Re: [Xen-devel] [PATCH v3 4/8] x86/vm_event/monitor/cr: check for vm-event subscriber on domctl

[Corneliu ZUZU <czuzu@xxxxxxxxxxxxxxx>]

On 7/6/2016 7:01 PM, Jan Beulich wrote:
> > > > On 06.07.16 at 17:52, <czuzu@xxxxxxxxxxxxxxx> wrote:
> > Also adjust returned error code for similar check from -EINVAL to more
> > descriptive -ENOSYS (XEN_DOMCTL_MONITOR_OP_EMULATE_EACH_REP).
> I'm not sure that's more descriptive, and we really try to not use
> ENOSYS for other than unimplemented hypercalls. EOPNOTSUPP
> perhaps?
>
> Jan
Well,  it's not quite an 'unsupported operation' and neither does the 
toolstack user communicate an 'invalid value', he must just be noticed 
that something (the vm-event subsystem) needs to be initialised before 
the operation can be done.
But I agree ENOSYS is not acceptable either (I only now see it signifies 
"Function not implemented", my bad for not peeking at that before using 
it, I expected differently).
What about ENODEV, i.e. "No such device."? We need an error code saying 
"Device not initialised"...
Thanks,
Corneliu.

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