Remove silly messages from input layer.

Remove silly messages from input layer.

[Dave Jones]

There are two messages in the input layer that seem to be
triggerable very easily, and they confuse end-users to no end.
"too many keys pressed? Should I press less keys?"
I actually got a complaint from one user that he had only
hit one key before being told to type less.

The latter message seems to trigger with certain keyboard switchers
and again, does nothing but confuse people.

Best of all, asides from the silly messages, none of the people suffering
them report any other misbehaviour, their keyboards work just fine.

Signed-off-by: Dave Jones <davej@redhat.com>

--- linux-2.6.16.noarch/drivers/input/keyboard/atkbd.c~	2006-05-03 22:37:20.000000000 -0400
+++ linux-2.6.16.noarch/drivers/input/keyboard/atkbd.c	2006-05-03 22:39:58.000000000 -0400
@@ -340,7 +340,6 @@ static irqreturn_t atkbd_interrupt(struc
 			atkbd_report_key(atkbd->dev, regs, KEY_HANJA, 3);
 			goto out;
 		case ATKBD_RET_ERR:
-			printk(KERN_DEBUG "atkbd.c: Keyboard on %s reports too many keys pressed.\n", serio->phys);
 			goto out;
 	}
 
@@ -359,11 +358,7 @@ static irqreturn_t atkbd_interrupt(struc
 		case ATKBD_KEY_NULL:
 			break;
 		case ATKBD_KEY_UNKNOWN:
-			if (data == ATKBD_RET_ACK || data == ATKBD_RET_NAK) {
-				printk(KERN_WARNING "atkbd.c: Spurious %s on %s. Some program, "
-				       "like XFree86, might be trying access hardware directly.\n",
-				       data == ATKBD_RET_ACK ? "ACK" : "NAK", serio->phys);
-			} else {
+			if (data != ATKBD_RET_ACK && data != ATKBD_RET_NAK) {
 				printk(KERN_WARNING "atkbd.c: Unknown key %s "
 				       "(%s set %d, code %#x on %s).\n",
 				       atkbd->release ? "released" : "pressed",

Re: Remove silly messages from input layer.

[Pavel Machek]

Hi!

> There are two messages in the input layer that seem to be
> triggerable very easily, and they confuse end-users to no end.
> "too many keys pressed? Should I press less keys?"
> I actually got a complaint from one user that he had only
> hit one key before being told to type less.

Actually, that message is useful for me. It tells me which keyboards
are crap; I have non-standard keymap (ctrl<->capslock) and some
keyboards can handle it while some other can't :-( (thinkpad x32
complains a bit). And yes, there is
problem hidden behind this message: if see this too often, your
keyboard will drop some keypresses, too (logitech keyboard with
touchpad).
					Pavel
-- 
Thanks, Sharp!

Re: Remove silly messages from input layer.

[Pavel Machek]

On Wed 03-05-06 22:44:04, Dave Jones wrote:
> There are two messages in the input layer that seem to be
> triggerable very easily, and they confuse end-users to no end.
> "too many keys pressed? Should I press less keys?"

It actually means 'type more slowly' or 'use standard keymap' or 'get
a better keyboard' :-) or 'no, you are not imagining it, I've seen
your keypress and dropped it'.
									Pavel

-- 
Thanks, Sharp!

Re: Remove silly messages from input layer.

[Nuri Jawad]

On Wed, 3 May 2006, Dave Jones wrote:

> The latter message seems to trigger with certain keyboard switchers
> and again, does nothing but confuse people.

I rather think it's showing us there was a glitch when switching.

I have a mechanical switch that sometimes produces this message, and every 
now and then, the keyboard loses its key repeat time/rate setting.
It often happens when the switch is not turned quickly or firmly enough 
and in such a case I had the keyboard port lock up completely a few times. 
I think the kernel should report such a situation.

If people are "confused" by valid error messages, they can use certain 
proprietary operating systems that hide the ugly truth from them. What's 
next, removing "access beyond end of device"? I want to stay informed if 
my mechanical switch produces glitches. There are electronic ones that 
don't.

It would be nice if kernel messages had a structure that allowed a 
verbosity setting, but until then, non-confused users want the 
information they can get.

Regards, Nuri

Re: Remove silly messages from input layer.

[Martin J. Bligh]

Pavel Machek wrote:
> On Wed 03-05-06 22:44:04, Dave Jones wrote:
> 
>>There are two messages in the input layer that seem to be
>>triggerable very easily, and they confuse end-users to no end.
>>"too many keys pressed? Should I press less keys?"
> 
> 
> It actually means 'type more slowly' or 'use standard keymap' or 'get
> a better keyboard' :-) or 'no, you are not imagining it, I've seen
> your keypress and dropped it'.

Perhaps it should say that then ;-)

M.

Re: Remove silly messages from input layer.

[David Greaves]

Pavel Machek wrote:
> On Wed 03-05-06 22:44:04, Dave Jones wrote:
>   
>> There are two messages in the input layer that seem to be
>> triggerable very easily, and they confuse end-users to no end.
>> "too many keys pressed? Should I press less keys?"
>>     
>
> It actually means 'type more slowly' or 'use standard keymap' or 'get
> a better keyboard' :-) or 'no, you are not imagining it, I've seen
> your keypress and dropped it'.
> 									Pavel
>
>   
or 'the cat walked on the keyboard again...'

David

-- 

Re: Remove silly messages from input layer.

[Dmitry Torokhov]

On 5/4/06, Martin J. Bligh <mbligh@mbligh.org> wrote:
> Pavel Machek wrote:
> > On Wed 03-05-06 22:44:04, Dave Jones wrote:
> >
> >>There are two messages in the input layer that seem to be
> >>triggerable very easily, and they confuse end-users to no end.
> >>"too many keys pressed? Should I press less keys?"
> >
> >
> > It actually means 'type more slowly' or 'use standard keymap' or 'get
> > a better keyboard' :-) or 'no, you are not imagining it, I've seen
> > your keypress and dropped it'.
>
> Perhaps it should say that then ;-)
>

Do you have a beter wording in mind? "Keyboard reports too many keys
were pessed at once, some keystrokes might be dropped"?

Also I don't understand what people have against this message, it's at
KERN_DEBUG level after all.

--
Dmitry

Re: Remove silly messages from input layer.

[Dave Jones]

On Thu, May 04, 2006 at 02:34:34PM -0400, Dmitry Torokhov wrote:

 > >Perhaps it should say that then ;-)
 > 
 > Do you have a beter wording in mind? "Keyboard reports too many keys
 > were pessed at once, some keystrokes might be dropped"?

It still doesn't make sense when the user only pressed a single key,
or in some cases, never pressed *any* key (don't have that report to hand,
but it was a laptop keyboard)

 > Also I don't understand what people have against this message, it's at
 > KERN_DEBUG level after all.

When you're on the recieving end of distro kernel bug reports, it becomes clearer :)
Users read dmesg from time to time, and freak out when they see something
like this that looks like an error that they can't do anything about.
Until I silenced these in the Fedora kernel I was getting quite a few reports
from concerned users.

		Dave
-- 
http://www.codemonkey.org.uk

Re: Remove silly messages from input layer.

[Hugh Dickins]

On Thu, 4 May 2006, Dmitry Torokhov wrote:
> > > On Wed 03-05-06 22:44:04, Dave Jones wrote:
> > >
> > > >There are two messages in the input layer that seem to be
> > > >triggerable very easily, and they confuse end-users to no end.
> > > >"too many keys pressed? Should I press less keys?"
> 
> Also I don't understand what people have against this message, it's at
> KERN_DEBUG level after all.

I often type at the text console; and I like to see KERN_DEBUG messages;
but I do NOT like to see kernel messages telling me I've mistyped - they
don't hurt my pride, but they do mess up my screen.

If I hit a wrong key, it's normal for that wrong key to be echoed, not a
kernel message saying "Are you sure you didn't want to hit another key?".
If I don't hit a key hard enough, it's normal for that key not to be
echoed.  So why, if I happen to hit a combination of keys by mistake,
do I get that wretched "too many keys pressed" message?  The appropriate
response is for what is accepted to be echoed, and that is all.

I keep #ifdef ATKBD_DEBUG around both the messages Dave highlighted
(but I forget when it was that I used to get the second one); the -mm
tree has had the "too many keys" message commented out for many months.

Hugh

RE: Remove silly messages from input layer.

[Hua Zhong]

> If people are "confused" by valid error messages, they can 
> use certain proprietary operating systems that hide the ugly 
> truth from them. What's next, removing "access beyond end of 
> device"? I want to stay informed if my mechanical switch 
> produces glitches. There are electronic ones that don't.

Why not maintain an error counter? You can then easily identify whether your keyboard is funky by querying the counter.

For others who don't care, life goes on.

Hua

Re: Remove silly messages from input layer.

[Pavel Machek]

On Čt 04-05-06 14:38:40, Dave Jones wrote:
> On Thu, May 04, 2006 at 02:34:34PM -0400, Dmitry Torokhov wrote:
> 
>  > >Perhaps it should say that then ;-)
>  > 
>  > Do you have a beter wording in mind? "Keyboard reports too many keys
>  > were pessed at once, some keystrokes might be dropped"?
> 
> It still doesn't make sense when the user only pressed a single key,
> or in some cases, never pressed *any* key (don't have that report to hand,
> but it was a laptop keyboard)

If you only pressed single key -- your keyboard is crap or there's
some problem in the driver.

If you never pressed any key -- your keyboard is crap or there's
some problem in the driver.

...in both of these cases the message is actually useful. Where it is
not useful is when cat walks over your keyboard, because you can
actually expect this message when pressing 10 keys at once.

...in both of these cases failure should be investigated to find out
what is going on.

OTOH for example I now know that capslock-x-c is combination X32 does
not like. Unfortunately, if you swap capslock and ctrl, and type too
fast... you'll press this combination while exiting emacs.

								Pavel
-- 
Thanks for all the (sleeping) penguins.

Re: Remove silly messages from input layer.

[Sergei Organov]

Pavel Machek <pavel@ucw.cz> writes:
[...]
> OTOH for example I now know that capslock-x-c is combination X32 does
> not like. Unfortunately, if you swap capslock and ctrl, and type too
> fast... you'll press this combination while exiting emacs.

Fortunately nobody ever needs to exit emacs :)

-- 
Sergei.

Re: Remove silly messages from input layer.

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 594 bytes --]

On Fri, 05 May 2006 14:39:27 +0400, Sergei Organov said:
> Pavel Machek <pavel@ucw.cz> writes:
> [...]
> > OTOH for example I now know that capslock-x-c is combination X32 does
> > not like. Unfortunately, if you swap capslock and ctrl, and type too
> > fast... you'll press this combination while exiting emacs.
> 
> Fortunately nobody ever needs to exit emacs :)

M-x save-buffers-shutdown-h-now

I'm not sure which surprised and worried me more - that I saw somebody try that
with the expectation that it would work, or that he was running emacs in a
configuration where it *did* work. ;)



[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: Remove silly messages from input layer.

[Dave Jones]

On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:

 > If you only pressed single key -- your keyboard is crap or there's
 > some problem in the driver.
 > 
 > If you never pressed any key -- your keyboard is crap or there's
 > some problem in the driver.

That's hardly a constructive answer when the keyboard is a part of
a laptop.  Crap hardware exists, get used to it.

		Dave

-- 
http://www.codemonkey.org.uk

Re: Remove silly messages from input layer.

[Martin Mares]

> On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:
> 
>  > If you only pressed single key -- your keyboard is crap or there's
>  > some problem in the driver.
>  > 
>  > If you never pressed any key -- your keyboard is crap or there's
>  > some problem in the driver.
> 
> That's hardly a constructive answer when the keyboard is a part of
> a laptop.  Crap hardware exists, get used to it.

Yes, but removing a message which can be sometimes useful is hardly
justified by crappy hardware sometimes triggering it. If it's triggered
too often, it should be rate-limited, not removed.

				Have a nice fortnight
-- 
Martin `MJ' Mares   <mj@ucw.cz>   http://atrey.karlin.mff.cuni.cz/~mj/
Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth
"It's a lemon tree, my dear Watson." -- Sherlock Holmes (?)

Re: Remove silly messages from input layer.

[Dave Jones]

On Fri, May 05, 2006 at 05:37:28PM +0200, Martin Mares wrote:
 > > On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:
 > > 
 > >  > If you only pressed single key -- your keyboard is crap or there's
 > >  > some problem in the driver.
 > >  > 
 > >  > If you never pressed any key -- your keyboard is crap or there's
 > >  > some problem in the driver.
 > > 
 > > That's hardly a constructive answer when the keyboard is a part of
 > > a laptop.  Crap hardware exists, get used to it.
 > 
 > Yes, but removing a message which can be sometimes useful is hardly
 > justified by crappy hardware sometimes triggering it. If it's triggered
 > too often, it should be rate-limited, not removed.

I'd argue that anything that triggers that many false positives is worthless.

		Dave

-- 
http://www.codemonkey.org.uk

Re: Remove silly messages from input layer.

[Martin Mares]

> I'd argue that anything that triggers that many false positives is worthless.

Why do you think these are false positives? They usually report real
problems. Unfortunately a significant fraction of keyboards is crappy
these days, but it's still good to know if the keyboard you are currently
testing is broken or not.

				Have a nice fortnight
-- 
Martin `MJ' Mares   <mj@ucw.cz>   http://atrey.karlin.mff.cuni.cz/~mj/
Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth
If a train station is where the train stops, what is a work station?

Re: Remove silly messages from input layer.

[Dave Jones]

On Fri, May 05, 2006 at 05:54:52PM +0200, Martin Mares wrote:
 > > I'd argue that anything that triggers that many false positives is worthless.
 >
 > Why do you think these are false positives? They usually report real
 > problems.

Did you read my earlier posts?
Users are seeing this *during boot*, before they've even pressed *ANY* keys.
They're seeing it after pressing a *single* key.

How on earth is "too many keys pressed" a useful message in this context?
Yes, maybe their keyboard is crap, but what is the user to do?
Go buy a new laptop because someone else has a utopian view on how hardware should be?

 > Unfortunately a significant fraction of keyboards is crappy
 > these days, but it's still good to know if the keyboard you are currently
 > testing is broken or not.

When a user can't do *anything* about it, it's useless, and serves
as nothing but a cause for concern. "Oh no, is my laptop dying?".

		Dave

-- 
http://www.codemonkey.org.uk

Re: Remove silly messages from input layer.

[Martin Mares]

Hello!

>  > Why do you think these are false positives? They usually report real
>  > problems.
> 
> Did you read my earlier posts?
> Users are seeing this *during boot*, before they've even pressed *ANY* keys.
> They're seeing it after pressing a *single* key.

If it is so (sorry for catching the thread in the middle), then it means
that there is either a bug in the keyboard driver, which should be fixed,
or a bug in the keyboard controller, which can have other side-effects
and so it should be at least reported.

> How on earth is "too many keys pressed" a useful message in this context?

I am not telling that the message has to stay intact -- it should be reworded
and rate-limited, but not hidden.

> Yes, maybe their keyboard is crap, but what is the user to do?
> Go buy a new laptop because someone else has a utopian view on how hardware should be?

No, just notice that his hardware is buggy and that it's probably a harmless
bug -- like we already do in many other cases (e.g., the 3Com drivers sometime
tell the user that his card has broken DMA, so it's going to use PIO instead).

> When a user can't do *anything* about it, it's useless, and serves
> as nothing but a cause for concern. "Oh no, is my laptop dying?".

If it's not a laptop, they can do.

When I am considering a keyboard to buy and I see this message, I usually
choose a different one immediately, because it often means that the keyboard
will react badly to fast typing.

				Have a nice fortnight
-- 
Martin `MJ' Mares   <mj@ucw.cz>   http://atrey.karlin.mff.cuni.cz/~mj/
Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth
A student who changes the course of history is probably taking an exam.

Re: Remove silly messages from input layer.

[Nuri Jawad]

On Fri, 5 May 2006, Dave Jones wrote:

> Did you read my earlier posts?

Sorry to interrupt, but did you read my reply?

> How on earth is "too many keys pressed" a useful message in this context?

It doesn't say "user pressed too many keys", it says the keyboard *reports 
too many keys pressed* which is most likely what's happening. This is not 
a terribly useful, but valid error message.

> Yes, maybe their keyboard is crap, but what is the user to do?

Nothing, take notice of the fact.

> Go buy a new laptop because someone else has a utopian view on how 
> hardware should be?

You mean deciding not to silently ignore errors is having a utopian view? 
Are we talking about Linux or kernel32.dll?

> When a user can't do *anything* about it, it's useless, and serves
> as nothing but a cause for concern. "Oh no, is my laptop dying?".

Laptops come with Windows XP pre-installed for those users, what was your 
problem again?

Regards, Nuri

Re: Remove silly messages from input layer.

[Martin Bligh]

>> Go buy a new laptop because someone else has a utopian view on how 
>> hardware should be?
> 
> You mean deciding not to silently ignore errors is having a utopian 
> view? Are we talking about Linux or kernel32.dll?
> 
>> When a user can't do *anything* about it, it's useless, and serves
>> as nothing but a cause for concern. "Oh no, is my laptop dying?".
> 
> Laptops come with Windows XP pre-installed for those users, what was 
> your problem again?

Sorry, but these comparisons to Windows are just childish.

Linux is not obliged to spit out meaningless, unhelpful error messages,
and be as user-hostile as possible. It's not a good trait. The current
error message is wrong ... if we can come up with something useful to
print, then great. But throwing 'Windows' around is not useful, and
neither are crappy, incorrect errors.

M.

Re: Remove silly messages from input layer.

[Nuri Jawad]

On Fri, 5 May 2006, Martin Bligh wrote:

> Sorry, but these comparisons to Windows are just childish.

No, in this case they're not. Clear text error messages, compared to 
obscure numerical error codes or no reports at all, are a very useful 
attribute of free operating systems that I do not want to miss because 
somebody is trying to please clueless Joe User.

> Linux is not obliged to spit out meaningless, unhelpful error messages,

I already explained in a previous reply why this message IS undisputably 
useful for me, and others have done the same. Why are we discussing if you 
don't read what we write but have obviously already made up your mind?

> The current error message is wrong

No it's not, the keyboard is in fact reporting too many pressed keys.

> But throwing 'Windows' around is not useful

It is useful to show you what kind of environment many users do not want 
to have. Hiding information is not user-friendly for the experienced user, 
only for the novice.

Regards, Nuri

Re: Remove silly messages from input layer.

[Jindrich Makovicka]

On Thu, 04 May 2006 16:45:58 +0100
David Greaves <david@dgreaves.com> wrote:

> Pavel Machek wrote:
> > On Wed 03-05-06 22:44:04, Dave Jones wrote:
> >   
> >> There are two messages in the input layer that seem to be
> >> triggerable very easily, and they confuse end-users to no end.
> >> "too many keys pressed? Should I press less keys?"
> >>     
> >
> > It actually means 'type more slowly' or 'use standard keymap' or
> > 'get a better keyboard' :-) or 'no, you are not imagining it, I've
> > seen your keypress and dropped it'.
> > 									Pavel
> >
> >   
> or 'the cat walked on the keyboard again...'

Note that this might be already patented:

http://www.bitboost.com/pawsense/pawsense-faq.html

-- 
Jindrich Makovicka

Re: Remove silly messages from input layer.

[Joseph Fannin]

On Fri, May 05, 2006 at 10:30:00PM +0200, Nuri Jawad wrote:
>
> It is useful to show you what kind of environment many users do not want
> to have. Hiding information is not user-friendly for the experienced user,
> only for the novice.

    Spamming my logs with these messages so often that my dmesg buffer
soon contains nothing but and I have to use grep -v to read my syslog
is not user-friendly, and I am hardly a novice.

    At most, a reworded message should be emitted on the first
occurance of the error and not again until a reboot.  If you want to
know more, there are debugging tools for that sort of thing.

--
Joseph Fannin
jhf@rivenstone.net

 /* So there I am, in the middle of my `netfilter-is-wonderful'
talk in Sydney, and someone asks `What happens if you try
to enlarge a 64k packet here?'. I think I said something
eloquent like `fuck'. - RR */

Re: Remove silly messages from input layer.

[Helge Hafting]

Dave Jones wrote:

>On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:
>
> > If you only pressed single key -- your keyboard is crap or there's
> > some problem in the driver.
> > 
> > If you never pressed any key -- your keyboard is crap or there's
> > some problem in the driver.
>
>That's hardly a constructive answer when the keyboard is a part of
>a laptop.  Crap hardware exists, get used to it.
>  
>
If some laptop comes with a bad keyboard, please blacklist
it so future linux users can avoid the brand when shopping
for hardware.

Helge Hafting

Re: Remove silly messages from input layer.

[Con Kolivas]

On Monday 08 May 2006 16:43, Helge Hafting wrote:
> Dave Jones wrote:
> >On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:
> > > If you only pressed single key -- your keyboard is crap or there's
> > > some problem in the driver.
> > >
> > > If you never pressed any key -- your keyboard is crap or there's
> > > some problem in the driver.
> >
> >That's hardly a constructive answer when the keyboard is a part of
> >a laptop.  Crap hardware exists, get used to it.
>
> If some laptop comes with a bad keyboard, please blacklist
> it so future linux users can avoid the brand when shopping
> for hardware.

This is great in theory but if we end up blacklisting half of the hardware out 
there we're stuffed. The truth is most hardware out there is cheap and nasty 
and sells in vast quantities. We have workarounds for timer code being buggy 
on virtually half the motherboards out there on amd64 for example...

-- 
-ck

Re: Remove silly messages from input layer.

[Helge Hafting]

Con Kolivas wrote:

>On Monday 08 May 2006 16:43, Helge Hafting wrote:
>  
>
>>Dave Jones wrote:
>>    
>>
>>>On Fri, May 05, 2006 at 12:31:23PM +0200, Pavel Machek wrote:
>>>      
>>>
>>>>If you only pressed single key -- your keyboard is crap or there's
>>>>some problem in the driver.
>>>>
>>>>If you never pressed any key -- your keyboard is crap or there's
>>>>some problem in the driver.
>>>>        
>>>>
>>>That's hardly a constructive answer when the keyboard is a part of
>>>a laptop.  Crap hardware exists, get used to it.
>>>      
>>>
>>If some laptop comes with a bad keyboard, please blacklist
>>it so future linux users can avoid the brand when shopping
>>for hardware.
>>    
>>
>
>This is great in theory but if we end up blacklisting half of the hardware out 
>there we're stuffed. The truth is most hardware out there is cheap and nasty 
>and sells in vast quantities. We have workarounds for timer code being buggy 
>on virtually half the motherboards out there on amd64 for example...
>  
>
Well, it depends on how broken they are then.
These keyboards actually work, so it is not so much a case
of being broken, more a case of  "too cheap to follow the spec,
but we can still get the keypresses"? 

I should have been more clear.  With blacklisting, I didn't mean
to make the driver refuse them.  I was thinking about
http://www.tldp.org/HOWTO/Hardware-HOWTO/index.html
and similiar sites, where a buyer can go and look for any linux issues
with the hardware he plans to buy.  There is no problem if 50%
of all hardware ends up here - one can then read about the issues
and decide if they matter enough to get a different brand instead.

Helge Hafting