diff options
Diffstat (limited to 'lib/Target/ARM/ARMInstrNEON.td')
-rw-r--r-- | lib/Target/ARM/ARMInstrNEON.td | 28 |
1 files changed, 25 insertions, 3 deletions
diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td index c02bb3b..a0c627c 100644 --- a/lib/Target/ARM/ARMInstrNEON.td +++ b/lib/Target/ARM/ARMInstrNEON.td @@ -34,6 +34,14 @@ def nImmSplatI32 : Operand<i32> { let PrintMethod = "printNEONModImmOperand"; let ParserMatchClass = nImmSplatI32AsmOperand; } +def nImmSplatNotI16AsmOperand : AsmOperandClass { let Name = "NEONi16splatNot"; } +def nImmSplatNotI16 : Operand<i32> { + let ParserMatchClass = nImmSplatNotI16AsmOperand; +} +def nImmSplatNotI32AsmOperand : AsmOperandClass { let Name = "NEONi32splatNot"; } +def nImmSplatNotI32 : Operand<i32> { + let ParserMatchClass = nImmSplatNotI32AsmOperand; +} def nImmVMOVI32AsmOperand : AsmOperandClass { let Name = "NEONi32vmov"; } def nImmVMOVI32 : Operand<i32> { let PrintMethod = "printNEONModImmOperand"; @@ -4376,7 +4384,7 @@ defm VMLSLslu : N3VLMulOpSL_HS<1, 0b0110, "vmlsl", "u", NEONvmullu, sub>; // VQDMLSL : Vector Saturating Doubling Multiply Subtract Long (Q -= D * D) defm VQDMLSL : N3VLInt3_HS<0, 1, 0b1011, 0, IIC_VMACi16D, IIC_VMACi32D, "vqdmlsl", "s", null_frag>; -defm VQDMLSLsl: N3VLInt3SL_HS<0, 0b111, "vqdmlsl", "s", null_frag>; +defm VQDMLSLsl: N3VLInt3SL_HS<0, 0b0111, "vqdmlsl", "s", null_frag>; def : Pat<(v4i32 (int_arm_neon_vqsubs (v4i32 QPR:$src1), (v4i32 (int_arm_neon_vqdmull (v4i16 DPR:$Vn), @@ -5429,7 +5437,7 @@ def VGETLNi32 : NVGetLane<{1,1,1,0,0,0,?,1}, 0b1011, 0b00, IIC_VMOVSI, "vmov", "32", "$R, $V$lane", [(set GPR:$R, (extractelt (v2i32 DPR:$V), imm:$lane))]>, - Requires<[HasNEON, HasFastVGETLNi32]> { + Requires<[HasVFP2, HasFastVGETLNi32]> { let Inst{21} = lane{0}; } // def VGETLNf32: see FMRDH and FMRDL in ARMInstrVFP.td @@ -5497,8 +5505,12 @@ def VSETLNi32 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, 0b00, (outs DPR:$V), (ins DPR:$src1, GPR:$R, VectorIndex32:$lane), IIC_VMOVISL, "vmov", "32", "$V$lane, $R", [(set DPR:$V, (insertelt (v2i32 DPR:$src1), - GPR:$R, imm:$lane))]> { + GPR:$R, imm:$lane))]>, + Requires<[HasVFP2]> { let Inst{21} = lane{0}; + // This instruction is equivalent as + // $V = INSERT_SUBREG $src1, $R, translateImmToSubIdx($imm) + let isInsertSubreg = 1; } } def : Pat<(vector_insert (v16i8 QPR:$src1), GPR:$src2, imm:$lane), @@ -6635,6 +6647,16 @@ defm : NEONDTAnyInstAlias<"vorr${p}", "$Vdn, $Vm", (VORRd DPR:$Vdn, DPR:$Vdn, DPR:$Vm, pred:$p)>; defm : NEONDTAnyInstAlias<"vorr${p}", "$Vdn, $Vm", (VORRq QPR:$Vdn, QPR:$Vdn, QPR:$Vm, pred:$p)>; +// ... immediates +def : NEONInstAlias<"vand${p}.i16 $Vd, $imm", + (VBICiv4i16 DPR:$Vd, nImmSplatNotI16:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i32 $Vd, $imm", + (VBICiv2i32 DPR:$Vd, nImmSplatNotI32:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i16 $Vd, $imm", + (VBICiv8i16 QPR:$Vd, nImmSplatNotI16:$imm, pred:$p)>; +def : NEONInstAlias<"vand${p}.i32 $Vd, $imm", + (VBICiv4i32 QPR:$Vd, nImmSplatNotI32:$imm, pred:$p)>; + // VLD1 single-lane pseudo-instructions. These need special handling for // the lane index that an InstAlias can't handle, so we use these instead. |