[wasm] Table.Grow should grow dispatch tables

[wasm] Table.Grow should grow dispatch tables
 - Table.Grow updates function, signature table sizes
 - Updates generated code with new base addresses for function, signature tables
 - Relocates size references for correct bounds check

R=bradnelson@chromium.org, titzer@chromium.org

Review-Url: https://codereview.chromium.org/2637643002
Cr-Commit-Position: refs/heads/master@{#42349}
Committed: https://chromium.googlesource.com/v8/v8/+/ccf0998d6327db9ef48aa8ca0531bf25b9cb13c8

[gdeepti, 2017-01-15 01:03:35]

The CQ bit was checked by gdeepti@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-01-15 01:03:50]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[gdeepti, 2017-01-15 01:28:29]

Refactoring to use WeakLinks for instances that share tables is still causing
failures on gc stress, abstracting this out for now.

[commit-bot: I haz the power, 2017-01-15 01:30:40]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-01-15 01:30:41]

Dry run: This issue passed the CQ dry run.

[bradnelson, 2017-01-15 03:00:14]

lgtm

https://codereview.chromium.org/2637643002/diff/1/src/wasm/wasm-js.cc
File src/wasm/wasm-js.cc (right):

https://codereview.chromium.org/2637643002/diff/1/src/wasm/wasm-js.cc#newcode563
src/wasm/wasm-js.cc:563: // TODO(gdeepti): use weak links for instances
Just to make sure I understand correctly. By this you mean you'd like to switch
this to instance chains like memory? Or do instances bound to a table get held
alive until the table is also garbage?

https://codereview.chromium.org/2637643002/diff/1/test/mjsunit/wasm/indirect-...
File test/mjsunit/wasm/indirect-tables.js (right):

https://codereview.chromium.org/2637643002/diff/1/test/mjsunit/wasm/indirect-...
test/mjsunit/wasm/indirect-tables.js:402: let table = new
WebAssembly.Table({element: "anyfunc", initial: 10, maximum: 30});
>80

[gdeepti, 2017-01-15 05:57:02]

The CQ bit was checked by gdeepti@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-01-15 05:57:12]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[gdeepti, 2017-01-15 05:57:54]

https://codereview.chromium.org/2637643002/diff/1/src/wasm/wasm-js.cc
File src/wasm/wasm-js.cc (right):

https://codereview.chromium.org/2637643002/diff/1/src/wasm/wasm-js.cc#newcode563
src/wasm/wasm-js.cc:563: // TODO(gdeepti): use weak links for instances
On 2017/01/15 03:00:14, bradnelson wrote:
> Just to make sure I understand correctly. By this you mean you'd like to
switch
> this to instance chains like memory? Or do instances bound to a table get held
> alive until the table is also garbage?

Ideally, I would like to switch this to instance chains - the refactoring CL I
was working on does exactly that, but there's been some test failures on that I
haven't been able to fix yet. Currently the instances bound to a table get held
till the garbage is collected. The other approach is to refactor the dispatch
tables to be their own class and have state about the instance and vice versa,
and on garbage collection have a new array of dispatch tables created without
the dispatch table for that particular instance - Though it's a little more
wasteful than the first approach, I'm contemplating trying the second approach
to see if that gives me the same stability issues.

https://codereview.chromium.org/2637643002/diff/1/test/mjsunit/wasm/indirect-...
File test/mjsunit/wasm/indirect-tables.js (right):

https://codereview.chromium.org/2637643002/diff/1/test/mjsunit/wasm/indirect-...
test/mjsunit/wasm/indirect-tables.js:402: let table = new
WebAssembly.Table({element: "anyfunc", initial: 10, maximum: 30});
On 2017/01/15 03:00:14, bradnelson wrote:
> >80

Done.

[commit-bot: I haz the power, 2017-01-15 06:29:55]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-01-15 06:29:55]

Dry run: This issue passed the CQ dry run.

[gdeepti, 2017-01-15 07:44:26]

The CQ bit was checked by gdeepti@chromium.org

[gdeepti, 2017-01-15 07:44:26]

The patchset sent to the CQ was uploaded after l-g-t-m from
bradnelson@chromium.org
Link to the patchset: https://codereview.chromium.org/2637643002/#ps20001
(title: "Rebase + review")

[commit-bot: I haz the power, 2017-01-15 07:44:32]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-15 07:46:14]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1484466266132130,
"parent_rev": "6fad1ad27d84ec7cce275d5c2c8bffd48ab8a654", "commit_rev":
"ccf0998d6327db9ef48aa8ca0531bf25b9cb13c8"}

[commit-bot: I haz the power, 2017-01-15 07:46:19]

Description was changed from

==========
[wasm] Table.Grow should grow dispatch tables
 - Table.Grow updates function, signature table sizes
 - Updates generated code with new base addresses for function, signature tables
 - Relocates size references for correct bounds check

R=bradnelson@chromium.org, titzer@chromium.org
==========

to

==========
[wasm] Table.Grow should grow dispatch tables
 - Table.Grow updates function, signature table sizes
 - Updates generated code with new base addresses for function, signature tables
 - Relocates size references for correct bounds check

R=bradnelson@chromium.org, titzer@chromium.org

Review-Url: https://codereview.chromium.org/2637643002
Cr-Commit-Position: refs/heads/master@{#42349}
Committed:
https://chromium.googlesource.com/v8/v8/+/ccf0998d6327db9ef48aa8ca0531bf25b9c...
==========

[commit-bot: I haz the power, 2017-01-15 07:46:20]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/v8/v8/+/ccf0998d6327db9ef48aa8ca0531bf25b9c...