Landing for Zaheer Ahmad.

Landing for Zaheer Ahmad.

Direct call api functions (arm implementation)

See: http://codereview.chromium.org/6170001/

Committed: http://code.google.com/p/v8/source/detail?r=6639

[antonm, 2011-02-03 11:52:53]

Last sanity check, if you please

[Zaheer, 2011-02-04 06:32:42]

On 2011/02/03 11:52:53, antonm wrote:
> Last sanity check, if you please
ARM simulator
./tools/test.py --mode=debug --simulator=arm
[144:05|% 100|+ 4200|-   0]: Done                                              
Browsing/File cycler works fine.

[Erik Corry, 2011-02-04 09:53:57]

On 2011/02/03 11:52:53, antonm wrote:
> Last sanity check, if you please

Sorry to have more nits at this late stage

[Erik Corry, 2011-02-04 09:55:55]

http://codereview.chromium.org/6286078/diff/1/AUTHORS
File AUTHORS (right):

http://codereview.chromium.org/6286078/diff/1/AUTHORS#newcode38
AUTHORS:38: Mike Gilbert <floppymaster@gmail.com>
Perhaps you could move Mike who is not in alphabetical order here.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5743: 
Missing blank line.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5747: 
and here

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5755: __ mov(r2,
Operand(reinterpret_cast<intptr_t>(ref.address())));
There is a constructor for Operand that takes an external reference.  I think
you have to use that otherwise this doesn't work correctly with snapshots.  The
snapshot code needs the reloc information from a real Operand in case the
destination is different in mksnapshot and the final executable.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5756: __ add(ip, pc, Operand(4));
It seems to me that omitting this add and just using ip in the next instruction
would achieve the same.  Pc is always 8 bytes ahead of the actual instruction
executing so that should be the same.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5760: 
and missing blank line here

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.h
File src/arm/code-stubs-arm.h (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.h#newcod...
src/arm/code-stubs-arm.h:572: 
Missing blank line.

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc
File src/arm/macro-assembler-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1830: 
Blank line

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1840: 
Blank line

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1857: 
Blank line

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1868: 
Blank line

http://codereview.chromium.org/6286078/diff/1/src/heap.cc
File src/heap.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/heap.cc#newcode1944
src/heap.cc:1944: 
Blank lines.

http://codereview.chromium.org/6286078/diff/1/src/heap.cc#newcode1951
src/heap.cc:1951: 
Blank lines.

[antonm, 2011-02-04 12:59:45]

All comments except for one are addressed,

http://codereview.chromium.org/6286078/diff/1/AUTHORS
File AUTHORS (right):

http://codereview.chromium.org/6286078/diff/1/AUTHORS#newcode38
AUTHORS:38: Mike Gilbert <floppymaster@gmail.com>
On 2011/02/04 09:55:55, Erik Corry wrote:
> Perhaps you could move Mike who is not in alphabetical order here.

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5743: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Missing blank line.

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5747: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> and here

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5755: __ mov(r2,
Operand(reinterpret_cast<intptr_t>(ref.address())));
On 2011/02/04 09:55:55, Erik Corry wrote:
> There is a constructor for Operand that takes an external reference.  I think
> you have to use that otherwise this doesn't work correctly with snapshots. 
The
> snapshot code needs the reloc information from a real Operand in case the
> destination is different in mksnapshot and the final executable.

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5756: __ add(ip, pc, Operand(4));
On 2011/02/04 09:55:55, Erik Corry wrote:
> It seems to me that omitting this add and just using ip in the next
instruction
> would achieve the same.  Pc is always 8 bytes ahead of the actual instruction
> executing so that should be the same.

Hmm, ip would have any garbage then.  Am I missing something.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5760: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> and missing blank line here

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.h
File src/arm/code-stubs-arm.h (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.h#newcod...
src/arm/code-stubs-arm.h:572: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Missing blank line.

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc
File src/arm/macro-assembler-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1830: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank line

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1840: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank line

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1857: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank line

Done.

http://codereview.chromium.org/6286078/diff/1/src/arm/macro-assembler-arm.cc#...
src/arm/macro-assembler-arm.cc:1868: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank line

Done.

http://codereview.chromium.org/6286078/diff/1/src/heap.cc
File src/heap.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/heap.cc#newcode1944
src/heap.cc:1944: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank lines.

Done.

http://codereview.chromium.org/6286078/diff/1/src/heap.cc#newcode1951
src/heap.cc:1951: 
On 2011/02/04 09:55:55, Erik Corry wrote:
> Blank lines.

Done.

[Zaheer, 2011-02-04 13:08:05]

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5756: __ add(ip, pc, Operand(4));
On 2011/02/04 12:59:45, antonm wrote:
> On 2011/02/04 09:55:55, Erik Corry wrote:
> > It seems to me that omitting this add and just using ip in the next
> instruction
> > would achieve the same.  Pc is always 8 bytes ahead of the actual
instruction
> > executing so that should be the same.
> 
> Hmm, ip would have any garbage then.  Am I missing something.

I think Erik meant pc.

[Erik Corry, 2011-02-04 13:19:52]

lgtm

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/6286078/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5756: __ add(ip, pc, Operand(4));
On 2011/02/04 12:59:45, antonm wrote:
> On 2011/02/04 09:55:55, Erik Corry wrote:
> > It seems to me that omitting this add and just using ip in the next
> instruction
> > would achieve the same.  Pc is always 8 bytes ahead of the actual
instruction
> > executing so that should be the same.
> 
> Hmm, ip would have any garbage then.  Am I missing something.

No you are right of course.