Merge tag 'for-5.3-rc4-tag' of git://git.kernel.org/pub/scm/linux/ker…

…nel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "Two fixes that popped up during testing:

   - fix for sysfs-related code that adds/removes block groups, warnings
     appear during several fstests in connection with sysfs updates in
     5.3, the fix essentially replaces a workaround with scope NOFS and
     applies to 5.2-based branch too

   - add sanity check of trim range"

* tag 'for-5.3-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: trim: Check the range passed into to prevent overflow
  Btrfs: fix sysfs warning and missing raid sysfs directories

fs/btrfs/ctree.h

@@ -401,7 +401,6 @@ struct btrfs_dev_replace {
 struct raid_kobject {
 	u64 flags;
 	struct kobject kobj;
-	struct list_head list;
 };
 
 /*
@@ -915,8 +914,6 @@ struct btrfs_fs_info {
 	u32 thread_pool_size;
 
 	struct kobject *space_info_kobj;
-	struct list_head pending_raid_kobjs;
-	spinlock_t pending_raid_kobjs_lock; /* uncontended */
 
 	u64 total_pinned;
 
@@ -2698,7 +2695,6 @@ int btrfs_can_relocate(struct btrfs_fs_info *fs_info, u64 bytenr);
 int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 			   u64 bytes_used, u64 type, u64 chunk_offset,
 			   u64 size);
-void btrfs_add_raid_kobjects(struct btrfs_fs_info *fs_info);
 struct btrfs_trans_handle *btrfs_start_trans_remove_block_group(
 				struct btrfs_fs_info *fs_info,
 				const u64 chunk_offset);

fs/btrfs/disk-io.c

@@ -2683,8 +2683,6 @@ int open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->delayed_iputs);
 	INIT_LIST_HEAD(&fs_info->delalloc_roots);
 	INIT_LIST_HEAD(&fs_info->caching_block_groups);
-	INIT_LIST_HEAD(&fs_info->pending_raid_kobjs);
-	spin_lock_init(&fs_info->pending_raid_kobjs_lock);
 	spin_lock_init(&fs_info->delalloc_root_lock);
 	spin_lock_init(&fs_info->trans_lock);
 	spin_lock_init(&fs_info->fs_roots_radix_lock);

fs/btrfs/extent-tree.c

@@ -4,6 +4,7 @@
  */
 
 #include <linux/sched.h>
+#include <linux/sched/mm.h>
 #include <linux/sched/signal.h>
 #include <linux/pagemap.h>
 #include <linux/writeback.h>
@@ -7888,33 +7889,6 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	return 0;
 }
 
-/* link_block_group will queue up kobjects to add when we're reclaim-safe */
-void btrfs_add_raid_kobjects(struct btrfs_fs_info *fs_info)
-{
-	struct btrfs_space_info *space_info;
-	struct raid_kobject *rkobj;
-	LIST_HEAD(list);
-	int ret = 0;
-
-	spin_lock(&fs_info->pending_raid_kobjs_lock);
-	list_splice_init(&fs_info->pending_raid_kobjs, &list);
-	spin_unlock(&fs_info->pending_raid_kobjs_lock);
-
-	list_for_each_entry(rkobj, &list, list) {
-		space_info = btrfs_find_space_info(fs_info, rkobj->flags);
-
-		ret = kobject_add(&rkobj->kobj, &space_info->kobj,
-				"%s", btrfs_bg_type_to_raid_name(rkobj->flags));
-		if (ret) {
-			kobject_put(&rkobj->kobj);
-			break;
-		}
-	}
-	if (ret)
-		btrfs_warn(fs_info,
-			   "failed to add kobject for block cache, ignoring");
-}
-
 static void link_block_group(struct btrfs_block_group_cache *cache)
 {
 	struct btrfs_space_info *space_info = cache->space_info;
@@ -7929,18 +7903,36 @@ static void link_block_group(struct btrfs_block_group_cache *cache)
 	up_write(&space_info->groups_sem);
 
 	if (first) {
-		struct raid_kobject *rkobj = kzalloc(sizeof(*rkobj), GFP_NOFS);
+		struct raid_kobject *rkobj;
+		unsigned int nofs_flag;
+		int ret;
+
+		/*
+		 * Setup a NOFS context because kobject_add(), deep in its call
+		 * chain, does GFP_KERNEL allocations, and we are often called
+		 * in a context where if reclaim is triggered we can deadlock
+		 * (we are either holding a transaction handle or some lock
+		 * required for a transaction commit).
+		 */
+		nofs_flag = memalloc_nofs_save();
+		rkobj = kzalloc(sizeof(*rkobj), GFP_KERNEL);
 		if (!rkobj) {
+			memalloc_nofs_restore(nofs_flag);
 			btrfs_warn(cache->fs_info,
 				"couldn't alloc memory for raid level kobject");
 			return;
 		}
 		rkobj->flags = cache->flags;
 		kobject_init(&rkobj->kobj, &btrfs_raid_ktype);
-
-		spin_lock(&fs_info->pending_raid_kobjs_lock);
-		list_add_tail(&rkobj->list, &fs_info->pending_raid_kobjs);
-		spin_unlock(&fs_info->pending_raid_kobjs_lock);
+		ret = kobject_add(&rkobj->kobj, &space_info->kobj, "%s",
+				  btrfs_bg_type_to_raid_name(rkobj->flags));
+		memalloc_nofs_restore(nofs_flag);
+		if (ret) {
+			kobject_put(&rkobj->kobj);
+			btrfs_warn(fs_info,
+			   "failed to add kobject for block cache, ignoring");
+			return;
+		}
 		space_info->block_group_kobjs[index] = &rkobj->kobj;
 	}
 }
@@ -8206,7 +8198,6 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
 			inc_block_group_ro(cache, 1);
 	}
 
-	btrfs_add_raid_kobjects(info);
 	btrfs_init_global_block_rsv(info);
 	ret = check_chunk_block_group_mappings(info);
 error:
@@ -8975,6 +8966,7 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	struct btrfs_device *device;
 	struct list_head *devices;
 	u64 group_trimmed;
+	u64 range_end = U64_MAX;
 	u64 start;
 	u64 end;
 	u64 trimmed = 0;
@@ -8984,16 +8976,23 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	int dev_ret = 0;
 	int ret = 0;
 
+	/*
+	 * Check range overflow if range->len is set.
+	 * The default range->len is U64_MAX.
+	 */
+	if (range->len != U64_MAX &&
+	    check_add_overflow(range->start, range->len, &range_end))
+		return -EINVAL;
+
 	cache = btrfs_lookup_first_block_group(fs_info, range->start);
 	for (; cache; cache = next_block_group(cache)) {
-		if (cache->key.objectid >= (range->start + range->len)) {
+		if (cache->key.objectid >= range_end) {
 			btrfs_put_block_group(cache);
 			break;
 		}
 
 		start = max(range->start, cache->key.objectid);
-		end = min(range->start + range->len,
-				cache->key.objectid + cache->key.offset);
+		end = min(range_end, cache->key.objectid + cache->key.offset);
 
 		if (end - start >= range->minlen) {
 			if (!block_group_cache_done(cache)) {

fs/btrfs/volumes.c

@@ -3087,16 +3087,6 @@ static int btrfs_relocate_chunk(struct btrfs_fs_info *fs_info, u64 chunk_offset)
 	if (ret)
 		return ret;
 
-	/*
-	 * We add the kobjects here (and after forcing data chunk creation)
-	 * since relocation is the only place we'll create chunks of a new
-	 * type at runtime.  The only place where we'll remove the last
-	 * chunk of a type is the call immediately below this one.  Even
-	 * so, we're protected against races with the cleaner thread since
-	 * we're covered by the delete_unused_bgs_mutex.
-	 */
-	btrfs_add_raid_kobjects(fs_info);
-
 	trans = btrfs_start_trans_remove_block_group(root->fs_info,
 						     chunk_offset);
 	if (IS_ERR(trans)) {
@@ -3223,9 +3213,6 @@ static int btrfs_may_alloc_data_chunk(struct btrfs_fs_info *fs_info,
 			btrfs_end_transaction(trans);
 			if (ret < 0)
 				return ret;
-
-			btrfs_add_raid_kobjects(fs_info);
-
 			return 1;
 		}
 	}