On Mon, Jun 16, 2025 at 07:21:28PM -0700, Nicolin Chen wrote: > On Mon, Jun 16, 2025 at 10:33:05AM -0300, Jason Gunthorpe wrote: > > On Sat, Jun 14, 2025 at 12:14:31AM -0700, Nicolin Chen wrote: > > > @@ -1321,7 +1328,7 @@ int iommufd_access_pin_pages(struct iommufd_access > > > *access, unsigned long iova, > > > > > > /* Driver's ops don't support pin_pages */ > > > if (IS_ENABLED(CONFIG_IOMMUFD_TEST) && > > > - WARN_ON(access->iova_alignment != PAGE_SIZE || !access->ops->unmap)) > > > + WARN_ON(access->iova_alignment != PAGE_SIZE)) > > > return -EINVAL; > > > > I don't want to loose this check, continuing blocking mdevs is still > > important. Only the internal access should be able to use this > > mechanism. > > OK. So, it probably should be > if (IS_ENABLED(CONFIG_IOMMUFD_TEST) && > WARN_ON(access->iova_alignment != PAGE_SIZE || > (access->ictx && !access->ops->unmap)) >
Nit: Probably worth mentioning in a comment that access->ops shouldn't be NULL during selftests and hence we avoid checking for !access->ops? If you feel it's too verbose, you can let be as well. Reviewed-by: Pranjal Shrivastava <pr...@google.com> > Thanks > Nicolin
