[PATCH 2.6.21-rc1] ibmebus: Support dynamic addition and removal of adapters

Joachim Fenkes FENKES at de.ibm.com
Wed Feb 21 09:36:16 EST 2007


John Rose <johnrose at austin.ibm.com> wrote on 20.02.2007 12:25:06:

> > If the probe operation succeeds, the respective device will show up 
> > beneath
> > /sys/bus/ibmebus/devices.
> 
> This approach is not particularly synchronous.  Take the case of an add
> failure: how long would an application wait before deciding that the new
> device is not going to appear?
>
> It might be preferable to have the write to the probe function fail. 
> For example, if we can't find the device node, return -EINVAL or
> something.

Can be done, sure. Just a matter of putting some more sophisticated error
handling into my code. Patch follows in a bit.
 
> > This is already taken care of by of_device_register().
> 
> If we are scrapping the use of of_device for ibmebus devices, can we
> still create a devspec file for OF correlation purposes?

We don't. My recent patch only changes the fake root device.
Ibmebus-based devices will still be of_devices, of course.

Regards,
  Joachim

---
Joachim Fenkes  --  eHCA Linux Driver Developer and Hardware Tamer
IBM Deutschland Entwicklung GmbH  --  Dept. 3627 (I/O Firmware Dev. 2)
Schoenaicher Strasse 220  --  71032 Boeblingen  --  Germany
eMail: fenkes at de.ibm.com  --  Phone: +49 7031 16 1239 

Sitz der Gesellschaft: Böblingen  --  Geschftsfhrung: Herbert Kircher
Vorsitzender des Aufsichtsrats: Johann Weihen
Registergericht: Amtsgericht Stuttgart, HRB 243294



More information about the Linuxppc-dev mailing list