Re: [Xen-devel] Re: [Xen-API] [PATCH 2 of 4] xc: split xc non-upstream bindings into xcext module

[Ian Campbell <Ian.Campbell@xxxxxxxxxxxxx>]

On Fri, 2010-11-19 at 17:43 +0000, Ian Jackson wrote:
> Ian Campbell writes ("[Xen-devel] Re: [Xen-API] [PATCH 2 of 4] xc: split xc 
> non-upstream bindings into xcext module"):
> >       * Add functionality to libxc to allow it to dlopen a backend (e.g.
> >         pointed to by an envvar) containing the hook implementation with
> >         explicit calls to the layer as necessary.
> 
> I think this is the best approach.  You could put the hook function
> pointers in the xc handle struct.

I've currently made them global (and it works ok for the cases I've
tried) but I think adding them to the xc_handle would indeed be better.

Next I need to decide what to do with the event channel interfaces in
libxenctrl which also need this treatment but which currently use a
straight integer as the handle.

I shall probably convert the evtchn interfaces to use an xc_interface
style opaque pointer as a handle, possibly/probably even reusing the
existing xc_handle type in some way since it already contains broadly
the right set of stuff.

> > Probably the second two are pretty much equivalent modulo the name of
> > the environment variable being either LD_PRELOAD or something else.
> 
> LD_PRELOAD is a sledgehammer to crack a nut for this, I think.  Also,
> I generally think that the existence of LD_PRELOAD should not be used
> as an excuse for not providing a properly-supported interface.

Agreed.

Ian.


_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api