Reimplement the padding of relocation information for lazy deoptimization on ia32.

Reimplement the padding of relocation information for lazy deoptimization on ia32.

The previous implementation attempted to keep track of the needed
relocation size for deoptimization while generating the optimized
code. That was error prone. This patch moves the relocation resizing
to the deoptimizer as the last step of creating an optimized code
object.

The down side to this approach is that two relocation information byte
arrays are created for all optimized functions that do not have enough
relocation space for lazy deoptimization.

R=sgjesse@chromium.org

Committed: http://code.google.com/p/v8/source/detail?r=7360

[Søren Thygesen Gjesse, 2011-03-25 09:44:43]

LGTM

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc
File src/ia32/deoptimizer-ia32.cc (right):

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc#ne...
src/ia32/deoptimizer-ia32.cc:96: // Padding needed.
How about just calculating padding here (avoid the min_padding varaible).

// Find the padding needed rounded up to a whole number of comment relocs.
int padding = (min_reloc_size - reloc_length + comment_reloc_size - 1) /
comment_reloc_size * comment_reloc_size
ASSERT_EQ(0, padding % comment_reloc_size);
int additional_comments = padding / comment_reloc_size;

Maybe also loose comment_reloc_size and just use
RelocInfo::kMinRelocCommentSize.

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc#ne...
src/ia32/deoptimizer-ia32.cc:120: reloc_info_writer.Write(&rinfo);
ASSERT that this always write RelocInfo::kMinRelocCommentSize bytes?

[Mads Ager (chromium), 2011-03-25 10:15:03]

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc
File src/ia32/deoptimizer-ia32.cc (right):

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc#ne...
src/ia32/deoptimizer-ia32.cc:96: // Padding needed.
On 2011/03/25 09:44:43, Søren Gjesse wrote:
> How about just calculating padding here (avoid the min_padding varaible).
> 
> // Find the padding needed rounded up to a whole number of comment relocs.
> int padding = (min_reloc_size - reloc_length + comment_reloc_size - 1) /
> comment_reloc_size * comment_reloc_size
> ASSERT_EQ(0, padding % comment_reloc_size);
> int additional_comments = padding / comment_reloc_size;
> 
> Maybe also loose comment_reloc_size and just use
> RelocInfo::kMinRelocCommentSize.

I had that initially. Decided on some more variables to make it more readable.

http://codereview.chromium.org/6730050/diff/1/src/ia32/deoptimizer-ia32.cc#ne...
src/ia32/deoptimizer-ia32.cc:120: reloc_info_writer.Write(&rinfo);
On 2011/03/25 09:44:43, Søren Gjesse wrote:
> ASSERT that this always write RelocInfo::kMinRelocCommentSize bytes?

Done.