[BRLTTY] Brailliant BI sticky input

[S. Massy]

On Tue, Mar 24, 2015 at 01:29:45PM -0400, Stéphane Doyon wrote:
> I'm having a strange and annoying issue with my Brailliant BI 40, where
> brltty seems to think some navigation keys get stuck.
> I wanted to ask about known / frequent issues before trying the hardware
> repair route.
> 
> I can't repro this on demand. It happens usually a couple times a day,
> sometimes not at all sometimes 5-6 times. It'll start repeating a window up
> or down command, continuously making the bounce sound, as if it never gets
> the release event. The physical key is not stuck though. At that point none
> of the keys appear to do anything any more, but the braille output part
> still works fine.
> 
> I usually power cycle the display and reconnect, but it's usually not
> detected quickly, so I usually end up killing brltty too and then
> reconnecting. (In fact I feel there's some kind of slowness / flakiness with
> USB initial detection/connection / re-connection, but I can't quite give
> very specific facts about the problem.)
> 
> I can't quite tell whether there's some glitch in brltty's state keeping, in
> the display's firmware, in the USB communication in between...
> 
> Any known problems along those lines? Useful investigation tips?
I've never actually seen the specific issue you describe with keys being
stuck, but I have definitely felt the pains related to slow detection
and reconnection. I've banged at this problem on and off and never came
away very successful. At this point, my suspicion is as follows:

The Brailliant BI gets detected as a modem (this happens both on Linux
and Mac OS X 10.6). I'm not a big Mac user, so I never bothered to
investigate the Mac OS side; I however, found that blacklisting the
cdc-acm module on Linux helps mitigate the problem, without fully
solving it. 

I've also come to suspect that there is some kind of bug in the actual
display firmware, possibly related to auto-poweroff. This bug seems to
happen after a sufficient number of poweroff/poweron/reconnect cycles and
causes loss of response to any input events whilst maintaining braille
output. Performing a device reset usually solves this problem. This bug
does not seem to occur with the auto-poweroff option disabled. 

I know this is not tremendously helpful, but, hopefully, those are some
pointers in the right direction. 

Let me know if I can help any further.

Cheers,
S.M.
-- 
Sebastien Massy
Montreal, Canada
Website: http://www.wolfdream.ca
Twitter: http://www.twitter.com/SMassy1
LinkedIn: http://ca.linkedin.com/pub/sebastien-massy/47/5a5/81a
Google+: https://plus.google.com/113474861330993124986/posts