Bug 222392 – Hold assertion while waiting for WebContent process to finish initializing

• Status: NEW
• Component: WebKit Misc.

[Per Arne Vollan]

Make sure the WebContent process is not suspended during initialization by holding assertion while waiting for WebContent process to finish initializing.

[Per Arne Vollan]

Created attachment 421480 [details]
Patch

[Per Arne Vollan]

Created attachment 421485 [details]
Patch

[Chris Dumez]

Comment on attachment 421485 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421485&action=review

> Source/WebKit/UIProcess/WebProcessPool.cpp:849
> +    process.sendWithAsyncReply(Messages::WebProcess::InitializeWebProcess(parameters), [initializationActivity = WTFMove(initializationActivity)] {

So the thing is. I tried this exact change in the past and had to revert it due to regression. I suggest you take a look at:
Unreviewed, revert r255662 as it seems to cause blank views in SVC.
https://bugs.webkit.org/show_bug.cgi?id=207144
<rdar://problem/59334011>

[Chris Dumez]

Comment on attachment 421485 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421485&action=review

> Source/WebKit/ChangeLog:3
> +        Hold assertion while waiting for WebContent process to finish initializing

The title is misleading since we are already going that. You are merely doing it in a different way.

[Per Arne Vollan]

(In reply to Chris Dumez from comment #3)
> Comment on attachment 421485 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=421485&action=review
> 
> > Source/WebKit/UIProcess/WebProcessPool.cpp:849
> > +    process.sendWithAsyncReply(Messages::WebProcess::InitializeWebProcess(parameters), [initializationActivity = WTFMove(initializationActivity)] {
> 
> So the thing is. I tried this exact change in the past and had to revert it
> due to regression. I suggest you take a look at:
> Unreviewed, revert r255662 as it seems to cause blank views in SVC.
> https://bugs.webkit.org/show_bug.cgi?id=207144
> <rdar://problem/59334011>

Ah, that is interesting!

Do you think it is possible that the process could suspend before the assertion is taken?

Thanks for reviewing!

[Chris Dumez]

(In reply to Per Arne Vollan from comment #5)
> (In reply to Chris Dumez from comment #3)
> > Comment on attachment 421485 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=421485&action=review
> > 
> > > Source/WebKit/UIProcess/WebProcessPool.cpp:849
> > > +    process.sendWithAsyncReply(Messages::WebProcess::InitializeWebProcess(parameters), [initializationActivity = WTFMove(initializationActivity)] {
> > 
> > So the thing is. I tried this exact change in the past and had to revert it
> > due to regression. I suggest you take a look at:
> > Unreviewed, revert r255662 as it seems to cause blank views in SVC.
> > https://bugs.webkit.org/show_bug.cgi?id=207144
> > <rdar://problem/59334011>
> 
> Ah, that is interesting!
> 
> Do you think it is possible that the process could suspend before the
> assertion is taken?

No idea. RunningBoard logging would tell you though.

[Radar WebKit Bug Importer]

<rdar://problem/75008899>