Skip to contentSkip to conversation

ALPS psmouse_reset on reconnect confusing Tecra M2

DISCUSSION Mike Waychison 5 participants 11 messages Jul 4, 2005 – Jul 10, 2005 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

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

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Before resume, this is the output I get:

Jul  3 01:25:11 lapdance kernel: alps.c: E6 report: 00 00 64
Jul  3 01:25:11 lapdance kernel: alps.c: E7 report: 22 02 14
Jul  3 01:25:11 lapdance kernel: alps.c: E6 report: 00 00 64
Jul  3 01:25:11 lapdance kernel: alps.c: E7 report: 22 02 14
Jul  3 01:25:12 lapdance kernel: alps.c: Status: 05 01 0a
Jul  3 01:25:12 lapdance kernel: alps.c: Enabling hardware tapping
Jul  3 01:25:12 lapdance kernel: alps.c: Status: 05 01 0a
Jul  3 01:25:12 lapdance kernel: input: DualPoint Stick on isa0060/serio1
Jul  3 01:25:12 lapdance kernel: input: AlpsPS/2 ALPS DualPoint TouchPad

on isa0060/serio1

If I then suspend and resume:

Jul  3 01:26:13 lapdance kernel: alps.c: E6 report: 00 00 64
Jul  3 01:26:13 lapdance kernel: alps.c: E7 report: 73 02 0a
Jul  3 01:26:14 lapdance kernel: alps.c: E6 report: 00 00 64
Jul  3 01:26:14 lapdance kernel: alps.c: E7 report: 73 02 0a
Jul  3 01:26:14 lapdance kernel: alps.c: Status: 15 01 0a
Jul  3 01:26:14 lapdance kernel: alps.c: Enabling hardware tapping
Jul  3 01:26:14 lapdance kernel: alps.c: Status: 15 01 0a
Jul  3 01:26:14 lapdance kernel: input: PS/2 Mouse on isa0060/serio1
Jul  3 01:26:14 lapdance kernel: input: AlpsPS/2 ALPS GlidePoint on

isa0060/serio1

The pointer is confused, and a hard shutdown and boot is required to restore it. A reboot alone won't do the trick.

FWIW, I was using the regular IMPS/2 X mouse driver, though I later tried using the synaptics X driver and that didn't work either.

Other things I've tried include trying psmouse_reset until it returns successfully, though sometimes it did, and other times it didn't, even after 100 retries. In either case the pointer didn't come back.

Removing the psmouse_reset makes the pointer come back properly 100% of the time.

Are you sure that the psmouse_reset is really the right thing to do?

(worked fine in 2.6.11-rc2).

Thanks,

Mike Waychison
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCyaaadQs4kOxk3/MRAqvBAKCeaYgsMFPrubb+7p8Se/Z6BwEmcwCfXPOB yVQkUCORLUKyGGO5ttnfjIs=
=B1O/
-----END PGP SIGNATURE-----

#1 lore raw
Quoted 9 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

Quoted 1 line

Are you sure that the psmouse_reset is really the right thing to do?

Yes, it helps on some other laptops.

Signature · 1 line
Dmitry
#3 lore raw
In reply to Dmitry Torokhov

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

Quoted 10 lines · Dmitry Torokhov
Quoted 8 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

Quoted 10 lines
Quoted 1 line

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

Yup. This did the trick :D

Thanks,

Mike Waychison
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCyrfldQs4kOxk3/MRAs6CAJwMeMy8uZK3wVxwptihGtyRQUKVPQCaAnSO XP1TQKemZRAbwiy/0UovARM=
=YVQS
-----END PGP SIGNATURE-----

#4 lore raw
In reply to Dmitry Torokhov
Quoted 10 lines · Dmitry Torokhov
Quoted 8 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

Quoted 9 lines
Quoted 1 line

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

This fixed the problem for mylaptop, but ONLY if I had #define DEBUG in alps.c, switch it back to the usual #undef and the exact same problem happens. I've got no idea what's going on, I'll poke at it more when I'm awake....

Signature · 3 lines
Michael Marineau
[email protected]
Oregon State University
#5 lore raw
Quoted 18 lines · Micheal Marineau
Quoted 17 lines · Dmitry Torokhov
Quoted 9 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

Quoted 8 lines
Quoted 2 lines

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

This fixed the problem for mylaptop, but ONLY if I had #define DEBUG in alps.c, switch it back to the usual #undef and the exact same problem happens. I've got no idea what's going on, I'll poke at it more when I'm awake....

Hi,

Sorry, any update on this topic? Does it still only work with DEBUG?

Thanks!

Signature · 1 line
Dmitry
#6 lore raw
Quoted 32 lines · Dmitry Torokhov
Quoted 20 lines · Micheal Marineau
Quoted 18 lines · Dmitry Torokhov
Quoted 9 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

Quoted 8 lines
Quoted 2 lines

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

This fixed the problem for mylaptop, but ONLY if I had #define DEBUG in alps.c, switch it back to the usual #undef and the exact same problem happens. I've got no idea what's going on, I'll poke at it more when I'm awake....

Quoted 5 lines

Hi,

Sorry, any update on this topic? Does it still only work with DEBUG?

Thanks!

I have no clue why DEBUG helped, but it seems to have been simply dumb luck as the issue came up again with DEBUG still on. An easy workaround was to compile psmouse as a module and unload/load it during the suspend cycle (automatic with the hibernate script, it has psmouse in it's blacklist). I haven't dug into it to see what exactly doesn't get initilized during resume that should.

Signature · 3 lines
Michael Marineau
[email protected]
Oregon State University

A message that is not in the archive (sent off-list or not yet received): [email protected]

#8 lore raw
Quoted 44 lines · Micheal Marineau
Quoted 35 lines · Dmitry Torokhov
Quoted 22 lines · Micheal Marineau
Quoted 19 lines · Dmitry Torokhov
Quoted 9 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

Quoted 8 lines
Quoted 2 lines

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

This fixed the problem for mylaptop, but ONLY if I had #define DEBUG in alps.c, switch it back to the usual #undef and the exact same problem happens. I've got no idea what's going on, I'll poke at it more when I'm awake....

Quoted 5 lines

Hi,

Sorry, any update on this topic? Does it still only work with DEBUG?

Thanks!

Quoted 8 lines

I have no clue why DEBUG helped, but it seems to have been simply dumb luck as the issue came up again with DEBUG still on. An easy workaround was to compile psmouse as a module and unload/load it during the suspend cycle (automatic with the hibernate script, it has psmouse in it's blacklist). I haven't dug into it to see what exactly doesn't get initilized during resume that should.

Here's a little more info though:
dmesg on boot/re-load module
alps.c: E6 report: 00 00 64
alps.c: E7 report: 63 03 c8
alps.c: E6 report: 00 00 64
alps.c: E7 report: 63 03 c8
alps.c: Status: 10 00 64

dmesg after a resume (no reloading of the module)
alps.c: E6 report: 00 00 64
alps.c: E7 report: 63 03 c8
alps.c: Status: 10 00 64
alps.c: Status: 10 00 64

Signature · 3 lines
Michael Marineau
[email protected]
Oregon State University
#9 lore raw
In reply to Dmitry Torokhov
Quoted 32 lines · Dmitry Torokhov
Quoted 20 lines · Micheal Marineau
Quoted 18 lines · Dmitry Torokhov
Quoted 9 lines · Mike Waychison

Hi,

I just upgrade my Tecra M2 this weekend to the latest GIT tree and noticed that my mouse pointer/touchpad is now broken on resume.

Investigating, it appears that mouse device gets confused due to the introduced psmouse_reset(psmouse) during reconnect:

http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=f3a5c73d5ecb40909db662c4d2ace497b25c5940

Hi,

Please try the following patch:

Quoted 8 lines
Quoted 2 lines

http://www.ucw.cz/~vojtech/input/alps-suspend-typo

This fixed the problem for mylaptop, but ONLY if I had #define DEBUG in alps.c, switch it back to the usual #undef and the exact same problem happens. I've got no idea what's going on, I'll poke at it more when I'm awake....

Quoted 6 lines

Hi,

Sorry, any update on this topic? Does it still only work with DEBUG?

Thanks!

Yey! fixed it, simple little patch, just updates the alps_model_info struct. Here's a link so my mail client won't mess up the white space:
http://dev.gentoo.org/~marineam/files/alps-dell8500-dualpoint.patch

(note, this sill requires the alps-suspend-typo fix)

Signature · 3 lines
Michael Marineau
[email protected]
Oregon State University
#10 lore raw
Quoted 5 lines · Micheal Marineau

Yey! fixed it, simple little patch, just updates the alps_model_info struct. Here's a link so my mail client won't mess up the white space:
http://dev.gentoo.org/~marineam/files/alps-dell8500-dualpoint.patch

(note, this sill requires the alps-suspend-typo fix)

It's hard to believe your patch

drivers/input/mouse/alps.c +1 -1
@@ -48,1 +48,1 @@
-	{ { 0x63, 0x03, 0xc8 }, 0xf8, 0xf8, ALPS_PASS },		/* Dell Latitude D800 */
+	{ { 0x63, 0x03, 0xc8 }, 0xf8, 0xf8, ALPS_PASS | ALPS_DUALPOINT }, /* Dell Latitude D800, Inspiron 8500 */

can fix it, because the ALPS_DUALPOINT constant doesn't affect the initalization behavior at all, only the way how TouchPoint data are decoded.

Signature · 2 lines
Vojtech Pavlik
SuSE Labs, SuSE CR
#11 lore raw
Quoted 18 lines · Vojtech Pavlik
Quoted 5 lines · Micheal Marineau

Yey! fixed it, simple little patch, just updates the alps_model_info struct. Here's a link so my mail client won't mess up the white space:
http://dev.gentoo.org/~marineam/files/alps-dell8500-dualpoint.patch

(note, this sill requires the alps-suspend-typo fix)

It's hard to believe your patch

drivers/input/mouse/alps.c +1 -1
@@ -48,1 +48,1 @@
-	{ { 0x63, 0x03, 0xc8 }, 0xf8, 0xf8, ALPS_PASS },		/* Dell Latitude D800 */
+	{ { 0x63, 0x03, 0xc8 }, 0xf8, 0xf8, ALPS_PASS | ALPS_DUALPOINT }, /* Dell Latitude D800, Inspiron 8500 */
Quoted 4 lines

can fix it, because the ALPS_DUALPOINT constant doesn't affect the initalization behavior at all, only the way how TouchPoint data are decoded.

I don't really understand what's going on with that. It's not the root issue for sure, just now after a run of about 15 suspends the mouse finally capped out. Somehow that made it slightly more reliable? Or it was just dumb luck that I couldn't reproduce it yesterday, but I managed to do it once today.

My best guess is that something more needs to be done to priv->dev2 to make sure it initilizes properly but I don't know what, I'm having some difficulty understanding the alps driver.

But it could be something independent of resume, I have a gentoo user with another inspiron 8500 who says even with 2.6.12, the dualpoint stick doesn't initilize a lot of the time even on boot.

Signature · 3 lines
Michael Marineau
[email protected]
Oregon State University