On (10/08/12 12:49), Paul E. McKenney wrote: > > ------------------------------------------------------------------------ > > device_cgroup: Restore rcu_read_lock() protection to devcgroup_inode_mknod() > > Commit ad676077 (device_cgroup: convert device_cgroup internally to > policy + exceptions) restructured devcgroup_inode_mknod(), removing > rcu_read_lock() in the process. However, RCU read-side protection > is required by the call to task_devcgroup(), so this commit restores > the rcu_read_lock() and rcu_read_unlock(). > > Signed-off-by: Paul E. McKenney <paul...@linux.vnet.ibm.com> > > diff --git a/security/device_cgroup.c b/security/device_cgroup.c > index 44dfc41..c686110 100644 > --- a/security/device_cgroup.c > +++ b/security/device_cgroup.c > @@ -576,9 +576,12 @@ int __devcgroup_inode_permission(struct inode *inode, > int mask) > > int devcgroup_inode_mknod(int mode, dev_t dev) > { > - struct dev_cgroup *dev_cgroup = task_devcgroup(current); > + struct dev_cgroup *dev_cgroup; > + int ret; > short type; > > + rcu_read_lock(); > + dev_cgroup = task_devcgroup(current); > if (!S_ISBLK(mode) && !S_ISCHR(mode)) > return 0; > > @@ -587,7 +590,9 @@ int devcgroup_inode_mknod(int mode, dev_t dev) > else > type = DEV_CHAR; > > - return __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev), > + ret = __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev), > MINOR(dev), ACC_MKNOD); > + rcu_read_unlock(); > + return ret; > > } >
I believe the same should be done for __devcgroup_inode_permission() as well. And we probably can call task_devcgroup() and rcu_read_lock() after "S_ISBLK(mode) && !S_ISCHR(mode)" checks (I guess we also need to unlock RCU on `return 0'). ---------------------------------------------------------------------------- Commit ad676077 | Author: Aristeu Rozanski <a...@redhat.com> | Date: Thu Oct 4 17:15:17 2012 -0700 | device_cgroup: convert device_cgroup internally to policy + exceptions moved RCU read-side protection from devcgroup_inode_mknod(), which, however is required by task_devcgroup(). Patch also add RCU read-side protection to __devcgroup_inode_permission() function, introduced in commit ad676077. [ 0.946303] include/linux/cgroup.h:566 suspicious rcu_dereference_check() usage! [ 0.946511] [ 0.946606] 2 locks held by kdevtmpfs/28: [ 0.946684] #0: (sb_writers){.+.+.+}, at: [<ffffffff81144bcb>] mnt_want_write+0x24/0x4b [ 0.947083] #1: (&sb->s_type->i_mutex_key#3/1){+.+.+.}, at: [<ffffffff81133d04>] kern_path_create+0x83/0x144 [ 0.947598] [ 0.947787] Call Trace: [ 0.947868] [<ffffffff81089644>] lockdep_rcu_suspicious+0x109/0x112 [ 0.947958] [<ffffffff81258fa0>] devcgroup_inode_mknod+0x9e/0xee [ 0.948043] [<ffffffff81132ee7>] vfs_mknod+0x8a/0xed [ 0.948129] [<ffffffff813b98af>] handle_create.isra.2+0x144/0x1b5 [ 0.948214] [<ffffffff813b99bf>] ? devtmpfsd+0x9f/0x138 [ 0.948298] [<ffffffff81295d5c>] ? do_raw_spin_lock+0x67/0xde [ 0.948384] [<ffffffff81295e92>] ? do_raw_spin_unlock+0x8f/0x98 [ 0.948469] [<ffffffff813b9920>] ? handle_create.isra.2+0x1b5/0x1b5 [ 0.948554] [<ffffffff813b9a04>] devtmpfsd+0xe4/0x138 [ 0.948638] [<ffffffff813b9920>] ? handle_create.isra.2+0x1b5/0x1b5 [ 0.948724] [<ffffffff810582b6>] kthread+0xd5/0xdd [ 0.948814] [<ffffffff814db664>] kernel_thread_helper+0x4/0x10 [ 0.948900] [<ffffffff814d2973>] ? retint_restore_args+0x13/0x13 [ 0.948985] [<ffffffff810581e1>] ? __init_kthread_worker+0x5a/0x5a [ 0.949069] [<ffffffff814db660>] ? gs_change+0x13/0x13 devcgroup_inode_mknod() part submitted by Paul E. McKenney. Signed-off-by: Sergey Senozhatsky <sergey.senozhat...@gmail.com> --- security/device_cgroup.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) diff --git a/security/device_cgroup.c b/security/device_cgroup.c index 44dfc41..043eb00 100644 --- a/security/device_cgroup.c +++ b/security/device_cgroup.c @@ -558,7 +558,8 @@ static int __devcgroup_check_permission(struct dev_cgroup *dev_cgroup, int __devcgroup_inode_permission(struct inode *inode, int mask) { - struct dev_cgroup *dev_cgroup = task_devcgroup(current); + struct dev_cgroup *dev_cgroup; + int ret; short type, access = 0; if (S_ISBLK(inode->i_mode)) @@ -570,13 +571,20 @@ int __devcgroup_inode_permission(struct inode *inode, int mask) if (mask & MAY_READ) access |= ACC_READ; - return __devcgroup_check_permission(dev_cgroup, type, imajor(inode), + rcu_read_lock(); + + dev_cgroup = task_devcgroup(current); + ret = __devcgroup_check_permission(dev_cgroup, type, imajor(inode), iminor(inode), access); + + rcu_read_unlock(); + return ret; } int devcgroup_inode_mknod(int mode, dev_t dev) { - struct dev_cgroup *dev_cgroup = task_devcgroup(current); + struct dev_cgroup *dev_cgroup; + int ret; short type; if (!S_ISBLK(mode) && !S_ISCHR(mode)) @@ -587,7 +595,12 @@ int devcgroup_inode_mknod(int mode, dev_t dev) else type = DEV_CHAR; - return __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev), + rcu_read_lock(); + + dev_cgroup = task_devcgroup(current); + ret = __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev), MINOR(dev), ACC_MKNOD); + rcu_read_unlock(); + return ret; } -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/