lockdep: Change mark_held_locks() to check hlock->check instead of lockdep_no_validate

The __lockdep_no_validate check in mark_held_locks() adds the subtle
and (afaics) unnecessary difference between no-validate and check==0.
And this looks even more inconsistent because __lock_acquire() skips
mark_irqflags()->mark_lock() if !check.

Change mark_held_locks() to check hlock->check instead.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140120182013.GA26505@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
index f7eba92..bf0c6b0 100644 (file)
--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -2517,7 +2517,7 @@ mark_held_locks(struct task_struct *curr, enum mark_type mark)
 
                BUG_ON(usage_bit >= LOCK_USAGE_STATES);
 
-               if (hlock_class(hlock)->key == __lockdep_no_validate__.subkeys)
+               if (!hlock->check)
                        continue;
 
                if (!mark_lock(curr, hlock, usage_bit))