maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/Target/Mips/Mips64r6InstrInfo.td

238 lines
11 KiB
TableGen
Raw Normal View History

[mips][mips64r6] Add experimental support for MIPS32r6 and MIPS64r6 Summary: Adds MIPS32r6/MIPS64r6 and checks the compatibility requirements for these processors. I've also included comments to describe removed and re-encoded instructions, along with placeholder def's for the new instructions but there are no functional changes to codegen at this point. Reviewers: jkolek, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3622 llvm-svn: 208399
2014-05-09 17:46:21 +08:00
//=- Mips64r6InstrInfo.td - Mips64r6 Instruction Information -*- 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 Mips64r6 instructions.
//
//===----------------------------------------------------------------------===//
// Notes about removals/changes from MIPS32r6:
// Reencoded: dclo, dclz
[mips][mips64r6] Added mul/mulu/muh/muhu Summary: The 'mul' line of the test is temporarily commented out because it currently matches the MIPS32 mul instead of the MIPS32r6 mul. This line will be uncommented when we disable the MIPS32 mul on MIPS32r6. Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3668 llvm-svn: 208576
2014-05-12 23:12:45 +08:00
//===----------------------------------------------------------------------===//
//
// Instruction Encodings
//
//===----------------------------------------------------------------------===//
[mips][mips64r6] Add align and dalign Summary: Depends on D3689 Reviewers: vmedic, zoran.jovanovic, jkolek Reviewed By: jkolek Differential Revision: http://reviews.llvm.org/D3728 llvm-svn: 208872
2014-05-15 20:06:36 +08:00
class DALIGN_ENC : SPECIAL3_DALIGN_FM<OPCODE6_DALIGN>;
[mips][mips64r6] Add aui, daui, dahi, and dati Summary: Depends on D3671 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3759 llvm-svn: 208857
2014-05-15 18:27:19 +08:00
class DAUI_ENC : DAUI_FM;
class DAHI_ENC : REGIMM_FM<OPCODE5_DAHI>;
class DATI_ENC : REGIMM_FM<OPCODE5_DATI>;
[mips][mips64r6] Add bitswap, and dbitswap Summary: Depends on D3728 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3729 llvm-svn: 208877
2014-05-15 20:18:23 +08:00
class DBITSWAP_ENC : SPECIAL3_2R_FM<OPCODE6_DBITSWAP>;
[mips][mips64r6] cl[oz], and dcl[oz] are re-encoded in MIPS32r6/MIPS64r6 Summary: There is no change to the restrictions, just the result register is stored once in the encoding rather than twice. The rt field is zero in MIPS32r6/MIPS64r6. Depends on D4119 Reviewers: zoran.jovanovic, jkolek, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4120 llvm-svn: 211019
2014-06-16 21:18:59 +08:00
class DCLO_R6_ENC : SPECIAL_2R_FM<OPCODE6_DCLO>;
class DCLZ_R6_ENC : SPECIAL_2R_FM<OPCODE6_DCLZ>;
[mips][mips64r6] Add d?div, d?mod, d?divu, d?modu Summary: Depends on D3668 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3669 llvm-svn: 208579
2014-05-12 23:24:16 +08:00
class DDIV_ENC : SPECIAL_3R_FM<0b00010, 0b011110>;
class DDIVU_ENC : SPECIAL_3R_FM<0b00010, 0b011111>;
ps][mips64r6] Added LSA/DLSA instructions Differential Revision: http://reviews.llvm.org/D3897 llvm-svn: 211346
2014-06-20 17:28:09 +08:00
class DLSA_R6_ENC : SPECIAL_LSA_FM<OPCODE6_DLSA>;
[mips][mips64r6] Add d?div, d?mod, d?divu, d?modu Summary: Depends on D3668 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3669 llvm-svn: 208579
2014-05-12 23:24:16 +08:00
class DMOD_ENC : SPECIAL_3R_FM<0b00011, 0b011110>;
class DMODU_ENC : SPECIAL_3R_FM<0b00011, 0b011111>;
[mips][mips64r6] Correct the encoding of dmuh, dmuhu, dmul, and dmulu. We have detected a documentation bug in the encoding tables of the released MIPS64r6 specification that has resulted in the wrong encodings being used for these instructions in LLVM. This commit corrects them. llvm-svn: 212330
2014-07-04 18:08:27 +08:00
class DMUH_ENC : SPECIAL_3R_FM<0b00011, 0b011100>;
class DMUHU_ENC : SPECIAL_3R_FM<0b00011, 0b011101>;
class DMUL_R6_ENC : SPECIAL_3R_FM<0b00010, 0b011100>;
class DMULU_ENC : SPECIAL_3R_FM<0b00010, 0b011101>;
[mips][mips64r6] Add LDPC instruction Differential Revision: http://reviews.llvm.org/D3822 llvm-svn: 210460
2014-06-09 17:49:51 +08:00
class LDPC_ENC : PCREL18_FM<OPCODE3_LDPC>;
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 llvm-svn: 211018
2014-06-16 21:13:03 +08:00
class LLD_R6_ENC : SPECIAL3_LL_SC_FM<OPCODE6_LLD>;
class SCD_R6_ENC : SPECIAL3_LL_SC_FM<OPCODE6_SCD>;
[mips][mips64r6] Added mul/mulu/muh/muhu Summary: The 'mul' line of the test is temporarily commented out because it currently matches the MIPS32 mul instead of the MIPS32r6 mul. This line will be uncommented when we disable the MIPS32 mul on MIPS32r6. Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3668 llvm-svn: 208576
2014-05-12 23:12:45 +08:00
//===----------------------------------------------------------------------===//
//
// Instruction Descriptions
//
//===----------------------------------------------------------------------===//
[mips] MIPSR6 Instruction itineraries Summary: Defines instruction itineraries for common MIPSR6 instructions. Patch by Simon Dardis. Reviewers: vkalintiris Subscribers: MatzeB, dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D17198 llvm-svn: 263229
2016-03-11 21:05:06 +08:00
class AHI_ATI_DESC_BASE<string instr_asm, RegisterOperand GPROpnd, InstrItinClass itin> {
[mips][mips64r6] Add b[on]vc Summary: This required me to implement the disassembler for MIPS64r6 since the encodings are ambiguous with other instructions. This in turn revealed a few assembly/disassembly bugs which I have fixed. * da[ht]i only take two operands according to the spec, not three. * DecodeBranchTarget2[16] correctly handles wider immediates than simm16 * Also made non-functional change to DecodeBranchTarget and DecodeBranchTargetMM to keep implementation style consistent between them. * Difficult encodings are handled by a custom decode method on the most general encoding in the group. This method will convert the MCInst to a different opcode if necessary. DecodeBranchTarget is not currently the inverse of getBranchTargetOpValue so disassembling some branch instructions emit incorrect output. This seems to affect branches with delay slots on all MIPS ISA's. I've left this bug for now and temporarily removed the check for the immediate on bc[12]eqz/bc[12]nez in the MIPS32r6/MIPS64r6 tests. jialc and jic crash the disassembler for some reason. I've left these instructions commented out for the moment. Depends on D3760 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3761 llvm-svn: 209415
2014-05-22 19:23:21 +08:00
dag OutOperandList = (outs GPROpnd:$rs);
[mips] Range check simm16 Summary: There are too many instructions to exhaustively test so addiu and lwc2 are used as representative examples. It should be noted that many memory instructions that should have simm16 range checking do not because it is also necessary to support the macro of the same name which accepts simm32. The range checks for these occur in the macro expansion. Reviewers: vkalintiris Subscribers: dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D18437 llvm-svn: 265019
2016-03-31 22:34:00 +08:00
dag InOperandList = (ins GPROpnd:$rt, simm16_relaxed:$imm);
[mips][mips64r6] Add b[on]vc Summary: This required me to implement the disassembler for MIPS64r6 since the encodings are ambiguous with other instructions. This in turn revealed a few assembly/disassembly bugs which I have fixed. * da[ht]i only take two operands according to the spec, not three. * DecodeBranchTarget2[16] correctly handles wider immediates than simm16 * Also made non-functional change to DecodeBranchTarget and DecodeBranchTargetMM to keep implementation style consistent between them. * Difficult encodings are handled by a custom decode method on the most general encoding in the group. This method will convert the MCInst to a different opcode if necessary. DecodeBranchTarget is not currently the inverse of getBranchTargetOpValue so disassembling some branch instructions emit incorrect output. This seems to affect branches with delay slots on all MIPS ISA's. I've left this bug for now and temporarily removed the check for the immediate on bc[12]eqz/bc[12]nez in the MIPS32r6/MIPS64r6 tests. jialc and jic crash the disassembler for some reason. I've left these instructions commented out for the moment. Depends on D3760 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3761 llvm-svn: 209415
2014-05-22 19:23:21 +08:00
string AsmString = !strconcat(instr_asm, "\t$rt, $imm");
string Constraints = "$rs = $rt";
[mips] MIPSR6 Instruction itineraries Summary: Defines instruction itineraries for common MIPSR6 instructions. Patch by Simon Dardis. Reviewers: vkalintiris Subscribers: MatzeB, dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D17198 llvm-svn: 263229
2016-03-11 21:05:06 +08:00
InstrItinClass Itinerary = itin;
[mips][mips64r6] Add b[on]vc Summary: This required me to implement the disassembler for MIPS64r6 since the encodings are ambiguous with other instructions. This in turn revealed a few assembly/disassembly bugs which I have fixed. * da[ht]i only take two operands according to the spec, not three. * DecodeBranchTarget2[16] correctly handles wider immediates than simm16 * Also made non-functional change to DecodeBranchTarget and DecodeBranchTargetMM to keep implementation style consistent between them. * Difficult encodings are handled by a custom decode method on the most general encoding in the group. This method will convert the MCInst to a different opcode if necessary. DecodeBranchTarget is not currently the inverse of getBranchTargetOpValue so disassembling some branch instructions emit incorrect output. This seems to affect branches with delay slots on all MIPS ISA's. I've left this bug for now and temporarily removed the check for the immediate on bc[12]eqz/bc[12]nez in the MIPS32r6/MIPS64r6 tests. jialc and jic crash the disassembler for some reason. I've left these instructions commented out for the moment. Depends on D3760 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3761 llvm-svn: 209415
2014-05-22 19:23:21 +08:00
}
[mips] MIPSR6 Instruction itineraries Summary: Defines instruction itineraries for common MIPSR6 instructions. Patch by Simon Dardis. Reviewers: vkalintiris Subscribers: MatzeB, dsanders, llvm-commits Differential Revision: http://reviews.llvm.org/D17198 llvm-svn: 263229
2016-03-11 21:05:06 +08:00
class DALIGN_DESC : ALIGN_DESC_BASE<"dalign", GPR64Opnd, uimm3, II_DALIGN>;
class DAHI_DESC : AHI_ATI_DESC_BASE<"dahi", GPR64Opnd, II_DAHI>;
class DATI_DESC : AHI_ATI_DESC_BASE<"dati", GPR64Opnd, II_DATI>;
class DAUI_DESC : AUI_DESC_BASE<"daui", GPR64Opnd, II_DAUI>;
class DBITSWAP_DESC : BITSWAP_DESC_BASE<"dbitswap", GPR64Opnd, II_DBITSWAP>;
class DCLO_R6_DESC : CLO_R6_DESC_BASE<"dclo", GPR64Opnd, II_DCLO>;
class DCLZ_R6_DESC : CLZ_R6_DESC_BASE<"dclz", GPR64Opnd, II_DCLZ>;
class DDIV_DESC : DIVMOD_DESC_BASE<"ddiv", GPR64Opnd, II_DDIV, sdiv>;
class DDIVU_DESC : DIVMOD_DESC_BASE<"ddivu", GPR64Opnd, II_DDIVU, udiv>;
class DLSA_R6_DESC : LSA_R6_DESC_BASE<"dlsa", GPR64Opnd, uimm2_plus1, II_DLSA>;
class DMOD_DESC : DIVMOD_DESC_BASE<"dmod", GPR64Opnd, II_DMOD, srem>;
class DMODU_DESC : DIVMOD_DESC_BASE<"dmodu", GPR64Opnd, II_DMODU, urem>;
class DMUH_DESC : MUL_R6_DESC_BASE<"dmuh", GPR64Opnd, II_DMUH, mulhs>;
class DMUHU_DESC : MUL_R6_DESC_BASE<"dmuhu", GPR64Opnd, II_DMUHU, mulhu>;
class DMUL_R6_DESC : MUL_R6_DESC_BASE<"dmul", GPR64Opnd, II_DMUL, mul>;
class DMULU_DESC : MUL_R6_DESC_BASE<"dmulu", GPR64Opnd, II_DMUL>;
class LDPC_DESC : PCREL_DESC_BASE<"ldpc", GPR64Opnd, simm18_lsl3, II_LDPC>;
class LLD_R6_DESC : LL_R6_DESC_BASE<"lld", GPR64Opnd, II_LLD>;
class SCD_R6_DESC : SC_R6_DESC_BASE<"scd", GPR64Opnd, II_SCD>;
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
class SELEQZ64_DESC : SELEQNE_Z_DESC_BASE<"seleqz", GPR64Opnd>;
class SELNEZ64_DESC : SELEQNE_Z_DESC_BASE<"selnez", GPR64Opnd>;
[mips] MIPSR6 Compact jump support This patch adds support for compact jumps similiar to the previous compact branch support for MIPSR6. Unlike compact branches, compact jumps do not have a forbidden slot. As MipsInstrInfo::getEquivalentCompactForm can determine the correct expansion for jumps and branches for both microMIPS and MIPSR6, remove the unnecessary distinction in the delay slot filler. Reviewers: vkalintiris Subscribers: llvm-commits, dsanders llvm-svn: 265390
2016-04-05 20:50:29 +08:00
class JIALC64_DESC : JMP_IDX_COMPACT_DESC_BASE<"jialc", calloffset16,
GPR64Opnd> {
bit isCall = 1;
list<Register> Defs = [RA];
}
class JIC64_DESC : JMP_IDX_COMPACT_DESC_BASE<"jic", jmpoffset16, GPR64Opnd> {
bit isBarrier = 1;
bit isTerminator = 1;
list<Register> Defs = [AT];
}
[mips][mips64r6] Added mul/mulu/muh/muhu Summary: The 'mul' line of the test is temporarily commented out because it currently matches the MIPS32 mul instead of the MIPS32r6 mul. This line will be uncommented when we disable the MIPS32 mul on MIPS32r6. Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3668 llvm-svn: 208576
2014-05-12 23:12:45 +08:00
//===----------------------------------------------------------------------===//
//
// Instruction Definitions
//
//===----------------------------------------------------------------------===//
[mips][microMIPS] Create microMIPS64r6 subtarget and implement DALIGN, DAUI, DAHI, DATI, DEXT, DEXTM and DEXTU instructions Differential Revision: http://reviews.llvm.org/D10923 llvm-svn: 244744
2015-08-12 20:45:16 +08:00
let AdditionalPredicates = [NotInMicroMips] in {
def DATI : DATI_ENC, DATI_DESC, ISA_MIPS64R6;
def DAHI : DAHI_ENC, DAHI_DESC, ISA_MIPS64R6;
def DAUI : DAUI_ENC, DAUI_DESC, ISA_MIPS64R6;
def DALIGN : DALIGN_ENC, DALIGN_DESC, ISA_MIPS64R6;
}
[mips][mips64r6] Add bitswap, and dbitswap Summary: Depends on D3728 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3729 llvm-svn: 208877
2014-05-15 20:18:23 +08:00
def DBITSWAP : DBITSWAP_ENC, DBITSWAP_DESC, ISA_MIPS64R6;
[mips][mips64r6] cl[oz], and dcl[oz] are re-encoded in MIPS32r6/MIPS64r6 Summary: There is no change to the restrictions, just the result register is stored once in the encoding rather than twice. The rt field is zero in MIPS32r6/MIPS64r6. Depends on D4119 Reviewers: zoran.jovanovic, jkolek, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4120 llvm-svn: 211019
2014-06-16 21:18:59 +08:00
def DCLO_R6 : DCLO_R6_ENC, DCLO_R6_DESC, ISA_MIPS64R6;
def DCLZ_R6 : DCLZ_R6_ENC, DCLZ_R6_DESC, ISA_MIPS64R6;
[mips][microMIPS] Add CodeGen support for DIV, MOD, DIVU, MODU, DDIV, DMOD, DDIVU and DMODU instructions Differential Revision: http://reviews.llvm.org/D17137 This patch was reverted after the revertion of dependant patch http://reviews.llvm.org/D17068. There was the problem with test-suite failure. The problem is hopefully solved with dependant patch so this patch is commited again. llvm-svn: 266179
2016-04-13 16:02:26 +08:00
let AdditionalPredicates = [NotInMicroMips] in {
def DDIV : DDIV_ENC, DDIV_DESC, ISA_MIPS64R6;
def DDIVU : DDIVU_ENC, DDIVU_DESC, ISA_MIPS64R6;
def DMOD : DMOD_ENC, DMOD_DESC, ISA_MIPS64R6;
def DMODU : DMODU_ENC, DMODU_DESC, ISA_MIPS64R6;
}
ps][mips64r6] Added LSA/DLSA instructions Differential Revision: http://reviews.llvm.org/D3897 llvm-svn: 211346
2014-06-20 17:28:09 +08:00
def DLSA_R6 : DLSA_R6_ENC, DLSA_R6_DESC, ISA_MIPS64R6;
[mips][mips64r6] Added mul/mulu/muh/muhu Summary: The 'mul' line of the test is temporarily commented out because it currently matches the MIPS32 mul instead of the MIPS32r6 mul. This line will be uncommented when we disable the MIPS32 mul on MIPS32r6. Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D3668 llvm-svn: 208576
2014-05-12 23:12:45 +08:00
def DMUH: DMUH_ENC, DMUH_DESC, ISA_MIPS64R6;
def DMUHU: DMUHU_ENC, DMUHU_DESC, ISA_MIPS64R6;
def DMUL_R6: DMUL_R6_ENC, DMUL_R6_DESC, ISA_MIPS64R6;
def DMULU: DMULU_ENC, DMULU_DESC, ISA_MIPS64R6;
[mips][mips64r6] Add LDPC instruction Differential Revision: http://reviews.llvm.org/D3822 llvm-svn: 210460
2014-06-09 17:49:51 +08:00
def LDPC: LDPC_ENC, LDPC_DESC, ISA_MIPS64R6;
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
def LLD_R6 : LLD_R6_ENC, LLD_R6_DESC, ISA_MIPS32R6;
def SCD_R6 : SCD_R6_ENC, SCD_R6_DESC, ISA_MIPS32R6;
let DecoderNamespace = "Mips32r6_64r6_GP64" in {
def SELEQZ64 : SELEQZ_ENC, SELEQZ64_DESC, ISA_MIPS32R6, GPR_64;
def SELNEZ64 : SELNEZ_ENC, SELNEZ64_DESC, ISA_MIPS32R6, GPR_64;
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
}
[mips] MIPSR6 Compact jump support This patch adds support for compact jumps similiar to the previous compact branch support for MIPSR6. Unlike compact branches, compact jumps do not have a forbidden slot. As MipsInstrInfo::getEquivalentCompactForm can determine the correct expansion for jumps and branches for both microMIPS and MIPSR6, remove the unnecessary distinction in the delay slot filler. Reviewers: vkalintiris Subscribers: llvm-commits, dsanders llvm-svn: 265390
2016-04-05 20:50:29 +08:00
let isCodeGenOnly = 1 in {
def JIALC64 : JIALC_ENC, JIALC64_DESC, ISA_MIPS64R6;
def JIC64 : JIC_ENC, JIC64_DESC, ISA_MIPS64R6;
}
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
//===----------------------------------------------------------------------===//
//
[mips][mips64r6] Use JALR for returns instead of JR (which is not available on MIPS32r6/MIPS64r6) Summary: RET, and RET_MM have been replaced by a pseudo named PseudoReturn. In addition a version with a 64-bit GPR named PseudoReturn64 has been added. Instruction selection for a return matches RetRA, which is expanded post register allocation to PseudoReturn/PseudoReturn64. During MipsAsmPrinter, this PseudoReturn/PseudoReturn64 are emitted as: - (JALR64 $zero, $rs) on MIPS64r6 - (JALR $zero, $rs) on MIPS32r6 - (JR_MM $rs) on microMIPS - (JR $rs) otherwise On MIPS32r6/MIPS64r6, 'jr $rs' is an alias for 'jalr $zero, $rs'. To aid development and review (specifically, to ensure all cases of jr are updated), these aliases are temporarily named 'r6.jr' instead of 'jr'. A follow up patch will change them back to the correct mnemonic. Added (JALR $zero, $rs) to MipsNaClELFStreamer's definition of an indirect jump, and removed it from its definition of a call. Note: I haven't accounted for MIPS64 in MipsNaClELFStreamer since it's doesn't appear to account for any MIPS64-specifics. The return instruction created as part of eh_return expansion is now expanded using expandRetRA() so we use the right return instruction on MIPS32r6/MIPS64r6 ('jalr $zero, $rs'). Also, fixed a misuse of isABI_N64() to detect 64-bit wide registers in expandEhReturn(). Reviewers: jkolek, vmedic, mseaborn, zoran.jovanovic, dsanders Reviewed By: dsanders Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D4268 llvm-svn: 212604
2014-07-09 18:16:07 +08:00
// Instruction Aliases
//
//===----------------------------------------------------------------------===//
def : MipsInstAlias<"jr $rs", (JALR64 ZERO_64, GPR64Opnd:$rs), 1>, ISA_MIPS64R6;
//===----------------------------------------------------------------------===//
//
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
// Patterns and Pseudo Instructions
//
//===----------------------------------------------------------------------===//
// i64 selects
def : MipsPat<(select i64:$cond, i64:$t, i64:$f),
(OR64 (SELNEZ64 i64:$t, i64:$cond),
(SELEQZ64 i64:$f, i64:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i64:$cond, immz)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t, i64:$cond),
(SELNEZ64 i64:$f, i64:$cond))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i64:$cond, immz)), i64:$t, i64:$f),
(OR64 (SELNEZ64 i64:$t, i64:$cond),
(SELEQZ64 i64:$f, i64:$cond))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i64:$cond, immZExt16_64:$imm)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t, (XORi64 i64:$cond, immZExt16_64:$imm)),
(SELNEZ64 i64:$f, (XORi64 i64:$cond, immZExt16_64:$imm)))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i64:$cond, immZExt16_64:$imm)), i64:$t, i64:$f),
(OR64 (SELNEZ64 i64:$t, (XORi64 i64:$cond, immZExt16_64:$imm)),
(SELEQZ64 i64:$f, (XORi64 i64:$cond, immZExt16_64:$imm)))>,
ISA_MIPS64R6;
def : MipsPat<
(select (i32 (setgt i64:$cond, immSExt16Plus1:$imm)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t,
(SUBREG_TO_REG (i64 0), (SLTi64 i64:$cond, (Plus1 imm:$imm)),
sub_32)),
(SELNEZ64 i64:$f,
(SUBREG_TO_REG (i64 0), (SLTi64 i64:$cond, (Plus1 imm:$imm)),
sub_32)))>,
ISA_MIPS64R6;
def : MipsPat<
(select (i32 (setugt i64:$cond, immSExt16Plus1:$imm)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t,
(SUBREG_TO_REG (i64 0), (SLTiu64 i64:$cond, (Plus1 imm:$imm)),
sub_32)),
(SELNEZ64 i64:$f,
(SUBREG_TO_REG (i64 0), (SLTiu64 i64:$cond, (Plus1 imm:$imm)),
sub_32)))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i64:$cond, immz)), i64:$t, immz),
(SELNEZ64 i64:$t, i64:$cond)>, ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i64:$cond, immz)), i64:$t, immz),
(SELEQZ64 i64:$t, i64:$cond)>, ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i64:$cond, immz)), immz, i64:$f),
(SELEQZ64 i64:$f, i64:$cond)>, ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i64:$cond, immz)), immz, i64:$f),
(SELNEZ64 i64:$f, i64:$cond)>, ISA_MIPS64R6;
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
// i64 selects from an i32 comparison
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
// One complicating factor here is that bits 32-63 of an i32 are undefined.
// FIXME: Ideally, setcc would always produce an i64 on MIPS64 targets.
// This would allow us to remove the sign-extensions here.
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
def : MipsPat<(select i32:$cond, i64:$t, i64:$f),
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
(OR64 (SELNEZ64 i64:$t, (SLL64_32 i32:$cond)),
(SELEQZ64 i64:$f, (SLL64_32 i32:$cond)))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i32:$cond, immz)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t, (SLL64_32 i32:$cond)),
(SELNEZ64 i64:$f, (SLL64_32 i32:$cond)))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i32:$cond, immz)), i64:$t, i64:$f),
(OR64 (SELNEZ64 i64:$t, (SLL64_32 i32:$cond)),
(SELEQZ64 i64:$f, (SLL64_32 i32:$cond)))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i32:$cond, immZExt16:$imm)), i64:$t, i64:$f),
(OR64 (SELEQZ64 i64:$t, (SLL64_32 (XORi i32:$cond,
immZExt16:$imm))),
(SELNEZ64 i64:$f, (SLL64_32 (XORi i32:$cond,
immZExt16:$imm))))>,
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i32:$cond, immZExt16:$imm)), i64:$t, i64:$f),
(OR64 (SELNEZ64 i64:$t, (SLL64_32 (XORi i32:$cond,
immZExt16:$imm))),
(SELEQZ64 i64:$f, (SLL64_32 (XORi i32:$cond,
immZExt16:$imm))))>,
ISA_MIPS64R6;
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
def : MipsPat<(select i32:$cond, i64:$t, immz),
(SELNEZ64 i64:$t, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i32:$cond, immz)), i64:$t, immz),
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
(SELNEZ64 i64:$t, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i32:$cond, immz)), i64:$t, immz),
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
(SELEQZ64 i64:$t, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
def : MipsPat<(select i32:$cond, immz, i64:$f),
(SELEQZ64 i64:$f, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
def : MipsPat<(select (i32 (setne i32:$cond, immz)), immz, i64:$f),
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
(SELEQZ64 i64:$f, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;
def : MipsPat<(select (i32 (seteq i32:$cond, immz)), immz, i64:$f),
[mips][mips64r6] c.cond.fmt, mov[fntz], and mov[fntz].[ds] are not available on MIPS32r6/MIPS64r6 Summary: c.cond.fmt has been replaced by cmp.cond.fmt. Where c.cond.fmt wrote to dedicated condition registers, cmp.cond.fmt writes 1 or 0 to normal FGR's (like the GPR comparisons). mov[fntz] have been replaced by seleqz and selnez. These instructions conditionally zero a register based on a bool in a GPR. The results can then be or'd together to act as a select without, for example, requiring a third register read port. mov[fntz].[ds] have been replaced with sel.[ds] MIPS64r6 currently generates unnecessary sign-extensions for most selects. This is because the result of a SETCC is currently an i32. Bits 32-63 are undefined in i32 and the behaviour of seleqz/selnez would otherwise depend on undefined bits. Later, we will fix this by making the result of SETCC an i64 on MIPS64 targets. Depends on D3958 Reviewers: jkolek, vmedic, zoran.jovanovic Reviewed By: vmedic, zoran.jovanovic Differential Revision: http://reviews.llvm.org/D4003 llvm-svn: 210777
2014-06-12 21:39:06 +08:00
(SELNEZ64 i64:$f, (SLL64_32 i32:$cond))>,
Revert "[mips] Promote the result of SETCC nodes to GPR width." This reverts commit r262316. It seems that my change breaks an out-of-tree chromium buildbot, so I'm reverting this in order to investigate the situation further. llvm-svn: 262387
2016-03-02 04:25:43 +08:00
ISA_MIPS64R6;