On Wed, 2015-03-18 at 13:21 +0000, Ian Campbell wrote: > On Fri, 2015-03-13 at 10:34 +0000, Wei Liu wrote: > > ... because it is the right place to clean up device model stuffs. > > ... and not devices_destroy_cb because it is the right ... > > (also "stuff"). > > > And the path should use LIBXL_TOOLSTACK_DOMID instead of hardcoded 0. > > Between this and what is in the next patch I think you probably should > refactor into a helper function to get the correct path for a given > domid. > > FWIW I think you might also be able to do: > GCSPRINTF("/local/domain/" #LIBXL_TOOLSTACK_DOMID "/device-model/%d", > domid); > > If you want, although perhaps the intention is for it to eventually not > be a hard-define of 0 and become e.g. a function call or a global > variable reference, in which case best not?
Having reread patch #0 more carefully I suppose the helper would need to take the dm_domid too, which would then rule out that cpp trick. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel