Re: [Xen-devel] [PATCH] libxl: do not fail device removal if backend domain is gone

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Fri, Feb 09, 2018 at 02:08:33PM +0100, Marek Marczykowski-Górecki wrote:
> On Fri, Feb 09, 2018 at 12:10:39PM +0000, Roger Pau Monné wrote:
> > On Fri, Feb 09, 2018 at 12:41:58PM +0100, Marek Marczykowski-Górecki wrote:
> > > On Fri, Feb 09, 2018 at 11:27:04AM +0000, Roger Pau Monné wrote:
> > > > I'm also wondering, if you jump to 'out' here, you avoid the call to
> > > > libxl__xs_transaction_commit and instead end up calling
> > > > libxl__xs_transaction_abort, which means the above call to
> > > > libxl__xs_path_cleanup will not be committed to xenstore, is this
> > > > really desired?
> > > >
> > > > It seems to me libxl might leak xenstore frontend entries in that
> > > > case.
> > > 
> > > That call is only if aodev->force. In other cases cleanup is done in
> > > device_hotplug_done()->libxl__device_destroy(), which have its own 
> > > transaction.
> > 
> > Hm, right, but this would still be incorrect in the force case then?
> > Or is this simply not needed for the 'force' case?
> 
> In that case, the first libxl__xs_path_cleanup will indeed be aborted.
> But then it will be cleaned up the same way as in !force case.
> Anyway, this is about the case when backend is already gone, so 'force'
> doesn't really change anything - it was forcefully removed already, by
> shutting down the backend domain (or removing backend using something
> else)...

Are you sure? The libxl__xs_path_cleanup call is not for removing the
backend entries but the frontend ones. Ie: the backend entries not
being present doesn't imply the frontend entries also not being
present.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel