More ULE bugs fixed.

[Eirik Oeverby]

Matthew D. Fuller wrote:
> On Wed, Nov 05, 2003 at 11:28:50AM +0100 I heard the voice of
> Eirik Oeverby, and lo! it spake thus:
> 
>>The second is that mouse messages are actually *lost*, or bogus ones are 
>>being generated. I guess it's the first, making moused or X misinterpret 
>>the messages it gets. Where along the chain it fails I obviously have no 
>>clue. The consequence of this is that when the mouse stops (like in #1) 
>>but then resumes from an entirely different point - be it 10 pixels away 
>>or at the other end of the screen - possibly even generating a button 
>>push (but not necessarily the corresponding button release) message.
> 
> 
> Note that I've had this to a greater or lesser extent for as long as I
> can remember (certainly back to 3.0-CURRENT).  It corresponds with
> syslog'd messages on my xconsole along the lines of:
> 
> Nov  3 12:46:13 mortis kernel: psmintr: out of sync (00c0 != 0000).
> Nov  3 12:46:13 mortis kernel: psmintr: discard a byte (12).
> 
> It's certainly a lot more common (by orders of magnitude) on 5.x in the
> past...   oh, I dunno, year-ish, than it was previously.  I lose mouse
> function for maybe a second, then it squirms itself off somewhere on the
> screen and sends some button press events.
> 
> I'm currently running 5.1-R, the traditional scheduler, a PS/2 mouse with
> no moused.  And since I got them (much more rarely) with earlier
> 5-CURRENT's, and with 4-CURRENT's, etc, I can't see how it's scheduler
> related.
> 

No idea, but I never got messages like the ones you mention, and it has 
absolutely never happened on 4.x or with SCHED_4BSD.
Weirdness. :)

/Eirik

> 
> 
>>When you say you get the bogus mouse events (which I believe you are 
>>saying atleast ;) only during load, I'm immediately thinking that yes, 
>>that might make sense.
> 
> 
> I don't get it only under load; sometimes from flat idle.  However, it's
> usually when I first move the mouse, after it sitting still for a while
> (where 'while' can vary from a few seconds to a few days, of course); it
> hardly ever happens in mid-move.
> 
> 
>