chrisha@ mentioned that we might not have prioritization of the foreground main frame process today?  gab@, do you know if there's anything making that tab load faster than background tabs today?

oysteine@: gab@ mentioned that you're working on lowering process priority of renderers for solely invisible widgets.  Is there a bug for that?  I'm wondering if this can be marked as a duplicate, assuming that your change will cover the case of invisible OOPIF widgets as well as invisible main frame widgets.

(That way, visible OOPIFs will stay at normal priority, along with visible main frames.)

If that work will be ready soon, we could try to schedule our trial in issue 780133 to include it, but otherwise we can proceed without it.

#3: The bug is https://bugs.chromium.org/p/chromium/issues/detail?id=560446 and the CL is https://chromium-review.googlesource.com/c/chromium/src/+/754299

It's not a complete fix (for now) and there's edge-cases where it'll fail (process-sharing will mean some background frames will still be set with foreground priority).

It *should* help with the session restore OOPIF case though, I think? As the CL will cause more renderer processes to get background priority set.

Thanks!  At a high level it should help with OOPIFs, though I have some questions about the pending_views_ part.  I'll post a comment to the CL in a minute.  Glad to hear we're making progress on it, though!

<adding more page load metrics folks to CC>

The plan from #c3 to #c5 SGTM for addressing the SessionRestore.*Foreground*TabFirstLoaded regression, but I wonder if in the long-term we might also want to apply different process priority to 1) foreground main frame VS 2) foreground OOPIFs.  It seems to me that such differentiation might help with metrics like PageLoad.Experimental.PaintTiming.NavigationToFirstMeaningfulPaint (which AFAIU track only the painting/layout of the *main frame*).  OTOH, I understand that NavigationToFirstMeaningfulPaint considerations maybe be postponed and/or treated as a separate bug.

QUESTION: do we want to open a NavigationToFirstMeaningfulPaint-focused bug now or do we want to wait until OOPIFs trials gather enough data to see a statistically significant impact on NavigationToFirstMeaningfulPaint and similar metrics?