> -----Original Message----- > From: Jan Beulich <[email protected]> > Sent: 19 October 2020 08:30 > To: [email protected] > Cc: 'Julien Grall' <[email protected]>; [email protected]; 'Paul > Durrant' > <[email protected]>; 'Daniel De Graaf' <[email protected]>; 'Ian > Jackson' <[email protected]>; > 'Wei Liu' <[email protected]>; 'Andrew Cooper' <[email protected]>; > 'George Dunlap' > <[email protected]>; 'Stefano Stabellini' <[email protected]> > Subject: Re: [PATCH 2/5] iommu / domctl: introduce XEN_DOMCTL_iommu_ctl > > On 19.10.2020 09:23, Paul Durrant wrote: > >> From: Julien Grall <[email protected]> > >> Sent: 16 October 2020 16:47 > >> > >> On 05/10/2020 10:49, Paul Durrant wrote: > >>> diff --git a/xen/include/public/domctl.h b/xen/include/public/domctl.h > >>> index 791f0a2592..75e855625a 100644 > >>> --- a/xen/include/public/domctl.h > >>> +++ b/xen/include/public/domctl.h > >>> @@ -1130,6 +1130,18 @@ struct xen_domctl_vuart_op { > >>> */ > >>> }; > >>> > >>> +/* > >>> + * XEN_DOMCTL_iommu_ctl > >>> + * > >>> + * Control of VM IOMMU settings > >>> + */ > >>> + > >>> +#define XEN_DOMCTL_IOMMU_INVALID 0 > >> > >> I can't find any user of XEN_DOMCTL_IOMMU_INVALID. What's the purpose > >> for it? > >> > > > > It's just a placeholder. I think it's generally safer that a zero opcode > > value is invalid. > > But does this then need a #define? Starting valid command from 1 > ought to be sufficient? >
Seems harmless enough, and it also seemed the best way since to reserve 0 since this patch doesn't actually introduce any ops. As it has caused so much controversy though, I'll remove it. Paul > Jan
