[PATCH] per-task-delay-accounting: delay accounting usage of taskstats interface

Usage of taskstats interface by delay accounting.
Signed-off-by: Shailabh Nagar <nagar@us.ibm.com>
Signed-off-by: Balbir Singh <balbir@in.ibm.com>
Cc: Jes Sorensen <jes@sgi.com>
Cc: Peter Chubb <peterc@gelato.unsw.edu.au>
Cc: Erich Focht <efocht@ess.nec.de>
Cc: Levent Serinol <lserinol@gmail.com>
Cc: Jay Lan <jlan@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

include/linux/delayacct.h

@@ -18,6 +18,7 @@
 #define _LINUX_DELAYACCT_H
 
 #include <linux/sched.h>
+#include <linux/taskstats_kern.h>
 
 /*
  * Per-task flags relevant to delay accounting
@@ -35,6 +36,7 @@ extern void __delayacct_tsk_init(struct task_struct *);
 extern void __delayacct_tsk_exit(struct task_struct *);
 extern void __delayacct_blkio_start(void);
 extern void __delayacct_blkio_end(void);
+extern int __delayacct_add_tsk(struct taskstats *, struct task_struct *);
 
 static inline void delayacct_set_flag(int flag)
 {
@@ -74,6 +76,16 @@ static inline void delayacct_blkio_end(void)
 		__delayacct_blkio_end();
 }
 
+static inline int delayacct_add_tsk(struct taskstats *d,
+					struct task_struct *tsk)
+{
+	if (likely(!delayacct_on))
+		return -EINVAL;
+	if (!tsk->delays)
+		return 0;
+	return __delayacct_add_tsk(d, tsk);
+}
+
 #else
 static inline void delayacct_set_flag(int flag)
 {}
@@ -89,6 +101,9 @@ static inline void delayacct_blkio_start(void)
 {}
 static inline void delayacct_blkio_end(void)
 {}
+static inline int delayacct_add_tsk(struct taskstats *d,
+					struct task_struct *tsk)
+{ return 0; }
 #endif /* CONFIG_TASK_DELAY_ACCT */
 
 #endif

include/linux/sched.h

@@ -990,6 +990,7 @@ struct task_struct {
 	 */
 	struct pipe_inode_info *splice_pipe;
 #ifdef	CONFIG_TASK_DELAY_ACCT
+	spinlock_t delays_lock;
 	struct task_delay_info *delays;
 #endif
 };

include/linux/taskstats.h

@@ -34,7 +34,60 @@
 struct taskstats {
 
 	/* Version 1 */
-	__u64	version;
+	__u16	version;
+	__u16	padding[3];	/* Userspace should not interpret the padding
+				 * field which can be replaced by useful
+				 * fields if struct taskstats is extended.
+				 */
+
+	/* Delay accounting fields start
+	 *
+	 * All values, until comment "Delay accounting fields end" are
+	 * available only if delay accounting is enabled, even though the last
+	 * few fields are not delays
+	 *
+	 * xxx_count is the number of delay values recorded
+	 * xxx_delay_total is the corresponding cumulative delay in nanoseconds
+	 *
+	 * xxx_delay_total wraps around to zero on overflow
+	 * xxx_count incremented regardless of overflow
+	 */
+
+	/* Delay waiting for cpu, while runnable
+	 * count, delay_total NOT updated atomically
+	 */
+	__u64	cpu_count;
+	__u64	cpu_delay_total;
+
+	/* Following four fields atomically updated using task->delays->lock */
+
+	/* Delay waiting for synchronous block I/O to complete
+	 * does not account for delays in I/O submission
+	 */
+	__u64	blkio_count;
+	__u64	blkio_delay_total;
+
+	/* Delay waiting for page fault I/O (swap in only) */
+	__u64	swapin_count;
+	__u64	swapin_delay_total;
+
+	/* cpu "wall-clock" running time
+	 * On some architectures, value will adjust for cpu time stolen
+	 * from the kernel in involuntary waits due to virtualization.
+	 * Value is cumulative, in nanoseconds, without a corresponding count
+	 * and wraps around to zero silently on overflow
+	 */
+	__u64	cpu_run_real_total;
+
+	/* cpu "virtual" running time
+	 * Uses time intervals seen by the kernel i.e. no adjustment
+	 * for kernel's involuntary waits due to virtualization.
+	 * Value is cumulative, in nanoseconds, without a corresponding count
+	 * and wraps around to zero silently on overflow
+	 */
+	__u64	cpu_run_virtual_total;
+	/* Delay accounting fields end */
+	/* version 1 ends here */
 };
 
 

include/linux/taskstats_kern.h

@@ -17,6 +17,7 @@ enum {
 
 #ifdef CONFIG_TASKSTATS
 extern kmem_cache_t *taskstats_cache;
+extern struct mutex taskstats_exit_mutex;
 
 static inline void taskstats_exit_alloc(struct taskstats **ptidstats,
 					struct taskstats **ptgidstats)

init/Kconfig

@@ -173,6 +173,7 @@ config TASKSTATS
 
 config TASK_DELAY_ACCT
 	bool "Enable per-task delay accounting (EXPERIMENTAL)"
+	depends on TASKSTATS
 	help
 	  Collect information on time spent by a task waiting for system
 	  resources like cpu, synchronous block I/O completion and swapping

kernel/delayacct.c

@@ -41,6 +41,10 @@ void delayacct_init(void)
 
 void __delayacct_tsk_init(struct task_struct *tsk)
 {
+	spin_lock_init(&tsk->delays_lock);
+	/* No need to acquire tsk->delays_lock for allocation here unless
+	   __delayacct_tsk_init called after tsk is attached to tasklist
+	*/
 	tsk->delays = kmem_cache_zalloc(delayacct_cache, SLAB_KERNEL);
 	if (tsk->delays)
 		spin_lock_init(&tsk->delays->lock);
@@ -48,8 +52,11 @@ void __delayacct_tsk_init(struct task_struct *tsk)
 
 void __delayacct_tsk_exit(struct task_struct *tsk)
 {
-	kmem_cache_free(delayacct_cache, tsk->delays);
+	struct task_delay_info *delays = tsk->delays;
+	spin_lock(&tsk->delays_lock);
 	tsk->delays = NULL;
+	spin_unlock(&tsk->delays_lock);
+	kmem_cache_free(delayacct_cache, delays);
 }
 
 /*
@@ -104,3 +111,56 @@ void __delayacct_blkio_end(void)
 			&current->delays->blkio_delay,
 			&current->delays->blkio_count);
 }
+
+int __delayacct_add_tsk(struct taskstats *d, struct task_struct *tsk)
+{
+	s64 tmp;
+	struct timespec ts;
+	unsigned long t1,t2,t3;
+
+	spin_lock(&tsk->delays_lock);
+
+	/* Though tsk->delays accessed later, early exit avoids
+	 * unnecessary returning of other data
+	 */
+	if (!tsk->delays)
+		goto done;
+
+	tmp = (s64)d->cpu_run_real_total;
+	cputime_to_timespec(tsk->utime + tsk->stime, &ts);
+	tmp += timespec_to_ns(&ts);
+	d->cpu_run_real_total = (tmp < (s64)d->cpu_run_real_total) ? 0 : tmp;
+
+	/*
+	 * No locking available for sched_info (and too expensive to add one)
+	 * Mitigate by taking snapshot of values
+	 */
+	t1 = tsk->sched_info.pcnt;
+	t2 = tsk->sched_info.run_delay;
+	t3 = tsk->sched_info.cpu_time;
+
+	d->cpu_count += t1;
+
+	jiffies_to_timespec(t2, &ts);
+	tmp = (s64)d->cpu_delay_total + timespec_to_ns(&ts);
+	d->cpu_delay_total = (tmp < (s64)d->cpu_delay_total) ? 0 : tmp;
+
+	tmp = (s64)d->cpu_run_virtual_total + (s64)jiffies_to_usecs(t3) * 1000;
+	d->cpu_run_virtual_total =
+		(tmp < (s64)d->cpu_run_virtual_total) ?	0 : tmp;
+
+	/* zero XXX_total, non-zero XXX_count implies XXX stat overflowed */
+
+	spin_lock(&tsk->delays->lock);
+	tmp = d->blkio_delay_total + tsk->delays->blkio_delay;
+	d->blkio_delay_total = (tmp < d->blkio_delay_total) ? 0 : tmp;
+	tmp = d->swapin_delay_total + tsk->delays->swapin_delay;
+	d->swapin_delay_total = (tmp < d->swapin_delay_total) ? 0 : tmp;
+	d->blkio_count += tsk->delays->blkio_count;
+	d->swapin_count += tsk->delays->swapin_count;
+	spin_unlock(&tsk->delays->lock);
+
+done:
+	spin_unlock(&tsk->delays_lock);
+	return 0;
+}

kernel/taskstats.c

@@ -18,13 +18,13 @@
 
 #include <linux/kernel.h>
 #include <linux/taskstats_kern.h>
+#include <linux/delayacct.h>
 #include <net/genetlink.h>
 #include <asm/atomic.h>
 
 static DEFINE_PER_CPU(__u32, taskstats_seqnum) = { 0 };
 static int family_registered;
 kmem_cache_t *taskstats_cache;
-static DEFINE_MUTEX(taskstats_exit_mutex);
 
 static struct genl_family family = {
 	.id		= GENL_ID_GENERATE,
@@ -120,7 +120,10 @@ static int fill_pid(pid_t pid, struct task_struct *pidtsk,
 	 *		goto err;
 	 */
 
-err:
+	rc = delayacct_add_tsk(stats, tsk);
+	stats->version = TASKSTATS_VERSION;
+
+	/* Define err: label here if needed */
 	put_task_struct(tsk);
 	return rc;
 
@@ -152,8 +155,14 @@ static int fill_tgid(pid_t tgid, struct task_struct *tgidtsk,
 		 *		break;
 		 */
 
+		rc = delayacct_add_tsk(stats, tsk);
+		if (rc)
+			break;
+
 	} while_each_thread(first, tsk);
 	read_unlock(&tasklist_lock);
+	stats->version = TASKSTATS_VERSION;
+
 
 	/*
 	 * Accounting subsytems can also add calls here if they don't
@@ -233,8 +242,6 @@ void taskstats_exit_send(struct task_struct *tsk, struct taskstats *tidstats,
 	if (!family_registered || !tidstats)
 		return;
 
-	mutex_lock(&taskstats_exit_mutex);
-
 	is_thread_group = !thread_group_empty(tsk);
 	rc = 0;
 
@@ -292,7 +299,6 @@ nla_put_failure:
 err_skb:
 	nlmsg_free(rep_skb);
 ret:
-	mutex_unlock(&taskstats_exit_mutex);
 	return;
 }