Enable allocator shim for Android

Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886, 598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org

Committed: https://crrev.com/ebb95496c73dc0d5ce83968ac619921f154305f7
Cr-Commit-Position: refs/heads/master@{#386386}

[Primiano Tucci (use gerrit), 2016-04-11 14:28:55]

Description was changed from

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
==========

to

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org
==========

[Primiano Tucci (use gerrit), 2016-04-11 14:28:55]

primiano@chromium.org changed reviewers:
+ thakis@chromium.org

[Primiano Tucci (use gerrit), 2016-04-11 14:29:45]

Nico, TBR-ing this as this is just split out of crrev.com/1719433002.
Let's see how the waterfall react to this, it will be another fun week. Take #1
landing in 3... 2... 1...

[Nico, 2016-04-11 14:42:10]

lgtm

[Primiano Tucci (use gerrit), 2016-04-11 15:04:36]

The CQ bit was checked by primiano@chromium.org

[commit-bot: I haz the power, 2016-04-11 15:04:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1875043003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1875043003/20001

[commit-bot: I haz the power, 2016-04-11 15:09:58]

Description was changed from

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org
==========

to

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org
==========

[commit-bot: I haz the power, 2016-04-11 15:09:59]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-04-11 15:11:06]

Description was changed from

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org
==========

to

==========
Enable allocator shim for Android

This is a follow-up to crrev.com/1719433002, which introduced the
shim for Android, and enables it by default by setting
use_experimental_allocator_shim=true for Android.

Build/Perf sheriffs heads up
----------------------------
If you see any build error or crash related with __wrap_malloc,
__wrap_free, __real_malloc, __real_free, etc this CL is to blame.

Performance considerations
------------------------
Binary size diff (GN, arm, static, official build): 24k

I did a mixture of local and trybots run to estimate the perf impact
of this change. Didn't get any conclusive data, everything I tried
seems in the same ballpark, below noise levels. More in details:

cc_perftests.PrepareTiles on a Nexus 4.
Rationale of the choice: in a previous CL (crbug.com/593344), this
benchmark revealed the presence of two mfences in the malloc path.
Results: https://goo.gl/8VC3Jp in the same ballpark.

page-cycler on Nexus 9 via trybots:
Results: http://goo.gl/J3i50a seems to suggest that this CL improves
both warm and cold times in most cases. I doubt it, more likely it's
noise.

All the other perf trybots failed. The perf waterfall seems to be in a
bad state in these days.

BUG=550886,598075
TEST=base_unittests --gtest_filter=AllocatorShimTest.*
TBR=thakis@chromium.org

Committed: https://crrev.com/ebb95496c73dc0d5ce83968ac619921f154305f7
Cr-Commit-Position: refs/heads/master@{#386386}
==========

[commit-bot: I haz the power, 2016-04-11 15:11:07]

Patchset 2 (id:??) landed as
https://crrev.com/ebb95496c73dc0d5ce83968ac619921f154305f7
Cr-Commit-Position: refs/heads/master@{#386386}

[DaleCurtis, 2016-04-11 17:52:34]

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

[DaleCurtis, 2016-04-11 17:52:34]

I think this broke the Android clang build:

In file included from ../../base/allocator/allocator_shim.cc:233:
../../base/allocator/allocator_shim_override_cpp_symbols.h:20:25: error:
function previously declared with an explicit exception specification redeclared
with an implicit exception specification
[-Werror,-Wimplicit-exception-spec-mismatch]
SHIM_ALWAYS_EXPORT void operator delete(void* p) __THROW
                        ^
../../third_party/android_tools/ndk/sources/cxx-stl/llvm-libc++/libcxx/include/new:133:30:
note: previous declaration is here
_LIBCPP_NEW_DELETE_VIS void  operator delete(void* __p) _NOEXCEPT;
                             ^
Or at least it has for me locally on a clean ToT build.

[DaleCurtis, 2016-04-11 17:53:30]

dcheck_always_on = true
disable_incremental_isolated_processes = true
enable_nacl = false
target_os = "android"
is_debug = false
use_goma = true
proprietary_codecs = true
is_component_build = true
is_clang = true
ffmpeg_branding = "Chrome"

[Nico, 2016-04-11 18:11:14]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/1881523003/ by thakis@chromium.org.

The reason for reverting is: Looks like this breaks android clang builds..

[Nico, 2016-04-11 18:11:33]

Huh weird,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_clan...
is happy. Maybe gn-specific? Anyhoo, reverted for now.

[DaleCurtis, 2016-04-11 18:12:15]

Sheriff says primiano@ is on the fix, clang bots build with allocator disabled
for asan.

[Primiano Tucci (use gerrit), 2016-04-11 18:19:53]

On 2016/04/11 18:12:15, DaleCurtis wrote:
> Sheriff says primiano@ is on the fix, clang bots build with allocator disabled
> for asan.

I think I know what's going on: this, as is, breaks 100% clang builds. This did
pass the CQ and the main waterfall because all the clang builds seem to build
with asan=1, which disables all the logic of this CL.
I have a possible fix here:
https://codereview.chromium.org/1874413002

[Primiano Tucci (use gerrit), 2016-04-11 18:49:42]

Reland w/ fix coming in https://codereview.chromium.org/1875173002/