GR32, load, 4>;
+ def CLHF : CompareRXY<"clhf", 0xE3CF, z_ucmp, GRH32, load, 4>,
+ Requires<[FeatureHighWord]>;
def CLGF : CompareRXY<"clgf", 0xE331, z_ucmp, GR64, azextloadi32, 4>;
def CLG : CompareRXY<"clg", 0xE321, z_ucmp, GR64, load, 8>;
def CLHRL : CompareRILPC<"clhrl", 0xC67, z_ucmp, GR32,
defm CLC : MemorySS<"clc", 0xD5, z_clc, z_clc_loop>;
// String comparison.
-let mayLoad = 1, Defs = [CC], Uses = [R0L] in
+let mayLoad = 1, Defs = [CC] in
defm CLST : StringRRE<"clst", 0xB25D, z_strcmp>;
// Test under mask.
let Defs = [CC] in {
+ // TMxMux expands to TM[LH]x, depending on the choice of register.
+ def TMLMux : CompareRIPseudo<z_tm_reg, GRX32, imm32ll16>,
+ Requires<[FeatureHighWord]>;
+ def TMHMux : CompareRIPseudo<z_tm_reg, GRX32, imm32lh16>,
+ Requires<[FeatureHighWord]>;
def TMLL : CompareRI<"tmll", 0xA71, z_tm_reg, GR32, imm32ll16>;
def TMLH : CompareRI<"tmlh", 0xA70, z_tm_reg, GR32, imm32lh16>;
+ def TMHL : CompareRI<"tmhl", 0xA73, z_tm_reg, GRH32, imm32ll16>;
+ def TMHH : CompareRI<"tmhh", 0xA72, z_tm_reg, GRH32, imm32lh16>;
- def TMHL : CompareRI<"tmhl", 0xA73, z_tm_reg, GR64, imm64hl16>;
- def TMHH : CompareRI<"tmhh", 0xA72, z_tm_reg, GR64, imm64hh16>;
+ def TMLL64 : CompareAliasRI<z_tm_reg, GR64, imm64ll16>;
+ def TMLH64 : CompareAliasRI<z_tm_reg, GR64, imm64lh16>;
+ def TMHL64 : CompareAliasRI<z_tm_reg, GR64, imm64hl16>;
+ def TMHH64 : CompareAliasRI<z_tm_reg, GR64, imm64hh16>;
defm TM : CompareSIPair<"tm", 0x91, 0xEB51, z_tm_mem, anyextloadi8, imm32zx8>;
}
-def : CompareGR64RI<TMLL, z_tm_reg, imm64ll16>;
-def : CompareGR64RI<TMLH, z_tm_reg, imm64lh16>;
//===----------------------------------------------------------------------===//
// Prefetch
// Atomic operations
//===----------------------------------------------------------------------===//
-def ATOMIC_SWAPW : AtomicLoadWBinaryReg<z_atomic_swapw>;
-def ATOMIC_SWAP_32 : AtomicLoadBinaryReg32<atomic_swap_32>;
-def ATOMIC_SWAP_64 : AtomicLoadBinaryReg64<atomic_swap_64>;
-
-def ATOMIC_LOADW_AR : AtomicLoadWBinaryReg<z_atomic_loadw_add>;
-def ATOMIC_LOADW_AFI : AtomicLoadWBinaryImm<z_atomic_loadw_add, simm32>;
-def ATOMIC_LOAD_AR : AtomicLoadBinaryReg32<atomic_load_add_32>;
-def ATOMIC_LOAD_AHI : AtomicLoadBinaryImm32<atomic_load_add_32, imm32sx16>;
-def ATOMIC_LOAD_AFI : AtomicLoadBinaryImm32<atomic_load_add_32, simm32>;
-def ATOMIC_LOAD_AGR : AtomicLoadBinaryReg64<atomic_load_add_64>;
-def ATOMIC_LOAD_AGHI : AtomicLoadBinaryImm64<atomic_load_add_64, imm64sx16>;
-def ATOMIC_LOAD_AGFI : AtomicLoadBinaryImm64<atomic_load_add_64, imm64sx32>;
-
-def ATOMIC_LOADW_SR : AtomicLoadWBinaryReg<z_atomic_loadw_sub>;
-def ATOMIC_LOAD_SR : AtomicLoadBinaryReg32<atomic_load_sub_32>;
-def ATOMIC_LOAD_SGR : AtomicLoadBinaryReg64<atomic_load_sub_64>;
-
-def ATOMIC_LOADW_NR : AtomicLoadWBinaryReg<z_atomic_loadw_and>;
-def ATOMIC_LOADW_NILH : AtomicLoadWBinaryImm<z_atomic_loadw_and, imm32lh16c>;
-def ATOMIC_LOAD_NR : AtomicLoadBinaryReg32<atomic_load_and_32>;
-def ATOMIC_LOAD_NILL : AtomicLoadBinaryImm32<atomic_load_and_32, imm32ll16c>;
-def ATOMIC_LOAD_NILH : AtomicLoadBinaryImm32<atomic_load_and_32, imm32lh16c>;
-def ATOMIC_LOAD_NILF : AtomicLoadBinaryImm32<atomic_load_and_32, uimm32>;
-def ATOMIC_LOAD_NGR : AtomicLoadBinaryReg64<atomic_load_and_64>;
-def ATOMIC_LOAD_NILL64 : AtomicLoadBinaryImm64<atomic_load_and_64, imm64ll16c>;
-def ATOMIC_LOAD_NILH64 : AtomicLoadBinaryImm64<atomic_load_and_64, imm64lh16c>;
-def ATOMIC_LOAD_NIHL : AtomicLoadBinaryImm64<atomic_load_and_64, imm64hl16c>;
-def ATOMIC_LOAD_NIHH : AtomicLoadBinaryImm64<atomic_load_and_64, imm64hh16c>;
-def ATOMIC_LOAD_NILF64 : AtomicLoadBinaryImm64<atomic_load_and_64, imm64lf32c>;
-def ATOMIC_LOAD_NIHF : AtomicLoadBinaryImm64<atomic_load_and_64, imm64hf32c>;
+// A serialization instruction that acts as a barrier for all memory
+// accesses, which expands to "bcr 14, 0".
+let hasSideEffects = 1 in
+def Serialize : Alias<2, (outs), (ins), [(z_serialize)]>;
+
+let Predicates = [FeatureInterlockedAccess1], Defs = [CC] in {
+ def LAA : LoadAndOpRSY<"laa", 0xEBF8, atomic_load_add_32, GR32>;
+ def LAAG : LoadAndOpRSY<"laag", 0xEBE8, atomic_load_add_64, GR64>;
+ def LAAL : LoadAndOpRSY<"laal", 0xEBFA, null_frag, GR32>;
+ def LAALG : LoadAndOpRSY<"laalg", 0xEBEA, null_frag, GR64>;
+ def LAN : LoadAndOpRSY<"lan", 0xEBF4, atomic_load_and_32, GR32>;
+ def LANG : LoadAndOpRSY<"lang", 0xEBE4, atomic_load_and_64, GR64>;
+ def LAO : LoadAndOpRSY<"lao", 0xEBF6, atomic_load_or_32, GR32>;
+ def LAOG : LoadAndOpRSY<"laog", 0xEBE6, atomic_load_or_64, GR64>;
+ def LAX : LoadAndOpRSY<"lax", 0xEBF7, atomic_load_xor_32, GR32>;
+ def LAXG : LoadAndOpRSY<"laxg", 0xEBE7, atomic_load_xor_64, GR64>;
+}
+
+def ATOMIC_SWAPW : AtomicLoadWBinaryReg<z_atomic_swapw>;
+def ATOMIC_SWAP_32 : AtomicLoadBinaryReg32<atomic_swap_32>;
+def ATOMIC_SWAP_64 : AtomicLoadBinaryReg64<atomic_swap_64>;
+
+def ATOMIC_LOADW_AR : AtomicLoadWBinaryReg<z_atomic_loadw_add>;
+def ATOMIC_LOADW_AFI : AtomicLoadWBinaryImm<z_atomic_loadw_add, simm32>;
+let Predicates = [FeatureNoInterlockedAccess1] in {
+ def ATOMIC_LOAD_AR : AtomicLoadBinaryReg32<atomic_load_add_32>;
+ def ATOMIC_LOAD_AHI : AtomicLoadBinaryImm32<atomic_load_add_32, imm32sx16>;
+ def ATOMIC_LOAD_AFI : AtomicLoadBinaryImm32<atomic_load_add_32, simm32>;
+ def ATOMIC_LOAD_AGR : AtomicLoadBinaryReg64<atomic_load_add_64>;
+ def ATOMIC_LOAD_AGHI : AtomicLoadBinaryImm64<atomic_load_add_64, imm64sx16>;
+ def ATOMIC_LOAD_AGFI : AtomicLoadBinaryImm64<atomic_load_add_64, imm64sx32>;
+}
+
+def ATOMIC_LOADW_SR : AtomicLoadWBinaryReg<z_atomic_loadw_sub>;
+def ATOMIC_LOAD_SR : AtomicLoadBinaryReg32<atomic_load_sub_32>;
+def ATOMIC_LOAD_SGR : AtomicLoadBinaryReg64<atomic_load_sub_64>;
+
+def ATOMIC_LOADW_NR : AtomicLoadWBinaryReg<z_atomic_loadw_and>;
+def ATOMIC_LOADW_NILH : AtomicLoadWBinaryImm<z_atomic_loadw_and, imm32lh16c>;
+let Predicates = [FeatureNoInterlockedAccess1] in {
+ def ATOMIC_LOAD_NR : AtomicLoadBinaryReg32<atomic_load_and_32>;
+ def ATOMIC_LOAD_NILL : AtomicLoadBinaryImm32<atomic_load_and_32,
+ imm32ll16c>;
+ def ATOMIC_LOAD_NILH : AtomicLoadBinaryImm32<atomic_load_and_32,
+ imm32lh16c>;
+ def ATOMIC_LOAD_NILF : AtomicLoadBinaryImm32<atomic_load_and_32, uimm32>;
+ def ATOMIC_LOAD_NGR : AtomicLoadBinaryReg64<atomic_load_and_64>;
+ def ATOMIC_LOAD_NILL64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64ll16c>;
+ def ATOMIC_LOAD_NILH64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64lh16c>;
+ def ATOMIC_LOAD_NIHL64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64hl16c>;
+ def ATOMIC_LOAD_NIHH64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64hh16c>;
+ def ATOMIC_LOAD_NILF64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64lf32c>;
+ def ATOMIC_LOAD_NIHF64 : AtomicLoadBinaryImm64<atomic_load_and_64,
+ imm64hf32c>;
+}
def ATOMIC_LOADW_OR : AtomicLoadWBinaryReg<z_atomic_loadw_or>;
def ATOMIC_LOADW_OILH : AtomicLoadWBinaryImm<z_atomic_loadw_or, imm32lh16>;
-def ATOMIC_LOAD_OR : AtomicLoadBinaryReg32<atomic_load_or_32>;
-def ATOMIC_LOAD_OILL : AtomicLoadBinaryImm32<atomic_load_or_32, imm32ll16>;
-def ATOMIC_LOAD_OILH : AtomicLoadBinaryImm32<atomic_load_or_32, imm32lh16>;
-def ATOMIC_LOAD_OILF : AtomicLoadBinaryImm32<atomic_load_or_32, uimm32>;
-def ATOMIC_LOAD_OGR : AtomicLoadBinaryReg64<atomic_load_or_64>;
-def ATOMIC_LOAD_OILL64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64ll16>;
-def ATOMIC_LOAD_OILH64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64lh16>;
-def ATOMIC_LOAD_OIHL : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hl16>;
-def ATOMIC_LOAD_OIHH : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hh16>;
-def ATOMIC_LOAD_OILF64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64lf32>;
-def ATOMIC_LOAD_OIHF : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hf32>;
+let Predicates = [FeatureNoInterlockedAccess1] in {
+ def ATOMIC_LOAD_OR : AtomicLoadBinaryReg32<atomic_load_or_32>;
+ def ATOMIC_LOAD_OILL : AtomicLoadBinaryImm32<atomic_load_or_32, imm32ll16>;
+ def ATOMIC_LOAD_OILH : AtomicLoadBinaryImm32<atomic_load_or_32, imm32lh16>;
+ def ATOMIC_LOAD_OILF : AtomicLoadBinaryImm32<atomic_load_or_32, uimm32>;
+ def ATOMIC_LOAD_OGR : AtomicLoadBinaryReg64<atomic_load_or_64>;
+ def ATOMIC_LOAD_OILL64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64ll16>;
+ def ATOMIC_LOAD_OILH64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64lh16>;
+ def ATOMIC_LOAD_OIHL64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hl16>;
+ def ATOMIC_LOAD_OIHH64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hh16>;
+ def ATOMIC_LOAD_OILF64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64lf32>;
+ def ATOMIC_LOAD_OIHF64 : AtomicLoadBinaryImm64<atomic_load_or_64, imm64hf32>;
+}
def ATOMIC_LOADW_XR : AtomicLoadWBinaryReg<z_atomic_loadw_xor>;
def ATOMIC_LOADW_XILF : AtomicLoadWBinaryImm<z_atomic_loadw_xor, uimm32>;
-def ATOMIC_LOAD_XR : AtomicLoadBinaryReg32<atomic_load_xor_32>;
-def ATOMIC_LOAD_XILF : AtomicLoadBinaryImm32<atomic_load_xor_32, uimm32>;
-def ATOMIC_LOAD_XGR : AtomicLoadBinaryReg64<atomic_load_xor_64>;
-def ATOMIC_LOAD_XILF64 : AtomicLoadBinaryImm64<atomic_load_xor_64, imm64lf32>;
-def ATOMIC_LOAD_XIHF : AtomicLoadBinaryImm64<atomic_load_xor_64, imm64hf32>;
+let Predicates = [FeatureNoInterlockedAccess1] in {
+ def ATOMIC_LOAD_XR : AtomicLoadBinaryReg32<atomic_load_xor_32>;
+ def ATOMIC_LOAD_XILF : AtomicLoadBinaryImm32<atomic_load_xor_32, uimm32>;
+ def ATOMIC_LOAD_XGR : AtomicLoadBinaryReg64<atomic_load_xor_64>;
+ def ATOMIC_LOAD_XILF64 : AtomicLoadBinaryImm64<atomic_load_xor_64, imm64lf32>;
+ def ATOMIC_LOAD_XIHF64 : AtomicLoadBinaryImm64<atomic_load_xor_64, imm64hf32>;
+}
def ATOMIC_LOADW_NRi : AtomicLoadWBinaryReg<z_atomic_loadw_nand>;
def ATOMIC_LOADW_NILHi : AtomicLoadWBinaryImm<z_atomic_loadw_nand,
imm64ll16c>;
def ATOMIC_LOAD_NILH64i : AtomicLoadBinaryImm64<atomic_load_nand_64,
imm64lh16c>;
-def ATOMIC_LOAD_NIHL
i : AtomicLoadBinaryImm64<atomic_load_nand_64,
+def ATOMIC_LOAD_NIHL
64i : AtomicLoadBinaryImm64<atomic_load_nand_64,
imm64hl16c>;
-def ATOMIC_LOAD_NIHH
i : AtomicLoadBinaryImm64<atomic_load_nand_64,
+def ATOMIC_LOAD_NIHH
64i : AtomicLoadBinaryImm64<atomic_load_nand_64,
imm64hh16c>;
def ATOMIC_LOAD_NILF64i : AtomicLoadBinaryImm64<atomic_load_nand_64,
imm64lf32c>;
-def ATOMIC_LOAD_NIHF
i : AtomicLoadBinaryImm64<atomic_load_nand_64,
+def ATOMIC_LOAD_NIHF
64i : AtomicLoadBinaryImm64<atomic_load_nand_64,
imm64hf32c>;
def ATOMIC_LOADW_MIN : AtomicLoadWBinaryReg<z_atomic_loadw_min>;
def CSG : CmpSwapRSY<"csg", 0xEB30, atomic_cmp_swap_64, GR64>;
}
+//===----------------------------------------------------------------------===//
+// Transactional execution
+//===----------------------------------------------------------------------===//
+
+let Predicates = [FeatureTransactionalExecution] in {
+ // Transaction Begin
+ let hasSideEffects = 1, mayStore = 1,
+ usesCustomInserter = 1, Defs = [CC] in {
+ def TBEGIN : InstSIL<0xE560,
+ (outs), (ins bdaddr12only:$BD1, imm32zx16:$I2),
+ "tbegin\t$BD1, $I2",
+ [(z_tbegin bdaddr12only:$BD1, imm32zx16:$I2)]>;
+ def TBEGIN_nofloat : Pseudo<(outs), (ins bdaddr12only:$BD1, imm32zx16:$I2),
+ [(z_tbegin_nofloat bdaddr12only:$BD1,
+ imm32zx16:$I2)]>;
+ def TBEGINC : InstSIL<0xE561,
+ (outs), (ins bdaddr12only:$BD1, imm32zx16:$I2),
+ "tbeginc\t$BD1, $I2",
+ [(int_s390_tbeginc bdaddr12only:$BD1,
+ imm32zx16:$I2)]>;
+ }
+
+ // Transaction End
+ let hasSideEffects = 1, Defs = [CC], BD2 = 0 in
+ def TEND : InstS<0xB2F8, (outs), (ins), "tend", [(z_tend)]>;
+
+ // Transaction Abort
+ let hasSideEffects = 1, isTerminator = 1, isBarrier = 1 in
+ def TABORT : InstS<0xB2FC, (outs), (ins bdaddr12only:$BD2),
+ "tabort\t$BD2",
+ [(int_s390_tabort bdaddr12only:$BD2)]>;
+
+ // Nontransactional Store
+ let hasSideEffects = 1 in
+ def NTSTG : StoreRXY<"ntstg", 0xE325, int_s390_ntstg, GR64, 8>;
+
+ // Extract Transaction Nesting Depth
+ let hasSideEffects = 1 in
+ def ETND : InherentRRE<"etnd", 0xB2EC, GR32, (int_s390_etnd)>;
+}
+
+//===----------------------------------------------------------------------===//
+// Processor assist
+//===----------------------------------------------------------------------===//
+
+let Predicates = [FeatureProcessorAssist] in {
+ let hasSideEffects = 1, R4 = 0 in
+ def PPA : InstRRF<0xB2E8, (outs), (ins GR64:$R1, GR64:$R2, imm32zx4:$R3),
+ "ppa\t$R1, $R2, $R3", []>;
+ def : Pat<(int_s390_ppa_txassist GR32:$src),
+ (PPA (INSERT_SUBREG (i64 (IMPLICIT_DEF)), GR32:$src, subreg_l32),
+ 0, 1)>;
+}
+
//===----------------------------------------------------------------------===//
// Miscellaneous Instructions.
//===----------------------------------------------------------------------===//
def : Pat<(ctlz GR64:$src),
(EXTRACT_SUBREG (FLOGR GR64:$src), subreg_h64)>;
+// Population count. Counts bits set per byte.
+let Predicates = [FeaturePopulationCount], Defs = [CC] in {
+ def POPCNT : InstRRE<0xB9E1, (outs GR64:$R1), (ins GR64:$R2),
+ "popcnt\t$R1, $R2",
+ [(set GR64:$R1, (z_popcnt GR64:$R2))]>;
+}
+
// Use subregs to populate the "don't care" bits in a 32-bit to 64-bit anyext.
def : Pat<(i64 (anyext GR32:$src)),
(INSERT_SUBREG (i64 (IMPLICIT_DEF)), GR32:$src, subreg_l32)>;
}
// Search a block of memory for a character.
-let mayLoad = 1, Defs = [CC], Uses = [R0L] in
+let mayLoad = 1, Defs = [CC] in
defm SRST : StringRRE<"srst", 0xb25e, z_search_string>;
+// Other instructions for inline assembly
+let hasSideEffects = 1, Defs = [CC], mayStore = 1 in
+ def STCK : InstS<0xB205, (outs), (ins bdaddr12only:$BD2),
+ "stck\t$BD2",
+ []>;
+let hasSideEffects = 1, Defs = [CC], mayStore = 1 in
+ def STCKF : InstS<0xB27C, (outs), (ins bdaddr12only:$BD2),
+ "stckf\t$BD2",
+ []>;
+let hasSideEffects = 1, Defs = [CC], mayStore = 1 in
+ def STCKE : InstS<0xB278, (outs), (ins bdaddr12only:$BD2),
+ "stcke\t$BD2",
+ []>;
+let hasSideEffects = 1, Defs = [CC], mayStore = 1 in
+ def STFLE : InstS<0xB2B0, (outs), (ins bdaddr12only:$BD2),
+ "stfle\t$BD2",
+ []>;
+
+
+
//===----------------------------------------------------------------------===//
// Peepholes.
//===----------------------------------------------------------------------===//
// Optimize sign-extended 1/0 selects to -1/0 selects. This is important
// for vector legalization.
-def : Pat<(sra (shl (i32 (z_select_ccmask 1, 0, uimm8zx4:$valid, uimm8zx4:$cc)),
+def : Pat<(sra (shl (i32 (z_select_ccmask 1, 0, imm32zx4:$valid, imm32zx4:$cc)),
(i32 31)),
(i32 31)),
- (Select32 (LHI -1), (LHI 0), uimm8zx4:$valid, uimm8zx4:$cc)>;
-def : Pat<(sra (shl (i64 (anyext (i32 (z_select_ccmask 1, 0, uimm8zx4:$valid,
- uimm8zx4:$cc)))),
+ (Select32 (LHI -1), (LHI 0), imm32zx4:$valid, imm32zx4:$cc)>;
+def : Pat<(sra (shl (i64 (anyext (i32 (z_select_ccmask 1, 0, imm32zx4:$valid,
+ imm32zx4:$cc)))),
(i32 63)),
(i32 63)),
- (Select64 (LGHI -1), (LGHI 0), uimm8zx4:$valid, uimm8zx4:$cc)>;
+ (Select64 (LGHI -1), (LGHI 0), imm32zx4:$valid, imm32zx4:$cc)>;
// Peepholes for turning scalar operations into block operations.
defm : BlockLoadStore<anyextloadi8, i32, MVCSequence, NCSequence, OCSequence,
projects / oota-llvm.git / blobdiff