vfs.git
Al Viro
maintainerAll kinds of stuff this time around; some more notable parts:
* RCU'd vfsmounts handling
* new primitives for coredump handling
* files_lock is gone
* Bruce's delegations handling series
* exportfs fixes
plus misc stuff all over the place. There's one trivial conflict with mainline - drivers/usb/core/file.c (whitespace noise on mainline side vs. switch to replace_fops() on vfs.git one); no other conflicts. Please, pull fromgit://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus
Shortlog:
Al Viro (72):
9p: make v9fs_cache_inode_{get,put,set}_cookie empty inlines for !9P_CACHEFS
ecryptfs: check DCACHE_OP_REVALIDATE instead of ->d_op
ecryptfs: don't leave RCU pathwalk immediately
ecryptfs: get rid of ecryptfs_set_dentry_lower{,_mnt}
ecryptfs: ->lower_path.dentry is never NULL
libfs: get exports to definitions of objects being exported... new helper: kfree_put_link()
befs: split symlink iops in two - for short and long symlinks resp.
nfs: use %p[dD] instead of open-coded (and often racy) equivalents
nfsd: switch to %p[dD]
sunrpc: switch to %pd
ubifs: switch to %pd
ncpfs: switch to %p[dD]
put_mnt_ns(): use drop_collected_mounts()
consolidate the reassignments of ->f_op in ->open() instances
dmxdev: get rid of pointless clearing ->f_op
rtl8188eu: remove dead code
file->f_op is never NULL...
initialize namespace_sem statically
fs_is_visible only needs namespace_sem held shared
dup_mnt_ns(): get rid of pointless grabbing of vfsmount_lock do_remount(): pull touch_mnt_namespace() up
fold mntfree() into mntput_no_expire()
fs/namespace.c: bury long-dead define
finish_automount() doesn't need vfsmount_lock for removal from expiry list mnt_set_expiry() doesn't need vfsmount_lock
fold dup_mnt_ns() into its only surviving caller
namespace.c: get rid of mnt_ghosts
don't bother with vfsmount_lock in mounts_poll()
new helpers: lock_mount_hash/unlock_mount_hash
isofs: don't pass dentry to isofs_hash{i,}_common()
uninline destroy_super(), consolidate alloc_super()
split __lookup_mnt() in two functions
move taking vfsmount_lock down into prepend_path()
make freeing super_block rcu-delayed
hpfs: make freeing sbi and codetables rcu-delayed
adfs: delayed freeing of sbi
autofs4: make freeing sbi rcu-delayed
cifs: rcu-delay unload_nls() and freeing sbi
fat: rcu-delay unloading nls and freeing sbi
ncpfs: rcu-delay unload_nls() and freeing ncp_server
pid_namespace: make freeing struct pid_namespace rcu-delayed
fuse: rcu-delay freeing fuse_conn
switch shrink_dcache_for_umount() to use of d_walk()
RCU'd vfsmounts
get rid of {lock,unlock}_rcu_walk()
get rid of s_files and files_lock
fold __d_shrink() into its only remaining caller coda_revalidate_inode(): switch to passing inode...
no need to keep brlock macros anymore...
restore 32bit aout coredump
new helper: dump_emit()
switch elf_core_write_extra_phdrs() to dump_emit()
switch elf_core_write_extra_data() to dump_emit()
binfmt_elf: convert writing actual dump pages to dump_emit() convert the rest of binfmt_elf_fdpic to dump_emit()
switch elf_coredump_extra_notes_write() to dump_emit()
aout: switch to dump_emit
binfmt_elf: count notes towards coredump limit
make dump_emit() use vfs_write() instead of banging at ->f_op->write directly dump_skip(): dump_seek() replacement taking coredump_params
spufs: get rid of dump_emit() wrappers
new helper: dump_align()
take anon inode allocation to libfs.c
... and kill anon_inode_getfile_private()
constify copy_siginfo_to_user{,32}()
constify do_coredump() argument
elf{,_fdpic} coredump: get rid of pointless if (siginfo->si_signo) iget/iget5: don't bother with ->i_lock until we find a match
qnx4: i_sb is never NULL
ocfs2: get rid of impossible checks
ecryptfs: ->f_op is never NULL
Benjamin LaHaise (1):
rework aio migrate pages to use aio fs
Christoph Hellwig (1):
exportfs: BUG_ON in crazy corner case
David Howells (1):
VFS: Put a small type field into struct dentry::d_flags
J. Bruce Fields (25):
vfs: split out vfs_getattr_nosec
exportfs: fix 32-bit nfsd handling of 64-bit inode numbers
dcache: use IS_ROOT to decide where dentry is hashed
dcache: Don't set DISCONNECTED on "pseudo filesystem" dentries
dcache: don't clear DCACHE_DISCONNECTED too early
dcache: fix outdated DCACHE_NEED_LOOKUP comment
exportfs: more detailed comment for path_reconnect
exportfs: clear DISCONNECTED on all parents sooner
exportfs: stop retrying once we race with rename/remove
exportfs: eliminate unused "noprogress" counter
exportfs: move most of reconnect_path to helper function
exportfs: better variable name
exportfs: fix quadratic behavior in filehandle lookup
vfs: pull ext4's double-i_mutex-locking into common code
vfs: don't use PARENT/CHILD lock classes for non-directories
vfs: rename I_MUTEX_QUOTA now that it's not used for quotas
vfs: take i_mutex on renamed file
locks: introduce new FL_DELEG lock flag
locks: implement delegations
namei: minor vfs_unlink cleanup
locks: break delegations on unlink
locks: helper functions for delegation breaking
locks: break delegations on rename
locks: break delegations on link
locks: break delegations on any attribute modification
Miklos Szeredi (1):
vfs: introduce d_instantiate_no_diralias()
Diffstat:
Documentation/filesystems/directory-locking | 31 ++- Documentation/filesystems/porting | 8 + arch/arm64/kernel/signal32.c | 2 +- arch/ia64/kernel/elfcore.c | 12 +- arch/ia64/kernel/signal.c | 2 +- arch/mips/kernel/signal32.c | 2 +- arch/parisc/kernel/signal32.c | 2 +- arch/parisc/kernel/signal32.h | 2 +- arch/powerpc/include/asm/spu.h | 3 +- arch/powerpc/kernel/signal_32.c | 2 +- arch/powerpc/platforms/cell/spu_syscalls.c | 5 +- arch/powerpc/platforms/cell/spufs/coredump.c | 89 ++---- arch/powerpc/platforms/cell/spufs/spufs.h | 3 +- arch/s390/kernel/compat_signal.c | 2 +- arch/sparc/kernel/signal32.c | 2 +- arch/tile/kernel/compat_signal.c | 2 +- arch/x86/ia32/ia32_aout.c | 86 +++--- arch/x86/ia32/ia32_signal.c | 2 +- arch/x86/um/elfcore.c | 15 +- drivers/base/devtmpfs.c | 6 +- drivers/char/misc.c | 12 +- drivers/gpu/drm/drm_fops.c | 17 +- drivers/media/dvb-core/dmxdev.c | 4 - drivers/media/dvb-core/dvbdev.c | 19 +- drivers/mtd/nand/nandsim.c | 2 +- drivers/staging/comedi/comedi_compat32.c | 3 - .../lustre/lustre/include/linux/lustre_compat25.h | 4 +- drivers/staging/lustre/lustre/llite/namei.c | 2 +- drivers/staging/lustre/lustre/lvfs/lvfs_linux.c | 2 +- drivers/staging/rtl8188eu/include/osdep_service.h | 5 - drivers/staging/rtl8188eu/os_dep/osdep_service.c | 208 ----------- drivers/usb/core/file.c | 16 +- fs/9p/cache.h | 12 + fs/9p/vfs_file.c | 2 - fs/9p/vfs_inode.c | 6 - fs/9p/vfs_inode_dotl.c | 4 - fs/adfs/adfs.h | 9 +- fs/adfs/super.c | 3 +- fs/aio.c | 63 +++- fs/anon_inodes.c | 114 +------ fs/attr.c | 25 ++- fs/autofs4/autofs_i.h | 3 +- fs/autofs4/dev-ioctl.c | 6 - fs/autofs4/inode.c | 13 +- fs/befs/linuxvfs.c | 61 ++-- fs/binfmt_aout.c | 13 +- fs/binfmt_elf.c | 127 +++---- fs/binfmt_elf_fdpic.c | 152 +++------ fs/binfmt_em86.c | 2 +- fs/cachefiles/interface.c | 4 +- fs/cachefiles/namei.c | 4 +- fs/char_dev.c | 6 +- fs/cifs/cifs_fs_sb.h | 1 + fs/cifs/cifsfs.c | 2 +- fs/cifs/cifsfs.h | 2 - fs/cifs/connect.c | 10 +- fs/cifs/link.c | 7 - fs/coda/coda_linux.h | 2 +- fs/coda/dir.c | 6 +- fs/coda/file.c | 6 +- fs/coda/inode.c | 2 +- fs/compat_ioctl.c | 4 +- fs/coredump.c | 71 +++-- fs/dcache.c | 342 ++++++++++-------- fs/ecryptfs/dentry.c | 29 +- fs/ecryptfs/ecryptfs_kernel.h | 19 +- fs/ecryptfs/file.c | 8 +- fs/ecryptfs/inode.c | 29 +- fs/ecryptfs/main.c | 3 +- fs/eventpoll.c | 2 +- fs/exec.c | 35 +- fs/exportfs/expfs.c | 269 ++++++++------ fs/ext4/ext4.h | 2 - fs/ext4/ioctl.c | 4 +- fs/ext4/move_extent.c | 40 +-- fs/fat/fat.h | 1 + fs/fat/inode.c | 19 +- fs/fcntl.c | 5 +- fs/file_table.c | 129 +------- fs/fs-writeback.c | 1 + fs/fuse/cuse.c | 2 +- fs/fuse/dir.c | 40 +-- fs/fuse/fuse_i.h | 5 +- fs/fuse/inode.c | 4 +- fs/gfs2/inode.c | 9 +- fs/hpfs/hpfs_fn.h | 1 + fs/hpfs/namei.c | 2 +- fs/hpfs/super.c | 28 +- fs/inode.c | 62 +++- fs/internal.h | 7 - fs/ioctl.c | 4 +- fs/isofs/inode.c | 12 +- fs/libfs.c | 122 +++++-- fs/locks.c | 69 +++- fs/mount.h | 20 +- fs/namei.c | 322 ++++++++++------- fs/namespace.c | 390 ++++++++++---------- fs/ncpfs/dir.c | 55 +-- fs/ncpfs/file.c | 12 +- fs/ncpfs/inode.c | 19 +- fs/ncpfs/ncp_fs_sb.h | 2 +- fs/nfs/dir.c | 119 +++---- fs/nfs/direct.c | 17 +- fs/nfs/file.c | 117 +++---- fs/nfs/namespace.c | 5 +- fs/nfs/nfs3proc.c | 8 +- fs/nfs/nfs4file.c | 4 +- fs/nfs/nfs4namespace.c | 7 +- fs/nfs/nfs4proc.c | 5 +- fs/nfs/proc.c | 8 +- fs/nfs/unlink.c | 9 +- fs/nfs/write.c | 6 +- fs/nfsd/nfs4recover.c | 12 +- fs/nfsd/nfs4state.c | 17 +- fs/nfsd/nfsfh.c | 28 +- fs/nfsd/nfsfh.h | 4 +- fs/nfsd/vfs.c | 23 +- fs/ntfs/inode.c | 2 +- fs/ocfs2/inode.c | 10 - fs/open.c | 32 ++- fs/pnode.c | 13 +- fs/proc/self.c | 10 +- fs/proc_namespace.c | 8 +- fs/qnx4/namei.c | 4 - fs/read_write.c | 25 +- fs/readdir.c | 2 +- fs/select.c | 4 +- fs/splice.c | 6 +- fs/stat.c | 31 ++- fs/super.c | 201 ++++------ fs/sync.c | 2 +- fs/ubifs/dir.c | 41 +-- fs/ubifs/journal.c | 6 +- fs/ubifs/xattr.c | 16 +- fs/utimes.c | 9 +- include/asm-generic/siginfo.h | 2 +- include/linux/anon_inodes.h | 3 - include/linux/binfmts.h | 3 +- include/linux/compat.h | 2 +- include/linux/coredump.h | 10 +- include/linux/dcache.h | 104 +++++- include/linux/elf.h | 6 +- include/linux/elfcore.h | 7 +- include/linux/fs.h | 106 ++++-- include/linux/lglock.h | 10 - include/linux/mount.h | 2 + include/linux/namei.h | 2 +- include/linux/pid_namespace.h | 1 + ipc/mqueue.c | 2 +- kernel/elfcore.c | 10 +- kernel/pid_namespace.c | 8 +- kernel/signal.c | 2 +- mm/memory.c | 2 +- mm/mmap.c | 4 +- mm/nommu.c | 2 +- net/9p/trans_fd.c | 4 +- net/sunrpc/rpc_pipe.c | 12 +- sound/core/sound.c | 22 +- sound/sound_core.c | 17 +- 159 files changed, 2100 insertions(+), 2492 deletions(-)
Al Viro
maintainer[git pull] vfs.git fixes
Several fixes, mostly for regressions in the last pile. Howeover, prepend_path() forgetting to reininitalize dentry/vfsmount is in 3.12 as well and qib_fs had been leaking all along... Please, pull fromgit://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus
Shortlog:
Al Viro (2):
prepend_path() needs to reinitialize dentry/vfsmount/mnt on restarts
qib_fs: fix (some) dcache abuses
Dan Carpenter (2):
aio: checking for NULL instead of IS_ERR
locks: missing unlock on error in generic_add_lease()
Li Zhong (1):
fix unpaired rcu lock in prepend_path()
Diffstat:
drivers/infiniband/hw/qib/qib_fs.c | 7 +++++-- fs/aio.c | 4 ++-- fs/dcache.c | 13 ++++++++++--- fs/locks.c | 1 + 4 files changed, 18 insertions(+), 7 deletions(-)
Quoted 6 lines · Al Viro
All kinds of stuff this time around; some more notable parts:
* RCU'd vfsmounts handling
* new primitives for coredump handling
*files_lockis gone
* Bruce's delegations handling series
* exportfs fixes
Thanks. The one thing I have left for you right now is jlayton's fix for a lease/open race--is that still on your queue someplace?
It's at
git://linux-nfs.org/~bfields/linux.git for-viro
if you need it.
(Nothing urgent, the race has been there forever, just want to make sure it doesn't fall through the cracks.)
--b.