[turbofan] Add the StackSlot operator to turbofan.

[turbofan] Add the StackSlot operator to turbofan.

The StackSlot operator allows to allocate a spill slot on the stack. We
are going to use this operator to pass floats through pointers to c
functions, which we need for floating point rounding in the case where
the architecture does not provide rounding instructions.

R=titzer@chromium.org, v8-arm-ports@googlegroups.com, v8-ppc-ports@googlegroups.com, v8-mips-ports@googlegroups.com

Committed: https://crrev.com/7a693437787090d62d937b862e29521debcc5223
Cr-Commit-Position: refs/heads/master@{#33600}

Committed: https://crrev.com/645880373b13a96b812e3e4038c8e4044b2b34ca
Cr-Commit-Position: refs/heads/master@{#33606}

[jbramley, 2016-01-27 17:28:58]

Out of curiosity, why do we need to pass floats by pointer?

[titzer, 2016-01-27 17:31:26]

On 2016/01/27 17:28:58, jbramley wrote:
> Out of curiosity, why do we need to pass floats by pointer?

It's easier than dealing with the x87 FP stack :-/

[jbramley, 2016-01-27 17:32:54]

On 2016/01/27 17:31:26, titzer wrote:
> On 2016/01/27 17:28:58, jbramley wrote:
> > Out of curiosity, why do we need to pass floats by pointer?
> 
> It's easier than dealing with the x87 FP stack :-/

Is it possible to do that just for x87?

[ahaas, 2016-01-28 08:35:21]

On 2016/01/27 at 17:32:54, jacob.bramley wrote:
> On 2016/01/27 17:31:26, titzer wrote:
> > On 2016/01/27 17:28:58, jbramley wrote:
> > > Out of curiosity, why do we need to pass floats by pointer?
> > 
> > It's easier than dealing with the x87 FP stack :-/
> 
> Is it possible to do that just for x87?

Passing floats to C functions does not work on any architecture at the moment,
not just x87.

[titzer, 2016-01-28 09:02:19]

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/instructio...
File src/compiler/instruction-selector.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/instructio...
src/compiler/instruction-selector.cc:1140: MachineRepresentation rep =
StackSlotRepresentationOf(node->op());
You should probably use ElemSizeLog2Of(MachineRepresentation) here.

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/pipeline.cc
File src/compiler/pipeline.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/pipeline.c...
src/compiler/pipeline.cc:1393: void Pipeline::InitializeFrame(CallDescriptor*
descriptor) {
You can just inline this method into the callers.

https://codereview.chromium.org/1645653002/diff/20001/test/cctest/compiler/te...
File test/cctest/compiler/test-run-machops.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/test/cctest/compiler/te...
test/cctest/compiler/test-run-machops.cc:75: TEST(RunStackSlot) {
Can you write another test that tries the same for all machine types?

[ahaas, 2016-01-28 14:55:44]

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/instructio...
File src/compiler/instruction-selector.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/instructio...
src/compiler/instruction-selector.cc:1140: MachineRepresentation rep =
StackSlotRepresentationOf(node->op());
On 2016/01/28 at 09:02:18, titzer wrote:
> You should probably use ElemSizeLog2Of(MachineRepresentation) here.

Done.

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/pipeline.cc
File src/compiler/pipeline.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/src/compiler/pipeline.c...
src/compiler/pipeline.cc:1393: void Pipeline::InitializeFrame(CallDescriptor*
descriptor) {
On 2016/01/28 at 09:02:18, titzer wrote:
> You can just inline this method into the callers.

Done.

https://codereview.chromium.org/1645653002/diff/20001/test/cctest/compiler/te...
File test/cctest/compiler/test-run-machops.cc (right):

https://codereview.chromium.org/1645653002/diff/20001/test/cctest/compiler/te...
test/cctest/compiler/test-run-machops.cc:75: TEST(RunStackSlot) {
On 2016/01/28 at 09:02:18, titzer wrote:
> Can you write another test that tries the same for all machine types?

Done.

[titzer, 2016-01-28 15:17:02]

lgtm other than one comment

https://codereview.chromium.org/1645653002/diff/40001/src/compiler/instructio...
File src/compiler/instruction-scheduler.cc (right):

https://codereview.chromium.org/1645653002/diff/40001/src/compiler/instructio...
src/compiler/instruction-scheduler.cc:208: case kArchStackSlot:
Technically this is not a load, since it just computes the address of the stack
slot.

[ahaas, 2016-01-29 08:05:47]

On 2016/01/28 at 15:17:02, titzer wrote:
> lgtm other than one comment
> 
>
https://codereview.chromium.org/1645653002/diff/40001/src/compiler/instructio...
> File src/compiler/instruction-scheduler.cc (right):
> 
>
https://codereview.chromium.org/1645653002/diff/40001/src/compiler/instructio...
> src/compiler/instruction-scheduler.cc:208: case kArchStackSlot:
> Technically this is not a load, since it just computes the address of the
stack slot.

Done.

[ahaas, 2016-01-29 08:05:51]

The CQ bit was checked by ahaas@chromium.org

[ahaas, 2016-01-29 08:05:52]

The patchset sent to the CQ was uploaded after l-g-t-m from titzer@chromium.org
Link to the patchset: https://codereview.chromium.org/1645653002/#ps60001
(title: "fix comments.")

[commit-bot: I haz the power, 2016-01-29 08:06:03]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1645653002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1645653002/60001

[commit-bot: I haz the power, 2016-01-29 08:07:44]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-01-29 08:08:32]

Description was changed from

==========
[turbofan] Add the StackSlot operator to turbofan.

The StackSlot operator allows to allocate a spill slot on the stack. We
are going to use this operator to pass floats through pointers to c
functions, which we need for floating point rounding in the case where
the architecture does not provide rounding instructions.

R=titzer@chromium.org, v8-arm-ports@googlegroups.com,
v8-ppc-ports@googlegroups.com, v8-mips-ports@googlegroups.com
==========

to

==========
[turbofan] Add the StackSlot operator to turbofan.

The StackSlot operator allows to allocate a spill slot on the stack. We
are going to use this operator to pass floats through pointers to c
functions, which we need for floating point rounding in the case where
the architecture does not provide rounding instructions.

R=titzer@chromium.org, v8-arm-ports@googlegroups.com,
v8-ppc-ports@googlegroups.com, v8-mips-ports@googlegroups.com

Committed: https://crrev.com/7a693437787090d62d937b862e29521debcc5223
Cr-Commit-Position: refs/heads/master@{#33600}
==========

[commit-bot: I haz the power, 2016-01-29 08:08:33]

Patchset 4 (id:??) landed as
https://crrev.com/7a693437787090d62d937b862e29521debcc5223
Cr-Commit-Position: refs/heads/master@{#33600}

[ahaas, 2016-01-29 08:18:26]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/1644283002/ by ahaas@chromium.org.

The reason for reverting is: problems on Mac64.

[Michael Achenbach, 2016-01-29 08:22:03]

A case of outdated trybots. Just needs a rebase.

[Michael Achenbach, 2016-01-29 08:22:45]

On 2016/01/29 08:22:03, Michael Achenbach wrote:
> A case of outdated trybots. Just needs a rebase.

Or not. Maybe a mac64-only problem...

[Michael Achenbach, 2016-01-29 08:24:54]

On 2016/01/29 08:22:45, Michael Achenbach wrote:
> On 2016/01/29 08:22:03, Michael Achenbach wrote:
> > A case of outdated trybots. Just needs a rebase.
> 
> Or not. Maybe a mac64-only problem...

You can manually include the v8_mac64_rel trybot no a reland. It is not usually
run - this is the first ever problem of that kind since we have the bots...

[ahaas, 2016-01-29 09:41:11]

The CQ bit was checked by ahaas@chromium.org

[ahaas, 2016-01-29 09:41:11]

The patchset sent to the CQ was uploaded after l-g-t-m from titzer@chromium.org
Link to the patchset: https://codereview.chromium.org/1645653002/#ps80001
(title: "Explicitly state the type of test values.")

[commit-bot: I haz the power, 2016-01-29 09:41:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1645653002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1645653002/80001

[commit-bot: I haz the power, 2016-01-29 09:42:58]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-01-29 09:44:01]

Description was changed from

==========
[turbofan] Add the StackSlot operator to turbofan.

The StackSlot operator allows to allocate a spill slot on the stack. We
are going to use this operator to pass floats through pointers to c
functions, which we need for floating point rounding in the case where
the architecture does not provide rounding instructions.

R=titzer@chromium.org, v8-arm-ports@googlegroups.com,
v8-ppc-ports@googlegroups.com, v8-mips-ports@googlegroups.com

Committed: https://crrev.com/7a693437787090d62d937b862e29521debcc5223
Cr-Commit-Position: refs/heads/master@{#33600}
==========

to

==========
[turbofan] Add the StackSlot operator to turbofan.

The StackSlot operator allows to allocate a spill slot on the stack. We
are going to use this operator to pass floats through pointers to c
functions, which we need for floating point rounding in the case where
the architecture does not provide rounding instructions.

R=titzer@chromium.org, v8-arm-ports@googlegroups.com,
v8-ppc-ports@googlegroups.com, v8-mips-ports@googlegroups.com

Committed: https://crrev.com/7a693437787090d62d937b862e29521debcc5223
Cr-Commit-Position: refs/heads/master@{#33600}

Committed: https://crrev.com/645880373b13a96b812e3e4038c8e4044b2b34ca
Cr-Commit-Position: refs/heads/master@{#33606}
==========

[commit-bot: I haz the power, 2016-01-29 09:44:02]

Patchset 5 (id:??) landed as
https://crrev.com/645880373b13a96b812e3e4038c8e4044b2b34ca
Cr-Commit-Position: refs/heads/master@{#33606}