Update Linux to v5.4.148

Sourced from [1]

[1] https://cdn.kernel.org/pub/linux/kernel/v5.x/linux-5.4.148.tar.gz

Change-Id: Ib3d26c5ba9b022e2e03533005c4fed4d7c30b61b
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 123ac54..fc688af 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -23,6 +23,15 @@
 #include "btrfs_inode.h"
 #include "transaction.h"
 #include "compression.h"
+#include "xattr.h"
+
+/*
+ * Maximum number of references an extent can have in order for us to attempt to
+ * issue clone operations instead of write operations. This currently exists to
+ * avoid hitting limitations of the backreference walking code (taking a lot of
+ * time and using too much memory for extents with large number of references).
+ */
+#define SEND_MAX_EXTENT_REFS	64
 
 /*
  * A fs_path is a helper to dynamically build path names with unknown size.
@@ -229,6 +238,7 @@
 	 * after this directory is moved, we can try to rmdir the ino rmdir_ino.
 	 */
 	u64 rmdir_ino;
+	u64 rmdir_gen;
 	bool orphanized;
 };
 
@@ -314,7 +324,7 @@
 static struct waiting_dir_move *
 get_waiting_dir_move(struct send_ctx *sctx, u64 ino);
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino);
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 gen);
 
 static int need_send_hole(struct send_ctx *sctx)
 {
@@ -1248,12 +1258,21 @@
 	 */
 	if (found->root == bctx->sctx->send_root) {
 		/*
-		 * TODO for the moment we don't accept clones from the inode
-		 * that is currently send. We may change this when
-		 * BTRFS_IOC_CLONE_RANGE supports cloning from and to the same
-		 * file.
+		 * If the source inode was not yet processed we can't issue a
+		 * clone operation, as the source extent does not exist yet at
+		 * the destination of the stream.
 		 */
-		if (ino >= bctx->cur_objectid)
+		if (ino > bctx->cur_objectid)
+			return 0;
+		/*
+		 * We clone from the inode currently being sent as long as the
+		 * source extent is already processed, otherwise we could try
+		 * to clone from an extent that does not exist yet at the
+		 * destination of the stream.
+		 */
+		if (ino == bctx->cur_objectid &&
+		    offset + bctx->extent_len >
+		    bctx->sctx->cur_inode_next_write_offset)
 			return 0;
 	}
 
@@ -1302,6 +1321,7 @@
 	struct clone_root *cur_clone_root;
 	struct btrfs_key found_key;
 	struct btrfs_path *tmp_path;
+	struct btrfs_extent_item *ei;
 	int compressed;
 	u32 i;
 
@@ -1349,7 +1369,6 @@
 	ret = extent_from_logical(fs_info, disk_byte, tmp_path,
 				  &found_key, &flags);
 	up_read(&fs_info->commit_root_sem);
-	btrfs_release_path(tmp_path);
 
 	if (ret < 0)
 		goto out;
@@ -1358,6 +1377,21 @@
 		goto out;
 	}
 
+	ei = btrfs_item_ptr(tmp_path->nodes[0], tmp_path->slots[0],
+			    struct btrfs_extent_item);
+	/*
+	 * Backreference walking (iterate_extent_inodes() below) is currently
+	 * too expensive when an extent has a large number of references, both
+	 * in time spent and used memory. So for now just fallback to write
+	 * operations instead of clone operations when an extent has more than
+	 * a certain amount of references.
+	 */
+	if (btrfs_extent_refs(tmp_path->nodes[0], ei) > SEND_MAX_EXTENT_REFS) {
+		ret = -ENOENT;
+		goto out;
+	}
+	btrfs_release_path(tmp_path);
+
 	/*
 	 * Setup the clone roots.
 	 */
@@ -2273,7 +2307,7 @@
 
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino)) {
+		if (is_waiting_for_rm(sctx, ino, gen)) {
 			ret = gen_unique_name(sctx, ino, gen, name);
 			if (ret < 0)
 				goto out;
@@ -2832,8 +2866,8 @@
 	return ret;
 }
 
-static struct orphan_dir_info *
-add_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
+static struct orphan_dir_info *add_orphan_dir_info(struct send_ctx *sctx,
+						   u64 dir_ino, u64 dir_gen)
 {
 	struct rb_node **p = &sctx->orphan_dirs.rb_node;
 	struct rb_node *parent = NULL;
@@ -2842,20 +2876,23 @@
 	while (*p) {
 		parent = *p;
 		entry = rb_entry(parent, struct orphan_dir_info, node);
-		if (dir_ino < entry->ino) {
+		if (dir_ino < entry->ino)
 			p = &(*p)->rb_left;
-		} else if (dir_ino > entry->ino) {
+		else if (dir_ino > entry->ino)
 			p = &(*p)->rb_right;
-		} else {
+		else if (dir_gen < entry->gen)
+			p = &(*p)->rb_left;
+		else if (dir_gen > entry->gen)
+			p = &(*p)->rb_right;
+		else
 			return entry;
-		}
 	}
 
 	odi = kmalloc(sizeof(*odi), GFP_KERNEL);
 	if (!odi)
 		return ERR_PTR(-ENOMEM);
 	odi->ino = dir_ino;
-	odi->gen = 0;
+	odi->gen = dir_gen;
 	odi->last_dir_index_offset = 0;
 
 	rb_link_node(&odi->node, parent, p);
@@ -2863,8 +2900,8 @@
 	return odi;
 }
 
-static struct orphan_dir_info *
-get_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
+static struct orphan_dir_info *get_orphan_dir_info(struct send_ctx *sctx,
+						   u64 dir_ino, u64 gen)
 {
 	struct rb_node *n = sctx->orphan_dirs.rb_node;
 	struct orphan_dir_info *entry;
@@ -2875,15 +2912,19 @@
 			n = n->rb_left;
 		else if (dir_ino > entry->ino)
 			n = n->rb_right;
+		else if (gen < entry->gen)
+			n = n->rb_left;
+		else if (gen > entry->gen)
+			n = n->rb_right;
 		else
 			return entry;
 	}
 	return NULL;
 }
 
-static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino)
+static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino, u64 gen)
 {
-	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino);
+	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino, gen);
 
 	return odi != NULL;
 }
@@ -2928,7 +2969,7 @@
 	key.type = BTRFS_DIR_INDEX_KEY;
 	key.offset = 0;
 
-	odi = get_orphan_dir_info(sctx, dir);
+	odi = get_orphan_dir_info(sctx, dir, dir_gen);
 	if (odi)
 		key.offset = odi->last_dir_index_offset;
 
@@ -2959,7 +3000,7 @@
 
 		dm = get_waiting_dir_move(sctx, loc.objectid);
 		if (dm) {
-			odi = add_orphan_dir_info(sctx, dir);
+			odi = add_orphan_dir_info(sctx, dir, dir_gen);
 			if (IS_ERR(odi)) {
 				ret = PTR_ERR(odi);
 				goto out;
@@ -2967,12 +3008,13 @@
 			odi->gen = dir_gen;
 			odi->last_dir_index_offset = found_key.offset;
 			dm->rmdir_ino = dir;
+			dm->rmdir_gen = dir_gen;
 			ret = 0;
 			goto out;
 		}
 
 		if (loc.objectid > send_progress) {
-			odi = add_orphan_dir_info(sctx, dir);
+			odi = add_orphan_dir_info(sctx, dir, dir_gen);
 			if (IS_ERR(odi)) {
 				ret = PTR_ERR(odi);
 				goto out;
@@ -3012,6 +3054,7 @@
 		return -ENOMEM;
 	dm->ino = ino;
 	dm->rmdir_ino = 0;
+	dm->rmdir_gen = 0;
 	dm->orphanized = orphanized;
 
 	while (*p) {
@@ -3157,7 +3200,7 @@
 	while (ino != BTRFS_FIRST_FREE_OBJECTID) {
 		fs_path_reset(name);
 
-		if (is_waiting_for_rm(sctx, ino))
+		if (is_waiting_for_rm(sctx, ino, gen))
 			break;
 		if (is_waiting_for_move(sctx, ino)) {
 			if (*ancestor_ino == 0)
@@ -3197,6 +3240,7 @@
 	u64 parent_ino, parent_gen;
 	struct waiting_dir_move *dm = NULL;
 	u64 rmdir_ino = 0;
+	u64 rmdir_gen;
 	u64 ancestor;
 	bool is_orphan;
 	int ret;
@@ -3211,6 +3255,7 @@
 	dm = get_waiting_dir_move(sctx, pm->ino);
 	ASSERT(dm);
 	rmdir_ino = dm->rmdir_ino;
+	rmdir_gen = dm->rmdir_gen;
 	is_orphan = dm->orphanized;
 	free_waiting_dir_move(sctx, dm);
 
@@ -3247,6 +3292,7 @@
 			dm = get_waiting_dir_move(sctx, pm->ino);
 			ASSERT(dm);
 			dm->rmdir_ino = rmdir_ino;
+			dm->rmdir_gen = rmdir_gen;
 		}
 		goto out;
 	}
@@ -3265,7 +3311,7 @@
 		struct orphan_dir_info *odi;
 		u64 gen;
 
-		odi = get_orphan_dir_info(sctx, rmdir_ino);
+		odi = get_orphan_dir_info(sctx, rmdir_ino, rmdir_gen);
 		if (!odi) {
 			/* already deleted */
 			goto finish;
@@ -3780,6 +3826,72 @@
 }
 
 /*
+ * When processing the new references for an inode we may orphanize an existing
+ * directory inode because its old name conflicts with one of the new references
+ * of the current inode. Later, when processing another new reference of our
+ * inode, we might need to orphanize another inode, but the path we have in the
+ * reference reflects the pre-orphanization name of the directory we previously
+ * orphanized. For example:
+ *
+ * parent snapshot looks like:
+ *
+ * .                                     (ino 256)
+ * |----- f1                             (ino 257)
+ * |----- f2                             (ino 258)
+ * |----- d1/                            (ino 259)
+ *        |----- d2/                     (ino 260)
+ *
+ * send snapshot looks like:
+ *
+ * .                                     (ino 256)
+ * |----- d1                             (ino 258)
+ * |----- f2/                            (ino 259)
+ *        |----- f2_link/                (ino 260)
+ *        |       |----- f1              (ino 257)
+ *        |
+ *        |----- d2                      (ino 258)
+ *
+ * When processing inode 257 we compute the name for inode 259 as "d1", and we
+ * cache it in the name cache. Later when we start processing inode 258, when
+ * collecting all its new references we set a full path of "d1/d2" for its new
+ * reference with name "d2". When we start processing the new references we
+ * start by processing the new reference with name "d1", and this results in
+ * orphanizing inode 259, since its old reference causes a conflict. Then we
+ * move on the next new reference, with name "d2", and we find out we must
+ * orphanize inode 260, as its old reference conflicts with ours - but for the
+ * orphanization we use a source path corresponding to the path we stored in the
+ * new reference, which is "d1/d2" and not "o259-6-0/d2" - this makes the
+ * receiver fail since the path component "d1/" no longer exists, it was renamed
+ * to "o259-6-0/" when processing the previous new reference. So in this case we
+ * must recompute the path in the new reference and use it for the new
+ * orphanization operation.
+ */
+static int refresh_ref_path(struct send_ctx *sctx, struct recorded_ref *ref)
+{
+	char *name;
+	int ret;
+
+	name = kmemdup(ref->name, ref->name_len, GFP_KERNEL);
+	if (!name)
+		return -ENOMEM;
+
+	fs_path_reset(ref->full_path);
+	ret = get_cur_path(sctx, ref->dir, ref->dir_gen, ref->full_path);
+	if (ret < 0)
+		goto out;
+
+	ret = fs_path_add(ref->full_path, name, ref->name_len);
+	if (ret < 0)
+		goto out;
+
+	/* Update the reference's base name pointer. */
+	set_ref_path(ref, ref->full_path);
+out:
+	kfree(name);
+	return ret;
+}
+
+/*
  * This does all the move/link/unlink/rmdir magic.
  */
 static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
@@ -3847,52 +3959,56 @@
 			goto out;
 	}
 
+	/*
+	 * Before doing any rename and link operations, do a first pass on the
+	 * new references to orphanize any unprocessed inodes that may have a
+	 * reference that conflicts with one of the new references of the current
+	 * inode. This needs to happen first because a new reference may conflict
+	 * with the old reference of a parent directory, so we must make sure
+	 * that the path used for link and rename commands don't use an
+	 * orphanized name when an ancestor was not yet orphanized.
+	 *
+	 * Example:
+	 *
+	 * Parent snapshot:
+	 *
+	 * .                                                      (ino 256)
+	 * |----- testdir/                                        (ino 259)
+	 * |          |----- a                                    (ino 257)
+	 * |
+	 * |----- b                                               (ino 258)
+	 *
+	 * Send snapshot:
+	 *
+	 * .                                                      (ino 256)
+	 * |----- testdir_2/                                      (ino 259)
+	 * |          |----- a                                    (ino 260)
+	 * |
+	 * |----- testdir                                         (ino 257)
+	 * |----- b                                               (ino 257)
+	 * |----- b2                                              (ino 258)
+	 *
+	 * Processing the new reference for inode 257 with name "b" may happen
+	 * before processing the new reference with name "testdir". If so, we
+	 * must make sure that by the time we send a link command to create the
+	 * hard link "b", inode 259 was already orphanized, since the generated
+	 * path in "valid_path" already contains the orphanized name for 259.
+	 * We are processing inode 257, so only later when processing 259 we do
+	 * the rename operation to change its temporary (orphanized) name to
+	 * "testdir_2".
+	 */
 	list_for_each_entry(cur, &sctx->new_refs, list) {
-		/*
-		 * We may have refs where the parent directory does not exist
-		 * yet. This happens if the parent directories inum is higher
-		 * than the current inum. To handle this case, we create the
-		 * parent directory out of order. But we need to check if this
-		 * did already happen before due to other refs in the same dir.
-		 */
 		ret = get_cur_inode_state(sctx, cur->dir, cur->dir_gen);
 		if (ret < 0)
 			goto out;
-		if (ret == inode_state_will_create) {
-			ret = 0;
-			/*
-			 * First check if any of the current inodes refs did
-			 * already create the dir.
-			 */
-			list_for_each_entry(cur2, &sctx->new_refs, list) {
-				if (cur == cur2)
-					break;
-				if (cur2->dir == cur->dir) {
-					ret = 1;
-					break;
-				}
-			}
-
-			/*
-			 * If that did not happen, check if a previous inode
-			 * did already create the dir.
-			 */
-			if (!ret)
-				ret = did_create_dir(sctx, cur->dir);
-			if (ret < 0)
-				goto out;
-			if (!ret) {
-				ret = send_create_inode(sctx, cur->dir);
-				if (ret < 0)
-					goto out;
-			}
-		}
+		if (ret == inode_state_will_create)
+			continue;
 
 		/*
-		 * Check if this new ref would overwrite the first ref of
-		 * another unprocessed inode. If yes, orphanize the
-		 * overwritten inode. If we find an overwritten ref that is
-		 * not the first ref, simply unlink it.
+		 * Check if this new ref would overwrite the first ref of another
+		 * unprocessed inode. If yes, orphanize the overwritten inode.
+		 * If we find an overwritten ref that is not the first ref,
+		 * simply unlink it.
 		 */
 		ret = will_overwrite_ref(sctx, cur->dir, cur->dir_gen,
 				cur->name, cur->name_len,
@@ -3909,6 +4025,12 @@
 				struct name_cache_entry *nce;
 				struct waiting_dir_move *wdm;
 
+				if (orphanized_dir) {
+					ret = refresh_ref_path(sctx, cur);
+					if (ret < 0)
+						goto out;
+				}
+
 				ret = orphanize_inode(sctx, ow_inode, ow_gen,
 						cur->full_path);
 				if (ret < 0)
@@ -3965,12 +4087,66 @@
 				if (ret < 0)
 					goto out;
 			} else {
+				/*
+				 * If we previously orphanized a directory that
+				 * collided with a new reference that we already
+				 * processed, recompute the current path because
+				 * that directory may be part of the path.
+				 */
+				if (orphanized_dir) {
+					ret = refresh_ref_path(sctx, cur);
+					if (ret < 0)
+						goto out;
+				}
 				ret = send_unlink(sctx, cur->full_path);
 				if (ret < 0)
 					goto out;
 			}
 		}
 
+	}
+
+	list_for_each_entry(cur, &sctx->new_refs, list) {
+		/*
+		 * We may have refs where the parent directory does not exist
+		 * yet. This happens if the parent directories inum is higher
+		 * than the current inum. To handle this case, we create the
+		 * parent directory out of order. But we need to check if this
+		 * did already happen before due to other refs in the same dir.
+		 */
+		ret = get_cur_inode_state(sctx, cur->dir, cur->dir_gen);
+		if (ret < 0)
+			goto out;
+		if (ret == inode_state_will_create) {
+			ret = 0;
+			/*
+			 * First check if any of the current inodes refs did
+			 * already create the dir.
+			 */
+			list_for_each_entry(cur2, &sctx->new_refs, list) {
+				if (cur == cur2)
+					break;
+				if (cur2->dir == cur->dir) {
+					ret = 1;
+					break;
+				}
+			}
+
+			/*
+			 * If that did not happen, check if a previous inode
+			 * did already create the dir.
+			 */
+			if (!ret)
+				ret = did_create_dir(sctx, cur->dir);
+			if (ret < 0)
+				goto out;
+			if (!ret) {
+				ret = send_create_inode(sctx, cur->dir);
+				if (ret < 0)
+					goto out;
+			}
+		}
+
 		if (S_ISDIR(sctx->cur_inode_mode) && sctx->parent_root) {
 			ret = wait_for_dest_dir_move(sctx, cur, is_orphan);
 			if (ret < 0)
@@ -4513,6 +4689,10 @@
 	struct fs_path *p;
 	struct posix_acl_xattr_header dummy_acl;
 
+	/* Capabilities are emitted by finish_inode_if_needed */
+	if (!strncmp(name, XATTR_NAME_CAPS, name_len))
+		return 0;
+
 	p = fs_path_alloc();
 	if (!p)
 		return -ENOMEM;
@@ -5075,6 +5255,64 @@
 	return 0;
 }
 
+/*
+ * Search for a capability xattr related to sctx->cur_ino. If the capability is
+ * found, call send_set_xattr function to emit it.
+ *
+ * Return 0 if there isn't a capability, or when the capability was emitted
+ * successfully, or < 0 if an error occurred.
+ */
+static int send_capabilities(struct send_ctx *sctx)
+{
+	struct fs_path *fspath = NULL;
+	struct btrfs_path *path;
+	struct btrfs_dir_item *di;
+	struct extent_buffer *leaf;
+	unsigned long data_ptr;
+	char *buf = NULL;
+	int buf_len;
+	int ret = 0;
+
+	path = alloc_path_for_send();
+	if (!path)
+		return -ENOMEM;
+
+	di = btrfs_lookup_xattr(NULL, sctx->send_root, path, sctx->cur_ino,
+				XATTR_NAME_CAPS, strlen(XATTR_NAME_CAPS), 0);
+	if (!di) {
+		/* There is no xattr for this inode */
+		goto out;
+	} else if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	buf_len = btrfs_dir_data_len(leaf, di);
+
+	fspath = fs_path_alloc();
+	buf = kmalloc(buf_len, GFP_KERNEL);
+	if (!fspath || !buf) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, fspath);
+	if (ret < 0)
+		goto out;
+
+	data_ptr = (unsigned long)(di + 1) + btrfs_dir_name_len(leaf, di);
+	read_extent_buffer(leaf, buf, data_ptr, buf_len);
+
+	ret = send_set_xattr(sctx, fspath, XATTR_NAME_CAPS,
+			strlen(XATTR_NAME_CAPS), buf, buf_len);
+out:
+	kfree(buf);
+	fs_path_free(fspath);
+	btrfs_free_path(path);
+	return ret;
+}
+
 static int clone_range(struct send_ctx *sctx,
 		       struct clone_root *clone_root,
 		       const u64 disk_byte,
@@ -5292,6 +5530,21 @@
 			break;
 		offset += clone_len;
 		clone_root->offset += clone_len;
+
+		/*
+		 * If we are cloning from the file we are currently processing,
+		 * and using the send root as the clone root, we must stop once
+		 * the current clone offset reaches the current eof of the file
+		 * at the receiver, otherwise we would issue an invalid clone
+		 * operation (source range going beyond eof) and cause the
+		 * receiver to fail. So if we reach the current eof, bail out
+		 * and fallback to a regular write.
+		 */
+		if (clone_root->root == sctx->send_root &&
+		    clone_root->ino == sctx->cur_ino &&
+		    clone_root->offset >= sctx->cur_inode_next_write_offset)
+			break;
+
 		data_offset += clone_len;
 next:
 		path->slots[0]++;
@@ -5978,6 +6231,10 @@
 			goto out;
 	}
 
+	ret = send_capabilities(sctx);
+	if (ret < 0)
+		goto out;
+
 	/*
 	 * If other directory inodes depended on our current directory
 	 * inode's move/rename, now do their move/rename operations.
@@ -7053,12 +7310,6 @@
 	spin_unlock(&send_root->root_item_lock);
 
 	/*
-	 * This is done when we lookup the root, it should already be complete
-	 * by the time we get here.
-	 */
-	WARN_ON(send_root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE);
-
-	/*
 	 * Userspace tools do the checks and warn the user if it's
 	 * not RO.
 	 */
@@ -7140,7 +7391,7 @@
 
 	alloc_size = sizeof(struct clone_root) * (arg->clone_sources_count + 1);
 
-	sctx->clone_roots = kzalloc(alloc_size, GFP_KERNEL);
+	sctx->clone_roots = kvzalloc(alloc_size, GFP_KERNEL);
 	if (!sctx->clone_roots) {
 		ret = -ENOMEM;
 		goto out;