[Patch Part2 V2 17/17] iommu/vt-d: update IOMMU state when memory hotplug happens

[Jiang Liu <jiang.liu@linux.intel.com>]

If static identity domain is created, IOMMU driver needs to update
si_domain page table when memory hotplug event happens. Otherwise
PCI device DMA operations can't access the hot-added memory regions.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/iommu/intel-iommu.c |   71 ++++++++++++++++++++++++++++++++++++++++++-
 drivers/iommu/iova.c        |   64 ++++++++++++++++++++++++++++++++++----
 include/linux/iova.h        |    2 ++
 3 files changed, 130 insertions(+), 7 deletions(-)

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index dd576c0..484d669 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -33,6 +33,7 @@
 #include <linux/dmar.h>
 #include <linux/dma-mapping.h>
 #include <linux/mempool.h>
+#include <linux/memory.h>
 #include <linux/timer.h>
 #include <linux/iova.h>
 #include <linux/iommu.h>
@@ -3683,6 +3684,73 @@ static struct notifier_block device_nb = {
 	.notifier_call = device_notifier,
 };
 
+static int intel_iommu_memory_notifier(struct notifier_block *nb,
+				       unsigned long val, void *v)
+{
+	struct memory_notify *mhp = v;
+	unsigned long long start, end;
+	unsigned long start_vpfn, last_vpfn;
+
+	switch (val) {
+	case MEM_GOING_ONLINE:
+		start = mhp->start_pfn << PAGE_SHIFT;
+		end = ((mhp->start_pfn + mhp->nr_pages) << PAGE_SHIFT) - 1;
+		if (iommu_domain_identity_map(si_domain, start, end)) {
+			pr_warn("dmar: failed to build identity map for [%llx-%llx]\n",
+				start, end);
+			return NOTIFY_BAD;
+		}
+		break;
+
+	case MEM_OFFLINE:
+	case MEM_CANCEL_ONLINE:
+		start_vpfn = mm_to_dma_pfn(mhp->start_pfn);
+		last_vpfn = mm_to_dma_pfn(mhp->start_pfn + mhp->nr_pages - 1);
+		while (start_vpfn <= last_vpfn) {
+			struct iova *iova;
+			struct dmar_drhd_unit *drhd;
+			struct intel_iommu *iommu;
+
+			iova = find_iova(&si_domain->iovad, start_vpfn);
+			if (iova == NULL) {
+				pr_debug("dmar: failed get IOVA for PFN %lx\n",
+					 start_vpfn);
+				break;
+			}
+
+			iova = split_and_remove_iova(&si_domain->iovad, iova,
+						     start_vpfn, last_vpfn);
+			if (iova == NULL) {
+				pr_warn("dmar: failed to split IOVA PFN [%lx-%lx]\n",
+					start_vpfn, last_vpfn);
+				return NOTIFY_BAD;
+			}
+
+			rcu_read_lock();
+			for_each_active_iommu(iommu, drhd)
+				iommu_flush_iotlb_psi(iommu, si_domain->id,
+					iova->pfn_lo,
+					iova->pfn_hi - iova->pfn_lo + 1, 0);
+			rcu_read_unlock();
+			dma_pte_clear_range(si_domain, iova->pfn_lo,
+					    iova->pfn_hi);
+			dma_pte_free_pagetable(si_domain, iova->pfn_lo,
+					       iova->pfn_hi);
+
+			start_vpfn = iova->pfn_hi + 1;
+			free_iova_mem(iova);
+		}
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block intel_iommu_memory_nb = {
+	.notifier_call = intel_iommu_memory_notifier,
+	.priority = 0
+};
+
 int __init intel_iommu_init(void)
 {
 	int ret = -ENODEV;
@@ -3755,8 +3823,9 @@ int __init intel_iommu_init(void)
 	init_iommu_pm_ops();
 
 	bus_set_iommu(&pci_bus_type, &intel_iommu_ops);
-
 	bus_register_notifier(&pci_bus_type, &device_nb);
+	if (si_domain && !hw_pass_through)
+		register_memory_notifier(&intel_iommu_memory_nb);
 
 	intel_iommu_enabled = 1;
 
diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c
index 67da6cff..f6b17e6 100644
--- a/drivers/iommu/iova.c
+++ b/drivers/iommu/iova.c
@@ -342,19 +342,30 @@ __is_range_overlap(struct rb_node *node,
 	return 0;
 }
 
+static inline struct iova *
+alloc_and_init_iova(unsigned long pfn_lo, unsigned long pfn_hi)
+{
+	struct iova *iova;
+
+	iova = alloc_iova_mem();
+	if (iova) {
+		iova->pfn_lo = pfn_lo;
+		iova->pfn_hi = pfn_hi;
+	}
+
+	return iova;
+}
+
 static struct iova *
 __insert_new_range(struct iova_domain *iovad,
 	unsigned long pfn_lo, unsigned long pfn_hi)
 {
 	struct iova *iova;
 
-	iova = alloc_iova_mem();
-	if (!iova)
-		return iova;
+	iova = alloc_and_init_iova(pfn_lo, pfn_hi);
+	if (iova)
+		iova_insert_rbtree(&iovad->rbroot, iova);
 
-	iova->pfn_hi = pfn_hi;
-	iova->pfn_lo = pfn_lo;
-	iova_insert_rbtree(&iovad->rbroot, iova);
 	return iova;
 }
 
@@ -433,3 +444,44 @@ copy_reserved_iova(struct iova_domain *from, struct iova_domain *to)
 	}
 	spin_unlock_irqrestore(&from->iova_rbtree_lock, flags);
 }
+
+struct iova *
+split_and_remove_iova(struct iova_domain *iovad, struct iova *iova,
+		      unsigned long pfn_lo, unsigned long pfn_hi)
+{
+	unsigned long flags;
+	struct iova *prev = NULL, *next = NULL;
+
+	spin_lock_irqsave(&iovad->iova_rbtree_lock, flags);
+	if (iova->pfn_lo < pfn_lo) {
+		prev = alloc_and_init_iova(iova->pfn_lo, pfn_lo - 1);
+		if (prev == NULL)
+			goto error;
+	}
+	if (iova->pfn_hi > pfn_hi) {
+		next = alloc_and_init_iova(pfn_hi + 1, iova->pfn_hi);
+		if (next == NULL)
+			goto error;
+	}
+
+	__cached_rbnode_delete_update(iovad, iova);
+	rb_erase(&iova->node, &iovad->rbroot);
+
+	if (prev) {
+		iova_insert_rbtree(&iovad->rbroot, prev);
+		iova->pfn_lo = pfn_lo;
+	}
+	if (next) {
+		iova_insert_rbtree(&iovad->rbroot, next);
+		iova->pfn_hi = pfn_hi;
+	}
+	spin_unlock_irqrestore(&iovad->iova_rbtree_lock, flags);
+
+	return iova;
+
+error:
+	spin_unlock_irqrestore(&iovad->iova_rbtree_lock, flags);
+	if (prev)
+		free_iova_mem(prev);
+	return NULL;
+}
diff --git a/include/linux/iova.h b/include/linux/iova.h
index 76a0759..3277f47 100644
--- a/include/linux/iova.h
+++ b/include/linux/iova.h
@@ -47,5 +47,7 @@ void copy_reserved_iova(struct iova_domain *from, struct iova_domain *to);
 void init_iova_domain(struct iova_domain *iovad, unsigned long pfn_32bit);
 struct iova *find_iova(struct iova_domain *iovad, unsigned long pfn);
 void put_iova_domain(struct iova_domain *iovad);
+struct iova *split_and_remove_iova(struct iova_domain *iovad,
+	struct iova *iova, unsigned long pfn_lo, unsigned long pfn_hi);
 
 #endif
-- 
1.7.10.4