irq: only update affinity if ->set_affinity() is sucessfull

irq_set_affinity() and move_masked_irq() try to assign affinity
before calling chip set_affinity(). Some archs are assigning it
in ->set_affinity() again.

We do something like:

 cpumask_cpy(desc->affinity, mask);
 desc->chip->set_affinity(mask);

But in the failure path, affinity should not be touched - otherwise
we'll end up with a different affinity mask despite the failure to
migrate the IRQ.

So try to update the afffinity only if set_affinity returns with 0.
Also call irq_set_thread_affinity accordingly.

v2: update after "irq, x86: Remove IRQ_DISABLED check in process context IRQ move"
v3: according to Ingo, change set_affinity() in irq_chip should return int.
v4: update comments by removing moving irq_desc code.

[ Impact: fix /proc/irq/*/smp_affinity setting corner case bug ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
LKML-Reference: <49F65509.60307@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h
index 01ce20e..de5f412 100644 (file)
--- a/kernel/irq/internals.h
+++ b/kernel/irq/internals.h
@@ -42,6 +42,9 @@ static inline void unregister_handler_proc(unsigned int irq,
 
 extern int irq_select_affinity_usr(unsigned int irq);
 
+extern void
+irq_set_thread_affinity(struct irq_desc *desc, const struct cpumask *cpumask);
+
 /*
  * Debugging printout:
  */

diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 2734eca..aaf5c9d 100644 (file)
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -80,7 +80,7 @@ int irq_can_set_affinity(unsigned int irq)
        return 1;
 }
 
-static void
+void
 irq_set_thread_affinity(struct irq_desc *desc, const struct cpumask *cpumask)
 {
        struct irqaction *action = desc->action;
@@ -109,17 +109,22 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask)
        spin_lock_irqsave(&desc->lock, flags);
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
-       if (desc->status & IRQ_MOVE_PCNTXT)
-               desc->chip->set_affinity(irq, cpumask);
+       if (desc->status & IRQ_MOVE_PCNTXT) {
+               if (!desc->chip->set_affinity(irq, cpumask)) {
+                       cpumask_copy(desc->affinity, cpumask);
+                       irq_set_thread_affinity(desc, cpumask);
+               }
+       }
        else {
                desc->status |= IRQ_MOVE_PENDING;
                cpumask_copy(desc->pending_mask, cpumask);
        }
 #else
-       cpumask_copy(desc->affinity, cpumask);
-       desc->chip->set_affinity(irq, cpumask);
+       if (!desc->chip->set_affinity(irq, cpumask)) {
+               cpumask_copy(desc->affinity, cpumask);
+               irq_set_thread_affinity(desc, cpumask);
+       }
 #endif
-       irq_set_thread_affinity(desc, cpumask);
        desc->status |= IRQ_AFFINITY_SET;
        spin_unlock_irqrestore(&desc->lock, flags);
        return 0;

diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c
index e05ad9b..cfe767c 100644 (file)
--- a/kernel/irq/migration.c
+++ b/kernel/irq/migration.c
@@ -1,5 +1,8 @@
 
 #include <linux/irq.h>
+#include <linux/interrupt.h>
+
+#include "internals.h"
 
 void move_masked_irq(int irq)
 {
@@ -39,11 +42,12 @@ void move_masked_irq(int irq)
         * masking the irqs.
         */
        if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask)
-                  < nr_cpu_ids)) {
-               cpumask_and(desc->affinity,
-                           desc->pending_mask, cpu_online_mask);
-               desc->chip->set_affinity(irq, desc->affinity);
-       }
+                  < nr_cpu_ids))
+               if (!desc->chip->set_affinity(irq, desc->pending_mask)) {
+                       cpumask_copy(desc->affinity, desc->pending_mask);
+                       irq_set_thread_affinity(desc, desc->pending_mask);
+               }
+
        cpumask_clear(desc->pending_mask);
 }