Bug 346938 - Lockscreen breaks unlock with Nvidia 349.16
Summary: Lockscreen breaks unlock with Nvidia 349.16
Status: RESOLVED UPSTREAM
Alias: None
Product: ksmserver
Classification: Plasma
Component: lockscreen (show other bugs)
Version: 5.2.1
Platform: Arch Linux Linux
: NOR grave
Target Milestone: ---
Assignee: David Edmundson
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-04-30 07:27 UTC by Ben
Modified: 2015-05-26 06:19 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
journalctl contents between login and reboot after unable to unlock (284.14 KB, text/plain)
2015-04-30 07:32 UTC, Ben
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Ben 2015-04-30 07:27:48 UTC
As the title says. If I roll back to 346.59 it works fine. There are others who have had the same results both after updating to the newer nvidia driver and it returning to normal after rollback. Attached is my journalctl. Hopefully I put the report in the right place.

Reproducible: Always

Steps to Reproduce:
1. Upgrade nvidia and associated packages ((lib32)libgl, (lib32)utils) from 346.59 to 349.16
2. Lock screen
3. Unable to log back in

Actual Results:  
Cannot unlock

Expected Results:  
Screen is unlocked
Comment 1 Ben 2015-04-30 07:32:09 UTC
Created attachment 92335 [details]
journalctl contents between login and reboot after unable to unlock
Comment 2 Martin Flöser 2015-05-19 07:56:19 UTC
As unfortunate as it is: there's not much we can do about it if it's related to NVIDIA's blob driver. It's closed to us, we cannot investigate. I'm sorry.
Comment 3 Ben 2015-05-21 03:28:38 UTC
Martin, can we even be sure without further investigation? Binary blob or not, this crash problem is a major problem and unless it actually cannot be fixed due to something broken in the binary driver, it should be looked into further.
Comment 4 Martin Flöser 2015-05-21 06:02:57 UTC
Please tell me how we should look into it further. I don't have an NVIDIA and even if I had I would run the FLOSS drivers. We have no chance to investigate issues caused by the non-free drivers.
Comment 5 Ben 2015-05-22 00:40:26 UTC
It's pretty childish to say "tell me how" when you damn well know how. The latter problem of not having a card is of course problematic. I'll send you my card if that would change things, but I'm pretty sure you're mind is made up no matter the circumstances. Which is pretty unfortunate. Typical Linux crap of "doesn't fit my needs, so it doesn't matter" kind of crap.
Comment 6 Martin Flöser 2015-05-26 05:56:32 UTC
I'm sorry that we cannot meet your expectations. Unfortunately we are not in a position to get the driver code from NVIDIA to investigate such issues. In that regard this Linux crap just cannot do what you expect.
Comment 7 Ben 2015-05-26 06:05:24 UTC
That was not in reference to Linux being crap. It was in reference to the typical attitude that you've exhibited and has been exhibited in the Linux dev community at large.

It's not my expectations that everything can be solved by you or someone specific. My expectation is that investigation happens that can determine whether or not it is an issue with the binary driver (possible) or, possibly, some KDE component (equally possible). Unfortunately, you're wholly unwilling to investigate. I am offering up my own card for development and I don't want it back if that is what is it would take. I also know you could purchase a card for much cheaper than I can send mine to you, and as such, one should be just purchased for testing. It doesn't need to be a brand new just released card. This problem is exhibited on my meager 560TI and at least one other that is a 970. I am sure it's happening with others too.

But alas, we're still at the "doesn't fit my needs, so it doesn't matter" shit.
Comment 8 Martin Flöser 2015-05-26 06:16:15 UTC
I did all the investigation I can do with the information you provided. Which is a useless log (doesn't show any problem) and the information that it is NVIDIA driver related. My investigation showed that it is a problem caused by the driver. This investigation is of course based on my experience of OpenGL and driver related issues.

Of course you can provide more information to allow us to investigate better. E.g. it's unclear to me what you mean by "unable to log in". That could mean anything from UI not responding to password is incorrectly assumed as false.

Unfortunately after having been accused of laziness I do not want to help you any more with the issue.

I just give you one hint: look at bug 346525 whether you have the same issue (also Archlinux).
Comment 9 Ben 2015-05-26 06:18:16 UTC
Already looked at that and it is not the issue.
Comment 10 Ben 2015-05-26 06:19:31 UTC
Also, for the record, I do not see you as lazy. That was not what I said, nor what I implied.