1693445 - GLX: UI not responding after wake up from suspend or sleep

Status: RESOLVED DUPLICATE of bug 1710400

(Core :: Graphics: WebRender, defect, P3)

Description

[xenonelive]

Attachment: FF-85-TSI

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:85.0) Gecko/20100101 Firefox/85.0

Steps to reproduce:

• Start and run FF as usual for few moment.
• Suspend the Ubuntu OS or leave it to go to sleep
• After a few moment, resume the OS and switch to already opened FF window

Actual results:

FF user interface isn't responding to mouse clicks or keyboard shortcuts anymore. Even if it's responding that's taking few seconds to get a mouse click to work. ex: opening a new tab, closing a tab, etc

Expected results:

FF should be usable after wake up or suspending of the OS

Comment 1

[Marcela (Please NI request to Brindusa Tot)]

Hi xenonelive!
I was unable to reproduce this issue on latest Nightly version 87.0a1 (2021-02-22)(64-bit) on Ubuntu 20.04.
Please could you try to reproduce it on Nightly and tell us your input? You can download it from here: https://nightly.mozilla.org/
I will set a component to have a starting point from this.

Comment 2

[Marcela (Please NI request to Brindusa Tot)]

Hi,
I'm marking this as Resolved-Incomplete due to lack of response. If the issue is still reproducible with the latest Firefox version, feel free to reopen the bug with more information.
Thanks,

Comment 3

[xenonelive]

I can confirm, I installed Mozilla Firefox 89.0.2 Still the above issue is there.
Also tried Nightly build Mozilla Firefox 90.0a1 it also has the same issue @Marcela
Maybe this due to the graphics hardware?

Comment 4

[xenonelive]

Another important update on this matter. I've resorted to FF 84.0.2 and it doesn't have this issue.

Comment 5

[xenonelive]

Another development, I tried 85, 87, 89 and all got this issue. HOwever 84.0.2 works fine

Comment 6

[Matthew Gaudet (he/him) [:mgaudet]]

Might take a bit of time, but if you could use mozregression to narrow it to a specific change that would really help this become actionable: https://mozilla.github.io/mozregression/

Comment 7

[xenonelive]

(In reply to Matthew Gaudet (he/him) [:mgaudet] from comment #6)

Might take a bit of time, but if you could use mozregression to narrow it to a specific change that would really help this become actionable: https://mozilla.github.io/mozregression/

I've done a bisection on working and bad versions. However this is the first time I used this tool (mozregression) and I don't know how to collect the report. However as a blind guess I paste the https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=d9b0105477475ee9d411fdd67d37b1774aedb3da&tochange=c22c3f6c8eadfaf2cf1678f1d78d4f2de1093e74

Let me know if I can perform any other helpful things on this.

Comment 8

[xenonelive]

(In reply to Matthew Gaudet (he/him) [:mgaudet] from comment #6)

Might take a bit of time, but if you could use mozregression to narrow it to a specific change that would really help this become actionable: https://mozilla.github.io/mozregression/

Here, I've shorten the distance between good and bad versions
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=d9b0105477475ee9d411fdd67d37b1774aedb3da&tochange=921737066346db5708ea27b28a7ec2ef01cc4286

Comment 9

[xenonelive]

(In reply to Matthew Gaudet (he/him) [:mgaudet] from comment #6)

Might take a bit of time, but if you could use mozregression to narrow it to a specific change that would really help this become actionable: https://mozilla.github.io/mozregression/

Here I've compared the FIrefox 84 with the broken changeset, then I get this

https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=efceae9186cd456bb70240aa851855efb5afd316&tochange=0d82c3f140ba2110d3d40512c9f04b4cadcb4883

Comment 10

[xenonelive]

@Matthew Gaudet (he/him) [:mgaudet (he/him)
Just checking guys, can I expect something on this matter? I'm ready to perform any tests needed to fix this issue. Worst part is to stuck at version 84 with less security and less cool look.

Comment 11

[Marcela (Please NI request to Brindusa Tot)]

I'll set a product and a component to have more visibility.
Also, could be related to this issue? : 1712665

Comment 12

[xenonelive]

1712665
It seems almost similar, however this particular problem happens even from version FF 85.

Comment 13

[xenonelive]

How could I track the progress of this?

Comment 15

[Robert Mader [:rmader]]

This could be a duplicate of bug 1710400 - Xenonelive, could you post the output of xrandr --listproviders here?

Comment 16

[xenonelive]

Robert Mader, Here it goes...

$ xrandr --listproviders
Providers: number : 1
Provider 0: id: 0x4c cap: 0x9, Source Output, Sink Offload crtcs: 4 outputs: 9 associated providers: 0 name:Intel

Comment 17

[Robert Mader [:rmader]]

Thanks!

name:Intel

Most likely a duplicate of bug 1710400 then. This driver appears to be simply broken, please try changing to the modesetting driver. Future version of Firefox will default to software rendering when detecting the Intel DDX driver.