[turbofan] Better and more sane support for tail calls

[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics (currently only in TF).
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n

Committed: https://crrev.com/ff283f7ded84df5a5a77d088e37fd69cf6271a8c
Cr-Commit-Position: refs/heads/master@{#31987}

[danno, 2015-11-12 09:22:42]

Description was changed from

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function, %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics.
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gab resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

to

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function, %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics.
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

[danno, 2015-11-12 09:27:27]

PTAL

[danno, 2015-11-12 09:28:27]

Description was changed from

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function, %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics.
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

to

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics.
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

[Jarin, 2015-11-12 13:45:11]

jarin@chromium.org changed reviewers:
+ jarin@chromium.org

[Jarin, 2015-11-12 13:45:12]

Several drive-by comments.

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
File src/compiler/instruction-selector.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
src/compiler/instruction-selector.cc:491:
buffer->descriptor->GetInputLocation(index), stack_param_delta);
Replace 'buffer->descriptor->GetInputLocation(index)' with 'location' here.

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
src/compiler/instruction-selector.cc:1176: InitializeCallBuffer(node, &buffer,
true, true, false);
Hmm, the bool parameters are getting a bit out of hand, perhaps it is time to
introduce enums or flags?

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/linkage.cc
File src/compiler/linkage.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/linkage.cc...
src/compiler/linkage.cc:96: // number and content.
Update the comment, please.

https://codereview.chromium.org/1439613003/diff/80001/src/runtime/runtime-fun...
File src/runtime/runtime-function.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/runtime/runtime-fun...
src/runtime/runtime-function.cc:532: UNREACHABLE();
Why do not we just do the same thing as Runtime_Call here? Ultimately, we would
want to fuzz this.

https://codereview.chromium.org/1439613003/diff/80001/test/unittests/compiler...
File test/unittests/compiler/linkage-tail-call-unittest.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/test/unittests/compiler...
test/unittests/compiler/linkage-tail-call-unittest.cc:175:
EXPECT_TRUE(desc1->CanTailCall(node, &stack_param_delta));
How about checking the delta here and below?

[danno, 2015-11-13 10:04:17]

Feedback addressed. Please take another look.

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
File src/compiler/instruction-selector.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
src/compiler/instruction-selector.cc:491:
buffer->descriptor->GetInputLocation(index), stack_param_delta);
On 2015/11/12 13:45:12, Jarin wrote:
> Replace 'buffer->descriptor->GetInputLocation(index)' with 'location' here.

Done.

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/instructio...
src/compiler/instruction-selector.cc:1176: InitializeCallBuffer(node, &buffer,
true, true, false);
On 2015/11/12 13:45:12, Jarin wrote:
> Hmm, the bool parameters are getting a bit out of hand, perhaps it is time to
> introduce enums or flags?

Done.

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/linkage.cc
File src/compiler/linkage.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/compiler/linkage.cc...
src/compiler/linkage.cc:96: // number and content.
On 2015/11/12 13:45:12, Jarin wrote:
> Update the comment, please.

Done.

https://codereview.chromium.org/1439613003/diff/80001/src/runtime/runtime-fun...
File src/runtime/runtime-function.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/src/runtime/runtime-fun...
src/runtime/runtime-function.cc:532: UNREACHABLE();
On 2015/11/12 13:45:12, Jarin wrote:
> Why do not we just do the same thing as Runtime_Call here? Ultimately, we
would
> want to fuzz this.

Done.

https://codereview.chromium.org/1439613003/diff/80001/test/unittests/compiler...
File test/unittests/compiler/linkage-tail-call-unittest.cc (right):

https://codereview.chromium.org/1439613003/diff/80001/test/unittests/compiler...
test/unittests/compiler/linkage-tail-call-unittest.cc:175:
EXPECT_TRUE(desc1->CanTailCall(node, &stack_param_delta));
On 2015/11/12 13:45:12, Jarin wrote:
> How about checking the delta here and below?

Done.

[Jarin, 2015-11-13 10:42:49]

Nit.

https://codereview.chromium.org/1439613003/diff/100001/src/compiler/instructi...
File src/compiler/instruction-selector.h (right):

https://codereview.chromium.org/1439613003/diff/100001/src/compiler/instructi...
src/compiler/instruction-selector.h:171: };
Please use "typedef base::Flags<CallBufferFlags> CallBufferInitFlags;" and then
refer to CallBufferInitFlags. That way you can avoid the ugly casts.

Also, the convention for enums is hungarian/camel (kCallCodeImmediate).

[danno, 2015-11-13 11:54:11]

PTAL

https://codereview.chromium.org/1439613003/diff/100001/src/compiler/instructi...
File src/compiler/instruction-selector.h (right):

https://codereview.chromium.org/1439613003/diff/100001/src/compiler/instructi...
src/compiler/instruction-selector.h:171: };
On 2015/11/13 10:42:49, Jarin wrote:
> Please use "typedef base::Flags<CallBufferFlags> CallBufferInitFlags;" and
then
> refer to CallBufferInitFlags. That way you can avoid the ugly casts.
> 
> Also, the convention for enums is hungarian/camel (kCallCodeImmediate).
> 

Done.

[Jarin, 2015-11-13 13:29:39]

LGTM. Thanks!

[danno, 2015-11-13 14:29:02]

The CQ bit was checked by danno@chromium.org

[commit-bot: I haz the power, 2015-11-13 14:29:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1439613003/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1439613003/120001

[danno, 2015-11-13 14:29:40]

Description was changed from

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics.
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

to

==========
[turbofan] Better and more sane support for tail calls

* Limit triggering of tail calls to explicit use of a new inline runtime
  function %_TailCall. %_TailCall works just like %_Call except for using
  tail-calling mechanics (currently only in TF).
* Remove hack that recognized some specific usages of %_Call and converted them
  into tail calls.
* Support tail calls for all calls where the number of callee stack parameters
  is less than or equal to the number of caller stack parameters.
* Use the gap resolver to swizzle parameters and registers to tail calls.

BUG=v8:4076
LOG=n
==========

[commit-bot: I haz the power, 2015-11-13 14:30:00]

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

[commit-bot: I haz the power, 2015-11-13 14:30:01]

Try jobs failed on following builders:
  v8_android_arm_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_android_arm_compile_rel/...)
  v8_linux64_asan_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_asan_rel/builds/...)
  v8_linux_arm_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel/builds/10111)
  v8_linux_dbg on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_dbg/builds/10729)
  v8_linux_gcc_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)
  v8_linux_mips64el_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_mips64el_compile_r...)
  v8_linux_mipsel_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_mipsel_compile_rel...)
  v8_mac_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_rel/builds/11826)
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/7710)

[danno, 2015-11-13 14:33:56]

The CQ bit was checked by danno@chromium.org

[danno, 2015-11-13 14:33:56]

The patchset sent to the CQ was uploaded after l-g-t-m from jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/1439613003/#ps140001
(title: "Rebase")

[commit-bot: I haz the power, 2015-11-13 14:34:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1439613003/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1439613003/140001

[commit-bot: I haz the power, 2015-11-13 14:49:13]

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

[commit-bot: I haz the power, 2015-11-13 14:49:14]

Try jobs failed on following builders:
  v8_linux_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel/builds/11704)

[danno, 2015-11-13 15:49:05]

The CQ bit was checked by danno@chromium.org

[danno, 2015-11-13 15:49:05]

The patchset sent to the CQ was uploaded after l-g-t-m from jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/1439613003/#ps160001
(title: "Skip test in ignition configuration")

[commit-bot: I haz the power, 2015-11-13 15:49:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1439613003/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1439613003/160001

[commit-bot: I haz the power, 2015-11-13 16:08:16]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2015-11-14 23:21:40]

Patchset 9 (id:??) landed as
https://crrev.com/ff283f7ded84df5a5a77d088e37fd69cf6271a8c
Cr-Commit-Position: refs/heads/master@{#31987}