kernel/git/stable/linux-stable.git/fs/ext4/ext4_jbd2.c, branch master

Merge tag 'ext4_for_linus-7.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

2026-08-20T20:17:11+00:00

Pull ext4 updates from Ted Ts'o: - Improve performance by allowing parallel DIO writes when we were previously being overly conservative when checking whether it was safe to avoid requiring an exclusive lock - Improve the performance of ext4_mb_prefetch() used by fallocate() by avoiding work when it is not needed - Remove the unnecessary custom end_io function ext4_end_buffer_io_sync() - Improve performance when performing an overwrite to an already uptodate folio - Clean up how we handle deallocating EA inodes to avoid a potential lock ordering issue when there is a failed mount while an EA inode is still being evicted - Use str_plural() instead of a custom macro - Avoid soft lockups or RCU stalls if there are many busy buffers (caused by heavy I/O) while checkpointing - Use scoped NOFS when starting a handle in nojournal mode - Align fields in handle structure to optimize setting and getting the h_type and h_line_no fields - Fix documentation of the meta_bg block group layout - Bug fixes: - Fix a potential out-of-bounds read in ext4_read_inline_dir() - Fix a potential deadlock when concurrent xattr operations are racing with each other when some of the xattrs are using the ea_inode feature - Fix a spurious warning with data=journal that can be triggered when writeback races with remounting the file system read-only - Fix a potential deadlock when EXT4_IOC_MIGRATE races with a file system freeze operation - Make sure all in-flight direct I/O operations are complete before falling back to buffered I/O - Handle IOCB_NOWAIT properly when performing a extending DAX write - Prevent potentially sleeping on a block allocation when IOCB_NOWAIT is set - Fix potential races when racing an inline data write with a page fault - Propagate errors when adding or removing extent ranges during a fast commit replay - Avoid trying to expand an inode's extra size when it is being evicted to avoid a number of corner case or deadlocks - Avoid spurious error when retrying inode extra size expansion - Fix corner cases where we underestimate the number of journal credits needed - Avoid hangs/crashes/WARNINGS caused by maliciously corrupted file systems - Don't issue spurious orphan clean message on RO file systems - Avoid leaving the file system in an inconsistent state after a crash when a WRITE_ZEROS in progress converting an unwritten extent to a written extent - Handle WRITE_ZEROS correctly when there are some partially dirtied regions in the page cache - Pass errors during zero-rage, truncate, or punch hole to the caller if ext4_get_block() fails - Wait for writeback to finish when triggered by zero-range or zero-range for those devices that require stable writes - If the reserved gid superblock field is set, set the reserved gid instead of the reserved uid * tag 'ext4_for_linus-7.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (56 commits) ext4: fix estimate extent index blocks in ext4_ext_index_trans_blocks() ext4: fix transaction overflow during writeback ext4: teach ext4_meta_trans_blocks() about number of allocated extents ext4: guard against NULL s_group_info in ext4_get_group_info ext4: fix spurious message about orphan cleanup on RO fs ext4: stop retrying saturated xattr cache entries ext4: don't enable DAX on new encrypted files ext4: protect WRITE_ZEROES written extents with orphan list ext4: export converted block count from ext4_convert_unwritten_extents() ext4: fix incorrect function call when initializing s_resgid ext4: validate EA inode i_nlink in ext4_xattr_inode_iget jbd2: align h_type and h_line_no in the handle structure on byte boundaries ext4: enable scoped NOFS when starting a handle in nojournal mode ext4: write back partial-zeroed edges in WRITE_ZEROES ext4: zero out whole block for clean edges in WRITE_ZEROES ext4: track partial-zero outcome per edge in ext4_zero_partial_blocks() ext4: clarify return semantics of ext4_load_tail_bh() ext4: move partial block zeroing earlier in ext4_zero_range() ext4: check return value of ext4_get_block() in ext4_load_tail_bh() ext4: skip tail block zeroing for inline data files ...

ext4: Allocate mapping_metadata_bhs struct on demand

2026-07-27T14:25:31+00:00

Currently every ext4 inode gets mapping_metadata_bhs struct although it is only needed when running without a journal and only for inodes where any metadata was dirtied. Allocate mapping_metadata_bhs struct on demand when dirtying the first metadata buffer for the inode. Acked-by: Theodore Ts'o Signed-off-by: Jan Kara Link: https://patch.msgid.link/20260727104923.3828017-24-jack@suse.cz Signed-off-by: Christian Brauner (Amutable)

ext4: enable scoped NOFS when starting a handle in nojournal mode

2026-07-23T03:47:06+00:00

The jbd2 layer enables NOFS mode using memalloc_nofs_{save,restore}() while a handle is active. We need to do the same in nojournal mode so that it is safe to remove GFP_NOFS flags while a jbd2 handle is active. This will require that we actually allocate a real handle, but with an h_invalid flag set, so there is a place to put the saved memalloc context. Signed-off-by: Theodore Ts'o Reviewed-by: Andreas Dilger

ext4: Track metadata bhs in fs-private inode part

2026-03-26T14:03:32+00:00

Track metadata bhs for an inode in fs-private part of the inode. We need the tracking only for nojournal mode so this is somewhat wasteful. We can relatively easily make the mapping_metadata_bhs struct dynamically allocated similarly to how we treat jbd2_inode but let's leave that for ext4 specific series once the dust settles a bit. Signed-off-by: Jan Kara Link: https://patch.msgid.link/20260326095354.16340-82-jack@suse.cz Signed-off-by: Christian Brauner

ext4: make data=journal support large block size

2025-11-29T03:35:28+00:00

Currently, ext4_set_inode_mapping_order() does not set max folio order for files with the data journalling flag. For files that already have large folios enabled, ext4_inode_journal_mode() ignores the data journalling flag once max folio order is set. This is not because data journalling cannot work with large folios, but because credit estimates will go through the roof if there are too many blocks per folio. Since the real constraint is blocks-per-folio, to support data=journal under LBS, we now set max folio order to be equal to min folio order for files with the journalling flag. When LBS is disabled, the max folio order remains unset as before. Therefore, before ext4_change_inode_journal_flag() switches the journalling mode, we call truncate_pagecache() to drop all page cache for that inode, and filemap_write_and_wait() is called unconditionally. After that, once the journalling mode has been switched, we can safely reset the inode mapping order, and the mapping_large_folio_support() check in ext4_inode_journal_mode() can be removed. Suggested-by: Jan Kara Suggested-by: Dan Carpenter Signed-off-by: Baokun Li Reviewed-by: Zhang Yi Reviewed-by: Jan Kara Reviewed-by: Ojaswin Mujoo Message-ID: <20251121090654.631996-22-libaokun@huaweicloud.com> Signed-off-by: Theodore Ts'o

ext4: wait for ongoing I/O to complete before freeing blocks

2025-10-10T17:34:35+00:00

When freeing metadata blocks in nojournal mode, ext4_forget() calls bforget() to clear the dirty flag on the buffer_head and remvoe associated mappings. This is acceptable if the metadata has not yet begun to be written back. However, if the write-back has already started but is not yet completed, ext4_forget() will have no effect. Subsequently, ext4_mb_clear_bb() will immediately return the block to the mb allocator. This block can then be reallocated immediately, potentially causing an data corruption issue. Fix this by clearing the buffer's dirty flag and waiting for the ongoing I/O to complete, ensuring that no further writes to stale data will occur. Fixes: 16e08b14a455 ("ext4: cleanup clean_bdev_aliases() calls") Cc: stable@kernel.org Reported-by: Gao Xiang Closes: https://lore.kernel.org/linux-ext4/a9417096-9549-4441-9878-b1955b899b4e@huaweicloud.com/ Signed-off-by: Zhang Yi Reviewed-by: Jan Kara Message-ID: <20250916093337.3161016-3-yi.zhang@huaweicloud.com> Signed-off-by: Theodore Ts'o

ext4: enable large folio for regular file

2025-05-20T14:31:12+00:00

Besides fsverity, fscrypt, and the data=journal mode, ext4 now supports large folios for regular files. Enable this feature by default. However, since we cannot change the folio order limitation of mappings on active inodes, setting the journal=data mode via ioctl on an active inode will not take immediate effect in non-delalloc mode. Signed-off-by: Zhang Yi Link: https://patch.msgid.link/20250512063319.3539411-9-yi.zhang@huaweicloud.com Signed-off-by: Theodore Ts'o

ext4: remove redundant function ext4_has_metadata_csum

2025-03-17T15:19:41+00:00

Since commit f2b4fa19647e ("ext4: switch to using the crc32c library"), ext4_has_metadata_csum() is just an alias for ext4_has_feature_metadata_csum(). ext4_has_feature_metadata_csum() is generated by EXT4_FEATURE_RO_COMPAT_FUNCS and uses the regular naming convention for checking a single ext4 feature. Therefore, remove ext4_has_metadata_csum() and update all its callers to use ext4_has_feature_metadata_csum() directly. Signed-off-by: Eric Biggers Reviewed-by: Darrick J. Wong Link: https://patch.msgid.link/20250207031335.42637-1-ebiggers@kernel.org Signed-off-by: Theodore Ts'o

ext4: add ext4_emergency_state() helper function

2025-03-13T14:16:34+00:00

Since both SHUTDOWN and EMERGENCY_RO are emergency states of the ext4 file system, and they are checked in similar locations, we have added a helper function, ext4_emergency_state(), to determine whether the current file system is in one of these two emergency states. Then, replace calls to ext4_forced_shutdown() with ext4_emergency_state() in those functions that could potentially trigger write operations. Signed-off-by: Baokun Li Reviewed-by: Jan Kara Reviewed-by: Zhang Yi Link: https://patch.msgid.link/20250122114130.229709-4-libaokun@huaweicloud.com Signed-off-by: Theodore Ts'o

use ->bd_mapping instead of ->bd_inode->i_mapping

2024-05-03T06:36:51+00:00

Just the low-hanging fruit... Signed-off-by: Al Viro Link: https://lore.kernel.org/r/20240411145346.2516848-2-viro@zeniv.linux.org.uk Signed-off-by: Christian Brauner