merge to the -K3 scheduler.

arch/i386/kernel/entry.S

@@ -190,9 +190,11 @@ ENTRY(lcall27)
 
 
 ENTRY(ret_from_fork)
+#if CONFIG_SMP
 	pushl %ebx
 	call SYMBOL_NAME(schedule_tail)
 	addl $4, %esp
+#endif
 	GET_THREAD_INFO(%ebx)
 	jmp syscall_exit
 

fs/binfmt_elf.c

@@ -1119,7 +1119,7 @@ static int elf_core_dump(long signr, struct pt_regs * regs, struct file * file)
 	psinfo.pr_state = i;
 	psinfo.pr_sname = (i < 0 || i > 5) ? '.' : "RSDZTD"[i];
 	psinfo.pr_zomb = psinfo.pr_sname == 'Z';
-	psinfo.pr_nice = current->__nice;
+	psinfo.pr_nice = task_nice(current);
 	psinfo.pr_flag = current->flags;
 	psinfo.pr_uid = NEW_TO_OLD_UID(current->uid);
 	psinfo.pr_gid = NEW_TO_OLD_GID(current->gid);

fs/proc/array.c

@@ -336,12 +336,8 @@ int proc_pid_stat(struct task_struct *task, char * buffer)
 
 	/* scale priority and nice values from timeslices to -20..20 */
 	/* to make it look like a "normal" Unix priority/nice value  */
-	priority = task->prio;
-	if (priority >= MAX_RT_PRIO)
-		priority -= MAX_RT_PRIO;
-	else
-		priority = priority-100;
-	nice = task->__nice;
+	priority = task_prio(task);
+	nice = task_nice(task);
 
 	read_lock(&tasklist_lock);
 	ppid = task->pid ? task->p_opptr->pid : 0;

include/asm-i386/bitops.h

@@ -358,7 +358,7 @@ static __inline__ int find_next_zero_bit (void * addr, int size, int offset)
  * @offset: The bitnumber to start searching at
  * @size: The maximum size to search
  */
-static __inline__ int find_next_bit (void * addr, int size, int offset)
+static __inline__ int find_next_bit(void * addr, int size, int offset)
 {
 	unsigned long * p = ((unsigned long *) addr) + (offset >> 5);
 	int set = 0, bit = offset & 31, res;

include/asm-i386/mmu_context.h

@@ -8,14 +8,10 @@
 
 /*
  * Every architecture must define this function. It's the fastest
- * way of searching a 168-bit bitmap where the first 128 bits are
- * unlikely to be set. It's guaranteed that at least one of the 168
+ * way of searching a 140-bit bitmap where the first 100 bits are
+ * unlikely to be set. It's guaranteed that at least one of the 140
  * bits is cleared.
  */
-#if MAX_RT_PRIO != 128 || MAX_PRIO != 168
-# error update this function.
-#endif
-
 static inline int sched_find_first_bit(unsigned long *b)
 {
 	if (unlikely(b[0]))
@@ -24,11 +20,9 @@ static inline int sched_find_first_bit(unsigned long *b)
 		return __ffs(b[1]) + 32;
 	if (unlikely(b[2]))
 		return __ffs(b[2]) + 64;
-	if (unlikely(b[3]))
+	if (b[3])
 		return __ffs(b[3]) + 96;
-	if (b[4])
 	return __ffs(b[4]) + 128;
-	return __ffs(b[5]) + 32 + 128;
 }
 /*
  * possibly do the LDT unload here?

include/linux/init_task.h

@@ -45,7 +45,8 @@
     thread_info:	&init_thread_info,				\
     flags:		0,						\
     lock_depth:		-1,						\
-    __nice:		DEF_USER_NICE,					\
+    prio:		120,						\
+    static_prio:	120,						\
     policy:		SCHED_OTHER,					\
     cpus_allowed:	-1,						\
     mm:			NULL,						\

include/linux/sched.h

@@ -150,7 +150,7 @@ extern void trap_init(void);
 extern void update_process_times(int user);
 extern void update_one_process(struct task_struct *p, unsigned long user,
 			       unsigned long system, int cpu);
-extern void scheduler_tick(struct task_struct *p);
+extern void scheduler_tick(int user_tick, int system);
 extern void sched_task_migrated(struct task_struct *p);
 extern void smp_migrate_task(int cpu, task_t *task);
 extern unsigned long cache_decay_ticks;
@@ -241,18 +241,16 @@ struct task_struct {
 
 	int lock_depth;		/* Lock depth */
 
-	int prio;
-	long __nice;
+	int prio, static_prio;
 	list_t run_list;
 	prio_array_t *array;
 
-	unsigned int time_slice;
-
 	unsigned long sleep_avg;
 	unsigned long sleep_timestamp;
 
 	unsigned long policy;
 	unsigned long cpus_allowed;
+	unsigned int time_slice;
 
 	struct task_struct *next_task, *prev_task;
 
@@ -385,66 +383,12 @@ do { if (atomic_dec_and_test(&(tsk)->usage)) __put_task_struct(tsk); } while(0)
  */
 #define _STK_LIM	(8*1024*1024)
 
-/*
- * RT priorites go from 0 to 99, but internally we max
- * them out at 128 to make it easier to search the
- * scheduler bitmap.
- */
-#define MAX_RT_PRIO		128
-/*
- * The lower the priority of a process, the more likely it is
- * to run. Priority of a process goes from 0 to 167. The 0-99
- * priority range is allocated to RT tasks, the 128-167 range
- * is for SCHED_OTHER tasks.
- */
-#define MAX_PRIO		(MAX_RT_PRIO + 40)
-
-/*
- * Scales user-nice values [ -20 ... 0 ... 19 ]
- * to static priority [ 128 ... 167 (MAX_PRIO-1) ]
- *
- * User-nice value of -20 == static priority 128, and
- * user-nice value 19 == static priority 167. The lower
- * the priority value, the higher the task's priority.
- */
-#define NICE_TO_PRIO(n)		(MAX_RT_PRIO + (n) + 20)
-#define DEF_USER_NICE		0
-
-/*
- * Default timeslice is 150 msecs, maximum is 300 msecs.
- * Minimum timeslice is 10 msecs.
- *
- * These are the 'tuning knobs' of the scheduler:
- */
-#define MIN_TIMESLICE		( 10 * HZ / 1000)
-#define MAX_TIMESLICE		(300 * HZ / 1000)
-#define CHILD_FORK_PENALTY	95
-#define PARENT_FORK_PENALTY	100
-#define EXIT_WEIGHT		3
-#define PRIO_INTERACTIVE_RATIO	20
-#define PRIO_CPU_HOG_RATIO	60
-#define PRIO_BONUS_RATIO	70
-#define INTERACTIVE_DELTA	3
-#define MAX_SLEEP_AVG		(2*HZ)
-#define STARVATION_LIMIT	(2*HZ)
-
-#define USER_PRIO(p)		((p)-MAX_RT_PRIO)
-#define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))
-
-/*
- * NICE_TO_TIMESLICE scales nice values [ -20 ... 19 ]
- * to time slice values.
- *
- * The higher a process's priority, the bigger timeslices
- * it gets during one round of execution. But even the lowest
- * priority process gets MIN_TIMESLICE worth of execution time.
- */
-
-#define NICE_TO_TIMESLICE(n) (MIN_TIMESLICE + \
-	((MAX_TIMESLICE - MIN_TIMESLICE) * (19-(n))) / 39)
-
 extern void set_cpus_allowed(task_t *p, unsigned long new_mask);
 extern void set_user_nice(task_t *p, long nice);
+extern int task_prio(task_t *p);
+extern int task_nice(task_t *p);
+extern int idle_cpu(int cpu);
+
 asmlinkage long sys_sched_yield(void);
 #define yield() sys_sched_yield()
 
@@ -526,6 +470,7 @@ extern long FASTCALL(interruptible_sleep_on_timeout(wait_queue_head_t *q,
 						    signed long timeout));
 extern int FASTCALL(wake_up_process(struct task_struct * tsk));
 extern void FASTCALL(wake_up_forked_process(struct task_struct * tsk));
+extern void FASTCALL(sched_exit(task_t * p));
 
 #define wake_up(x)			__wake_up((x),TASK_UNINTERRUPTIBLE | TASK_INTERRUPTIBLE, 1)
 #define wake_up_nr(x, nr)		__wake_up((x),TASK_UNINTERRUPTIBLE | TASK_INTERRUPTIBLE, nr)

kernel/exit.c

@@ -31,8 +31,6 @@ int getrusage(struct task_struct *, int, struct rusage *);
 
 static void release_task(struct task_struct * p)
 {
-	unsigned long flags;
-
 	if (p == current)
 		BUG();
 #ifdef CONFIG_SMP
@@ -46,25 +44,7 @@ static void release_task(struct task_struct * p)
 	current->cmin_flt += p->min_flt + p->cmin_flt;
 	current->cmaj_flt += p->maj_flt + p->cmaj_flt;
 	current->cnswap += p->nswap + p->cnswap;
-	/*
-	 * Potentially available timeslices are retrieved
-	 * here - this way the parent does not get penalized
-	 * for creating too many processes.
-	 *
-	 * (this cannot be used to artificially 'generate'
-	 * timeslices, because any timeslice recovered here
-	 * was given away by the parent in the first place.)
-	 */
-	__save_flags(flags);
-	__cli();
-	current->time_slice += p->time_slice;
-	if (current->time_slice > MAX_TIMESLICE)
-		current->time_slice = MAX_TIMESLICE;
-	if (p->sleep_avg < current->sleep_avg)
-		current->sleep_avg = (current->sleep_avg * EXIT_WEIGHT +
-				p->sleep_avg) / (EXIT_WEIGHT + 1);
-	__restore_flags(flags);
-
+	sched_exit(p);
 	p->pid = 0;
 	put_task_struct(p);
 }
@@ -172,9 +152,8 @@ void reparent_to_init(void)
 	current->exit_signal = SIGCHLD;
 
 	current->ptrace = 0;
-	if ((current->policy == SCHED_OTHER) &&
-			(current->__nice < DEF_USER_NICE))
-		set_user_nice(current, DEF_USER_NICE);
+	if ((current->policy == SCHED_OTHER) && (task_nice(current) < 0))
+		set_user_nice(current, 0);
 	/* cpus_allowed? */
 	/* rt_priority? */
 	/* signals? */

kernel/fork.c

@@ -749,14 +749,11 @@ int do_fork(unsigned long clone_flags, unsigned long stack_start,
 		 * runqueue lock is not a problem.
 		 */
 		current->time_slice = 1;
-		scheduler_tick(current);
+		scheduler_tick(0, 0);
 	}
 	p->sleep_timestamp = jiffies;
 	__restore_flags(flags);
 
-	if (p->policy == SCHED_OTHER)
-		p->prio = MAX_PRIO - 1 - ((MAX_PRIO - 1 - p->prio) * 1) / 3;
-
 	/*
 	 * Ok, add it to the run-queues and make it
 	 * visible to the rest of the system.

kernel/ksyms.c

@@ -455,6 +455,8 @@ EXPORT_SYMBOL(preempt_schedule);
 EXPORT_SYMBOL(schedule_timeout);
 EXPORT_SYMBOL(sys_sched_yield);
 EXPORT_SYMBOL(set_user_nice);
+EXPORT_SYMBOL(task_nice);
+EXPORT_SYMBOL_GPL(idle_cpu);
 EXPORT_SYMBOL(jiffies);
 EXPORT_SYMBOL(xtime);
 EXPORT_SYMBOL(do_gettimeofday);

kernel/sys.c

@@ -221,7 +221,7 @@ asmlinkage long sys_setpriority(int which, int who, int niceval)
 		}
 		if (error == -ESRCH)
 			error = 0;
-		if (niceval < p->__nice && !capable(CAP_SYS_NICE))
+		if (niceval < task_nice(p) && !capable(CAP_SYS_NICE))
 			error = -EACCES;
 		else
 			set_user_nice(p, niceval);
@@ -250,7 +250,7 @@ asmlinkage long sys_getpriority(int which, int who)
 		long niceval;
 		if (!proc_sel(p, which, who))
 			continue;
-		niceval = 20 - p->__nice;
+		niceval = 20 - task_nice(p);
 		if (niceval > retval)
 			retval = niceval;
 	}

kernel/timer.c

@@ -584,17 +584,7 @@ void update_process_times(int user_tick)
 	int cpu = smp_processor_id(), system = user_tick ^ 1;
 
 	update_one_process(p, user_tick, system, cpu);
-	if (p->pid) {
-		if (p->__nice > 0)
-			kstat.per_cpu_nice[cpu] += user_tick;
-		else
-			kstat.per_cpu_user[cpu] += user_tick;
-		kstat.per_cpu_system[cpu] += system;
-	} else {
-		if (local_bh_count(cpu) || local_irq_count(cpu) > 1)
-			kstat.per_cpu_system[cpu] += system;
-	}
-	scheduler_tick(p);
+	scheduler_tick(user_tick, system);
 }
 
 /*

mm/oom_kill.c

@@ -82,7 +82,7 @@ static int badness(struct task_struct *p)
 	 * Niced processes are most likely less important, so double
 	 * their badness points.
 	 */
-	if (p->__nice > 0)
+	if (task_nice(p) > 0)
 		points *= 2;
 
 	/*
@@ -146,7 +146,7 @@ void oom_kill_task(struct task_struct *p)
 	 * all the memory it needs. That way it should be able to
 	 * exit() and clear out its resources quickly...
 	 */
-	p->time_slice = 2 * MAX_TIMESLICE;
+	p->time_slice = HZ;
 	p->flags |= PF_MEMALLOC | PF_MEMDIE;
 
 	/* This process has hardware access, be more careful. */