linux/fs, branch v6.15

Merge tag 'v6.15-rc8-ksmbd-server-fixes' of git://git.samba.org/ksmbd

2025-05-23T15:42:29Z

Pull smb server fixes from Steve French: - Fix for rename regression due to the recent VFS lookup changes - Fix write failure - locking fix for oplock handling * tag 'v6.15-rc8-ksmbd-server-fixes' of git://git.samba.org/ksmbd: ksmbd: use list_first_entry_or_null for opinfo_get_list() ksmbd: fix rename failure ksmbd: fix stream write failure

Merge tag 'vfs-6.15-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

2025-05-23T14:51:05Z

Pull vfs fixes from Christian Brauner: "This contains a small set of fixes for the blocking buffer lookup conversion done earlier this cycle. It adds a missing conversion in the getblk slowpath and a few minor optimizations and cleanups" * tag 'vfs-6.15-rc8.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: fs/buffer: optimize discard_buffer() fs/buffer: remove superfluous statements fs/buffer: avoid redundant lookup in getblk slowpath fs/buffer: use sleeping lookup in __getblk_slowpath()

Merge tag 'bcachefs-2025-05-22' of git://evilpiepirate.org/bcachefs

2025-05-23T02:17:55Z

Pull bcachefs fixes from Kent Overstreet: "Small stuff, main ones users will be interested in: - Couple more casefolding fixes; we can now detect and repair casefolded dirents in non-casefolded dir and vice versa - Fix for massive write inflation with mmapped io, which hit certain databases" * tag 'bcachefs-2025-05-22' of git://evilpiepirate.org/bcachefs: bcachefs: Check for casefolded dirents in non casefolded dirs bcachefs: Fix bch2_dirent_create_snapshot() for casefolding bcachefs: Fix casefold opt via xattr interface bcachefs: mkwrite() now only dirties one page bcachefs: fix extent_has_stripe_ptr() bcachefs: Fix bch2_btree_path_traverse_cached() when paths realloced

Merge tag '6.15-rc8-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

2025-05-22T19:35:16Z

Pull smb client fixes from Steve French: - Two fixes for use after free in readdir code paths * tag '6.15-rc8-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6: smb: client: Reset all search buffer pointers when releasing buffer smb: client: Fix use-after-free in cifs_fill_dirent

ksmbd: use list_first_entry_or_null for opinfo_get_list()

2025-05-22T03:30:39Z

The list_first_entry() macro never returns NULL. If the list is empty then it returns an invalid pointer. Use list_first_entry_or_null() to check if the list is empty. Reported-by: kernel test robot Reported-by: Dan Carpenter Closes: https://lore.kernel.org/r/202505080231.7OXwq4Te-lkp@intel.com/ Signed-off-by: Namjae Jeon Signed-off-by: Steve French

ksmbd: fix rename failure

2025-05-22T03:30:30Z

I found that rename fails after cifs mount due to update of lookup_one_qstr_excl(). mv a/c b/ mv: cannot move 'a/c' to 'b/c': No such file or directory In order to rename to a new name regardless of whether the dentry is negative, we need to get the dentry through lookup_one_qstr_excl(). So It will not return error if the name doesn't exist. Fixes: 204a575e91f3 ("VFS: add common error checks to lookup_one_qstr_excl()") Signed-off-by: Namjae Jeon Signed-off-by: Steve French

bcachefs: Check for casefolded dirents in non casefolded dirs

2025-05-22T00:13:14Z

Check for mismatches between casefold dirents and casefold directories. A mismatch will cause lookups to fail, as we'll be doing the lookup with the casefolded name, which won't match the non-casefolded dirent, and vice versa. Reported-by: Christopher Snowhill Signed-off-by: Kent Overstreet

bcachefs: Fix bch2_dirent_create_snapshot() for casefolding

2025-05-22T00:13:13Z

bch2_dirent_create_snapshot(), used in fsck, neglected to create a casefolded dirent. Just move this into dirent_create_key(). Signed-off-by: Kent Overstreet

bcachefs: Fix casefold opt via xattr interface

2025-05-22T00:13:09Z

Changing the casefold option requires extra checks/work - factor out a helper from bch2_fileattr_set() for the xattr code to use. Signed-off-by: Kent Overstreet

fs/buffer: optimize discard_buffer()

2025-05-21T07:34:29Z

While invalidating, the clearing of the bits in discard_buffer() is done in one fully ordered CAS operation. In the past this was done via individual clear_bit(), until e7470ee89f0 (fs: buffer: do not use unnecessary atomic operations when discarding buffers). This implies that there were never strong ordering requirements outside of being serialized by the buffer lock. As such relax the ordering for archs that can benefit. Further, the implied ordering in buffer_unlock() makes current cmpxchg implied barrier redundant due to release semantics. And while in theory the unlock could be part of the bulk clearing, it is best to leave it explicit, but without the double barriers. Signed-off-by: Davidlohr Bueso Link: https://lore.kernel.org/20250515173925.147823-5-dave@stgolabs.net Reviewed-by: Jan Kara Signed-off-by: Christian Brauner