On 2023/10/25 14:46, Tian, Kevin wrote:
From: Liu, Yi L <yi.l.liu@xxxxxxxxx> Sent: Tuesday, October 24, 2023 11:06 PM + +static int iommufd_group_do_replace_paging(struct iommufd_group *igroup, + struct iommufd_hw_pagetable *hwpt) +{ + struct iommufd_hw_pagetable *old_hwpt = igroup->hwpt; + struct iommufd_device *cur; + int rc; + + lockdep_assert_held(&igroup->lock); + + if (hwpt_is_paging(old_hwpt) && hwpt->ioas != old_hwpt->ioas) { + list_for_each_entry(cur, &igroup->device_list, group_item) { + rc = iopt_table_enforce_dev_resv_regions( + &hwpt->ioas->iopt, cur->dev, NULL); + if (rc) + goto err_unresv; + }should be: if (!hwpt_is_paging(old_hwpt) || hwpt->ioas != old_hwpt->ioas) { ...
oh, yes. The original logic is to add resv region when the ioas are different between new and old hwpts. But now, if the old hwpt is not paging, then it's already needed to add resv regions in the ioas. Regards, Yi Liu
