Re: [Xen-devel] [PATCH 2/2] libxl: fix stale timeout event callback race

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Jim Fehlig writes ("Re: [Xen-devel] [PATCH 2/2] libxl: fix stale timeout event 
callback race"):
> Yes, after thinking about it some more, I agree.
> 
> As for the modify callback, do you agree that it is fine to ignore the
> timeval parameter and just update the timer in libvirt's event loop to
> fire immediately?  I.e. always treat the timeval parameter as containing
> {0,0} regardless of "old" or "new" libxl?

Yes.  That is fine.

The reason is that old libxl doesn't ever call this function and new
libxl always calls it with {0,0}.  If you're worried about this, add
an assertion :-).  But it's theoretical.

> I think my patch here is correct
> http://lists.xen.org/archives/html/xen-devel/2012-12/msg00985.html

Having thought about it, I agree, provided that you also fix the
potential integer overflow in libxlTimeoutRegisterEventHook.

> > Right.  Well then yes you need to call virEventRemoveTimeout here.
> > But that applies to both old and new libxl.
> 
> Right.
> 
> Thanks for the help with getting this issue resolved!

Thank you and I'm sorry to have caused trouble with my inadequate
analysis.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel