linux-stable.git/fs/udf, branch v6.9 Linux kernel stable tree udf: remove SLAB_MEM_SPREAD flag usage 2024-03-05T14:16:15+00:00 Chengming Zhou zhouchengming@bytedance.com 2024-02-24T13:52:29+00:00 a78e41a67bef099ca3ffee78c7eda8d43b693f27 The SLAB_MEM_SPREAD flag is already a no-op after removal of SLAB allocator and in [1] it was fully deprecated. Remove its usage so we can delete it from slab. No functional change. Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com> Signed-off-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/all/20240223-slab-cleanup-flags-v2-1-02f1753e8303@suse.cz/ Message-Id: <20240224135229.830356-1-chengming.zhou@linux.dev> 
The SLAB_MEM_SPREAD flag is already a no-op after removal of SLAB
allocator and in [1] it was fully deprecated. Remove its usage so we can
delete it from slab. No functional change.

Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/all/20240223-slab-cleanup-flags-v2-1-02f1753e8303@suse.cz/
Message-Id: <20240224135229.830356-1-chengming.zhou@linux.dev>
udf: convert to new mount API 2024-02-21T12:55:30+00:00 Eric Sandeen sandeen@redhat.com 2024-02-20T21:44:52+00:00 c4e89cc674ac36743f9c28db8c1b565039e9e08b Convert the UDF filesystem to the new mount API. UDF is slightly unique in that it always preserves prior mount options across a remount, so that's handled by udf_init_options(). Signed-off-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Jan Kara <jack@suse.cz> Message-Id: <fcb3faf3-e2df-450c-b37a-11000c274585@redhat.com> 
Convert the UDF filesystem to the new mount API.

UDF is slightly unique in that it always preserves prior mount
options across a remount, so that's handled by udf_init_options().

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Message-Id: <fcb3faf3-e2df-450c-b37a-11000c274585@redhat.com>
udf: convert novrs to an option flag 2024-02-21T12:55:30+00:00 Eric Sandeen sandeen@redhat.com 2024-02-20T21:43:21+00:00 7a8e72c16e734f2ccdfe8d86bcd2e176aa5e978a There's no reason to treat novers specially, convert it to a flag in uopt->flags like other flag options. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Jan Kara <jack@suse.cz> Message-Id: <b5d53492-b99a-4b3c-82c0-581df9a9e384@redhat.com> 
There's no reason to treat novers specially, convert it
to a flag in uopt->flags like other flag options.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Message-Id: <b5d53492-b99a-4b3c-82c0-581df9a9e384@redhat.com>
udf: Avoid invalid LVID used on mount 2024-02-05T15:51:59+00:00 Jan Kara jack@suse.cz 2024-02-05T15:24:22+00:00 c8f1140cb82dc843e72697dc6f8c7ee1acce5d28 udf_load_logicalvolint() loads logical volume integrity descriptors. Since there can be multiple blocks with LVIDs, we verify the contents of only the last (prevailing) LVID found. However if we fail to load the last LVID (either due to IO error or because it's checksum fails to match), we never perform the verification of validity of the LVID we are going to use. If such LVID contains invalid data, we can hit out-of-bounds access or similar issues. Fix the problem by verifying each LVID we are potentially going to accept. Reported-by: Robert Morris <rtm@csail.mit.edu> Signed-off-by: Jan Kara <jack@suse.cz> 
udf_load_logicalvolint() loads logical volume integrity descriptors.
Since there can be multiple blocks with LVIDs, we verify the contents of
only the last (prevailing) LVID found. However if we fail to load the
last LVID (either due to IO error or because it's checksum fails to
match), we never perform the verification of validity of the LVID we are
going to use. If such LVID contains invalid data, we can hit
out-of-bounds access or similar issues. Fix the problem by verifying
each LVID we are potentially going to accept.

Reported-by: Robert Morris <rtm@csail.mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
udf: Remove GFP_NOFS allocation in udf_expand_file_adinicb() 2024-01-23T18:21:10+00:00 Jan Kara jack@suse.cz 2024-01-09T10:15:43+00:00 38f8af2a7191e5da21c557210d810c6d0d34f6c4 udf_expand_file_adinicb() is called under inode->i_rwsem and mapping->invalidate_lock. i_rwsem is safe wrt fs reclaim, invalidate_lock on this inode is safe as well (we hold inode reference so reclaim will not touch it, furthermore even lockdep should not complain as invalidate_lock is acquired from udf_evict_inode() only when truncating inode which should not happen from fs reclaim). Signed-off-by: Jan Kara <jack@suse.cz> 
udf_expand_file_adinicb() is called under inode->i_rwsem and
mapping->invalidate_lock. i_rwsem is safe wrt fs reclaim,
invalidate_lock on this inode is safe as well (we hold inode reference
so reclaim will not touch it, furthermore even lockdep should not
complain as invalidate_lock is acquired from udf_evict_inode() only when
truncating inode which should not happen from fs reclaim).

Signed-off-by: Jan Kara <jack@suse.cz>
udf: Avoid GFP_NOFS allocation in udf_load_pvoldesc() 2024-01-23T18:21:10+00:00 Jan Kara jack@suse.cz 2024-01-09T10:10:14+00:00 b27ffdc17c2b4cb93eee37aa7a11487f03a2a8ac udf_load_pvoldesc() is called only during mount when it is safe to enter fs reclaim (we hold only s_umount semaphore). Change GFP_NOFS to GFP_KERNEL allocation. Signed-off-by: Jan Kara <jack@suse.cz> 
udf_load_pvoldesc() is called only during mount when it is safe to
enter fs reclaim (we hold only s_umount semaphore). Change GFP_NOFS to
GFP_KERNEL allocation.

Signed-off-by: Jan Kara <jack@suse.cz>
udf: Avoid GFP_NOFS allocation in udf_symlink() 2024-01-23T18:21:10+00:00 Jan Kara jack@suse.cz 2024-01-09T10:06:57+00:00 2ed0d3d4fee199869e9aef09bc86ceed9d79b978 The GFP_NOFS allocation in udf_symlink() is called only under inode->i_rwsem and UDF_I(inode)->i_data_sem. The first is safe wrt reclaim, the second should be as well but allocating unde this lock is actually unnecessary. Move the allocation from under i_data_sem and change it to GFP_KERNEL. Signed-off-by: Jan Kara <jack@suse.cz> 
The GFP_NOFS allocation in udf_symlink() is called only under
inode->i_rwsem and UDF_I(inode)->i_data_sem. The first is safe wrt
reclaim, the second should be as well but allocating unde this lock is
actually unnecessary. Move the allocation from under i_data_sem and
change it to GFP_KERNEL.

Signed-off-by: Jan Kara <jack@suse.cz>
udf: Remove GFP_NOFS from dir iteration code 2024-01-23T18:21:10+00:00 Jan Kara jack@suse.cz 2024-01-09T10:00:46+00:00 f6766303c0f2ffbcc640ef2701984a81a274aff2 Directory iteration code was using GFP_NOFS allocations in two places. However the code is called only under inode->i_rwsem which is generally safe wrt reclaim. So we can do the allocations with GFP_KERNEL instead. Signed-off-by: Jan Kara <jack@suse.cz> 
Directory iteration code was using GFP_NOFS allocations in two places.
However the code is called only under inode->i_rwsem which is generally
safe wrt reclaim. So we can do the allocations with GFP_KERNEL instead.

Signed-off-by: Jan Kara <jack@suse.cz>
Merge tag 'pull-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2024-01-12T04:23:50+00:00 Linus Torvalds torvalds@linux-foundation.org 2024-01-12T04:23:50+00:00 488926926a1653adfda3f662355907c896524487 Pull misc filesystem updates from Al Viro: "Misc cleanups (the part that hadn't been picked by individual fs trees)" * tag 'pull-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: apparmorfs: don't duplicate kfree_link() orangefs: saner arguments passing in readdir guts ocfs2_find_match(): there's no such thing as NULL or negative ->d_parent reiserfs_add_entry(): get rid of pointless namelen checks __ocfs2_add_entry(), ocfs2_prepare_dir_for_insert(): namelen checks ext4_add_entry(): ->d_name.len is never 0 befs: d_obtain_alias(ERR_PTR(...)) will do the right thing affs: d_obtain_alias(ERR_PTR(...)) will do the right thing /proc/sys: use d_splice_alias() calling conventions to simplify failure exits hostfs: use d_splice_alias() calling conventions to simplify failure exits udf_fiiter_add_entry(): check for zero ->d_name.len is bogus... udf: d_obtain_alias(ERR_PTR(...)) will do the right thing... udf: d_splice_alias() will do the right thing on ERR_PTR() inode nfsd: kill stale comment about simple_fill_super() requirements bfs_add_entry(): get rid of pointless ->d_name.len checks nilfs2: d_obtain_alias(ERR_PTR(...)) will do the right thing... zonefs: d_splice_alias() will do the right thing on ERR_PTR() inode 
Pull misc filesystem updates from Al Viro:
 "Misc cleanups (the part that hadn't been picked by individual fs
  trees)"

* tag 'pull-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  apparmorfs: don't duplicate kfree_link()
  orangefs: saner arguments passing in readdir guts
  ocfs2_find_match(): there's no such thing as NULL or negative ->d_parent
  reiserfs_add_entry(): get rid of pointless namelen checks
  __ocfs2_add_entry(), ocfs2_prepare_dir_for_insert(): namelen checks
  ext4_add_entry(): ->d_name.len is never 0
  befs: d_obtain_alias(ERR_PTR(...)) will do the right thing
  affs: d_obtain_alias(ERR_PTR(...)) will do the right thing
  /proc/sys: use d_splice_alias() calling conventions to simplify failure exits
  hostfs: use d_splice_alias() calling conventions to simplify failure exits
  udf_fiiter_add_entry(): check for zero ->d_name.len is bogus...
  udf: d_obtain_alias(ERR_PTR(...)) will do the right thing...
  udf: d_splice_alias() will do the right thing on ERR_PTR() inode
  nfsd: kill stale comment about simple_fill_super() requirements
  bfs_add_entry(): get rid of pointless ->d_name.len checks
  nilfs2: d_obtain_alias(ERR_PTR(...)) will do the right thing...
  zonefs: d_splice_alias() will do the right thing on ERR_PTR() inode
udf_fiiter_add_entry(): check for zero ->d_name.len is bogus... 2023-12-21T17:50:31+00:00 Al Viro viro@zeniv.linux.org.uk 2023-11-12T22:12:26+00:00 5e7582f6e3a89d2fecf0a08e44ee02780c4cb4cf Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>