Revision: 3709
Author: [email protected]
Date: Tue Jan 26 03:08:42 2010
Log: Don't pass the "at start" parameter to native RegExp
As the start index is already passed it is easy to calculate the "at start"
boolean in generated code. Also as direct entry has been implemented this
needs to be done in generated code anyway, and therefore might as well be
moved to the generated code for RegExp. The "at start" value is now
calcualted as a local variable on the native RegExp frame based on the
value of the start index argument.
The x64 version have been tested on both Linux and 64-bit Windows Vista.
For ARM I have tested cctest/test-regexp on ARM hardware, but the rest of
the tests have only been run on the ARM simulator.
Review URL: http://codereview.chromium.org/554078
http://code.google.com/p/v8/source/detail?r=3709
Modified:
/branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.cc
/branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.h
/branches/bleeding_edge/src/arm/simulator-arm.h
/branches/bleeding_edge/src/ia32/codegen-ia32.cc
/branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.cc
/branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.h
/branches/bleeding_edge/src/ia32/simulator-ia32.h
/branches/bleeding_edge/src/regexp-macro-assembler.cc
/branches/bleeding_edge/src/regexp-macro-assembler.h
/branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.cc
/branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.h
/branches/bleeding_edge/src/x64/simulator-x64.h
/branches/bleeding_edge/test/cctest/test-regexp.cc
=======================================
--- /branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.cc Mon Jan
18 01:49:50 2010
+++ /branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.cc Tue Jan
26 03:08:42 2010
@@ -63,8 +63,6 @@
* through the runtime system)
* - stack_area_base (High end of the memory area to use as
* backtracking stack)
- * - at_start (if 1, we are starting at the start of the
- * string, otherwise 0)
* - int* capture_array (int[num_saved_registers_], for output).
* --- sp when called ---
* - link address
@@ -76,6 +74,8 @@
* - void* input_string (location of a handle containing the string)
* - Offset of location before start of input (effectively character
* position -1). Used to initialize capture registers to a
non-position.
+ * - At start (if 1, we are starting at the start of the
+ * string, otherwise 0)
* - register 0 (Only positions must be stored in the first
* - register 1 num_saved_registers_ registers)
* - ...
@@ -610,6 +610,7 @@
// Set frame pointer just above the arguments.
__ add(frame_pointer(), sp, Operand(4 * kPointerSize));
__ push(r0); // Make room for "position - 1" constant (value is
irrelevant).
+ __ push(r0); // Make room for "at start" constant (value is irrelevant).
// Check if we have space on the stack for registers.
Label stack_limit_hit;
@@ -653,6 +654,15 @@
// Store this value in a local variable, for use when clearing
// position registers.
__ str(r0, MemOperand(frame_pointer(), kInputStartMinusOne));
+
+ // Determine whether the start index is zero, that is at the start of the
+ // string, and store that value in a local variable.
+ __ ldr(r1, MemOperand(frame_pointer(), kStartIndex));
+ __ tst(r1, Operand(r1));
+ __ mov(r1, Operand(1), LeaveCC, eq);
+ __ mov(r1, Operand(0), LeaveCC, ne);
+ __ str(r1, MemOperand(frame_pointer(), kAtStart));
+
if (num_saved_registers_ > 0) { // Always is, if generated from a
regexp.
// Fill saved registers with initial value = start offset - 1
=======================================
--- /branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.h Thu Jan 7
11:01:23 2010
+++ /branches/bleeding_edge/src/arm/regexp-macro-assembler-arm.h Tue Jan 26
03:08:42 2010
@@ -123,8 +123,7 @@
static const int kReturnAddress = kStoredRegisters + 8 * kPointerSize;
// Stack parameters placed by caller.
static const int kRegisterOutput = kReturnAddress + kPointerSize;
- static const int kAtStart = kRegisterOutput + kPointerSize;
- static const int kStackHighEnd = kAtStart + kPointerSize;
+ static const int kStackHighEnd = kRegisterOutput + kPointerSize;
static const int kDirectCall = kStackHighEnd + kPointerSize;
// Below the frame pointer.
@@ -136,8 +135,9 @@
// When adding local variables remember to push space for them in
// the frame in GetCode.
static const int kInputStartMinusOne = kInputString - kPointerSize;
+ static const int kAtStart = kInputStartMinusOne - kPointerSize;
// First register address. Following registers are below it on the stack.
- static const int kRegisterZero = kInputStartMinusOne - kPointerSize;
+ static const int kRegisterZero = kAtStart - kPointerSize;
// Initial size of code buffer.
static const size_t kRegExpCodeSize = 1024;
=======================================
--- /branches/bleeding_edge/src/arm/simulator-arm.h Wed Jan 6 03:09:30 2010
+++ /branches/bleeding_edge/src/arm/simulator-arm.h Tue Jan 26 03:08:42 2010
@@ -63,8 +63,8 @@
// Call the generated regexp code directly. The entry function pointer
should
// expect eight int/pointer sized arguments and return an int.
-#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6, p7) \
- entry(p0, p1, p2, p3, p4, p5, p6, p7)
+#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6) \
+ entry(p0, p1, p2, p3, p4, p5, p6)
#define TRY_CATCH_FROM_ADDRESS(try_catch_address) \
reinterpret_cast<TryCatch*>(try_catch_address)
@@ -79,9 +79,9 @@
assembler::arm::Simulator::current()->Call(FUNCTION_ADDR(entry), 5, \
p0, p1, p2, p3, p4))
-#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6, p7) \
+#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6) \
assembler::arm::Simulator::current()->Call( \
- FUNCTION_ADDR(entry), 8, p0, p1, p2, p3, p4, p5, p6, p7)
+ FUNCTION_ADDR(entry), 7, p0, p1, p2, p3, p4, p5, p6)
#define TRY_CATCH_FROM_ADDRESS(try_catch_address) \
try_catch_address == NULL ? \
=======================================
--- /branches/bleeding_edge/src/ia32/codegen-ia32.cc Tue Jan 26 02:27:27
2010
+++ /branches/bleeding_edge/src/ia32/codegen-ia32.cc Tue Jan 26 03:08:42
2010
@@ -8453,20 +8453,14 @@
// All checks done. Now push arguments for native regexp code.
__ IncrementCounter(&Counters::regexp_entry_native, 1);
- // Argument 8: Indicate that this is a direct call from JavaScript.
+ // Argument 7: Indicate that this is a direct call from JavaScript.
__ push(Immediate(1));
- // Argument 7: Start (high end) of backtracking stack memory area.
+ // Argument 6: Start (high end) of backtracking stack memory area.
__ mov(ecx,
Operand::StaticVariable(address_of_regexp_stack_memory_address));
__ add(ecx,
Operand::StaticVariable(address_of_regexp_stack_memory_size));
__ push(ecx);
- // Argument 6: At start of string?
- __ xor_(Operand(ecx), ecx); // setcc only operated on cl (lower byte of
ecx).
- __ test(ebx, Operand(ebx));
- __ setcc(zero, ecx); // 1 if 0 (start of string), 0 if positive.
- __ push(ecx);
-
// Argument 5: static offsets vector buffer.
__
push(Immediate(ExternalReference::address_of_static_offsets_vector()));
@@ -8501,7 +8495,7 @@
__ add(Operand(edx), Immediate(Code::kHeaderSize - kHeapObjectTag));
__ call(Operand(edx));
// Remove arguments.
- __ add(Operand(esp), Immediate(8 * kPointerSize));
+ __ add(Operand(esp), Immediate(7 * kPointerSize));
// Check the result.
Label success;
=======================================
--- /branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.cc Mon Jan
18 01:49:50 2010
+++ /branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.cc Tue Jan
26 03:08:42 2010
@@ -59,8 +59,6 @@
* call through the runtime system)
* - stack_area_base (High end of the memory area to use as
* backtracking stack)
- * - at_start (if 1, we are starting at the start of the
- * string, otherwise 0)
* - int* capture_array (int[num_saved_registers_], for output).
* - end of input (Address of end of string)
* - start of input (Address of first character in string)
@@ -74,6 +72,8 @@
* - backup of caller ebx
* - Offset of location before start of input (effectively character
* position -1). Used to initialize capture registers to a
non-position.
+ * - Boolean at start (if 1, we are starting at the start of the
string,
+ * otherwise 0)
* - register 0 ebp[-4] (Only positions must be stored in the first
* - register 1 ebp[-8] num_saved_registers_ registers)
* - ...
@@ -625,6 +625,7 @@
__ push(edi);
__ push(ebx); // Callee-save on MacOS.
__ push(Immediate(0)); // Make room for "input start - 1" constant.
+ __ push(Immediate(0)); // Make room for "at start" constant.
// Check if we have space on the stack for registers.
Label stack_limit_hit;
@@ -667,6 +668,15 @@
// Store this value in a local variable, for use when clearing
// position registers.
__ mov(Operand(ebp, kInputStartMinusOne), eax);
+
+ // Determine whether the start index is zero, that is at the start of the
+ // string, and store that value in a local variable.
+ __ mov(ebx, Operand(ebp, kStartIndex));
+ __ xor_(Operand(ecx), ecx); // setcc only operates on cl (lower byte of
ecx).
+ __ test(ebx, Operand(ebx));
+ __ setcc(zero, ecx); // 1 if 0 (start of string), 0 if positive.
+ __ mov(Operand(ebp, kAtStart), ecx);
+
if (num_saved_registers_ > 0) { // Always is, if generated from a
regexp.
// Fill saved registers with initial value = start offset - 1
// Fill in stack push order, to avoid accessing across an unwritten
=======================================
--- /branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.h Thu Jan
7 11:01:23 2010
+++ /branches/bleeding_edge/src/ia32/regexp-macro-assembler-ia32.h Tue Jan
26 03:08:42 2010
@@ -123,8 +123,7 @@
static const int kInputStart = kStartIndex + kPointerSize;
static const int kInputEnd = kInputStart + kPointerSize;
static const int kRegisterOutput = kInputEnd + kPointerSize;
- static const int kAtStart = kRegisterOutput + kPointerSize;
- static const int kStackHighEnd = kAtStart + kPointerSize;
+ static const int kStackHighEnd = kRegisterOutput + kPointerSize;
static const int kDirectCall = kStackHighEnd + kPointerSize;
// Below the frame pointer - local stack variables.
// When adding local variables remember to push space for them in
@@ -133,8 +132,9 @@
static const int kBackup_edi = kBackup_esi - kPointerSize;
static const int kBackup_ebx = kBackup_edi - kPointerSize;
static const int kInputStartMinusOne = kBackup_ebx - kPointerSize;
+ static const int kAtStart = kInputStartMinusOne - kPointerSize;
// First register address. Following registers are below it on the stack.
- static const int kRegisterZero = kInputStartMinusOne - kPointerSize;
+ static const int kRegisterZero = kAtStart - kPointerSize;
// Initial size of code buffer.
static const size_t kRegExpCodeSize = 1024;
=======================================
--- /branches/bleeding_edge/src/ia32/simulator-ia32.h Wed Jan 6 03:09:30
2010
+++ /branches/bleeding_edge/src/ia32/simulator-ia32.h Tue Jan 26 03:08:42
2010
@@ -53,8 +53,8 @@
// Call the generated regexp code directly. The entry function pointer
should
// expect eight int/pointer sized arguments and return an int.
-#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6, p7) \
- entry(p0, p1, p2, p3, p4, p5, p6, p7)
+#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6) \
+ entry(p0, p1, p2, p3, p4, p5, p6)
#define TRY_CATCH_FROM_ADDRESS(try_catch_address) \
reinterpret_cast<TryCatch*>(try_catch_address)
=======================================
--- /branches/bleeding_edge/src/regexp-macro-assembler.cc Fri Jan 22
00:31:33 2010
+++ /branches/bleeding_edge/src/regexp-macro-assembler.cc Tue Jan 26
03:08:42 2010
@@ -144,8 +144,7 @@
start_offset,
input_start,
input_end,
- offsets_vector,
- previous_index == 0);
+ offsets_vector);
return res;
}
@@ -156,14 +155,11 @@
int start_offset,
const byte* input_start,
const byte* input_end,
- int* output,
- bool at_start) {
+ int* output) {
typedef int (*matcher)(String*, int, const byte*,
- const byte*, int*, int, Address, int);
+ const byte*, int*, Address, int);
matcher matcher_func = FUNCTION_CAST<matcher>(code->entry());
- int at_start_val = at_start ? 1 : 0;
-
// Ensure that the minimum stack has been allocated.
RegExpStack stack;
Address stack_base = RegExpStack::stack_base();
@@ -175,7 +171,6 @@
input_start,
input_end,
output,
- at_start_val,
stack_base,
direct_call);
ASSERT(result <= SUCCESS);
=======================================
--- /branches/bleeding_edge/src/regexp-macro-assembler.h Mon Jan 18
01:49:50 2010
+++ /branches/bleeding_edge/src/regexp-macro-assembler.h Tue Jan 26
03:08:42 2010
@@ -218,8 +218,7 @@
int start_offset,
const byte* input_start,
const byte* input_end,
- int* output,
- bool at_start);
+ int* output);
};
#endif // V8_NATIVE_REGEXP
=======================================
--- /branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.cc Mon Jan
18 01:49:50 2010
+++ /branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.cc Tue Jan
26 03:08:42 2010
@@ -71,8 +71,6 @@
* through the runtime system)
* - stack_area_base (High end of the memory area to use as
* backtracking stack)
- * - at_start (if 1, we are starting at the start of the
- * string, otherwise 0)
* - int* capture_array (int[num_saved_registers_], for output).
* - end of input (Address of end of string)
* - start of input (Address of first character in string)
@@ -82,6 +80,8 @@
* - backup of callee save registers (rbx, possibly rsi and rdi).
* - Offset of location before start of input (effectively character
* position -1). Used to initialize capture registers to a
non-position.
+ * - At start of string (if 1, we are starting at the start of the
+ * string, otherwise 0)
* - register 0 rbp[-n] (Only positions must be stored in the first
* - register 1 rbp[-n-8] num_saved_registers_ registers)
* - ...
@@ -661,7 +661,7 @@
ASSERT_EQ(kInputStart, -3 * kPointerSize);
ASSERT_EQ(kInputEnd, -4 * kPointerSize);
ASSERT_EQ(kRegisterOutput, -5 * kPointerSize);
- ASSERT_EQ(kAtStart, -6 * kPointerSize);
+ ASSERT_EQ(kStackHighEnd, -6 * kPointerSize);
__ push(rdi);
__ push(rsi);
__ push(rdx);
@@ -672,6 +672,7 @@
__ push(rbx); // Callee-save
#endif
__ push(Immediate(0)); // Make room for "input start - 1" constant.
+ __ push(Immediate(0)); // Make room for "at start" constant.
// Check if we have space on the stack for registers.
Label stack_limit_hit;
@@ -716,6 +717,15 @@
// Store this value in a local variable, for use when clearing
// position registers.
__ movq(Operand(rbp, kInputStartMinusOne), rax);
+
+ // Determine whether the start index is zero, that is at the start of the
+ // string, and store that value in a local variable.
+ __ movq(rbx, Operand(rbp, kStartIndex));
+ __ xor_(rcx, rcx); // setcc only operates on cl (lower byte of rcx).
+ __ testq(rbx, rbx);
+ __ setcc(zero, rcx); // 1 if 0 (start of string), 0 if positive.
+ __ movq(Operand(rbp, kAtStart), rcx);
+
if (num_saved_registers_ > 0) {
// Fill saved registers with initial value = start offset - 1
// Fill in stack push order, to avoid accessing across an unwritten
=======================================
--- /branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.h Thu Jan 7
11:01:23 2010
+++ /branches/bleeding_edge/src/x64/regexp-macro-assembler-x64.h Tue Jan 26
03:08:42 2010
@@ -138,9 +138,7 @@
static const int kInputStart = kStartIndex + kPointerSize;
static const int kInputEnd = kInputStart + kPointerSize;
static const int kRegisterOutput = kInputEnd + kPointerSize;
- // AtStart is passed as 32 bit int (values 0 or 1).
- static const int kAtStart = kRegisterOutput + kPointerSize;
- static const int kStackHighEnd = kAtStart + kPointerSize;
+ static const int kStackHighEnd = kRegisterOutput + kPointerSize;
// DirectCall is passed as 32 bit int (values 0 or 1).
static const int kDirectCall = kStackHighEnd + kPointerSize;
#else
@@ -152,9 +150,8 @@
static const int kInputStart = kStartIndex - kPointerSize;
static const int kInputEnd = kInputStart - kPointerSize;
static const int kRegisterOutput = kInputEnd - kPointerSize;
- static const int kAtStart = kRegisterOutput - kPointerSize;
- static const int kStackHighEnd = kFrameAlign;
- static const int kDirectCall = kStackHighEnd + kPointerSize;
+ static const int kStackHighEnd = kRegisterOutput - kPointerSize;
+ static const int kDirectCall = kFrameAlign;
#endif
#ifdef _WIN64
@@ -168,7 +165,7 @@
// AMD64 Calling Convention has only one callee-save register that
// we use. We push this after the frame pointer (and after the
// parameters).
- static const int kBackup_rbx = kAtStart - kPointerSize;
+ static const int kBackup_rbx = kStackHighEnd - kPointerSize;
static const int kLastCalleeSaveRegister = kBackup_rbx;
#endif
@@ -176,9 +173,10 @@
// the frame in GetCode.
static const int kInputStartMinusOne =
kLastCalleeSaveRegister - kPointerSize;
+ static const int kAtStart = kInputStartMinusOne - kPointerSize;
// First register address. Following registers are below it on the stack.
- static const int kRegisterZero = kInputStartMinusOne - kPointerSize;
+ static const int kRegisterZero = kAtStart - kPointerSize;
// Initial size of code buffer.
static const size_t kRegExpCodeSize = 1024;
=======================================
--- /branches/bleeding_edge/src/x64/simulator-x64.h Wed Jan 6 03:09:30 2010
+++ /branches/bleeding_edge/src/x64/simulator-x64.h Tue Jan 26 03:08:42 2010
@@ -54,8 +54,8 @@
// Call the generated regexp code directly. The entry function pointer
should
// expect eight int/pointer sized arguments and return an int.
-#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6, p7) \
- entry(p0, p1, p2, p3, p4, p5, p6, p7)
+#define CALL_GENERATED_REGEXP_CODE(entry, p0, p1, p2, p3, p4, p5, p6) \
+ entry(p0, p1, p2, p3, p4, p5, p6)
#define TRY_CATCH_FROM_ADDRESS(try_catch_address) \
reinterpret_cast<TryCatch*>(try_catch_address)
=======================================
--- /branches/bleeding_edge/test/cctest/test-regexp.cc Mon Jan 25 04:56:49
2010
+++ /branches/bleeding_edge/test/cctest/test-regexp.cc Tue Jan 26 03:08:42
2010
@@ -679,16 +679,14 @@
int start_offset,
const byte* input_start,
const byte* input_end,
- int* captures,
- bool at_start) {
+ int* captures) {
return NativeRegExpMacroAssembler::Execute(
code,
input,
start_offset,
input_start,
input_end,
- captures,
- at_start);
+ captures);
}
@@ -716,8 +714,7 @@
0,
start_adr,
start_adr + seq_input->length(),
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(-1, captures[0]);
@@ -760,8 +757,7 @@
0,
start_adr,
start_adr + input->length(),
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, captures[0]);
@@ -778,8 +774,7 @@
0,
start_adr,
start_adr + input->length(),
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::FAILURE, result);
}
@@ -820,8 +815,7 @@
0,
start_adr,
start_adr + input->length(),
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, captures[0]);
@@ -839,8 +833,7 @@
0,
start_adr,
start_adr + input->length() * 2,
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::FAILURE, result);
}
@@ -877,8 +870,7 @@
0,
start_adr,
start_adr + input->length(),
- NULL,
- true);
+ NULL);
CHECK_EQ(NativeRegExpMacroAssembler::FAILURE, result);
}
@@ -920,8 +912,7 @@
0,
start_adr,
start_adr + input->length(),
- output,
- true);
+ output);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, output[0]);
@@ -969,8 +960,7 @@
0,
start_adr,
start_adr + input->length() * 2,
- output,
- true);
+ output);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, output[0]);
@@ -1022,8 +1012,7 @@
0,
start_adr,
start_adr + input->length(),
- NULL,
- true);
+ NULL);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
@@ -1032,8 +1021,7 @@
3,
start_adr + 3,
start_adr + input->length(),
- NULL,
- false);
+ NULL);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
}
@@ -1084,8 +1072,7 @@
0,
start_adr,
start_adr + input->length(),
- output,
- true);
+ output);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, output[0]);
@@ -1184,8 +1171,7 @@
0,
start_adr,
start_adr + input->length(),
- output,
- true);
+ output);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, output[0]);
@@ -1225,8 +1211,7 @@
0,
start_adr,
start_adr + input->length(),
- NULL,
- true);
+ NULL);
CHECK_EQ(NativeRegExpMacroAssembler::EXCEPTION, result);
CHECK(Top::has_pending_exception());
@@ -1271,8 +1256,7 @@
0,
start_adr,
start_adr + input->length(),
- captures,
- true);
+ captures);
CHECK_EQ(NativeRegExpMacroAssembler::SUCCESS, result);
CHECK_EQ(0, captures[0]);
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
