Re: [PATCH for Xen 4.6 1/5] tools/libxl: introduce libxl_socket_bitmap_fill

[Chao Peng <chao.p.peng@linux.intel.com>]

On Mon, Sep 28, 2015 at 04:53:58PM +0200, Dario Faggioli wrote:
> On Mon, 2015-09-28 at 15:12 +0100, Wei Liu wrote:
> > On Mon, Sep 28, 2015 at 07:54:49PM +0800, Chao Peng wrote:
> 
> > > diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
> > > index 5f9047c..5a91687 100644
> > > --- a/tools/libxl/libxl.h
> > > +++ b/tools/libxl/libxl.h
> > >  
> > >  /*
> > > - * LIBXL_HAVE_SOCKET_BITMAP_ALLOC
> > > + * LIBXL_HAVE_SOCKET_BITMAP
> > >   *
> > > - * If this is defined, then libxl_socket_bitmap_alloc exists.
> > > + * If this is defined, then libxl_socket_bitmap_alloc and
> > > + * libxl_socket_bitmap_Fill exist.
> > 
> > _Fill -> _fill.
> > 
> Right.
> 
> However, it seems to me that the function would be better named
> libxl_get_online_sockets() or something like that.
> 
> I see that you want the actual map. For CPUs and NUMA nodes, we do have
> libxl_get_online_{cpus,nodes}(), but they return the number of online
> CPUs and nodes, so for consistency, libxl_get_online_sockets() would
> better (if necessary at some point) behave similarly.

Actually for the similar purpose of libxl_get_online_{cpus,nodes}, we have
libxl__count_physical_sockets() already, which would ideally be called
libxl_get_online_sockets(). Since it's not public so we have liberty to
change it in the future, if everyone agrees.

> 
> What about libxl_get_online_socketmap() ?  This is still a nice (IMO)
> name for what you're after here, and it leaves us room to implement
> libxl_get_online_cpumap() and libxl_get_online_nodemap(), if we'll ever
> need those.
> 
> Thoughs?

I have been strugging with the naming, but libxl_get_online_socketmap() looks
right for me. Thanks for suggestion.

Chao