[Interpreter] Make dispatch table point to code entry instead of code objects.

[Interpreter] Make dispatch table point to code entry instead of code objects.

Modifies Ignition to store code entry addresses in the dispatch table
rather than code objects. This allows the interpreter to avoid
calculating the code entry address from the code object on every
dispatch and provides a ~5-7% performance improvement on Octane with
Ignition.

This change adds ArchOpcode::kArchTailCallAddress to TurboFan to enable
tail call dispatch using these code addresses. It also adds a Dispatch
linkage creator (distinct from the stub linkage type used previously) to
allow targetting a code address target (which will diverge further from
the stub linkage type when we remove the context machine register in
Ignition).

BUG=v8:4280
LOG=N

Committed: https://crrev.com/0c05e02f25912c0ce510dac3077be3f72728aa34
Cr-Commit-Position: refs/heads/master@{#35480}

[rmcilroy, 2016-04-12 13:14:42]

rmcilroy@chromium.org changed reviewers:
+ bmeurer@chromium.org

[rmcilroy, 2016-04-12 13:14:42]

Benedkit, this is the CL I discussed over chat. PTAL, if you think it is
reasonable I'll port to other architectures.

[Benedikt Meurer, 2016-04-12 13:20:12]

I think the approach is OK, but I don't like the naming. The "Dispatch" is
confusing IMHO. Can we find a better name? Is "Address" better?

https://codereview.chromium.org/1882073002/diff/1/src/compiler/code-stub-asse...
File src/compiler/code-stub-assembler.h (right):

https://codereview.chromium.org/1882073002/diff/1/src/compiler/code-stub-asse...
src/compiler/code-stub-assembler.h:270: Node* TailCallDispatch(const
CallInterfaceDescriptor& descriptor,
How about TailCallAddress instead?

https://codereview.chromium.org/1882073002/diff/1/src/compiler/linkage.h
File src/compiler/linkage.h (right):

https://codereview.chromium.org/1882073002/diff/1/src/compiler/linkage.h#newc...
src/compiler/linkage.h:311: // Call[Dispatch]   code_entry, arg 1, arg 2, arg 3,
[...],            context
Nit: s/Dispatch/Address ?

[rmcilroy, 2016-04-12 15:25:28]

Patchset #2 (id:20001) has been deleted

[rmcilroy, 2016-04-12 15:25:58]

Done and added ports, PTAL.

https://codereview.chromium.org/1882073002/diff/1/src/compiler/code-stub-asse...
File src/compiler/code-stub-assembler.h (right):

https://codereview.chromium.org/1882073002/diff/1/src/compiler/code-stub-asse...
src/compiler/code-stub-assembler.h:270: Node* TailCallDispatch(const
CallInterfaceDescriptor& descriptor,
On 2016/04/12 13:20:12, Benedikt Meurer wrote:
> How about TailCallAddress instead?

As discussed offline, changed to TailCallBytecodeDispatch.

https://codereview.chromium.org/1882073002/diff/1/src/compiler/linkage.h
File src/compiler/linkage.h (right):

https://codereview.chromium.org/1882073002/diff/1/src/compiler/linkage.h#newc...
src/compiler/linkage.h:311: // Call[Dispatch]   code_entry, arg 1, arg 2, arg 3,
[...],            context
On 2016/04/12 13:20:12, Benedikt Meurer wrote:
> Nit: s/Dispatch/Address ?

Also BytecodeDispatch.

[rmcilroy, 2016-04-12 15:34:38]

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

[commit-bot: I haz the power, 2016-04-12 15:34:45]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/40001

[commit-bot: I haz the power, 2016-04-12 15:40:36]

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

[commit-bot: I haz the power, 2016-04-12 15:40:39]

Dry run: Try jobs failed on following builders:
  v8_linux_gcc_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)

[rmcilroy, 2016-04-12 15:46:38]

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

[commit-bot: I haz the power, 2016-04-12 15:46:45]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/40001

[commit-bot: I haz the power, 2016-04-12 15:51:59]

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

[commit-bot: I haz the power, 2016-04-12 15:51:59]

Dry run: Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/13512)

[rmcilroy, 2016-04-12 16:30:32]

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

[commit-bot: I haz the power, 2016-04-12 16:30:39]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/60001

[commit-bot: I haz the power, 2016-04-12 16:52:39]

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

[commit-bot: I haz the power, 2016-04-12 16:52:40]

Dry run: Try jobs failed on following builders:
  v8_linux_dbg_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng/builds/4155)
  v8_linux_dbg_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng_triggered/b...)

[rmcilroy, 2016-04-12 17:16:06]

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

[commit-bot: I haz the power, 2016-04-12 17:16:11]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/60001

[commit-bot: I haz the power, 2016-04-12 17:34:42]

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

[commit-bot: I haz the power, 2016-04-12 17:34:43]

Dry run: Try jobs failed on following builders:
  v8_linux_dbg_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng/builds/4158)
  v8_linux_dbg_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg_ng_triggered/b...)

[rmcilroy, 2016-04-13 12:12:46]

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

[commit-bot: I haz the power, 2016-04-13 12:12:49]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/80001

[commit-bot: I haz the power, 2016-04-13 12:35:41]

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

[commit-bot: I haz the power, 2016-04-13 12:35:43]

Dry run: This issue passed the CQ dry run.

[rmcilroy, 2016-04-14 09:39:27]

Ping?

[Benedikt Meurer, 2016-04-14 09:54:05]

LGTM

[rmcilroy, 2016-04-14 10:01:35]

The CQ bit was checked by rmcilroy@chromium.org

[commit-bot: I haz the power, 2016-04-14 10:01:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1882073002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1882073002/80001

[commit-bot: I haz the power, 2016-04-14 10:03:55]

Committed patchset #4 (id:80001)

[commit-bot: I haz the power, 2016-04-14 10:04:30]

Description was changed from

==========
[Interpreter] Make dispatch table point to code entry instead of code objects.

Modifies Ignition to store code entry addresses in the dispatch table
rather than code objects. This allows the interpreter to avoid
calculating the code entry address from the code object on every
dispatch and provides a ~5-7% performance improvement on Octane with
Ignition.

This change adds ArchOpcode::kArchTailCallAddress to TurboFan to enable
tail call dispatch using these code addresses. It also adds a Dispatch
linkage creator (distinct from the stub linkage type used previously) to
allow targetting a code address target (which will diverge further from
the stub linkage type when we remove the context machine register in
Ignition).

BUG=v8:4280
LOG=N
==========

to

==========
[Interpreter] Make dispatch table point to code entry instead of code objects.

Modifies Ignition to store code entry addresses in the dispatch table
rather than code objects. This allows the interpreter to avoid
calculating the code entry address from the code object on every
dispatch and provides a ~5-7% performance improvement on Octane with
Ignition.

This change adds ArchOpcode::kArchTailCallAddress to TurboFan to enable
tail call dispatch using these code addresses. It also adds a Dispatch
linkage creator (distinct from the stub linkage type used previously) to
allow targetting a code address target (which will diverge further from
the stub linkage type when we remove the context machine register in
Ignition).

BUG=v8:4280
LOG=N

Committed: https://crrev.com/0c05e02f25912c0ce510dac3077be3f72728aa34
Cr-Commit-Position: refs/heads/master@{#35480}
==========

[commit-bot: I haz the power, 2016-04-14 10:04:31]

Patchset 4 (id:??) landed as
https://crrev.com/0c05e02f25912c0ce510dac3077be3f72728aa34
Cr-Commit-Position: refs/heads/master@{#35480}