[compiler] Add relocatable pointer constants for wasm memory references.

[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when wasm GrowMemory is used. Code generator changes to accept relocatable constants as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}

Committed: https://crrev.com/52148c41c948f4e70778de758fb1a0ae118cecc2
Cr-Commit-Position: refs/heads/master@{#35407}

[gdeepti1, 2016-03-04 22:29:01]

Patchset #2 (id:20001) has been deleted

[bradnelson, 2016-03-06 21:53:14]

Some general questions, but I strongly defer to other reviewers on this as it
touches a lot of places.

https://codereview.chromium.org/1759383003/diff/40001/src/assembler.cc
File src/assembler.cc (right):

https://codereview.chromium.org/1759383003/diff/40001/src/assembler.cc#newcod...
src/assembler.cc:841: case WASM_MEMORY_REFERENCE:
Have you thought about what you'll want to call the heap size? WASM_MEMORY_SIZE?
I figure you'll be back through some of these for this purpose, so it might be
worth thinking about.

Actually, given that will touch many of the same spots, did you want to consider
adding the heap size in the same change?

https://codereview.chromium.org/1759383003/diff/40001/src/compiler/common-nod...
File src/compiler/common-node-cache.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/compiler/common-nod...
src/compiler/common-node-cache.h:55: Node** FindRelocatableInt32Constant(int32_t
value) {
Is Relocatable still a good name here?
Did you want to call this WasmMemory ... something?

[Yang, 2016-03-07 06:21:22]

https://codereview.chromium.org/1759383003/diff/40001/src/arm/assembler-arm-i...
File src/arm/assembler-arm-inl.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/arm/assembler-arm-i...
src/arm/assembler-arm-inl.h:121: void RelocInfo::update_wasm_memory_reference(
I think Ben's idea of "update" was to provide the original memory start, the new
memory start, and let the method take care of the offset calculation.

The way it is currently implemented ought to be called set_wasm_memory_reference
to be inline with other setters and getters.

https://codereview.chromium.org/1759383003/diff/40001/src/arm64/assembler-arm...
File src/arm64/assembler-arm64.cc (right):

https://codereview.chromium.org/1759383003/diff/40001/src/arm64/assembler-arm...
src/arm64/assembler-arm64.cc:444: (mode >= RelocInfo::WASM_MEMORY_REFERENCE));
This seems wrong. Previously a RelocInfo::CELL is considered shareable. Now it
no longer is.

Can we introduce a new entry to the enum called
RelocInfo::FIRST_SHAREABLE_RELOC_MODE that is equals to RelocInfo::CELL and use
that?

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h
File src/x64/assembler-x64.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h...
src/x64/assembler-x64.h:338: value_ =
static_cast<int32_t>(reinterpret_cast<intptr_t>(value));
Is this useful at all? Any address that you pass in here will be truncated to
32-bit. This would only make sense for x32, and I'm not sure x32 even works at
this point.

[titzer, 2016-03-07 19:34:09]

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h
File src/x64/assembler-x64.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h...
src/x64/assembler-x64.h:338: value_ =
static_cast<int32_t>(reinterpret_cast<intptr_t>(value));
On 2016/03/07 06:21:22, Yang wrote:
> Is this useful at all? Any address that you pass in here will be truncated to
> 32-bit. This would only make sense for x32, and I'm not sure x32 even works at
> this point.

Yeah, I don't think we can allow addresses as immediates here.

[gdeepti1, 2016-03-08 04:22:51]

https://codereview.chromium.org/1759383003/diff/40001/src/arm/assembler-arm-i...
File src/arm/assembler-arm-inl.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/arm/assembler-arm-i...
src/arm/assembler-arm-inl.h:121: void RelocInfo::update_wasm_memory_reference(
On 2016/03/07 06:21:22, Yang wrote:
> I think Ben's idea of "update" was to provide the original memory start, the
new
> memory start, and let the method take care of the offset calculation.
> 
> The way it is currently implemented ought to be called
set_wasm_memory_reference
> to be inline with other setters and getters.

Changed the updated interface -
https://codereview.chromium.org/1759873002/diff/180001/src/arm/assembler-arm-...

https://codereview.chromium.org/1759383003/diff/40001/src/arm64/assembler-arm...
File src/arm64/assembler-arm64.cc (right):

https://codereview.chromium.org/1759383003/diff/40001/src/arm64/assembler-arm...
src/arm64/assembler-arm64.cc:444: (mode >= RelocInfo::WASM_MEMORY_REFERENCE));
On 2016/03/07 06:21:22, Yang wrote:
> This seems wrong. Previously a RelocInfo::CELL is considered shareable. Now it
> no longer is.
> 
> Can we introduce a new entry to the enum called
> RelocInfo::FIRST_SHAREABLE_RELOC_MODE that is equals to RelocInfo::CELL and
use
> that?

Fixed now - 
https://codereview.chromium.org/1759873002/diff/180001/src/assembler.h
https://codereview.chromium.org/1759873002/diff/180001/src/arm64/assembler-ar...

https://codereview.chromium.org/1759383003/diff/40001/src/assembler.cc
File src/assembler.cc (right):

https://codereview.chromium.org/1759383003/diff/40001/src/assembler.cc#newcod...
src/assembler.cc:841: case WASM_MEMORY_REFERENCE:
On 2016/03/06 21:53:14, bradnelson wrote:
> Have you thought about what you'll want to call the heap size?
WASM_MEMORY_SIZE?
> I figure you'll be back through some of these for this purpose, so it might be
> worth thinking about.
> 
> Actually, given that will touch many of the same spots, did you want to
consider
> adding the heap size in the same change?

I was thinking about calling it WASM_MEMORY_SIZE. I want to keep it separate
from this change mostly because I'm a little fuzzy on exactly what is needed for
memory size. Specifically about what needs to be done for wasm if the
update_wasm_memory_reference takes care of bounds check and if anything needs to
be added for the asm.js CheckedLoad/Store semantics.

https://codereview.chromium.org/1759383003/diff/40001/src/compiler/common-nod...
File src/compiler/common-node-cache.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/compiler/common-nod...
src/compiler/common-node-cache.h:55: Node** FindRelocatableInt32Constant(int32_t
value) {
On 2016/03/06 21:53:14, bradnelson wrote:
> Is Relocatable still a good name here?
> Did you want to call this WasmMemory ... something?

As these Int32/Int64 constants that take a RelocInfo::Mode I figured it was
still a good enough name. Would WasmInt32Constant/WasmMemoryInt32Constant be
better?

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h
File src/x64/assembler-x64.h (right):

https://codereview.chromium.org/1759383003/diff/40001/src/x64/assembler-x64.h...
src/x64/assembler-x64.h:338: value_ =
static_cast<int32_t>(reinterpret_cast<intptr_t>(value));
On 2016/03/07 19:34:09, titzer wrote:
> On 2016/03/07 06:21:22, Yang wrote:
> > Is this useful at all? Any address that you pass in here will be truncated
to
> > 32-bit. This would only make sense for x32, and I'm not sure x32 even works
at
> > this point.
> 
> Yeah, I don't think we can allow addresses as immediates here.

Fixed to not use Address -
https://codereview.chromium.org/1759873002/diff/180001/src/x64/assembler-x64.h
It works on ia32 and x64 because for both the Immediate class is only used for
32-bit values, movq on x64 with the 64-bit case is handled separately.

[gdeepti1, 2016-03-09 03:45:33]

Patchset #7 (id:140001) has been deleted

[gdeepti1, 2016-03-18 23:47:45]

PTAL

https://codereview.chromium.org/1759383003/diff/240001/src/x64/macro-assemble...
File src/x64/macro-assembler-x64.cc (left):

https://codereview.chromium.org/1759383003/diff/240001/src/x64/macro-assemble...
src/x64/macro-assembler-x64.cc:1117: void MacroAssembler::Set(Register dst,
int64_t x, RelocInfo::Mode rmode) {
No longer required as the optimizations and check for x being 0 are removed,
src/compiler/x64/code-generator-x64.cc now generates the movq instruction.
Additional checking removed because correct patchable constants are being
generated, x = 0 can be allowed in the future to check for Wasm with modules
with no memory.

[titzer, 2016-03-23 20:35:36]

lgtm other than splitting up tests.

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier.cc
File src/compiler/verifier.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier....
src/compiler/verifier.cc:350: // TODO(gdeepti): Add accurate checking for
relocatable constants.
You can probably drop this TODO. This looks fine to me.

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
File src/compiler/x64/instruction-selector-x64.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
src/compiler/x64/instruction-selector-x64.cc:27: case
IrOpcode::kRelocatableInt64Constant:
Probably don't want this case here.

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
File test/cctest/compiler/test-run-wasm-machops.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
test/cctest/compiler/test-run-wasm-machops.cc:21: static void
RunLoadStoreRelocation(MachineType rep) {
Can you split out some simple cases (e.g. offset 0) into their own separate
tests?

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
test/cctest/compiler/test-run-wasm-machops.cc:59: Handle<Code> code =
m.GetCode();
This is OK for now, but this operation probably needs to be pulled out to an
operation in the production code that you can simply call.

[Yang, 2016-03-24 05:27:31]

On 2016/03/23 20:35:36, titzer wrote:
> lgtm other than splitting up tests.
> 
>
https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier.cc
> File src/compiler/verifier.cc (right):
> 
>
https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier....
> src/compiler/verifier.cc:350: // TODO(gdeepti): Add accurate checking for
> relocatable constants.
> You can probably drop this TODO. This looks fine to me.
> 
>
https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
> File src/compiler/x64/instruction-selector-x64.cc (right):
> 
>
https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
> src/compiler/x64/instruction-selector-x64.cc:27: case
> IrOpcode::kRelocatableInt64Constant:
> Probably don't want this case here.
> 
>
https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
> File test/cctest/compiler/test-run-wasm-machops.cc (right):
> 
>
https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
> test/cctest/compiler/test-run-wasm-machops.cc:21: static void
> RunLoadStoreRelocation(MachineType rep) {
> Can you split out some simple cases (e.g. offset 0) into their own separate
> tests?
> 
>
https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
> test/cctest/compiler/test-run-wasm-machops.cc:59: Handle<Code> code =
> m.GetCode();
> This is OK for now, but this operation probably needs to be pulled out to an
> operation in the production code that you can simply call.

I don't think this requires my review. Everything in this CL is in src/compiler/
after the assembler changes were split off.

[Yang, 2016-03-29 12:06:39]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

[Yang, 2016-03-29 12:06:39]

yangguo@chromium.org changed reviewers:
- yangguo@chromium.org

[gdeepti1, 2016-03-31 22:29:54]

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier.cc
File src/compiler/verifier.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/verifier....
src/compiler/verifier.cc:350: // TODO(gdeepti): Add accurate checking for
relocatable constants.
On 2016/03/23 20:35:36, titzer wrote:
> You can probably drop this TODO. This looks fine to me.

Done.

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
File src/compiler/x64/instruction-selector-x64.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/src/compiler/x64/instr...
src/compiler/x64/instruction-selector-x64.cc:27: case
IrOpcode::kRelocatableInt64Constant:
On 2016/03/23 20:35:36, titzer wrote:
> Probably don't want this case here.

Done.

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
File test/cctest/compiler/test-run-wasm-machops.cc (right):

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
test/cctest/compiler/test-run-wasm-machops.cc:21: static void
RunLoadStoreRelocation(MachineType rep) {
On 2016/03/23 20:35:36, titzer wrote:
> Can you split out some simple cases (e.g. offset 0) into their own separate
> tests?

Added a simple load/store test with 0 offset.

https://codereview.chromium.org/1759383003/diff/240001/test/cctest/compiler/t...
test/cctest/compiler/test-run-wasm-machops.cc:59: Handle<Code> code =
m.GetCode();
On 2016/03/23 20:35:36, titzer wrote:
> This is OK for now, but this operation probably needs to be pulled out to an
> operation in the production code that you can simply call.

Pulled out a method to update memory references.

[gdeepti1, 2016-03-31 22:31:33]

The CQ bit was checked by gdeepti@google.com

[gdeepti1, 2016-03-31 22:31:34]

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/1759383003/#ps300001
(title: "Fix bad test")

[commit-bot: I haz the power, 2016-03-31 22:31:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1759383003/300001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1759383003/300001

[commit-bot: I haz the power, 2016-04-01 00:39:26]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

[commit-bot: I haz the power, 2016-04-01 00:39:28]

Committed patchset #14 (id:300001)

[commit-bot: I haz the power, 2016-04-01 00:41:40]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

[commit-bot: I haz the power, 2016-04-01 00:41:41]

Patchset 14 (id:??) landed as
https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

[Yang, 2016-04-01 05:52:18]

A revert of this CL (patchset #14 id:300001) has been created in
https://codereview.chromium.org/1846083005/ by yangguo@chromium.org.

The reason for reverting is: Test failures:
https://build.chromium.org/p/client.v8/builders/V8%20Mac64/builds/8046.

[Benedikt Meurer, 2016-04-01 05:54:48]

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

[Benedikt Meurer, 2016-04-01 05:54:49]

I'm probably a bit late to the discussion, somehow missed it so far, but have we
considered putting the address into the header of the Code object? That way we
don't need all the relocinfos and the constant pool entries, plus we still have
space left in the header anyway (or we just use a field that is otherwise unused
for WASM functions). For example materializing the memory reference on x64 would
then be

movq reg, [rip-<offset to header entry>]

instead of

movq reg, <64-bit constant>

plus RelocInfo for every materialization. GrowMemory only needs to update the
field in the Code header then.

[bradn, 2016-04-01 18:46:05]

bradnelson@google.com changed reviewers:
+ bradnelson@google.com

[bradn, 2016-04-01 18:46:06]

Hi Benedikt,

Good suggestions are always welcome!

So we'd talked about it way back (titzer and I), but hadn't revisited it.

titzer (correct me if I'm off ben) believes that burning immediates for constant
offsets from the heap base is a win (we have more of those than asm.js because
wasm doesn't have globals, so those go to the heap).
But it's an excellent point that for offset 0, and the heap size (which deepti
was going to land in a separate CL), that putting it in the header would likely
be better.

Deepti, how about start by doing the heap size the way Benedikt proposes.
There's a bunch of unused fields (maybe grab one of the type feedback offsets
and share under a new name).

Then circle back and add it for heap base with offset 0, as I assume you'll
reland this (I think the revert was spurious but double check that).

How does that sound bmeurer?
And thoughts titzer?

Thanks

[titzer, 2016-04-01 19:00:57]

On 2016/04/01 18:46:06, bradn wrote:
> Hi Benedikt,
> 
> Good suggestions are always welcome!
> 
> So we'd talked about it way back (titzer and I), but hadn't revisited it.
> 
> titzer (correct me if I'm off ben) believes that burning immediates for
constant
> offsets from the heap base is a win (we have more of those than asm.js because
> wasm doesn't have globals, so those go to the heap).
> But it's an excellent point that for offset 0, and the heap size (which deepti
> was going to land in a separate CL), that putting it in the header would
likely
> be better.
> 
> Deepti, how about start by doing the heap size the way Benedikt proposes.
> There's a bunch of unused fields (maybe grab one of the type feedback offsets
> and share under a new name).
> 
> Then circle back and add it for heap base with offset 0, as I assume you'll
> reland this (I think the revert was spurious but double check that).
> 
> How does that sound bmeurer?
> And thoughts titzer?
> 

I discussed with Benedikt and I think it's OK to get everything working with
what Deepti has now. In general we want to be smarter about how TurboFan
backends materialize constants so that we can share entries. In the long run
we'll want to probably generate one massive code object and have only a few
places where the constants are stored and need to be patched.

[gdeepti1, 2016-04-11 21:34:28]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

[gdeepti1, 2016-04-11 22:29:48]

On 2016/04/01 05:52:18, Yang wrote:
> A revert of this CL (patchset #14 id:300001) has been created in
> https://codereview.chromium.org/1846083005/ by mailto:yangguo@chromium.org.
> 
> The reason for reverting is: Test failures:
> https://build.chromium.org/p/client.v8/builders/V8%20Mac64/builds/8046.

Tweaked to fix mac tests, reland attempt as the tests pass on my macbook.

[gdeepti1, 2016-04-11 22:30:00]

The CQ bit was checked by gdeepti@google.com

[gdeepti1, 2016-04-11 22:30:01]

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/1759383003/#ps320001
(title: "Tweaks to fix mac tests")

[commit-bot: I haz the power, 2016-04-11 22:30:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1759383003/320001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1759383003/320001

[commit-bot: I haz the power, 2016-04-11 22:33:19]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

[commit-bot: I haz the power, 2016-04-11 22:33:21]

Committed patchset #15 (id:320001)

[commit-bot: I haz the power, 2016-04-11 22:35:10]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

[commit-bot: I haz the power, 2016-04-11 22:35:11]

Patchset 15 (id:??) landed as
https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}

[Michael Achenbach, 2016-04-12 07:48:21]

A revert of this CL (patchset #15 id:320001) has been created in
https://codereview.chromium.org/1881913002/ by machenbach@chromium.org.

The reason for reverting is: [Sheriff] Breaks msan:
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm64%20-%20....

[gdeepti1, 2016-04-12 08:01:08]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

[gdeepti1, 2016-04-12 09:02:50]

On 2016/04/12 07:48:21, Michael Achenbach wrote:
> A revert of this CL (patchset #15 id:320001) has been created in
> https://codereview.chromium.org/1881913002/ by mailto:machenbach@chromium.org.
> 
> The reason for reverting is: [Sheriff] Breaks msan:
>
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm64%20-%20....

Fixed msan, v8_linux64_msan_rel now passes.

[gdeepti1, 2016-04-12 09:03:03]

The CQ bit was checked by gdeepti@google.com

[gdeepti1, 2016-04-12 09:03:03]

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/1759383003/#ps340001
(title: "Fix uninitialized variable causing msan failure")

[commit-bot: I haz the power, 2016-04-12 09:03:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1759383003/340001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1759383003/340001

[commit-bot: I haz the power, 2016-04-12 09:05:12]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

[commit-bot: I haz the power, 2016-04-12 09:05:14]

Committed patchset #16 (id:340001)

[commit-bot: I haz the power, 2016-04-12 09:07:05]

Description was changed from

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}
==========

to

==========
[compiler] Add relocatable pointer constants for wasm memory references.

Add relocatable pointers for wasm memory references that need to be updated when
wasm GrowMemory is used. Code generator changes to accept relocatable constants
as immediates.

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

Committed: https://crrev.com/eb5fe0df64ec0add423b2a1f6fb62d5a33dce2a5
Cr-Commit-Position: refs/heads/master@{#35182}

Committed: https://crrev.com/297932a302ce0b73c3618ef9e4eba9d9d241f2b3
Cr-Commit-Position: refs/heads/master@{#35400}

Committed: https://crrev.com/52148c41c948f4e70778de758fb1a0ae118cecc2
Cr-Commit-Position: refs/heads/master@{#35407}
==========

[commit-bot: I haz the power, 2016-04-12 09:07:07]

Patchset 16 (id:??) landed as
https://crrev.com/52148c41c948f4e70778de758fb1a0ae118cecc2
Cr-Commit-Position: refs/heads/master@{#35407}