Fix arc_adapt() spinning in iterate_supers_type()

[behlendorf]

The iterate_supers_type() function which was introduced in the
3.0 kernel was supposed to provide a safe way to call an arbitrary
function on all super blocks of a specific type. Unfortunately,
because a list_head was used a bug was introduced which made it
possible for iterate_supers_type() to get stuck spinning on a
super block which was just deactivated.

The bug was fixed in the 3.3 kernel by converting the list_head
to an hlist_node. However, to resolve the issue for existing
3.0 - 3.2 kernels we detect when a list_head is used. Then to
prevent the spinning from occurring the .next pointer is set to
the fs_supers list_head which ensures the iterate_supers_type()
function will always terminate.

Signed-off-by: Brian Behlendorf behlendorf1@llnl.gov
Issue #1045
Issue #861
Issue #790

[behlendorf]

@ryao @prakashsurya Can you please review this.

[prakashsurya]

I'm not trilled by this, but options are limited with kernel defects. Running with a debug kernel might cause issues since we're essentially corruption the sb->s_instances.next field, but that might just fall into the "don't do that" category for 3.0 - 3.2 kernels. Since the impact is minimal, and it targets only a select few kernels, I'm not opposed to fixing this issue this way. The only other solution I can think of, is to use the ugly hack we do for pre-3.0 kernels.

Ack.

[behlendorf]

dba1d70 Fix arc_adapt() spinning in iterate_supers_type()