genirq: Validate action before dereferencing it in handle_irq_event_percpu()
authorThomas Gleixner <tglx@linutronix.de>
Wed, 13 Jan 2016 13:07:25 +0000 (14:07 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 3 Mar 2016 23:07:11 +0000 (15:07 -0800)
commit 570540d50710ed192e98e2f7f74578c9486b6b05 upstream.

commit 71f64340fc0e changed the handling of irq_desc->action from

CPU 0                   CPU 1
free_irq()              lock(desc)
  lock(desc)            handle_edge_irq()
                        if (desc->action) {
                          handle_irq_event()
                            action = desc->action
                            unlock(desc)
  desc->action = NULL       handle_irq_event_percpu(desc, action)
                              action->xxx
to

CPU 0                   CPU 1
free_irq()              lock(desc)
  lock(desc)            handle_edge_irq()
                        if (desc->action) {
                          handle_irq_event()
                            unlock(desc)
  desc->action = NULL       handle_irq_event_percpu(desc, action)
                              action = desc->action
                              action->xxx

So if free_irq manages to set the action to NULL between the unlock and before
the readout, we happily dereference a null pointer.

We could simply revert 71f64340fc0e, but we want to preserve the better code
generation. A simple solution is to change the action loop from a do {} while
to a while {} loop.

This is safe because we either see a valid desc->action or NULL. If the action
is about to be removed it is still valid as free_irq() is blocked on
synchronize_irq().

CPU 0                   CPU 1
free_irq()              lock(desc)
  lock(desc)            handle_edge_irq()
                          handle_irq_event(desc)
                            set(INPROGRESS)
                            unlock(desc)
                            handle_irq_event_percpu(desc)
                            action = desc->action
  desc->action = NULL           while (action) {
                                  action->xxx
                                  ...
                                  action = action->next;
  sychronize_irq()
    while(INPROGRESS);      lock(desc)
                            clr(INPROGRESS)
free(action)

That's basically the same mechanism as we have for shared
interrupts. action->next can become NULL while handle_irq_event_percpu()
runs. Either it sees the action or NULL. It does not matter, because action
itself cannot go away before the interrupt in progress flag has been cleared.

Fixes: commit 71f64340fc0e "genirq: Remove the second parameter from handle_irq_event_percpu()"
Reported-by: zyjzyj2000@gmail.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Huang Shijie <shijie.huang@arm.com>
Cc: Jiang Liu <jiang.liu@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/alpine.DEB.2.11.1601131224190.3575@nanos
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
kernel/irq/handle.c

index a302cf9a2126c8a911ff4da1a944c88c6d3b3f8a..57bff7857e879bf2301a92723ade0a968870e637 100644 (file)
@@ -138,7 +138,8 @@ irqreturn_t handle_irq_event_percpu(struct irq_desc *desc)
        unsigned int flags = 0, irq = desc->irq_data.irq;
        struct irqaction *action = desc->action;
 
-       do {
+       /* action might have become NULL since we dropped the lock */
+       while (action) {
                irqreturn_t res;
 
                trace_irq_handler_entry(irq, action);
@@ -173,7 +174,7 @@ irqreturn_t handle_irq_event_percpu(struct irq_desc *desc)
 
                retval |= res;
                action = action->next;
-       } while (action);
+       }
 
        add_interrupt_randomness(irq, flags);