Perfectly normal (but old) keyboard dysfunctional on a VT (Breezy)

[kko]

This bug is related to the keyboard-setup in linux, filing under "hotkey-setup" (due to reasons explained later), unsure in what way this is related to "console-data".

EDIT: I thought I filed this under "hotkey-setup" already at first, but apparently this was not registered. I have now changed the report to refer to the binary package "hotkey-setup".

In a virtual terminal, when pressing (e.g.) LeftShift-PageUp (or LeftShift-PageDown, LeftShift + any arrow key) to scroll up, I get an error message on the following lines:
"atkbd.c: Unknown key pressed (translated set 2, code 0xaa on isa0060/serio0).
atkbd.c: Use 'setkeycodes e02a <keycode>' to make it known."

When I set NumLock on, this particular issue disappears, but PgUp, PgDn, all arrow keys on their own generate identical error messages.

I have a trusty old Finnish 102-key keyboard (not even a Win-key) that has worked for me on every system I've used (DOS, Win 3.11, Slackware...), except, sadly, on Kubuntu.

Now, the bug _seems to be_ hotkey-setup related. A search on ubuntuforums.org with the terms "setkeycodes e02a" throws up a profusion of topics (most of which just contain dmesg output pasted to solve other, unrelated problems), in which several solutions are suggested. Among the solutions suggested are re-assigning the keycode e02a in /usr/share/hotkey-setup/generic.hk, or removing the keycode from there, or simply stopping hotkey-setup.

EDITed: Stopping hotkey-setup seems to work for me (at least I have yet to found any negative effects due to that). However, as keyboard functionality is such a basic issue, I do hope that this issue can be addressed for people to be able to use their regular, "no extra-keys" -keyboards.

EDIT: I have recently noticed, after filing this report, that Ubuntu-bugzilla 17569 claims that this issue is fixed (in a version of hotkey-setup not released for Breezy, so I am not able to test, thus I think filing this report was necessary).

Besides, it is unexpected that Shift-UpArrow and Shift-DownArrow do _not_ scroll the display buffer one line at a time. Is this an expected behaviour in Ubuntu, or should I file a separate report on that (against "console-data"), or will it best be handled in conjunction with this report?

Now, is it possible that this could in some way be fixed (in Breezy), if I were to supply you with any information that you need?

[kko (kko)]

In case it's of any use, as I noticed that "/etc/init.d/hotkey-setup" uses dmidecode.

[kko (kko)]

Marking as confirmed (for Breezy) based on the ubuntu.com bugzilla 17569 plus several ubuntuforums.org reports. I will quote those from this year (i.e. after the bugzilla bug report was marked as fixed): http://www.ubuntuforums.org/showthread.php?t=119449&highlight=setkeycodes+e02a+dmesg
http://www.ubuntuforums.org/showthread.php?t=127458&highlight=setkeycodes+e02a+dmesg
http://ubuntuforums.org/showthread.php?t=119499

[kko (kko)]

Reverting, will let the developers mark this as "confirmed".

[Felix Miata (mrmazda)]

Why is this new bug not marked as a duplicate of the much older Bug #27458?

[kko (kko)]

I did the marking, so I will provide my reasoning. Basically I considered the issue like this:
- These two reports apparently address the same issue.
- However, the title of the older report was not quite as descriptive as to the cause of the problem.
- It is apparently possible to edit another person's bug report (even its title) to be more descriptive, but I did not feel comfortable touching someone else's report.
- Further, I do not believe that a bug report's age is a definite criterion for which of them should be marked as a duplicate. The previous bug report hadn't received attention in a long time, so I felt that the issue is more likely to receive positive attention if tracked under this newer report (with the more descriptive title).

- A further note: As I noted in my edit to the original bug report (the edit appears as the current text), I found out that this is (apparently) fixed in Dapper, but the fix has not been published for Breezy.
- It is unclear to me, what the Ubuntu policy regarding these kinds of fixes (backports) is, so if a developer considers this report (for Breezy), I do appreciate if someone is able to clarify the policy.

[Felix Miata (mrmazda)]

I don't know the official policy here, but I've been using Mozilla's bugzilla, and doing QA there, for more than 5 years. Marking an older bug a duplicate of a newer bug is what lazy programmers do. They don't care about the QA people. Unless a programmer files the newer bug and attaches a patch to it, or the earlier bug is a clearly inferior report to the later report, the newer bug is supposed to be duped to the earlier bug. To do otherwise:

1-Makes open bugs look like they're newer than they are (date originally filed, real age, is not normally included in query lists)
2-provides incentive for people to be lazy and not first look for existing bugs, since they're no less likely to be fixed than older bugs
3-provides disincentive for people to try to file only good bug reports that are not duplicates, since there's reduced likelihood they'll get looked at before the latest filed bugs
4-provides incentive for QA people to quit doing QA due to all the extra work resulting from 2 & 3 above

That someone finds only the summary deficient in an existing bug, is no justification to file a new bug. Instead, the existing bug should have its summary adjusted. If you wish to alienate the few people interested in doing quality QA, by all means continue to permit old bugs to be duped to new bugs.

From the description of this bug, I would not likely have found it in a search prior to filing bug 27458. The age of a keyboard has absolutely nothing to do with vttys being unusable.

[kko (kko)]

Felix: I notice now that you are the person who filed bug 27458. I sincerely apologise if I have done you wrong. Also, I am not a programmer, neither lazy nor productive. :-)

I did a search for this issue prior to filing the bug, and checked about a dozen bug reports that seemed to possibly describe this issue. However, I failed to find your report then - "I was not likely to find it", I'd say.

Now, only after filing the bug did I find your report, at which point I had to decide what to do. As I wrote, I did not feel comfortable editing someone else's text then, and considered my report to be more descriptive, as I had identified the package that caused the bug.

You write that "The age of a keyboard has nothing to do with vttys being unusable". True. However, I don't think the VT's _are_ unusable. (I admit, they are hard to use. You can still press Ctrl-L to redraw the screen.) In parallel to your claim, I could write that "claiming 'tty[1-6] unusable' has nothing to do with 'unknown key' error messages thrown to the screen" - I hope you see what I mean. :-)

Then again, actually both of our reports are duplicates. They are duplicates of the two old ubuntu.com bugzilla reports that I have referenced here. And those have even been marked as fixed (well, one as fixed and the other as duplicate) - the fix just isn't available in Breezy.

Now, in the end: I appreciate your reasoning that older bug reports should be valued to respect the bug's real age. However, a "good bugreport" also has other qualities than "not being a duplicate".

If you consider it worth your while, it is still possible to incorporate the information from this report to the older report. In order to do so, the description (and title) of that report need to be edited to reflect what is known about the bug. In addition, that report has to be marked with the appropriate package ("hotkey-setup"), so it gets identified better.

Regardless of how this issue is handled in the Launchpad, I hope that it does get addressed.

[kko (kko)]

Reporting this on "hotkey-setup" doesn't seem to work, trying again.

[Matthew Garrett (mjg59)]

This is fixed in dapper. A fix will not be released for breezy.

[kko (kko)]

Thank you.

For further reference, are you able to point me to a ubuntu.com page that would outline, what exactly does "support during a release's lifetime" include?

I would be grateful in knowing what policies are followed, more specifically:
- Are only security-related updates known as and included in "support"?
- If not, what types of bugfixes may a user expect without upgrading to a newer version, and how major issues will be addressed?
- And further, is there an official policy for what is called "backports"?

If such a policy does not (yet) exist, I would expect one at least to be created published for the Ubuntu 6.06 Long Term Support release, i.e. Dapper Drake.

I will be grateful for clarification in this matter.

[Matthew Garrett (mjg59)]

In general, only security-related issues are released as updates for released versions. https://wiki.ubuntu.com/UbuntuBackports describes the current state of backports.

[Memo (memo-cj-ro)]

The bug came somehow to life again ...

I have this issue when using 3 keys on my keyboard: Right SHIFT, Left SHIFT and Left ALT. I have a generic 101 key PC keyboard.

Pressing and releasing Left Shift:
==================================
Jun 29 15:45:36 memo-desktop kernel: [103435.145068] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:45:36 memo-desktop kernel: [103435.145075] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.
Jun 29 15:45:36 memo-desktop kernel: [103435.150265] atkbd.c: Unknown key pressed (translated set 2, code 0x92 on isa0060/serio0).
Jun 29 15:45:36 memo-desktop kernel: [103435.150271] atkbd.c: Use 'setkeycodes e012 <keycode>' to make it known.
Jun 29 15:45:36 memo-desktop kernel: [103435.574586] atkbd.c: Unknown key released (translated set 2, code 0x92 on isa0060/serio0).
Jun 29 15:45:36 memo-desktop kernel: [103435.574593] atkbd.c: Use 'setkeycodes e012 <keycode>' to make it known.
Jun 29 15:45:36 memo-desktop kernel: [103435.578203] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:45:36 memo-desktop kernel: [103435.578209] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.

Pressing and releasing Right Shift:
===================================
Jun 29 15:48:04 memo-desktop kernel: [103582.948920] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:48:04 memo-desktop kernel: [103582.948927] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.
Jun 29 15:48:04 memo-desktop kernel: [103583.188423] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:48:04 memo-desktop kernel: [103583.188429] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.

Pressing and releasing Left ALT:
================================
Jun 29 15:49:35 memo-desktop kernel: [103673.665006] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:49:35 memo-desktop kernel: [103673.665013] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.
Jun 29 15:49:35 memo-desktop kernel: [103674.099733] atkbd.c: Unknown key released (translated set 2, code 0x62 on isa0060/serio0).
Jun 29 15:49:35 memo-desktop kernel: [103674.099739] atkbd.c: Use 'setkeycodes 62 <keycode>' to make it known.