On systems where hw.pagesize != vm.pagesize, we get goofy looking memory numbers.
Created attachment 266726 [details] Patch
Comment on attachment 266726 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=266726&action=review > Source/WebCore/page/cocoa/ResourceUsageOverlayCocoa.mm:79 > + sysctlbyname("vm.pagesize", &pageSize, &outputSize, nullptr, 0); Could you ASSERT that outputSize still equals sizeof(pageSize) after this call? You might also want to either handle an error or ASSERT that none occurred. I'm imagining a case where we could end up diving by zero later on.
Created attachment 266892 [details] Patch for landing
Comment on attachment 266892 [details] Patch for landing Clearing flags on attachment: 266892 Committed r193757: <http://trac.webkit.org/changeset/193757>
All reviewed patches have been landed. Closing bug.