Bug #17235

User input is prone to hangs since Tails 4.0

Added by kdr4 2019-11-15 09:17:21 . Updated 2019-11-22 09:10:23 .

Status:
New
Priority:
Normal
Assignee:
Category:
Hardware support
Target version:
Start date:
Due date:
% Done:

0%

Feature Branch:
Type of work:
Test
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description

Since I have upgraded to Tails 4.0 from Tails 3.16, I have noticed a new behaviour where user input by keyboard, mouse or trackpoint (this is a Lenovo device) will quite often result in no visible change on-screen, e.g. the mouse pointer will not move or text will not be entered.

NB - The behaviour I am about to describe is present on every normal boot, though it is sporadic, and I have not been able to identify a way of reproducing it. I have not tested Troubleshooting mode, sorry. I have encountered these problems on a Tails install updated from 3.16 and also a fresh clean install.

Keyboard: one striking example is when I used Ctrl-W to close a tab in Thunderbird. The tab did not close, so I pressed Ctrl-W again. Suddenly Thunderbird closed completely, which I thought was a crash, but testing showed that what had happened was that the tab I had attempted to close was one of only two open tabs (a message viewer tab, and the main viewer window tab), and that the tab I had targeted with the keystroke had closed, but the screen had not updated to reflect this. So, when I pressed Ctrl-W again, it targeted the final tab, resulting in Thunderbird quitting.

Keyboard: another example happened while typing this bug report. I paused very briefly in my typing, and when I restarted the text did not immediately appear. I think it did not appear until I typed another character, which caused the screen to update and display the previous and the current character. This behaviour of not refreshing the screen until receiving additional input after the first is common to this issue.

Trackpoint: Response of the cursor to my trackpoint is similar. Interacting with the trackpoint can result in no visible movement of the cursor onscreen, though it will be updated following additional interaction, which means the mouse cursor will suddenly re-position itself to account for the first interaction. This is actually a minor security risk for me because it makes cursor location unreliable and can cause me to initiate clicks in undesirable parts of webpages, for example. This behaviour was noticed at least once during the composing of this bug report.

On at least three occasions, I believe the hangs I’ve described have escalated into full display hangs for up to a minute, where no keyboard, trackpad or trackpoint interaction at all is reproduced on screen, and no automated change is drawn either (e.g. a progress bar will not update to reflect data transfer even though data transfer continues on a hardware level). On these occasions it seems the whole system has hung, but eventually it comes back to life.


Subtasks


History

#1 Updated by kdr4 2019-11-15 09:31:30

I meant to say that the trackpoint behaviour described here is also evident in Windows 10 on the same device, though the other behaviours are not. I installed an official trackpoint patch (released by Lenovo and produced by Synaptics I think) on Windows 10 that I think claimed to address this behaviour, but it didn’t produce any improvement that I could see.

#2 Updated by kdr4 2019-11-16 08:28:07

Update: I attempted to boot into Troubleshooting mode twice, and both times the device failed to reach the Tails Greeter. Tails has never failed to reach the Tails Greeter in Normal boot mode so far, so I have no explanation for this.

There was a long pause after the ‘Started GNOME Display Manager.’ message and every service reported starting ‘[ OK ]’ up to that point too. Then there was 17 messages related to ‘snd_hda_codec’ and other sound device modules, and at that point the boot stalled. I waited at this point for 3 minutes, but the boot didn’t progress any further. I used Ctrl-Alt-Delete to shut down cleanly at this point and this was successful. No other keystrokes produced any effect.

#3 Updated by intrigeri 2019-11-16 10:03:35

  • Category set to Hardware support

kdr4 wrote:
> Update: I attempted to boot into Troubleshooting mode twice, and both times the device failed to reach the Tails Greeter. Tails has never failed to reach the Tails Greeter in Normal boot mode so far, so I have no explanation for this.

Troubleshooting mode is not a “failsafe” mode: it can help in some cases, but in other cases it can break stuff. In your case it breaks stuff, so I would say: forget about it.

Regarding the problems this ticket is about: please send a bug report with WhisperBack next time it happens, so we hopefully have enough technical info to work with. Please reference this ticket ID (Bug #17235) in that bug report. Thanks in advance! (In general, reporting issues directly on Redmine is a valid option only if you’re 200% sure you can provide enough information by yourself to allow us to work on the problem; as we can see here, that’s a risky bet :)

#4 Updated by kdr4 2019-11-22 08:42:27

If I submit a Whisperback report will it automatically pull any relevant logs or is there something I should attach manually?

#5 Updated by intrigeri 2019-11-22 09:09:17

> If I submit a Whisperback report will it automatically pull any relevant logs

Yes, as long as you don’t voluntarily opt out :)

#6 Updated by kdr4 2019-11-22 09:10:23

Just noticed this post in the Tails-ux list, might be a similar issue:

https://lists.autistici.org/message/20190809.115000.f1a6d5e9.en.html