1 files changed, 315 insertions, 346 deletions

diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
index 5efc5eee9544..2c960f187f7c 100644
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@ -22,6 +22,7 @@
 
 #define pr_fmt(fmt) "xen:" KBUILD_MODNAME ": " fmt
 
+#include <linux/dma-mapping.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
@@ -34,9 +35,7 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/refcount.h>
-#ifdef CONFIG_XEN_GRANT_DMA_ALLOC
-#include <linux/of_device.h>
-#endif
+#include <linux/workqueue.h>
 
 #include <xen/xen.h>
 #include <xen/grant_table.h>
@@ -57,26 +56,33 @@ MODULE_AUTHOR("Derek G. Murray <Derek.Murray@cl.cam.ac.uk>, "
 	      "Gerd Hoffmann <kraxel@redhat.com>");
 MODULE_DESCRIPTION("User-space granted page access driver");
 
-static int limit = 1024*1024;
-module_param(limit, int, 0644);
-MODULE_PARM_DESC(limit, "Maximum number of grants that may be mapped by "
-		"the gntdev device");
+#define GNTDEV_COPY_BATCH 16
 
-static atomic_t pages_mapped = ATOMIC_INIT(0);
+struct gntdev_copy_batch {
+	struct gnttab_copy ops[GNTDEV_COPY_BATCH];
+	struct page *pages[GNTDEV_COPY_BATCH];
+	s16 __user *status[GNTDEV_COPY_BATCH];
+	unsigned int nr_ops;
+	unsigned int nr_pages;
+	bool writeable;
+	struct gntdev_copy_batch *next;
+};
 
-static int use_ptemod;
-#define populate_freeable_maps use_ptemod
+static unsigned int limit = 64*1024;
+module_param(limit, uint, 0644);
+MODULE_PARM_DESC(limit,
+	"Maximum number of grants that may be mapped by one mapping request");
 
-static int unmap_grant_pages(struct gntdev_grant_map *map,
-			     int offset, int pages);
+static void unmap_grant_pages(struct gntdev_grant_map *map,
+			      int offset, int pages);
 
 static struct miscdevice gntdev_miscdev;
 
 /* ------------------------------------------------------------------ */
 
-bool gntdev_account_mapped_pages(int count)
+bool gntdev_test_page_count(unsigned int count)
 {
-	return atomic_add_return(count, &pages_mapped) > limit;
+	return !count || count > limit;
 }
 
 static void gntdev_print_maps(struct gntdev_priv *priv,
@@ -117,14 +123,15 @@ static void gntdev_free_map(struct gntdev_grant_map *map)
 		gnttab_free_pages(map->count, map->pages);
 
 #ifdef CONFIG_XEN_GRANT_DMA_ALLOC
-	kfree(map->frames);
+	kvfree(map->frames);
 #endif
-	kfree(map->pages);
-	kfree(map->grants);
-	kfree(map->map_ops);
-	kfree(map->unmap_ops);
-	kfree(map->kmap_ops);
-	kfree(map->kunmap_ops);
+	kvfree(map->pages);
+	kvfree(map->grants);
+	kvfree(map->map_ops);
+	kvfree(map->unmap_ops);
+	kvfree(map->kmap_ops);
+	kvfree(map->kunmap_ops);
+	kvfree(map->being_removed);
 	kfree(map);
 }
 
@@ -138,19 +145,29 @@ struct gntdev_grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count,
 	if (NULL == add)
 		return NULL;
 
-	add->grants    = kcalloc(count, sizeof(add->grants[0]), GFP_KERNEL);
-	add->map_ops   = kcalloc(count, sizeof(add->map_ops[0]), GFP_KERNEL);
-	add->unmap_ops = kcalloc(count, sizeof(add->unmap_ops[0]), GFP_KERNEL);
-	add->kmap_ops  = kcalloc(count, sizeof(add->kmap_ops[0]), GFP_KERNEL);
-	add->kunmap_ops = kcalloc(count, sizeof(add->kunmap_ops[0]), GFP_KERNEL);
-	add->pages     = kcalloc(count, sizeof(add->pages[0]), GFP_KERNEL);
+	add->grants    = kvmalloc_array(count, sizeof(add->grants[0]),
+					GFP_KERNEL);
+	add->map_ops   = kvmalloc_array(count, sizeof(add->map_ops[0]),
+					GFP_KERNEL);
+	add->unmap_ops = kvmalloc_array(count, sizeof(add->unmap_ops[0]),
+					GFP_KERNEL);
+	add->pages     = kvcalloc(count, sizeof(add->pages[0]), GFP_KERNEL);
+	add->being_removed =
+		kvcalloc(count, sizeof(add->being_removed[0]), GFP_KERNEL);
 	if (NULL == add->grants    ||
 	    NULL == add->map_ops   ||
 	    NULL == add->unmap_ops ||
-	    NULL == add->kmap_ops  ||
-	    NULL == add->kunmap_ops ||
-	    NULL == add->pages)
+	    NULL == add->pages     ||
+	    NULL == add->being_removed)
 		goto err;
+	if (xen_pv_domain()) {
+		add->kmap_ops   = kvmalloc_array(count, sizeof(add->kmap_ops[0]),
+						 GFP_KERNEL);
+		add->kunmap_ops = kvmalloc_array(count, sizeof(add->kunmap_ops[0]),
+						 GFP_KERNEL);
+		if (NULL == add->kmap_ops || NULL == add->kunmap_ops)
+			goto err;
+	}
 
 #ifdef CONFIG_XEN_GRANT_DMA_ALLOC
 	add->dma_flags = dma_flags;
@@ -162,8 +179,8 @@ struct gntdev_grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count,
 	if (dma_flags & (GNTDEV_DMA_FLAG_WC | GNTDEV_DMA_FLAG_COHERENT)) {
 		struct gnttab_dma_alloc_args args;
 
-		add->frames = kcalloc(count, sizeof(add->frames[0]),
-				      GFP_KERNEL);
+		add->frames = kvcalloc(count, sizeof(add->frames[0]),
+				       GFP_KERNEL);
 		if (!add->frames)
 			goto err;
 
@@ -187,10 +204,14 @@ struct gntdev_grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count,
 		goto err;
 
 	for (i = 0; i < count; i++) {
-		add->map_ops[i].handle = -1;
-		add->unmap_ops[i].handle = -1;
-		add->kmap_ops[i].handle = -1;
-		add->kunmap_ops[i].handle = -1;
+		add->grants[i].domid = DOMID_INVALID;
+		add->grants[i].ref = INVALID_GRANT_REF;
+		add->map_ops[i].handle = INVALID_GRANT_HANDLE;
+		add->unmap_ops[i].handle = INVALID_GRANT_HANDLE;
+		if (xen_pv_domain()) {
+			add->kmap_ops[i].handle = INVALID_GRANT_HANDLE;
+			add->kunmap_ops[i].handle = INVALID_GRANT_HANDLE;
+		}
 	}
 
 	add->index = 0;
@@ -244,69 +265,76 @@ void gntdev_put_map(struct gntdev_priv *priv, struct gntdev_grant_map *map)
 	if (!refcount_dec_and_test(&map->users))
 		return;
 
-	atomic_sub(map->count, &pages_mapped);
+	if (map->pages && !xen_pv_domain()) {
+		/*
+		 * Increment the reference count.  This ensures that the
+		 * subsequent call to unmap_grant_pages() will not wind up
+		 * re-entering itself.  It *can* wind up calling
+		 * gntdev_put_map() recursively, but such calls will be with a
+		 * reference count greater than 1, so they will return before
+		 * this code is reached.  The recursion depth is thus limited to
+		 * 1.  Do NOT use refcount_inc() here, as it will detect that
+		 * the reference count is zero and WARN().
+		 */
+		refcount_set(&map->users, 1);
+
+		/*
+		 * Unmap the grants.  This may or may not be asynchronous, so it
+		 * is possible that the reference count is 1 on return, but it
+		 * could also be greater than 1.
+		 */
+		unmap_grant_pages(map, 0, map->count);
+
+		/* Check if the memory now needs to be freed */
+		if (!refcount_dec_and_test(&map->users))
+			return;
+
+		/*
+		 * All pages have been returned to the hypervisor, so free the
+		 * map.
+		 */
+	}
+
+	if (xen_pv_domain() && map->notifier_init)
+		mmu_interval_notifier_remove(&map->notifier);
 
 	if (map->notify.flags & UNMAP_NOTIFY_SEND_EVENT) {
 		notify_remote_via_evtchn(map->notify.event);
 		evtchn_put(map->notify.event);
 	}
-
-	if (populate_freeable_maps && priv) {
-		mutex_lock(&priv->lock);
-		list_del(&map->next);
-		mutex_unlock(&priv->lock);
-	}
-
-	if (map->pages && !use_ptemod)
-		unmap_grant_pages(map, 0, map->count);
 	gntdev_free_map(map);
 }
 
 /* ------------------------------------------------------------------ */
 
-static int find_grant_ptes(pte_t *pte, pgtable_t token,
-		unsigned long addr, void *data)
+static int find_grant_ptes(pte_t *pte, unsigned long addr, void *data)
 {
 	struct gntdev_grant_map *map = data;
-	unsigned int pgnr = (addr - map->vma->vm_start) >> PAGE_SHIFT;
-	int flags = map->flags | GNTMAP_application_map | GNTMAP_contains_pte;
+	unsigned int pgnr = (addr - map->pages_vm_start) >> PAGE_SHIFT;
+	int flags = map->flags | GNTMAP_application_map | GNTMAP_contains_pte |
+		    (1 << _GNTMAP_guest_avail0);
 	u64 pte_maddr;
 
 	BUG_ON(pgnr >= map->count);
 	pte_maddr = arbitrary_virt_to_machine(pte).maddr;
 
-	/*
-	 * Set the PTE as special to force get_user_pages_fast() fall
-	 * back to the slow path.  If this is not supported as part of
-	 * the grant map, it will be done afterwards.
-	 */
-	if (xen_feature(XENFEAT_gnttab_map_avail_bits))
-		flags |= (1 << _GNTMAP_guest_avail0);
-
+	/* Note: this will perform a pte_mkspecial() through the hypercall. */
 	gnttab_set_map_op(&map->map_ops[pgnr], pte_maddr, flags,
 			  map->grants[pgnr].ref,
 			  map->grants[pgnr].domid);
 	gnttab_set_unmap_op(&map->unmap_ops[pgnr], pte_maddr, flags,
-			    -1 /* handle */);
+			    INVALID_GRANT_HANDLE);
 	return 0;
 }
 
-#ifdef CONFIG_X86
-static int set_grant_ptes_as_special(pte_t *pte, pgtable_t token,
-				     unsigned long addr, void *data)
-{
-	set_pte_at(current->mm, addr, pte, pte_mkspecial(*pte));
-	return 0;
-}
-#endif
-
 int gntdev_map_grant_pages(struct gntdev_grant_map *map)
 {
+	size_t alloced = 0;
 	int i, err = 0;
 
-	if (!use_ptemod) {
+	if (!xen_pv_domain()) {
 		/* Note: it could already be mapped */
-		if (map->map_ops[0].handle != -1)
+		if (map->map_ops[0].handle != INVALID_GRANT_HANDLE)
 			return 0;
 		for (i = 0; i < map->count; i++) {
 			unsigned long addr = (unsigned long)
@@ -315,7 +343,7 @@ int gntdev_map_grant_pages(struct gntdev_grant_map *map)
 				map->grants[i].ref,
 				map->grants[i].domid);
 			gnttab_set_unmap_op(&map->unmap_ops[i], addr,
-				map->flags, -1 /* handle */);
+				map->flags, INVALID_GRANT_HANDLE);
 		}
 	} else {
 		/*
@@ -323,111 +351,157 @@ int gntdev_map_grant_pages(struct gntdev_grant_map *map)
 		 * to the kernel linear addresses of the struct pages.
 		 * These ptes are completely different from the user ptes dealt
 		 * with find_grant_ptes.
+		 * Note that GNTMAP_device_map isn't needed here: The
+		 * dev_bus_addr output field gets consumed only from ->map_ops,
+		 * and by not requesting it when mapping we also avoid needing
+		 * to mirror dev_bus_addr into ->unmap_ops (and holding an extra
+		 * reference to the page in the hypervisor).
 		 */
+		unsigned int flags = (map->flags & ~GNTMAP_device_map) |
+				     GNTMAP_host_map;
+
 		for (i = 0; i < map->count; i++) {
 			unsigned long address = (unsigned long)
 				pfn_to_kaddr(page_to_pfn(map->pages[i]));
 			BUG_ON(PageHighMem(map->pages[i]));
 
-			gnttab_set_map_op(&map->kmap_ops[i], address,
-				map->flags | GNTMAP_host_map,
+			gnttab_set_map_op(&map->kmap_ops[i], address, flags,
 				map->grants[i].ref,
 				map->grants[i].domid);
 			gnttab_set_unmap_op(&map->kunmap_ops[i], address,
-				map->flags | GNTMAP_host_map, -1);
+				flags, INVALID_GRANT_HANDLE);
 		}
 	}
 
 	pr_debug("map %d+%d\n", map->index, map->count);
-	err = gnttab_map_refs(map->map_ops, use_ptemod ? map->kmap_ops : NULL,
-			map->pages, map->count);
-	if (err)
-		return err;
+	err = gnttab_map_refs(map->map_ops, map->kmap_ops, map->pages,
+			map->count);
 
 	for (i = 0; i < map->count; i++) {
-		if (map->map_ops[i].status) {
+		if (map->map_ops[i].status == GNTST_okay) {
+			map->unmap_ops[i].handle = map->map_ops[i].handle;
+			alloced++;
+		} else if (!err)
 			err = -EINVAL;
-			continue;
-		}
 
-		map->unmap_ops[i].handle = map->map_ops[i].handle;
-		if (use_ptemod)
-			map->kunmap_ops[i].handle = map->kmap_ops[i].handle;
-#ifdef CONFIG_XEN_GRANT_DMA_ALLOC
-		else if (map->dma_vaddr) {
-			unsigned long bfn;
+		if (map->flags & GNTMAP_device_map)
+			map->unmap_ops[i].dev_bus_addr = map->map_ops[i].dev_bus_addr;
 
-			bfn = pfn_to_bfn(page_to_pfn(map->pages[i]));
-			map->unmap_ops[i].dev_bus_addr = __pfn_to_phys(bfn);
+		if (xen_pv_domain()) {
+			if (map->kmap_ops[i].status == GNTST_okay) {
+				alloced++;
+				map->kunmap_ops[i].handle = map->kmap_ops[i].handle;
+			} else if (!err)
+				err = -EINVAL;
 		}
-#endif
 	}
+	atomic_add(alloced, &map->live_grants);
 	return err;
 }
 
-static int __unmap_grant_pages(struct gntdev_grant_map *map, int offset,
-			       int pages)
+static void __unmap_grant_pages_done(int result,
+		struct gntab_unmap_queue_data *data)
 {
-	int i, err = 0;
-	struct gntab_unmap_queue_data unmap_data;
+	unsigned int i;
+	struct gntdev_grant_map *map = data->data;
+	unsigned int offset = data->unmap_ops - map->unmap_ops;
+	int successful_unmaps = 0;
+	int live_grants;
+
+	for (i = 0; i < data->count; i++) {
+		if (map->unmap_ops[offset + i].status == GNTST_okay &&
+		    map->unmap_ops[offset + i].handle != INVALID_GRANT_HANDLE)
+			successful_unmaps++;
+
+		WARN_ON(map->unmap_ops[offset + i].status != GNTST_okay &&
+			map->unmap_ops[offset + i].handle != INVALID_GRANT_HANDLE);
+		pr_debug("unmap handle=%d st=%d\n",
+			map->unmap_ops[offset+i].handle,
+			map->unmap_ops[offset+i].status);
+		map->unmap_ops[offset+i].handle = INVALID_GRANT_HANDLE;
+		if (xen_pv_domain()) {
+			if (map->kunmap_ops[offset + i].status == GNTST_okay &&
+			    map->kunmap_ops[offset + i].handle != INVALID_GRANT_HANDLE)
+				successful_unmaps++;
+
+			WARN_ON(map->kunmap_ops[offset + i].status != GNTST_okay &&
+				map->kunmap_ops[offset + i].handle != INVALID_GRANT_HANDLE);
+			pr_debug("kunmap handle=%u st=%d\n",
+				 map->kunmap_ops[offset+i].handle,
+				 map->kunmap_ops[offset+i].status);
+			map->kunmap_ops[offset+i].handle = INVALID_GRANT_HANDLE;
+		}
+	}
+
+	/*
+	 * Decrease the live-grant counter.  This must happen after the loop to
+	 * prevent premature reuse of the grants by gnttab_mmap().
+	 */
+	live_grants = atomic_sub_return(successful_unmaps, &map->live_grants);
+	if (WARN_ON(live_grants < 0))
+		pr_err("%s: live_grants became negative (%d) after unmapping %d pages!\n",
+		       __func__, live_grants, successful_unmaps);
+
+	/* Release reference taken by __unmap_grant_pages */
+	gntdev_put_map(NULL, map);
+}
 
+static void __unmap_grant_pages(struct gntdev_grant_map *map, int offset,
+			       int pages)
+{
 	if (map->notify.flags & UNMAP_NOTIFY_CLEAR_BYTE) {
 		int pgno = (map->notify.addr >> PAGE_SHIFT);
+
 		if (pgno >= offset && pgno < offset + pages) {
 			/* No need for kmap, pages are in lowmem */
 			uint8_t *tmp = pfn_to_kaddr(page_to_pfn(map->pages[pgno]));
+
 			tmp[map->notify.addr & (PAGE_SIZE-1)] = 0;
 			map->notify.flags &= ~UNMAP_NOTIFY_CLEAR_BYTE;
 		}
 	}
 
-	unmap_data.unmap_ops = map->unmap_ops + offset;
-	unmap_data.kunmap_ops = use_ptemod ? map->kunmap_ops + offset : NULL;
-	unmap_data.pages = map->pages + offset;
-	unmap_data.count = pages;
+	map->unmap_data.unmap_ops = map->unmap_ops + offset;
+	map->unmap_data.kunmap_ops = xen_pv_domain() ? map->kunmap_ops + offset : NULL;
+	map->unmap_data.pages = map->pages + offset;
+	map->unmap_data.count = pages;
+	map->unmap_data.done = __unmap_grant_pages_done;
+	map->unmap_data.data = map;
+	refcount_inc(&map->users); /* to keep map alive during async call below */
 
-	err = gnttab_unmap_refs_sync(&unmap_data);
-	if (err)
-		return err;
-
-	for (i = 0; i < pages; i++) {
-		if (map->unmap_ops[offset+i].status)
-			err = -EINVAL;
-		pr_debug("unmap handle=%d st=%d\n",
-			map->unmap_ops[offset+i].handle,
-			map->unmap_ops[offset+i].status);
-		map->unmap_ops[offset+i].handle = -1;
-	}
-	return err;
+	gnttab_unmap_refs_async(&map->unmap_data);
 }
 
-static int unmap_grant_pages(struct gntdev_grant_map *map, int offset,
-			     int pages)
+static void unmap_grant_pages(struct gntdev_grant_map *map, int offset,
+			      int pages)
 {
-	int range, err = 0;
+	int range;
+
+	if (atomic_read(&map->live_grants) == 0)
+		return; /* Nothing to do */
 
 	pr_debug("unmap %d+%d [%d+%d]\n", map->index, map->count, offset, pages);
 
 	/* It is possible the requested range will have a "hole" where we
 	 * already unmapped some of the grants. Only unmap valid ranges.
 	 */
-	while (pages && !err) {
-		while (pages && map->unmap_ops[offset].handle == -1) {
+	while (pages) {
+		while (pages && map->being_removed[offset]) {
 			offset++;
 			pages--;
 		}
 		range = 0;
 		while (range < pages) {
-			if (map->unmap_ops[offset+range].handle == -1)
+			if (map->being_removed[offset + range])
 				break;
+			map->being_removed[offset + range] = true;
 			range++;
 		}
-		err = __unmap_grant_pages(map, offset, range);
+		if (range)
+			__unmap_grant_pages(map, offset, range);
 		offset += range;
 		pages -= range;
 	}
-
-	return err;
 }
 
 /* ------------------------------------------------------------------ */
@@ -447,23 +521,12 @@ static void gntdev_vma_close(struct vm_area_struct *vma)
 	struct gntdev_priv *priv = file->private_data;
 
 	pr_debug("gntdev_vma_close %p\n", vma);
-	if (use_ptemod) {
-		/* It is possible that an mmu notifier could be running
-		 * concurrently, so take priv->lock to ensure that the vma won't
-		 * vanishing during the unmap_grant_pages call, since we will
-		 * spin here until that completes. Such a concurrent call will
-		 * not do any unmapping, since that has been done prior to
-		 * closing the vma, but it may still iterate the unmap_ops list.
-		 */
-		mutex_lock(&priv->lock);
-		map->vma = NULL;
-		mutex_unlock(&priv->lock);
-	}
+
 	vma->vm_private_data = NULL;
 	gntdev_put_map(priv, map);
 }
 
-static struct page *gntdev_vma_find_special_page(struct vm_area_struct *vma,
+static struct page *gntdev_vma_find_normal_page(struct vm_area_struct *vma,
 						 unsigned long addr)
 {
 	struct gntdev_grant_map *map = vma->vm_private_data;
@@ -474,114 +537,48 @@ static struct page *gntdev_vma_find_special_page(struct vm_area_struct *vma,
 static const struct vm_operations_struct gntdev_vmops = {
 	.open = gntdev_vma_open,
 	.close = gntdev_vma_close,
-	.find_special_page = gntdev_vma_find_special_page,
+	.find_normal_page = gntdev_vma_find_normal_page,
 };
 
 /* ------------------------------------------------------------------ */
 
-static bool in_range(struct gntdev_grant_map *map,
-			      unsigned long start, unsigned long end)
-{
-	if (!map->vma)
-		return false;
-	if (map->vma->vm_start >= end)
-		return false;
-	if (map->vma->vm_end <= start)
-		return false;
-
-	return true;
-}
-
-static int unmap_if_in_range(struct gntdev_grant_map *map,
-			      unsigned long start, unsigned long end,
-			      bool blockable)
+static bool gntdev_invalidate(struct mmu_interval_notifier *mn,
+			      const struct mmu_notifier_range *range,
+			      unsigned long cur_seq)
 {
+	struct gntdev_grant_map *map =
+		container_of(mn, struct gntdev_grant_map, notifier);
 	unsigned long mstart, mend;
-	int err;
-
-	if (!in_range(map, start, end))
-		return 0;
-
-	if (!blockable)
-		return -EAGAIN;
-
-	mstart = max(start, map->vma->vm_start);
-	mend   = min(end,   map->vma->vm_end);
-	pr_debug("map %d+%d (%lx %lx), range %lx %lx, mrange %lx %lx\n",
-			map->index, map->count,
-			map->vma->vm_start, map->vma->vm_end,
-			start, end, mstart, mend);
-	err = unmap_grant_pages(map,
-				(mstart - map->vma->vm_start) >> PAGE_SHIFT,
-				(mend - mstart) >> PAGE_SHIFT);
-	WARN_ON(err);
-
-	return 0;
-}
-
-static int mn_invl_range_start(struct mmu_notifier *mn,
-			       const struct mmu_notifier_range *range)
-{
-	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
-	struct gntdev_grant_map *map;
-	int ret = 0;
+	unsigned long map_start, map_end;
 
-	if (range->blockable)
-		mutex_lock(&priv->lock);
-	else if (!mutex_trylock(&priv->lock))
-		return -EAGAIN;
-
-	list_for_each_entry(map, &priv->maps, next) {
-		ret = unmap_if_in_range(map, range->start, range->end,
-					range->blockable);
-		if (ret)
-			goto out_unlock;
-	}
-	list_for_each_entry(map, &priv->freeable_maps, next) {
-		ret = unmap_if_in_range(map, range->start, range->end,
-					range->blockable);
-		if (ret)
-			goto out_unlock;
-	}
+	if (!mmu_notifier_range_blockable(range))
+		return false;
 
-out_unlock:
-	mutex_unlock(&priv->lock);
+	map_start = map->pages_vm_start;
+	map_end = map->pages_vm_start + (map->count << PAGE_SHIFT);
 
-	return ret;
-}
+	/*
+	 * If the VMA is split or otherwise changed the notifier is not
+	 * updated, but we don't want to process VA's outside the modified
+	 * VMA. FIXME: It would be much more understandable to just prevent
+	 * modifying the VMA in the first place.
+	 */
+	if (map_start >= range->end || map_end <= range->start)
+		return true;
 
-static void mn_release(struct mmu_notifier *mn,
-		       struct mm_struct *mm)
-{
-	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
-	struct gntdev_grant_map *map;
-	int err;
+	mstart = max(range->start, map_start);
+	mend = min(range->end, map_end);
+	pr_debug("map %d+%d (%lx %lx), range %lx %lx, mrange %lx %lx\n",
+		 map->index, map->count, map_start, map_end,
+		 range->start, range->end, mstart, mend);
+	unmap_grant_pages(map, (mstart - map_start) >> PAGE_SHIFT,
+			  (mend - mstart) >> PAGE_SHIFT);
 
-	mutex_lock(&priv->lock);
-	list_for_each_entry(map, &priv->maps, next) {
-		if (!map->vma)
-			continue;
-		pr_debug("map %d+%d (%lx %lx)\n",
-				map->index, map->count,
-				map->vma->vm_start, map->vma->vm_end);
-		err = unmap_grant_pages(map, /* offset */ 0, map->count);
-		WARN_ON(err);
-	}
-	list_for_each_entry(map, &priv->freeable_maps, next) {
-		if (!map->vma)
-			continue;
-		pr_debug("map %d+%d (%lx %lx)\n",
-				map->index, map->count,
-				map->vma->vm_start, map->vma->vm_end);
-		err = unmap_grant_pages(map, /* offset */ 0, map->count);
-		WARN_ON(err);
-	}
-	mutex_unlock(&priv->lock);
+	return true;
 }
 
-static const struct mmu_notifier_ops gntdev_mmu_ops = {
-	.release                = mn_release,
-	.invalidate_range_start = mn_invl_range_start,
+static const struct mmu_interval_notifier_ops gntdev_mmu_ops = {
+	.invalidate = gntdev_invalidate,
 };
 
 /* ------------------------------------------------------------------ */
@@ -589,52 +586,30 @@ static const struct mmu_notifier_ops gntdev_mmu_ops = {
 static int gntdev_open(struct inode *inode, struct file *flip)
 {
 	struct gntdev_priv *priv;
-	int ret = 0;
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv)
 		return -ENOMEM;
 
 	INIT_LIST_HEAD(&priv->maps);
-	INIT_LIST_HEAD(&priv->freeable_maps);
 	mutex_init(&priv->lock);
 
+	mutex_init(&priv->batch_lock);
+
 #ifdef CONFIG_XEN_GNTDEV_DMABUF
-	priv->dmabuf_priv = gntdev_dmabuf_init();
+	priv->dmabuf_priv = gntdev_dmabuf_init(flip);
 	if (IS_ERR(priv->dmabuf_priv)) {
-		ret = PTR_ERR(priv->dmabuf_priv);
-		kfree(priv);
-		return ret;
-	}
-#endif
+		int ret = PTR_ERR(priv->dmabuf_priv);
 
-	if (use_ptemod) {
-		priv->mm = get_task_mm(current);
-		if (!priv->mm) {
-			kfree(priv);
-			return -ENOMEM;
-		}
-		priv->mn.ops = &gntdev_mmu_ops;
-		ret = mmu_notifier_register(&priv->mn, priv->mm);
-		mmput(priv->mm);
-	}
-
-	if (ret) {
 		kfree(priv);
 		return ret;
 	}
+#endif
 
 	flip->private_data = priv;
 #ifdef CONFIG_XEN_GRANT_DMA_ALLOC
 	priv->dma_dev = gntdev_miscdev.this_device;
-
-	/*
-	 * The device is not spawn from a device tree, so arch_setup_dma_ops
-	 * is not called, thus leaving the device with dummy DMA ops.
-	 * Fix this by calling of_dma_configure() with a NULL node to set
-	 * default DMA ops.
-	 */
-	of_dma_configure(priv->dma_dev, NULL, true);
+	dma_coerce_mask_and_coherent(priv->dma_dev, DMA_BIT_MASK(64));
 #endif
 	pr_debug("priv %p\n", priv);
 
@@ -645,6 +620,7 @@ static int gntdev_release(struct inode *inode, struct file *flip)
 {
 	struct gntdev_priv *priv = flip->private_data;
 	struct gntdev_grant_map *map;
+	struct gntdev_copy_batch *batch;
 
 	pr_debug("priv %p\n", priv);
 
@@ -655,16 +631,20 @@ static int gntdev_release(struct inode *inode, struct file *flip)
 		list_del(&map->next);
 		gntdev_put_map(NULL /* already removed */, map);
 	}
-	WARN_ON(!list_empty(&priv->freeable_maps));
 	mutex_unlock(&priv->lock);
 
+	mutex_lock(&priv->batch_lock);
+	while (priv->batch) {
+		batch = priv->batch;
+		priv->batch = batch->next;
+		kfree(batch);
+	}
+	mutex_unlock(&priv->batch_lock);
+
 #ifdef CONFIG_XEN_GNTDEV_DMABUF
 	gntdev_dmabuf_fini(priv->dmabuf_priv);
 #endif
 
-	if (use_ptemod)
-		mmu_notifier_unregister(&priv->mn, priv->mm);
-
 	kfree(priv);
 	return 0;
 }
@@ -679,7 +659,7 @@ static long gntdev_ioctl_map_grant_ref(struct gntdev_priv *priv,
 	if (copy_from_user(&op, u, sizeof(op)) != 0)
 		return -EFAULT;
 	pr_debug("priv %p, add %d\n", priv, op.count);
-	if (unlikely(op.count <= 0))
+	if (unlikely(gntdev_test_page_count(op.count)))
 		return -EINVAL;
 
 	err = -ENOMEM;
@@ -687,12 +667,6 @@ static long gntdev_ioctl_map_grant_ref(struct gntdev_priv *priv,
 	if (!map)
 		return err;
 
-	if (unlikely(gntdev_account_mapped_pages(op.count))) {
-		pr_debug("can't map: over limit\n");
-		gntdev_put_map(NULL, map);
-		return err;
-	}
-
 	if (copy_from_user(map->grants, &u->refs,
 			   sizeof(map->grants[0]) * op.count) != 0) {
 		gntdev_put_map(NULL, map);
@@ -725,8 +699,6 @@ static long gntdev_ioctl_unmap_grant_ref(struct gntdev_priv *priv,
 	map = gntdev_find_map_index(priv, op.index >> PAGE_SHIFT, op.count);
 	if (map) {
 		list_del(&map->next);
-		if (populate_freeable_maps)
-			list_add_tail(&map->next, &priv->freeable_maps);
 		err = 0;
 	}
 	mutex_unlock(&priv->lock);
@@ -747,7 +719,7 @@ static long gntdev_ioctl_get_offset_for_vaddr(struct gntdev_priv *priv,
 		return -EFAULT;
 	pr_debug("priv %p, offset for vaddr %lx\n", priv, (unsigned long)op.vaddr);
 
-	down_read(&current->mm->mmap_sem);
+	mmap_read_lock(current->mm);
 	vma = find_vma(current->mm, op.vaddr);
 	if (!vma || vma->vm_ops != &gntdev_vmops)
 		goto out_unlock;
@@ -761,7 +733,7 @@ static long gntdev_ioctl_get_offset_for_vaddr(struct gntdev_priv *priv,
 	rv = 0;
 
  out_unlock:
-	up_read(&current->mm->mmap_sem);
+	mmap_read_unlock(current->mm);
 
 	if (rv == 0 && copy_to_user(u, &op, sizeof(op)) != 0)
 		return -EFAULT;
@@ -774,7 +746,7 @@ static long gntdev_ioctl_notify(struct gntdev_priv *priv, void __user *u)
 	struct gntdev_grant_map *map;
 	int rc;
 	int out_flags;
-	unsigned int out_event;
+	evtchn_port_t out_event;
 
 	if (copy_from_user(&op, u, sizeof(op)))
 		return -EFAULT;
@@ -834,25 +806,15 @@ static long gntdev_ioctl_notify(struct gntdev_priv *priv, void __user *u)
 	return rc;
 }
 
-#define GNTDEV_COPY_BATCH 16
-
-struct gntdev_copy_batch {
-	struct gnttab_copy ops[GNTDEV_COPY_BATCH];
-	struct page *pages[GNTDEV_COPY_BATCH];
-	s16 __user *status[GNTDEV_COPY_BATCH];
-	unsigned int nr_ops;
-	unsigned int nr_pages;
-};
-
 static int gntdev_get_page(struct gntdev_copy_batch *batch, void __user *virt,
-			   bool writeable, unsigned long *gfn)
+				unsigned long *gfn)
 {
 	unsigned long addr = (unsigned long)virt;
 	struct page *page;
 	unsigned long xen_pfn;
 	int ret;
 
-	ret = get_user_pages_fast(addr, 1, writeable, &page);
+	ret = pin_user_pages_fast(addr, 1, batch->writeable ? FOLL_WRITE : 0, &page);
 	if (ret < 0)
 		return ret;
 
@@ -866,11 +828,9 @@ static int gntdev_get_page(struct gntdev_copy_batch *batch, void __user *virt,
 
 static void gntdev_put_pages(struct gntdev_copy_batch *batch)
 {
-	unsigned int i;
-
-	for (i = 0; i < batch->nr_pages; i++)
-		put_page(batch->pages[i]);
+	unpin_user_pages_dirty_lock(batch->pages, batch->nr_pages, batch->writeable);
 	batch->nr_pages = 0;
+	batch->writeable = false;
 }
 
 static int gntdev_copy(struct gntdev_copy_batch *batch)
@@ -959,8 +919,9 @@ static int gntdev_grant_copy_seg(struct gntdev_copy_batch *batch,
 			virt = seg->source.virt + copied;
 			off = (unsigned long)virt & ~XEN_PAGE_MASK;
 			len = min(len, (size_t)XEN_PAGE_SIZE - off);
+			batch->writeable = false;
 
-			ret = gntdev_get_page(batch, virt, false, &gfn);
+			ret = gntdev_get_page(batch, virt, &gfn);
 			if (ret < 0)
 				return ret;
 
@@ -978,8 +939,9 @@ static int gntdev_grant_copy_seg(struct gntdev_copy_batch *batch,
 			virt = seg->dest.virt + copied;
 			off = (unsigned long)virt & ~XEN_PAGE_MASK;
 			len = min(len, (size_t)XEN_PAGE_SIZE - off);
+			batch->writeable = true;
 
-			ret = gntdev_get_page(batch, virt, true, &gfn);
+			ret = gntdev_get_page(batch, virt, &gfn);
 			if (ret < 0)
 				return ret;
 
@@ -1001,36 +963,53 @@ static int gntdev_grant_copy_seg(struct gntdev_copy_batch *batch,
 static long gntdev_ioctl_grant_copy(struct gntdev_priv *priv, void __user *u)
 {
 	struct ioctl_gntdev_grant_copy copy;
-	struct gntdev_copy_batch batch;
+	struct gntdev_copy_batch *batch;
 	unsigned int i;
 	int ret = 0;
 
 	if (copy_from_user(&copy, u, sizeof(copy)))
 		return -EFAULT;
 
-	batch.nr_ops = 0;
-	batch.nr_pages = 0;
+	mutex_lock(&priv->batch_lock);
+	if (!priv->batch) {
+		batch = kmalloc(sizeof(*batch), GFP_KERNEL);
+	} else {
+		batch = priv->batch;
+		priv->batch = batch->next;
+	}
+	mutex_unlock(&priv->batch_lock);
+	if (!batch)
+		return -ENOMEM;
+
+	batch->nr_ops = 0;
+	batch->nr_pages = 0;
 
 	for (i = 0; i < copy.count; i++) {
 		struct gntdev_grant_copy_segment seg;
 
 		if (copy_from_user(&seg, &copy.segments[i], sizeof(seg))) {
 			ret = -EFAULT;
+			gntdev_put_pages(batch);
 			goto out;
 		}
 
-		ret = gntdev_grant_copy_seg(&batch, &seg, &copy.segments[i].status);
-		if (ret < 0)
+		ret = gntdev_grant_copy_seg(batch, &seg, &copy.segments[i].status);
+		if (ret < 0) {
+			gntdev_put_pages(batch);
 			goto out;
+		}
 
 		cond_resched();
 	}
-	if (batch.nr_ops)
-		ret = gntdev_copy(&batch);
-	return ret;
+	if (batch->nr_ops)
+		ret = gntdev_copy(batch);
+
+ out:
+	mutex_lock(&priv->batch_lock);
+	batch->next = priv->batch;
+	priv->batch = batch;
+	mutex_unlock(&priv->batch_lock);
 
-  out:
-	gntdev_put_pages(&batch);
 	return ret;
 }
 
@@ -1058,7 +1037,7 @@ static long gntdev_ioctl(struct file *flip,
 
 #ifdef CONFIG_XEN_GNTDEV_DMABUF
 	case IOCTL_GNTDEV_DMABUF_EXP_FROM_REFS:
-		return gntdev_ioctl_dmabuf_exp_from_refs(priv, use_ptemod, ptr);
+		return gntdev_ioctl_dmabuf_exp_from_refs(priv, ptr);
 
 	case IOCTL_GNTDEV_DMABUF_EXP_WAIT_RELEASED:
 		return gntdev_ioctl_dmabuf_exp_wait_released(priv, ptr);
@@ -1084,39 +1063,31 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma)
 	int index = vma->vm_pgoff;
 	int count = vma_pages(vma);
 	struct gntdev_grant_map *map;
-	int i, err = -EINVAL;
+	int err = -EINVAL;
 
 	if ((vma->vm_flags & VM_WRITE) && !(vma->vm_flags & VM_SHARED))
 		return -EINVAL;
 
 	pr_debug("map %d+%d at %lx (pgoff %lx)\n",
-			index, count, vma->vm_start, vma->vm_pgoff);
+		 index, count, vma->vm_start, vma->vm_pgoff);
 
 	mutex_lock(&priv->lock);
 	map = gntdev_find_map_index(priv, index, count);
 	if (!map)
 		goto unlock_out;
-	if (use_ptemod && map->vma)
+	if (!atomic_add_unless(&map->in_use, 1, 1))
 		goto unlock_out;
-	if (use_ptemod && priv->mm != vma->vm_mm) {
-		pr_warn("Huh? Other mm?\n");
-		goto unlock_out;
-	}
 
 	refcount_inc(&map->users);
 
 	vma->vm_ops = &gntdev_vmops;
 
-	vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP | VM_MIXEDMAP;
+	vm_flags_set(vma, VM_DONTEXPAND | VM_DONTDUMP | VM_MIXEDMAP);
 
-	if (use_ptemod)
-		vma->vm_flags |= VM_DONTCOPY;
+	if (xen_pv_domain())
+		vm_flags_set(vma, VM_DONTCOPY);
 
 	vma->vm_private_data = map;
-
-	if (use_ptemod)
-		map->vma = vma;
-
 	if (map->flags) {
 		if ((vma->vm_flags & VM_WRITE) &&
 				(map->flags & GNTMAP_readonly))
@@ -1127,10 +1098,32 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma)
 			map->flags |= GNTMAP_readonly;
 	}
 
+	map->pages_vm_start = vma->vm_start;
+
+	if (xen_pv_domain()) {
+		err = mmu_interval_notifier_insert_locked(
+			&map->notifier, vma->vm_mm, vma->vm_start,
+			vma->vm_end - vma->vm_start, &gntdev_mmu_ops);
+		if (err)
+			goto out_unlock_put;
+
+		map->notifier_init = true;
+	}
 	mutex_unlock(&priv->lock);
 
-	if (use_ptemod) {
-		map->pages_vm_start = vma->vm_start;
+	if (xen_pv_domain()) {
+		/*
+		 * gntdev takes the address of the PTE in find_grant_ptes() and
+		 * passes it to the hypervisor in gntdev_map_grant_pages(). The
+		 * purpose of the notifier is to prevent the hypervisor pointer
+		 * to the PTE from going stale.
+		 *
+		 * Since this vma's mappings can't be touched without the
+		 * mmap_lock, and we are holding it now, there is no need for
+		 * the notifier_range locking pattern.
+		 */
+		mmu_interval_read_begin(&map->notifier);
+
 		err = apply_to_page_range(vma->vm_mm, vma->vm_start,
 					  vma->vm_end - vma->vm_start,
 					  find_grant_ptes, map);
@@ -1144,30 +1137,10 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma)
 	if (err)
 		goto out_put_map;
 
-	if (!use_ptemod) {
-		for (i = 0; i < count; i++) {
-			err = vm_insert_page(vma, vma->vm_start + i*PAGE_SIZE,
-				map->pages[i]);
-			if (err)
-				goto out_put_map;
-		}
-	} else {
-#ifdef CONFIG_X86
-		/*
-		 * If the PTEs were not made special by the grant map
-		 * hypercall, do so here.
-		 *
-		 * This is racy since the mapping is already visible
-		 * to userspace but userspace should be well-behaved
-		 * enough to not touch it until the mmap() call
-		 * returns.
-		 */
-		if (!xen_feature(XENFEAT_gnttab_map_avail_bits)) {
-			apply_to_page_range(vma->vm_mm, vma->vm_start,
-					    vma->vm_end - vma->vm_start,
-					    set_grant_ptes_as_special, NULL);
-		}
-#endif
+	if (!xen_pv_domain()) {
+		err = vm_map_pages_zero(vma, map->pages, map->count);
+		if (err)
+			goto out_put_map;
 	}
 
 	return 0;
@@ -1179,10 +1152,8 @@ unlock_out:
 out_unlock_put:
 	mutex_unlock(&priv->lock);
 out_put_map:
-	if (use_ptemod) {
-		map->vma = NULL;
+	if (xen_pv_domain())
 		unmap_grant_pages(map, 0, map->count);
-	}
 	gntdev_put_map(priv, map);
 	return err;
 }
@@ -1210,8 +1181,6 @@ static int __init gntdev_init(void)
 	if (!xen_domain())
 		return -ENODEV;
 
-	use_ptemod = !xen_feature(XENFEAT_auto_translated_physmap);
-
 	err = misc_register(&gntdev_miscdev);
 	if (err != 0) {
 		pr_err("Could not register gntdev device\n");