Re: [rtl] bug in rtf handler ??

[Nathan Paul Simons <npsimons@fsmlabs.com>]

On Fri, Mar 02, 2001 at 01:47:59PM +0100, Olaf Petzold wrote:
> Imo it's a bug/design fault- if I need to write to the rtf to kick the thread
> to fill me the buffer I don't need a handler on user read ops.
> 
> To Nathan: Imo to poll() is an access. I ask for new data - if there are new I
> get it from the buffer. If I don't want to get the prepered data I can leave
> the handler - the task is reponsible for filling the rtf with data even is
> there is no callback. Therefore poll() gets only data if the kernel task has
> data. handler callback is asyncronous to the rt task.

	Okay, I'll see if I can get a patch coded up right quick.  I'll
(hopefully) post it to the list in about an hour.