Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1361196 > unrolled thread

[git pull] vfs.git

Back to article view | Back to linux.kernel

Contents

  [git pull] vfs.git Al Viro <viro@ZenIV.linux.org.uk> - 2016-03-20 02:50 +0100
    Re: [git pull] vfs.git Linus Torvalds <torvalds@linux-foundation.org> - 2016-03-20 03:00 +0100
      Re: [git pull] vfs.git Al Viro <viro@ZenIV.linux.org.uk> - 2016-03-20 03:10 +0100

#1361196 — [git pull] vfs.git

	Preparations of parallel lookups (the remaining main obstacle is the
need to move security_d_instantiate(); once that becomes safe, the rest will
be a matter of rather short series local to fs/*.c) + preadv2/pwritev2 series
from Christoph + assorted fixes.

The following changes since commit f93812846f31381d35c04c6c577d724254355e7f:

  jffs2: reduce the breakage on recovery from halfway failed rename() (2016-03-07 23:07:10 -0500)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus

for you to fetch changes up to 8b23a8ce1094f25a85826e25217c5b9779a4f5aa:

  Merge branches 'work.lookups', 'work.misc' and 'work.preadv2' into for-next (2016-03-18 16:07:38 -0400)

----------------------------------------------------------------
Al Viro (25):
      ecryptfs_lookup(): use lookup_one_len_unlocked()
      ecryptfs_encrypt_and_encode_filename(): drop unused argument
      Merge branch 'for-linus' into work.lookups
      do_last(): reorder and simplify a bit
      lookup_dcache(): lift d_alloc() into callers
      namei: untanlge lookup_fast()
      namei: change calling conventions for lookup_{fast,slow} and follow_managed()
      namei: simplify invalidation logics in lookup_dcache()
      lookup_one_len_unlocked(): use lookup_dcache()
      namei: massage lookup_slow() to be usable by lookup_one_len_unlocked()
      namei: teach lookup_slow() to skip revalidate
      ceph: don't bother with d_rehash() in splice_dentry()
      configfs: move d_rehash() into configfs_create() for regular files
      autofs4: don't bother with d_instantiate(dentry, NULL) in ->lookup()
      ceph_fill_trace(): don't bother with d_instantiate(dn, NULL)
      kill dentry_unhash()
      nfs_lookup: don't bother with d_instantiate(dentry, NULL)
      cifs_get_root(): use lookup_one_len_unlocked()
      quota: use lookup_one_len_unlocked()
      replace d_add_unique() with saner primitive
      uninline d_add()
      untangle fsnotify_d_instantiate() a bit
      don't bother with __d_instantiate(dentry, NULL)
      dcache.c: new helper: __d_add()
      Merge branches 'work.lookups', 'work.misc' and 'work.preadv2' into for-next

Christoph Hellwig (5):
      vfs: pass a flags argument to vfs_readv/vfs_writev
      x86: wire up preadv2 and pwritev2
      vfs: add the RWF_HIPRI flag for preadv2/pwritev2
      direct-io: only use block polling if explicitly requested
      blk-mq: enable polling support by default

David Howells (1):
      CacheFiles: Provide read-and-reset release counters for cachefilesd

Dmitry V. Levin (1):
      vfs: show_vfsstat: do not ignore errors from show_devname method

Milosz Tanski (1):
      vfs: vfs: Define new syscalls preadv2,pwritev2

Rabin Vincent (1):
      splice: handle zero nr_pages in splice_to_pipe()

 arch/x86/entry/syscalls/syscall_32.tbl |   2 +
 arch/x86/entry/syscalls/syscall_64.tbl |   2 +
 fs/autofs4/root.c                      |   2 -
 fs/cachefiles/daemon.c                 |  13 +-
 fs/cachefiles/interface.c              |  11 +-
 fs/cachefiles/internal.h               |   4 +
 fs/cachefiles/namei.c                  |  28 ++-
 fs/ceph/inode.c                        |  21 +--
 fs/cifs/cifsfs.c                       |   4 +-
 fs/configfs/dir.c                      |   9 +-
 fs/configfs/inode.c                    |  12 +-
 fs/dcache.c                            | 177 ++++++++++---------
 fs/direct-io.c                         |   3 +-
 fs/ecryptfs/crypto.c                   |  27 +--
 fs/ecryptfs/ecryptfs_kernel.h          |   1 -
 fs/ecryptfs/inode.c                    |  11 +-
 fs/namei.c                             | 309 ++++++++++++++-------------------
 fs/nfs/dir.c                           |  12 +-
 fs/nfs/nfs4proc.c                      |  13 +-
 fs/nfsd/vfs.c                          |   4 +-
 fs/proc_namespace.c                    |   2 +
 fs/quota/dquot.c                       |   4 +-
 fs/read_write.c                        | 197 ++++++++++++++++-----
 fs/splice.c                            |   5 +-
 include/linux/blkdev.h                 |   3 +-
 include/linux/compat.h                 |   6 +
 include/linux/dcache.h                 |  33 +---
 include/linux/fs.h                     |  10 +-
 include/linux/fsnotify.h               |   9 -
 include/linux/fsnotify_backend.h       |   9 +-
 include/linux/namei.h                  |   1 +
 include/linux/syscalls.h               |   6 +
 include/uapi/linux/fs.h                |   3 +
 33 files changed, 488 insertions(+), 465 deletions(-)

[toc] | [next] | [standalone]

#1361202

On Sat, Mar 19, 2016 at 6:44 PM, Al Viro <viro@zeniv.linux.org.uk> wrote:
>         Preparations of parallel lookups (the remaining main obstacle is the
> need to move security_d_instantiate(); once that becomes safe, the rest will
> be a matter of rather short series local to fs/*.c) + preadv2/pwritev2 series
> from Christoph + assorted fixes.

So this:

> Al Viro:
>       replace d_add_unique() with saner primitive

ends up conflicting with commit d9dfd8d74168 ("NFSv4: Fix a dentry
leak on alias use").

From what I can tell, your version doesn't have the leak issue, so my
merge resolution is to just take your new code as-is, but I'd ask you
to double-check my logic.

(I haven't pushed out the merge yet, it's still going through my build
test. Soon, but I guess you can just take a look at that dentry leak
commit regardless).

                  Linus

[toc] | [prev] | [next] | [standalone]

#1361204

On Sat, Mar 19, 2016 at 06:55:21PM -0700, Linus Torvalds wrote:
> On Sat, Mar 19, 2016 at 6:44 PM, Al Viro <viro@zeniv.linux.org.uk> wrote:
> >         Preparations of parallel lookups (the remaining main obstacle is the
> > need to move security_d_instantiate(); once that becomes safe, the rest will
> > be a matter of rather short series local to fs/*.c) + preadv2/pwritev2 series
> > from Christoph + assorted fixes.
> 
> So this:
> 
> > Al Viro:
> >       replace d_add_unique() with saner primitive
> 
> ends up conflicting with commit d9dfd8d74168 ("NFSv4: Fix a dentry
> leak on alias use").
> 
> >From what I can tell, your version doesn't have the leak issue, so my
> merge resolution is to just take your new code as-is, but I'd ask you
> to double-check my logic.
> 
> (I haven't pushed out the merge yet, it's still going through my build
> test. Soon, but I guess you can just take a look at that dentry leak
> commit regardless).

It does contain an equivalent of the leak fix, and yes, resolution should
be "take the variant from vfs.git in the area of conflict".

In fact, I was just pondering whether to send a followup describing the
conflict resolution (== take the variant from vfs.git in the area of
conflict) or leave it be as too trivial.  Should've sent...

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web