diff --git a/arch/arm/src/armv7-a/arm_assert.c b/arch/arm/src/armv7-a/arm_assert.c
index 3c798d33a0d..c6742d2a6a5 100644
--- a/arch/arm/src/armv7-a/arm_assert.c
+++ b/arch/arm/src/armv7-a/arm_assert.c
@@ -53,6 +53,7 @@
 
 #include "up_arch.h"
 #include "sched/sched.h"
+#include "irq/irq.h"
 #include "up_internal.h"
 
 /****************************************************************************
@@ -346,6 +347,10 @@ static void _up_assert(int errorcode)
   if (CURRENT_REGS || this_task()->pid == 0)
     {
       (void)up_irq_save();
+#ifdef SMP
+      (void)spin_trylock(&g_cpu_irqlock);
+#endif
+
       for (; ; )
         {
 #ifdef CONFIG_ARCH_LEDS
diff --git a/arch/arm/src/armv7-a/arm_sigdeliver.c b/arch/arm/src/armv7-a/arm_sigdeliver.c
index 672648a63a5..5d895832820 100644
--- a/arch/arm/src/armv7-a/arm_sigdeliver.c
+++ b/arch/arm/src/armv7-a/arm_sigdeliver.c
@@ -117,6 +117,9 @@ void up_sigdeliver(void)
    * on the local CPU.  We do not want to call enter_critical_section()
    * here, however, because we don't want this state to stick after the
    * call to up_fullcontextrestore().
+   *
+   * I would prefer that all interrupts are disabled when
+   * up_fullcontextrestore() is called, but that may not be necessary.
    */
 
   sinfo("Resuming\n");
diff --git a/arch/arm/src/common/up_exit.c b/arch/arm/src/common/up_exit.c
index 4d16f2a8a55..ec5a88498be 100644
--- a/arch/arm/src/common/up_exit.c
+++ b/arch/arm/src/common/up_exit.c
@@ -50,6 +50,7 @@
 #include "task/task.h"
 #include "sched/sched.h"
 #include "group/group.h"
+#include "irq/irq.h"
 #include "up_internal.h"
 
 /****************************************************************************
@@ -140,11 +141,14 @@ void _exit(int status)
 {
   struct tcb_s *tcb;
 
-  /* Disable interrupts.  They will be restored when the next
-   * task is started.
+  /* Disable interrupts.  They will be restored when the next task is
+   * started.
    */
 
   (void)up_irq_save();
+#ifdef SMP
+  (void)spin_trylock(&g_cpu_irqlock);
+#endif
 
   sinfo("TCB=%p exiting\n", this_task());
 
@@ -177,4 +181,3 @@ void _exit(int status)
 
   up_fullcontextrestore(tcb->xcp.regs);
 }
-
diff --git a/configs/sabre-6quad/README.txt b/configs/sabre-6quad/README.txt
index ec41ee62719..c7d7fc57c7d 100644
--- a/configs/sabre-6quad/README.txt
+++ b/configs/sabre-6quad/README.txt
@@ -555,11 +555,6 @@ Open Issues:
    Update: Cache inconsistencies seem to be the root cause of all current SMP
    issues.
 
-5. Assertions.  On a fatal assertions, other CPUs need to be stopped.  The SCR,
-   however, only supports disabling CPUs 1 through 3.  Perhaps if the assertion
-   occurs on CPUn, n > 0, then it should use and SGI to perform the assertion
-   on CPU0 always.  From CPU0, CPU1-3 can be disabled.
-
 Configurations
 ==============
 
diff --git a/include/nuttx/spinlock.h b/include/nuttx/spinlock.h
index 1ad7fef5253..ee6bf65b648 100644
--- a/include/nuttx/spinlock.h
+++ b/include/nuttx/spinlock.h
@@ -179,6 +179,27 @@ void spin_initializer(FAR struct spinlock_s *lock);
 
 void spin_lock(FAR volatile spinlock_t *lock);
 
+/****************************************************************************
+ * Name: spin_trylock
+ *
+ * Description:
+ *   Try once to lock the spinlock.  Do not wait if the spinlock is already
+ *   locked.
+ *
+ * Input Parameters:
+ *   lock - A reference to the spinlock object to lock.
+ *
+ * Returned Value:
+ *   SP_LOCKED   - Failure, the spinlock was already locked
+ *   SP_UNLOCKED - Success, the spinlock was successfully locked
+ *
+ * Assumptions:
+ *   Not running at the interrupt level.
+ *
+ ****************************************************************************/
+
+#define spin_trylock(l) up_testset(l)
+
 /****************************************************************************
  * Name: spin_lockr
  *