From b4b1054f6cdd08b040fc5132936a5700c1d2d05b Mon Sep 17 00:00:00 2001 From: Tina Zhang Date: Tue, 27 Feb 2024 10:14:37 +0800 Subject: iommu/vt-d: Remove treatment for revoking PASIDs with pending page faults Commit 2f26e0a9c986 ("iommu/vt-d: Add basic SVM PASID support") added a special treatment to mandate that no page faults may be outstanding for the PASID after intel_svm_unbind_mm() is called, as the PASID will be released and reused after unbind. This is unnecessary anymore as no outstanding page faults have been ensured in the driver's remove_dev_pasid path: - Tear down the pasid entry, which guarantees that new page faults for the PASID will be rejected by the iommu hardware. - All outstanding page faults have been responded to. - All hardware pending faults are drained in intel_drain_pasid_prq(). Remove this unnecessary code. Signed-off-by: Tina Zhang Link: https://lore.kernel.org/r/20240219125723.1645703-2-tina.zhang@intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/intel/svm.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'drivers/iommu') diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index 40edd282903f..a815362c8e60 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -408,13 +408,6 @@ void intel_svm_remove_dev_pasid(struct device *dev, u32 pasid) if (svm->notifier.ops) mmu_notifier_unregister(&svm->notifier, mm); pasid_private_remove(svm->pasid); - /* - * We mandate that no page faults may be outstanding - * for the PASID when intel_svm_unbind_mm() is called. - * If that is not obeyed, subtle errors will happen. - * Let's make them less subtle... - */ - memset(svm, 0x6b, sizeof(*svm)); kfree(svm); } } -- cgit v1.2.3