RE: [Xen-devel] [PATCH] Arch-neutral balloon driver

["Magenheimer, Dan (HP Labs Fort Collins)" <dan.magenheimer@xxxxxx>]

> > +   if (xen_init() < 0)
> > +           return -1;
> > +
> 
> I had some suggestions to obviate this "xen_init()" stuff; 
> you can read them again at  
> http://lists.xensource.com/archives/html/xen-devel/2005-09/ 
> msg00881.html . I see that xen_init() was introduced anyways, but I  
> would not like to see it spread. Instead let's correct the 
> problem it's working around.

While I'm sympathetic to your argument, I'd prefer for xen_init
to be replaced at some point in the future when more major
surgery is done on the drivers.  In the meantime, it has the
advantage that:

1) It has zero impact on Xen/x86.
2) It serves two purposes on Xen/ia64:
   a) ensures that certain xen driver init routines execute
      before other xen driver init routines
   b) returns a no-impact failure if called on a kernel that
      is not running on Xen (is running on bare metal)

It could be argued that we are now "post-3.0", but these are
really just minor cleanups on (pre-)3.0 drivers to get networking
working on Xen/ia64, preferably on 3.0.x.  So I'll be submitting
a patch to add it to netback also.

> > +// for drivers/xen/balloon/balloon.c
> 
> C++ comment

Good point.
 
> Other than that this looks ok to me. I think your mailer 
> line-wrapped  
> the patch though.

Stupid Outlook. :-(  Keir, should I resend as an attachment?

Thanks,
Dan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel