[PATCH v10 65/69] nommu: remove uses of VMA linked list

[Liam Howlett]

From: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx>

Use the maple tree or VMA iterator instead.  This is faster and will allow
us to shrink the VMA.

Link: https://lkml.kernel.org/r/20220504011345.662299-50-Liam.Howlett@xxxxxxxxxx
Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Signed-off-by: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx>
Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
Cc: David Howells <dhowells@xxxxxxxxxx>
Cc: SeongJae Park <sj@xxxxxxxxxx>
Cc: Will Deacon <will@xxxxxxxxxx>
Cc: Davidlohr Bueso <dave@xxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
 mm/nommu.c | 135 +++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 101 insertions(+), 34 deletions(-)

diff --git a/mm/nommu.c b/mm/nommu.c
index 4ae252a785dc..f6b187090d95 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -557,26 +557,14 @@ void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas)
 	mas_store_prealloc(mas, NULL);
 }
 
-/*
- * add a VMA into a process's mm_struct in the appropriate place in the list
- * and tree and add to the address space's page tree also if not an anonymous
- * page
- * - should be called with mm->mmap_lock held writelocked
- */
-static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma)
+static void setup_vma_to_mm(struct vm_area_struct *vma, struct mm_struct *mm)
 {
-	struct address_space *mapping;
-	struct vm_area_struct *prev;
-	MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_end);
-
-	BUG_ON(!vma->vm_region);
-
 	mm->map_count++;
 	vma->vm_mm = mm;
 
 	/* add the VMA to the mapping */
 	if (vma->vm_file) {
-		mapping = vma->vm_file->f_mapping;
+		struct address_space *mapping = vma->vm_file->f_mapping;
 
 		i_mmap_lock_write(mapping);
 		flush_dcache_mmap_lock(mapping);
@@ -584,21 +572,52 @@ static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma)
 		flush_dcache_mmap_unlock(mapping);
 		i_mmap_unlock_write(mapping);
 	}
+}
 
-	prev = mas_prev(&mas, 0);
-	mas_reset(&mas);
+/*
+ * mas_add_vma_to_mm() - Maple state variant of add_mas_to_mm().
+ * @mas: The maple state with preallocations.
+ * @mm: The mm_struct
+ * @vma: The vma to add
+ *
+ */
+static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm,
+			      struct vm_area_struct *vma)
+{
+	struct vm_area_struct *prev;
+
+	BUG_ON(!vma->vm_region);
+
+	setup_vma_to_mm(vma, mm);
+
+	prev = mas_prev(mas, 0);
+	mas_reset(mas);
 	/* add the VMA to the tree */
-	vma_mas_store(vma, &mas);
+	vma_mas_store(vma, mas);
 	__vma_link_list(mm, vma, prev);
 }
 
 /*
- * delete a VMA from its owning mm_struct and address space
+ * add a VMA into a process's mm_struct in the appropriate place in the list
+ * and tree and add to the address space's page tree also if not an anonymous
+ * page
+ * - should be called with mm->mmap_lock held writelocked
  */
-static void delete_vma_from_mm(struct vm_area_struct *vma)
+static int add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma)
 {
-	MA_STATE(mas, &vma->vm_mm->mm_mt, 0, 0);
+	MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_end);
+
+	if (mas_preallocate(&mas, vma, GFP_KERNEL)) {
+		pr_warn("Allocation of vma tree for process %d failed\n",
+		       current->pid);
+		return -ENOMEM;
+	}
+	mas_add_vma_to_mm(&mas, mm, vma);
+	return 0;
+}
 
+static void cleanup_vma_from_mm(struct vm_area_struct *vma)
+{
 	vma->vm_mm->map_count--;
 	/* remove the VMA from the mapping */
 	if (vma->vm_file) {
@@ -611,10 +630,25 @@ static void delete_vma_from_mm(struct vm_area_struct *vma)
 		flush_dcache_mmap_unlock(mapping);
 		i_mmap_unlock_write(mapping);
 	}
+}
+/*
+ * delete a VMA from its owning mm_struct and address space
+ */
+static int delete_vma_from_mm(struct vm_area_struct *vma)
+{
+	MA_STATE(mas, &vma->vm_mm->mm_mt, 0, 0);
+
+	if (mas_preallocate(&mas, vma, GFP_KERNEL)) {
+		pr_warn("Allocation of vma tree for process %d failed\n",
+		       current->pid);
+		return -ENOMEM;
+	}
+	cleanup_vma_from_mm(vma);
 
 	/* remove from the MM's tree and list */
 	vma_mas_remove(vma, &mas);
 	__vma_unlink_list(vma->vm_mm, vma);
+	return 0;
 }
 
 /*
@@ -1024,6 +1058,7 @@ unsigned long do_mmap(struct file *file,
 	vm_flags_t vm_flags;
 	unsigned long capabilities, result;
 	int ret;
+	MA_STATE(mas, &current->mm->mm_mt, 0, 0);
 
 	*populate = 0;
 
@@ -1042,6 +1077,7 @@ unsigned long do_mmap(struct file *file,
 	 * now know into VMA flags */
 	vm_flags = determine_vm_flags(file, prot, flags, capabilities);
 
+
 	/* we're going to need to record the mapping */
 	region = kmem_cache_zalloc(vm_region_jar, GFP_KERNEL);
 	if (!region)
@@ -1051,6 +1087,9 @@ unsigned long do_mmap(struct file *file,
 	if (!vma)
 		goto error_getting_vma;
 
+	if (mas_preallocate(&mas, vma, GFP_KERNEL))
+		goto error_maple_preallocate;
+
 	region->vm_usage = 1;
 	region->vm_flags = vm_flags;
 	region->vm_pgoff = pgoff;
@@ -1191,7 +1230,7 @@ unsigned long do_mmap(struct file *file,
 	current->mm->total_vm += len >> PAGE_SHIFT;
 
 share:
-	add_vma_to_mm(current->mm, vma);
+	mas_add_vma_to_mm(&mas, current->mm, vma);
 
 	/* we flush the region from the icache only when the first executable
 	 * mapping of it is made  */
@@ -1217,6 +1256,7 @@ unsigned long do_mmap(struct file *file,
 
 sharing_violation:
 	up_write(&nommu_region_sem);
+	mas_destroy(&mas);
 	pr_warn("Attempt to share mismatched mappings\n");
 	ret = -EINVAL;
 	goto error;
@@ -1233,6 +1273,14 @@ unsigned long do_mmap(struct file *file,
 			len, current->pid);
 	show_free_areas(0, NULL);
 	return -ENOMEM;
+
+error_maple_preallocate:
+	kmem_cache_free(vm_region_jar, region);
+	vm_area_free(vma);
+	pr_warn("Allocation of vma tree for process %d failed\n", current->pid);
+	show_free_areas(0, NULL);
+	return -ENOMEM;
+
 }
 
 unsigned long ksys_mmap_pgoff(unsigned long addr, unsigned long len,
@@ -1298,6 +1346,7 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
 	struct vm_area_struct *new;
 	struct vm_region *region;
 	unsigned long npages;
+	MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_end);
 
 	/* we're only permitted to split anonymous regions (these should have
 	 * only a single usage on the region) */
@@ -1333,7 +1382,6 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
 	if (new->vm_ops && new->vm_ops->open)
 		new->vm_ops->open(new);
 
-	delete_vma_from_mm(vma);
 	down_write(&nommu_region_sem);
 	delete_nommu_region(vma->vm_region);
 	if (new_below) {
@@ -1346,8 +1394,17 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
 	add_nommu_region(vma->vm_region);
 	add_nommu_region(new->vm_region);
 	up_write(&nommu_region_sem);
-	add_vma_to_mm(mm, vma);
-	add_vma_to_mm(mm, new);
+	if (mas_preallocate(&mas, vma, GFP_KERNEL)) {
+		pr_warn("Allocation of vma tree for process %d failed\n",
+		       current->pid);
+		return -ENOMEM;
+	}
+
+	setup_vma_to_mm(vma, mm);
+	setup_vma_to_mm(new, mm);
+	mas_set_range(&mas, vma->vm_start, vma->vm_end - 1);
+	mas_store(&mas, vma);
+	vma_mas_store(new, &mas);
 	return 0;
 }
 
@@ -1363,12 +1420,14 @@ static int shrink_vma(struct mm_struct *mm,
 
 	/* adjust the VMA's pointers, which may reposition it in the MM's tree
 	 * and list */
-	delete_vma_from_mm(vma);
+	if (delete_vma_from_mm(vma))
+		return -ENOMEM;
 	if (from > vma->vm_start)
 		vma->vm_end = from;
 	else
 		vma->vm_start = to;
-	add_vma_to_mm(mm, vma);
+	if (add_vma_to_mm(mm, vma))
+		return -ENOMEM;
 
 	/* cut the backing region down to size */
 	region = vma->vm_region;
@@ -1396,9 +1455,10 @@ static int shrink_vma(struct mm_struct *mm,
  */
 int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list_head *uf)
 {
+	MA_STATE(mas, &mm->mm_mt, start, start);
 	struct vm_area_struct *vma;
 	unsigned long end;
-	int ret;
+	int ret = 0;
 
 	len = PAGE_ALIGN(len);
 	if (len == 0)
@@ -1407,7 +1467,7 @@ int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list
 	end = start + len;
 
 	/* find the first potentially overlapping VMA */
-	vma = find_vma(mm, start);
+	vma = mas_find(&mas, end - 1);
 	if (!vma) {
 		static int limit;
 		if (limit < 5) {
@@ -1426,7 +1486,7 @@ int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list
 				return -EINVAL;
 			if (end == vma->vm_end)
 				goto erase_whole_vma;
-			vma = vma->vm_next;
+			vma = mas_next(&mas, end - 1);
 		} while (vma);
 		return -EINVAL;
 	} else {
@@ -1448,9 +1508,10 @@ int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list
 	}
 
 erase_whole_vma:
-	delete_vma_from_mm(vma);
+	if (delete_vma_from_mm(vma))
+		ret = -ENOMEM;
 	delete_vma(mm, vma);
-	return 0;
+	return ret;
 }
 
 int vm_munmap(unsigned long addr, size_t len)
@@ -1475,6 +1536,7 @@ SYSCALL_DEFINE2(munmap, unsigned long, addr, size_t, len)
  */
 void exit_mmap(struct mm_struct *mm)
 {
+	VMA_ITERATOR(vmi, mm, 0);
 	struct vm_area_struct *vma;
 
 	if (!mm)
@@ -1482,13 +1544,18 @@ void exit_mmap(struct mm_struct *mm)
 
 	mm->total_vm = 0;
 
-	while ((vma = mm->mmap)) {
-		mm->mmap = vma->vm_next;
-		delete_vma_from_mm(vma);
+	/*
+	 * Lock the mm to avoid assert complaining even though this is the only
+	 * user of the mm
+	 */
+	mmap_write_lock(mm);
+	for_each_vma(vmi, vma) {
+		cleanup_vma_from_mm(vma);
 		delete_vma(mm, vma);
 		cond_resched();
 	}
 	__mt_destroy(&mm->mm_mt);
+	mmap_write_unlock(mm);
 }
 
 int vm_brk(unsigned long addr, unsigned long len)
-- 
2.35.1