Authentication Required When Resuming From Sleep

I used my laptop for a long spell without suspending to RAM. For whatever reasons I had the laptop connected to the wall during this period. I was powering down every night. Just a habit quirk — in my own little proverbial rut.

Then one day I pressed Fn+F4 to invoke suspend to RAM. Upon returning I was prompted with an authentication dialog to enter a password.

This was with Fedora 22 with the MATE desktop. I did not recall needing to authenticate with Fedora 21.

As I have other distros installed on this laptop, I rebooted into CentOS 7. Same authentication dialog required. I do not use CentOS as much as Fedora and do not recall whether previously authentication was required.

In both systems I use the lightdm login manager and MATE desktop.

I do not have the lock screen option enabled when using the screen saver.

I found no obvious related GUI control in the MATE control center.

To add to the confusion, I logged out to the login manager and invoked sleep. I closed the lid and pressed Fn+F4. Nothing happened. Logging in and once the desktop startup settled the system suspended to RAM with no action on my part. Apparently my invocation of sleep with the login manager carried over into the desktop session.

Although an apples-to-oranges comparison, rebooting into LMDE resulted in the expected behavior. No password required on resume and suspending worked correctly with the MDM login manager. Nominally this would seem to eliminate MATE as the culprit.

I restored a Fedora 21 backup to a testing partition. My memory had not gone sour. I did not have to authenticate to resume my desktop.

I installed the Fedora 22 MATE spin. There was no requirement to type a password. After I fully updated with dnf I had to type a password when resuming from suspend.

What caused the change? Systemd? Polkit? Pam.d? GConf? I use the laptop for testing — did I inadvertently do something?

I performed a dnf history undo of the updates. The problem disappeared. Slowly I again updated packages. After a few hours I isolated the cause to the mate-screensaver package.

There seem to be two related dconf options, lock-suspend and lock-use-screensaver. Playing with those options makes no difference.

While the locking feature is a sane idea in an untrusted environment, whether that be a cube farm or roaming cats, the lack of GUI controls to change the behavior is a usability frustration.

After addressing the issue on the Fedora forum and discussing with the Fedora MATE maintainer, the good news was the patch was reverted until the issues could be addressed.

The patch was from the upstream MATE developers and not Fedora or CentOS. This patch looked good “on paper” but proved otherwise in the real world. These are the kinds of changes that annoy technical users, let alone non technical users. Thankfully the developers and maintainers involved responded in a positive manner rather than push agendas.

Posted: Category: Usability Tagged: General, MATE

Next: What Changed This Time

Previous: Living Under a Rock