Adding dump provider for dumping java heap sizes on android.

Adding dump provider for dumping java heap sizes on android.

This CL follows crrev.com/1086943003. This CL adds a dump provider to
dump the values given by java.lang.Runtime class on dalvik heap
statistics into tracing.

BUG=477464
Committed: https://crrev.com/74add93a0342bc2f67a7f397b3881ef08d364a0a
Cr-Commit-Position: refs/heads/master@{#326817}

[ssid, 2015-04-16 09:39:14]

ssid@chromium.org changed reviewers:
+ picksi@chromium.org, primiano@chromium.org

[ssid, 2015-04-16 09:39:15]

Adding android java heap numbers.

[Primiano Tucci (use gerrit), 2015-04-21 18:36:01]

Almost good, few nits.

P.S: have you check that it actually works? :)

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
File base/trace_event/java_heap_dump_provider_android.cc (right):

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:15: const char
kDumperFriendlyName[] = "AndroidJavaHeap";
just "Java Heap"

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:16: const char kDumperName[]
= "android_java_heap";
"java_heap"

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:21:
AndrdoiJavaHeapDumpProvider* AndrdoiJavaHeapDumpProvider::GetInstance() {
Just call it JavaHeapDumpProvider, remove android (but keep it in the file name
to get the auto filtering)

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:38: // The total size used
by java heap and available memory out of the total that
I'd probably just say:
"These numbers come from java.lang.Runtime() stats"

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:40: long total_heap_memory,
free_heap_memory;
Initialize them to zero.
Also I think that our code style doesn't allow multiple decl in the same line,
just have two distinct lines.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:40: long total_heap_memory,
free_heap_memory;
remove "memory", as should be pretty clear.
just total_heap_size and free_heap_size

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
File base/trace_event/java_heap_dump_provider_android.h (right):

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.h:15: class BASE_EXPORT
AndrdoiJavaHeapDumpProvider : public MemoryDumpProvider {
This doesn't need to be BASE_EXPORT-ed. It's internal in base.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.h:15: class BASE_EXPORT
AndrdoiJavaHeapDumpProvider : public MemoryDumpProvider {
Nit: you miss a "d" here (Androi).

[ssid, 2015-04-22 10:40:11]

Addressed comments.

Yes it works, example values :
     (pid) : total free used:
    ( 2332): 12582012 3979800 8602212
    ( 2332): 12582012 3979800 8602212
    ( 2332): 12582012 3979800 860221

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
File base/trace_event/java_heap_dump_provider_android.cc (right):

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:15: const char
kDumperFriendlyName[] = "AndroidJavaHeap";
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> just "Java Heap"

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:16: const char kDumperName[]
= "android_java_heap";
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> "java_heap"

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:21:
AndrdoiJavaHeapDumpProvider* AndrdoiJavaHeapDumpProvider::GetInstance() {
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> Just call it JavaHeapDumpProvider, remove android (but keep it in the file
name
> to get the auto filtering)

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:38: // The total size used
by java heap and available memory out of the total that
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> I'd probably just say:
> "These numbers come from java.lang.Runtime() stats"

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.cc:40: long total_heap_memory,
free_heap_memory;
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> remove "memory", as should be pretty clear.
> just total_heap_size and free_heap_size

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
File base/trace_event/java_heap_dump_provider_android.h (right):

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.h:15: class BASE_EXPORT
AndrdoiJavaHeapDumpProvider : public MemoryDumpProvider {
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> This doesn't need to be BASE_EXPORT-ed. It's internal in base.

Done.

https://codereview.chromium.org/1050233004/diff/60001/base/trace_event/java_h...
base/trace_event/java_heap_dump_provider_android.h:15: class BASE_EXPORT
AndrdoiJavaHeapDumpProvider : public MemoryDumpProvider {
On 2015/04/21 18:36:01, Primiano Tucci wrote:
> Nit: you miss a "d" here (Androi).

Actually i have the d in middle :P. changed to JavaHeapDumpProvider

[Primiano Tucci (use gerrit), 2015-04-24 14:26:59]

LGTM but please send another CL soon to add a test.
You don't need to do anything in particular, Just call DumpInto, along the lines
of
https://chromium.googlesource.com/chromium/src/base/+/master/trace_event/winh...

This is to avoid that somebody accidentally breaks the android runtime JNI
plumbing.

[ssid, 2015-04-24 15:18:55]

On 2015/04/24 14:26:59, Primiano Tucci wrote:
> LGTM but please send another CL soon to add a test.
> You don't need to do anything in particular, Just call DumpInto, along the
lines
> of
>
https://chromium.googlesource.com/chromium/src/base/+/master/trace_event/winh...
> 
> This is to avoid that somebody accidentally breaks the android runtime JNI
> plumbing.

Will add test soon.

[ssid, 2015-04-24 15:19:53]

The CQ bit was checked by ssid@chromium.org

[commit-bot: I haz the power, 2015-04-24 15:20:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1050233004/100001

[commit-bot: I haz the power, 2015-04-24 16:35:23]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2015-04-24 16:36:12]

Patchset 6 (id:??) landed as
https://crrev.com/74add93a0342bc2f67a7f397b3881ef08d364a0a
Cr-Commit-Position: refs/heads/master@{#326817}