rcu: add comment saying why DEBUG_OBJECTS_RCU_HEAD depends on PREEMPT.
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Mon, 31 Jan 2011 07:56:46 +0000 (23:56 -0800)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Fri, 4 Mar 2011 16:05:41 +0000 (08:05 -0800)
The build will break if you change the Kconfig to allow
DEBUG_OBJECTS_RCU_HEAD and !PREEMPT, so document the reasoning
near where the breakage would occur.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
kernel/rcupdate.c

index afd21d1..f3240e9 100644 (file)
@@ -214,6 +214,11 @@ static int rcuhead_fixup_free(void *addr, enum debug_obj_state state)
                 * Ensure that queued callbacks are all executed.
                 * If we detect that we are nested in a RCU read-side critical
                 * section, we should simply fail, otherwise we would deadlock.
+                * Note that the machinery to reliably determine whether
+                * or not we are in an RCU read-side critical section
+                * exists only in the preemptible RCU implementations
+                * (TINY_PREEMPT_RCU and TREE_PREEMPT_RCU), which is why
+                * DEBUG_OBJECTS_RCU_HEAD is disallowed if !PREEMPT.
                 */
                if (rcu_preempt_depth() != 0 || preempt_count() != 0 ||
                    irqs_disabled()) {