[PATCH] sched: disable preempt in idle tasks

Run idle threads with preempt disabled.

Also corrected a bugs in arm26's cpu_idle (make it actually call schedule()).
How did it ever work before?

Might fix the CPU hotplugging hang which Nigel Cunningham noted.

We think the bug hits if the idle thread is preempted after checking
need_resched() and before going to sleep, then the CPU offlined.

After calling stop_machine_run, the CPU eventually returns from preemption and
into the idle thread and goes to sleep.  The CPU will continue executing
previous idle and have no chance to call play_dead.

By disabling preemption until we are ready to explicitly schedule, this bug is
fixed and the idle threads generally become more robust.

From: alexs <ashepard@u.washington.edu>

  PPC build fix

From: Yoichi Yuasa <yuasa@hh.iij4u.or.jp>

  MIPS build fix

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

arch/arm/kernel/process.c

@@ -116,13 +116,13 @@ void cpu_idle(void)
 
 		if (!idle)
 			idle = default_idle;
-		preempt_disable();
 		leds_event(led_idle_start);
 		while (!need_resched())
 			idle();
 		leds_event(led_idle_end);
-		preempt_enable();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/arm/kernel/smp.c

@@ -256,7 +256,9 @@ void __cpuexit cpu_die(void)
 asmlinkage void __cpuinit secondary_start_kernel(void)
 {
 	struct mm_struct *mm = &init_mm;
-	unsigned int cpu = smp_processor_id();
+	unsigned int cpu;
+
+	cpu = smp_processor_id();
 
 	printk("CPU%u: Booted secondary processor\n", cpu);
 
@@ -273,6 +275,7 @@ asmlinkage void __cpuinit secondary_start_kernel(void)
 	local_flush_tlb_all();
 
 	cpu_init();
+	preempt_disable();
 
 	/*
 	 * Give the platform a chance to do its own initialisation.

arch/arm26/kernel/process.c

@@ -74,15 +74,13 @@ __setup("hlt", hlt_setup);
 void cpu_idle(void)
 {
 	/* endless idle loop with no priority at all */
-	preempt_disable();
 	while (1) {
-		while (!need_resched()) {
-			local_irq_disable();
-			if (!need_resched() && !hlt_counter)
-				local_irq_enable();
-		}
-	}
+		while (!need_resched())
+			cpu_relax();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
+	}
 }
 
 static char reboot_mode = 'h';

arch/cris/arch-v32/kernel/smp.c

@@ -161,6 +161,7 @@ void __init smp_callin(void)
 	REG_WR(intr_vect, irq_regs[cpu], rw_mask, vect_mask);
 	unmask_irq(IPI_INTR_VECT);
 	unmask_irq(TIMER_INTR_VECT);
+	preempt_disable();
 	local_irq_enable();
 
 	cpu_set(cpu, cpu_online_map);

arch/cris/kernel/process.c

@@ -218,7 +218,9 @@ void cpu_idle (void)
 				idle = default_idle;
 			idle();
 		}
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/frv/kernel/process.c

@@ -77,16 +77,20 @@ void (*idle)(void) = core_sleep_idle;
  */
 void cpu_idle(void)
 {
+	int cpu = smp_processor_id();
+
 	/* endless idle loop with no priority at all */
 	while (1) {
 		while (!need_resched()) {
-			irq_stat[smp_processor_id()].idle_timestamp = jiffies;
+			irq_stat[cpu].idle_timestamp = jiffies;
 
 			if (!frv_dma_inprogress && idle)
 				idle();
 		}
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/h8300/kernel/process.c

@@ -53,22 +53,18 @@ asmlinkage void ret_from_fork(void);
 #if !defined(CONFIG_H8300H_SIM) && !defined(CONFIG_H8S_SIM)
 void default_idle(void)
 {
-	while(1) {
+	local_irq_disable();
 	if (!need_resched()) {
 		local_irq_enable();
+		/* XXX: race here! What if need_resched() gets set now? */
 		__asm__("sleep");
-			local_irq_disable();
-		}
-		schedule();
-	}
+	} else
+		local_irq_enable();
 }
 #else
 void default_idle(void)
 {
-	while(1) {
-		if (need_resched())
-			schedule();
-	}
+	cpu_relax();
 }
 #endif
 void (*idle)(void) = default_idle;
@@ -81,7 +77,13 @@ void (*idle)(void) = default_idle;
  */
 void cpu_idle(void)
 {
+	while (1) {
+		while (!need_resched())
 			idle();
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
 }
 
 void machine_restart(char * __unused)

arch/i386/kernel/process.c

@@ -179,7 +179,7 @@ static inline void play_dead(void)
  */
 void cpu_idle(void)
 {
-	int cpu = raw_smp_processor_id();
+	int cpu = smp_processor_id();
 
 	/* endless idle loop with no priority at all */
 	while (1) {
@@ -201,7 +201,9 @@ void cpu_idle(void)
 			__get_cpu_var(irq_stat).idle_timestamp = jiffies;
 			idle();
 		}
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/i386/kernel/smpboot.c

@@ -485,6 +485,7 @@ static void __devinit start_secondary(void *unused)
 	 * things done here to the most necessary things.
 	 */
 	cpu_init();
+	preempt_disable();
 	smp_callin();
 	while (!cpu_isset(smp_processor_id(), smp_commenced_mask))
 		rep_nop();

arch/ia64/kernel/process.c

@@ -292,7 +292,9 @@ cpu_idle (void)
 #ifdef CONFIG_SMP
 		normal_xtp();
 #endif
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 		check_pgt_cache();
 		if (cpu_is_offline(smp_processor_id()))
 			play_dead();

arch/ia64/kernel/smpboot.c

@@ -399,6 +399,7 @@ start_secondary (void *unused)
 	Dprintk("start_secondary: starting CPU 0x%x\n", hard_smp_processor_id());
 	efi_map_pal_code();
 	cpu_init();
+	preempt_disable();
 	smp_callin();
 
 	cpu_idle();

arch/m32r/kernel/process.c

@@ -104,7 +104,9 @@ void cpu_idle (void)
 
 			idle();
 		}
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/m32r/kernel/smpboot.c

@@ -426,6 +426,7 @@ void __init smp_cpus_done(unsigned int max_cpus)
 int __init start_secondary(void *unused)
 {
 	cpu_init();
+	preempt_disable();
 	smp_callin();
 	while (!cpu_isset(smp_processor_id(), smp_commenced_mask))
 		cpu_relax();

arch/m68k/kernel/process.c

@@ -102,7 +102,9 @@ void cpu_idle(void)
 	while (1) {
 		while (!need_resched())
 			idle();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/mips/kernel/process.c

@@ -52,7 +52,9 @@ ATTRIB_NORET void cpu_idle(void)
 		while (!need_resched())
 			if (cpu_wait)
 				(*cpu_wait)();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/mips/kernel/smp.c

@@ -82,7 +82,7 @@ extern ATTRIB_NORET void cpu_idle(void);
  */
 asmlinkage void start_secondary(void)
 {
-	unsigned int cpu = smp_processor_id();
+	unsigned int cpu;
 
 	cpu_probe();
 	cpu_report();
@@ -95,6 +95,8 @@ asmlinkage void start_secondary(void)
 	 */
 
 	calibrate_delay();
+	preempt_disable();
+	cpu = smp_processor_id();
 	cpu_data[cpu].udelay_val = loops_per_jiffy;
 
 	prom_smp_finish();

arch/parisc/kernel/process.c

@@ -92,7 +92,9 @@ void cpu_idle(void)
 	while (1) {
 		while (!need_resched())
 			barrier();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 		check_pgt_cache();
 	}
 }

arch/parisc/kernel/smp.c

@@ -463,6 +463,7 @@ void __init smp_callin(void)
 #endif
 
 	smp_cpu_init(slave_id);
+	preempt_disable();
 
 #if 0	/* NOT WORKING YET - see entry.S */
 	istack = (void *)__get_free_pages(GFP_KERNEL,ISTACK_ORDER);

arch/powerpc/platforms/iseries/setup.c

@@ -694,7 +694,9 @@ static void iseries_shared_idle(void)
 		if (hvlpevent_is_pending())
 			process_iSeries_events();
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 
@@ -726,7 +728,9 @@ static void iseries_dedicated_idle(void)
 		}
 
 		ppc64_runlatch_on();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/powerpc/platforms/pseries/setup.c

@@ -539,7 +539,9 @@ static void pseries_dedicated_idle(void)
 		lpaca->lppaca.idle = 0;
 		ppc64_runlatch_on();
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 
 		if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING)
 			cpu_die();
@@ -583,7 +585,9 @@ static void pseries_shared_idle(void)
 		lpaca->lppaca.idle = 0;
 		ppc64_runlatch_on();
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 
 		if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING)
 			cpu_die();

arch/ppc/kernel/idle.c

@@ -53,10 +53,6 @@ void default_idle(void)
 		}
 #endif
 	}
-	if (need_resched())
-		schedule();
-	if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING)
-		cpu_die();
 }
 
 /*
@@ -64,11 +60,22 @@ void default_idle(void)
  */
 void cpu_idle(void)
 {
-	for (;;)
+	int cpu = smp_processor_id();
+
+	for (;;) {
 		if (ppc_md.idle != NULL)
 			ppc_md.idle();
 		else
 			default_idle();
+		if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING)
+			cpu_die();
+		if (need_resched()) {
+			preempt_enable_no_resched();
+			schedule();
+			preempt_disable();
+		}
+
+	}
 }
 
 #if defined(CONFIG_SYSCTL) && defined(CONFIG_6xx)

arch/ppc/kernel/smp.c

@@ -341,6 +341,7 @@ int __devinit start_secondary(void *unused)
 	cpu = smp_processor_id();
         smp_store_cpu_info(cpu);
 	set_dec(tb_ticks_per_jiffy);
+	preempt_disable();
 	cpu_callin_map[cpu] = 1;
 
 	printk("CPU %d done callin...\n", cpu);

arch/ppc64/kernel/idle.c

@@ -61,7 +61,9 @@ void default_idle(void)
 		}
 
 		ppc64_runlatch_on();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 		if (cpu_is_offline(cpu) && system_state == SYSTEM_RUNNING)
 			cpu_die();
 	}
@@ -77,7 +79,9 @@ void native_idle(void)
 
 		if (need_resched()) {
 			ppc64_runlatch_on();
+			preempt_enable_no_resched();
 			schedule();
+			preempt_disable();
 		}
 
 		if (cpu_is_offline(smp_processor_id()) &&

arch/s390/kernel/process.c

@@ -102,7 +102,6 @@ void default_idle(void)
 	local_irq_disable();
         if (need_resched()) {
 		local_irq_enable();
-                schedule();
                 return;
         }
 
@@ -139,8 +138,14 @@ void default_idle(void)
 
 void cpu_idle(void)
 {
-	for (;;)
+	for (;;) {
+		while (!need_resched())
 			default_idle();
+
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
 }
 
 void show_regs(struct pt_regs *regs)

arch/s390/kernel/smp.c

@@ -533,6 +533,7 @@ int __devinit start_secondary(void *cpuvoid)
 {
         /* Setup the cpu */
         cpu_init();
+	preempt_disable();
         /* init per CPU timer */
         init_cpu_timer();
 #ifdef CONFIG_VIRT_TIMER

arch/sh/kernel/process.c

@@ -64,7 +64,9 @@ void default_idle(void)
 				cpu_sleep();
 		}
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/sh/kernel/smp.c

@@ -112,7 +112,9 @@ int __cpu_up(unsigned int cpu)
 
 int start_secondary(void *unused)
 {
-	unsigned int cpu = smp_processor_id();
+	unsigned int cpu;
+
+	cpu = smp_processor_id();
 
 	atomic_inc(&init_mm.mm_count);
 	current->active_mm = &init_mm;
@@ -120,6 +122,7 @@ int start_secondary(void *unused)
 	smp_store_cpu_info(cpu);
 
 	__smp_slave_init(cpu);
+	preempt_disable();
 	per_cpu_trap_init();
 	
 	atomic_inc(&cpus_booted);

arch/sh64/kernel/process.c

@@ -334,7 +334,9 @@ void default_idle(void)
 			}
 			local_irq_enable();
 		}
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/sparc/kernel/process.c

@@ -120,7 +120,9 @@ void cpu_idle(void)
 			(*pm_idle)();
 		}
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 		check_pgt_cache();
 	}
 }
@@ -133,7 +135,9 @@ void cpu_idle(void)
 	/* endless idle loop with no priority at all */
 	while(1) {
 		if(need_resched()) {
+			preempt_enable_no_resched();
 			schedule();
+			preempt_disable();
 			check_pgt_cache();
 		}
 		barrier(); /* or else gcc optimizes... */

arch/sparc64/kernel/process.c

@@ -74,7 +74,9 @@ void cpu_idle(void)
 		while (!need_resched())
 			barrier();
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 		check_pgt_cache();
 	}
 }
@@ -93,7 +95,9 @@ void cpu_idle(void)
 		if (need_resched()) {
 			unidle_me();
 			clear_thread_flag(TIF_POLLING_NRFLAG);
+			preempt_enable_no_resched();
 			schedule();
+			preempt_disable();
 			set_thread_flag(TIF_POLLING_NRFLAG);
 			check_pgt_cache();
 		}

arch/sparc64/kernel/smp.c

@@ -168,6 +168,9 @@ void __init smp_callin(void)
 		rmb();
 
 	cpu_set(cpuid, cpu_online_map);
+
+	/* idle thread is expected to have preempt disabled */
+	preempt_disable();
 }
 
 void cpu_panic(void)

arch/v850/kernel/process.c

@@ -36,11 +36,8 @@ extern void ret_from_fork (void);
 /* The idle loop.  */
 void default_idle (void)
 {
-	while (1) {
 	while (! need_resched ())
 		asm ("halt; nop; nop; nop; nop; nop" ::: "cc");
-		schedule ();
-	}
 }
 
 void (*idle)(void) = default_idle;
@@ -54,7 +51,14 @@ void (*idle)(void) = default_idle;
 void cpu_idle (void)
 {
 	/* endless idle loop with no priority at all */
+	while (1) {
+		while (!need_resched())
 			(*idle) ();
+
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
 }
 
 /*

arch/x86_64/kernel/process.c

@@ -204,7 +204,9 @@ void cpu_idle (void)
 			idle();
 		}
 
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

arch/x86_64/kernel/smpboot.c

@@ -472,6 +472,7 @@ void __cpuinit start_secondary(void)
 	 * things done here to the most necessary things.
 	 */
 	cpu_init();
+	preempt_disable();
 	smp_callin();
 
 	/* otherwise gcc will move up the smp_processor_id before the cpu_init */

arch/xtensa/kernel/process.c

@@ -96,8 +96,9 @@ void cpu_idle(void)
 	while (1) {
 		while (!need_resched())
 			platform_idle();
-		preempt_enable();
+		preempt_enable_no_resched();
 		schedule();
+		preempt_disable();
 	}
 }
 

init/main.c

@@ -394,14 +394,16 @@ static void noinline rest_init(void)
 	kernel_thread(init, NULL, CLONE_FS | CLONE_SIGHAND);
 	numa_default_policy();
 	unlock_kernel();
-	preempt_enable_no_resched();
 
 	/*
 	 * The boot idle thread must execute schedule()
 	 * at least one to get things moving:
 	 */
+	preempt_enable_no_resched();
 	schedule();
+	preempt_disable();
 
+	/* Call into cpu_idle with preempt disabled */
 	cpu_idle();
 }