> From: Nicolin Chen <nicolinc@xxxxxxxxxx> > Sent: Monday, October 23, 2023 8:18 AM > > On Sat, Oct 21, 2023 at 01:38:04PM -0300, Jason Gunthorpe wrote: > > On Fri, Oct 20, 2023 at 11:59:13AM -0700, Nicolin Chen wrote: > > > On Fri, Oct 20, 2023 at 10:55:01AM -0300, Jason Gunthorpe wrote: > > > > On Fri, Oct 20, 2023 at 02:43:58AM +0000, Tian, Kevin wrote: > > > > > > > > > But the user shouldn't assume such explicit consistency since it's not > > > > > defined in our uAPI. All we defined is that the attaching may > > > > > fail due to incompatibility for whatever reason then the user can > > > > > always try creating a new hwpt for the to-be-attached device. From > > > > > this regard I don't see providing consistency of result is > > > > > necessary. 😊 > > > > > > > > Anyhow, OK, lets add a comment summarizing your points and remove > the > > > > cc upgrade at attach time (sorry Nicolin/Yi!) > > > > > > Ack. I will send a small removal series. I assume it should CC > > > stable tree also? > > > > No, it seems more like tidying that fixing a functional issue, do I > > misunderstand? > > Hmm. Maybe the misunderstanding is mine -- Kevin was asking if > it was already a bug and you answered yes: > https://lore.kernel.org/linux-iommu/20231016115736.GP3952@xxxxxxxxxx/ > currently intel-iommu driver already rejects 1) enforcing CC on a domain which is already attached to non-CC device and 2) attaching a non-CC device to a domain which has enforce_cc. so there is no explorable bug to fix in stable tree. Just logically intel-iommu driver doesn't support enforcing Cc on a domain which is capable of enforce-cc and already has valid mappings. Then it should add proper check to prevent it from being attempted by future usages.
