On Friday 10 November 2006 9:10 am, Peter Wippich wrote:

> did not expect that something like this is around for such a long time 
> without being fixed. 

It needs "someone" to find out what's really going on.


> I'll check against the Montavista kernel 2.6.10 which does not have this 
> problem. If I come up with some better solution I'll post it here. Maybe 
> we have some silicon bug here ?!? 

The serial core has changed a bunch since 2.6.10 and it's not
obvious why those changes affected DaVinci the way they have.

If I were to guess, I'd say it's largely an artifact of a bad
assumption in that serial core code, namely that all 8250 style
can safely be probed for capabilities.  In general it's safer
to let board-specific code declare the chip type, but that's
not possible with the current serial core.  (Sometimes probing
can be better; yes.  But not always; today, there's no option.)

The DaVinci serial port has some specific init sequencing issues
that I'm not sure the serial core obeys.  

- Dave
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source

Reply via email to