Wei Liu writes ("Re: [PATCH v8 21/21] libxl: add support for migrating HVM guests without a device model"): > On Fri, Nov 06, 2015 at 05:05:55PM +0100, Roger Pau Monne wrote: > > @@ -387,6 +396,7 @@ static void emulator_xenstore_record_done(libxl__egc > > *egc, > > libxl__stream_write_state > > *stream) > > { > > libxl__domain_suspend_state *dss = stream->dss; > > + STATE_AO_GC(stream->ao); > > > > This is redundant.
`Unneeded' calls to STATE_AO_GC (ie ones in functions which do not actually use anything from ao or gc) are at worst harmless. In fact, such uses of STATE_AO_GC provide additional safety, because they call libxl__ao_inprogress_gc which checks that the ao is in the right state. So I think there is nothing wrong with adding this call here (even if it is a left-over addition from a previous version or draft of this patch, or from wanting to add some calls to LOG* for debugging). There would also be nothing wrong with not adding it. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel