linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Re: Keyboard oddness.
@ 2003-09-26  9:43 Nicolas Mailhot
  2003-09-26 10:24 ` Vojtech Pavlik
  0 siblings, 1 reply; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26  9:43 UTC (permalink / raw)
  To: linux-kernel; +Cc: Vojtech Pavlik

[-- Attachment #1: Type: text/plain, Size: 1282 bytes --]

Vojtech Pavlik  wrote:

|> Many people have reported missing key releases, and, as a consequence
| of that,
| > stuck keys. Your reports feel a bit different: the e0 is sometimes lost from
| > a key press, sometimes from a key release.

| I'm wondering if it could be a bug in the i8042.c driver ...

This is worse than that. I'm seing the same kind of bug on a pure HID+EHCI setup 
and I've seen other reports of USB problems on the lists these past months.

I can't just believe everyone and his cat has suddenly a faulty/broken keyboard.

The fact is software autorepeat seems extremely brittle in 2.6 and goes nuts every
once in a while for *everyone* (and I also seem to remember this was not always the 
case - the first 2.5 I tried didn't go mad on my setup this only happened later in 
2.5.7x times I think). Now maybe the underlying keyboard drivers are feeding it junk
I don't know but this is no justification for the way it's been misbehaving (crap 
hardware happens, glitchy hardware is common but the autorepeat code seems to expect
ideal behaviour that only happens on paper)

Couldn't it at least detect there's a problem ? Most people I know do not press a key
2000+ times in a row during normal activity.

Cheers,

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26  9:43 Keyboard oddness Nicolas Mailhot
@ 2003-09-26 10:24 ` Vojtech Pavlik
  2003-09-26 10:41   ` Nicolas Mailhot
  2003-10-01 10:09   ` Pavel Machek
  0 siblings, 2 replies; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 10:24 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: linux-kernel, Vojtech Pavlik

On Fri, Sep 26, 2003 at 11:43:43AM +0200, Nicolas Mailhot wrote:
> Vojtech Pavlik  wrote:
> 
> |> Many people have reported missing key releases, and, as a consequence
> | of that,
> | > stuck keys. Your reports feel a bit different: the e0 is sometimes lost from
> | > a key press, sometimes from a key release.
> 
> | I'm wondering if it could be a bug in the i8042.c driver ...
> 
> This is worse than that. I'm seing the same kind of bug on a pure HID+EHCI setup 
> and I've seen other reports of USB problems on the lists these past months.

Well, the autorepeat-gone-bonkers is just a symptom, and a likely one,
not a cause. Few people will notice a missed keypress, just assume a bad
contact on the key, but a missing key release simply must cause infinite
autorepeat until another (or the same) key is pressed and released.

> I can't just believe everyone and his cat has suddenly a faulty/broken keyboard.
> 
> The fact is software autorepeat seems extremely brittle in 2.6 and goes nuts every
> once in a while for *everyone* (and I also seem to remember this was not always the 
> case - the first 2.5 I tried didn't go mad on my setup this only happened later in 
> 2.5.7x times I think). Now maybe the underlying keyboard drivers are feeding it junk
> I don't know but this is no justification for the way it's been misbehaving (crap 
> hardware happens, glitchy hardware is common but the autorepeat code seems to expect
> ideal behaviour that only happens on paper)
> 
> Couldn't it at least detect there's a problem ? Most people I know do not press a key
> 2000+ times in a row during normal activity.

You do. Scrolling up/down in a document is one example. And there is no
point to limit the repeat to say 80 or 200 characters. You would still
hate having 80 repeated characters and then it stopping.

The problem really is there is no way to detect it. My latest patches
should fix this for AT keyboards by not using software autorepeat for
them.

Of course this won't fix any problems with USB, if there are still any.
My USB keyboard works just perfectly, no problems with the autorepeat.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 10:24 ` Vojtech Pavlik
@ 2003-09-26 10:41   ` Nicolas Mailhot
  2003-09-26 10:51     ` Vojtech Pavlik
                       ` (2 more replies)
  2003-10-01 10:09   ` Pavel Machek
  1 sibling, 3 replies; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 10:41 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 1246 bytes --]

Le ven 26/09/2003 à 12:24, Vojtech Pavlik a écrit :
> On Fri, Sep 26, 2003 at 11:43:43AM +0200, Nicolas Mailhot wrote:
> > Vojtech Pavlik  wrote:

[...]

> > Couldn't it at least detect there's a problem ? Most people I know do not press a key
> > 2000+ times in a row during normal activity.
> 
> You do. Scrolling up/down in a document is one example. And there is no
> point to limit the repeat to say 80 or 200 characters. You would still
> hate having 80 repeated characters and then it stopping.

Well then only allow monster autorepeats for arrows then.
(they are never stuck in my board anyway;)

> The problem really is there is no way to detect it. My latest patches
> should fix this for AT keyboards by not using software autorepeat for
> them.
> 
> Of course this won't fix any problems with USB, if there are still any.
> My USB keyboard works just perfectly, no problems with the autorepeat.

Well mine doesn't:(. I seem to have gathered from past threads that HID
makes the full keyboard status available at all time (unlike AT which
only provides push/release events). Couldn't the repeat code just
double-check the key is really stuck every ten repeats for example ?

Cheers

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 10:41   ` Nicolas Mailhot
@ 2003-09-26 10:51     ` Vojtech Pavlik
  2003-09-26 12:48     ` jw schultz
  2003-09-26 13:08     ` Nicolas Mailhot
  2 siblings, 0 replies; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 10:51 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 12:41:38PM +0200, Nicolas Mailhot wrote:
> Le ven 26/09/2003 ? 12:24, Vojtech Pavlik a écrit :
> > On Fri, Sep 26, 2003 at 11:43:43AM +0200, Nicolas Mailhot wrote:
> > > Vojtech Pavlik  wrote:
> 
> [...]
> 
> > > Couldn't it at least detect there's a problem ? Most people I know do not press a key
> > > 2000+ times in a row during normal activity.
> > 
> > You do. Scrolling up/down in a document is one example. And there is no
> > point to limit the repeat to say 80 or 200 characters. You would still
> > hate having 80 repeated characters and then it stopping.
> 
> Well then only allow monster autorepeats for arrows then.
> (they are never stuck in my board anyway;)
> 
> > The problem really is there is no way to detect it. My latest patches
> > should fix this for AT keyboards by not using software autorepeat for
> > them.
> > 
> > Of course this won't fix any problems with USB, if there are still any.
> > My USB keyboard works just perfectly, no problems with the autorepeat.
> 
> Well mine doesn't:(. I seem to have gathered from past threads that HID
> makes the full keyboard status available at all time (unlike AT which
> only provides push/release events). Couldn't the repeat code just
> double-check the key is really stuck every ten repeats for example ?

It provides the full state when the state changes. If you miss the
change on release ...

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 10:41   ` Nicolas Mailhot
  2003-09-26 10:51     ` Vojtech Pavlik
@ 2003-09-26 12:48     ` jw schultz
  2003-09-26 13:08     ` Nicolas Mailhot
  2 siblings, 0 replies; 35+ messages in thread
From: jw schultz @ 2003-09-26 12:48 UTC (permalink / raw)
  To: linux-kernel

On Fri, Sep 26, 2003 at 12:41:38PM +0200, Nicolas Mailhot wrote:
> Le ven 26/09/2003 à 12:24, Vojtech Pavlik a écrit :
> > On Fri, Sep 26, 2003 at 11:43:43AM +0200, Nicolas Mailhot wrote:
> > > Vojtech Pavlik  wrote:
> 
> [...]
> 
> > > Couldn't it at least detect there's a problem ? Most people I know do not press a key
> > > 2000+ times in a row during normal activity.
> > 
> > You do. Scrolling up/down in a document is one example. And there is no
> > point to limit the repeat to say 80 or 200 characters. You would still
> > hate having 80 repeated characters and then it stopping.
> 
> Well then only allow monster autorepeats for arrows then.
> (they are never stuck in my board anyway;)

And j, k, w, b, ., all function keys, <bs>, <del>, <cr>,
<sp>, <tab> and any other key used by any editor or game for
navigation, level control or other function where the same
key would be used scores of times in in rapid sequence.
Hmm, i've just described most of the keyboard.

END SARCASM




-- 
________________________________________________________________
	J.W. Schultz            Pegasystems Technologies
	email address:		jw@pegasys.ws

		Remember Cernan and Schmitt

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 10:41   ` Nicolas Mailhot
  2003-09-26 10:51     ` Vojtech Pavlik
  2003-09-26 12:48     ` jw schultz
@ 2003-09-26 13:08     ` Nicolas Mailhot
  2003-09-26 13:41       ` Vojtech Pavlik
  2003-09-26 13:56       ` Helge Hafting
  2 siblings, 2 replies; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 13:08 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 1317 bytes --]


| > > > Couldn't it at least detect there's a problem ? Most people I know do not press a key
| > > > 2000+ times in a row during normal activity.
| > > 
| > > You do. Scrolling up/down in a document is one example. And there is no
| > > point to limit the repeat to say 80 or 200 characters. You would still
| > > hate having 80 repeated characters and then it stopping.
| > 
| > Well then only allow monster autorepeats for arrows then.
| > (they are never stuck in my board anyway;)

| And j, k, w, b, ., all function keys, <bs>, <del>, <cr>,
| <sp>, <tab> and any other key used by any editor or game for
| navigation, level control or other function where the same
| key would be used scores of times in in rapid sequence.

score << 2k+

I wrote about monster autorepeats not every single duplicated keypress.
I fully agree it's stupid to expect detecting every single bogus repeat.

However saying the system has no way to guess monster
autorepeats=problem is just plain wrong. There *are* thresholds after
which one can be 99% sure there is a problem (autorepeat gone mad or cat
sitting on the keyboard). No one is going to complain he has to release
a key every hundred or so repeats to confirm there's a human on the
other side of the keyboard.

Cheers,

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 13:08     ` Nicolas Mailhot
@ 2003-09-26 13:41       ` Vojtech Pavlik
  2003-09-26 14:12         ` Nicolas Mailhot
  2003-09-26 13:56       ` Helge Hafting
  1 sibling, 1 reply; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 13:41 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 03:08:36PM +0200, Nicolas Mailhot wrote:

> | > > > Couldn't it at least detect there's a problem ? Most people I know do not press a key
> | > > > 2000+ times in a row during normal activity.
> | > > 
> | > > You do. Scrolling up/down in a document is one example. And there is no
> | > > point to limit the repeat to say 80 or 200 characters. You would still
> | > > hate having 80 repeated characters and then it stopping.
> | > 
> | > Well then only allow monster autorepeats for arrows then.
> | > (they are never stuck in my board anyway;)
> 
> | And j, k, w, b, ., all function keys, <bs>, <del>, <cr>,
> | <sp>, <tab> and any other key used by any editor or game for
> | navigation, level control or other function where the same
> | key would be used scores of times in in rapid sequence.
> 
> score << 2k+
> 
> I wrote about monster autorepeats not every single duplicated keypress.
> I fully agree it's stupid to expect detecting every single bogus repeat.
> 
> However saying the system has no way to guess monster
> autorepeats=problem is just plain wrong. There *are* thresholds after
> which one can be 99% sure there is a problem (autorepeat gone mad or cat
> sitting on the keyboard). No one is going to complain he has to release
> a key every hundred or so repeats to confirm there's a human on the
> other side of the keyboard.

But what use would be this? You'd still get a screenful of 'j's for
example, maybe only 200 instead of 2000, but where is the difference?

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 13:08     ` Nicolas Mailhot
  2003-09-26 13:41       ` Vojtech Pavlik
@ 2003-09-26 13:56       ` Helge Hafting
  2003-09-26 14:01         ` Mike Dresser
  1 sibling, 1 reply; 35+ messages in thread
From: Helge Hafting @ 2003-09-26 13:56 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: linux-kernel

Nicolas Mailhot wrote:
[...]
> I wrote about monster autorepeats not every single duplicated keypress.
> I fully agree it's stupid to expect detecting every single bogus repeat.
> 
> However saying the system has no way to guess monster
> autorepeats=problem is just plain wrong. There *are* thresholds after
> which one can be 99% sure there is a problem (autorepeat gone mad or cat
> sitting on the keyboard). No one is going to complain he has to release
> a key every hundred or so repeats to confirm there's a human on the
> other side of the keyboard.
> 
First, such detection is kind of useless.  If I get 200 W's before
the system detects, well I'll fix it long before detection kicks
in by tapping the stuck key.  That tends to unstick it.
Keys don't usually get stuck when nobody's there, they stick because
of a missed release, not a bogus press.

Second - yes, people are going to get impressively pissed off
if they have to release a key now and then.  Scrolling on a heavily
loaded machine - it stops from time to time anyway - but now we
have to release the key all the time?

And don't even think of having to release a key now and then in
a action game.  Games may use any key, so no restriction on
which keys may repeat is useable.

Gamers press keys for a long time, movement keys in quake, the
accelerator key in a car game, the fire key in space invaders.
Bogus unsticking of keys isn't acceptable - there will certainly
be a storm of patches for removing the misfeature.

Helge Hafting





^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 13:56       ` Helge Hafting
@ 2003-09-26 14:01         ` Mike Dresser
  2003-09-28 10:58           ` Helge Hafting
  0 siblings, 1 reply; 35+ messages in thread
From: Mike Dresser @ 2003-09-26 14:01 UTC (permalink / raw)
  To: linux-kernel

On Fri, 26 Sep 2003, Helge Hafting wrote:

> Gamers press keys for a long time, movement keys in quake, the
> accelerator key in a car game, the fire key in space invaders.

And I throw something heavy on the d key on a few of my data gathering
email accounts, and go away for a few hours.

So there's a non-game application.

Mike

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 13:41       ` Vojtech Pavlik
@ 2003-09-26 14:12         ` Nicolas Mailhot
  2003-09-26 14:17           ` Vojtech Pavlik
  0 siblings, 1 reply; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 14:12 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 2138 bytes --]

Le ven 26/09/2003 à 15:41, Vojtech Pavlik a écrit :
> On Fri, Sep 26, 2003 at 03:08:36PM +0200, Nicolas Mailhot wrote:
> 
> > | > > > Couldn't it at least detect there's a problem ? Most people I know do not press a key
> > | > > > 2000+ times in a row during normal activity.
> > | > > 
> > | > > You do. Scrolling up/down in a document is one example. And there is no
> > | > > point to limit the repeat to say 80 or 200 characters. You would still
> > | > > hate having 80 repeated characters and then it stopping.
> > | > 
> > | > Well then only allow monster autorepeats for arrows then.
> > | > (they are never stuck in my board anyway;)
> > 
> > | And j, k, w, b, ., all function keys, <bs>, <del>, <cr>,
> > | <sp>, <tab> and any other key used by any editor or game for
> > | navigation, level control or other function where the same
> > | key would be used scores of times in in rapid sequence.
> > 
> > score << 2k+
> > 
> > I wrote about monster autorepeats not every single duplicated keypress.
> > I fully agree it's stupid to expect detecting every single bogus repeat.
> > 
> > However saying the system has no way to guess monster
> > autorepeats=problem is just plain wrong. There *are* thresholds after
> > which one can be 99% sure there is a problem (autorepeat gone mad or cat
> > sitting on the keyboard). No one is going to complain he has to release
> > a key every hundred or so repeats to confirm there's a human on the
> > other side of the keyboard.
> 
> But what use would be this? You'd still get a screenful of 'j's for
> example, maybe only 200 instead of 2000, but where is the difference?

The difference being the system can then try to rescue my keyboard;)
Right now the only fix I have is to reboot the system because there is
precious little I can do with a stuck keyboard. Thank god software
reboot is always possible be it with the mouse or the acpi button.

(and this also solves the case when something falls on a keyboard which
does happen now and then. I don't mind a screen of j's when the
alternative is 200 j's screenfulls)

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:12         ` Nicolas Mailhot
@ 2003-09-26 14:17           ` Vojtech Pavlik
  2003-09-26 14:21             ` Nicolas Mailhot
  0 siblings, 1 reply; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 14:17 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 04:12:47PM +0200, Nicolas Mailhot wrote:
 
> The difference being the system can then try to rescue my keyboard;)
> Right now the only fix I have is to reboot the system because there is
> precious little I can do with a stuck keyboard. Thank god software
> reboot is always possible be it with the mouse or the acpi button.
> 
> (and this also solves the case when something falls on a keyboard which
> does happen now and then. I don't mind a screen of j's when the
> alternative is 200 j's screenfulls)

You can simply press any key and it'll stop repeating. 

If that doesn't work, you have a more severe problem than a stuck key,
that wouldn't be solved by stopping the repeat.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:17           ` Vojtech Pavlik
@ 2003-09-26 14:21             ` Nicolas Mailhot
  2003-09-26 14:26               ` Vojtech Pavlik
  0 siblings, 1 reply; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 14:21 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 913 bytes --]

Le ven 26/09/2003 à 16:17, Vojtech Pavlik a écrit :
> On Fri, Sep 26, 2003 at 04:12:47PM +0200, Nicolas Mailhot wrote:
>  
> > The difference being the system can then try to rescue my keyboard;)
> > Right now the only fix I have is to reboot the system because there is
> > precious little I can do with a stuck keyboard. Thank god software
> > reboot is always possible be it with the mouse or the acpi button.
> > 
> > (and this also solves the case when something falls on a keyboard which
> > does happen now and then. I don't mind a screen of j's when the
> > alternative is 200 j's screenfulls)
> 
> You can simply press any key and it'll stop repeating. 
> 
> If that doesn't work, you have a more severe problem than a stuck key,
> that wouldn't be solved by stopping the repeat.

It stops the repeat all right.
The problem is the keyboard is dead afterwards:(

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:21             ` Nicolas Mailhot
@ 2003-09-26 14:26               ` Vojtech Pavlik
  2003-09-26 14:50                 ` Nicolas Mailhot
  0 siblings, 1 reply; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 14:26 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 04:21:57PM +0200, Nicolas Mailhot wrote:
> Le ven 26/09/2003 ? 16:17, Vojtech Pavlik a écrit :
> > On Fri, Sep 26, 2003 at 04:12:47PM +0200, Nicolas Mailhot wrote:
> >  
> > > The difference being the system can then try to rescue my keyboard;)
> > > Right now the only fix I have is to reboot the system because there is
> > > precious little I can do with a stuck keyboard. Thank god software
> > > reboot is always possible be it with the mouse or the acpi button.
> > > 
> > > (and this also solves the case when something falls on a keyboard which
> > > does happen now and then. I don't mind a screen of j's when the
> > > alternative is 200 j's screenfulls)
> > 
> > You can simply press any key and it'll stop repeating. 
> > 
> > If that doesn't work, you have a more severe problem than a stuck key,
> > that wouldn't be solved by stopping the repeat.
> 
> It stops the repeat all right.
> The problem is the keyboard is dead afterwards:(

That's very interesting. Can you enable DEBUG in i8042.c and post a log?

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:26               ` Vojtech Pavlik
@ 2003-09-26 14:50                 ` Nicolas Mailhot
  2003-09-26 15:06                   ` Vojtech Pavlik
  0 siblings, 1 reply; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 14:50 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 1262 bytes --]

Le ven 26/09/2003 à 16:26, Vojtech Pavlik a écrit :
> On Fri, Sep 26, 2003 at 04:21:57PM +0200, Nicolas Mailhot wrote:
> > Le ven 26/09/2003 ? 16:17, Vojtech Pavlik a écrit :
> > > On Fri, Sep 26, 2003 at 04:12:47PM +0200, Nicolas Mailhot wrote:
> > >  
> > > > The difference being the system can then try to rescue my keyboard;)
> > > > Right now the only fix I have is to reboot the system because there is
> > > > precious little I can do with a stuck keyboard. Thank god software
> > > > reboot is always possible be it with the mouse or the acpi button.
> > > > 
> > > > (and this also solves the case when something falls on a keyboard which
> > > > does happen now and then. I don't mind a screen of j's when the
> > > > alternative is 200 j's screenfulls)
> > > 
> > > You can simply press any key and it'll stop repeating. 
> > > 
> > > If that doesn't work, you have a more severe problem than a stuck key,
> > > that wouldn't be solved by stopping the repeat.
> > 
> > It stops the repeat all right.
> > The problem is the keyboard is dead afterwards:(
> 
> That's very interesting. Can you enable DEBUG in i8042.c and post a log?

Will it be of any use for an USB keyboard ? (just asking)

Cheers,

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:50                 ` Nicolas Mailhot
@ 2003-09-26 15:06                   ` Vojtech Pavlik
  2003-09-26 15:21                     ` Nicolas Mailhot
  0 siblings, 1 reply; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 15:06 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: linux-kernel

On Fri, Sep 26, 2003 at 04:50:25PM +0200, Nicolas Mailhot wrote:
> Le ven 26/09/2003 ? 16:26, Vojtech Pavlik a écrit :
> > On Fri, Sep 26, 2003 at 04:21:57PM +0200, Nicolas Mailhot wrote:
> > > Le ven 26/09/2003 ? 16:17, Vojtech Pavlik a écrit :
> > > > On Fri, Sep 26, 2003 at 04:12:47PM +0200, Nicolas Mailhot wrote:
> > > >  
> > > > > The difference being the system can then try to rescue my keyboard;)
> > > > > Right now the only fix I have is to reboot the system because there is
> > > > > precious little I can do with a stuck keyboard. Thank god software
> > > > > reboot is always possible be it with the mouse or the acpi button.
> > > > > 
> > > > > (and this also solves the case when something falls on a keyboard which
> > > > > does happen now and then. I don't mind a screen of j's when the
> > > > > alternative is 200 j's screenfulls)
> > > > 
> > > > You can simply press any key and it'll stop repeating. 
> > > > 
> > > > If that doesn't work, you have a more severe problem than a stuck key,
> > > > that wouldn't be solved by stopping the repeat.
> > > 
> > > It stops the repeat all right.
> > > The problem is the keyboard is dead afterwards:(
> > 
> > That's very interesting. Can you enable DEBUG in i8042.c and post a log?
> 
> Will it be of any use for an USB keyboard ? (just asking)

No. For an USB keyboard I'd suggest unplugging it, then re-plugging and
then it should work. Then look at what 'dmesg' says.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 15:06                   ` Vojtech Pavlik
@ 2003-09-26 15:21                     ` Nicolas Mailhot
  2003-09-26 15:24                       ` Vojtech Pavlik
  0 siblings, 1 reply; 35+ messages in thread
From: Nicolas Mailhot @ 2003-09-26 15:21 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

[-- Attachment #1: Type: text/plain, Size: 1147 bytes --]

Le ven 26/09/2003 à 17:06, Vojtech Pavlik a écrit :
> On Fri, Sep 26, 2003 at 04:50:25PM +0200, Nicolas Mailhot wrote:
> > Le ven 26/09/2003 ? 16:26, Vojtech Pavlik a écrit :
> > > On Fri, Sep 26, 2003 at 04:21:57PM +0200, Nicolas Mailhot wrote:
> > > > Le ven 26/09/2003 ? 16:17, Vojtech Pavlik a écrit :

> > > > > If that doesn't work, you have a more severe problem than a stuck key,
> > > > > that wouldn't be solved by stopping the repeat.
> > > > 
> > > > It stops the repeat all right.
> > > > The problem is the keyboard is dead afterwards:(
> > > 
> > > That's very interesting. Can you enable DEBUG in i8042.c and post a log?
> > 
> > Will it be of any use for an USB keyboard ? (just asking)
> 
> No. For an USB keyboard I'd suggest unplugging it, then re-plugging and
> then it should work. Then look at what 'dmesg' says.

Ok. I think I've already tried this and the outcome was not satisfying,
and I sort of remember dmesg was empty. I'll try this with usb debug
enabled this evening or next monday. (I just hope something comes out of
this - 2.6 is great till the keyboard goes mad).

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 15:21                     ` Nicolas Mailhot
@ 2003-09-26 15:24                       ` Vojtech Pavlik
  0 siblings, 0 replies; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26 15:24 UTC (permalink / raw)
  To: Nicolas Mailhot; +Cc: Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 05:21:26PM +0200, Nicolas Mailhot wrote:
> Le ven 26/09/2003 ? 17:06, Vojtech Pavlik a écrit :
> > On Fri, Sep 26, 2003 at 04:50:25PM +0200, Nicolas Mailhot wrote:
> > > Le ven 26/09/2003 ? 16:26, Vojtech Pavlik a écrit :
> > > > On Fri, Sep 26, 2003 at 04:21:57PM +0200, Nicolas Mailhot wrote:
> > > > > Le ven 26/09/2003 ? 16:17, Vojtech Pavlik a écrit :
> 
> > > > > > If that doesn't work, you have a more severe problem than a stuck key,
> > > > > > that wouldn't be solved by stopping the repeat.
> > > > > 
> > > > > It stops the repeat all right.
> > > > > The problem is the keyboard is dead afterwards:(
> > > > 
> > > > That's very interesting. Can you enable DEBUG in i8042.c and post a log?
> > > 
> > > Will it be of any use for an USB keyboard ? (just asking)
> > 
> > No. For an USB keyboard I'd suggest unplugging it, then re-plugging and
> > then it should work. Then look at what 'dmesg' says.
> 
> Ok. I think I've already tried this and the outcome was not satisfying,
> and I sort of remember dmesg was empty. I'll try this with usb debug
> enabled this evening or next monday. (I just hope something comes out of
> this - 2.6 is great till the keyboard goes mad).

You can also enable DEBUG in hid-core.c.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 14:01         ` Mike Dresser
@ 2003-09-28 10:58           ` Helge Hafting
  2003-10-02 12:40             ` Pavel Machek
  0 siblings, 1 reply; 35+ messages in thread
From: Helge Hafting @ 2003-09-28 10:58 UTC (permalink / raw)
  To: Mike Dresser; +Cc: linux-kernel

On Fri, Sep 26, 2003 at 10:01:36AM -0400, Mike Dresser wrote:
> On Fri, 26 Sep 2003, Helge Hafting wrote:
> 
> > Gamers press keys for a long time, movement keys in quake, the
> > accelerator key in a car game, the fire key in space invaders.
> 
> And I throw something heavy on the d key on a few of my data gathering
> email accounts, and go away for a few hours.
> 
> So there's a non-game application.
> 
1. Do we want to protect people against leaving something heavy
   on the keyboard - when this protection will cause other
   trouble?
   
2. Will this help - you'll get 200 or so d's before the stuck
   key is detected.  Enough to "d"elete a lot...
   
Helge Hafting

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26 10:24 ` Vojtech Pavlik
  2003-09-26 10:41   ` Nicolas Mailhot
@ 2003-10-01 10:09   ` Pavel Machek
  2003-10-01 13:06     ` Nicolas Mailhot
  1 sibling, 1 reply; 35+ messages in thread
From: Pavel Machek @ 2003-10-01 10:09 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: Nicolas Mailhot, linux-kernel

Hi!

> Of course this won't fix any problems with USB, if there are still any.
> My USB keyboard works just perfectly, no problems with the autorepeat.

Can you try running your system with interrupts disabled for 100ms+
sometimes? That should show any bugs/races in keyboard code. Perhaps
you have good hardware that never ever disables interrupts for that
long, but other people have more broken stuff.
								Pavel
-- 
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-10-01 10:09   ` Pavel Machek
@ 2003-10-01 13:06     ` Nicolas Mailhot
  0 siblings, 0 replies; 35+ messages in thread
From: Nicolas Mailhot @ 2003-10-01 13:06 UTC (permalink / raw)
  To: Pavel Machek; +Cc: Vojtech Pavlik, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 691 bytes --]

Le mer 01/10/2003 à 12:09, Pavel Machek a écrit :
> Hi!
> 
> > Of course this won't fix any problems with USB, if there are still any.
> > My USB keyboard works just perfectly, no problems with the autorepeat.
> 
> Can you try running your system with interrupts disabled for 100ms+
> sometimes? That should show any bugs/races in keyboard code. Perhaps
> you have good hardware that never ever disables interrupts for that
> long, but other people have more broken stuff.

Will try - as soon as I restore the raid that the last 2.6.0-test5-bk
snapshots ate:(

I had something setup last evening unfortunately the system's gone AWOL
since.

Cheers,

-- 
Nicolas Mailhot

[-- Attachment #2: Ceci est une partie de message numériquement signée --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-28 10:58           ` Helge Hafting
@ 2003-10-02 12:40             ` Pavel Machek
  0 siblings, 0 replies; 35+ messages in thread
From: Pavel Machek @ 2003-10-02 12:40 UTC (permalink / raw)
  To: Helge Hafting; +Cc: Mike Dresser, linux-kernel

Hi!

> > So there's a non-game application.
> > 
> 1. Do we want to protect people against leaving something heavy
>    on the keyboard - when this protection will cause other
>    trouble?

No, thats a feature. Last night I was testing swsusp and had one
heavy object on ctrl (reboot instead of powerdown)
and second on enter (get through lilo).
-- 
				Pavel
Written on sharp zaurus, because my Velo1 broke. If you have Velo you don't need...


^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26  8:15         ` Andries Brouwer
  2003-09-26  8:43           ` Rob Landley
@ 2003-09-26  8:46           ` Vojtech Pavlik
  1 sibling, 0 replies; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-26  8:46 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: Rob Landley, Vojtech Pavlik, linux-kernel

On Fri, Sep 26, 2003 at 10:15:42AM +0200, Andries Brouwer wrote:
> On Thu, Sep 25, 2003 at 08:27:57PM -0500, Rob Landley wrote:
> > Okay, a little fresh data:
> 
> > Sep 25 20:22:22 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd1, 
> > on isa0060/serio0) pressed.
> > Sep 25 20:22:22 localhost kernel: i8042 history: d1 e0 51 e0 d1 e0 51 e0 d1 e0 
> > 51 e0 d1 e0 51 d1
> 
> e0 51 is PageDown press
> e0 d1 is PageDown release
> 
> You see here (apart from the first byte, which probably is the second half
> of a PageDown release): PageDown press, release, press, release, press, release,
> press, broken release.
> 
> A byte e0 was lost, and the release was not seen as a PageDown release.
> 
> > The page down key is the one that stuck.  I pressed another key (possibly 
> > either cursor up or page up) to unstick it, and then the next time I pressed 
> > page down it didn't register, but the time after that it did.
> 
> > You're talking about missed keypresses, but the end-user symptom I'm seeing
> > is definitely a missed key release
> 
> Yes - here a release was garbled.
> 
> Many people have reported missing key releases, and, as a consequence of that,
> stuck keys. Your reports feel a bit different: the e0 is sometimes lost from
> a key press, sometimes from a key release.

I'm wondering if it could be a bug in the i8042.c driver ...

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26  8:15         ` Andries Brouwer
@ 2003-09-26  8:43           ` Rob Landley
  2003-09-26  8:46           ` Vojtech Pavlik
  1 sibling, 0 replies; 35+ messages in thread
From: Rob Landley @ 2003-09-26  8:43 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: Vojtech Pavlik, linux-kernel

On Friday 26 September 2003 03:15, Andries Brouwer wrote:

> > You're talking about missed keypresses, but the end-user symptom I'm
> > seeing is definitely a missed key release
>
> Yes - here a release was garbled.
>
> Many people have reported missing key releases, and, as a consequence of
> that, stuck keys. Your reports feel a bit different: the e0 is sometimes
> lost from a key press, sometimes from a key release.

I don't know what to tell you.  When I compiled 2.4 (no detectable problems), 
I wasn't using the input (and in fact had it disabled).  Not necessarily a 
useful piece of information.  The symptom that I notice is software 
autorepeat going bananas when the hardware at least knows the key has been 
released and would not be sending hardware autorepeat events...

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-26  1:27       ` Rob Landley
@ 2003-09-26  8:15         ` Andries Brouwer
  2003-09-26  8:43           ` Rob Landley
  2003-09-26  8:46           ` Vojtech Pavlik
  0 siblings, 2 replies; 35+ messages in thread
From: Andries Brouwer @ 2003-09-26  8:15 UTC (permalink / raw)
  To: Rob Landley; +Cc: Vojtech Pavlik, linux-kernel

On Thu, Sep 25, 2003 at 08:27:57PM -0500, Rob Landley wrote:
> Okay, a little fresh data:

> Sep 25 20:22:22 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd1, 
> on isa0060/serio0) pressed.
> Sep 25 20:22:22 localhost kernel: i8042 history: d1 e0 51 e0 d1 e0 51 e0 d1 e0 
> 51 e0 d1 e0 51 d1

e0 51 is PageDown press
e0 d1 is PageDown release

You see here (apart from the first byte, which probably is the second half
of a PageDown release): PageDown press, release, press, release, press, release,
press, broken release.

A byte e0 was lost, and the release was not seen as a PageDown release.

> The page down key is the one that stuck.  I pressed another key (possibly 
> either cursor up or page up) to unstick it, and then the next time I pressed 
> page down it didn't register, but the time after that it did.

> You're talking about missed keypresses, but the end-user symptom I'm seeing
> is definitely a missed key release

Yes - here a release was garbled.

Many people have reported missing key releases, and, as a consequence of that,
stuck keys. Your reports feel a bit different: the e0 is sometimes lost from
a key press, sometimes from a key release.


^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-22 22:06     ` Andries Brouwer
  2003-09-23  0:23       ` Rob Landley
  2003-09-25 23:59       ` Rob Landley
@ 2003-09-26  1:27       ` Rob Landley
  2003-09-26  8:15         ` Andries Brouwer
  2 siblings, 1 reply; 35+ messages in thread
From: Rob Landley @ 2003-09-26  1:27 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: Vojtech Pavlik, linux-kernel

On Monday 22 September 2003 17:06, Andries Brouwer wrote:

> > Any clues?  (This happens to me at least once an hour...)
>
> Some people have been reporting missing key releases (maybe also you),
> but these are all missing key presses. It is easiest to blame the
> keyboard, even though I could imagine ways to blame the kernel.
>
> What about 2.4? Do you have to go back once an hour and add a symbol
> that was not transmitted correctly? Does 2.4 work perfectly for you?
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Okay, a little fresh data:

I just had the keyboard key sticking problem occur, 2.6.0-test5-mm4.

Sep 25 20:06:25 localhost dhclient: DHCPREQUEST on eth0 to 128.83.185.55 port 
67
Sep 25 20:06:25 localhost dhclient: DHCPACK from 128.83.185.55
Sep 25 20:06:25 localhost dhclient: bound to 128.62.240.239 -- renewal in 1608 
seconds.
Sep 25 20:22:22 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd1, 
on isa0060/serio0) pressed.
Sep 25 20:22:22 localhost kernel: i8042 history: d1 e0 51 e0 d1 e0 51 e0 d1 e0 
51 e0 d1 e0 51 d1
Sep 25 20:22:42 localhost su(pam_unix)[2047]: session opened for user root by 
landley(uid=500)

The page down key is the one that stuck.  I pressed another key (possibly 
either cursor up or page up) to unstick it, and then the next time I pressed 
page down it didn't register, but the time after that it did.

Checking the log, I saw the above message.  This is the first time since the 
last reboot the problem has occurred, and the first such message since the 
last reboot.  The message IS occurring when I have the key sticking problem, 
as shown by the timestamps.

You're talking about missed keypresses, but the end-user symptom I'm seeing is 
definitely a missed key release, and it's associated with the above log 
message.

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-22 22:06     ` Andries Brouwer
  2003-09-23  0:23       ` Rob Landley
@ 2003-09-25 23:59       ` Rob Landley
  2003-09-26  1:27       ` Rob Landley
  2 siblings, 0 replies; 35+ messages in thread
From: Rob Landley @ 2003-09-25 23:59 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: Vojtech Pavlik, linux-kernel

On Monday 22 September 2003 17:06, Andries Brouwer wrote:

> > Any clues?  (This happens to me at least once an hour...)
>
> Some people have been reporting missing key releases (maybe also you),
> but these are all missing key presses. It is easiest to blame the
> keyboard, even though I could imagine ways to blame the kernel.

If the press is the first event, how does it know that it's missing?  Are you 
saying t got a key release that didn't match an active keypress?  This seems 
like a similar problem.  Are missing key releases flagged as errors and 
logged?  (Maybe the next time that key is pressed, if the system thinks it's 
already down?)

I'm certainly seeing missing releases.  Missing presses are just typos, I 
usually have enough gunk in the keyboard that it misses the occasional key 
anyway.  (This laptop lives in my backpack, with books and papers and Wendy's 
coupons and pens and...)  But "enter", "delete", or "cursor up" holding down 
for an extra second and change can be quite annoying.

> What about 2.4? Do you have to go back once an hour and add a symbol
> that was not transmitted correctly? Does 2.4 work perfectly for you?

Okay, over the past few days I've booted into 2.4.21 three times and done my 
standard text-intensive programming, email, and word processing.  I didn't 
notice any keys sticking.  (It's intermittent enough problem that this isn't 
difinitive, but it's the best I can do.)

On the other hand, the key repeat rate is a lot slower under 2.4.21, and I 
suspect it's using the hardware repeat instead of the software repeat.  The 
keyboard hardware definitely knows when a key gets released. :)

I.E. Under 2.4.21 I don't experience a problem.  (Maybe there is a technical 
problem with the hardware, but it doesn't manifest in a way that I ever 
notice.)  I'd be happy to live with the slower repeat rate if I could get 2.6 
to use 2.4's keyboard access method.

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-23  0:23       ` Rob Landley
@ 2003-09-23 10:51         ` Greg Louis
  0 siblings, 0 replies; 35+ messages in thread
From: Greg Louis @ 2003-09-23 10:51 UTC (permalink / raw)
  To: linux-kernel

On 20030922 (Mon) at 1923:28 -0500, Rob Landley wrote:
> On Monday 22 September 2003 17:06, Andries Brouwer wrote:
> > On Mon, Sep 22, 2003 at 03:06:08PM -0500, Rob Landley wrote:
> 
> > > Any clues?  (This happens to me at least once an hour...)
> >
> > Some people have been reporting missing key releases (maybe also you),
> > but these are all missing key presses. It is easiest to blame the
> > keyboard, even though I could imagine ways to blame the kernel.
> >
> > What about 2.4?
> 
> 2.4 worked for me when I used it.  I haven't booted 2.4 in weeks

I've been missing keypresses for at least six weeks; at first I too
thought the keyboard was the culprit, but I've anecdotal grumblings in
email from several 2.4 users who thought the same of their own machines.
Not very likely to be a keyboard-infesting virus out there... 
Frequency seems to be of the order of 0.001 on average.

It might not be only keyboard interrupts that are being missed, but I
have no hard data -- just a gut feeling that occasional ethernet
packets are going astray too for no known good reason.  Hasn't been
enough of a problem to trigger serious investigation.

-- 
| G r e g  L o u i s         | gpg public key: 0x400B1AA86D9E3E64 |
|  http://www.bgl.nu/~glouis |   (on my website or any keyserver) |
|  http://wecanstopspam.org in signatures helps fight junk email. |

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-22 22:06     ` Andries Brouwer
@ 2003-09-23  0:23       ` Rob Landley
  2003-09-23 10:51         ` Greg Louis
  2003-09-25 23:59       ` Rob Landley
  2003-09-26  1:27       ` Rob Landley
  2 siblings, 1 reply; 35+ messages in thread
From: Rob Landley @ 2003-09-23  0:23 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: Vojtech Pavlik, linux-kernel

On Monday 22 September 2003 17:06, Andries Brouwer wrote:
> On Mon, Sep 22, 2003 at 03:06:08PM -0500, Rob Landley wrote:

> > Any clues?  (This happens to me at least once an hour...)
>
> Some people have been reporting missing key releases (maybe also you),
> but these are all missing key presses. It is easiest to blame the
> keyboard, even though I could imagine ways to blame the kernel.
>
> What about 2.4? Do you have to go back once an hour and add a symbol
> that was not transmitted correctly? Does 2.4 work perfectly for you?

2.4 worked for me when I used it.  I haven't booted 2.4 in weeks, but I don't 
remember losing any keypresses under 2.4.  (I suppose I could fire up a 2.4 
kernel and answer email with it for a couple hours, though...)

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-22 20:06   ` Rob Landley
@ 2003-09-22 22:06     ` Andries Brouwer
  2003-09-23  0:23       ` Rob Landley
                         ` (2 more replies)
  0 siblings, 3 replies; 35+ messages in thread
From: Andries Brouwer @ 2003-09-22 22:06 UTC (permalink / raw)
  To: Rob Landley; +Cc: Vojtech Pavlik, linux-kernel

On Mon, Sep 22, 2003 at 03:06:08PM -0500, Rob Landley wrote:

> Here's a cut and paste from a tail of /var/log/messages:
> 
> Sep 22 14:12:26 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x94, 
> on isa0060/serio0) pressed.
> Sep 22 14:12:26 localhost kernel: i8042 history: 18 98 18 98 26 a6 39 b9 1f 9f 
> 15 1f 95 9f 12 94

ool sys e <t release>

> Sep 22 14:22:05 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xa6, 
> on isa0060/serio0) pressed.
> Sep 22 14:22:05 localhost kernel: i8042 history: 94 15 95 39 14 b9 18 94 98 39 
> b9 23 12 a3 92 a6

y to he <l release>

>From "try to help"? The l press was lost.

> Sep 22 14:39:06 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1cb, 
> on isa0060/serio0) pressed.
> Sep 22 14:39:06 localhost kernel: i8042 history: cb e0 4b e0 cb e0 4b e0 cb e0 
> 4b e0 cb 4b e0 cb

[Left] [Left] [Left] KP-4 <[Left] release>

This is a rather clear one. You press the grey [Left] key four times.
It has make code e0 4b and break code e0 cb.
Three times we get e0 4b e0 cb but once we get 4b e0 cb.
Surely, the e0 was lost here.

> Sep 22 14:58:05 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
> on isa0060/serio0) pressed.
> Sep 22 14:58:05 localhost kernel: i8042 history: 1c 90 9c e0 48 e0 c8 23 0f a3 
> 8f 1c 9c 50 e0 d0

[Enter] [Up] h [Tab] [Enter] KP-2 <[Down] release>

Rather similar. The e0 part of e0 50 was lost.

> Sep 22 14:59:46 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xb1, 
> on isa0060/serio0) pressed.
> Sep 22 14:59:46 localhost kernel: i8042 history: 2d ad 2d ad 17 97 19 99 1e 15 
> 9e 95 17 97 22 b1

xxipayig <n release>

You were typing "paying" but the n press was not seen.

> Any clues?  (This happens to me at least once an hour...)

Some people have been reporting missing key releases (maybe also you),
but these are all missing key presses. It is easiest to blame the
keyboard, even though I could imagine ways to blame the kernel.

What about 2.4? Do you have to go back once an hour and add a symbol
that was not transmitted correctly? Does 2.4 work perfectly for you?


^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-21 10:04 ` Vojtech Pavlik
  2003-09-21 20:20   ` Rob Landley
@ 2003-09-22 20:06   ` Rob Landley
  2003-09-22 22:06     ` Andries Brouwer
  1 sibling, 1 reply; 35+ messages in thread
From: Rob Landley @ 2003-09-22 20:06 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

On Sunday 21 September 2003 05:04, Vojtech Pavlik wrote:
> On Sat, Sep 20, 2003 at 04:33:22PM -0400, Rob Landley wrote:
> > I've mentioned my keyboard repeat problems before.  I grepped through the
> > logs and found a whole bunch of these type messages:
> >
> > Aug 17 05:28:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0x1d0, on isa0060/serio0) pressed.
> > Aug 19 09:06:51 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0x8e, on isa0060/serio0) pressed.
> > Aug 22 04:33:36 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0xcd,
> >
> > (There's more, it just goes on and on...)
> >
> > Any clues?  (Thinkpad iSeries...  1200, I think.)
>
> What kernel version? Can you test with latest?

So the key repeat problem is still happening with -test5-mm4.  I just had the 
return key stick on me, and gave it a good three seconds to stop repeating 
before I hit another key.  It didn't.   I noticed after it stuck that the 
next time I hit it, it didn't register.  (I unstuck it with the up-arrow key, 
I believe.  But after that one failure to press, it seemed to be nicely 
reset...)

Here's a cut and paste from a tail of /var/log/messages:

Sep 22 12:50:38 localhost  -- landley[984]: LOGIN ON tty1 BY landley
Sep 22 13:24:05 localhost kernel: spurious 8259A interrupt: IRQ7.
Sep 22 14:12:26 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x94, 
on isa0060/serio0) pressed.
Sep 22 14:12:26 localhost kernel: i8042 history: 18 98 18 98 26 a6 39 b9 1f 9f 
15 1f 95 9f 12 94
Sep 22 14:22:05 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xa6, 
on isa0060/serio0) pressed.
Sep 22 14:22:05 localhost kernel: i8042 history: 94 15 95 39 14 b9 18 94 98 39 
b9 23 12 a3 92 a6
Sep 22 14:39:06 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1cb, 
on isa0060/serio0) pressed.
Sep 22 14:39:06 localhost kernel: i8042 history: cb e0 4b e0 cb e0 4b e0 cb e0 
4b e0 cb 4b e0 cb
Sep 22 14:58:05 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Sep 22 14:58:05 localhost kernel: i8042 history: 1c 90 9c e0 48 e0 c8 23 0f a3 
8f 1c 9c 50 e0 d0
Sep 22 14:59:46 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xb1, 
on isa0060/serio0) pressed.
Sep 22 14:59:46 localhost kernel: i8042 history: 2d ad 2d ad 17 97 19 99 1e 15 
9e 95 17 97 22 b1
Sep 22 15:02:11 localhost su(pam_unix)[1649]: session opened for user root by 
landley(uid=500)

Any clues?  (This happens to me at least once an hour...)

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-21 10:04 ` Vojtech Pavlik
@ 2003-09-21 20:20   ` Rob Landley
  2003-09-22 20:06   ` Rob Landley
  1 sibling, 0 replies; 35+ messages in thread
From: Rob Landley @ 2003-09-21 20:20 UTC (permalink / raw)
  To: Vojtech Pavlik; +Cc: linux-kernel

On Sunday 21 September 2003 06:04, Vojtech Pavlik wrote:
> On Sat, Sep 20, 2003 at 04:33:22PM -0400, Rob Landley wrote:
> > I've mentioned my keyboard repeat problems before.  I grepped through the
> > logs and found a whole bunch of these type messages:
> >
> > Aug 17 05:28:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0x1d0, on isa0060/serio0) pressed.
> > Aug 19 09:06:51 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0x8e, on isa0060/serio0) pressed.
> > Aug 22 04:33:36 localhost kernel: atkbd.c: Unknown key (set 2, scancode
> > 0xcd,
> >
> > (There's more, it just goes on and on...)
> >
> > Any clues?  (Thinkpad iSeries...  1200, I think.)
>
> What kernel version? Can you test with latest?

Currently 2.6.0-test5-mm2.  It's done it since at least -test3.

Which would be "latest"?  -bk8 or -mm3? :)

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-20 22:18 ` Andries Brouwer
@ 2003-09-21 18:51   ` Rob Landley
  0 siblings, 0 replies; 35+ messages in thread
From: Rob Landley @ 2003-09-21 18:51 UTC (permalink / raw)
  To: Andries Brouwer; +Cc: linux-kernel

On Saturday 20 September 2003 18:18, Andries Brouwer wrote:
> On Sat, Sep 20, 2003 at 04:33:22PM -0400, Rob Landley wrote:
> > I've mentioned my keyboard repeat problems before.  I grepped through the
> > logs and found a whole bunch of these type messages:
> >
> > Aug 17 05:28:48 atkbd.c: Unknown key (set 2, scancode 0x1d0,
> > on isa0060/serio0) pressed.
> > Aug 19 09:06:51 atkbd.c: Unknown key (set 2, scancode 0x8e,
> > on isa0060/serio0) pressed.
>
> ...
>
> These are key releases for keys i8042.c didnt know were down.
> If otherwise your keyboard functions well, this is harmless.

It doesn't.  It's missing key release events left and right.  About twice an 
hour a key well get "stuck".  X is okay once you press a second key, but if a 
VT gets a key stuck, it doesn't come back.

> > Sep  2 13:37:52 atkbd.c: Unknown key (set 0, scancode 0xfc,
> > on isa0060/serio1) pressed.
> > Sep  2 13:37:52 atkbd.c: Unknown key (set 0, scancode 0xfc,
> > on isa0060/serio1) pressed.
>
> I suppose these are error codes from your mouse.
> If so, it is a bug that they ever went to atkbd.c.

The mouse sometimes sticks as well, just like the keyboard.  (Click and the 
button is held down for no reason.)

I suspect the bug is actually in the new input core...

> Andries

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-20 20:33 Rob Landley
  2003-09-20 22:18 ` Andries Brouwer
@ 2003-09-21 10:04 ` Vojtech Pavlik
  2003-09-21 20:20   ` Rob Landley
  2003-09-22 20:06   ` Rob Landley
  1 sibling, 2 replies; 35+ messages in thread
From: Vojtech Pavlik @ 2003-09-21 10:04 UTC (permalink / raw)
  To: Rob Landley; +Cc: linux-kernel

On Sat, Sep 20, 2003 at 04:33:22PM -0400, Rob Landley wrote:
> I've mentioned my keyboard repeat problems before.  I grepped through the logs 
> and found a whole bunch of these type messages:
> 
> Aug 17 05:28:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
> on isa0060/serio0) pressed.
> Aug 19 09:06:51 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x8e, 
> on isa0060/serio0) pressed.
> Aug 22 04:33:36 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcd, 
> 
> (There's more, it just goes on and on...)
> 
> Any clues?  (Thinkpad iSeries...  1200, I think.)

What kernel version? Can you test with latest?

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

^ permalink raw reply	[flat|nested] 35+ messages in thread

* Re: Keyboard oddness.
  2003-09-20 20:33 Rob Landley
@ 2003-09-20 22:18 ` Andries Brouwer
  2003-09-21 18:51   ` Rob Landley
  2003-09-21 10:04 ` Vojtech Pavlik
  1 sibling, 1 reply; 35+ messages in thread
From: Andries Brouwer @ 2003-09-20 22:18 UTC (permalink / raw)
  To: Rob Landley; +Cc: linux-kernel

On Sat, Sep 20, 2003 at 04:33:22PM -0400, Rob Landley wrote:
> I've mentioned my keyboard repeat problems before.  I grepped through the logs 
> and found a whole bunch of these type messages:
> 
> Aug 17 05:28:48 atkbd.c: Unknown key (set 2, scancode 0x1d0, 
> on isa0060/serio0) pressed.
> Aug 19 09:06:51 atkbd.c: Unknown key (set 2, scancode 0x8e, 
> on isa0060/serio0) pressed.
...

These are key releases for keys i8042.c didnt know were down.
If otherwise your keyboard functions well, this is harmless.

> Sep  2 13:37:52 atkbd.c: Unknown key (set 0, scancode 0xfc, 
> on isa0060/serio1) pressed.
> Sep  2 13:37:52 atkbd.c: Unknown key (set 0, scancode 0xfc, 
> on isa0060/serio1) pressed.

I suppose these are error codes from your mouse.
If so, it is a bug that they ever went to atkbd.c.

Andries


^ permalink raw reply	[flat|nested] 35+ messages in thread

* Keyboard oddness.
@ 2003-09-20 20:33 Rob Landley
  2003-09-20 22:18 ` Andries Brouwer
  2003-09-21 10:04 ` Vojtech Pavlik
  0 siblings, 2 replies; 35+ messages in thread
From: Rob Landley @ 2003-09-20 20:33 UTC (permalink / raw)
  To: linux-kernel

I've mentioned my keyboard repeat problems before.  I grepped through the logs 
and found a whole bunch of these type messages:

Aug 17 05:28:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Aug 19 09:06:51 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x8e, 
on isa0060/serio0) pressed.
Aug 22 04:33:36 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcd, 
on isa0060/serio0) pressed.
Aug 22 10:29:45 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x9f, 
on isa0060/serio0) pressed.
Aug 22 10:43:45 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x9f, 
on isa0060/serio0) pressed.
Aug 24 00:29:57 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Aug 26 19:46:44 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcd, 
on isa0060/serio0) pressed.
Aug 26 22:18:52 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1cb, 
on isa0060/serio0) pressed.
Aug 26 22:36:32 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Aug 28 05:18:28 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x97, 
on isa0060/serio0) pressed.
Aug 28 07:28:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Sep  1 19:18:52 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcb, 
on isa0060/serio0) pressed.
Sep  1 19:50:52 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  2 00:52:01 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x94, 
on isa0060/serio0) pressed.
Sep  2 13:37:52 localhost kernel: atkbd.c: Unknown key (set 0, scancode 0xfc, 
on isa0060/serio1) pressed.
Sep  2 13:37:52 localhost kernel: atkbd.c: Unknown key (set 0, scancode 0xfc, 
on isa0060/serio1) pressed.
Sep  2 22:31:25 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xb9, 
on isa0060/serio0) pressed.
Sep  3 04:58:47 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Sep  3 14:26:41 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  3 22:03:33 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  3 22:06:53 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  3 22:27:33 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x9e, 
on isa0060/serio0) pressed.
Sep  4 16:00:48 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x9d, 
on isa0060/serio0) pressed.
Sep  5 00:29:00 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  5 00:42:40 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcd, 
on isa0060/serio0) pressed.
Sep  5 00:43:40 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x92, 
on isa0060/serio0) pressed.
Sep  5 00:52:00 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Sep  5 00:54:40 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1d0, 
on isa0060/serio0) pressed.
Sep  5 01:49:00 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  5 02:24:40 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  5 03:06:00 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xcd, 
on isa0060/serio0) pressed.
Sep  5 14:36:01 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.
Sep  5 14:37:42 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xae, 
on isa0060/serio0) pressed.
Sep  5 15:23:42 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x90, 
on isa0060/serio0) pressed.
Sep  5 15:38:22 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x97, 
on isa0060/serio0) pressed.
Sep  5 16:03:42 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0x1cb, 
on isa0060/serio0) pressed.
Sep  5 18:16:22 localhost kernel: atkbd.c: Unknown key (set 2, scancode 0xd0, 
on isa0060/serio0) pressed.

(There's more, it just goes on and on...)

Any clues?  (Thinkpad iSeries...  1200, I think.)

Rob

^ permalink raw reply	[flat|nested] 35+ messages in thread

end of thread, other threads:[~2003-10-04 12:23 UTC | newest]

Thread overview: 35+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-09-26  9:43 Keyboard oddness Nicolas Mailhot
2003-09-26 10:24 ` Vojtech Pavlik
2003-09-26 10:41   ` Nicolas Mailhot
2003-09-26 10:51     ` Vojtech Pavlik
2003-09-26 12:48     ` jw schultz
2003-09-26 13:08     ` Nicolas Mailhot
2003-09-26 13:41       ` Vojtech Pavlik
2003-09-26 14:12         ` Nicolas Mailhot
2003-09-26 14:17           ` Vojtech Pavlik
2003-09-26 14:21             ` Nicolas Mailhot
2003-09-26 14:26               ` Vojtech Pavlik
2003-09-26 14:50                 ` Nicolas Mailhot
2003-09-26 15:06                   ` Vojtech Pavlik
2003-09-26 15:21                     ` Nicolas Mailhot
2003-09-26 15:24                       ` Vojtech Pavlik
2003-09-26 13:56       ` Helge Hafting
2003-09-26 14:01         ` Mike Dresser
2003-09-28 10:58           ` Helge Hafting
2003-10-02 12:40             ` Pavel Machek
2003-10-01 10:09   ` Pavel Machek
2003-10-01 13:06     ` Nicolas Mailhot
  -- strict thread matches above, loose matches on Subject: below --
2003-09-20 20:33 Rob Landley
2003-09-20 22:18 ` Andries Brouwer
2003-09-21 18:51   ` Rob Landley
2003-09-21 10:04 ` Vojtech Pavlik
2003-09-21 20:20   ` Rob Landley
2003-09-22 20:06   ` Rob Landley
2003-09-22 22:06     ` Andries Brouwer
2003-09-23  0:23       ` Rob Landley
2003-09-23 10:51         ` Greg Louis
2003-09-25 23:59       ` Rob Landley
2003-09-26  1:27       ` Rob Landley
2003-09-26  8:15         ` Andries Brouwer
2003-09-26  8:43           ` Rob Landley
2003-09-26  8:46           ` Vojtech Pavlik

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).