DragonFly BSD
DragonFly kernel List (threaded) for 2004-03
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: My first real attempt at kernel hacking!


From: Chris Pressey <cpressey@xxxxxxxxxxxxxxx>
Date: Sat, 6 Mar 2004 00:35:00 -0800

On Fri, 5 Mar 2004 22:49:11 -0800 (PST)
Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx> wrote:

> :  http://catseye.webhop.net/projects/belld/
> :
> :-Chris
> 
>    I like the idea!

Yay!  Glad that the idea has some appeal.  Also glad you're feeling
better.

>                     I'll take a look at the patch tomorrow.  One thing
>    I might do (if the code doesn't already do it) is limit the number
>    of bells that can build up in the device driver to something
>    reasonable, like 2.
>
> 					-Matt
> 					Matthew Dillon 
> 					<dillon@xxxxxxxxxxxxx>

Yes - I didn't intend for it to be that way, but that's approximately
how it worked out.  I think because there's really no queue, only a flag
- so if the syscons bell is rung three times in quick succession:

	- the flag is set
	- belld reads it
	- belld starts playing the sound
	- the flag is cleared
	- the flag is set again
	- belld is still playing the sound
	- the flag is set again (no change)
	- belld eventually finishes playing the sound
	- belld reads the flag
	- belld plays the sound again

What does need to be done on it: the flag needs to be set in a better
place in the code (with maybe an option to syscons to enable it?)  Some
consideration might be given to *what* byte syscons should place on
/dev/consolectl when a bell happens.... Also, Hiten pointed out to me
that belld could use a kqueue instead of reading /dev/consolectl.  For
that matter, the 'wait_on' port could be used for this purpose, and
belld could be turned into a simple shell script (FWIW.)

-Chris



[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]