618533 - forward oom notifications to child content process

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Doug Turner (:dougt)]

when we run low on memory, we want to let the child content process know about it.  We already have a FlushMemory method on nsIMemory which has an associated observer topic.  We should forward that.

Comment 1

[Doug Turner (:dougt)]

Attachment: patch v.1

Comment 2

[Josh Matthews [:jdm]]

Comment on attachment 497036 [details] [diff] [review]
patch v.1

You're missing any actual sending of the message.  Add an observer in ContentParent?

Comment 3

[Doug Turner (:dougt)]

yeah, patch queue fail.  when i get back, i'll figure it out

Comment 4

[Doug Turner (:dougt)]

Attachment: patch v.2

merge == better.

Comment 5

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 497091 [details] [diff] [review]
patch v.2

>diff --git a/dom/ipc/ContentChild.cpp b/dom/ipc/ContentChild.cpp
>--- a/dom/ipc/ContentChild.cpp
>+++ b/dom/ipc/ContentChild.cpp
>@@ -581,10 +581,20 @@ ContentChild::RecvScreenSizeChanged(cons
> #ifdef ANDROID
>     mScreenSize = size;
> #else
>     NS_RUNTIMEABORT("Message currently only expected on android");
> #endif
>   return true;
> }
> 
>+bool
>+ContentChild::RecvFlushMemory(const nsString& reason)
>+{
>+    nsCOMPtr<nsIObserverService> os =
>+        mozilla::services::GetObserverService();
>+    if (os)
>+	os->NotifyObservers(nsnull, "memory-pressure", reason.get());

This is a tab. Please fix.

>+  return true;
>+}
>+
> } // namespace dom
> } // namespace mozilla

Comment 6

[Doug Turner (:dougt)]

oh, i missed it when i pushed

http://hg.mozilla.org/mozilla-central/rev/888d64457836

but a new ws-removal patch is in my queue.  I will push it next time I push something else.

Comment 7

[Doug Turner (:dougt)]

ws - http://hg.mozilla.org/mozilla-central/rev/d999c9baa0b7

Comment 8

[Doug Turner (:dougt)]

err.. wrong link.

the ws change was http://hg.mozilla.org/mozilla-central/rev/d4456ada0374