First pile out of several (there _definitely_ will be more). Stuff in this
one:
* unification of d_splice_alias()/d_materialize_unique()
* iov_iter rewrite
* killing a bunch of ->f_path.dentry users (and f_dentry macro).
Getting that completed will make life much simpler for unionmount/overlayfs,
since then we'll be able to limit the places sensitive to file _dentry_
to reasonably few. Which allows to have file_inode(file) pointing to
inode in a covered layer, with dentry pointing to (negative) dentry in
union one. Still not complete, but much closer now.
* crapectomy in lustre (dead code removal, mostly)
* "let's make seq_printf return nothing" preparations
* assorted cleanups and fixes
There _definitely_ will be more piles. Please, pull from
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus
Shortlog:
Al Viro (53):
new helper: is_root_inode()
move d_rcu from overlapping d_child to overlapping d_alias
lustre: switch ll_intent_file_open() to struct dentry *
ll_setxattr(): get rid of struct file on stack
lustre: opened file can't have negative dentry
lustre: use %p[dD]
lustre: use file_inode()
lustre: switch ll_release_openhandle() to struct inode *
kill ll_mknod_generic()
kill ll_mkdir_generic()
kill ll_symlink_generic()
ll_new_inode(): don't bother with name - it's always &dentry->d_name
kill ll_rmdir_generic()
kill ll_unlink_generic()
kill ll_rename_generic()
ll_get_child_fid(): callers already have the child
kill ll_link_generic()
lustre: get rid of duplicate mountpoint checks
lustre: use is_root_inode()
deal with deadlock in d_walk()
gfs2: bugger off early if O_CREAT open finds a directory
gfs2_create_inode(): don't bother with d_splice_alias()
gfs2_atomic_open(): simplify the use of finish_no_open()
d_add_ci() should just accept a hashed exact match if it finds one
merge d_materialise_unique() into d_splice_alias()
switch d_materialise_unique() users to d_splice_alias()
assorted conversions to %p[dD]
nfsd/nfsctl.c: new helper
btrfs: get rid of f_dentry use
cifs: get rid of ->f_path.dentry->d_sb uses, add a new helper
afs_fsync: don't bother with ->f_path.dentry
rpc_pipefs.c: get rid of f_dentry
nfsd: get rid of ->f_dentry
procfs: get rid of ->f_dentry
lockd: get rid of ->f_path.dentry->d_sb
kill f_dentry uses
ncpfs: use file_inode()
nfsd_vfs_write(): use file_inode()
new helper: audit_file()
kill f_dentry macro
iov_iter.c: macros for iterating over iov_iter
iov_iter.c: iterate_and_advance
iov_iter.c: convert iov_iter_npages() to iterate_all_kinds
iov_iter.c: convert iov_iter_get_pages() to iterate_all_kinds
iov_iter.c: convert iov_iter_get_pages_alloc() to iterate_all_kinds
iov_iter.c: convert iov_iter_zero() to iterate_and_advance
iov_iter.c: get rid of bvec_copy_page_{to,from}_iter()
iov_iter.c: convert copy_from_iter() to iterate_and_advance
iov_iter.c: convert copy_to_iter() to iterate_and_advance
iov_iter.c: handle ITER_KVEC directly
csum_and_copy_..._iter()
new helper: iov_iter_kvec()
copy_from_iter_nocache()
Joe Perches (6):
seq_file: Rename seq_overflow() to seq_has_overflowed() and make public
netfilter: Convert print_tuple functions to return void
dlm: Remove seq_printf() return checks and use seq_has_overflowed()
dlm: Use seq_puts() instead of seq_printf() for constant strings
fs: Convert show_fdinfo functions to void
debugfs: Have debugfs_print_regs32() return void
Miklos Szeredi (1):
vfs: make first argument of dir_context.actor typed
Mikulas Patocka (1):
dcache: fix kmemcheck warning in switch_names
Steven Rostedt (Red Hat) (2):
netfilter: Remove return values for print_conntrack callbacks
netfilter: Remove checks of seq_printf() return values
Diffstat:
Documentation/filesystems/debugfs.txt | 2 +-
Documentation/filesystems/nfs/Exporting | 23 +-
Documentation/filesystems/porting | 8 +
Documentation/filesystems/seq_file.txt | 22 +-
Documentation/filesystems/vfs.txt | 2 +-
arch/alpha/kernel/osf_sys.c | 7 +-
arch/parisc/hpux/fs.c | 7 +-
arch/powerpc/oprofile/cell/spu_task_sync.c | 10 +-
arch/powerpc/platforms/cell/spufs/inode.c | 2 +-
arch/s390/hypfs/hypfs_dbfs.c | 3 +-
arch/x86/ia32/ia32_aout.c | 8 +-
drivers/block/drbd/drbd_debugfs.c | 6 +-
drivers/gpu/drm/armada/armada_gem.c | 2 +-
drivers/media/pci/zoran/zoran_procfs.c | 4 +-
drivers/misc/genwqe/card_dev.c | 2 +-
drivers/net/tun.c | 4 +-
drivers/s390/char/hmcdrv_dev.c | 16 +-
drivers/scsi/lpfc/lpfc_debugfs.c | 8 +-
drivers/staging/lustre/lustre/libcfs/tracefile.c | 4 +-
drivers/staging/lustre/lustre/llite/dcache.c | 21 +-
drivers/staging/lustre/lustre/llite/dir.c | 9 +-
drivers/staging/lustre/lustre/llite/file.c | 158 ++-
.../staging/lustre/lustre/llite/llite_internal.h | 10 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 8 +-
drivers/staging/lustre/lustre/llite/llite_mmap.c | 12 +-
drivers/staging/lustre/lustre/llite/llite_nfs.c | 8 +-
drivers/staging/lustre/lustre/llite/lloop.c | 7 +-
drivers/staging/lustre/lustre/llite/namei.c | 388 +++----
drivers/staging/lustre/lustre/llite/statahead.c | 33 +-
drivers/staging/lustre/lustre/llite/vvp_io.c | 4 +-
drivers/staging/lustre/lustre/llite/xattr.c | 9 +-
fs/9p/vfs_inode.c | 2 +-
fs/9p/vfs_inode_dotl.c | 4 +-
fs/affs/amigaffs.c | 2 +-
fs/affs/inode.c | 4 +-
fs/affs/namei.c | 40 +-
fs/afs/dir.c | 80 +-
fs/afs/inode.c | 4 +-
fs/afs/mntpt.c | 22 +-
fs/afs/write.c | 9 +-
fs/autofs4/expire.c | 42 +-
fs/autofs4/root.c | 25 +-
fs/befs/linuxvfs.c | 16 +-
fs/binfmt_aout.c | 8 +-
fs/btrfs/inode.c | 2 +-
fs/btrfs/ioctl.c | 2 +-
fs/cachefiles/namei.c | 21 +-
fs/cachefiles/xattr.c | 15 +-
fs/ceph/debugfs.c | 14 +-
fs/ceph/dir.c | 56 +-
fs/ceph/file.c | 6 +-
fs/ceph/inode.c | 18 +-
fs/cifs/cifsfs.c | 3 +-
fs/cifs/cifsglob.h | 6 +
fs/cifs/file.c | 12 +-
fs/cifs/inode.c | 2 +-
fs/cifs/readdir.c | 10 +-
fs/coda/cache.c | 2 +-
fs/coda/coda_linux.c | 6 -
fs/coda/coda_linux.h | 1 -
fs/coda/dir.c | 12 +-
fs/compat.c | 21 +-
fs/configfs/dir.c | 2 +-
fs/dcache.c | 271 ++---
fs/debugfs/file.c | 15 +-
fs/debugfs/inode.c | 2 +-
fs/dlm/debug_fs.c | 263 +++--
fs/ecryptfs/crypto.c | 2 +-
fs/ecryptfs/file.c | 6 +-
fs/ecryptfs/mmap.c | 2 +-
fs/efivarfs/file.c | 4 +-
fs/eventfd.c | 9 +-
fs/eventpoll.c | 13 +-
fs/exportfs/expfs.c | 7 +-
fs/fat/dir.c | 5 +-
fs/fuse/dir.c | 4 +-
fs/fuse/file.c | 2 +-
fs/gfs2/export.c | 8 +-
fs/gfs2/inode.c | 26 +-
fs/hppfs/hppfs.c | 5 +-
fs/jfs/namei.c | 18 +-
fs/kernfs/dir.c | 2 +-
fs/libfs.c | 12 +-
fs/lockd/svcsubs.c | 2 +-
fs/ncpfs/dir.c | 12 +-
fs/ncpfs/file.c | 14 +-
fs/ncpfs/mmap.c | 4 +-
fs/ncpfs/ncplib_kernel.h | 4 +-
fs/nfs/blocklayout/rpc_pipefs.c | 2 +-
fs/nfs/dir.c | 6 +-
fs/nfs/getroot.c | 4 +-
fs/nfsd/nfs4recover.c | 7 +-
fs/nfsd/nfs4xdr.c | 2 +-
fs/nfsd/nfsctl.c | 37 +-
fs/nfsd/vfs.c | 16 +-
fs/nfsd/vfs.h | 4 +-
fs/notify/fdinfo.c | 78 +-
fs/notify/fdinfo.h | 4 +-
fs/notify/fsnotify.c | 4 +-
fs/ntfs/namei.c | 4 +-
fs/ocfs2/dcache.c | 20 +-
fs/ocfs2/dir.c | 8 +-
fs/ocfs2/dlmfs/dlmfs.c | 4 +-
fs/ocfs2/dlmglue.c | 3 +-
fs/ocfs2/journal.c | 8 +-
fs/open.c | 4 +-
fs/overlayfs/readdir.c | 8 +-
fs/proc/base.c | 4 +-
fs/proc/fd.c | 3 +-
fs/readdir.c | 21 +-
fs/reiserfs/xattr.c | 21 +-
fs/seq_file.c | 15 +-
fs/signalfd.c | 4 +-
fs/sync.c | 2 +-
fs/timerfd.c | 27 +-
fs/xattr.c | 16 +-
include/linux/audit.h | 9 +
include/linux/cgroup.h | 4 +-
include/linux/dcache.h | 9 +-
include/linux/debugfs.h | 7 +-
include/linux/fs.h | 13 +-
include/linux/seq_file.h | 15 +
include/linux/uio.h | 6 +
include/net/netfilter/nf_conntrack_core.h | 2 +-
include/net/netfilter/nf_conntrack_l3proto.h | 4 +-
include/net/netfilter/nf_conntrack_l4proto.h | 6 +-
ipc/mqueue.c | 4 +-
kernel/auditsc.c | 7 +-
kernel/events/core.c | 2 +-
kernel/taskstats.c | 2 +-
kernel/trace/trace.c | 4 +-
kernel/trace/trace_events.c | 2 +-
mm/iov_iter.c | 1058 ++++++++------------
mm/memcontrol.c | 4 +-
net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c | 6 +-
.../netfilter/nf_conntrack_l3proto_ipv4_compat.c | 53 +-
net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 10 +-
net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c | 6 +-
net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 10 +-
net/netfilter/nf_conntrack_l3proto_generic.c | 5 +-
net/netfilter/nf_conntrack_proto_dccp.c | 14 +-
net/netfilter/nf_conntrack_proto_generic.c | 5 +-
net/netfilter/nf_conntrack_proto_gre.c | 18 +-
net/netfilter/nf_conntrack_proto_sctp.c | 14 +-
net/netfilter/nf_conntrack_proto_tcp.c | 14 +-
net/netfilter/nf_conntrack_proto_udp.c | 10 +-
net/netfilter/nf_conntrack_proto_udplite.c | 10 +-
net/netfilter/nf_conntrack_standalone.c | 77 +-
net/netfilter/nf_log.c | 30 +-
net/netfilter/nfnetlink_queue_core.c | 13 +-
net/netfilter/x_tables.c | 19 +-
net/netfilter/xt_hashlimit.c | 36 +-
security/commoncap.c | 2 +-
security/integrity/ima/ima_api.c | 4 +-
security/integrity/ima/ima_appraise.c | 4 +-
security/integrity/ima/ima_template_lib.c | 2 +-
security/selinux/selinuxfs.c | 6 +-
security/smack/smack_lsm.c | 8 +-
158 files changed, 1675 insertions(+), 2183 deletions(-)
On Wed, Dec 10, 2014 at 8:13 PM, Al Viro <[email protected]> wrote:
> * killing a bunch of ->f_path.dentry users (and f_dentry macro).
> Getting that completed will make life much simpler for unionmount/overlayfs,
> since then we'll be able to limit the places sensitive to file _dentry_
> to reasonably few. Which allows to have file_inode(file) pointing to
> inode in a covered layer, with dentry pointing to (negative) dentry in
> union one. Still not complete, but much closer now.
If we want f_path to point to union path, then shouldn't we be hiding
it from all but core VFS? Filesystems have no business looking at the
mount anyway. And looking at the union dentry will be just an endless
source of bugs.
Also some API changes are in order, like passing dentry to ->iterate()
for "readdirplus" like operations, since f_path.dentry won't be usable
for that purpose. Alternatively replace f_inode with f_dentry and
make file_inode() use that.
Thanks,
Miklos
On Thu, Dec 11, 2014 at 05:18:21PM +0100, Miklos Szeredi wrote:
> On Wed, Dec 10, 2014 at 8:13 PM, Al Viro <[email protected]> wrote:
>
> > * killing a bunch of ->f_path.dentry users (and f_dentry macro).
> > Getting that completed will make life much simpler for unionmount/overlayfs,
> > since then we'll be able to limit the places sensitive to file _dentry_
> > to reasonably few. Which allows to have file_inode(file) pointing to
> > inode in a covered layer, with dentry pointing to (negative) dentry in
> > union one. Still not complete, but much closer now.
>
> If we want f_path to point to union path, then shouldn't we be hiding
> it from all but core VFS? Filesystems have no business looking at the
> mount anyway. And looking at the union dentry will be just an endless
> source of bugs.
>
> Also some API changes are in order, like passing dentry to ->iterate()
> for "readdirplus" like operations, since f_path.dentry won't be usable
> for that purpose. Alternatively replace f_inode with f_dentry and
> make file_inode() use that.
Check the remaining users of ->f_path.dentry. I agree that they ought to
be encapsulated into a sane set of helpers - raw access to ->f_path in
filesystems is asking for serious trouble. And we have very few such places
remaining.
FWIW, I would prefer to have the whole stack of underlying dentries reachable
in a uniform way...
PS: I'll send review of your multilayer stuff today.
On Thu, Dec 11, 2014 at 06:06:24PM +0000, Al Viro wrote:
> Check the remaining users of ->f_path.dentry. I agree that they ought to
> be encapsulated into a sane set of helpers - raw access to ->f_path in
> filesystems is asking for serious trouble. And we have very few such places
> remaining.
Sloppy writing: insert "with the exception of filesystems that are not
going to be suitable for any form of stacking" - there _is_ a bunch of
debugfs, configfs, kernfs places using that. They'd also better be
taken care of, but it's an independent story; for overlayfs/unionmount/etc.
purposes they are non-issue.
BTW, there are several places where ->f_path.mnt access is the right thing
to do - in autofs. Also a non-issue here...