PartitionAlloc: fixes and improvements to large-chunk size tracking.

PartitionAlloc: fixes an improvements to large-chunk size tracking.

- Fix a bug where the size for a chunk wasn't being updated correctly.
- Use accurate allocation size, if known, for more accurate cookie placement in
debug. This results in strong out-of-bounds write detection.
- Track accurate allocation size for direct mapped chunks, with the same cookie
benefits.
- Tighten up ASSERTing in this area.

BUG=501878
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197520

[Chris Evans, 2015-06-18 21:38:30]

cevans@chromium.org changed reviewers:
+ haraken@chromium.org

[haraken, 2015-06-19 12:48:17]

LGTM

https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
File Source/wtf/PartitionAlloc.cpp (right):

https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
Source/wtf/PartitionAlloc.cpp:622: static ALWAYS_INLINE void*
partitionDirectMap(PartitionRootBase* root, int flags, size_t origSize)

origSize => rawSize

https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
Source/wtf/PartitionAlloc.cpp:911: page->freelistHead = nullptr;

I don't fully understand this, but I believe you understand what you're doing :)

https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
Source/wtf/PartitionAlloc.cpp:913: // If it's the current active page, attempt
to change it. We'd prefer t

t => to

https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
Source/wtf/PartitionAlloc.cpp:958: bool
partitionReallocDirectMappedInPlace(PartitionRootGeneric* root, PartitionPage*
page, size_t origSize)

origSize => rawSize

[Chris Evans, 2015-06-19 21:13:03]

On 2015/06/19 12:48:17, haraken wrote:
> LGTM
> 
>
https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
> File Source/wtf/PartitionAlloc.cpp (right):
> 
>
https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
> Source/wtf/PartitionAlloc.cpp:622: static ALWAYS_INLINE void*
> partitionDirectMap(PartitionRootBase* root, int flags, size_t origSize)
> 
> origSize => rawSize

Done.

> 
>
https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
> Source/wtf/PartitionAlloc.cpp:911: page->freelistHead = nullptr;
> 
> I don't fully understand this, but I believe you understand what you're doing
:)

It does deserve a bit of a comment, now that you mention it. Added.
I'll be reworking page state transition logic in a future CL, which should
remove this temporary tweak and also resolve the existing TODO a bit lower down.

> 
>
https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
> Source/wtf/PartitionAlloc.cpp:913: // If it's the current active page, attempt
> to change it. We'd prefer t
> 
> t => to

Done.

> 
>
https://codereview.chromium.org/1195543005/diff/20001/Source/wtf/PartitionAll...
> Source/wtf/PartitionAlloc.cpp:958: bool
> partitionReallocDirectMappedInPlace(PartitionRootGeneric* root, PartitionPage*
> page, size_t origSize)
> 
> origSize => rawSize

Done.

[Chris Evans, 2015-06-19 21:13:10]

The CQ bit was checked by cevans@chromium.org

[Chris Evans, 2015-06-19 21:13:11]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/1195543005/#ps40001
(title: "Review feedback.")

[commit-bot: I haz the power, 2015-06-19 21:13:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1195543005/40001

[commit-bot: I haz the power, 2015-06-19 22:59:59]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-06-19 23:00:01]

Try jobs failed on following builders:
  win_blink_rel on tryserver.blink (JOB_FAILED,
http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/67530)

[Chris Evans, 2015-06-19 23:06:31]

The CQ bit was checked by cevans@chromium.org

[commit-bot: I haz the power, 2015-06-19 23:07:02]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1195543005/40001

[haraken, 2015-06-20 00:03:09]

https://codereview.chromium.org/1195543005/diff/40001/Source/wtf/PartitionAll...
File Source/wtf/PartitionAlloc.cpp (right):

https://codereview.chromium.org/1195543005/diff/40001/Source/wtf/PartitionAll...
Source/wtf/PartitionAlloc.cpp:912: // page as having a single unprovisioned
slot.

Thanks, now I understand :)

[commit-bot: I haz the power, 2015-06-20 02:38:00]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197520