Nicolas Souchu (firstname.lastname@example.org)
Fri, 11 Dec 1998 23:12:52 +0000
On Wed, Dec 09, 1998 at 10:00:18PM +0100, Andrea Arcangeli wrote:
>On Wed, 9 Dec 1998, Nicolas Souchu wrote:
>>Couldn't the IEEE1284.3 standard be implemented as a regular driver? It
>Hmm, for the idea I have of IEEE1284.3 no, there are many issues for every
>compliant pardevice (with irqs and other things). But I have not read it
>in details and I have only an old draft of IEEE1284.3 to read so I could
>be wrong here :(.
So, if you consider devices are not totally compliant, which is mostly
the case I believe, how to you adapt your parport stack? If you
give a more thin granularity to your interfaces, you can change your stack
for device particularities.
Anyway, parallel port problems will disappear with USB. A good point for
users, unemployment for developers :)
>>would just share the port with others, at least with drivers that
>>operate with the terminal peripheral of the daisy chain.
>My point is exactly the daisy chain. Right now it' s the pardevice that
>handle the switch considering all other devices as legacy devices. This
>works fine because we do all in software and we don' t care at all about
>IEEE or legacy devices. If you want to take advantage of the irq chaining
>of IEEE1284.3 instead we should do something of more high level in parport
That's what I was suggesting.
>I think... This is a bit sad, since parport scheduling is perfect right
>now and will instead need to change it. For IEEE1284 the change seems
>right and good though. For 1284.3 it' s more tricky.
Right, and each time you add improvments to parport you break the toy.
That's why I've choosen to implement 1284 support at the top of the existing
-- email@example.com / firstname.lastname@example.org FreeBSD - Turning PCs into workstations - http://www.FreeBSD.org
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:53 EST