From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Mon, 13 Nov 2023 17:09:12 +0000 (-0800)
Subject: Merge tag 'for-6.7-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave... 
X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=9bacdd8996c77c42ca004440be610692275ff9d0;p=users%2Fjedix%2Flinux-maple.git

Merge tag 'for-6.7-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:

 - fix potential overflow in returned value from SEARCH_TREE_V2
   ioctl on 32bit architecture

 - zoned mode fixes:

     - drop unnecessary write pointer check for RAID0/RAID1/RAID10
       profiles, now it works because of raid-stripe-tree

     - wait for finishing the zone when direct IO needs a new
       allocation

 - simple quota fixes:

     - pass correct owning root pointer when cleaning up an
       aborted transaction

     - fix leaking some structures when processing delayed refs

     - change key type number of BTRFS_EXTENT_OWNER_REF_KEY,
       reorder it before inline refs that are supposed to be
       sorted, keeping the original number would complicate a lot
       of things; this change needs an updated version of
       btrfs-progs to work and filesystems need to be recreated

 - fix error pointer dereference after failure to allocate fs
   devices

 - fix race between accounting qgroup extents and removing a
   qgroup

* tag 'for-6.7-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: make OWNER_REF_KEY type value smallest among inline refs
  btrfs: fix qgroup record leaks when using simple quotas
  btrfs: fix race between accounting qgroup extents and removing a qgroup
  btrfs: fix error pointer dereference after failure to allocate fs devices
  btrfs: make found_logical_ret parameter mandatory for function queue_scrub_stripe()
  btrfs: get correct owning_root when dropping snapshot
  btrfs: zoned: wait for data BG to be finished on direct IO allocation
  btrfs: zoned: drop no longer valid write pointer check
  btrfs: directly return 0 on no error code in btrfs_insert_raid_extent()
  btrfs: use u64 for buffer sizes in the tree search ioctls
---

9bacdd8996c77c42ca004440be610692275ff9d0
diff --cc fs/btrfs/ctree.c
index 2a9344a3fcee,14cefeaf9622..35c1d24d4a78
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@@ -429,10 -418,10 +429,10 @@@ static noinline int update_ref_for_cow(
  	 * are only allowed for blocks use full backrefs.
  	 */
  
 -	if (btrfs_block_can_be_shared(root, buf)) {
 +	if (btrfs_block_can_be_shared(trans, root, buf)) {
  		ret = btrfs_lookup_extent_info(trans, fs_info, buf->start,
  					       btrfs_header_level(buf), 1,
- 					       &refs, &flags);
+ 					       &refs, &flags, NULL);
  		if (ret)
  			return ret;
  		if (unlikely(refs == 0)) {