Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(1157)

Unified Diff: src/s390/assembler-s390.h

Issue 1725243004: S390: Initial impl of S390 asm, masm, code-stubs,... (Closed) Base URL: https://chromium.googlesource.com/v8/v8.git@master
Patch Set: Updated BUILD.gn + cpu-s390.cc to addr @jochen's comments. Created 4 years, 10 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « src/base/macros.h ('k') | src/s390/assembler-s390.cc » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: src/s390/assembler-s390.h
diff --git a/src/s390/assembler-s390.h b/src/s390/assembler-s390.h
new file mode 100644
index 0000000000000000000000000000000000000000..05999611d75c12e8a3fcbac26c6ab14a0276ebd1
--- /dev/null
+++ b/src/s390/assembler-s390.h
@@ -0,0 +1,1467 @@
+// Copyright (c) 1994-2006 Sun Microsystems Inc.
+// All Rights Reserved.
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions
+// are met:
+//
+// - Redistributions of source code must retain the above copyright notice,
+// this list of conditions and the following disclaimer.
+//
+// - Redistribution in binary form must reproduce the above copyright
+// notice, this list of conditions and the following disclaimer in the
+// documentation and/or other materials provided with the
+// distribution.
+//
+// - Neither the name of Sun Microsystems or the names of contributors may
+// be used to endorse or promote products derived from this software without
+// specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+// FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+// COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+// INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+// (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+// SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+// HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+// OF THE POSSIBILITY OF SUCH DAMAGE.
+
+// The original source code covered by the above license above has been
+// modified significantly by Google Inc.
+// Copyright 2014 the V8 project authors. All rights reserved.
jochen (gone - plz use gerrit) 2016/02/26 12:11:14 please update the year in V8 copyright headers to
+
+// A light-weight S390 Assembler
+// Generates user mode instructions for z/Architecture
+
+#ifndef V8_S390_ASSEMBLER_S390_H_
+#define V8_S390_ASSEMBLER_S390_H_
+#include <stdio.h>
+#if V8_HOST_ARCH_S390
+// elf.h include is required for auxv check for STFLE facility used
+// for hardware detection, which is sensible only on s390 hosts.
+#include <elf.h>
+#endif
+
+#include <fcntl.h>
+#include <unistd.h>
+#include "src/assembler.h"
+#include "src/s390/constants-s390.h"
+
+#define ABI_USES_FUNCTION_DESCRIPTORS 0
+
+#define ABI_PASSES_HANDLES_IN_REGS 1
+
+// ObjectPair is defined under runtime/runtime-util.h.
+// On 31-bit, ObjectPair == uint64_t. ABI dictates long long
+// be returned with the lower addressed half in r2
+// and the higher addressed half in r3. (Returns in Regs)
+// On 64-bit, ObjectPair is a Struct. ABI dictaes Structs be
+// returned in a storage buffer allocated by the caller,
+// with the address of this buffer passed as a hidden
+// argument in r2. (Does NOT return in Regs)
+// For x86 linux, ObjectPair is returned in registers.
+#if V8_TARGET_ARCH_S390X
+#define ABI_RETURNS_OBJECTPAIR_IN_REGS 0
+#else
+#define ABI_RETURNS_OBJECTPAIR_IN_REGS 1
+#endif
+
+#define ABI_CALL_VIA_IP 1
+
+#define INSTR_AND_DATA_CACHE_COHERENCY LWSYNC
+
+namespace v8 {
+namespace internal {
+
+// clang-format off
+#define GENERAL_REGISTERS(V) \
+ V(r0) V(r1) V(r2) V(r3) V(r4) V(r5) V(r6) V(r7) \
+ V(r8) V(r9) V(r10) V(fp) V(ip) V(r13) V(r14) V(sp)
+
+#define ALLOCATABLE_GENERAL_REGISTERS(V) \
+ V(r2) V(r3) V(r4) V(r5) V(r6) V(r7) \
+ V(r8) V(r9) V(r13)
+
+#define DOUBLE_REGISTERS(V) \
+ V(d0) V(d1) V(d2) V(d3) V(d4) V(d5) V(d6) V(d7) \
+ V(d8) V(d9) V(d10) V(d11) V(d12) V(d13) V(d14) V(d15)
+
+#define ALLOCATABLE_DOUBLE_REGISTERS(V) \
+ V(d1) V(d2) V(d3) V(d4) V(d5) V(d6) V(d7) \
+ V(d8) V(d9) V(d10) V(d11) V(d12) V(d15) V(d0)
+// clang-format on
+
+// CPU Registers.
+//
+// 1) We would prefer to use an enum, but enum values are assignment-
+// compatible with int, which has caused code-generation bugs.
+//
+// 2) We would prefer to use a class instead of a struct but we don't like
+// the register initialization to depend on the particular initialization
+// order (which appears to be different on OS X, Linux, and Windows for the
+// installed versions of C++ we tried). Using a struct permits C-style
+// "initialization". Also, the Register objects cannot be const as this
+// forces initialization stubs in MSVC, making us dependent on initialization
+// order.
+//
+// 3) By not using an enum, we are possibly preventing the compiler from
+// doing certain constant folds, which may significantly reduce the
+// code generated for some assembly instructions (because they boil down
+// to a few constants). If this is a problem, we could change the code
+// such that we use an enum in optimized mode, and the struct in debug
+// mode. This way we get the compile-time error checking in debug mode
+// and best performance in optimized code.
+
+struct Register {
+ enum Code {
+#define REGISTER_CODE(R) kCode_##R,
+ GENERAL_REGISTERS(REGISTER_CODE)
+#undef REGISTER_CODE
+ kAfterLast,
+ kCode_no_reg = -1
+ };
+ static const int kNumRegisters = Code::kAfterLast;
+
+#define REGISTER_COUNT(R) 1 +
+ static const int kNumAllocatable =
+ ALLOCATABLE_GENERAL_REGISTERS(REGISTER_COUNT) 0;
+#undef REGISTER_COUNT
+
+#define REGISTER_BIT(R) 1 << kCode_##R |
+ static const RegList kAllocatable =
+ ALLOCATABLE_GENERAL_REGISTERS(REGISTER_BIT) 0;
+#undef REGISTER_BIT
+
+ static Register from_code(int code) {
+ DCHECK(code >= 0);
+ DCHECK(code < kNumRegisters);
+ Register r = {code};
+ return r;
+ }
+
+ const char* ToString();
+ bool IsAllocatable() const;
+ bool is_valid() const { return 0 <= reg_code && reg_code < kNumRegisters; }
+ bool is(Register reg) const { return reg_code == reg.reg_code; }
+ int code() const {
+ DCHECK(is_valid());
+ return reg_code;
+ }
+ int bit() const {
+ DCHECK(is_valid());
+ return 1 << reg_code;
+ }
+
+ void set_code(int code) {
+ reg_code = code;
+ DCHECK(is_valid());
+ }
+
+#if V8_TARGET_LITTLE_ENDIAN
+ static const int kMantissaOffset = 0;
+ static const int kExponentOffset = 4;
+#else
+ static const int kMantissaOffset = 4;
+ static const int kExponentOffset = 0;
+#endif
+
+ // Unfortunately we can't make this private in a struct.
+ int reg_code;
+};
+
+typedef struct Register Register;
+
+#define DECLARE_REGISTER(R) const Register R = {Register::kCode_##R};
+GENERAL_REGISTERS(DECLARE_REGISTER)
+#undef DECLARE_REGISTER
+const Register no_reg = {Register::kCode_no_reg};
+
+// Register aliases
+const Register kLithiumScratch = r1; // lithium scratch.
+const Register kRootRegister = r10; // Roots array pointer.
+const Register cp = r13; // JavaScript context pointer.
+
+// Double word FP register.
+struct DoubleRegister {
+ enum Code {
+#define REGISTER_CODE(R) kCode_##R,
+ DOUBLE_REGISTERS(REGISTER_CODE)
+#undef REGISTER_CODE
+ kAfterLast,
+ kCode_no_reg = -1
+ };
+
+ static const int kNumRegisters = Code::kAfterLast;
+ static const int kMaxNumRegisters = kNumRegisters;
+
+ const char* ToString();
+ bool IsAllocatable() const;
+ bool is_valid() const { return 0 <= reg_code && reg_code < kNumRegisters; }
+ bool is(DoubleRegister reg) const { return reg_code == reg.reg_code; }
+
+ int code() const {
+ DCHECK(is_valid());
+ return reg_code;
+ }
+
+ int bit() const {
+ DCHECK(is_valid());
+ return 1 << reg_code;
+ }
+
+ static DoubleRegister from_code(int code) {
+ DoubleRegister r = {code};
+ return r;
+ }
+
+ int reg_code;
+};
+
+typedef DoubleRegister DoubleRegister;
+
+#define DECLARE_REGISTER(R) \
+ const DoubleRegister R = {DoubleRegister::kCode_##R};
+DOUBLE_REGISTERS(DECLARE_REGISTER)
+#undef DECLARE_REGISTER
+const Register no_dreg = {Register::kCode_no_reg};
+
+// Aliases for double registers. Defined using #define instead of
+// "static const DoubleRegister&" because Clang complains otherwise when a
+// compilation unit that includes this header doesn't use the variables.
+#define kDoubleRegZero d14
+#define kScratchDoubleReg d13
+
+Register ToRegister(int num);
+
+// Coprocessor register
+struct CRegister {
+ bool is_valid() const { return 0 <= reg_code && reg_code < 16; }
+ bool is(CRegister creg) const { return reg_code == creg.reg_code; }
+ int code() const {
+ DCHECK(is_valid());
+ return reg_code;
+ }
+ int bit() const {
+ DCHECK(is_valid());
+ return 1 << reg_code;
+ }
+
+ // Unfortunately we can't make this private in a struct.
+ int reg_code;
+};
+
+const CRegister no_creg = {-1};
+
+const CRegister cr0 = {0};
+const CRegister cr1 = {1};
+const CRegister cr2 = {2};
+const CRegister cr3 = {3};
+const CRegister cr4 = {4};
+const CRegister cr5 = {5};
+const CRegister cr6 = {6};
+const CRegister cr7 = {7};
+const CRegister cr8 = {8};
+const CRegister cr9 = {9};
+const CRegister cr10 = {10};
+const CRegister cr11 = {11};
+const CRegister cr12 = {12};
+const CRegister cr13 = {13};
+const CRegister cr14 = {14};
+const CRegister cr15 = {15};
+
+// TODO(john.yan) Define SIMD registers.
+typedef DoubleRegister Simd128Register;
+
+// -----------------------------------------------------------------------------
+// Machine instruction Operands
+
+#if V8_TARGET_ARCH_S390X
+const RelocInfo::Mode kRelocInfo_NONEPTR = RelocInfo::NONE64;
+#else
+const RelocInfo::Mode kRelocInfo_NONEPTR = RelocInfo::NONE32;
+#endif
+
+// Class Operand represents a shifter operand in data processing instructions
+// defining immediate numbers and masks
+typedef uint8_t Length;
+
+struct Mask {
+ uint8_t mask;
+ uint8_t value() { return mask; }
+ static Mask from_value(uint8_t input) {
+ DCHECK(input <= 0x0F);
+ Mask m = {input};
+ return m;
+ }
+};
+
+class Operand BASE_EMBEDDED {
+ public:
+ // immediate
+ INLINE(explicit Operand(intptr_t immediate,
+ RelocInfo::Mode rmode = kRelocInfo_NONEPTR));
+ INLINE(static Operand Zero()) { return Operand(static_cast<intptr_t>(0)); }
+ INLINE(explicit Operand(const ExternalReference& f));
+ explicit Operand(Handle<Object> handle);
+ INLINE(explicit Operand(Smi* value));
+
+ // rm
+ INLINE(explicit Operand(Register rm));
+
+ // Return true if this is a register operand.
+ INLINE(bool is_reg() const);
+
+ bool must_output_reloc_info(const Assembler* assembler) const;
+
+ inline intptr_t immediate() const {
+ DCHECK(!rm_.is_valid());
+ return imm_;
+ }
+
+ inline void setBits(int n) {
+ imm_ = (static_cast<uint32_t>(imm_) << (32 - n)) >> (32 - n);
+ }
+
+ Register rm() const { return rm_; }
+
+ private:
+ Register rm_;
+ intptr_t imm_; // valid if rm_ == no_reg
+ RelocInfo::Mode rmode_;
+
+ friend class Assembler;
+ friend class MacroAssembler;
+};
+
+typedef int32_t Disp;
+
+// Class MemOperand represents a memory operand in load and store instructions
+// On S390, we have various flavours of memory operands:
+// 1) a base register + 16 bit unsigned displacement
+// 2) a base register + index register + 16 bit unsigned displacement
+// 3) a base register + index register + 20 bit signed displacement
+class MemOperand BASE_EMBEDDED {
+ public:
+ explicit MemOperand(Register rx, Disp offset = 0);
+ explicit MemOperand(Register rx, Register rb, Disp offset = 0);
+
+ int32_t offset() const { return offset_; }
+ uint32_t getDisplacement() const { return offset(); }
+
+ // Base register
+ Register rb() const {
+ DCHECK(!baseRegister.is(no_reg));
+ return baseRegister;
+ }
+
+ Register getBaseRegister() const { return rb(); }
+
+ // Index Register
+ Register rx() const {
+ DCHECK(!indexRegister.is(no_reg));
+ return indexRegister;
+ }
+ Register getIndexRegister() const { return rx(); }
+
+ private:
+ Register baseRegister; // base
+ Register indexRegister; // index
+ int32_t offset_; // offset
+
+ friend class Assembler;
+};
+
+class DeferredRelocInfo {
+ public:
+ DeferredRelocInfo() {}
+ DeferredRelocInfo(int position, RelocInfo::Mode rmode, intptr_t data)
+ : position_(position), rmode_(rmode), data_(data) {}
+
+ int position() const { return position_; }
+ RelocInfo::Mode rmode() const { return rmode_; }
+ intptr_t data() const { return data_; }
+
+ private:
+ int position_;
+ RelocInfo::Mode rmode_;
+ intptr_t data_;
+};
+
+class Assembler : public AssemblerBase {
+ public:
+ // Create an assembler. Instructions and relocation information are emitted
+ // into a buffer, with the instructions starting from the beginning and the
+ // relocation information starting from the end of the buffer. See CodeDesc
+ // for a detailed comment on the layout (globals.h).
+ //
+ // If the provided buffer is NULL, the assembler allocates and grows its own
+ // buffer, and buffer_size determines the initial buffer size. The buffer is
+ // owned by the assembler and deallocated upon destruction of the assembler.
+ //
+ // If the provided buffer is not NULL, the assembler uses the provided buffer
+ // for code generation and assumes its size to be buffer_size. If the buffer
+ // is too small, a fatal error occurs. No deallocation of the buffer is done
+ // upon destruction of the assembler.
+ Assembler(Isolate* isolate, void* buffer, int buffer_size);
+ virtual ~Assembler() {}
+
+ // GetCode emits any pending (non-emitted) code and fills the descriptor
+ // desc. GetCode() is idempotent; it returns the same result if no other
+ // Assembler functions are invoked in between GetCode() calls.
+ void GetCode(CodeDesc* desc);
+
+ // Label operations & relative jumps (PPUM Appendix D)
+ //
+ // Takes a branch opcode (cc) and a label (L) and generates
+ // either a backward branch or a forward branch and links it
+ // to the label fixup chain. Usage:
+ //
+ // Label L; // unbound label
+ // j(cc, &L); // forward branch to unbound label
+ // bind(&L); // bind label to the current pc
+ // j(cc, &L); // backward branch to bound label
+ // bind(&L); // illegal: a label may be bound only once
+ //
+ // Note: The same Label can be used for forward and backward branches
+ // but it may be bound only once.
+
+ void bind(Label* L); // binds an unbound label L to the current code position
+
+ // Links a label at the current pc_offset(). If already bound, returns the
+ // bound position. If already linked, returns the position of the prior link.
+ // Otherwise, returns the current pc_offset().
+ int link(Label* L);
+
+ // Determines if Label is bound and near enough so that a single
+ // branch instruction can be used to reach it.
+ bool is_near(Label* L, Condition cond);
+
+ // Returns the branch offset to the given label from the current code position
+ // Links the label to the current position if it is still unbound
+ int branch_offset(Label* L) { return link(L) - pc_offset(); }
+
+ // Puts a labels target address at the given position.
+ // The high 8 bits are set to zero.
+ void label_at_put(Label* L, int at_offset);
+ void load_label_offset(Register r1, Label* L);
+
+ // Read/Modify the code target address in the branch/call instruction at pc.
+ INLINE(static Address target_address_at(Address pc, Address constant_pool));
+ INLINE(static void set_target_address_at(
+ Isolate* isolate, Address pc, Address constant_pool, Address target,
+ ICacheFlushMode icache_flush_mode = FLUSH_ICACHE_IF_NEEDED));
+ INLINE(static Address target_address_at(Address pc, Code* code)) {
+ Address constant_pool = NULL;
+ return target_address_at(pc, constant_pool);
+ }
+ INLINE(static void set_target_address_at(
+ Isolate* isolate, Address pc, Code* code, Address target,
+ ICacheFlushMode icache_flush_mode = FLUSH_ICACHE_IF_NEEDED)) {
+ Address constant_pool = NULL;
+ set_target_address_at(isolate, pc, constant_pool, target,
+ icache_flush_mode);
+ }
+
+ // Return the code target address at a call site from the return address
+ // of that call in the instruction stream.
+ inline static Address target_address_from_return_address(Address pc);
+
+ // Given the address of the beginning of a call, return the address
+ // in the instruction stream that the call will return to.
+ INLINE(static Address return_address_from_call_start(Address pc));
+
+ inline Handle<Object> code_target_object_handle_at(Address pc);
+ // This sets the branch destination.
+ // This is for calls and branches within generated code.
+ inline static void deserialization_set_special_target_at(
+ Isolate* isolate, Address instruction_payload, Code* code,
+ Address target);
+
+ // This sets the internal reference at the pc.
+ inline static void deserialization_set_target_internal_reference_at(
+ Isolate* isolate, Address pc, Address target,
+ RelocInfo::Mode mode = RelocInfo::INTERNAL_REFERENCE);
+
+ // Here we are patching the address in the IIHF/IILF instruction pair.
+ // These values are used in the serialization process and must be zero for
+ // S390 platform, as Code, Embedded Object or External-reference pointers
+ // are split across two consecutive instructions and don't exist separately
+ // in the code, so the serializer should not step forwards in memory after
+ // a target is resolved and written.
+ static const int kSpecialTargetSize = 0;
+
+// Number of bytes for instructions used to store pointer sized constant.
+#if V8_TARGET_ARCH_S390X
+ static const int kBytesForPtrConstant = 12; // IIHF + IILF
+#else
+ static const int kBytesForPtrConstant = 6; // IILF
+#endif
+
+ // Distance between the instruction referring to the address of the call
+ // target and the return address.
+
+ // Offset between call target address and return address
+ // for BRASL calls
+ // Patch will be appiled to other FIXED_SEQUENCE call
+ static const int kCallTargetAddressOffset = 6;
+
+// The length of FIXED_SEQUENCE call
+// iihf r8, <address_hi> // <64-bit only>
+// iilf r8, <address_lo>
+// basr r14, r8
+#if V8_TARGET_ARCH_S390X
+ static const int kCallSequenceLength = 14;
+#else
+ static const int kCallSequenceLength = 8;
+#endif
+
+ // This is the length of the BreakLocationIterator::SetDebugBreakAtReturn()
+ // code patch FIXED_SEQUENCE in bytes!
+ // JS Return Sequence = Call Sequence + BKPT
+ // static const int kJSReturnSequenceLength = kCallSequenceLength + 2;
+
+ // This is the length of the code sequence from SetDebugBreakAtSlot()
+ // FIXED_SEQUENCE in bytes!
+ static const int kDebugBreakSlotLength = kCallSequenceLength;
+ static const int kPatchDebugBreakSlotReturnOffset = kCallTargetAddressOffset;
+
+ // Length to patch between the start of the JS return sequence
+ // from SetDebugBreakAtReturn and the address from
+ // break_address_from_return_address.
+ //
+ // frame->pc() in Debug::SetAfterBreakTarget will point to BKPT in
+ // JS return sequence, so the length to patch will not include BKPT
+ // instruction length.
+ // static const int kPatchReturnSequenceAddressOffset =
+ // kCallSequenceLength - kPatchDebugBreakSlotReturnOffset;
+
+ // Length to patch between the start of the FIXED call sequence from
+ // SetDebugBreakAtSlot() and the the address from
+ // break_address_from_return_address.
+ static const int kPatchDebugBreakSlotAddressOffset =
+ kDebugBreakSlotLength - kPatchDebugBreakSlotReturnOffset;
+
+ static inline int encode_crbit(const CRegister& cr, enum CRBit crbit) {
+ return ((cr.code() * CRWIDTH) + crbit);
+ }
+
+ // ---------------------------------------------------------------------------
+ // Code generation
+
+ // Helper for unconditional branch to Label with update to save register
+ void b(Register r, Label* l) {
+ positions_recorder()->WriteRecordedPositions();
+ int32_t halfwords = branch_offset(l) / 2;
+ brasl(r, Operand(halfwords));
+ }
+
+ // Conditional Branch Instruction - Generates either BRC / BRCL
+ void branchOnCond(Condition c, int branch_offset, bool is_bound = false);
+
+ // Helpers for conditional branch to Label
+ void b(Condition cond, Label* l, Label::Distance dist = Label::kFar) {
+ branchOnCond(cond, branch_offset(l),
+ l->is_bound() || (dist == Label::kNear));
+ }
+
+ void bc_short(Condition cond, Label* l, Label::Distance dist = Label::kFar) {
+ b(cond, l, Label::kNear);
+ }
+ // Helpers for conditional branch to Label
+ void beq(Label* l, Label::Distance dist = Label::kFar) { b(eq, l, dist); }
+ void bne(Label* l, Label::Distance dist = Label::kFar) { b(ne, l, dist); }
+ void blt(Label* l, Label::Distance dist = Label::kFar) { b(lt, l, dist); }
+ void ble(Label* l, Label::Distance dist = Label::kFar) { b(le, l, dist); }
+ void bgt(Label* l, Label::Distance dist = Label::kFar) { b(gt, l, dist); }
+ void bge(Label* l, Label::Distance dist = Label::kFar) { b(ge, l, dist); }
+ void b(Label* l, Label::Distance dist = Label::kFar) { b(al, l, dist); }
+ void jmp(Label* l, Label::Distance dist = Label::kFar) { b(al, l, dist); }
+ void bunordered(Label* l, Label::Distance dist = Label::kFar) {
+ b(unordered, l, dist);
+ }
+ void bordered(Label* l, Label::Distance dist = Label::kFar) {
+ b(ordered, l, dist);
+ }
+
+ // Helpers for conditional indirect branch off register
+ void b(Condition cond, Register r) { bcr(cond, r); }
+ void beq(Register r) { b(eq, r); }
+ void bne(Register r) { b(ne, r); }
+ void blt(Register r) { b(lt, r); }
+ void ble(Register r) { b(le, r); }
+ void bgt(Register r) { b(gt, r); }
+ void bge(Register r) { b(ge, r); }
+ void b(Register r) { b(al, r); }
+ void jmp(Register r) { b(al, r); }
+ void bunordered(Register r) { b(unordered, r); }
+ void bordered(Register r) { b(ordered, r); }
+
+ // ---------------------------------------------------------------------------
+ // Code generation
+
+ // Insert the smallest number of nop instructions
+ // possible to align the pc offset to a multiple
+ // of m. m must be a power of 2 (>= 4).
+ void Align(int m);
+ // Insert the smallest number of zero bytes possible to align the pc offset
+ // to a mulitple of m. m must be a power of 2 (>= 2).
+ void DataAlign(int m);
+ // Aligns code to something that's optimal for a jump target for the platform.
+ void CodeTargetAlign();
+
+ void breakpoint(bool do_print) {
+ if (do_print) {
+ printf("DebugBreak is inserted to %p\n", pc_);
+ }
+#if V8_HOST_ARCH_64_BIT
+ int64_t value = reinterpret_cast<uint64_t>(&v8::base::OS::DebugBreak);
+ int32_t hi_32 = static_cast<int64_t>(value) >> 32;
+ int32_t lo_32 = static_cast<int32_t>(value);
+
+ iihf(r1, Operand(hi_32));
+ iilf(r1, Operand(lo_32));
+#else
+ iilf(r1, Operand(reinterpret_cast<uint32_t>(&v8::base::OS::DebugBreak)));
+#endif
+ basr(r14, r1);
+ }
+
+ void call(Handle<Code> target, RelocInfo::Mode rmode,
+ TypeFeedbackId ast_id = TypeFeedbackId::None());
+ void jump(Handle<Code> target, RelocInfo::Mode rmode, Condition cond);
+
+// S390 instruction generation
+#define I_FORM(name) void name(const Operand& i)
+
+#define RR_FORM(name) void name(Register r1, Register r2)
+
+#define RR2_FORM(name) void name(Condition m1, Register r2)
+
+#define RX_FORM(name) \
+ void name(Register r1, Register x2, Register b2, Disp d2); \
+ void name(Register r1, const MemOperand& opnd)
+
+#define RI1_FORM(name) void name(Register r, const Operand& i)
+
+#define RI2_FORM(name) void name(Condition m, const Operand& i)
+
+#define RIE_FORM(name) void name(Register r1, Register R3, const Operand& i)
+
+#define RIE_F_FORM(name) \
+ void name(Register r1, Register r2, const Operand& i3, const Operand& i4, \
+ const Operand& i5)
+
+#define RIL1_FORM(name) void name(Register r1, const Operand& i2)
+
+#define RIL2_FORM(name) void name(Condition m1, const Operand& i2)
+
+#define RXE_FORM(name) \
+ void name(Register r1, const MemOperand& opnd); \
+ void name(Register r1, Register b2, Register x2, Disp d2)
+
+#define RXF_FORM(name) \
+ void name(Register r1, Register r3, const MemOperand& opnd); \
+ void name(Register r1, Register r3, Register b2, Register x2, Disp d2)
+
+#define RXY_FORM(name) \
+ void name(Register r1, Register x2, Register b2, Disp d2); \
+ void name(Register r1, const MemOperand& opnd)
+
+#define RSI_FORM(name) void name(Register r1, Register r3, const Operand& i)
+
+#define RIS_FORM(name) \
+ void name(Register r1, Condition m3, Register b4, Disp d4, \
+ const Operand& i2); \
+ void name(Register r1, const Operand& i2, Condition m3, \
+ const MemOperand& opnd)
+
+#define SI_FORM(name) \
+ void name(const MemOperand& opnd, const Operand& i); \
+ void name(const Operand& i2, Register b1, Disp d1)
+
+#define SIL_FORM(name) \
+ void name(Register b1, Disp d1, const Operand& i2); \
+ void name(const MemOperand& opnd, const Operand& i2)
+
+#define RRE_FORM(name) void name(Register r1, Register r2)
+
+#define RRF1_FORM(name) void name(Register r1, Register r2, Register r3)
+
+#define RRF2_FORM(name) void name(Condition m1, Register r1, Register r2)
+
+#define RRF3_FORM(name) \
+ void name(Register r3, Condition m4, Register r1, Register r2)
+
+#define RS1_FORM(name) \
+ void name(Register r1, Register r3, const MemOperand& opnd); \
+ void name(Register r1, Register r3, Register b2, Disp d2)
+
+#define RS2_FORM(name) \
+ void name(Register r1, Condition m3, const MemOperand& opnd); \
+ void name(Register r1, Condition m3, Register b2, Disp d2)
+
+#define RSE_FORM(name) \
+ void name(Register r1, Register r3, const MemOperand& opnd); \
+ void name(Register r1, Register r3, Register b2, Disp d2)
+
+#define RSL_FORM(name) \
+ void name(Length l, Register b2, Disp d2); \
+ void name(const MemOperand& opnd)
+
+#define RSY1_FORM(name) \
+ void name(Register r1, Register r3, Register b2, Disp d2); \
+ void name(Register r1, Register r3, const MemOperand& opnd)
+
+#define RSY2_FORM(name) \
+ void name(Register r1, Condition m3, Register b2, Disp d2); \
+ void name(Register r1, Condition m3, const MemOperand& opnd)
+
+#define RRD_FORM(name) void name(Register r1, Register r3, Register r2)
+
+#define RRS_FORM(name) \
+ void name(Register r1, Register r2, Register b4, Disp d4, Condition m3); \
+ void name(Register r1, Register r2, Condition m3, const MemOperand& opnd)
+
+#define S_FORM(name) \
+ void name(Register b2, Disp d2); \
+ void name(const MemOperand& opnd)
+
+#define SIY_FORM(name) \
+ void name(const Operand& i2, Register b1, Disp d1); \
+ void name(const MemOperand& opnd, const Operand& i)
+
+#define SS1_FORM(name) \
+ void name(Register b1, Disp d1, Register b3, Disp d2, Length length); \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2, Length length)
+
+#define SS2_FORM(name) \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2, Length length1, \
+ Length length2); \
+ void name(Register b1, Disp d1, Register b2, Disp d2, Length l1, Length l2)
+
+#define SS3_FORM(name) \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2, Length length); \
+ void name(const Operand& i3, Register b1, Disp d1, Register b2, Disp d2, \
+ Length l1)
+
+#define SS4_FORM(name) \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2); \
+ void name(Register r1, Register r3, Register b1, Disp d1, Register b2, \
+ Disp d2)
+
+#define SS5_FORM(name) \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2); \
+ void name(Register r1, Register r3, Register b3, Disp d2, Register b4, \
+ Disp d4)
+
+#define SSE_FORM(name) \
+ void name(Register b1, Disp d1, Register b2, Disp d2); \
+ void name(const MemOperand& opnd1, const MemOperand& opnd2)
+
+#define SSF_FORM(name) \
+ void name(Register r3, Register b1, Disp d1, Register b2, Disp d2); \
+ void name(Register r3, const MemOperand& opnd1, const MemOperand& opnd2)
+
+ // S390 instruction sets
+ RX_FORM(bc);
+ RR_FORM(bctr);
+ RX_FORM(cd);
+ RRE_FORM(cdr);
+ RXE_FORM(cdb);
+ RXE_FORM(ceb);
+ RRE_FORM(cefbr);
+ RXE_FORM(ddb);
+ RRE_FORM(ddbr);
+ SS1_FORM(ed);
+ RRE_FORM(epair);
+ RX_FORM(ex);
+ RRF2_FORM(fidbr);
+ RRE_FORM(flogr);
+ RX_FORM(ic_z);
+ RXY_FORM(icy);
+ RIL1_FORM(iihf);
+ RI1_FORM(iihh);
+ RI1_FORM(iihl);
+ RIL1_FORM(iilf);
+ RI1_FORM(iilh);
+ RI1_FORM(iill);
+ RRE_FORM(lcgr);
+ RR_FORM(lcr);
+ RX_FORM(le_z);
+ RXY_FORM(ley);
+ RIL1_FORM(llihf);
+ RIL1_FORM(llilf);
+ RRE_FORM(lngr);
+ RR_FORM(lnr);
+ RSY1_FORM(loc);
+ RXY_FORM(lrv);
+ RXY_FORM(lrvh);
+ RXE_FORM(mdb);
+ RRE_FORM(mdbr);
+ SS4_FORM(mvck);
+ SSF_FORM(mvcos);
+ SS4_FORM(mvcs);
+ SS1_FORM(mvn);
+ SS1_FORM(nc);
+ SI_FORM(ni);
+ RIL1_FORM(nihf);
+ RIL1_FORM(nilf);
+ RI1_FORM(nilh);
+ RI1_FORM(nill);
+ RIL1_FORM(oihf);
+ RIL1_FORM(oilf);
+ RI1_FORM(oill);
+ RRE_FORM(popcnt);
+ RXE_FORM(sdb);
+ RRE_FORM(sdbr);
+ RIL1_FORM(slfi);
+ RXY_FORM(slgf);
+ RIL1_FORM(slgfi);
+ RS1_FORM(srdl);
+ RX_FORM(ste);
+ RXY_FORM(stey);
+ RXY_FORM(strv);
+ RI1_FORM(tmll);
+ SS1_FORM(tr);
+ S_FORM(ts);
+ RIL1_FORM(xihf);
+ RIL1_FORM(xilf);
+
+ // Load Address Instructions
+ void la(Register r, const MemOperand& opnd);
+ void lay(Register r, const MemOperand& opnd);
+ void larl(Register r1, const Operand& opnd);
+ void larl(Register r, Label* l);
+
+ // Load Instructions
+ void lb(Register r, const MemOperand& src);
+ void lbr(Register r1, Register r2);
+ void lgb(Register r, const MemOperand& src);
+ void lgbr(Register r1, Register r2);
+ void lh(Register r, const MemOperand& src);
+ void lhy(Register r, const MemOperand& src);
+ void lhr(Register r1, Register r2);
+ void lgh(Register r, const MemOperand& src);
+ void lghr(Register r1, Register r2);
+ void l(Register r, const MemOperand& src);
+ void ly(Register r, const MemOperand& src);
+ void lr(Register r1, Register r2);
+ void lg(Register r, const MemOperand& src);
+ void lgr(Register r1, Register r2);
+ void lgf(Register r, const MemOperand& src);
+ void lgfr(Register r1, Register r2);
+ void lhi(Register r, const Operand& imm);
+ void lghi(Register r, const Operand& imm);
+
+ // Load And Test Instructions
+ void lt_z(Register r, const MemOperand& src);
+ void ltg(Register r, const MemOperand& src);
+ void ltr(Register r1, Register r2);
+ void ltgr(Register r1, Register r2);
+ void ltgfr(Register r1, Register r2);
+
+ // Load Logical Instructions
+ void llc(Register r, const MemOperand& src);
+ void llgc(Register r, const MemOperand& src);
+ void llgf(Register r, const MemOperand& src);
+ void llgfr(Register r1, Register r2);
+ void llh(Register r, const MemOperand& src);
+ void llgh(Register r, const MemOperand& src);
+ void llhr(Register r1, Register r2);
+ void llghr(Register r1, Register r2);
+
+ // Load Multiple Instructions
+ void lm(Register r1, Register r2, const MemOperand& src);
+ void lmy(Register r1, Register r2, const MemOperand& src);
+ void lmg(Register r1, Register r2, const MemOperand& src);
+
+ // Store Instructions
+ void st(Register r, const MemOperand& src);
+ void stc(Register r, const MemOperand& src);
+ void stcy(Register r, const MemOperand& src);
+ void stg(Register r, const MemOperand& src);
+ void sth(Register r, const MemOperand& src);
+ void sthy(Register r, const MemOperand& src);
+ void sty(Register r, const MemOperand& src);
+
+ // Store Multiple Instructions
+ void stm(Register r1, Register r2, const MemOperand& src);
+ void stmy(Register r1, Register r2, const MemOperand& src);
+ void stmg(Register r1, Register r2, const MemOperand& src);
+
+ // Compare Instructions
+ void c(Register r, const MemOperand& opnd);
+ void cy(Register r, const MemOperand& opnd);
+ void cr_z(Register r1, Register r2);
+ void cg(Register r, const MemOperand& opnd);
+ void cgr(Register r1, Register r2);
+ void ch(Register r, const MemOperand& opnd);
+ void chy(Register r, const MemOperand& opnd);
+ void chi(Register r, const Operand& opnd);
+ void cghi(Register r, const Operand& opnd);
+ void cfi(Register r, const Operand& opnd);
+ void cgfi(Register r, const Operand& opnd);
+
+ // Compare Logical Instructions
+ void cl(Register r, const MemOperand& opnd);
+ void cly(Register r, const MemOperand& opnd);
+ void clr(Register r1, Register r2);
+ void clg(Register r, const MemOperand& opnd);
+ void clgr(Register r1, Register r2);
+ void clfi(Register r, const Operand& opnd);
+ void clgfi(Register r, const Operand& opnd);
+ void cli(const MemOperand& mem, const Operand& imm);
+ void cliy(const MemOperand& mem, const Operand& imm);
+ void clc(const MemOperand& opnd1, const MemOperand& opnd2, Length length);
+
+ // Test Under Mask Instructions
+ void tm(const MemOperand& mem, const Operand& imm);
+ void tmy(const MemOperand& mem, const Operand& imm);
+
+ // Rotate Instructions
+ void rll(Register r1, Register r3, Register opnd);
+ void rll(Register r1, Register r3, const Operand& opnd);
+ void rll(Register r1, Register r3, Register r2, const Operand& opnd);
+ void rllg(Register r1, Register r3, const Operand& opnd);
+ void rllg(Register r1, Register r3, const Register opnd);
+ void rllg(Register r1, Register r3, Register r2, const Operand& opnd);
+
+ // Shift Instructions (32)
+ void sll(Register r1, Register opnd);
+ void sll(Register r1, const Operand& opnd);
+ void sllk(Register r1, Register r3, Register opnd);
+ void sllk(Register r1, Register r3, const Operand& opnd);
+ void srl(Register r1, Register opnd);
+ void srl(Register r1, const Operand& opnd);
+ void srlk(Register r1, Register r3, Register opnd);
+ void srlk(Register r1, Register r3, const Operand& opnd);
+ void sra(Register r1, Register opnd);
+ void sra(Register r1, const Operand& opnd);
+ void srak(Register r1, Register r3, Register opnd);
+ void srak(Register r1, Register r3, const Operand& opnd);
+ void sla(Register r1, Register opnd);
+ void sla(Register r1, const Operand& opnd);
+ void slak(Register r1, Register r3, Register opnd);
+ void slak(Register r1, Register r3, const Operand& opnd);
+
+ // Shift Instructions (64)
+ void sllg(Register r1, Register r3, const Operand& opnd);
+ void sllg(Register r1, Register r3, const Register opnd);
+ void srlg(Register r1, Register r3, const Operand& opnd);
+ void srlg(Register r1, Register r3, const Register opnd);
+ void srag(Register r1, Register r3, const Operand& opnd);
+ void srag(Register r1, Register r3, const Register opnd);
+ void srda(Register r1, const Operand& opnd);
+ void srdl(Register r1, const Operand& opnd);
+ void slag(Register r1, Register r3, const Operand& opnd);
+ void slag(Register r1, Register r3, const Register opnd);
+
+ // Rotate and Insert Selected Bits
+ void risbg(Register dst, Register src, const Operand& startBit,
+ const Operand& endBit, const Operand& shiftAmt,
+ bool zeroBits = true);
+ void risbgn(Register dst, Register src, const Operand& startBit,
+ const Operand& endBit, const Operand& shiftAmt,
+ bool zeroBits = true);
+
+ // Move Character (Mem to Mem)
+ void mvc(const MemOperand& opnd1, const MemOperand& opnd2, uint32_t length);
+
+ // Branch Instructions
+ void basr(Register r1, Register r2);
+ void bcr(Condition m, Register target);
+ void bct(Register r, const MemOperand& opnd);
+ void bctg(Register r, const MemOperand& opnd);
+ void bras(Register r, const Operand& opnd);
+ void brasl(Register r, const Operand& opnd);
+ void brc(Condition c, const Operand& opnd);
+ void brcl(Condition m, const Operand& opnd, bool isCodeTarget = false);
+ void brct(Register r1, const Operand& opnd);
+ void brctg(Register r1, const Operand& opnd);
+
+ // 32-bit Add Instructions
+ void a(Register r1, const MemOperand& opnd);
+ void ay(Register r1, const MemOperand& opnd);
+ void afi(Register r1, const Operand& opnd);
+ void ah(Register r1, const MemOperand& opnd);
+ void ahy(Register r1, const MemOperand& opnd);
+ void ahi(Register r1, const Operand& opnd);
+ void ahik(Register r1, Register r3, const Operand& opnd);
+ void ar(Register r1, Register r2);
+ void ark(Register r1, Register r2, Register r3);
+ void asi(const MemOperand&, const Operand&);
+
+ // 64-bit Add Instructions
+ void ag(Register r1, const MemOperand& opnd);
+ void agf(Register r1, const MemOperand& opnd);
+ void agfi(Register r1, const Operand& opnd);
+ void agfr(Register r1, Register r2);
+ void aghi(Register r1, const Operand& opnd);
+ void aghik(Register r1, Register r3, const Operand& opnd);
+ void agr(Register r1, Register r2);
+ void agrk(Register r1, Register r2, Register r3);
+ void agsi(const MemOperand&, const Operand&);
+
+ // 32-bit Add Logical Instructions
+ void al_z(Register r1, const MemOperand& opnd);
+ void aly(Register r1, const MemOperand& opnd);
+ void alfi(Register r1, const Operand& opnd);
+ void alr(Register r1, Register r2);
+ void alrk(Register r1, Register r2, Register r3);
+
+ // 64-bit Add Logical Instructions
+ void alg(Register r1, const MemOperand& opnd);
+ void algfi(Register r1, const Operand& opnd);
+ void algr(Register r1, Register r2);
+ void algrk(Register r1, Register r2, Register r3);
+
+ // 32-bit Subtract Instructions
+ void s(Register r1, const MemOperand& opnd);
+ void sy(Register r1, const MemOperand& opnd);
+ void sh(Register r1, const MemOperand& opnd);
+ void shy(Register r1, const MemOperand& opnd);
+ void sr(Register r1, Register r2);
+ void srk(Register r1, Register r2, Register r3);
+
+ // 64-bit Subtract Instructions
+ void sg(Register r1, const MemOperand& opnd);
+ void sgf(Register r1, const MemOperand& opnd);
+ void sgr(Register r1, Register r2);
+ void sgfr(Register r1, Register r2);
+ void sgrk(Register r1, Register r2, Register r3);
+
+ // 32-bit Subtract Logical Instructions
+ void sl(Register r1, const MemOperand& opnd);
+ void sly(Register r1, const MemOperand& opnd);
+ void slr(Register r1, Register r2);
+ void slrk(Register r1, Register r2, Register r3);
+
+ // 64-bit Subtract Logical Instructions
+ void slg(Register r1, const MemOperand& opnd);
+ void slgr(Register r1, Register r2);
+ void slgrk(Register r1, Register r2, Register r3);
+
+ // 32-bit Multiply Instructions
+ void m(Register r1, const MemOperand& opnd);
+ void mr_z(Register r1, Register r2);
+ void ml(Register r1, const MemOperand& opnd);
+ void mlr(Register r1, Register r2);
+ void ms(Register r1, const MemOperand& opnd);
+ void msy(Register r1, const MemOperand& opnd);
+ void msfi(Register r1, const Operand& opnd);
+ void msr(Register r1, Register r2);
+ void mh(Register r1, const MemOperand& opnd);
+ void mhy(Register r1, const MemOperand& opnd);
+ void mhi(Register r1, const Operand& opnd);
+
+ // 64-bit Multiply Instructions
+ void mlg(Register r1, const MemOperand& opnd);
+ void mlgr(Register r1, Register r2);
+ void mghi(Register r1, const Operand& opnd);
+ void msgfi(Register r1, const Operand& opnd);
+ void msg(Register r1, const MemOperand& opnd);
+ void msgr(Register r1, Register r2);
+
+ // 32-bit Divide Instructions
+ void d(Register r1, const MemOperand& opnd);
+ void dr(Register r1, Register r2);
+ void dl(Register r1, const MemOperand& opnd);
+ void dlr(Register r1, Register r2);
+
+ // 64-bit Divide Instructions
+ void dlgr(Register r1, Register r2);
+ void dsgr(Register r1, Register r2);
+
+ // Bitwise Instructions (AND / OR / XOR)
+ void n(Register r1, const MemOperand& opnd);
+ void ny(Register r1, const MemOperand& opnd);
+ void nr(Register r1, Register r2);
+ void nrk(Register r1, Register r2, Register r3);
+ void ng(Register r1, const MemOperand& opnd);
+ void ngr(Register r1, Register r2);
+ void ngrk(Register r1, Register r2, Register r3);
+ void o(Register r1, const MemOperand& opnd);
+ void oy(Register r1, const MemOperand& opnd);
+ void or_z(Register r1, Register r2);
+ void ork(Register r1, Register r2, Register r3);
+ void og(Register r1, const MemOperand& opnd);
+ void ogr(Register r1, Register r2);
+ void ogrk(Register r1, Register r2, Register r3);
+ void x(Register r1, const MemOperand& opnd);
+ void xy(Register r1, const MemOperand& opnd);
+ void xr(Register r1, Register r2);
+ void xrk(Register r1, Register r2, Register r3);
+ void xg(Register r1, const MemOperand& opnd);
+ void xgr(Register r1, Register r2);
+ void xgrk(Register r1, Register r2, Register r3);
+ void xc(const MemOperand& opnd1, const MemOperand& opnd2, Length length);
+
+ // Bitwise GPR <-> FPR Conversion Instructions
+ void lgdr(Register r1, DoubleRegister f2);
+ void ldgr(DoubleRegister f1, Register r2);
+
+ // Floating Point Load / Store Instructions
+ void ld(DoubleRegister r1, const MemOperand& opnd);
+ void ldy(DoubleRegister r1, const MemOperand& opnd);
+ void le_z(DoubleRegister r1, const MemOperand& opnd);
+ void ley(DoubleRegister r1, const MemOperand& opnd);
+ void ldr(DoubleRegister r1, DoubleRegister r2);
+ void ltdbr(DoubleRegister r1, DoubleRegister r2);
+ void ltebr(DoubleRegister r1, DoubleRegister r2);
+ void std(DoubleRegister r1, const MemOperand& opnd);
+ void stdy(DoubleRegister r1, const MemOperand& opnd);
+ void ste(DoubleRegister r1, const MemOperand& opnd);
+ void stey(DoubleRegister r1, const MemOperand& opnd);
+
+ // Floating Point Load Rounded/Positive Instructions
+ void ledbr(DoubleRegister r1, DoubleRegister r2);
+ void ldebr(DoubleRegister r1, DoubleRegister r2);
+ void lpebr(DoubleRegister r1, DoubleRegister r2);
+ void lpdbr(DoubleRegister r1, DoubleRegister r2);
+
+ // Floating <-> Fixed Point Conversion Instructions
+ void cdlfbr(Condition m3, Condition m4, DoubleRegister fltReg,
+ Register fixReg);
+ void cdlgbr(Condition m3, Condition m4, DoubleRegister fltReg,
+ Register fixReg);
+ void celgbr(Condition m3, Condition m4, DoubleRegister fltReg,
+ Register fixReg);
+ void celfbr(Condition m3, Condition m4, DoubleRegister fltReg,
+ Register fixReg);
+ void clfdbr(Condition m3, Condition m4, Register fixReg,
+ DoubleRegister fltReg);
+ void clfebr(Condition m3, Condition m4, Register fixReg,
+ DoubleRegister fltReg);
+ void clgdbr(Condition m3, Condition m4, Register fixReg,
+ DoubleRegister fltReg);
+ void clgebr(Condition m3, Condition m4, Register fixReg,
+ DoubleRegister fltReg);
+ void cfdbr(Condition m, Register fixReg, DoubleRegister fltReg);
+ void cdfbr(DoubleRegister fltReg, Register fixReg);
+ void cgebr(Condition m, Register fixReg, DoubleRegister fltReg);
+ void cgdbr(Condition m, Register fixReg, DoubleRegister fltReg);
+ void cegbr(DoubleRegister fltReg, Register fixReg);
+ void cdgbr(DoubleRegister fltReg, Register fixReg);
+ void cfebr(Condition m3, Register fixReg, DoubleRegister fltReg);
+ void cefbr(DoubleRegister fltReg, Register fixReg);
+
+ // Floating Point Compare Instructions
+ void cebr(DoubleRegister r1, DoubleRegister r2);
+ void cdb(DoubleRegister r1, const MemOperand& opnd);
+ void cdbr(DoubleRegister r1, DoubleRegister r2);
+
+ // Floating Point Arithmetic Instructions
+ void aebr(DoubleRegister r1, DoubleRegister r2);
+ void adb(DoubleRegister r1, const MemOperand& opnd);
+ void adbr(DoubleRegister r1, DoubleRegister r2);
+ void lzdr(DoubleRegister r1);
+ void sebr(DoubleRegister r1, DoubleRegister r2);
+ void sdb(DoubleRegister r1, const MemOperand& opnd);
+ void sdbr(DoubleRegister r1, DoubleRegister r2);
+ void meebr(DoubleRegister r1, DoubleRegister r2);
+ void mdb(DoubleRegister r1, const MemOperand& opnd);
+ void mdbr(DoubleRegister r1, DoubleRegister r2);
+ void debr(DoubleRegister r1, DoubleRegister r2);
+ void ddb(DoubleRegister r1, const MemOperand& opnd);
+ void ddbr(DoubleRegister r1, DoubleRegister r2);
+ void madbr(DoubleRegister r1, DoubleRegister r2, DoubleRegister r3);
+ void msdbr(DoubleRegister r1, DoubleRegister r2, DoubleRegister r3);
+ void sqebr(DoubleRegister r1, DoubleRegister r2);
+ void sqdb(DoubleRegister r1, const MemOperand& opnd);
+ void sqdbr(DoubleRegister r1, DoubleRegister r2);
+ void lcdbr(DoubleRegister r1, DoubleRegister r2);
+ void ldeb(DoubleRegister r1, const MemOperand& opnd);
+
+ enum FIDBRA_MASK3 {
+ FIDBRA_CURRENT_ROUNDING_MODE = 0,
+ FIDBRA_ROUND_TO_NEAREST_AWAY_FROM_0 = 1,
+ // ...
+ FIDBRA_ROUND_TOWARD_0 = 5,
+ FIDBRA_ROUND_TOWARD_POS_INF = 6,
+ FIDBRA_ROUND_TOWARD_NEG_INF = 7
+ };
+ void fiebra(DoubleRegister d1, DoubleRegister d2, FIDBRA_MASK3 m3);
+ void fidbra(DoubleRegister d1, DoubleRegister d2, FIDBRA_MASK3 m3);
+
+ // Move integer
+ void mvhi(const MemOperand& opnd1, const Operand& i2);
+ void mvghi(const MemOperand& opnd1, const Operand& i2);
+
+ // Exception-generating instructions and debugging support
+ void stop(const char* msg, Condition cond = al,
+ int32_t code = kDefaultStopCode, CRegister cr = cr7);
+
+ void bkpt(uint32_t imm16); // v5 and above
+
+ // Different nop operations are used by the code generator to detect certain
+ // states of the generated code.
+ enum NopMarkerTypes {
+ NON_MARKING_NOP = 0,
+ GROUP_ENDING_NOP,
+ DEBUG_BREAK_NOP,
+ // IC markers.
+ PROPERTY_ACCESS_INLINED,
+ PROPERTY_ACCESS_INLINED_CONTEXT,
+ PROPERTY_ACCESS_INLINED_CONTEXT_DONT_DELETE,
+ // Helper values.
+ LAST_CODE_MARKER,
+ FIRST_IC_MARKER = PROPERTY_ACCESS_INLINED
+ };
+
+ void nop(int type = 0); // 0 is the default non-marking type.
+
+ // Check the code size generated from label to here.
+ int SizeOfCodeGeneratedSince(Label* label) {
+ return pc_offset() - label->pos();
+ }
+
+ // Debugging
+
+ // Mark generator continuation.
+ void RecordGeneratorContinuation();
+
+ // Mark address of a debug break slot.
+ void RecordDebugBreakSlot(RelocInfo::Mode mode);
+
+ // Record the AST id of the CallIC being compiled, so that it can be placed
+ // in the relocation information.
+ void SetRecordedAstId(TypeFeedbackId ast_id) { recorded_ast_id_ = ast_id; }
+
+ TypeFeedbackId RecordedAstId() {
+ // roohack - another issue??? DCHECK(!recorded_ast_id_.IsNone());
+ return recorded_ast_id_;
+ }
+
+ void ClearRecordedAstId() { recorded_ast_id_ = TypeFeedbackId::None(); }
+
+ // Record a comment relocation entry that can be used by a disassembler.
+ // Use --code-comments to enable.
+ void RecordComment(const char* msg);
+
+ // Record a deoptimization reason that can be used by a log or cpu profiler.
+ // Use --trace-deopt to enable.
+ void RecordDeoptReason(const int reason, int raw_position);
+
+ // Writes a single byte or word of data in the code stream. Used
+ // for inline tables, e.g., jump-tables.
+ void db(uint8_t data);
+ void dd(uint32_t data);
+ void dq(uint64_t data);
+ void dp(uintptr_t data);
+
+ PositionsRecorder* positions_recorder() { return &positions_recorder_; }
+
+ void PatchConstantPoolAccessInstruction(int pc_offset, int offset,
+ ConstantPoolEntry::Access access,
+ ConstantPoolEntry::Type type) {
+ // No embedded constant pool support.
+ UNREACHABLE();
+ }
+
+ // Read/patch instructions
+ SixByteInstr instr_at(int pos) {
+ return Instruction::InstructionBits(buffer_ + pos);
+ }
+ template <typename T>
+ void instr_at_put(int pos, T instr) {
+ Instruction::SetInstructionBits<T>(buffer_ + pos, instr);
+ }
+
+ // Decodes instruction at pos, and returns its length
+ int32_t instr_length_at(int pos) {
+ return Instruction::InstructionLength(buffer_ + pos);
+ }
+
+ static SixByteInstr instr_at(byte* pc) {
+ return Instruction::InstructionBits(pc);
+ }
+
+ static Condition GetCondition(Instr instr);
+
+ static bool IsBranch(Instr instr);
+#if V8_TARGET_ARCH_S390X
+ static bool Is64BitLoadIntoIP(SixByteInstr instr1, SixByteInstr instr2);
+#else
+ static bool Is32BitLoadIntoIP(SixByteInstr instr);
+#endif
+
+ static bool IsCmpRegister(Instr instr);
+ static bool IsCmpImmediate(Instr instr);
+ static bool IsNop(SixByteInstr instr, int type = NON_MARKING_NOP);
+
+ // The code currently calls CheckBuffer() too often. This has the side
+ // effect of randomly growing the buffer in the middle of multi-instruction
+ // sequences.
+ //
+ // This function allows outside callers to check and grow the buffer
+ void EnsureSpaceFor(int space_needed);
+
+ void EmitRelocations();
+ void emit_label_addr(Label* label);
+
+ public:
+ byte* buffer_pos() const { return buffer_; }
+
+ protected:
+ // Relocation for a type-recording IC has the AST id added to it. This
+ // member variable is a way to pass the information from the call site to
+ // the relocation info.
+ TypeFeedbackId recorded_ast_id_;
+
+ int buffer_space() const { return reloc_info_writer.pos() - pc_; }
+
+ // Decode instruction(s) at pos and return backchain to previous
+ // label reference or kEndOfChain.
+ int target_at(int pos);
+
+ // Patch instruction(s) at pos to target target_pos (e.g. branch)
+ void target_at_put(int pos, int target_pos, bool* is_branch = nullptr);
+
+ // Record reloc info for current pc_
+ void RecordRelocInfo(RelocInfo::Mode rmode, intptr_t data = 0);
+
+ private:
+ // Code generation
+ // The relocation writer's position is at least kGap bytes below the end of
+ // the generated instructions. This is so that multi-instruction sequences do
+ // not have to check for overflow. The same is true for writes of large
+ // relocation info entries.
+ static const int kGap = 32;
+
+ // Relocation info generation
+ // Each relocation is encoded as a variable size value
+ static const int kMaxRelocSize = RelocInfoWriter::kMaxSize;
+ RelocInfoWriter reloc_info_writer;
+ std::vector<DeferredRelocInfo> relocations_;
+
+ // The bound position, before this we cannot do instruction elimination.
+ int last_bound_pos_;
+
+ // Code emission
+ inline void CheckBuffer();
+ void GrowBuffer(int needed = 0);
+ inline void TrackBranch();
+ inline void UntrackBranch();
+
+ inline int32_t emit_code_target(
+ Handle<Code> target, RelocInfo::Mode rmode,
+ TypeFeedbackId ast_id = TypeFeedbackId::None());
+
+ // Helpers to emit binary encoding of 2/4/6 byte instructions.
+ inline void emit2bytes(uint16_t x);
+ inline void emit4bytes(uint32_t x);
+ inline void emit6bytes(uint64_t x);
+
+ // Helpers to emit binary encoding for various instruction formats.
+
+ inline void rr_form(Opcode op, Register r1, Register r2);
+ inline void rr_form(Opcode op, DoubleRegister r1, DoubleRegister r2);
+ inline void rr_form(Opcode op, Condition m1, Register r2);
+ inline void rr2_form(uint8_t op, Condition m1, Register r2);
+
+ inline void rx_form(Opcode op, Register r1, Register x2, Register b2,
+ Disp d2);
+ inline void rx_form(Opcode op, DoubleRegister r1, Register x2, Register b2,
+ Disp d2);
+
+ inline void ri_form(Opcode op, Register r1, const Operand& i2);
+ inline void ri_form(Opcode op, Condition m1, const Operand& i2);
+
+ inline void rie_form(Opcode op, Register r1, Register r3, const Operand& i2);
+ inline void rie_f_form(Opcode op, Register r1, Register r2, const Operand& i3,
+ const Operand& i4, const Operand& i5);
+
+ inline void ril_form(Opcode op, Register r1, const Operand& i2);
+ inline void ril_form(Opcode op, Condition m1, const Operand& i2);
+
+ inline void ris_form(Opcode op, Register r1, Condition m3, Register b4,
+ Disp d4, const Operand& i2);
+
+ inline void rrd_form(Opcode op, Register r1, Register r3, Register r2);
+
+ inline void rre_form(Opcode op, Register r1, Register r2);
+ inline void rre_form(Opcode op, DoubleRegister r1, DoubleRegister r2);
+
+ inline void rrf1_form(Opcode op, Register r1, Register r2, Register r3);
+ inline void rrf1_form(uint32_t x);
+ inline void rrf2_form(uint32_t x);
+ inline void rrf3_form(uint32_t x);
+ inline void rrfe_form(Opcode op, Condition m3, Condition m4, Register r1,
+ Register r2);
+
+ inline void rrs_form(Opcode op, Register r1, Register r2, Register b4,
+ Disp d4, Condition m3);
+
+ inline void rs_form(Opcode op, Register r1, Condition m3, Register b2,
+ const Disp d2);
+ inline void rs_form(Opcode op, Register r1, Register r3, Register b2,
+ const Disp d2);
+
+ inline void rsi_form(Opcode op, Register r1, Register r3, const Operand& i2);
+ inline void rsl_form(Opcode op, Length l1, Register b2, Disp d2);
+
+ inline void rsy_form(Opcode op, Register r1, Register r3, Register b2,
+ const Disp d2);
+ inline void rsy_form(Opcode op, Register r1, Condition m3, Register b2,
+ const Disp d2);
+
+ inline void rxe_form(Opcode op, Register r1, Register x2, Register b2,
+ Disp d2);
+
+ inline void rxf_form(Opcode op, Register r1, Register r3, Register b2,
+ Register x2, Disp d2);
+
+ inline void rxy_form(Opcode op, Register r1, Register x2, Register b2,
+ Disp d2);
+ inline void rxy_form(Opcode op, DoubleRegister r1, Register x2, Register b2,
+ Disp d2);
+
+ inline void s_form(Opcode op, Register b1, Disp d2);
+
+ inline void si_form(Opcode op, const Operand& i2, Register b1, Disp d1);
+ inline void siy_form(Opcode op, const Operand& i2, Register b1, Disp d1);
+
+ inline void sil_form(Opcode op, Register b1, Disp d1, const Operand& i2);
+
+ inline void ss_form(Opcode op, Length l, Register b1, Disp d1, Register b2,
+ Disp d2);
+ inline void ss_form(Opcode op, Length l1, Length l2, Register b1, Disp d1,
+ Register b2, Disp d2);
+ inline void ss_form(Opcode op, Length l1, const Operand& i3, Register b1,
+ Disp d1, Register b2, Disp d2);
+ inline void ss_form(Opcode op, Register r1, Register r2, Register b1, Disp d1,
+ Register b2, Disp d2);
+ inline void sse_form(Opcode op, Register b1, Disp d1, Register b2, Disp d2);
+ inline void ssf_form(Opcode op, Register r3, Register b1, Disp d1,
+ Register b2, Disp d2);
+
+ // Labels
+ void print(Label* L);
+ int max_reach_from(int pos);
+ void bind_to(Label* L, int pos);
+ void next(Label* L);
+
+ friend class RegExpMacroAssemblerS390;
+ friend class RelocInfo;
+ friend class CodePatcher;
+
+ List<Handle<Code> > code_targets_;
+
+ PositionsRecorder positions_recorder_;
+ friend class PositionsRecorder;
+ friend class EnsureSpace;
+};
+
+class EnsureSpace BASE_EMBEDDED {
+ public:
+ explicit EnsureSpace(Assembler* assembler) { assembler->CheckBuffer(); }
+};
+
+} // namespace internal
+} // namespace v8
+
+#endif // V8_S390_ASSEMBLER_S390_H_
« no previous file with comments | « src/base/macros.h ('k') | src/s390/assembler-s390.cc » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698