1404764 - Add IPC check to TabChild::InitRenderingState()

Status: RESOLVED INVALID

(Core :: Graphics: Layers, enhancement, P2)

Comment 1

[Sotaro Ikeda [:sotaro]]

Adding IPC check to TabChild::InitRenderingState() could reduce crash of Bug 1391262.

Comment 2

[Sotaro Ikeda [:sotaro]]

(In reply to Sotaro Ikeda [:sotaro] from comment #1)
> Adding IPC check to TabChild::InitRenderingState() could reduce crash of Bug
> 1391262.

But I do not expect that all crashes of Bug 1391262 is addressed.

Comment 3

[Sotaro Ikeda [:sotaro]]

Attachment: patch - Add IPC check to TabChild::InitRenderingState()

Comment 4

[Sotaro Ikeda [:sotaro]]

Attachment: patch - Add IPC check to TabChild::InitRenderingState()

Comment 5

[Sotaro Ikeda [:sotaro]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=4a5b0ca9666cf38f85a1f7dd19079efcd5567c76

Comment 6

[Sotaro Ikeda [:sotaro]]

Attachment: patch - Add IPC check to TabChild::InitRenderingState()

Comment 7

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 8914152 [details] [diff] [review]
patch - Add IPC check to TabChild::InitRenderingState()

Review of attachment 8914152 [details] [diff] [review]:
-----------------------------------------------------------------

It would be great if you could describe a little more under what conditions we run into this problem. It seems to me like if we get to this point in the code then the IPC channel should be open. If it's not then maybe it's a bug elsewhere in the code?

::: dom/ipc/TabChild.cpp
@@ +2798,1 @@
>        NS_WARNING("failed to get CompositorBridgeChild instance");

Can we just set mLayersConnected to false here, instead of passing it in as parameter and only copying that to mLayersConnected later?

Comment 8

[Sotaro Ikeda [:sotaro]]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #7)
> Comment on attachment 8914152 [details] [diff] [review]
> patch - Add IPC check to TabChild::InitRenderingState()
> 
> Review of attachment 8914152 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> It would be great if you could describe a little more under what conditions
> we run into this problem. It seems to me like if we get to this point in the
> code then the IPC channel should be open. If it's not then maybe it's a bug
> elsewhere in the code?

From Bug 1391262, there could be a case that IPC between chrome process works, but IPC between GPU process does not work. attachment 8914152 [details] [diff] [review] was created as preparation for Bug 1391262. But it seems better to try to address whole problem.

Comment 9

[Sotaro Ikeda [:sotaro]]

bug 1391262 replaces this bug.