[PATCH 3/4] mm: Add kernel MMU notifier to manage remote TLB

[PATCH 3/4] mm: Add kernel MMU notifier to manage remote TLB

[Ashok Raj]

From: Huang Ying <ying.huang@intel.com>

Shared Virtual Memory (SVM) devices have TLBs that cache entries from
the CPU's page tables.  We need SVM device drivers to flush them at
the same time that we flush the CPU TLBs.  We can use the existing MMU
notifiers for userspace updates, but we lack a mechanism to get
notified when kernel page tables are updated.

To implement the MMU notification mechanism for the kernel address
space, a kernel MMU notifier chain is defined, and will be called when
the CPU TLB is flushed for the kernel address space.  The IOMMU SVM
driver can register on the notifier chain to flush the device TLBs
when necessary.

To: linux-kernel@vger.kernel.org
To: Joerg Roedel <joro@8bytes.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: CQ Tang <cq.tang@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: x86@kernel.org
Cc: linux-mm@kvack.org
Cc: iommu@lists.linux-foundation.org
Cc: David Woodhouse <dwmw2@infradead.org>
CC: Jean-Phillipe Brucker <jean-philippe.brucker@arm.com>
Signed-off-by: "Huang, Ying" <ying.huang@intel.com>
---
 arch/x86/include/asm/tlbflush.h |  1 +
 arch/x86/mm/tlb.c               |  1 +
 include/linux/mmu_notifier.h    | 33 +++++++++++++++++++++++++++++++++
 mm/mmu_notifier.c               | 25 +++++++++++++++++++++++++
 4 files changed, 60 insertions(+)

diff --git a/arch/x86/include/asm/tlbflush.h b/arch/x86/include/asm/tlbflush.h
index 50ea348..f5fd0b8 100644
--- a/arch/x86/include/asm/tlbflush.h
+++ b/arch/x86/include/asm/tlbflush.h
@@ -3,6 +3,7 @@
 
 #include <linux/mm.h>
 #include <linux/sched.h>
+#include <linux/mmu_notifier.h>
 
 #include <asm/processor.h>
 #include <asm/cpufeature.h>
diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c
index 014d07a..6dea8e9 100644
--- a/arch/x86/mm/tlb.c
+++ b/arch/x86/mm/tlb.c
@@ -314,6 +314,7 @@ void flush_tlb_kernel_range(unsigned long start, unsigned long end)
 		info.end = end;
 		on_each_cpu(do_kernel_range_flush, &info, 1);
 	}
+	kernel_mmu_notifier_invalidate_range(start, end);
 }
 
 void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch)
diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h
index c91b3bc..4a96089 100644
--- a/include/linux/mmu_notifier.h
+++ b/include/linux/mmu_notifier.h
@@ -418,6 +418,25 @@ extern void mmu_notifier_call_srcu(struct rcu_head *rcu,
 				   void (*func)(struct rcu_head *rcu));
 extern void mmu_notifier_synchronize(void);
 
+struct kernel_mmu_address_range {
+	unsigned long start;
+	unsigned long end;
+};
+
+/*
+ * Before the virtual address range managed by kernel (vmalloc/kmap)
+ * is reused, That is, remapped to the new physical addresses, the
+ * kernel MMU notifier will be called with KERNEL_MMU_INVALIDATE_RANGE
+ * and struct kernel_mmu_address_range as parameters.  This is used to
+ * manage the remote TLB.
+ */
+#define KERNEL_MMU_INVALIDATE_RANGE		1
+extern int kernel_mmu_notifier_register(struct notifier_block *nb);
+extern int kernel_mmu_notifier_unregister(struct notifier_block *nb);
+
+extern int kernel_mmu_notifier_invalidate_range(unsigned long start,
+						unsigned long end);
+
 #else /* CONFIG_MMU_NOTIFIER */
 
 static inline void mmu_notifier_release(struct mm_struct *mm)
@@ -479,6 +498,20 @@ static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
 #define pudp_huge_clear_flush_notify pudp_huge_clear_flush
 #define set_pte_at_notify set_pte_at
 
+static inline int kernel_mmu_notifier_register(struct notifier_block *nb)
+{
+	return 0;
+}
+
+static inline int kernel_mmu_notifier_unregister(struct notifier_block *nb)
+{
+	return 0;
+}
+
+static inline void kernel_mmu_notifier_invalidate_range(unsigned long start,
+							unsigned long end)
+{
+}
 #endif /* CONFIG_MMU_NOTIFIER */
 
 #endif /* _LINUX_MMU_NOTIFIER_H */
diff --git a/mm/mmu_notifier.c b/mm/mmu_notifier.c
index 54ca545..a919038 100644
--- a/mm/mmu_notifier.c
+++ b/mm/mmu_notifier.c
@@ -400,3 +400,28 @@ void mmu_notifier_unregister_no_release(struct mmu_notifier *mn,
 	mmdrop(mm);
 }
 EXPORT_SYMBOL_GPL(mmu_notifier_unregister_no_release);
+
+static ATOMIC_NOTIFIER_HEAD(kernel_mmu_notifier_list);
+
+int kernel_mmu_notifier_register(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_register(&kernel_mmu_notifier_list, nb);
+}
+
+int kernel_mmu_notifier_unregister(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&kernel_mmu_notifier_list, nb);
+}
+
+int kernel_mmu_notifier_invalidate_range(unsigned long start,
+					 unsigned long end)
+{
+	struct kernel_mmu_address_range range = {
+		.start	= start,
+		.end	= end,
+	};
+
+	return atomic_notifier_call_chain(&kernel_mmu_notifier_list,
+					  KERNEL_MMU_INVALIDATE_RANGE,
+					  &range);
+}
-- 
2.7.4

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 4/4] iommu/vt-d: Hooks to invalidate iotlb/devtlb when using supervisor PASID's.

[Ashok Raj]

When a kernel client uses intel_svm_bind_mm() and requests a supervisor
PASID, IOMMU needs to track changes to these addresses. Otherwise the device
tlb will be stale compared to what's on the cpu for kernel mappings. This
is similar to what's done for user space registrations via
mmu_notifier_register() api's.

To: linux-kernel@vger.kernel.org
To: Joerg Roedel <joro@8bytes.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc:	Huang Ying <ying.huang@intel.com>
Cc: CQ Tang <cq.tang@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: x86@kernel.org
Cc: linux-mm@kvack.org
Cc: iommu@lists.linux-foundation.org
Cc: David Woodhouse <dwmw2@infradead.org>
CC: Jean-Phillipe Brucker <jean-philippe.brucker@arm.com>

Signed-off-by: Ashok Raj <ashok.raj@intel.com>
---
 drivers/iommu/intel-svm.c   | 29 +++++++++++++++++++++++++++--
 include/linux/intel-iommu.h |  5 ++++-
 2 files changed, 31 insertions(+), 3 deletions(-)

diff --git a/drivers/iommu/intel-svm.c b/drivers/iommu/intel-svm.c
index 0c9f077..1758814 100644
--- a/drivers/iommu/intel-svm.c
+++ b/drivers/iommu/intel-svm.c
@@ -292,6 +292,26 @@ static const struct mmu_notifier_ops intel_mmuops = {
 
 static DEFINE_MUTEX(pasid_mutex);
 
+static int intel_init_mm_inval_range(struct notifier_block *nb,
+	unsigned long action, void *data)
+{
+	struct kernel_mmu_address_range *range;
+	struct intel_svm *svm = container_of(nb, struct intel_svm, init_mm_nb);
+	unsigned long start, end;
+	struct intel_iommu *iommu;
+
+	if (action == KERNEL_MMU_INVALIDATE_RANGE) {
+		range = data;
+		start = range->start;
+		end = range->end;
+		iommu = svm->iommu;
+
+		intel_flush_svm_range(svm, start,
+			(end - start + PAGE_SIZE - 1) >> VTD_PAGE_SHIFT, 0, 0);
+	}
+	return 0;
+}
+
 int intel_svm_bind_mm(struct device *dev, int *pasid, int flags, struct svm_dev_ops *ops)
 {
 	struct intel_iommu *iommu = intel_svm_device_to_iommu(dev);
@@ -391,12 +411,12 @@ int intel_svm_bind_mm(struct device *dev, int *pasid, int flags, struct svm_dev_
 			goto out;
 		}
 		svm->pasid = ret;
-		svm->notifier.ops = &intel_mmuops;
 		svm->mm = mm;
 		svm->flags = flags;
 		INIT_LIST_HEAD_RCU(&svm->devs);
 		ret = -ENOMEM;
 		if (mm) {
+			svm->notifier.ops = &intel_mmuops;
 			ret = mmu_notifier_register(&svm->notifier, mm);
 			if (ret) {
 				idr_remove(&svm->iommu->pasid_idr, svm->pasid);
@@ -405,8 +425,11 @@ 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;
-		} else
+		} else {
+			svm->init_mm_nb.notifier_call = intel_init_mm_inval_range;
+			kernel_mmu_notifier_register(&svm->init_mm_nb);
 			iommu->pasid_table[svm->pasid].val = (u64)__pa(init_mm.pgd) | 1 | (1ULL << 11);
+		}
 		wmb();
 		/* In caching mode, we still have to flush with PASID 0 when
 		 * a PASID table entry becomes present. Not entirely clear
@@ -471,6 +494,8 @@ int intel_svm_unbind_mm(struct device *dev, int pasid)
 					idr_remove(&svm->iommu->pasid_idr, svm->pasid);
 					if (svm->mm)
 						mmu_notifier_unregister(&svm->notifier, svm->mm);
+					else
+						kernel_mmu_notifier_unregister(&svm->init_mm_nb);
 
 					/* We mandate that no page faults may be outstanding
 					 * for the PASID when intel_svm_unbind_mm() is called.
diff --git a/include/linux/intel-iommu.h b/include/linux/intel-iommu.h
index 485a5b4..d6019b4 100644
--- a/include/linux/intel-iommu.h
+++ b/include/linux/intel-iommu.h
@@ -477,7 +477,10 @@ struct intel_svm_dev {
 };
 
 struct intel_svm {
-	struct mmu_notifier notifier;
+	union {
+		struct mmu_notifier notifier;
+		struct notifier_block init_mm_nb;
+	};
 	struct mm_struct *mm;
 	struct intel_iommu *iommu;
 	int flags;
-- 
2.7.4

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 4/4] iommu/vt-d: Hooks to invalidate iotlb/devtlb when using supervisor PASID's.

[Ashok Raj]

When a kernel client uses intel_svm_bind_mm() and requests a supervisor
PASID, IOMMU needs to track changes to these addresses. Otherwise the device
tlb will be stale compared to what's on the cpu for kernel mappings. This
is similar to what's done for user space registrations via
mmu_notifier_register() api's.

To: linux-kernel@vger.kernel.org
To: Joerg Roedel <joro@8bytes.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc:	Huang Ying <ying.huang@intel.com>
Cc: CQ Tang <cq.tang@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Vegard Nossum <vegard.nossum@oracle.com>
Cc: x86@kernel.org
Cc: linux-mm@kvack.org
Cc: iommu@lists-foundation.org
Cc: David Woodhouse <dwmw2@infradead.org>
CC: Jean-Phillipe Brucker <jean-philippe.brucker@arm.com>

Signed-off-by: Ashok Raj <ashok.raj@intel.com>
---
 drivers/iommu/intel-svm.c   | 29 +++++++++++++++++++++++++++--
 include/linux/intel-iommu.h |  5 ++++-
 2 files changed, 31 insertions(+), 3 deletions(-)

diff --git a/drivers/iommu/intel-svm.c b/drivers/iommu/intel-svm.c
index 0c9f077..1758814 100644
--- a/drivers/iommu/intel-svm.c
+++ b/drivers/iommu/intel-svm.c
@@ -292,6 +292,26 @@ static const struct mmu_notifier_ops intel_mmuops = {
 
 static DEFINE_MUTEX(pasid_mutex);
 
+static int intel_init_mm_inval_range(struct notifier_block *nb,
+	unsigned long action, void *data)
+{
+	struct kernel_mmu_address_range *range;
+	struct intel_svm *svm = container_of(nb, struct intel_svm, init_mm_nb);
+	unsigned long start, end;
+	struct intel_iommu *iommu;
+
+	if (action == KERNEL_MMU_INVALIDATE_RANGE) {
+		range = data;
+		start = range->start;
+		end = range->end;
+		iommu = svm->iommu;
+
+		intel_flush_svm_range(svm, start,
+			(end - start + PAGE_SIZE - 1) >> VTD_PAGE_SHIFT, 0, 0);
+	}
+	return 0;
+}
+
 int intel_svm_bind_mm(struct device *dev, int *pasid, int flags, struct svm_dev_ops *ops)
 {
 	struct intel_iommu *iommu = intel_svm_device_to_iommu(dev);
@@ -391,12 +411,12 @@ int intel_svm_bind_mm(struct device *dev, int *pasid, int flags, struct svm_dev_
 			goto out;
 		}
 		svm->pasid = ret;
-		svm->notifier.ops = &intel_mmuops;
 		svm->mm = mm;
 		svm->flags = flags;
 		INIT_LIST_HEAD_RCU(&svm->devs);
 		ret = -ENOMEM;
 		if (mm) {
+			svm->notifier.ops = &intel_mmuops;
 			ret = mmu_notifier_register(&svm->notifier, mm);
 			if (ret) {
 				idr_remove(&svm->iommu->pasid_idr, svm->pasid);
@@ -405,8 +425,11 @@ 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;
-		} else
+		} else {
+			svm->init_mm_nb.notifier_call = intel_init_mm_inval_range;
+			kernel_mmu_notifier_register(&svm->init_mm_nb);
 			iommu->pasid_table[svm->pasid].val = (u64)__pa(init_mm.pgd) | 1 | (1ULL << 11);
+		}
 		wmb();
 		/* In caching mode, we still have to flush with PASID 0 when
 		 * a PASID table entry becomes present. Not entirely clear
@@ -471,6 +494,8 @@ int intel_svm_unbind_mm(struct device *dev, int pasid)
 					idr_remove(&svm->iommu->pasid_idr, svm->pasid);
 					if (svm->mm)
 						mmu_notifier_unregister(&svm->notifier, svm->mm);
+					else
+						kernel_mmu_notifier_unregister(&svm->init_mm_nb);
 
 					/* We mandate that no page faults may be outstanding
 					 * for the PASID when intel_svm_unbind_mm() is called.
diff --git a/include/linux/intel-iommu.h b/include/linux/intel-iommu.h
index 485a5b4..d6019b4 100644
--- a/include/linux/intel-iommu.h
+++ b/include/linux/intel-iommu.h
@@ -477,7 +477,10 @@ struct intel_svm_dev {
 };
 
 struct intel_svm {
-	struct mmu_notifier notifier;
+	union {
+		struct mmu_notifier notifier;
+		struct notifier_block init_mm_nb;
+	};
 	struct mm_struct *mm;
 	struct intel_iommu *iommu;
 	int flags;
-- 
2.7.4

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>