[webkit-reviews] review granted: [Bug 112387] NetworkProcess should send vm_copied, mmap'ed memory to WebProcesses when a resource is already in the disk cache : [Attachment 193197] Patch v1

[bugzilla-daemon at webkit.org]

Geoffrey Garen <ggaren at apple.com> has granted Brady Eidson
<beidson at apple.com>'s request for review:
Bug 112387: NetworkProcess should send vm_copied, mmap'ed memory to
WebProcesses when a resource is already in the disk cache
https://bugs.webkit.org/show_bug.cgi?id=112387

Attachment 193197: Patch v1
https://bugs.webkit.org/attachment.cgi?id=193197&action=review

------- Additional Comments from Geoffrey Garen <ggaren at apple.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=193197&action=review


r=me

> Source/WebKit2/NetworkProcess/NetworkResourceLoader.h:87
> +    static size_t fileBackedResourceThreshold();

Might be nice to call this "minimum" or "minimumSize" instead of "threshold".

> Source/WebKit2/NetworkProcess/mac/NetworkResourceLoaderMac.mm:65
> +    // We only care about the vm_copy optimization for resources that should
be file backed.

This should really be a FIXME about using an API instead of a guess in future.

> Source/WebKit2/WebProcess/mac/WebProcessMac.mm:-119
> -	   [nsurlCache setDiskCapacity:0];

This is no longer necessary / correct. The WebProcess no longer shares the disk
cache with the network process, so it is appropriate for the WebProcess to have
a zero-sized disk cache, to save SQL memory.