vfs.git fixes
Al Viro
maintainer Assorted fixes all over the place. BTW, what is the preferred way
of sending a pull request on a series where an old commit has grown an ACKed-by? I've moved original branch to #fixes-old, git commit --amend on commit[s] in question and rebased the tail of the series on top of that; all trees involved are identical to ones in matching commits in original branch, so all testing results (including integration in -next, etc.) are still valid. Is that approach OK with you? Technically it is a rebase, but it seems that none of the usual objections apply here...
The following changes since commit 16a34adb9392b2fe4195267475ab5b472e55292c:
Don't leak MNT_INTERNAL away from internal mounts (2018-04-19 23:52:15 -0400)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git fixes
for you to fetch changes up to baf10564fbb66ea222cae66fbff11c444590ffd9:
aio: fix io_destroy(2) vs. lookup_ioctx() race (2018-05-21 14:30:11 -0400)
----------------------------------------------------------------
Al Viro (11):
do d_instantiate/unlock_new_inode combinations safely
fix breakage caused by d_find_alias() semantics change affs_lookup(): close a race with affs_remove_link()
affs_lookup: switch to d_splice_alias()
befs_lookup(): use d_splice_alias()
kernfs: deal with kernfs_fill_super() failures
unfuck sysfs_mount()
cachefiles: vfs_mkdir() might succeed leaving dentry negative unhashed
nfsd: vfs_mkdir() might succeed leaving dentry negative unhashed
ext2: fix a block leak
aio: fix io_destroy(2) vs. lookup_ioctx() race
Dave Chinner (1):
fs: don't scan the inode cache before SB_BORN is set
Ilya Dryomov (2):
iov_iter: fix return type of __pipe_get_pages()
iov_iter: fix memory leak in pipe_get_pages_alloc()
Joe Perches (1):
cramfs: Fix IS_ENABLED typo
fs/affs/namei.c | 15 +++++++++------ fs/aio.c | 4 ++-- fs/befs/linuxvfs.c | 17 +++++------------ fs/btrfs/inode.c | 16 ++++------------ fs/cachefiles/namei.c | 10 ++++++++++ fs/cramfs/inode.c | 2 +- fs/dcache.c | 22 ++++++++++++++++++++++ fs/ecryptfs/inode.c | 3 +-- fs/ext2/inode.c | 10 ---------- fs/ext2/namei.c | 6 ++---- fs/ext4/namei.c | 6 ++---- fs/f2fs/namei.c | 12 ++++-------- fs/jffs2/dir.c | 12 ++++-------- fs/jfs/namei.c | 12 ++++-------- fs/kernfs/mount.c | 1 + fs/nfsd/vfs.c | 22 ++++++++++++++++++++++ fs/nilfs2/namei.c | 6 ++---- fs/orangefs/namei.c | 9 +++------ fs/reiserfs/namei.c | 12 ++++-------- fs/super.c | 30 ++++++++++++++++++++++++------ fs/sysfs/mount.c | 6 +++--- fs/udf/namei.c | 6 ++---- fs/ufs/namei.c | 6 ++---- include/linux/dcache.h | 1 + lib/iov_iter.c | 4 ++-- security/selinux/hooks.c | 20 ++++++++++++++++---- 26 files changed, 152 insertions(+), 118 deletions(-)
Quoted 4 lines · Al Viro
BTW, what is the preferred way
of sending a pull request on a series where an old commit has grown an ACKed-by? I've moved original branch to #fixes-old, git commit --amend on commit[s] in question and rebased the tail of the series on top of
that;
That's exactly the same as just doing a "git rebase -i <commit>^" where 'commit' is the thing you want to amend, and then just picking that for 'r' ('reword').
In general, I'd prefer people not do it, unless there is a very strong reason to reword the commit message. I'm not sure just an ack is necessarily reason enough.
Anyway, pulled.
Linus