PartitionAlloc: Remove the 2 GB per-partition address space limit

PartitionAlloc: Remove the 2 GB per-partition address space limit

Depending on allocation patterns, the allocator sometimes uses a lot more
address space than actual committed memory, so the limit -- which was
enforced by crashing -- would sometimes be reached even though actual
memory usage was significantly lower.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=184081

[Jens Widell, 2014-10-16 12:06:23]

jl@opera.com changed reviewers:
+ cevans@chromium.org, tsepez@chromium.org

[Jens Widell, 2014-10-16 12:06:24]

PTAL

Removing the limit and kMaxPartitionSize constant, and necessarily restructuring
partition shutdown a bit.

[Tom Sepez, 2014-10-16 16:45:33]

lgtm

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cpp
File Source/wtf/PartitionAlloc.cpp (right):

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cp...
Source/wtf/PartitionAlloc.cpp:260: PartitionSuperPageExtentEntry* nextEntry;
nit: This can go inside the while loop, at 262 as
 PartitionSuperPageExtentEntry* nextEntry = entry->next;

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cp...
Source/wtf/PartitionAlloc.cpp:265: while (superPage != superPagesEnd) {
nit: Do you want < instead of != ?here?? At least you'll terminate if
entry->superPagesEnd somehow isn't at a multiple of kSuperPageSize?

[Jens Widell, 2014-10-16 16:49:25]

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cpp
File Source/wtf/PartitionAlloc.cpp (right):

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cp...
Source/wtf/PartitionAlloc.cpp:265: while (superPage != superPagesEnd) {
On 2014/10/16 16:45:33, Tom Sepez wrote:
> nit: Do you want < instead of != ?here?? At least you'll terminate if
> entry->superPagesEnd somehow isn't at a multiple of kSuperPageSize?

Old code did the same, so I guess it's never been aproblem.

[Chris Evans, 2014-10-16 22:16:49]

LGTM

Nice.

[Jens Widell, 2014-10-21 09:51:36]

Thank you both for reviewing!

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cpp
File Source/wtf/PartitionAlloc.cpp (right):

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cp...
Source/wtf/PartitionAlloc.cpp:260: PartitionSuperPageExtentEntry* nextEntry;
On 2014/10/16 16:45:33, Tom Sepez wrote:
> nit: This can go inside the while loop, at 262 as
>  PartitionSuperPageExtentEntry* nextEntry = entry->next;

Done.

https://codereview.chromium.org/658113002/diff/1/Source/wtf/PartitionAlloc.cp...
Source/wtf/PartitionAlloc.cpp:265: while (superPage != superPagesEnd) {
On 2014/10/16 16:45:33, Tom Sepez wrote:
> nit: Do you want < instead of != ?here?? At least you'll terminate if
> entry->superPagesEnd somehow isn't at a multiple of kSuperPageSize?

Done.

[Jens Widell, 2014-10-21 12:51:39]

The CQ bit was checked by jl@opera.com

[commit-bot: I haz the power, 2014-10-21 12:52:02]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/658113002/20001

[commit-bot: I haz the power, 2014-10-21 12:55:49]

Committed patchset #2 (id:20001) as 184081