This is the mail archive of the ecos-patches@sourceware.org mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: eSCI patch [was Re: Freescale unified driver]


On Thu, Aug 31, 2006 at 04:55:46PM +0200, Ilija Koco wrote:
> Andrew Lunn wrote:
> >> Now I would ask for a permission to add CYGPKG_IO_SERIAL_TTY_TTY2
> >> "/dev/tty3" component in tty.cdl. It will allow for a tty counterpart
> >> for /dev/ser3 (esci has 4 serial drivers and some other chips (now
> >> and/or in future) may have). Imho it wont compromise other eCos
> >> components. Please comment.
> >>     
> >
> > Sure, no problem. Please also add a similar termio device so that its
> > symmetrical at all levels in the serial stack.
> >   
> OK. I'll do it.
> 
> I made the #define as you suggested and it seem to work, but now we have
> another situation. eSCI driver is being silently discarded allthough
> selected by user. In previous situation user had a kind of warning that
> "something is missing" which was somewhat true: including driver is of
> little use without including at least 1 device since there isn't a
> (regular) way to access it.

In CDL you can have a requires statement that at least one device is
enable. Hopefully there will then be a conflict to indicate to the
user something is wrong. The danger is the inferance engine will solve
the conflict by simply enabling one of the devices. The one it picks
may however not be the one the user wants. When using ecosconfig this
will be clear, since it prints what inferences it has made. I don't
use the configtool, so i don't know if it throws up a dialog box
showing its inferences? 

    Andrew


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]