Fix the size dumped by child discardable manager for tracing.

Fix the size dumped by child discardable manager for tracing.

The child discardable manager dumps the size of the discardable memory
allocated as length of span, which is actually the number of blocks.
This CL fixes that using the length * block_size. The sub-allocations
are also added under allocated_objects dump.

BUG=503168
Committed: https://crrev.com/053b1f28b2a97bf1e0eaf11d88c291bef65733aa
Cr-Commit-Position: refs/heads/master@{#345846}

[ssid, 2015-08-26 15:48:06]

ssid@chromium.org changed reviewers:
+ avi@chromium.org

[ssid, 2015-08-26 15:48:07]

Small fix for the tracing. PTAL, thanks.

[Avi (use Gerrit), 2015-08-26 15:52:22]

lgtm

[reveman, 2015-08-26 15:56:17]

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

[reveman, 2015-08-26 15:56:18]

https://codereview.chromium.org/1308213005/diff/1/content/common/discardable_...
File content/common/discardable_shared_memory_heap.cc (right):

https://codereview.chromium.org/1308213005/diff/1/content/common/discardable_...
content/common/discardable_shared_memory_heap.cc:77:
static_cast<uint64_t>(span->shared_memory()->mapped_size()));
hm, mapped_size is the total size of the segment but we're creating a dump just
for the span, right? Should this not be "span->length() * block_size" ?

[ssid, 2015-08-26 16:00:35]

Yes done.

https://codereview.chromium.org/1308213005/diff/1/content/common/discardable_...
File content/common/discardable_shared_memory_heap.cc (right):

https://codereview.chromium.org/1308213005/diff/1/content/common/discardable_...
content/common/discardable_shared_memory_heap.cc:77:
static_cast<uint64_t>(span->shared_memory()->mapped_size()));
On 2015/08/26 15:56:18, reveman wrote:
> hm, mapped_size is the total size of the segment but we're creating a dump
just
> for the span, right? Should this not be "span->length() * block_size" ?

Ah yes, I got confused here because of the host side. Thanks

[reveman, 2015-08-26 16:02:09]

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
File content/common/discardable_shared_memory_heap.cc (right):

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
content/common/discardable_shared_memory_heap.cc:82:
base::StringPrintf("discardable/segment_%d/allocated_objects", id_));
why are you adding /allocated_objects here?

[reveman, 2015-08-26 16:04:59]

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
File content/common/discardable_shared_memory_heap.h (right):

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
content/common/discardable_shared_memory_heap.h:121: size_t block_size,
Maybe remove |span| and |block_size| args and just pass "size_t size" here?
we're not using |span| for anything but a DCHECK..

[ssid, 2015-08-26 16:05:57]

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
File content/common/discardable_shared_memory_heap.cc (right):

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
content/common/discardable_shared_memory_heap.cc:82:
base::StringPrintf("discardable/segment_%d/allocated_objects", id_));
On 2015/08/26 16:02:09, reveman wrote:
> why are you adding /allocated_objects here?

The sub-allocations are usually added under "allocated_objects". Like in the
case of malloc.

This is because the children of segment (if any in future) will not be double
counted along with these sub-allocations. But will be represented as "children
is same as allocated_objects" with ownership edge.

[reveman, 2015-08-26 16:21:05]

lgtm

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
File content/common/discardable_shared_memory_heap.cc (right):

https://codereview.chromium.org/1308213005/diff/20001/content/common/discarda...
content/common/discardable_shared_memory_heap.cc:82:
base::StringPrintf("discardable/segment_%d/allocated_objects", id_));
On 2015/08/26 at 16:05:57, ssid wrote:
> On 2015/08/26 16:02:09, reveman wrote:
> > why are you adding /allocated_objects here?
> 
> The sub-allocations are usually added under "allocated_objects". Like in the
case of malloc.
> 
> This is because the children of segment (if any in future) will not be double
counted along with these sub-allocations. But will be represented as "children
is same as allocated_objects" with ownership edge.

Ack.

[ssid, 2015-08-27 10:26:29]

The CQ bit was checked by ssid@chromium.org

[ssid, 2015-08-27 10:26:29]

The patchset sent to the CQ was uploaded after l-g-t-m from avi@chromium.org
Link to the patchset: https://codereview.chromium.org/1308213005/#ps20001
(title: "Change to block size.")

[commit-bot: I haz the power, 2015-08-27 10:26:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1308213005/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1308213005/20001

[commit-bot: I haz the power, 2015-08-27 11:15:04]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-08-27 11:15:40]

Patchset 2 (id:??) landed as
https://crrev.com/053b1f28b2a97bf1e0eaf11d88c291bef65733aa
Cr-Commit-Position: refs/heads/master@{#345846}