Merge tag 'for-linus-20160216' of git://git.infradead.org/intel-iommu

Pull IOMMU SVM fixes from David Woodhouse:
 "Minor register size and interrupt acknowledgement fixes which only
  showed up in testing on newer hardware, but mostly a fix to the MM
  refcount handling to prevent a recursive refcount issue when mmap() is
  used on the file descriptor associated with a bound PASID"

* tag 'for-linus-20160216' of git://git.infradead.org/intel-iommu:
  iommu/vt-d: Clear PPR bit to ensure we get more page request interrupts
  iommu/vt-d: Fix 64-bit accesses to 32-bit DMAR_GSTS_REG
  iommu/vt-d: Fix mm refcounting to hold mm_count not mm_users

3 files changed, 33 insertions, 8 deletions

diff --git a/drivers/iommu/dmar.c b/drivers/iommu/dmar.c
index 62a400c5ba06..fb092f3f11cb 100644
--- a/drivers/iommu/dmar.c
+++ b/drivers/iommu/dmar.c
@@ -1353,7 +1353,7 @@ void dmar_disable_qi(struct intel_iommu *iommu)
 
 	raw_spin_lock_irqsave(&iommu->register_lock, flags);
 
-	sts =  dmar_readq(iommu->reg + DMAR_GSTS_REG);
+	sts =  readl(iommu->reg + DMAR_GSTS_REG);
 	if (!(sts & DMA_GSTS_QIES))
 		goto end;
 
diff --git a/drivers/iommu/intel-svm.c b/drivers/iommu/intel-svm.c
index 50464833d0b8..d9939fa9b588 100644
--- a/drivers/iommu/intel-svm.c
+++ b/drivers/iommu/intel-svm.c
@@ -249,12 +249,30 @@ static void intel_flush_pasid_dev(struct intel_svm *svm, struct intel_svm_dev *s
 static void intel_mm_release(struct mmu_notifier *mn, struct mm_struct *mm)
 {
 	struct intel_svm *svm = container_of(mn, struct intel_svm, notifier);
+	struct intel_svm_dev *sdev;
 
+	/* This might end up being called from exit_mmap(), *before* the page
+	 * tables are cleared. And __mmu_notifier_release() will delete us from
+	 * the list of notifiers so that our invalidate_range() callback doesn't
+	 * get called when the page tables are cleared. So we need to protect
+	 * against hardware accessing those page tables.
+	 *
+	 * We do it by clearing the entry in the PASID table and then flushing
+	 * the IOTLB and the PASID table caches. This might upset hardware;
+	 * perhaps we'll want to point the PASID to a dummy PGD (like the zero
+	 * page) so that we end up taking a fault that the hardware really
+	 * *has* to handle gracefully without affecting other processes.
+	 */
 	svm->iommu->pasid_table[svm->pasid].val = 0;
+	wmb();
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(sdev, &svm->devs, list) {
+		intel_flush_pasid_dev(svm, sdev, svm->pasid);
+		intel_flush_svm_range_dev(svm, sdev, 0, -1, 0, !svm->mm);
+	}
+	rcu_read_unlock();
 
-	/* There's no need to do any flush because we can't get here if there
-	 * are any devices left anyway. */
-	WARN_ON(!list_empty(&svm->devs));
 }
 
 static const struct mmu_notifier_ops intel_mmuops = {
@@ -379,7 +397,6 @@ int intel_svm_bind_mm(struct device *dev, int *pasid, int flags, struct svm_dev_
 				goto out;
 			}
 			iommu->pasid_table[svm->pasid].val = (u64)__pa(mm->pgd) | 1;
-			mm = NULL;
 		} else
 			iommu->pasid_table[svm->pasid].val = (u64)__pa(init_mm.pgd) | 1 | (1ULL << 11);
 		wmb();
@@ -442,11 +459,11 @@ int intel_svm_unbind_mm(struct device *dev, int pasid)
 				kfree_rcu(sdev, rcu);
 
 				if (list_empty(&svm->devs)) {
-					mmu_notifier_unregister(&svm->notifier, svm->mm);
 
 					idr_remove(&svm->iommu->pasid_idr, svm->pasid);
 					if (svm->mm)
-						mmput(svm->mm);
+						mmu_notifier_unregister(&svm->notifier, svm->mm);
+
 					/* 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.
@@ -507,6 +524,10 @@ static irqreturn_t prq_event_thread(int irq, void *d)
 	struct intel_svm *svm = NULL;
 	int head, tail, handled = 0;
 
+	/* Clear PPR bit before reading head/tail registers, to
+	 * ensure that we get a new interrupt if needed. */
+	writel(DMA_PRS_PPR, iommu->reg + DMAR_PRS_REG);
+
 	tail = dmar_readq(iommu->reg + DMAR_PQT_REG) & PRQ_RING_MASK;
 	head = dmar_readq(iommu->reg + DMAR_PQH_REG) & PRQ_RING_MASK;
 	while (head != tail) {
@@ -551,6 +572,9 @@ static irqreturn_t prq_event_thread(int irq, void *d)
 		 * any faults on kernel addresses. */
 		if (!svm->mm)
 			goto bad_req;
+		/* If the mm is already defunct, don't handle faults. */
+		if (!atomic_inc_not_zero(&svm->mm->mm_users))
+			goto bad_req;
 		down_read(&svm->mm->mmap_sem);
 		vma = find_extend_vma(svm->mm, address);
 		if (!vma || address < vma->vm_start)
@@ -567,6 +591,7 @@ static irqreturn_t prq_event_thread(int irq, void *d)
 		result = QI_RESP_SUCCESS;
 	invalid:
 		up_read(&svm->mm->mmap_sem);
+		mmput(svm->mm);
 	bad_req:
 		/* Accounting for major/minor faults? */
 		rcu_read_lock();
diff --git a/drivers/iommu/intel_irq_remapping.c b/drivers/iommu/intel_irq_remapping.c
index c12ba4516df2..ac596928f6b4 100644
--- a/drivers/iommu/intel_irq_remapping.c
+++ b/drivers/iommu/intel_irq_remapping.c
@@ -629,7 +629,7 @@ static void iommu_disable_irq_remapping(struct intel_iommu *iommu)
 
 	raw_spin_lock_irqsave(&iommu->register_lock, flags);
 
-	sts = dmar_readq(iommu->reg + DMAR_GSTS_REG);
+	sts = readl(iommu->reg + DMAR_GSTS_REG);
 	if (!(sts & DMA_GSTS_IRES))
 		goto end;