kernel.git/fs/ufs, branch linux-6.15.y

Change inode_operations.mkdir to return struct dentry *

2025-02-27T19:00:17Z

Some filesystems, such as NFS, cifs, ceph, and fuse, do not have complete control of sequencing on the actual filesystem (e.g. on a different server) and may find that the inode created for a mkdir request already exists in the icache and dcache by the time the mkdir request returns. For example, if the filesystem is mounted twice the directory could be visible on the other mount before it is on the original mount, and a pair of name_to_handle_at(), open_by_handle_at() calls could instantiate the directory inode with an IS_ROOT() dentry before the first mkdir returns. This means that the dentry passed to ->mkdir() may not be the one that is associated with the inode after the ->mkdir() completes. Some callers need to interact with the inode after the ->mkdir completes and they currently need to perform a lookup in the (rare) case that the dentry is no longer hashed. This lookup-after-mkdir requires that the directory remains locked to avoid races. Planned future patches to lock the dentry rather than the directory will mean that this lookup cannot be performed atomically with the mkdir. To remove this barrier, this patch changes ->mkdir to return the resulting dentry if it is different from the one passed in. Possible returns are: NULL - the directory was created and no other dentry was used ERR_PTR() - an error occurred non-NULL - this other dentry was spliced in This patch only changes file-systems to return "ERR_PTR(err)" instead of "err" or equivalent transformations. Subsequent patches will make further changes to some file-systems to return a correct dentry. Not all filesystems reliably result in a positive hashed dentry: - NFS, cifs, hostfs will sometimes need to perform a lookup of the name to get inode information. Races could result in this returning something different. Note that this lookup is non-atomic which is what we are trying to avoid. Placing the lookup in filesystem code means it only happens when the filesystem has no other option. - kernfs and tracefs leave the dentry negative and the ->revalidate operation ensures that lookup will be called to correctly populate the dentry. This could be fixed but I don't think it is important to any of the users of vfs_mkdir() which look at the dentry. The recommendation to use d_drop();d_splice_alias() is ugly but fits with current practice. A planned future patch will change this. Reviewed-by: Jeff Layton Reviewed-by: Jan Kara Signed-off-by: NeilBrown Link: https://lore.kernel.org/r/20250227013949.536172-2-neilb@suse.de Signed-off-by: Christian Brauner

ufs: ufs_sb_private_info: remove unused s_{2,3}apb fields

2024-11-13T00:02:12Z

These two fields are populated and stored as a "frequently used value" in ufs_fill_super, but are not used afterwards in the driver. Moreover, one of the shifts triggers UBSAN: shift-out-of-bounds when apbshift is 12 because 12 * 3 = 36 and 1 << 36 does not fit in the 32 bit integer used to store the value. Closes: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2087853 Signed-off-by: Agathe Porte Signed-off-by: Al Viro

ufs: Convert ufs_change_blocknr() to take a folio

2024-10-18T21:35:31Z

Now that ufs_new_fragments() has a folio, pass it to ufs_change_blocknr() as a folio instead of converting it from folio to page to folio. This removes the last use of struct page in UFS. Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Al Viro

ufs: Pass a folio to ufs_new_fragments()

2024-10-18T21:35:31Z

All callers now have a folio, pass it to ufs_new_fragments() instead of converting back to a page. Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Al Viro

ufs: Convert ufs_inode_getfrag() to take a folio

2024-10-18T21:35:31Z

Pass bh->b_folio instead of bh->b_page. They're in a union, so no code change expected. Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Al Viro

ufs: Convert ufs_extend_tail() to take a folio

2024-10-18T21:35:31Z

Pass bh->b_folio instead of bh->b_page. They're in a union, so no code change expected. Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Al Viro

ufs: Convert ufs_inode_getblock() to take a folio

2024-10-18T21:35:31Z

Pass bh->b_folio instead of bh->b_page. They're in a union, so no code change expected. Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Al Viro

ufs: take the handling of free block counters into a helper

2024-10-18T21:35:31Z

There are 3 places where those counters (many and varied...) are adjusted - when we are freeing fragments and get an entire block freed, when we are freeing blocks and (in opposite direction) when we are grabbing a block. The logics is identical (modulo the sign of adjustment) in all three; better take it into a helper - less duplication and less clutter in the callers that way. Signed-off-by: Al Viro

clean ufs_trunc_direct() up a bit...

2024-10-18T21:35:31Z

For short files (== no indirect blocks needed) UFS allows the last block to be a partial one. That creates some complications for truncation down to "short file" lengths. ufs_trunc_direct() is called when we'd already made sure that new EOF is not in a hole; nothing needs to be done if we are extending the file and in case we are shrinking the file it needs to * shrink or free the old final block. * free all full direct blocks between the new and old EOF. * possibly shrink the new final block. The logics is needlessly complicated by trying to keep all cases handled by the same sequence of operations. if not shrinking nothing to do else if number of full blocks unchanged free the tail of possibly partial last block else free the tail of (currently full) new last block free all present (full) blocks in between free the (possibly partial) old last block is easier to follow than the result of trying to unify these cases. Signed-off-by: Al Viro

ufs: get rid of ubh_{ubhcpymem,memcpyubh}()

2024-10-18T21:35:31Z

used only in ufs_read_cylinder_structures()/ufs_put_super_internal() and there we can just as well avoid bothering with ufs_buffer_head and just deal with it fragment-by-fragment. Signed-off-by: Al Viro