padata: Make two separate cpumasks

The aim of this patch is to make two separate cpumasks
for padata parallel and serial workers respectively.
It allows user to make more thin and sophisticated configurations
of padata framework. For example user may bind parallel and serial workers to non-intersecting
CPU groups to gain better performance. Also each padata instance has notifiers chain for its
cpumasks now. If either parallel or serial or both masks were changed all
interested subsystems will get notification about that. It's especially useful
if padata user uses algorithm for callback CPU selection according to serial cpumask.

Signed-off-by: Dan Kruchinin <dkruchinin@acm.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

1 files changed, 80 insertions, 36 deletions

diff --git a/include/linux/padata.h b/include/linux/padata.h
index 8844b85..621e773 100644
--- a/include/linux/padata.h
+++ b/include/linux/padata.h
@@ -25,6 +25,10 @@
 #include <linux/spinlock.h>
 #include <linux/list.h>
 #include <linux/timer.h>
+#include <linux/notifier.h>
+
+#define PADATA_CPU_SERIAL   0x01
+#define PADATA_CPU_PARALLEL 0x02
 
 /**
  * struct padata_priv -  Embedded to the users data structure.
@@ -59,7 +63,20 @@ struct padata_list {
 };
 
 /**
- * struct padata_queue - The percpu padata queues.
+* struct padata_serial_queue - The percpu padata serial queue
+*
+* @serial: List to wait for serialization after reordering.
+* @work: work struct for serialization.
+* @pd: Backpointer to the internal control structure.
+*/
+struct padata_serial_queue {
+       struct padata_list    serial;
+       struct work_struct    work;
+       struct parallel_data *pd;
+};
+
+/**
+ * struct padata_parallel_queue - The percpu padata parallel queue
  *
  * @parallel: List to wait for parallelization.
  * @reorder: List to wait for reordering after parallel processing.
@@ -67,44 +84,52 @@ struct padata_list {
  * @pwork: work struct for parallelization.
  * @swork: work struct for serialization.
  * @pd: Backpointer to the internal control structure.
+ * @work: work struct for parallelization.
+ * @num_obj: Number of objects that are processed by this cpu.
  * @cpu_index: Index of the cpu.
  */
-struct padata_queue {
-	struct padata_list	parallel;
-	struct padata_list	reorder;
-	struct padata_list	serial;
-	struct work_struct	pwork;
-	struct work_struct	swork;
-	struct parallel_data    *pd;
-	int			cpu_index;
+struct padata_parallel_queue {
+       struct padata_list    parallel;
+       struct padata_list    reorder;
+       struct parallel_data *pd;
+       struct work_struct    work;
+       atomic_t              num_obj;
+       int                   cpu_index;
 };
 
+
 /**
  * struct parallel_data - Internal control structure, covers everything
  * that depends on the cpumask in use.
  *
  * @pinst: padata instance.
- * @queue: percpu padata queues.
+ * @pqueue: percpu padata queues used for parallelization.
+ * @squeue: percpu padata queues used for serialuzation.
  * @seq_nr: The sequence number that will be attached to the next object.
  * @reorder_objects: Number of objects waiting in the reorder queues.
  * @refcnt: Number of objects holding a reference on this parallel_data.
  * @max_seq_nr:  Maximal used sequence number.
- * @cpumask: cpumask in use.
+ * @cpumask: Contains two cpumasks: pcpu and cbcpu for
+ *           parallel and serial workers respectively.
  * @lock: Reorder lock.
  * @processed: Number of already processed objects.
  * @timer: Reorder timer.
  */
 struct parallel_data {
-	struct padata_instance	*pinst;
-	struct padata_queue	*queue;
-	atomic_t		seq_nr;
-	atomic_t		reorder_objects;
-	atomic_t                refcnt;
-	unsigned int		max_seq_nr;
-	cpumask_var_t		cpumask;
-	spinlock_t              lock ____cacheline_aligned;
-	unsigned int            processed;
-	struct timer_list       timer;
+	struct padata_instance		*pinst;
+	struct padata_parallel_queue	*pqueue;
+	struct padata_serial_queue	*squeue;
+	atomic_t			 seq_nr;
+	atomic_t			 reorder_objects;
+	atomic_t			 refcnt;
+	unsigned int			 max_seq_nr;
+	struct {
+		cpumask_var_t		 pcpu;
+		cpumask_var_t		 cbcpu;
+	} cpumask;
+	spinlock_t                       lock ____cacheline_aligned;
+	unsigned int			 processed;
+	struct timer_list		 timer;
 };
 
 /**
@@ -113,32 +138,51 @@ struct parallel_data {
  * @cpu_notifier: cpu hotplug notifier.
  * @wq: The workqueue in use.
  * @pd: The internal control structure.
- * @cpumask: User supplied cpumask.
+ * @cpumask: User supplied cpumask. Contains two cpumasks: pcpu and
+ *           cbcpu for parallel and serial works respectivly.
+ * @cpumask_change_notifier: Notifiers chain for user-defined notify
+ *            callbacks that will be called when either @pcpu or @cbcpu
+ *             or both cpumasks change.
  * @lock: padata instance lock.
  * @flags: padata flags.
  */
 struct padata_instance {
-	struct notifier_block   cpu_notifier;
-	struct workqueue_struct *wq;
-	struct parallel_data	*pd;
-	cpumask_var_t           cpumask;
-	struct mutex		lock;
-	u8			flags;
-#define	PADATA_INIT		1
-#define	PADATA_RESET		2
-#define	PADATA_INVALID		4
+	struct notifier_block		 cpu_notifier;
+	struct workqueue_struct		*wq;
+	struct parallel_data		*pd;
+	struct {
+		cpumask_var_t		 pcpu;
+		cpumask_var_t		 cbcpu;
+	} cpumask;
+	struct blocking_notifier_head	 cpumask_change_notifier;
+	struct mutex			 lock;
+	u8				 flags;
+#define	PADATA_INIT	1
+#define	PADATA_RESET	2
+#define	PADATA_INVALID	4
 };
 
-extern struct padata_instance *padata_alloc(const struct cpumask *cpumask,
-					    struct workqueue_struct *wq);
+extern struct padata_instance *padata_alloc(struct workqueue_struct *wq);
+extern struct padata_instance *__padata_alloc(struct workqueue_struct *wq,
+					      const struct cpumask *pcpumask,
+					      const struct cpumask *cbcpumask);
 extern void padata_free(struct padata_instance *pinst);
 extern int padata_do_parallel(struct padata_instance *pinst,
 			      struct padata_priv *padata, int cb_cpu);
 extern void padata_do_serial(struct padata_priv *padata);
-extern int padata_set_cpumask(struct padata_instance *pinst,
+extern int padata_get_cpumask(struct padata_instance *pinst,
+			      int cpumask_type, struct cpumask *out_mask);
+extern int padata_set_cpumask(struct padata_instance *pinst, int cpumask_type,
 			      cpumask_var_t cpumask);
-extern int padata_add_cpu(struct padata_instance *pinst, int cpu);
-extern int padata_remove_cpu(struct padata_instance *pinst, int cpu);
+extern int __padata_set_cpumasks(struct padata_instance *pinst,
+				 cpumask_var_t pcpumask,
+				 cpumask_var_t cbcpumask);
+extern int padata_add_cpu(struct padata_instance *pinst, int cpu, int mask);
+extern int padata_remove_cpu(struct padata_instance *pinst, int cpu, int mask);
 extern int padata_start(struct padata_instance *pinst);
 extern void padata_stop(struct padata_instance *pinst);
+extern int padata_register_cpumask_notifier(struct padata_instance *pinst,
+					    struct notifier_block *nblock);
+extern int padata_unregister_cpumask_notifier(struct padata_instance *pinst,
+					      struct notifier_block *nblock);
 #endif