kernfs: fix get_active failure handling in kernfs_seq_*()
authorTejun Heo <tj@kernel.org>
Fri, 10 Jan 2014 13:57:18 +0000 (08:57 -0500)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 10 Jan 2014 21:44:25 +0000 (13:44 -0800)
commitd92d2e6bd72b653f9811e0c9c46307c743b3fc58
tree3d9f3207bf1debe9fef016aea5e9eb23af91d8d6
parent08da2012e0bb0f3f1422cce3f76c36a90da366b5
kernfs: fix get_active failure handling in kernfs_seq_*()

When kernfs_seq_start() fails to obtain an active reference, it
returns ERR_PTR(-ENODEV).  kernfs_seq_stop() is then invoked with the
error pointer value; however, it still proceeds to invoke
kernfs_put_active() on the node leading to unbalanced put.

If kernfs_seq_stop() is called even after active ref failure, it
should skip invocation of @ops->seq_stop() and put_active.
Unfortunately, this is a bit complicated because active ref failure
isn't the only thing which may fail with ERR_PTR(-ENODEV).
@ops->seq_start/next() may also fail with the error value and
kernfs_seq_stop() doesn't have a way to tell apart those failures.

Work it around by factoring out the active part of kernfs_seq_stop()
into kernfs_seq_stop_active() and invoking it directly if
@ops->seq_start/next() fail with ERR_PTR(-ENODEV) and updating
kernfs_seq_stop() to skip kernfs_seq_stop_active() on
ERR_PTR(-ENODEV).  This is a bit nasty but ensures that the active put
is skipped iff get_active failed in kernfs_seq_start().

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/kernfs/file.c