From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Sat, 18 Apr 2026 00:08:31 +0000 (-0700)
Subject: Merge tag 'ext4_for_linux-7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git... 
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=a436a0b847c0fef9ead14f99bc03d8adbf66f15b;p=thirdparty%2Fkernel%2Flinux.git

Merge tag 'ext4_for_linux-7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:

 - Refactor code paths involved with partial block zero-out in
   prearation for converting ext4 to use iomap for buffered writes

 - Remove use of d_alloc() from ext4 in preparation for the deprecation
   of this interface

 - Replace some J_ASSERTS with a journal abort so we can avoid a kernel
   panic for a localized file system error

 - Simplify various code paths in mballoc, move_extent, and fast commit

 - Fix rare deadlock in jbd2_journal_cancel_revoke() that can be
   triggered by generic/013 when blocksize < pagesize

 - Fix memory leak when releasing an extended attribute when its value
   is stored in an ea_inode

 - Fix various potential kunit test bugs in fs/ext4/extents.c

 - Fix potential out-of-bounds access in check_xattr() with a corrupted
   file system

 - Make the jbd2_inode dirty range tracking safe for lockless reads

 - Avoid a WARN_ON when writeback files due to a corrupted file system;
   we already print an ext4 warning indicatign that data will be lost,
   so the WARN_ON is not necessary and doesn't add any new information

* tag 'ext4_for_linux-7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (37 commits)
  jbd2: fix deadlock in jbd2_journal_cancel_revoke()
  ext4: fix missing brelse() in ext4_xattr_inode_dec_ref_all()
  ext4: fix possible null-ptr-deref in mbt_kunit_exit()
  ext4: fix possible null-ptr-deref in extents_kunit_exit()
  ext4: fix the error handling process in extents_kunit_init).
  ext4: call deactivate_super() in extents_kunit_exit()
  ext4: fix miss unlock 'sb->s_umount' in extents_kunit_init()
  ext4: fix bounds check in check_xattrs() to prevent out-of-bounds access
  ext4: zero post-EOF partial block before appending write
  ext4: move pagecache_isize_extended() out of active handle
  ext4: remove ctime/mtime update from ext4_alloc_file_blocks()
  ext4: unify SYNC mode checks in fallocate paths
  ext4: ensure zeroed partial blocks are persisted in SYNC mode
  ext4: move zero partial block range functions out of active handle
  ext4: pass allocate range as loff_t to ext4_alloc_file_blocks()
  ext4: remove handle parameters from zero partial block functions
  ext4: move ordered data handling out of ext4_block_do_zero_range()
  ext4: rename ext4_block_zero_page_range() to ext4_block_zero_range()
  ext4: factor out journalled block zeroing range
  ext4: rename and extend ext4_block_truncate_page()
  ...
---

a436a0b847c0fef9ead14f99bc03d8adbf66f15b
diff --cc fs/jbd2/transaction.c
index a90f9092706cc,495f001298445..4885903bbd10c
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@@ -2651,18 -2702,24 +2702,24 @@@ static int jbd2_journal_file_inode(hand
  		return -EROFS;
  	journal = transaction->t_journal;
  
 -	jbd2_debug(4, "Adding inode %lu, tid:%d\n", jinode->i_vfs_inode->i_ino,
 +	jbd2_debug(4, "Adding inode %llu, tid:%d\n", jinode->i_vfs_inode->i_ino,
  			transaction->t_tid);
  
+ 	start_page = (pgoff_t)(start_byte >> PAGE_SHIFT);
+ 	end_page = (pgoff_t)(end_byte >> PAGE_SHIFT) + 1;
+ 
  	spin_lock(&journal->j_list_lock);
- 	jinode->i_flags |= flags;
+ 	WRITE_ONCE(jinode->i_flags, jinode->i_flags | flags);
  
- 	if (jinode->i_dirty_end) {
- 		jinode->i_dirty_start = min(jinode->i_dirty_start, start_byte);
- 		jinode->i_dirty_end = max(jinode->i_dirty_end, end_byte);
+ 	if (jinode->i_dirty_start_page != jinode->i_dirty_end_page) {
+ 		WRITE_ONCE(jinode->i_dirty_start_page,
+ 			   min(jinode->i_dirty_start_page, start_page));
+ 		WRITE_ONCE(jinode->i_dirty_end_page,
+ 			   max(jinode->i_dirty_end_page, end_page));
  	} else {
- 		jinode->i_dirty_start = start_byte;
- 		jinode->i_dirty_end = end_byte;
+ 		/* Publish a new non-empty range by making end visible first. */
+ 		WRITE_ONCE(jinode->i_dirty_end_page, end_page);
+ 		WRITE_ONCE(jinode->i_dirty_start_page, start_page);
  	}
  
  	/* Is inode already attached where we need it? */