mouse issues in 2.6.15-rc5-mm series

mouse issues in 2.6.15-rc5-mm series

[Joe Feise]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

[Note: please cc me on answers since I'm not subscribed to the kernel list]

I am experiencing problems with mouse resyncing in the -mm series.
This is a Logitech wheel mouse connected through a KVM.
Symptom: whenever the mouse isn't moved for some seconds, it doesn't
react to movement for a second, and then resyncs. Sometimes, the
resyncing results in the mouse pointer jumping, which as far as I
know is a protocol mismatch.
While searching for reports of similar problems, I came across
Frank Sorenson's post from Nov. 23 (http://lkml.org/lkml/2005/11/23/533).
Like in his case, reverting
input-attempt-to-re-synchronize-mouse-every-5-seconds.patch
resulted in a kernel without this problem.

-Joe





[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: mouse issues in 2.6.15-rc5-mm series

[Marc Koschewski]

* Joe Feise <jfeise@feise.com> [2005-12-23 22:43:32 -0800]:

> [Note: please cc me on answers since I'm not subscribed to the kernel list]
> 
> I am experiencing problems with mouse resyncing in the -mm series.
> This is a Logitech wheel mouse connected through a KVM.
> Symptom: whenever the mouse isn't moved for some seconds, it doesn't
> react to movement for a second, and then resyncs. Sometimes, the
> resyncing results in the mouse pointer jumping, which as far as I
> know is a protocol mismatch.
> While searching for reports of similar problems, I came across
> Frank Sorenson's post from Nov. 23 (http://lkml.org/lkml/2005/11/23/533).
> Like in his case, reverting
> input-attempt-to-re-synchronize-mouse-every-5-seconds.patch
> resulted in a kernel without this problem.
> 
> -Joe

Hi Joe,

read these ones:

http://lkml.org/lkml/2005/11/18/152
http://www.uwsg.iu.edu/hypermail/linux/kernel/0511.3/0019.html

Maybe they help you out.

I didn't get it working in the rc5 -mm tree so far. But non -mm work as good as
they always did.

Marc

Re: mouse issues in 2.6.15-rc5-mm series

[Frank Sorenson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Marc Koschewski wrote:
> * Joe Feise <jfeise@feise.com> [2005-12-23 22:43:32 -0800]:
> 
> 
>>[Note: please cc me on answers since I'm not subscribed to the kernel list]
>>
>>I am experiencing problems with mouse resyncing in the -mm series.
>>This is a Logitech wheel mouse connected through a KVM.
>>Symptom: whenever the mouse isn't moved for some seconds, it doesn't
>>react to movement for a second, and then resyncs. Sometimes, the
>>resyncing results in the mouse pointer jumping, which as far as I
>>know is a protocol mismatch.
>>While searching for reports of similar problems, I came across
>>Frank Sorenson's post from Nov. 23 (http://lkml.org/lkml/2005/11/23/533).
>>Like in his case, reverting
>>input-attempt-to-re-synchronize-mouse-every-5-seconds.patch
>>resulted in a kernel without this problem.
>>
>>-Joe

Joe,

I continue to see the same issues with the resync patch in -mm.  For me,
tapping stops working, and I'm now seeing both the mouse pointer jumping
 as well (a lesser issue for me, so it was probably present earlier as
well).

I switched to using the synaptics X driver as recommended by Dmitry
Torokhov (http://lkml.org/lkml/2005/11/25/130), and that has worked
splendidly for me.  However, that's really just a band-aid, and I'd be
happier with a working resync patch.

> Hi Joe,
> 
> read these ones:
> 
> http://lkml.org/lkml/2005/11/18/152
> http://www.uwsg.iu.edu/hypermail/linux/kernel/0511.3/0019.html
> 
> Maybe they help you out.
> 
> I didn't get it working in the rc5 -mm tree so far. But non -mm work as good as
> they always did.

Same issue for me as well.

Frank
- --
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFDrSmmaI0dwg4A47wRAidbAJ9rR7ZtgavKNw1aEamcdZ8Br/27GwCdEQcc
JHP9CL6TTlbhYm8kyzeXH+w=
=w3cG
-----END PGP SIGNATURE-----

Re: mouse issues in 2.6.15-rc5-mm series

[Dmitry Torokhov]

On Saturday 24 December 2005 01:43, Joe Feise wrote:
> [Note: please cc me on answers since I'm not subscribed to the kernel list]
> 
> I am experiencing problems with mouse resyncing in the -mm series.
> This is a Logitech wheel mouse connected through a KVM.
> Symptom: whenever the mouse isn't moved for some seconds, it doesn't
> react to movement for a second, and then resyncs. Sometimes, the
> resyncing results in the mouse pointer jumping, which as far as I
> know is a protocol mismatch.
> While searching for reports of similar problems, I came across
> Frank Sorenson's post from Nov. 23 (http://lkml.org/lkml/2005/11/23/533).
> Like in his case, reverting
> input-attempt-to-re-synchronize-mouse-every-5-seconds.patch
> resulted in a kernel without this problem.
>

Joe,

Instead of reverting input-attempt-to-re-synchronize-mouse-every-5-seconds
patch could youplease drop the patch below on top of -mm.

Jet me know if your mouse stays synchronized. Thanks!

-- 
Dmitry

 drivers/input/mouse/psmouse-base.c |   17 +++++++++++++++++
 1 file changed, 17 insertions(+)

Index: linux/drivers/input/mouse/psmouse-base.c
===================================================================
--- linux.orig/drivers/input/mouse/psmouse-base.c
+++ linux/drivers/input/mouse/psmouse-base.c
@@ -1029,6 +1029,23 @@ static int psmouse_switch_protocol(struc
 	else
 		psmouse->type = psmouse_extensions(psmouse, psmouse_max_proto, 1);
 
+	/*
+	 * If mouse's packet size is 3 there is no point in polling the
+	 * device in hopes to detect protocol reset - we won't get less
+	 * than 3 bytes response anyhow.
+	 */
+	if (psmouse->pktsize == 3)
+		psmouse->resync_time = 0;
+
+	/*
+	 * Some smart KVMs fake response to POLL command returning just
+	 * 3 bytes and messing up our resync logic, so if initial poll
+	 * fails we won't try polling the device anymore. Hopefully
+	 * such KVM will maintain initially selected protocol.
+	 */
+	if (psmouse->resync_time && psmouse->poll(psmouse))
+		psmouse->resync_time = 0;
+
 	sprintf(psmouse->devname, "%s %s %s",
 		psmouse_protocol_by_type(psmouse->type)->name, psmouse->vendor, psmouse->name);
 

Re: mouse issues in 2.6.15-rc5-mm series

[Joe Feise]

[-- Attachment #1: Type: text/plain, Size: 1156 bytes --]

Dmitry Torokhov wrote on 12/24/05 09:56:

> On Saturday 24 December 2005 01:43, Joe Feise wrote:
>> [Note: please cc me on answers since I'm not subscribed to the kernel list]
>>
>> I am experiencing problems with mouse resyncing in the -mm series.
>> This is a Logitech wheel mouse connected through a KVM.
>> Symptom: whenever the mouse isn't moved for some seconds, it doesn't
>> react to movement for a second, and then resyncs. Sometimes, the
>> resyncing results in the mouse pointer jumping, which as far as I
>> know is a protocol mismatch.
>> While searching for reports of similar problems, I came across
>> Frank Sorenson's post from Nov. 23 (http://lkml.org/lkml/2005/11/23/533).
>> Like in his case, reverting
>> input-attempt-to-re-synchronize-mouse-every-5-seconds.patch
>> resulted in a kernel without this problem.
>>
> 
> Joe,
> 
> Instead of reverting input-attempt-to-re-synchronize-mouse-every-5-seconds
> patch could youplease drop the patch below on top of -mm.
> 
> Jet me know if your mouse stays synchronized. Thanks!
> 


Dmitry,

thanks, that patch works. The mouse stays synchronized.

-Joe



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: mouse issues in 2.6.15-rc5-mm series

[Dmitry Torokhov]

On Saturday 24 December 2005 05:57, Frank Sorenson wrote:
> 
> I continue to see the same issues with the resync patch in -mm.  For me,
> tapping stops working, and I'm now seeing both the mouse pointer jumping
>  as well (a lesser issue for me, so it was probably present earlier as
> well).
> 

Frank,

Does the tapping not work period or it only does not work first time you
try to tap after not touching the pad for more than 5 seconds?

-- 
Dmitry

Re: mouse issues in 2.6.15-rc5-mm series

[Frank Sorenson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dmitry Torokhov wrote:
> On Saturday 24 December 2005 05:57, Frank Sorenson wrote:
> 
>>I continue to see the same issues with the resync patch in -mm.  For me,
>>tapping stops working, and I'm now seeing both the mouse pointer jumping
>> as well (a lesser issue for me, so it was probably present earlier as
>>well).
>>
> 
> Frank,
> 
> Does the tapping not work period or it only does not work first time you
> try to tap after not touching the pad for more than 5 seconds?

The tapping works initially, then stops.  I hadn't put 2+2 together with
the 5-second idle bit, but that seems the likely issue.

I applied that patch you sent out yesterday, and now tapping works and
I'm not seeing the mouse stall/jump problem.  I'm at 21+ hours uptime
now, with no mouse problems, so I think setting the resync_time to 0
looks like the right fix.

Thanks,

Frank
- --
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFDr3QuaI0dwg4A47wRApTrAKCznvvBq5lX66VPggjP6KTtbFQdDgCgmrhs
x+90pceDtoWznfjy6kBGaAs=
=e2Xa
-----END PGP SIGNATURE-----

Re: mouse issues in 2.6.15-rc5-mm series

[Dmitry Torokhov]

On Sunday 25 December 2005 23:40, Frank Sorenson wrote:
> Dmitry Torokhov wrote:
> > On Saturday 24 December 2005 05:57, Frank Sorenson wrote:
> >
> >>I continue to see the same issues with the resync patch in -mm.  For me,
> >>tapping stops working, and I'm now seeing both the mouse pointer jumping
> >> as well (a lesser issue for me, so it was probably present earlier as
> >>well).
> >>
> >
> > Frank,
> >
> > Does the tapping not work period or it only does not work first time you
> > try to tap after not touching the pad for more than 5 seconds?
> 
> The tapping works initially, then stops.  I hadn't put 2+2 together with
> the 5-second idle bit, but that seems the likely issue.
> 
> I applied that patch you sent out yesterday, and now tapping works and
> I'm not seeing the mouse stall/jump problem.  I'm at 21+ hours uptime
> now, with no mouse problems, so I think setting the resync_time to 0
> looks like the right fix.
>

Well, no, that's a band-aid ;) There are still touchpads speaking 4-byte
protocols so I still need to adjust resync code to work with tapping.

-- 
Dmitry

Re: mouse issues in 2.6.15-rc5-mm series

[Dmitry Torokhov]

On Sunday 25 December 2005 23:40, Frank Sorenson wrote:
> Dmitry Torokhov wrote:
> > On Saturday 24 December 2005 05:57, Frank Sorenson wrote:
> >
> >>I continue to see the same issues with the resync patch in -mm.  For me,
> >>tapping stops working, and I'm now seeing both the mouse pointer jumping
> >> as well (a lesser issue for me, so it was probably present earlier as
> >>well).
> >>
> >
> > Frank,
> >
> > Does the tapping not work period or it only does not work first time you
> > try to tap after not touching the pad for more than 5 seconds?
> 
> The tapping works initially, then stops.  I hadn't put 2+2 together with
> the 5-second idle bit, but that seems the likely issue.
> 
> I applied that patch you sent out yesterday, and now tapping works and
> I'm not seeing the mouse stall/jump problem.  I'm at 21+ hours uptime
> now, with no mouse problems, so I think setting the resync_time to 0
> looks like the right fix.
> 

Frank,

Could you please try the patch below and see if it makes tapping work?
Make sure you enable resynching by doing:

	echo -n 5 > /sys/bus/serio/devices/serioX/resync_time

Thanks!

-- 
Dmitry

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
---

 drivers/input/mouse/alps.c         |   11 ++++++-----
 drivers/input/mouse/psmouse-base.c |   27 ++++++++++++++++++++++++---
 2 files changed, 30 insertions(+), 8 deletions(-)

Index: work/drivers/input/mouse/alps.c
===================================================================
--- work.orig/drivers/input/mouse/alps.c
+++ work/drivers/input/mouse/alps.c
@@ -356,19 +356,20 @@ static int alps_poll(struct psmouse *psm
 {
 	struct alps_data *priv = psmouse->private;
 	unsigned char buf[6];
+	int poll_failed;
 
 	if (priv->i->flags & ALPS_PASS)
 		alps_passthrough_mode(psmouse, 1);
 
-	if (ps2_command(&psmouse->ps2dev, buf, PSMOUSE_CMD_POLL | (psmouse->pktsize << 8)))
-		return -1;
-
-	if ((buf[0] & priv->i->mask0) != priv->i->byte0)
-		return -1;
+	poll_failed = ps2_command(&psmouse->ps2dev, buf,
+				  PSMOUSE_CMD_POLL | (psmouse->pktsize << 8)) < 0;
 
 	if (priv->i->flags & ALPS_PASS)
 		alps_passthrough_mode(psmouse, 0);
 
+	if (poll_failed || (buf[0] & priv->i->mask0) != priv->i->byte0)
+		return -1;
+
 	if ((psmouse->badbyte & 0xc8) == 0x08) {
 /*
  * Poll the track stick ...
Index: work/drivers/input/mouse/psmouse-base.c
===================================================================
--- work.orig/drivers/input/mouse/psmouse-base.c
+++ work/drivers/input/mouse/psmouse-base.c
@@ -54,7 +54,7 @@ static unsigned int psmouse_smartscroll 
 module_param_named(smartscroll, psmouse_smartscroll, bool, 0644);
 MODULE_PARM_DESC(smartscroll, "Logitech Smartscroll autorepeat, 1 = enabled (default), 0 = disabled.");
 
-static unsigned int psmouse_resetafter;
+static unsigned int psmouse_resetafter = 5;
 module_param_named(resetafter, psmouse_resetafter, uint, 0644);
 MODULE_PARM_DESC(resetafter, "Reset device after so many bad packets (0 = never).");
 
@@ -850,8 +850,29 @@ static void psmouse_deactivate(struct ps
 
 static int psmouse_poll(struct psmouse *psmouse)
 {
-	return ps2_command(&psmouse->ps2dev, psmouse->packet,
-			   PSMOUSE_CMD_POLL | (psmouse->pktsize << 8));
+	if (ps2_command(&psmouse->ps2dev, psmouse->packet,
+			PSMOUSE_CMD_POLL | (psmouse->pktsize << 8)))
+		return -1;
+
+	switch (psmouse->type) {
+		case PSMOUSE_PS2:
+		case PSMOUSE_PS2PP:
+		case PSMOUSE_IMPS:
+		case PSMOUSE_IMEX:
+		case PSMOUSE_LIFEBOOK:
+		case PSMOUSE_TRACKPOINT:
+/*
+ * If protocol may be used by "tappable" device (touchpad, touchscreen) try to "restore"
+ * tap data from the halfway-discarded packet
+ */
+			if ((psmouse->badbyte & 0x08) == (psmouse->packet[0] & 0x08))
+				psmouse->packet[0] |= psmouse->badbyte & 0x07;
+
+		default:
+			break;
+	}
+
+	return 0;
 }
 
 

Re: mouse issues in 2.6.15-rc5-mm series

[Frank Sorenson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dmitry Torokhov wrote:
> On Sunday 25 December 2005 23:40, Frank Sorenson wrote:
> 
>>Dmitry Torokhov wrote:

>>>Does the tapping not work period or it only does not work first time you
>>>try to tap after not touching the pad for more than 5 seconds?
>>
>>The tapping works initially, then stops.  I hadn't put 2+2 together with
>>the 5-second idle bit, but that seems the likely issue.
>>
>>I applied that patch you sent out yesterday, and now tapping works and
>>I'm not seeing the mouse stall/jump problem.  I'm at 21+ hours uptime
>>now, with no mouse problems, so I think setting the resync_time to 0
>>looks like the right fix.

> Frank,
> 
> Could you please try the patch below and see if it makes tapping work?
> Make sure you enable resynching by doing:
> 
> 	echo -n 5 > /sys/bus/serio/devices/serioX/resync_time

With this patch (on top of 2.6.15-mm1, right?), I see the mouse
stall/jump problem, but tapping appears to continue working.  The
touchpad also seems to be extremely touchy.  I get spurious taps with
very little pressure, and sometimes double-tap will select, then
immediately deselect.

Frank
- --
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFDvio6aI0dwg4A47wRAmx5AJ9caJlziT8MfGdxSf/yzVhEGqxqfgCggE0M
P2wvTs9/Xbw/sXn8bc/Mk0Y=
=oGAz
-----END PGP SIGNATURE-----

Re: mouse issues in 2.6.15-rc5-mm series

[Dmitry Torokhov]

On 1/6/06, Frank Sorenson <frank@tuxrocks.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Dmitry Torokhov wrote:
> > On Sunday 25 December 2005 23:40, Frank Sorenson wrote:
> >
> >>Dmitry Torokhov wrote:
>
> >>>Does the tapping not work period or it only does not work first time you
> >>>try to tap after not touching the pad for more than 5 seconds?
> >>
> >>The tapping works initially, then stops.  I hadn't put 2+2 together with
> >>the 5-second idle bit, but that seems the likely issue.
> >>
> >>I applied that patch you sent out yesterday, and now tapping works and
> >>I'm not seeing the mouse stall/jump problem.  I'm at 21+ hours uptime
> >>now, with no mouse problems, so I think setting the resync_time to 0
> >>looks like the right fix.
>
> > Frank,
> >
> > Could you please try the patch below and see if it makes tapping work?
> > Make sure you enable resynching by doing:
> >
> >       echo -n 5 > /sys/bus/serio/devices/serioX/resync_time
>
> With this patch (on top of 2.6.15-mm1, right?), I see the mouse
> stall/jump problem, but tapping appears to continue working.  The
> touchpad also seems to be extremely touchy.  I get spurious taps with
> very little pressure, and sometimes double-tap will select, then
> immediately deselect.
>

I just want to confirm that when testing the patch you forced the
touchpad into PS/2 mode with psmouse.proto=exps, right?

Could you please enable i8042 debug (echo 1 >
/sys/modules/i8042/parameters/debug), move pointer and do coulple of
taps and touble-taps making sure there is 5 seconds intervals between
events and send me dmesg. Or is the mouse jerky and touchy even while
you using it continuously, without 5 seconds pauses?

--
Dmitry

Re: mouse issues in 2.6.15-rc5-mm series

[Frank Sorenson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Dmitry Torokhov wrote:
> I just want to confirm that when testing the patch you forced the
> touchpad into PS/2 mode with psmouse.proto=exps, right?

Correct

> Could you please enable i8042 debug (echo 1 >
> /sys/modules/i8042/parameters/debug), move pointer and do coulple of
> taps and touble-taps making sure there is 5 seconds intervals between
> events and send me dmesg. Or is the mouse jerky and touchy even while
> you using it continuously, without 5 seconds pauses?

Odd.  Looks like we've got quotes around the names in /sys/module (but
only for those compiled in--actual modules show correctly).

Anyway, I used /sys/module/"i8042"/parameters/debug, and I'll send you
the dmesg in another email.

The mouse is jerkey and touchy both with and without 5 second pauses.  I
only see the complete mouse "stall" after 5-second pauses, though.

Frank
- --
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFDvp59aI0dwg4A47wRAv4aAJ9+J80OvGQ1GMCFEAM/lQUyxx4+awCgjl7o
Y98BW+9mtxLyxclKtsb4qig=
=ON28
-----END PGP SIGNATURE-----

Re: mouse issues in 2.6.15-rc5-mm series

[Frank Sorenson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Frank Sorenson wrote:
> Dmitry Torokhov wrote:
>>>Frank,
>>>
>>>Could you please try the patch below and see if it makes tapping work?
>>>Make sure you enable resynching by doing:
>>>
>>>	echo -n 5 > /sys/bus/serio/devices/serioX/resync_time
> 
> With this patch (on top of 2.6.15-mm1, right?), I see the mouse
> stall/jump problem, but tapping appears to continue working.  The
> touchpad also seems to be extremely touchy.  I get spurious taps with
> very little pressure, and sometimes double-tap will select, then
> immediately deselect.

Sorry, all.  I should have paid closer attention to the protocol it was
using.  This behavior turned out to be the result of the command-line
parsing issue that Sam Ravnborg's patch addresses.

With the touchpad back in the right mode, I see the original behavior
(tapping stops working) with both resync patches.

Frank
- --
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.7 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFDvu+VaI0dwg4A47wRArgiAJ4qxpieqLXvn1cNozB+w8BDlbN2BACg4K7Z
RfjchjnmcSEBUrUao5+lw2A=
=YzdX
-----END PGP SIGNATURE-----