[GIT PULL] namespace related fixes for 4.10-rc3

[GIT PULL] namespace related fixes for 4.10-rc3

[Eric W. Biederman]

Linus,

Please pull the for-linus branch from the git tree:

   git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace.git for-linus

   HEAD: cd293bfacd57928627ef7da6194e0cbe24db3291 sysctl: Drop reference added by grab_header in proc_sys_readdir

This tree contains 5 fixes.

The first is a fix for a race that can causes oopses under the right
circumstances, and that someone just recently encountered.

The second is a fix that removes the possibility of creating unbounded
chains in the mount hash table.  After a lot of back and forth this fix
has finally matured into something mergeable.

Past that are several small trivial correct fixes.  A real issue that
was blocking development of an out of tree driver, but does not appear
to have caused any actual problems for in-tree code.  A potential
deadlock that was reported by lockdep.  And a deadlock people have
experienced and took the time to track down caused by a cleanup that
removed the code to drop a reference count.

Andrei Vagin (1):
      pid: fix lockdep deadlock warning due to ucount_lock

Eric W. Biederman (3):
      mnt: Protect the mountpoint hashtable with mount_lock
      mnt: Tuck mounts under others instead of creating shadow/side mounts.
      libfs: Modify mount_pseudo_xattr to be clear it is not a userspace mount

Zhou Chengming (1):
      sysctl: Drop reference added by grab_header in proc_sys_readdir

 fs/dcache.c            |   7 +-
 fs/libfs.c             |   3 +-
 fs/mount.h             |   1 -
 fs/namespace.c         | 174 +++++++++++++++++++++++++++++--------------------
 fs/pnode.c             |  27 ++++++--
 fs/pnode.h             |   2 +
 fs/proc/proc_sysctl.c  |   3 +-
 kernel/pid_namespace.c |  10 +--
 8 files changed, 142 insertions(+), 85 deletions(-)

Eric

Re: [GIT PULL] namespace related fixes for 4.10-rc3

[Al Viro]

On Sat, Jan 07, 2017 at 06:30:05PM +1300, Eric W. Biederman wrote:

> Eric W. Biederman (3):
>       mnt: Protect the mountpoint hashtable with mount_lock
ACK

>       mnt: Tuck mounts under others instead of creating shadow/side mounts.

NAK; you have reintroduced put_mountpoint() outside of mount lock in there
and you've got seriously confused umount side of the things.

>       libfs: Modify mount_pseudo_xattr to be clear it is not a userspace mount
ACK

> Zhou Chengming (1):
>       sysctl: Drop reference added by grab_header in proc_sys_readdir

ACK

The rest of the queue is fine.

[GIT PULL] namespace related fixes for 4.10-rc4

[Eric W. Biederman]

Linus,

Please pull the for-linus branch from the git tree:

   git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace.git for-linus

   HEAD: 93362fa47fe98b62e4a34ab408c4a418432e7939 sysctl: Drop reference added by grab_header in proc_sys_readdir

This tree contains 4 fixes.

The first is a fix for a race that can causes oopses under the right
circumstances, and that someone just recently encountered.

Past that are several small trivial correct fixes.  A real issue that
was blocking development of an out of tree driver, but does not appear
to have caused any actual problems for in-tree code.  A potential
deadlock that was reported by lockdep.  And a deadlock people have
experienced and took the time to track down caused by a cleanup that
removed the code to drop a reference count.

I have dropped the mount hash table fix that Al objected to and will
come back to that when the discussion around that patch is complete.

Andrei Vagin (1):
      pid: fix lockdep deadlock warning due to ucount_lock

Eric W. Biederman (2):
      mnt: Protect the mountpoint hashtable with mount_lock
      libfs: Modify mount_pseudo_xattr to be clear it is not a userspace mount

Zhou Chengming (1):
      sysctl: Drop reference added by grab_header in proc_sys_readdir

 fs/dcache.c            |  7 ++++--
 fs/libfs.c             |  3 ++-
 fs/namespace.c         | 64 +++++++++++++++++++++++++++++++++++---------------
 fs/proc/proc_sysctl.c  |  3 ++-
 kernel/pid_namespace.c | 10 ++++----
 5 files changed, 60 insertions(+), 27 deletions(-)

Eric

Re: [GIT PULL] namespace related fixes for 4.10-rc3

[Al Viro]

On Sat, Jan 07, 2017 at 06:29:52AM +0000, Al Viro wrote:
> On Sat, Jan 07, 2017 at 06:30:05PM +1300, Eric W. Biederman wrote:
> 
> > Eric W. Biederman (3):
> >       mnt: Protect the mountpoint hashtable with mount_lock
> ACK
> 
> >       mnt: Tuck mounts under others instead of creating shadow/side mounts.
> 
> NAK; you have reintroduced put_mountpoint() outside of mount lock in there
> and you've got seriously confused umount side of the things.

PS: details are in my reply (admittedly, sent only ~20 minutes
before your pull request) to your patch.  Check Message-ID
<20170107050644.GA12074-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>

My apologies for not getting to it yesterday (the patch was posted a couple of
days ago, AFAICS) - I'd been off-net since Wednesday, just got back ;-/

Re: [GIT PULL] namespace related fixes for 4.10-rc3

[Eric W. Biederman]

Al Viro <viro-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org> writes:

> On Sat, Jan 07, 2017 at 06:29:52AM +0000, Al Viro wrote:
>> On Sat, Jan 07, 2017 at 06:30:05PM +1300, Eric W. Biederman wrote:
>> 
>> > Eric W. Biederman (3):
>> >       mnt: Protect the mountpoint hashtable with mount_lock
>> ACK
>> 
>> >       mnt: Tuck mounts under others instead of creating shadow/side mounts.
>> 
>> NAK; you have reintroduced put_mountpoint() outside of mount lock in there
>> and you've got seriously confused umount side of the things.
>
> PS: details are in my reply (admittedly, sent only ~20 minutes
> before your pull request) to your patch.  Check Message-ID
> <20170107050644.GA12074-3bDd1+5oDREiFSDQTTA3OLVCufUGDwFn@public.gmane.org>

I suspect that was sent while I was composing my pull request, and the
locking around put_mountpoint I did fail to catch during my rebase.  I
will finish our conversation about that before I resend that patch.

> My apologies for not getting to it yesterday (the patch was posted a couple of
> days ago, AFAICS) - I'd been off-net since Wednesday, just got back
> ;-/

You and Ram Pai had said reasonably pleasant things about a previous
version of that patch.

Eric