Generated code for substring slices in x64 and arm.

Review URL: http://codereview.chromium.org/7795018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9111 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

diff --git a/src/arm/code-stubs-arm.cc b/src/arm/code-stubs-arm.cc
index 09d2c17..f6b87c3 100644 (file)
--- a/src/arm/code-stubs-arm.cc
+++ b/src/arm/code-stubs-arm.cc
@@ -4894,7 +4894,8 @@ void StringCharCodeAtGenerator::GenerateFast(MacroAssembler* masm) {
 
   // Check for 1-byte or 2-byte string.
   __ bind(&flat_string);
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ tst(result_, Operand(kStringEncodingMask));
   __ b(ne, &ascii_string);
 
@@ -5468,11 +5469,6 @@ void SubStringStub::Generate(MacroAssembler* masm) {
   Register to = r6;
   Register from = r7;
 
-  if (FLAG_string_slices) {
-    __ nop(0);  // Jumping as first instruction would crash the code generation.
-    __ jmp(&runtime);
-  }
-
   __ Ldrd(to, from, MemOperand(sp, kToOffset));
   STATIC_ASSERT(kFromOffset == kToOffset + 4);
   STATIC_ASSERT(kSmiTag == 0);
@@ -5490,64 +5486,79 @@ void SubStringStub::Generate(MacroAssembler* masm) {
   __ b(mi, &runtime);  // Fail if from > to.
   // Special handling of sub-strings of length 1 and 2. One character strings
   // are handled in the runtime system (looked up in the single character
-  // cache). Two character strings are looked for in the symbol cache.
+  // cache). Two character strings are looked for in the symbol cache in
+  // generated code.
   __ cmp(r2, Operand(2));
   __ b(lt, &runtime);
 
-  // r2: length
-  // r3: from index (untaged smi)
+  // r2: result string length
+  // r3: from index (untagged smi)
   // r6 (a.k.a. to): to (smi)
   // r7 (a.k.a. from): from offset (smi)
-
   // Make sure first argument is a sequential (or flat) string.
-  __ ldr(r5, MemOperand(sp, kStringOffset));
+  __ ldr(r0, MemOperand(sp, kStringOffset));
   STATIC_ASSERT(kSmiTag == 0);
-  __ JumpIfSmi(r5, &runtime);
-  Condition is_string = masm->IsObjectStringType(r5, r1);
+  __ JumpIfSmi(r0, &runtime);
+  Condition is_string = masm->IsObjectStringType(r0, r1);
   __ b(NegateCondition(is_string), &runtime);
 
+  // Short-cut for the case of trivial substring.
+  Label return_r0;
+  // r0: original string
+  // r2: result string length
+  __ ldr(r4, FieldMemOperand(r0, String::kLengthOffset));
+  __ cmp(r2, Operand(r4, ASR, 1));
+  __ b(eq, &return_r0);
+
+  Label create_slice;
+  if (FLAG_string_slices) {
+    __ cmp(r2, Operand(SlicedString::kMinLength));
+    __ b(ge, &create_slice);
+  }
+
+  // r0: original string
   // r1: instance type
-  // r2: length
+  // r2: result string length
   // r3: from index (untagged smi)
-  // r5: string
   // r6 (a.k.a. to): to (smi)
   // r7 (a.k.a. from): from offset (smi)
   Label seq_string;
   __ and_(r4, r1, Operand(kStringRepresentationMask));
   STATIC_ASSERT(kSeqStringTag < kConsStringTag);
   STATIC_ASSERT(kConsStringTag < kExternalStringTag);
+  STATIC_ASSERT(kConsStringTag < kSlicedStringTag);
   __ cmp(r4, Operand(kConsStringTag));
-  __ b(gt, &runtime);  // External strings go to runtime.
+  __ b(gt, &runtime);  // Slices and external strings go to runtime.
   __ b(lt, &seq_string);  // Sequential strings are handled directly.
 
   // Cons string. Try to recurse (once) on the first substring.
   // (This adds a little more generality than necessary to handle flattened
   // cons strings, but not much).
-  __ ldr(r5, FieldMemOperand(r5, ConsString::kFirstOffset));
-  __ ldr(r4, FieldMemOperand(r5, HeapObject::kMapOffset));
+  __ ldr(r0, FieldMemOperand(r0, ConsString::kFirstOffset));
+  __ ldr(r4, FieldMemOperand(r0, HeapObject::kMapOffset));
   __ ldrb(r1, FieldMemOperand(r4, Map::kInstanceTypeOffset));
   __ tst(r1, Operand(kStringRepresentationMask));
   STATIC_ASSERT(kSeqStringTag == 0);
-  __ b(ne, &runtime);  // Cons and External strings go to runtime.
+  __ b(ne, &runtime);  // Cons, slices and external strings go to runtime.
 
   // Definitly a sequential string.
   __ bind(&seq_string);
 
-  // r1: instance type.
-  // r2: length
-  // r3: from index (untaged smi)
-  // r5: string
+  // r0: original string
+  // r1: instance type
+  // r2: result string length
+  // r3: from index (untagged smi)
   // r6 (a.k.a. to): to (smi)
   // r7 (a.k.a. from): from offset (smi)
-  __ ldr(r4, FieldMemOperand(r5, String::kLengthOffset));
+  __ ldr(r4, FieldMemOperand(r0, String::kLengthOffset));
   __ cmp(r4, Operand(to));
   __ b(lt, &runtime);  // Fail if to > length.
   to = no_reg;
 
-  // r1: instance type.
-  // r2: result string length.
-  // r3: from index (untaged smi)
-  // r5: string.
+  // r0: original string or left hand side of the original cons string.
+  // r1: instance type
+  // r2: result string length
+  // r3: from index (untagged smi)
   // r7 (a.k.a. from): from offset (smi)
   // Check for flat ASCII string.
   Label non_ascii_flat;
@@ -5561,82 +5572,146 @@ void SubStringStub::Generate(MacroAssembler* masm) {
 
   // Sub string of length 2 requested.
   // Get the two characters forming the sub string.
-  __ add(r5, r5, Operand(r3));
-  __ ldrb(r3, FieldMemOperand(r5, SeqAsciiString::kHeaderSize));
-  __ ldrb(r4, FieldMemOperand(r5, SeqAsciiString::kHeaderSize + 1));
+  __ add(r0, r0, Operand(r3));
+  __ ldrb(r3, FieldMemOperand(r0, SeqAsciiString::kHeaderSize));
+  __ ldrb(r4, FieldMemOperand(r0, SeqAsciiString::kHeaderSize + 1));
 
   // Try to lookup two character string in symbol table.
   Label make_two_character_string;
   StringHelper::GenerateTwoCharacterSymbolTableProbe(
       masm, r3, r4, r1, r5, r6, r7, r9, &make_two_character_string);
   Counters* counters = masm->isolate()->counters();
-  __ IncrementCounter(counters->sub_string_native(), 1, r3, r4);
-  __ add(sp, sp, Operand(3 * kPointerSize));
-  __ Ret();
+  __ jmp(&return_r0);
 
   // r2: result string length.
   // r3: two characters combined into halfword in little endian byte order.
   __ bind(&make_two_character_string);
   __ AllocateAsciiString(r0, r2, r4, r5, r9, &runtime);
   __ strh(r3, FieldMemOperand(r0, SeqAsciiString::kHeaderSize));
-  __ IncrementCounter(counters->sub_string_native(), 1, r3, r4);
-  __ add(sp, sp, Operand(3 * kPointerSize));
-  __ Ret();
+  __ jmp(&return_r0);
 
   __ bind(&result_longer_than_two);
 
+  // Locate 'from' character of string.
+  __ add(r5, r0, Operand(SeqAsciiString::kHeaderSize - kHeapObjectTag));
+  __ add(r5, r5, Operand(from, ASR, 1));
+
   // Allocate the result.
   __ AllocateAsciiString(r0, r2, r3, r4, r1, &runtime);
 
-  // r0: result string.
-  // r2: result string length.
-  // r5: string.
+  // r0: result string
+  // r2: result string length
+  // r5: first character of substring to copy
   // r7 (a.k.a. from): from offset (smi)
   // Locate first character of result.
   __ add(r1, r0, Operand(SeqAsciiString::kHeaderSize - kHeapObjectTag));
-  // Locate 'from' character of string.
-  __ add(r5, r5, Operand(SeqAsciiString::kHeaderSize - kHeapObjectTag));
-  __ add(r5, r5, Operand(from, ASR, 1));
 
-  // r0: result string.
-  // r1: first character of result string.
-  // r2: result string length.
-  // r5: first character of sub string to copy.
+  // r0: result string
+  // r1: first character of result string
+  // r2: result string length
+  // r5: first character of substring to copy
   STATIC_ASSERT((SeqAsciiString::kHeaderSize & kObjectAlignmentMask) == 0);
   StringHelper::GenerateCopyCharactersLong(masm, r1, r5, r2, r3, r4, r6, r7, r9,
                                            COPY_ASCII | DEST_ALWAYS_ALIGNED);
-  __ IncrementCounter(counters->sub_string_native(), 1, r3, r4);
-  __ add(sp, sp, Operand(3 * kPointerSize));
-  __ Ret();
+  __ jmp(&return_r0);
 
   __ bind(&non_ascii_flat);
-  // r2: result string length.
-  // r5: string.
+  // r0: original string
+  // r2: result string length
   // r7 (a.k.a. from): from offset (smi)
   // Check for flat two byte string.
 
+  // Locate 'from' character of string.
+  __ add(r5, r0, Operand(SeqTwoByteString::kHeaderSize - kHeapObjectTag));
+  // As "from" is a smi it is 2 times the value which matches the size of a two
+  // byte character.
+  STATIC_ASSERT(kSmiTagSize == 1 && kSmiTag == 0);
+  __ add(r5, r5, Operand(from));
+
   // Allocate the result.
   __ AllocateTwoByteString(r0, r2, r1, r3, r4, &runtime);
 
-  // r0: result string.
-  // r2: result string length.
-  // r5: string.
+  // r0: result string
+  // r2: result string length
+  // r5: first character of substring to copy
   // Locate first character of result.
   __ add(r1, r0, Operand(SeqTwoByteString::kHeaderSize - kHeapObjectTag));
-  // Locate 'from' character of string.
-  __ add(r5, r5, Operand(SeqTwoByteString::kHeaderSize - kHeapObjectTag));
-  // As "from" is a smi it is 2 times the value which matches the size of a two
-  // byte character.
-  __ add(r5, r5, Operand(from));
+
   from = no_reg;
 
   // r0: result string.
   // r1: first character of result.
   // r2: result length.
-  // r5: first character of string to copy.
+  // r5: first character of substring to copy.
   STATIC_ASSERT((SeqTwoByteString::kHeaderSize & kObjectAlignmentMask) == 0);
   StringHelper::GenerateCopyCharactersLong(
       masm, r1, r5, r2, r3, r4, r6, r7, r9, DEST_ALWAYS_ALIGNED);
+  __ jmp(&return_r0);
+
+  if (FLAG_string_slices) {
+    __ bind(&create_slice);
+    // r0: original string
+    // r1: instance type
+    // r2: length
+    // r3: from index (untagged smi)
+    // r6 (a.k.a. to): to (smi)
+    // r7 (a.k.a. from): from offset (smi)
+    Label allocate_slice, sliced_string, seq_string;
+    STATIC_ASSERT(kSeqStringTag == 0);
+    __ tst(r1, Operand(kStringRepresentationMask));
+    __ b(eq, &seq_string);
+    STATIC_ASSERT(kIsIndirectStringMask == (kSlicedStringTag & kConsStringTag));
+    STATIC_ASSERT(kIsIndirectStringMask != 0);
+    __ tst(r1, Operand(kIsIndirectStringMask));
+    // External string.  Jump to runtime.
+    __ b(eq, &runtime);
+
+    __ tst(r1, Operand(kSlicedNotConsMask));
+    __ b(ne, &sliced_string);
+    // Cons string.  Check whether it is flat, then fetch first part.
+    __ ldr(r5, FieldMemOperand(r0, ConsString::kSecondOffset));
+    __ LoadRoot(r9, Heap::kEmptyStringRootIndex);
+    __ cmp(r5, r9);
+    __ b(ne, &runtime);
+    __ ldr(r5, FieldMemOperand(r0, ConsString::kFirstOffset));
+    __ jmp(&allocate_slice);
+
+    __ bind(&sliced_string);
+    // Sliced string.  Fetch parent and correct start index by offset.
+    __ ldr(r5, FieldMemOperand(r0, SlicedString::kOffsetOffset));
+    __ add(r7, r7, r5);
+    __ ldr(r5, FieldMemOperand(r0, SlicedString::kParentOffset));
+    __ jmp(&allocate_slice);
+
+    __ bind(&seq_string);
+    // Sequential string.  Just move string to the right register.
+    __ mov(r5, r0);
+
+    __ bind(&allocate_slice);
+    // r1: instance type of original string
+    // r2: length
+    // r5: underlying subject string
+    // r7 (a.k.a. from): from offset (smi)
+    // Allocate new sliced string.  At this point we do not reload the instance
+    // type including the string encoding because we simply rely on the info
+    // provided by the original string.  It does not matter if the original
+    // string's encoding is wrong because we always have to recheck encoding of
+    // the newly created string's parent anyways due to externalized strings.
+    Label two_byte_slice, set_slice_header;
+    STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+    STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+    __ tst(r1, Operand(kStringEncodingMask));
+    __ b(eq, &two_byte_slice);
+    __ AllocateAsciiSlicedString(r0, r2, r3, r4, &runtime);
+    __ jmp(&set_slice_header);
+    __ bind(&two_byte_slice);
+    __ AllocateTwoByteSlicedString(r0, r2, r3, r4, &runtime);
+    __ bind(&set_slice_header);
+    __ str(r7, FieldMemOperand(r0, SlicedString::kOffsetOffset));
+    __ str(r5, FieldMemOperand(r0, SlicedString::kParentOffset));
+  }
+
+  __ bind(&return_r0);
   __ IncrementCounter(counters->sub_string_native(), 1, r3, r4);
   __ add(sp, sp, Operand(3 * kPointerSize));
   __ Ret();

diff --git a/src/arm/lithium-codegen-arm.cc b/src/arm/lithium-codegen-arm.cc
index 976576b..03ea929 100644 (file)
--- a/src/arm/lithium-codegen-arm.cc
+++ b/src/arm/lithium-codegen-arm.cc
@@ -3509,7 +3509,8 @@ void LCodeGen::DoStringCharCodeAt(LStringCharCodeAt* instr) {
 
   // Dispatch on the encoding: ASCII or two-byte.
   Label ascii_string;
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ tst(result, Operand(kStringEncodingMask));
   __ b(ne, &ascii_string);
 

diff --git a/src/arm/macro-assembler-arm.cc b/src/arm/macro-assembler-arm.cc
index 88477bb..613a1f6 100644 (file)
--- a/src/arm/macro-assembler-arm.cc
+++ b/src/arm/macro-assembler-arm.cc
@@ -1725,6 +1725,46 @@ void MacroAssembler::AllocateAsciiConsString(Register result,
 }
 
 
+void MacroAssembler::AllocateTwoByteSlicedString(Register result,
+                                                 Register length,
+                                                 Register scratch1,
+                                                 Register scratch2,
+                                                 Label* gc_required) {
+  AllocateInNewSpace(SlicedString::kSize,
+                     result,
+                     scratch1,
+                     scratch2,
+                     gc_required,
+                     TAG_OBJECT);
+
+  InitializeNewString(result,
+                      length,
+                      Heap::kSlicedStringMapRootIndex,
+                      scratch1,
+                      scratch2);
+}
+
+
+void MacroAssembler::AllocateAsciiSlicedString(Register result,
+                                               Register length,
+                                               Register scratch1,
+                                               Register scratch2,
+                                               Label* gc_required) {
+  AllocateInNewSpace(SlicedString::kSize,
+                     result,
+                     scratch1,
+                     scratch2,
+                     gc_required,
+                     TAG_OBJECT);
+
+  InitializeNewString(result,
+                      length,
+                      Heap::kSlicedAsciiStringMapRootIndex,
+                      scratch1,
+                      scratch2);
+}
+
+
 void MacroAssembler::CompareObjectType(Register object,
                                        Register map,
                                        Register type_reg,

diff --git a/src/arm/macro-assembler-arm.h b/src/arm/macro-assembler-arm.h
index 9c653ad..9d66359 100644 (file)
--- a/src/arm/macro-assembler-arm.h
+++ b/src/arm/macro-assembler-arm.h
@@ -532,6 +532,16 @@ class MacroAssembler: public Assembler {
                                Register scratch1,
                                Register scratch2,
                                Label* gc_required);
+  void AllocateTwoByteSlicedString(Register result,
+                                   Register length,
+                                   Register scratch1,
+                                   Register scratch2,
+                                   Label* gc_required);
+  void AllocateAsciiSlicedString(Register result,
+                                 Register length,
+                                 Register scratch1,
+                                 Register scratch2,
+                                 Label* gc_required);
 
   // Allocates a heap number or jumps to the gc_required label if the young
   // space is full and a scavenge is needed. All registers are clobbered also

diff --git a/src/ia32/code-stubs-ia32.cc b/src/ia32/code-stubs-ia32.cc
index 85e74b8..f3eb09f 100644 (file)
--- a/src/ia32/code-stubs-ia32.cc
+++ b/src/ia32/code-stubs-ia32.cc
@@ -4907,7 +4907,8 @@ void StringCharCodeAtGenerator::GenerateFast(MacroAssembler* masm) {
 
   // Check for 1-byte or 2-byte string.
   __ bind(&flat_string);
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ test(result_, Immediate(kStringEncodingMask));
   __ j(not_zero, &ascii_string, Label::kNear);
 
@@ -5178,8 +5179,9 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   __ mov(edi, FieldOperand(edx, HeapObject::kMapOffset));
   __ movzx_b(edi, FieldOperand(edi, Map::kInstanceTypeOffset));
   __ and_(ecx, Operand(edi));
-  STATIC_ASSERT(kStringEncodingMask == kAsciiStringTag);
-  __ test(ecx, Immediate(kAsciiStringTag));
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+  __ test(ecx, Immediate(kStringEncodingMask));
   __ j(zero, &non_ascii);
   __ bind(&ascii_data);
   // Allocate an acsii cons string.
@@ -5210,7 +5212,7 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   __ cmp(edi, kAsciiStringTag | kAsciiDataHintTag);
   __ j(equal, &ascii_data);
   // Allocate a two byte cons string.
-  __ AllocateConsString(ecx, edi, no_reg, &string_add_runtime);
+  __ AllocateTwoByteConsString(ecx, edi, no_reg, &string_add_runtime);
   __ jmp(&allocated);
 
   // Handle creating a flat result. First check that both strings are not
@@ -5236,12 +5238,13 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   // ebx: length of resulting flat string as a smi
   // edx: second string
   Label non_ascii_string_add_flat_result;
-  STATIC_ASSERT(kStringEncodingMask == kAsciiStringTag);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ mov(ecx, FieldOperand(eax, HeapObject::kMapOffset));
-  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kAsciiStringTag);
+  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kStringEncodingMask);
   __ j(zero, &non_ascii_string_add_flat_result);
   __ mov(ecx, FieldOperand(edx, HeapObject::kMapOffset));
-  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kAsciiStringTag);
+  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kStringEncodingMask);
   __ j(zero, &string_add_runtime);
 
   // Both strings are ascii strings.  As they are short they are both flat.
@@ -5281,7 +5284,7 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   // edx: second string
   __ bind(&non_ascii_string_add_flat_result);
   __ mov(ecx, FieldOperand(edx, HeapObject::kMapOffset));
-  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kAsciiStringTag);
+  __ test_b(FieldOperand(ecx, Map::kInstanceTypeOffset), kStringEncodingMask);
   __ j(not_zero, &string_add_runtime);
   // Both strings are two byte strings. As they are short they are both
   // flat.
@@ -5759,13 +5762,14 @@ void SubStringStub::Generate(MacroAssembler* masm) {
     // string's encoding is wrong because we always have to recheck encoding of
     // the newly created string's parent anyways due to externalized strings.
     Label two_byte_slice, set_slice_header;
-    STATIC_ASSERT(kAsciiStringTag != 0);
-    __ test(ebx, Immediate(kAsciiStringTag));
+    STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+    STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+    __ test(ebx, Immediate(kStringEncodingMask));
     __ j(zero, &two_byte_slice, Label::kNear);
     __ AllocateAsciiSlicedString(eax, ebx, no_reg, &runtime);
     __ jmp(&set_slice_header, Label::kNear);
     __ bind(&two_byte_slice);
-    __ AllocateSlicedString(eax, ebx, no_reg, &runtime);
+    __ AllocateTwoByteSlicedString(eax, ebx, no_reg, &runtime);
     __ bind(&set_slice_header);
     __ mov(FieldOperand(eax, SlicedString::kOffsetOffset), edx);
     __ SmiTag(ecx);

diff --git a/src/ia32/lithium-codegen-ia32.cc b/src/ia32/lithium-codegen-ia32.cc
index 32e3074..7136105 100644 (file)
--- a/src/ia32/lithium-codegen-ia32.cc
+++ b/src/ia32/lithium-codegen-ia32.cc
@@ -3269,7 +3269,8 @@ void LCodeGen::DoStringCharCodeAt(LStringCharCodeAt* instr) {
 
   // Dispatch on the encoding: ASCII or two-byte.
   Label ascii_string;
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ test(result, Immediate(kStringEncodingMask));
   __ j(not_zero, &ascii_string, Label::kNear);
 

diff --git a/src/ia32/macro-assembler-ia32.cc b/src/ia32/macro-assembler-ia32.cc
index dff174c..6d3ce2b 100644 (file)
--- a/src/ia32/macro-assembler-ia32.cc
+++ b/src/ia32/macro-assembler-ia32.cc
@@ -1172,7 +1172,7 @@ void MacroAssembler::AllocateAsciiString(Register result,
 }
 
 
-void MacroAssembler::AllocateConsString(Register result,
+void MacroAssembler::AllocateTwoByteConsString(Register result,
                                         Register scratch1,
                                         Register scratch2,
                                         Label* gc_required) {
@@ -1208,7 +1208,7 @@ void MacroAssembler::AllocateAsciiConsString(Register result,
 }
 
 
-void MacroAssembler::AllocateSlicedString(Register result,
+void MacroAssembler::AllocateTwoByteSlicedString(Register result,
                                           Register scratch1,
                                           Register scratch2,
                                           Label* gc_required) {

diff --git a/src/ia32/macro-assembler-ia32.h b/src/ia32/macro-assembler-ia32.h
index de9361d..1906644 100644 (file)
--- a/src/ia32/macro-assembler-ia32.h
+++ b/src/ia32/macro-assembler-ia32.h
@@ -437,7 +437,7 @@ class MacroAssembler: public Assembler {
 
   // Allocate a raw cons string object. Only the map field of the result is
   // initialized.
-  void AllocateConsString(Register result,
+  void AllocateTwoByteConsString(Register result,
                           Register scratch1,
                           Register scratch2,
                           Label* gc_required);
@@ -448,7 +448,7 @@ class MacroAssembler: public Assembler {
 
   // Allocate a raw sliced string object. Only the map field of the result is
   // initialized.
-  void AllocateSlicedString(Register result,
+  void AllocateTwoByteSlicedString(Register result,
                             Register scratch1,
                             Register scratch2,
                             Label* gc_required);

diff --git a/src/mips/code-stubs-mips.cc b/src/mips/code-stubs-mips.cc
index 7c2a146..7ca78f6 100644 (file)
--- a/src/mips/code-stubs-mips.cc
+++ b/src/mips/code-stubs-mips.cc
@@ -5079,7 +5079,8 @@ void StringCharCodeAtGenerator::GenerateFast(MacroAssembler* masm) {
 
   // Check for 1-byte or 2-byte string.
   __ bind(&flat_string);
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ And(t0, result_, Operand(kStringEncodingMask));
   __ Branch(&ascii_string, ne, t0, Operand(zero_reg));
 

diff --git a/src/x64/code-stubs-x64.cc b/src/x64/code-stubs-x64.cc
index 9237a0a..2ae4d9f 100644 (file)
--- a/src/x64/code-stubs-x64.cc
+++ b/src/x64/code-stubs-x64.cc
@@ -3938,7 +3938,8 @@ void StringCharCodeAtGenerator::GenerateFast(MacroAssembler* masm) {
 
   // Check for 1-byte or 2-byte string.
   __ bind(&flat_string);
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ testb(result_, Immediate(kStringEncodingMask));
   __ j(not_zero, &ascii_string);
 
@@ -4195,8 +4196,9 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   Label non_ascii, allocated, ascii_data;
   __ movl(rcx, r8);
   __ and_(rcx, r9);
-  STATIC_ASSERT(kStringEncodingMask == kAsciiStringTag);
-  __ testl(rcx, Immediate(kAsciiStringTag));
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+  __ testl(rcx, Immediate(kStringEncodingMask));
   __ j(zero, &non_ascii);
   __ bind(&ascii_data);
   // Allocate an acsii cons string.
@@ -4225,7 +4227,7 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   __ cmpb(r8, Immediate(kAsciiStringTag | kAsciiDataHintTag));
   __ j(equal, &ascii_data);
   // Allocate a two byte cons string.
-  __ AllocateConsString(rcx, rdi, no_reg, &string_add_runtime);
+  __ AllocateTwoByteConsString(rcx, rdi, no_reg, &string_add_runtime);
   __ jmp(&allocated);
 
   // Handle creating a flat result. First check that both strings are not
@@ -4254,10 +4256,11 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   // r8: instance type of first string
   // r9: instance type of second string
   Label non_ascii_string_add_flat_result;
-  STATIC_ASSERT(kStringEncodingMask == kAsciiStringTag);
-  __ testl(r8, Immediate(kAsciiStringTag));
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+  __ testl(r8, Immediate(kStringEncodingMask));
   __ j(zero, &non_ascii_string_add_flat_result);
-  __ testl(r9, Immediate(kAsciiStringTag));
+  __ testl(r9, Immediate(kStringEncodingMask));
   __ j(zero, &string_add_runtime);
 
   __ bind(&make_flat_ascii_string);
@@ -4295,7 +4298,9 @@ void StringAddStub::Generate(MacroAssembler* masm) {
   // r8: instance type of first string
   // r9: instance type of first string
   __ bind(&non_ascii_string_add_flat_result);
-  __ and_(r9, Immediate(kAsciiStringTag));
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+  __ and_(r9, Immediate(kStringEncodingMask));
   __ j(not_zero, &string_add_runtime);
   // Both strings are two byte strings. As they are short they are both
   // flat.
@@ -4639,9 +4644,6 @@ void StringHelper::GenerateHashGetHash(MacroAssembler* masm,
 void SubStringStub::Generate(MacroAssembler* masm) {
   Label runtime;
 
-  if (FLAG_string_slices) {
-    __ jmp(&runtime);
-  }
   // Stack frame on entry.
   //  rsp[0]: return address
   //  rsp[8]: to
@@ -4707,7 +4709,82 @@ void SubStringStub::Generate(MacroAssembler* masm) {
   __ movzxbl(rbx, FieldOperand(rbx, Map::kInstanceTypeOffset));
   __ Set(rcx, 2);
 
-  __ bind(&result_longer_than_two);
+  if (FLAG_string_slices) {
+    Label copy_routine;
+    // If coming from the make_two_character_string path, the string
+    // is too short to be sliced anyways.
+    STATIC_ASSERT(2 < SlicedString::kMinLength);
+    __ jmp(&copy_routine);
+    __ bind(&result_longer_than_two);
+
+    // rax: string
+    // rbx: instance type
+    // rcx: sub string length
+    // rdx: from index (smi)
+    Label allocate_slice, sliced_string, seq_string;
+    __ cmpq(rcx, Immediate(SlicedString::kMinLength));
+    // Short slice.  Copy instead of slicing.
+    __ j(less, &copy_routine);
+    STATIC_ASSERT(kSeqStringTag == 0);
+    __ testb(rbx, Immediate(kStringRepresentationMask));
+    __ j(zero, &seq_string, Label::kNear);
+    STATIC_ASSERT(kIsIndirectStringMask == (kSlicedStringTag & kConsStringTag));
+    STATIC_ASSERT(kIsIndirectStringMask != 0);
+    __ testb(rbx, Immediate(kIsIndirectStringMask));
+    // External string.  Jump to runtime.
+    __ j(zero, &runtime);
+
+    __ testb(rbx, Immediate(kSlicedNotConsMask));
+    __ j(not_zero, &sliced_string, Label::kNear);
+    // Cons string.  Check whether it is flat, then fetch first part.
+    __ CompareRoot(FieldOperand(rax, ConsString::kSecondOffset),
+                   Heap::kEmptyStringRootIndex);
+    __ j(not_equal, &runtime);
+    __ movq(rdi, FieldOperand(rax, ConsString::kFirstOffset));
+    __ jmp(&allocate_slice, Label::kNear);
+
+    __ bind(&sliced_string);
+    // Sliced string.  Fetch parent and correct start index by offset.
+    __ addq(rdx, FieldOperand(rax, SlicedString::kOffsetOffset));
+    __ movq(rdi, FieldOperand(rax, SlicedString::kParentOffset));
+    __ jmp(&allocate_slice, Label::kNear);
+
+    __ bind(&seq_string);
+    // Sequential string.  Just move string to the right register.
+    __ movq(rdi, rax);
+
+    __ bind(&allocate_slice);
+    // edi: underlying subject string
+    // ebx: instance type of original subject string
+    // edx: offset
+    // ecx: length
+    // Allocate new sliced string.  At this point we do not reload the instance
+    // type including the string encoding because we simply rely on the info
+    // provided by the original string.  It does not matter if the original
+    // string's encoding is wrong because we always have to recheck encoding of
+    // the newly created string's parent anyways due to externalized strings.
+    Label two_byte_slice, set_slice_header;
+    STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+    STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
+    __ testb(rbx, Immediate(kStringEncodingMask));
+    __ j(zero, &two_byte_slice, Label::kNear);
+    __ AllocateAsciiSlicedString(rax, rbx, no_reg, &runtime);
+    __ jmp(&set_slice_header, Label::kNear);
+    __ bind(&two_byte_slice);
+    __ AllocateTwoByteSlicedString(rax, rbx, no_reg, &runtime);
+    __ bind(&set_slice_header);
+    __ movq(FieldOperand(rax, SlicedString::kOffsetOffset), rdx);
+    __ Integer32ToSmi(rcx, rcx);
+    __ movq(FieldOperand(rax, SlicedString::kLengthOffset), rcx);
+    __ movq(FieldOperand(rax, SlicedString::kParentOffset), rdi);
+    __ movq(FieldOperand(rax, SlicedString::kHashFieldOffset),
+           Immediate(String::kEmptyHashField));
+    __ jmp(&return_rax);
+
+    __ bind(&copy_routine);
+  } else {
+    __ bind(&result_longer_than_two);
+  }
 
   // rax: string
   // rbx: instance type

diff --git a/src/x64/lithium-codegen-x64.cc b/src/x64/lithium-codegen-x64.cc
index c182413..25863ed 100644 (file)
--- a/src/x64/lithium-codegen-x64.cc
+++ b/src/x64/lithium-codegen-x64.cc
@@ -3251,7 +3251,8 @@ void LCodeGen::DoStringCharCodeAt(LStringCharCodeAt* instr) {
 
   // Dispatch on the encoding: ASCII or two-byte.
   Label ascii_string;
-  STATIC_ASSERT(kAsciiStringTag != 0);
+  STATIC_ASSERT((kStringEncodingMask & kAsciiStringTag) != 0);
+  STATIC_ASSERT((kStringEncodingMask & kTwoByteStringTag) == 0);
   __ testb(result, Immediate(kStringEncodingMask));
   __ j(not_zero, &ascii_string, Label::kNear);
 

diff --git a/src/x64/macro-assembler-x64.cc b/src/x64/macro-assembler-x64.cc
index b51d531..2ee506d 100644 (file)
--- a/src/x64/macro-assembler-x64.cc
+++ b/src/x64/macro-assembler-x64.cc
@@ -3623,7 +3623,7 @@ void MacroAssembler::AllocateAsciiString(Register result,
 }
 
 
-void MacroAssembler::AllocateConsString(Register result,
+void MacroAssembler::AllocateTwoByteConsString(Register result,
                                         Register scratch1,
                                         Register scratch2,
                                         Label* gc_required) {
@@ -3659,6 +3659,42 @@ void MacroAssembler::AllocateAsciiConsString(Register result,
 }
 
 
+void MacroAssembler::AllocateTwoByteSlicedString(Register result,
+                                          Register scratch1,
+                                          Register scratch2,
+                                          Label* gc_required) {
+  // Allocate heap number in new space.
+  AllocateInNewSpace(SlicedString::kSize,
+                     result,
+                     scratch1,
+                     scratch2,
+                     gc_required,
+                     TAG_OBJECT);
+
+  // Set the map. The other fields are left uninitialized.
+  LoadRoot(kScratchRegister, Heap::kSlicedStringMapRootIndex);
+  movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
+}
+
+
+void MacroAssembler::AllocateAsciiSlicedString(Register result,
+                                               Register scratch1,
+                                               Register scratch2,
+                                               Label* gc_required) {
+  // Allocate heap number in new space.
+  AllocateInNewSpace(SlicedString::kSize,
+                     result,
+                     scratch1,
+                     scratch2,
+                     gc_required,
+                     TAG_OBJECT);
+
+  // Set the map. The other fields are left uninitialized.
+  LoadRoot(kScratchRegister, Heap::kSlicedAsciiStringMapRootIndex);
+  movq(FieldOperand(result, HeapObject::kMapOffset), kScratchRegister);
+}
+
+
 // Copy memory, byte-by-byte, from source to destination.  Not optimized for
 // long or aligned copies.  The contents of scratch and length are destroyed.
 // Destination is incremented by length, source, length and scratch are

diff --git a/src/x64/macro-assembler-x64.h b/src/x64/macro-assembler-x64.h
index 47ce01b..e7eb104 100644 (file)
--- a/src/x64/macro-assembler-x64.h
+++ b/src/x64/macro-assembler-x64.h
@@ -921,7 +921,7 @@ class MacroAssembler: public Assembler {
 
   // Allocate a raw cons string object. Only the map field of the result is
   // initialized.
-  void AllocateConsString(Register result,
+  void AllocateTwoByteConsString(Register result,
                           Register scratch1,
                           Register scratch2,
                           Label* gc_required);
@@ -930,6 +930,17 @@ class MacroAssembler: public Assembler {
                                Register scratch2,
                                Label* gc_required);
 
+  // Allocate a raw sliced string object. Only the map field of the result is
+  // initialized.
+  void AllocateTwoByteSlicedString(Register result,
+                            Register scratch1,
+                            Register scratch2,
+                            Label* gc_required);
+  void AllocateAsciiSlicedString(Register result,
+                                 Register scratch1,
+                                 Register scratch2,
+                                 Label* gc_required);
+
   // ---------------------------------------------------------------------------
   // Support functions.
 

diff --git a/test/cctest/test-strings.cc b/test/cctest/test-strings.cc
index 17020a3..55c2141 100644 (file)
--- a/test/cctest/test-strings.cc
+++ b/test/cctest/test-strings.cc
@@ -529,3 +529,32 @@ TEST(TrivialSlice) {
   CHECK(string->IsSlicedString());
   CHECK_EQ("bcdefghijklmnopqrstuvwxy", *(string->ToCString()));
 }
+
+
+TEST(SliceFromSlice) {
+  // This tests whether a slice that contains the entire parent string
+  // actually creates a new string (it should not).
+  FLAG_string_slices = true;
+  InitializeVM();
+  HandleScope scope;
+  v8::Local<v8::Value> result;
+  Handle<String> string;
+  const char* init = "var str = 'abcdefghijklmnopqrstuvwxyz';";
+  const char* slice = "var slice = str.slice(1,-1); slice";
+  const char* slice_from_slice = "slice.slice(1,-1);";
+
+  CompileRun(init);
+  result = CompileRun(slice);
+  CHECK(result->IsString());
+  string = v8::Utils::OpenHandle(v8::String::Cast(*result));
+  CHECK(string->IsSlicedString());
+  CHECK(SlicedString::cast(*string)->parent()->IsSeqString());
+  CHECK_EQ("bcdefghijklmnopqrstuvwxy", *(string->ToCString()));
+
+  result = CompileRun(slice_from_slice);
+  CHECK(result->IsString());
+  string = v8::Utils::OpenHandle(v8::String::Cast(*result));
+  CHECK(string->IsSlicedString());
+  CHECK(SlicedString::cast(*string)->parent()->IsSeqString());
+  CHECK_EQ("cdefghijklmnopqrstuvwx", *(string->ToCString()));
+}