#11171 NORM 1.75-ha: XO-1.75 touchpad jumps by X or Y coordinate

[Zarro Boogs per Child]

#11171: XO-1.75 touchpad jumps by X or Y coordinate
---------------------------------+------------------------------------------
           Reporter:  Quozl      |       Owner:  wad          
               Type:  defect     |      Status:  new          
           Priority:  normal     |   Milestone:  1.75-hardware
          Component:  hardware   |     Version:  1.75-B1      
         Resolution:             |    Keywords:               
        Next_action:  never set  |    Verified:  0            
Deployment_affected:             |   Blockedby:               
           Blocking:             |  
---------------------------------+------------------------------------------

Comment(by carrott):

 I have applied the 00fsp.tgz fix to Kid, an XO-1.75 C1 with Sentelic
 finger sensing pad and build os21 firmware Q4C08 and find the results to
 be quite mixed. It does seem to solve this problem but introduces another
 one.

 On first boot, before a suspend-resume, the jumpy cursor is easily
 reproducible by rolling a finger on the touchpad as described in in an
 earlier comment on this ticket. This is expected because this fix does not
 do anything until a resume occurs.

 Is this fix meant to help #11486 or maybe #11379?

 Suspend-resume performance with a finger on the touchpad is noticeably
 worse with this fix, indeed sometimes it is impossible to use the touch
 pad after a resume. On one resume cycle, the bottom third of the touch pad
 stopped working altogether. In every case so far, leaving the touch pad
 alone for a few seconds seems to return it to normal.

 Having said all this, it DOES seem to have fixed the jumpy touch pad!!
 After a resume cycle, the touch pad behave well, with no jumping and only
 a very small perceived error between what the finger and the pointer are
 doing. The jumping while rolling the finger appears to still be there but
 be much much smaller -- smaller than the size of the icons in the spiral
 home screen on a standard os21 install.

 Resume seems to follow this sequence of events:
 1. Wait for suspend[[BR]]
 2. Place finger on touchpad and move it[[BR]]
 3. Resume starts[[BR]]
 4. A short time later the pointer starts following the finger[[BR]]
 5. A shorter time later the pointer stops (I speculate that the post
 resume script is running/has run at this point)[[BR]]
 6. A short time after that the pointer follows the finger once more but
 now extremely badly[[BR]]
 7. If the user tires of this, and leaves the touchpad for a 10 seconds,
 the touchpad behaves much more reasonably on return to use[[BR]]

 Causing the resume to occur with the keyboard and waiting a few seconds
 before touching the touchpad seems to yield good behavior so I speculate
 that having a finger on the touchpad during the reset is causing a
 problem.

 I have repeated this across 3 reboots and confirmed that that each time

 The touchpad is jumpy on first boot[[BR]]
 After resume by finger on the touchpad the touchpad behaves extremely
 badly[[BR]]
 After a 10 second rest, the touchpad behaves well for the remainder of the
 session (several minutes tested so far)[[BR]]

 I have not attempted to discover the shortest rest time or the shortest
 time after a resume via keyboard before the touchpad behaves well.

-- 
Ticket URL: <http://dev.laptop.org/ticket/11171#comment:18>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system