[Webkit-unassigned] [Bug 188299] Web process never leaves memory pressured state if caused by per-process limit

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Fri Aug 3 08:26:41 PDT 2018


--- Comment #4 from Simon Fraser (smfr) <simon.fraser at apple.com> ---
Comment on attachment 346470
  --> https://bugs.webkit.org/attachment.cgi?id=346470

View in context: https://bugs.webkit.org/attachment.cgi?id=346470&action=review

> Source/WTF/wtf/cocoa/MemoryPressureHandlerCocoa.mm:90
> +            // Per process notification are one shot things while the system status is a state.

I think it's confusing to talk about "per process" notifications here, since the reader will probably think that refers to the "DISPATCH_MEMORYPRESSURE_PROC_LIMIT_" notifications, when really the DISPATCH_MEMORYPRESSURE_ are per-process for processes with an explicit jetsam limit (as I understand it).

> Source/WTF/wtf/cocoa/MemoryPressureHandlerCocoa.mm:96
> +                return MemoryPressureHandler::singleton().isUnderMemoryPressure();

Is this line only about simulated memory pressure? If so, would be nice to rename that function since it's really confusing that the input to setUnderMemoryPressure() uses isUnderMemoryPressure().

> Source/WTF/wtf/cocoa/MemoryPressureHandlerCocoa.mm:109
> +            MemoryPressureHandler::singleton().setUnderMemoryPressure(isUnderVMMemoryPressure);

I'm still not sure how you get out of memory pressure with this code, since isUnderVMMemoryPressure doesn't consult DISPATCH_MEMORYPRESSURE_PROC_LIMIT_CRITICAL.

> Source/WTF/wtf/cocoa/MemoryPressureHandlerCocoa.mm:118
> -                NSLog(@"Got memory pressure notification (%s)", critical ? "critical" : "non-critical");
> +                NSLog(@"Got memory pressure notification %lu (%s)", status, isCritical ? "critical" : "non-critical");

These NSLogs should be WTFLogAlways, so they show in iOS syslog. Are they even useful with all the release logging we have? Should they be release logs?

You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180803/e1444b3f/attachment.html>

More information about the webkit-unassigned mailing list