[PATCH] namespaces: fix exit race by splitting exit
authorSerge E. Hallyn <serue@us.ibm.com>
Mon, 29 Jan 2007 21:19:40 +0000 (13:19 -0800)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Tue, 30 Jan 2007 16:26:44 +0000 (08:26 -0800)
commit7a238fcba0629b6f2edbcd37458bae56fcf36be5
treead556f0ec00637df5b4c4a2063c6b3325666d2f1
parentc0d4d573feed199b16094c072e7cb07afb01c598
[PATCH] namespaces: fix exit race by splitting exit

Fix exit race by splitting the nsproxy putting into two pieces.  First
piece reduces the nsproxy refcount.  If we dropped the last reference, then
it puts the mnt_ns, and returns the nsproxy as a hint to the caller.  Else
it returns NULL.  The second piece of exiting task namespaces sets
tsk->nsproxy to NULL, and drops the references to other namespaces and
frees the nsproxy only if an nsproxy was passed in.

A little awkward and should probably be reworked, but hopefully it fixes
the NFS oops.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Cc: Herbert Poetzl <herbert@13thfloor.at>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Daniel Hokka Zakrisson <daniel@hozac.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/nsproxy.h
kernel/exit.c
kernel/fork.c
kernel/nsproxy.c