mm/mmap: use advanced maple tree API for mmap_region()

Changing mmap_region() to use the maple tree state and the advanced maple
tree interface allows for a lot less tree walking.

This change removes the last caller of munmap_vma_range(), so drop this
unused function.

Add vma_expand() to expand a VMA if possible by doing the necessary
hugepage check, uprobe_munmap of files, dcache flush, modifications then
undoing the detaches, etc.

Link: https://lkml.kernel.org/r/20220906194824.2110408-25-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Tested-by: Yu Zhao <yuzhao@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: SeongJae Park <sj@kernel.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

mm/mmap.c

@@ -470,28 +470,6 @@ static inline struct vm_area_struct *__vma_next(struct mm_struct *mm,
 	return vma->vm_next;
 }
 
-/*
- * munmap_vma_range() - munmap VMAs that overlap a range.
- * @mm: The mm struct
- * @start: The start of the range.
- * @len: The length of the range.
- * @pprev: pointer to the pointer that will be set to previous vm_area_struct
- *
- * Find all the vm_area_struct that overlap from @start to
- * @end and munmap them.  Set @pprev to the previous vm_area_struct.
- *
- * Returns: -ENOMEM on munmap failure or 0 on success.
- */
-static inline int
-munmap_vma_range(struct mm_struct *mm, unsigned long start, unsigned long len,
-		 struct vm_area_struct **pprev, struct list_head *uf)
-{
-	while (range_has_overlap(mm, start, start + len, pprev))
-		if (do_munmap(mm, start, len, uf))
-			return -ENOMEM;
-	return 0;
-}
-
 static unsigned long count_vma_pages_range(struct mm_struct *mm,
 		unsigned long addr, unsigned long end)
 {
@@ -618,6 +596,129 @@ static void __insert_vm_struct(struct mm_struct *mm, struct ma_state *mas,
 	mm->map_count++;
 }
 
+/*
+ * vma_expand - Expand an existing VMA
+ *
+ * @mas: The maple state
+ * @vma: The vma to expand
+ * @start: The start of the vma
+ * @end: The exclusive end of the vma
+ * @pgoff: The page offset of vma
+ * @next: The current of next vma.
+ *
+ * Expand @vma to @start and @end.  Can expand off the start and end.  Will
+ * expand over @next if it's different from @vma and @end == @next->vm_end.
+ * Checking if the @vma can expand and merge with @next needs to be handled by
+ * the caller.
+ *
+ * Returns: 0 on success
+ */
+inline int vma_expand(struct ma_state *mas, struct vm_area_struct *vma,
+		      unsigned long start, unsigned long end, pgoff_t pgoff,
+		      struct vm_area_struct *next)
+{
+	struct mm_struct *mm = vma->vm_mm;
+	struct address_space *mapping = NULL;
+	struct rb_root_cached *root = NULL;
+	struct anon_vma *anon_vma = vma->anon_vma;
+	struct file *file = vma->vm_file;
+	bool remove_next = false;
+
+	if (next && (vma != next) && (end == next->vm_end)) {
+		remove_next = true;
+		if (next->anon_vma && !vma->anon_vma) {
+			int error;
+
+			anon_vma = next->anon_vma;
+			vma->anon_vma = anon_vma;
+			error = anon_vma_clone(vma, next);
+			if (error)
+				return error;
+		}
+	}
+
+	/* Not merging but overwriting any part of next is not handled. */
+	VM_BUG_ON(next && !remove_next && next != vma && end > next->vm_start);
+	/* Only handles expanding */
+	VM_BUG_ON(vma->vm_start < start || vma->vm_end > end);
+
+	if (mas_preallocate(mas, vma, GFP_KERNEL))
+		goto nomem;
+
+	vma_adjust_trans_huge(vma, start, end, 0);
+
+	if (file) {
+		mapping = file->f_mapping;
+		root = &mapping->i_mmap;
+		uprobe_munmap(vma, vma->vm_start, vma->vm_end);
+		i_mmap_lock_write(mapping);
+	}
+
+	if (anon_vma) {
+		anon_vma_lock_write(anon_vma);
+		anon_vma_interval_tree_pre_update_vma(vma);
+	}
+
+	if (file) {
+		flush_dcache_mmap_lock(mapping);
+		vma_interval_tree_remove(vma, root);
+	}
+
+	vma->vm_start = start;
+	vma->vm_end = end;
+	vma->vm_pgoff = pgoff;
+	/* Note: mas must be pointing to the expanding VMA */
+	vma_mas_store(vma, mas);
+
+	if (file) {
+		vma_interval_tree_insert(vma, root);
+		flush_dcache_mmap_unlock(mapping);
+	}
+
+	/* Expanding over the next vma */
+	if (remove_next) {
+		/* Remove from mm linked list - also updates highest_vm_end */
+		__vma_unlink_list(mm, next);
+
+		/* Kill the cache */
+		vmacache_invalidate(mm);
+
+		if (file)
+			__remove_shared_vm_struct(next, file, mapping);
+
+	} else if (!next) {
+		mm->highest_vm_end = vm_end_gap(vma);
+	}
+
+	if (anon_vma) {
+		anon_vma_interval_tree_post_update_vma(vma);
+		anon_vma_unlock_write(anon_vma);
+	}
+
+	if (file) {
+		i_mmap_unlock_write(mapping);
+		uprobe_mmap(vma);
+	}
+
+	if (remove_next) {
+		if (file) {
+			uprobe_munmap(next, next->vm_start, next->vm_end);
+			fput(file);
+		}
+		if (next->anon_vma)
+			anon_vma_merge(vma, next);
+		mm->map_count--;
+		mpol_put(vma_policy(next));
+		vm_area_free(next);
+	}
+
+	validate_mm(mm);
+	return 0;
+
+nomem:
+	return -ENOMEM;
+}
+
 /*
  * We cannot adjust vm_start, vm_end, vm_pgoff fields of a vma that
  * is already present in an i_mmap tree without adjusting the tree.
@@ -1630,9 +1731,15 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 		struct list_head *uf)
 {
 	struct mm_struct *mm = current->mm;
-	struct vm_area_struct *vma, *prev, *merge;
+	struct vm_area_struct *vma = NULL;
-	int error;
+	struct vm_area_struct *next, *prev, *merge;
+	pgoff_t pglen = len >> PAGE_SHIFT;
 	unsigned long charged = 0;
+	unsigned long end = addr + len;
+	unsigned long merge_start = addr, merge_end = end;
+	pgoff_t vm_pgoff;
+	int error;
+	MA_STATE(mas, &mm->mm_mt, addr, end - 1);
 
 	/* Check against address space limit. */
 	if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) {
@@ -1642,16 +1749,17 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 		 * MAP_FIXED may remove pages of mappings that intersects with
 		 * requested mapping. Account for the pages it would unmap.
 		 */
-		nr_pages = count_vma_pages_range(mm, addr, addr + len);
+		nr_pages = count_vma_pages_range(mm, addr, end);
 
 		if (!may_expand_vm(mm, vm_flags,
 					(len >> PAGE_SHIFT) - nr_pages))
 			return -ENOMEM;
 	}
 
-	/* Clear old maps, set up prev and uf */
+	/* Unmap any existing mapping in the area */
-	if (munmap_vma_range(mm, addr, len, &prev, uf))
+	if (do_munmap(mm, addr, len, uf))
 		return -ENOMEM;
+
 	/*
 	 * Private writable mapping: check memory availability
 	 */
@@ -1662,14 +1770,43 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 		vm_flags |= VM_ACCOUNT;
 	}
 
-	/*
+	next = mas_next(&mas, ULONG_MAX);
-	 * Can we just expand an old mapping?
+	prev = mas_prev(&mas, 0);
-	 */
+	if (vm_flags & VM_SPECIAL)
-	vma = vma_merge(mm, prev, addr, addr + len, vm_flags,
+		goto cannot_expand;
-			NULL, file, pgoff, NULL, NULL_VM_UFFD_CTX, NULL);
-	if (vma)
-		goto out;
 
+	/* Attempt to expand an old mapping */
+	/* Check next */
+	if (next && next->vm_start == end && !vma_policy(next) &&
+	    can_vma_merge_before(next, vm_flags, NULL, file, pgoff+pglen,
+				 NULL_VM_UFFD_CTX, NULL)) {
+		merge_end = next->vm_end;
+		vma = next;
+		vm_pgoff = next->vm_pgoff - pglen;
+	}
+
+	/* Check prev */
+	if (prev && prev->vm_end == addr && !vma_policy(prev) &&
+	    (vma ? can_vma_merge_after(prev, vm_flags, vma->anon_vma, file,
+				       pgoff, vma->vm_userfaultfd_ctx, NULL) :
+		   can_vma_merge_after(prev, vm_flags, NULL, file, pgoff,
+				       NULL_VM_UFFD_CTX, NULL))) {
+		merge_start = prev->vm_start;
+		vma = prev;
+		vm_pgoff = prev->vm_pgoff;
+	}
+
+
+	/* Actually expand, if possible */
+	if (vma &&
+	    !vma_expand(&mas, vma, merge_start, merge_end, vm_pgoff, next)) {
+		khugepaged_enter_vma(vma, vm_flags);
+		goto expanded;
+	}
+
+	mas.index = addr;
+	mas.last = end - 1;
+cannot_expand:
 	/*
 	 * Determine the object being mapped and call the appropriate
 	 * specific mapper. the address has already been validated, but
@@ -1682,7 +1819,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 	}
 
 	vma->vm_start = addr;
-	vma->vm_end = addr + len;
+	vma->vm_end = end;
 	vma->vm_flags = vm_flags;
 	vma->vm_page_prot = vm_get_page_prot(vm_flags);
 	vma->vm_pgoff = pgoff;
@@ -1703,28 +1840,32 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 		 *
 		 * Answer: Yes, several device drivers can do it in their
 		 *         f_op->mmap method. -DaveM
-		 * Bug: If addr is changed, prev, rb_link, rb_parent should
-		 *      be updated for vma_link()
 		 */
 		WARN_ON_ONCE(addr != vma->vm_start);
 
 		addr = vma->vm_start;
+		mas_reset(&mas);
 
-		/* If vm_flags changed after call_mmap(), we should try merge vma again
+		/*
-		 * as we may succeed this time.
+		 * If vm_flags changed after call_mmap(), we should try merge
+		 * vma again as we may succeed this time.
 		 */
 		if (unlikely(vm_flags != vma->vm_flags && prev)) {
 			merge = vma_merge(mm, prev, vma->vm_start, vma->vm_end, vma->vm_flags,
 				NULL, vma->vm_file, vma->vm_pgoff, NULL, NULL_VM_UFFD_CTX, NULL);
 			if (merge) {
-				/* ->mmap() can change vma->vm_file and fput the original file. So
+				/*
-				 * fput the vma->vm_file here or we would add an extra fput for file
+				 * ->mmap() can change vma->vm_file and fput
-				 * and cause general protection fault ultimately.
+				 * the original file. So fput the vma->vm_file
+				 * here or we would add an extra fput for file
+				 * and cause general protection fault
+				 * ultimately.
 				 */
 				fput(vma->vm_file);
 				vm_area_free(vma);
 				vma = merge;
 				/* Update vm_flags to pick up the change. */
+				addr = vma->vm_start;
 				vm_flags = vma->vm_flags;
 				goto unmap_writable;
 			}
@@ -1748,7 +1889,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 			goto free_vma;
 	}
 
-	if (vma_link(mm, vma, prev)) {
+	if (mas_preallocate(&mas, vma, GFP_KERNEL)) {
 		error = -ENOMEM;
 		if (file)
 			goto unmap_and_free_vma;
@@ -1756,6 +1897,22 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
 			goto free_vma;
 	}
 
+	if (vma->vm_file)
+		i_mmap_lock_write(vma->vm_file->f_mapping);
+
+	vma_mas_store(vma, &mas);
+	__vma_link_list(mm, vma, prev);
+	mm->map_count++;
+	if (vma->vm_file) {
+		if (vma->vm_flags & VM_SHARED)
+			mapping_allow_writable(vma->vm_file->f_mapping);
+
+		flush_dcache_mmap_lock(vma->vm_file->f_mapping);
+		vma_interval_tree_insert(vma, &vma->vm_file->f_mapping->i_mmap);
+		flush_dcache_mmap_unlock(vma->vm_file->f_mapping);
+		i_mmap_unlock_write(vma->vm_file->f_mapping);
+	}
+
 	/*
 	 * vma_merge() calls khugepaged_enter_vma() either, the below
 	 * call covers the non-merge case.
@@ -1767,7 +1924,7 @@ unmap_writable:
 	if (file && vm_flags & VM_SHARED)
 		mapping_unmap_writable(file->f_mapping);
 	file = vma->vm_file;
-out:
+expanded:
 	perf_event_mmap(vma);
 
 	vm_stat_account(mm, vm_flags, len >> PAGE_SHIFT);
@@ -1794,6 +1951,7 @@ out:
 
 	vma_set_page_prot(vma);
 
+	validate_mm(mm);
 	return addr;
 
 unmap_and_free_vma:
@@ -1809,6 +1967,7 @@ free_vma:
 unacct_error:
 	if (charged)
 		vm_unacct_memory(charged);
+	validate_mm(mm);
 	return error;
 }
 
@@ -2632,10 +2791,6 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len,
 	prev = vma->vm_prev;
 	/* we have start < vma->vm_end  */
 
-	/* if it doesn't overlap, we have nothing.. */
-	if (vma->vm_start >= end)
-		return 0;
-
 	/*
 	 * If we need to split any vma, do it now to save pain later.
 	 *