[heap] Decouple old generation allocation limit from external memory.

[heap] Decouple old generation allocation limit from external memory.

We check for external memory limit in Heap::ReportExternalMemoryPressure.

BUG=chromium:616434
Committed: https://crrev.com/672d079ccba686019fa1457c83b42c2e692ef88b
Cr-Commit-Position: refs/heads/master@{#39374}

[ulan, 2016-09-12 13:50:41]

ulan@chromium.org changed reviewers:
+ hpayer@chromium.org

[ulan, 2016-09-12 13:50:41]

ptal

[Hannes Payer (out of office), 2016-09-13 06:31:48]

lgtm

[ulan, 2016-09-13 08:10:46]

The CQ bit was checked by ulan@chromium.org

[commit-bot: I haz the power, 2016-09-13 08:10:56]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-13 08:54:00]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-09-13 08:54:13]

Description was changed from

==========
[heap] Decouple old generation allocation limit from external memory.

We check for external memory limit in Heap::ReportExternalMemoryPressure.

BUG=chromium:616434
==========

to

==========
[heap] Decouple old generation allocation limit from external memory.

We check for external memory limit in Heap::ReportExternalMemoryPressure.

BUG=chromium:616434

Committed: https://crrev.com/672d079ccba686019fa1457c83b42c2e692ef88b
Cr-Commit-Position: refs/heads/master@{#39374}
==========

[commit-bot: I haz the power, 2016-09-13 08:54:14]

Patchset 1 (id:??) landed as
https://crrev.com/672d079ccba686019fa1457c83b42c2e692ef88b
Cr-Commit-Position: refs/heads/master@{#39374}

[Michael Achenbach, 2016-09-13 10:27:01]

Did this possibly make arm-sim flaky?
https://build.chromium.org/p/client.v8.ports/builders/V8%20Linux%20-%20arm%20...

[ulan, 2016-09-13 10:42:19]

On 2016/09/13 10:27:01, machenbach (slow) wrote:
> Did this possibly make arm-sim flaky?
>
https://build.chromium.org/p/client.v8.ports/builders/V8%20Linux%20-%20arm%20...

Unlikely. This CL changes only GC timing.

[Michael Achenbach, 2016-09-13 12:15:53]

Reproed locally that it comes from this CL. After this CL, 10 of 20 runs fail.
Bt:

(gdb) bt
#0  0xf7fdb440 in __kernel_vsyscall ()
#1  0xf7cba687 in raise () from /lib/i386-linux-gnu/libc.so.6
#2  0xf7cbdab3 in abort () from /lib/i386-linux-gnu/libc.so.6
#3  0xf7cf4fd3 in ?? () from /lib/i386-linux-gnu/libc.so.6
#4  0xf7cff4ba in ?? () from /lib/i386-linux-gnu/libc.so.6
#5  0xf7d0012d in ?? () from /lib/i386-linux-gnu/libc.so.6
#6  0x565ada2a in
v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>::Resize(v8::base::DefaultAllocationPolicy)
()
#7  0x565ad867 in
v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>::InsertNew(void*,
unsigned int, v8::base::DefaultAllocationPolicy) ()
#8  0x56cce7a3 in
v8::internal::Simulator::FlushICache(v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>*,
void*, unsigned int) ()
#9  0x565e7182 in
v8::internal::AssemblerBase::FlushICache(v8::internal::Isolate*, void*, unsigned
int) ()
#10 0x56a7c5b2 in v8::internal::Code::Relocate(int) ()
#11 0x56980628 in void
v8::internal::MarkCompactCollector::EvacuateVisitorBase::MigrateObject<(v8::internal::MarkCompactCollector::EvacuateVisitorBase::MigrationMode)0>(v8::internal::HeapObject*,
v8::internal::HeapObject*, int, v8::internal::AllocationSpace) ()
#12 0x5697fbaa in
v8::internal::MarkCompactCollector::EvacuateVisitorBase::TryEvacuateObject(v8::internal::PagedSpace*,
v8::internal::HeapObject*, v8::internal::HeapObject**) ()
#13 0x5698921b in bool
v8::internal::MarkCompactCollector::VisitLiveObjects<v8::internal::MarkCompactCollector::EvacuateOldSpaceVisitor>(v8::internal::MemoryChunk*,
v8::internal::MarkCompactCollector::EvacuateOldSpaceVisitor*,
v8::internal::MarkCompactCollector::IterationMode) ()
#14 0x56988dd4 in
v8::internal::MarkCompactCollector::Evacuator::EvacuatePage(v8::internal::Page*)
()
#15 0x56988acc in
v8::internal::PageParallelJob<v8::internal::EvacuationJobTraits>::Task::RunInternal()
()
#16 0x5667d104 in non-virtual thunk to v8::internal::CancelableTask::Run() ()
#17 0x565b5ca8 in v8::platform::WorkerThread::Run() ()
#18 0x56d52093 in v8::base::ThreadEntry(void*) ()
#19 0xf7f96f72 in start_thread () from /lib/i386-linux-gnu/libpthread.so.0
#20 0xf7d77f8e in clone () from /lib/i386-linux-gnu/libc.so.6

[ulan, 2016-09-13 13:26:30]

On 2016/09/13 12:15:53, machenbach (slow) wrote:
> Reproed locally that it comes from this CL. After this CL, 10 of 20 runs fail.
> Bt:
> 
> (gdb) bt
> #0  0xf7fdb440 in __kernel_vsyscall ()
> #1  0xf7cba687 in raise () from /lib/i386-linux-gnu/libc.so.6
> #2  0xf7cbdab3 in abort () from /lib/i386-linux-gnu/libc.so.6
> #3  0xf7cf4fd3 in ?? () from /lib/i386-linux-gnu/libc.so.6
> #4  0xf7cff4ba in ?? () from /lib/i386-linux-gnu/libc.so.6
> #5  0xf7d0012d in ?? () from /lib/i386-linux-gnu/libc.so.6
> #6  0x565ada2a in
>
v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>::Resize(v8::base::DefaultAllocationPolicy)
> ()
> #7  0x565ad867 in
>
v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>::InsertNew(void*,
> unsigned int, v8::base::DefaultAllocationPolicy) ()
> #8  0x56cce7a3 in
>
v8::internal::Simulator::FlushICache(v8::base::TemplateHashMapImpl<v8::base::DefaultAllocationPolicy>*,
> void*, unsigned int) ()
> #9  0x565e7182 in
> v8::internal::AssemblerBase::FlushICache(v8::internal::Isolate*, void*,
unsigned
> int) ()
> #10 0x56a7c5b2 in v8::internal::Code::Relocate(int) ()
> #11 0x56980628 in void
>
v8::internal::MarkCompactCollector::EvacuateVisitorBase::MigrateObject<(v8::internal::MarkCompactCollector::EvacuateVisitorBase::MigrationMode)0>(v8::internal::HeapObject*,
> v8::internal::HeapObject*, int, v8::internal::AllocationSpace) ()
> #12 0x5697fbaa in
>
v8::internal::MarkCompactCollector::EvacuateVisitorBase::TryEvacuateObject(v8::internal::PagedSpace*,
> v8::internal::HeapObject*, v8::internal::HeapObject**) ()
> #13 0x5698921b in bool
>
v8::internal::MarkCompactCollector::VisitLiveObjects<v8::internal::MarkCompactCollector::EvacuateOldSpaceVisitor>(v8::internal::MemoryChunk*,
> v8::internal::MarkCompactCollector::EvacuateOldSpaceVisitor*,
> v8::internal::MarkCompactCollector::IterationMode) ()
> #14 0x56988dd4 in
>
v8::internal::MarkCompactCollector::Evacuator::EvacuatePage(v8::internal::Page*)
> ()
> #15 0x56988acc in
>
v8::internal::PageParallelJob<v8::internal::EvacuationJobTraits>::Task::RunInternal()
> ()
> #16 0x5667d104 in non-virtual thunk to v8::internal::CancelableTask::Run() ()
> #17 0x565b5ca8 in v8::platform::WorkerThread::Run() ()
> #18 0x56d52093 in v8::base::ThreadEntry(void*) ()
> #19 0xf7f96f72 in start_thread () from /lib/i386-linux-gnu/libpthread.so.0
> #20 0xf7d77f8e in clone () from /lib/i386-linux-gnu/libc.so.6

This CL flushes out an existing race condition. Uploaded a fix:
https://codereview.chromium.org/2338793002/

[ulan, 2016-09-15 09:41:04]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/2339033005/ by ulan@chromium.org.

The reason for reverting is: Regressions in telemetry benchmarks:
crbug.com/646819. .