MIPS port initial commit

MIPS port initial commit

This is the first step in the MIPS port of V8. It adds assembler, disassembler and simulator for the MIPS32 architecture.

Contains stubbed out implementation of all the compiler/code generator infrastructure to make it all build.

Patch by Alexandre Rames from Sigma Designs Inc.

This is the landing of http://codereview.chromium.org/543161.

Committed: http://code.google.com/p/v8/source/detail?r=3799

[Søren Thygesen Gjesse, 2010-02-04 12:19:01]

Mads, are you ready to rubber stamp this, or should we have another person
reviewing it?

It adds about 10000 lines of code in the src/mips directory. The changes to the
platform independent files are minimal, mainly only handling of include files.

Running

  tools/test.py --simulator=mips

builds without error and runs all the enabled tests including the new ones in
test-assembler-mips.cc. No JavaScript tests are currently enabled.

[Mads Ager (chromium), 2010-02-04 15:05:06]

LGTM

This is mostly a rubberstamp.  I did have a quick look through it and have added
a number of nits.  Overall, it looks very nice.

http://codereview.chromium.org/561072/diff/1/40
File src/mips/assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/40#newcode177
src/mips/assembler-mips.cc:177: // Implementation of RelocInfo
Should do a comment period sweep on this file as well.

http://codereview.chromium.org/561072/diff/1/40#newcode310
src/mips/assembler-mips.cc:310: return    opcode == BEQ
Identation should be fixed for this return statement.

http://codereview.chromium.org/561072/diff/1/40#newcode336
src/mips/assembler-mips.cc:336: int32_t imm18 = (((int32_t)instr &
(int32_t)kImm16Mask) << 16) >> 14;
No C-style casts please.

http://codereview.chromium.org/561072/diff/1/40#newcode464
src/mips/assembler-mips.cc:464: Instr instr = opcode | (rs.code() << kRsShift) |
(rt.code() << kRtShift)
This should be indented according to Google style.  Same goes for the next
methods.

http://codereview.chromium.org/561072/diff/1/40#newcode842
src/mips/assembler-mips.cc:842: Instr instr = SPECIAL | TLTU | rs.code() <<
kRsShift |
Move the '|' to the next line for consistency with the rest of the code.

http://codereview.chromium.org/561072/diff/1/40#newcode997
src/mips/assembler-mips.cc:997: Instr instr = COP1 | fmt | ft.code() << 16 |
fs.code() << kFsShift |
Move the '|' to the next line for consistency with the rest of the code.

http://codereview.chromium.org/561072/diff/1/40#newcode1171
src/mips/assembler-mips.cc:1171: CHECK(((instr1 & kOpcodeMask) == LUI && (instr2
& kOpcodeMask) == ORI) ||
Let's remove the extra white-spacing in the formatting in the rest of this file?

http://codereview.chromium.org/561072/diff/1/47
File src/mips/assembler-mips.h (right):

http://codereview.chromium.org/561072/diff/1/47#newcode79
src/mips/assembler-mips.h:79: bool is_byte_register() const  { return false; }
If there is no such distinction, we should just remove this method and the
comment.  This method is not there in the ARM nor the x64 version either.

http://codereview.chromium.org/561072/diff/1/47#newcode89
src/mips/assembler-mips.h:89: // (unfortunately we can't make this private in a
struct)
This is just copied from the other assemblers but could we change this comment
in all the assemblers to remove the parenthesis and follow the convention of
capitalizing and ending with a period?

http://codereview.chromium.org/561072/diff/1/47#newcode132
src/mips/assembler-mips.h:132: // Coprocessor register
End with a period.

http://codereview.chromium.org/561072/diff/1/47#newcode145
src/mips/assembler-mips.h:145: // (unfortunately we can't make this private in a
struct)
Comment style.

http://codereview.chromium.org/561072/diff/1/47#newcode227
src/mips/assembler-mips.h:227: // Class Operand represents a shifter operand in
data processing instructions
End comment with period.

http://codereview.chromium.org/561072/diff/1/47#newcode295
src/mips/assembler-mips.h:295: // Label operations & relative jumps (PPUM
Appendix D)
End with period.

http://codereview.chromium.org/561072/diff/1/47#newcode353
src/mips/assembler-mips.h:353: // target and the return address
End with period.

http://codereview.chromium.org/561072/diff/1/47#newcode368
src/mips/assembler-mips.h:368: // We don't use likely variant of instructions
End with period.

http://codereview.chromium.org/561072/diff/1/47#newcode401
src/mips/assembler-mips.h:401: // Arithmetic
We should run through this file quickly and add periods at the end of comments.

http://codereview.chromium.org/561072/diff/1/47#newcode578
src/mips/assembler-mips.h:578: // The relocation writer's position is at least
kGap bytes Uless the end of
Uless -> below?

http://codereview.chromium.org/561072/diff/1/27
File src/mips/builtins-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/27#newcode1
src/mips/builtins-mips.cc:1: // Copyright 2006-2010 the V8 project authors. All
rights reserved.
All of the new files should just say "2010".

http://codereview.chromium.org/561072/diff/1/44
File src/mips/constants-mips.h (right):

http://codereview.chromium.org/561072/diff/1/44#newcode45
src/mips/constants-mips.h:45: // Try
www.cs.cornell.edu/courses/cs3410/2008fa/MIPS_Vol2.pdf  .
Remove spaces before period at the end.

http://codereview.chromium.org/561072/diff/1/44#newcode46
src/mips/constants-mips.h:46: // Else e quick Google search and you'll find it.
Remove, unneeded comment (people will do that anyway).  :)

http://codereview.chromium.org/561072/diff/1/44#newcode54
src/mips/constants-mips.h:54: // Number of general purpose registers
Periods at end of comments.

http://codereview.chromium.org/561072/diff/1/26
File src/mips/cpu-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/26#newcode54
src/mips/cpu-mips.cc:54: if (res)
Add braces around the body.

http://codereview.chromium.org/561072/diff/1/48
File src/mips/disasm-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/48#newcode777
src/mips/disasm-mips.cc:777: prev_pc, *reinterpret_cast<int32_t*>(prev_pc),
buffer.start());
Align with the other arguments to fprintf.

http://codereview.chromium.org/561072/diff/1/45
File src/mips/frames-mips.h (right):

http://codereview.chromium.org/561072/diff/1/45#newcode124
src/mips/frames-mips.h:124: // C/C++ argument slots size
Periods at end of comments.

http://codereview.chromium.org/561072/diff/1/31
File src/mips/jump-target-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/31#newcode29
src/mips/jump-target-mips.cc:29: // FROM ARM Code
Remove comment.

http://codereview.chromium.org/561072/diff/1/37
File src/mips/macro-assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/37#newcode374
src/mips/macro-assembler-mips.cc:374: addiu(sp, sp, -4*NumToPush);
Space around binary operator.

http://codereview.chromium.org/561072/diff/1/37#newcode375
src/mips/macro-assembler-mips.cc:375: for (int16_t i = 0; i< kNumRegisters; i++)
{
Space between 'i' and '<'.

http://codereview.chromium.org/561072/diff/1/37#newcode378
src/mips/macro-assembler-mips.cc:378: MemOperand(sp, 4*(NumToPush -
++NumSaved)));
Seems this will fit on the line above.

Space around binary operator.

http://codereview.chromium.org/561072/diff/1/37#newcode384
src/mips/macro-assembler-mips.cc:384: void
MacroAssembler::MultiPushReversed(RegList regs) {
Same comments as above.

http://codereview.chromium.org/561072/diff/1/37#newcode389
src/mips/macro-assembler-mips.cc:389: for (int16_t i = kNumRegisters; --i >= 0;)
{
for (int16_t i = kNumRegisters; i > 0; i--) {

http://codereview.chromium.org/561072/diff/1/37#newcode401
src/mips/macro-assembler-mips.cc:401: for (int16_t i = kNumRegisters; --i >= 0;)
{
for (int16_t i = kNumRegisters; i > 0; i--) {

http://codereview.chromium.org/561072/diff/1/41
File src/mips/macro-assembler-mips.h (right):

http://codereview.chromium.org/561072/diff/1/41#newcode61
src/mips/macro-assembler-mips.h:61: private:
Strange mix of public and private here.  I would prefer to have all the public
stuff at the beginning and move all the private stuff to the end.  We should do
that for the other macro assemblers too if they have the same structure.

http://codereview.chromium.org/561072/diff/1/41#newcode354
src/mips/macro-assembler-mips.h:354: Handle<Object> code_object_;  // This
handle will be patched with the code
Move comment to the line before the field.

http://codereview.chromium.org/561072/diff/1/51
File src/mips/register-allocator-mips-inl.h (right):

http://codereview.chromium.org/561072/diff/1/51#newcode1
src/mips/register-allocator-mips-inl.h:1: // Copyright 2006-2008 the V8 project
authors. All rights reserved.
2010

http://codereview.chromium.org/561072/diff/1/50
File src/mips/simulator-mips.h (right):

http://codereview.chromium.org/561072/diff/1/50#newcode241
src/mips/simulator-mips.h:241: "Eror:Unexpected %i opcode in a branch delay
slot.",
Indentation.

http://codereview.chromium.org/561072/diff/1/50#newcode256
src/mips/simulator-mips.h:256: // Exceptions
Period.

http://codereview.chromium.org/561072/diff/1/5
File test/cctest/test-assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/5#newcode34
test/cctest/test-assembler-mips.cc:34: // For macro-assembler support
Period at end of comment.  What is included for macro-assembler support?

http://codereview.chromium.org/561072/diff/1/5#newcode54
test/cctest/test-assembler-mips.cc:54: // The test framework does not accept
flags on the command line, so we set them
Periods at the end of comments.

[Søren Thygesen Gjesse, 2010-02-04 20:25:05]

http://codereview.chromium.org/561072/diff/1/40
File src/mips/assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/40#newcode177
src/mips/assembler-mips.cc:177: // Implementation of RelocInfo
On 2010/02/04 15:05:06, Mads Ager wrote:
> Should do a comment period sweep on this file as well.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode310
src/mips/assembler-mips.cc:310: return    opcode == BEQ
On 2010/02/04 15:05:06, Mads Ager wrote:
> Identation should be fixed for this return statement.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode336
src/mips/assembler-mips.cc:336: int32_t imm18 = (((int32_t)instr &
(int32_t)kImm16Mask) << 16) >> 14;
On 2010/02/04 15:05:06, Mads Ager wrote:
> No C-style casts please.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode464
src/mips/assembler-mips.cc:464: Instr instr = opcode | (rs.code() << kRsShift) |
(rt.code() << kRtShift)
On 2010/02/04 15:05:06, Mads Ager wrote:
> This should be indented according to Google style.  Same goes for the next
> methods.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode842
src/mips/assembler-mips.cc:842: Instr instr = SPECIAL | TLTU | rs.code() <<
kRsShift |
On 2010/02/04 15:05:06, Mads Ager wrote:
> Move the '|' to the next line for consistency with the rest of the code.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode997
src/mips/assembler-mips.cc:997: Instr instr = COP1 | fmt | ft.code() << 16 |
fs.code() << kFsShift |
On 2010/02/04 15:05:06, Mads Ager wrote:
> Move the '|' to the next line for consistency with the rest of the code.

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode1171
src/mips/assembler-mips.cc:1171: CHECK(((instr1 & kOpcodeMask) == LUI && (instr2
& kOpcodeMask) == ORI) ||
On 2010/02/04 15:05:06, Mads Ager wrote:
> Let's remove the extra white-spacing in the formatting in the rest of this
file?

Done.

http://codereview.chromium.org/561072/diff/1/40#newcode1205
src/mips/assembler-mips.cc:1205: 
Removed #undef of macros no longer defined.

http://codereview.chromium.org/561072/diff/1/47
File src/mips/assembler-mips.h (right):

http://codereview.chromium.org/561072/diff/1/47#newcode79
src/mips/assembler-mips.h:79: bool is_byte_register() const  { return false; }
On 2010/02/04 15:05:06, Mads Ager wrote:
> If there is no such distinction, we should just remove this method and the
> comment.  This method is not there in the ARM nor the x64 version either.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode89
src/mips/assembler-mips.h:89: // (unfortunately we can't make this private in a
struct)
On 2010/02/04 15:05:06, Mads Ager wrote:
> This is just copied from the other assemblers but could we change this comment
> in all the assemblers to remove the parenthesis and follow the convention of
> capitalizing and ending with a period?

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode132
src/mips/assembler-mips.h:132: // Coprocessor register
On 2010/02/04 15:05:06, Mads Ager wrote:
> End with a period.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode145
src/mips/assembler-mips.h:145: // (unfortunately we can't make this private in a
struct)
On 2010/02/04 15:05:06, Mads Ager wrote:
> Comment style.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode227
src/mips/assembler-mips.h:227: // Class Operand represents a shifter operand in
data processing instructions
On 2010/02/04 15:05:06, Mads Ager wrote:
> End comment with period.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode295
src/mips/assembler-mips.h:295: // Label operations & relative jumps (PPUM
Appendix D)
On 2010/02/04 15:05:06, Mads Ager wrote:
> End with period.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode353
src/mips/assembler-mips.h:353: // target and the return address
On 2010/02/04 15:05:06, Mads Ager wrote:
> End with period.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode368
src/mips/assembler-mips.h:368: // We don't use likely variant of instructions
On 2010/02/04 15:05:06, Mads Ager wrote:
> End with period.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode401
src/mips/assembler-mips.h:401: // Arithmetic
On 2010/02/04 15:05:06, Mads Ager wrote:
> We should run through this file quickly and add periods at the end of
comments.

Done.

http://codereview.chromium.org/561072/diff/1/47#newcode578
src/mips/assembler-mips.h:578: // The relocation writer's position is at least
kGap bytes Uless the end of
On 2010/02/04 15:05:06, Mads Ager wrote:
> Uless -> below?

Done.

http://codereview.chromium.org/561072/diff/1/27
File src/mips/builtins-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/27#newcode1
src/mips/builtins-mips.cc:1: // Copyright 2006-2010 the V8 project authors. All
rights reserved.
On 2010/02/04 15:05:06, Mads Ager wrote:
> All of the new files should just say "2010".

Done.

http://codereview.chromium.org/561072/diff/1/44
File src/mips/constants-mips.h (right):

http://codereview.chromium.org/561072/diff/1/44#newcode45
src/mips/constants-mips.h:45: // Try
www.cs.cornell.edu/courses/cs3410/2008fa/MIPS_Vol2.pdf  .
On 2010/02/04 15:05:06, Mads Ager wrote:
> Remove spaces before period at the end.

Done.

http://codereview.chromium.org/561072/diff/1/44#newcode46
src/mips/constants-mips.h:46: // Else e quick Google search and you'll find it.
On 2010/02/04 15:05:06, Mads Ager wrote:
> Remove, unneeded comment (people will do that anyway).  :)

Done.

http://codereview.chromium.org/561072/diff/1/44#newcode54
src/mips/constants-mips.h:54: // Number of general purpose registers
On 2010/02/04 15:05:06, Mads Ager wrote:
> Periods at end of comments.

Done.

http://codereview.chromium.org/561072/diff/1/26
File src/mips/cpu-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/26#newcode54
src/mips/cpu-mips.cc:54: if (res)
On 2010/02/04 15:05:06, Mads Ager wrote:
> Add braces around the body.

Done.

http://codereview.chromium.org/561072/diff/1/48
File src/mips/disasm-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/48#newcode777
src/mips/disasm-mips.cc:777: prev_pc, *reinterpret_cast<int32_t*>(prev_pc),
buffer.start());
On 2010/02/04 15:05:06, Mads Ager wrote:
> Align with the other arguments to fprintf.

Done.

http://codereview.chromium.org/561072/diff/1/45
File src/mips/frames-mips.h (right):

http://codereview.chromium.org/561072/diff/1/45#newcode124
src/mips/frames-mips.h:124: // C/C++ argument slots size
On 2010/02/04 15:05:06, Mads Ager wrote:
> Periods at end of comments.

Done.

http://codereview.chromium.org/561072/diff/1/31
File src/mips/jump-target-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/31#newcode29
src/mips/jump-target-mips.cc:29: // FROM ARM Code
On 2010/02/04 15:05:06, Mads Ager wrote:
> Remove comment.

Done.

http://codereview.chromium.org/561072/diff/1/37
File src/mips/macro-assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/37#newcode374
src/mips/macro-assembler-mips.cc:374: addiu(sp, sp, -4*NumToPush);
On 2010/02/04 15:05:06, Mads Ager wrote:
> Space around binary operator.

Done.

http://codereview.chromium.org/561072/diff/1/37#newcode375
src/mips/macro-assembler-mips.cc:375: for (int16_t i = 0; i< kNumRegisters; i++)
{
On 2010/02/04 15:05:06, Mads Ager wrote:
> Space between 'i' and '<'.

Done.

http://codereview.chromium.org/561072/diff/1/37#newcode378
src/mips/macro-assembler-mips.cc:378: MemOperand(sp, 4*(NumToPush -
++NumSaved)));
On 2010/02/04 15:05:06, Mads Ager wrote:
> Seems this will fit on the line above.
> 
> Space around binary operator.
> 

Done.

http://codereview.chromium.org/561072/diff/1/37#newcode384
src/mips/macro-assembler-mips.cc:384: void
MacroAssembler::MultiPushReversed(RegList regs) {
On 2010/02/04 15:05:06, Mads Ager wrote:
> Same comments as above.

Done.

http://codereview.chromium.org/561072/diff/1/37#newcode389
src/mips/macro-assembler-mips.cc:389: for (int16_t i = kNumRegisters; --i >= 0;)
{
On 2010/02/04 15:05:06, Mads Ager wrote:
> for (int16_t i = kNumRegisters; i > 0; i--) {

Done.

http://codereview.chromium.org/561072/diff/1/37#newcode401
src/mips/macro-assembler-mips.cc:401: for (int16_t i = kNumRegisters; --i >= 0;)
{
On 2010/02/04 15:05:06, Mads Ager wrote:
> for (int16_t i = kNumRegisters; i > 0; i--) {

Done.

http://codereview.chromium.org/561072/diff/1/41
File src/mips/macro-assembler-mips.h (right):

http://codereview.chromium.org/561072/diff/1/41#newcode61
src/mips/macro-assembler-mips.h:61: private:
On 2010/02/04 15:05:06, Mads Ager wrote:
> Strange mix of public and private here.  I would prefer to have all the public
> stuff at the beginning and move all the private stuff to the end.  We should
do
> that for the other macro assemblers too if they have the same structure.

Done.

http://codereview.chromium.org/561072/diff/1/41#newcode354
src/mips/macro-assembler-mips.h:354: Handle<Object> code_object_;  // This
handle will be patched with the code
On 2010/02/04 15:05:06, Mads Ager wrote:
> Move comment to the line before the field.

Done.

http://codereview.chromium.org/561072/diff/1/51
File src/mips/register-allocator-mips-inl.h (right):

http://codereview.chromium.org/561072/diff/1/51#newcode1
src/mips/register-allocator-mips-inl.h:1: // Copyright 2006-2008 the V8 project
authors. All rights reserved.
On 2010/02/04 15:05:06, Mads Ager wrote:
> 2010

Done.

http://codereview.chromium.org/561072/diff/1/50
File src/mips/simulator-mips.h (right):

http://codereview.chromium.org/561072/diff/1/50#newcode241
src/mips/simulator-mips.h:241: "Eror:Unexpected %i opcode in a branch delay
slot.",
On 2010/02/04 15:05:06, Mads Ager wrote:
> Indentation.

Done.

http://codereview.chromium.org/561072/diff/1/50#newcode256
src/mips/simulator-mips.h:256: // Exceptions
On 2010/02/04 15:05:06, Mads Ager wrote:
> Period.

Done.

http://codereview.chromium.org/561072/diff/1/5
File test/cctest/test-assembler-mips.cc (right):

http://codereview.chromium.org/561072/diff/1/5#newcode34
test/cctest/test-assembler-mips.cc:34: // For macro-assembler support
On 2010/02/04 15:05:06, Mads Ager wrote:
> Period at end of comment.  What is included for macro-assembler support?

Simplified include file list.

http://codereview.chromium.org/561072/diff/1/5#newcode54
test/cctest/test-assembler-mips.cc:54: // The test framework does not accept
flags on the command line, so we set them
On 2010/02/04 15:05:06, Mads Ager wrote:
> Periods at the end of comments.

Done.