Merge git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched: (96 commits)
  sched: keep total / count stats in addition to the max for
  sched, futex: detach sched.h and futex.h
  sched: fix: don't take a mutex from interrupt context
  sched: print backtrace of running tasks too
  printk: use ktime_get()
  softlockup: fix signedness
  sched: latencytop support
  sched: fix goto retry in pick_next_task_rt()
  timers: don't #error on higher HZ values
  sched: monitor clock underflows in /proc/sched_debug
  sched: fix rq->clock warps on frequency changes
  sched: fix, always create kernel threads with normal priority
  debug: clean up kernel/profile.c
  sched: remove the !PREEMPT_BKL code
  sched: make PREEMPT_BKL the default
  debug: track and print last unloaded module in the oops trace
  debug: show being-loaded/being-unloaded indicator for modules
  sched: rt-watchdog: fix .rlim_max = RLIM_INFINITY
  sched: rt-group: reduce rescheduling
  hrtimer: unlock hrtimer_wakeup
  ...

Documentation/RCU/RTFP.txt

@@ -9,8 +9,8 @@ The first thing resembling RCU was published in 1980, when Kung and Lehman
 [Kung80] recommended use of a garbage collector to defer destruction
 of nodes in a parallel binary search tree in order to simplify its
 implementation.  This works well in environments that have garbage
-collectors, but current production garbage collectors incur significant
-read-side overhead.
+collectors, but most production garbage collectors incur significant
+overhead.
 
 In 1982, Manber and Ladner [Manber82,Manber84] recommended deferring
 destruction until all threads running at that time have terminated, again
@@ -99,16 +99,25 @@ locking, reduces contention, reduces memory latency for readers, and
 parallelizes pipeline stalls and memory latency for writers.  However,
 these techniques still impose significant read-side overhead in the
 form of memory barriers.  Researchers at Sun worked along similar lines
-in the same timeframe [HerlihyLM02,HerlihyLMS03].  These techniques
-can be thought of as inside-out reference counts, where the count is
-represented by the number of hazard pointers referencing a given data
-structure (rather than the more conventional counter field within the
-data structure itself).
+in the same timeframe [HerlihyLM02].  These techniques can be thought
+of as inside-out reference counts, where the count is represented by the
+number of hazard pointers referencing a given data structure (rather than
+the more conventional counter field within the data structure itself).
+
+By the same token, RCU can be thought of as a "bulk reference count",
+where some form of reference counter covers all reference by a given CPU
+or thread during a set timeframe.  This timeframe is related to, but
+not necessarily exactly the same as, an RCU grace period.  In classic
+RCU, the reference counter is the per-CPU bit in the "bitmask" field,
+and each such bit covers all references that might have been made by
+the corresponding CPU during the prior grace period.  Of course, RCU
+can be thought of in other terms as well.
 
 In 2003, the K42 group described how RCU could be used to create
-hot-pluggable implementations of operating-system functions.  Later that
-year saw a paper describing an RCU implementation of System V IPC
-[Arcangeli03], and an introduction to RCU in Linux Journal [McKenney03a].
+hot-pluggable implementations of operating-system functions [Appavoo03a].
+Later that year saw a paper describing an RCU implementation of System
+V IPC [Arcangeli03], and an introduction to RCU in Linux Journal
+[McKenney03a].
 
 2004 has seen a Linux-Journal article on use of RCU in dcache
 [McKenney04a], a performance comparison of locking to RCU on several
@@ -117,10 +126,19 @@ number of operating-system kernels [PaulEdwardMcKenneyPhD], a paper
 describing how to make RCU safe for soft-realtime applications [Sarma04c],
 and a paper describing SELinux performance with RCU [JamesMorris04b].
 
-2005 has seen further adaptation of RCU to realtime use, permitting
+2005 brought further adaptation of RCU to realtime use, permitting
 preemption of RCU realtime critical sections [PaulMcKenney05a,
 PaulMcKenney05b].
 
+2006 saw the first best-paper award for an RCU paper [ThomasEHart2006a],
+as well as further work on efficient implementations of preemptible
+RCU [PaulEMcKenney2006b], but priority-boosting of RCU read-side critical
+sections proved elusive.  An RCU implementation permitting general
+blocking in read-side critical sections appeared [PaulEMcKenney2006c],
+Robert Olsson described an RCU-protected trie-hash combination
+[RobertOlsson2006a].
+
+
 Bibtex Entries
 
 @article{Kung80
@@ -203,6 +221,41 @@ Bibtex Entries
 ,Address="New Orleans, LA"
 }
 
+@conference{Pu95a,
+Author = "Calton Pu and Tito Autrey and Andrew Black and Charles Consel and
+Crispin Cowan and Jon Inouye and Lakshmi Kethana and Jonathan Walpole and
+Ke Zhang",
+Title = "Optimistic Incremental Specialization: Streamlining a Commercial
+Operating System",
+Booktitle = "15\textsuperscript{th} ACM Symposium on
+Operating Systems Principles (SOSP'95)",
+address = "Copper Mountain, CO",
+month="December",
+year="1995",
+pages="314-321",
+annotation="
+	Uses a replugger, but with a flag to signal when people are
+	using the resource at hand.  Only one reader at a time.
+"
+}
+
+@conference{Cowan96a,
+Author = "Crispin Cowan and Tito Autrey and Charles Krasic and
+Calton Pu and Jonathan Walpole",
+Title = "Fast Concurrent Dynamic Linking for an Adaptive Operating System",
+Booktitle = "International Conference on Configurable Distributed Systems
+(ICCDS'96)",
+address = "Annapolis, MD",
+month="May",
+year="1996",
+pages="108",
+isbn="0-8186-7395-8",
+annotation="
+	Uses a replugger, but with a counter to signal when people are
+	using the resource at hand.  Allows multiple readers.
+"
+}
+
 @techreport{Slingwine95
 ,author="John D. Slingwine and Paul E. McKenney"
 ,title="Apparatus and Method for Achieving Reduced Overhead Mutual
@@ -312,6 +365,49 @@ Andrea Arcangeli and Andi Kleen and Orran Krieger and Rusty Russell"
 [Viewed June 23, 2004]"
 }
 
+@conference{Michael02a
+,author="Maged M. Michael"
+,title="Safe Memory Reclamation for Dynamic Lock-Free Objects Using Atomic
+Reads and Writes"
+,Year="2002"
+,Month="August"
+,booktitle="{Proceedings of the 21\textsuperscript{st} Annual ACM
+Symposium on Principles of Distributed Computing}"
+,pages="21-30"
+,annotation="
+	Each thread keeps an array of pointers to items that it is
+	currently referencing.	Sort of an inside-out garbage collection
+	mechanism, but one that requires the accessing code to explicitly
+	state its needs.  Also requires read-side memory barriers on
+	most architectures.
+"
+}
+
+@conference{Michael02b
+,author="Maged M. Michael"
+,title="High Performance Dynamic Lock-Free Hash Tables and List-Based Sets"
+,Year="2002"
+,Month="August"
+,booktitle="{Proceedings of the 14\textsuperscript{th} Annual ACM
+Symposium on Parallel
+Algorithms and Architecture}"
+,pages="73-82"
+,annotation="
+	Like the title says...
+"
+}
+
+@InProceedings{HerlihyLM02
+,author={Maurice Herlihy and Victor Luchangco and Mark Moir}
+,title="The Repeat Offender Problem: A Mechanism for Supporting Dynamic-Sized,
+Lock-Free Data Structures"
+,booktitle={Proceedings of 16\textsuperscript{th} International
+Symposium on Distributed Computing}
+,year=2002
+,month="October"
+,pages="339-353"
+}
+
 @article{Appavoo03a
 ,author="J. Appavoo and K. Hui and C. A. N. Soules and R. W. Wisniewski and
 D. M. {Da Silva} and O. Krieger and M. A. Auslander and D. J. Edelsohn and
@@ -447,3 +543,95 @@ Oregon Health and Sciences University"
 	Realtime turns into making RCU yet more realtime friendly.
 "
 }
+
+@conference{ThomasEHart2006a
+,Author="Thomas E. Hart and Paul E. McKenney and Angela Demke Brown"
+,Title="Making Lockless Synchronization Fast: Performance Implications
+of Memory Reclamation"
+,Booktitle="20\textsuperscript{th} {IEEE} International Parallel and
+Distributed Processing Symposium"
+,month="April"
+,year="2006"
+,day="25-29"
+,address="Rhodes, Greece"
+,annotation="
+	Compares QSBR (AKA "classic RCU"), HPBR, EBR, and lock-free
+	reference counting.
+"
+}
+
+@Conference{PaulEMcKenney2006b
+,Author="Paul E. McKenney and Dipankar Sarma and Ingo Molnar and
+Suparna Bhattacharya"
+,Title="Extending RCU for Realtime and Embedded Workloads"
+,Booktitle="{Ottawa Linux Symposium}"
+,Month="July"
+,Year="2006"
+,pages="v2 123-138"
+,note="Available:
+\url{http://www.linuxsymposium.org/2006/view_abstract.php?content_key=184}
+\url{http://www.rdrop.com/users/paulmck/RCU/OLSrtRCU.2006.08.11a.pdf}
+[Viewed January 1, 2007]"
+,annotation="
+	Described how to improve the -rt implementation of realtime RCU.
+"
+}
+
+@unpublished{PaulEMcKenney2006c
+,Author="Paul E. McKenney"
+,Title="Sleepable {RCU}"
+,month="October"
+,day="9"
+,year="2006"
+,note="Available:
+\url{http://lwn.net/Articles/202847/}
+Revised:
+\url{http://www.rdrop.com/users/paulmck/RCU/srcu.2007.01.14a.pdf}
+[Viewed August 21, 2006]"
+,annotation="
+	LWN article introducing SRCU.
+"
+}
+
+@unpublished{RobertOlsson2006a
+,Author="Robert Olsson and Stefan Nilsson"
+,Title="{TRASH}: A dynamic {LC}-trie and hash data structure"
+,month="August"
+,day="18"
+,year="2006"
+,note="Available:
+\url{http://www.nada.kth.se/~snilsson/public/papers/trash/trash.pdf}
+[Viewed February 24, 2007]"
+,annotation="
+	RCU-protected dynamic trie-hash combination.
+"
+}
+
+@unpublished{ThomasEHart2007a
+,Author="Thomas E. Hart and Paul E. McKenney and Angela Demke Brown and Jonathan Walpole"
+,Title="Performance of memory reclamation for lockless synchronization"
+,journal="J. Parallel Distrib. Comput."
+,year="2007"
+,note="To appear in J. Parallel Distrib. Comput.
+       \url{doi=10.1016/j.jpdc.2007.04.010}"
+,annotation={
+	Compares QSBR (AKA "classic RCU"), HPBR, EBR, and lock-free
+	reference counting.  Journal version of ThomasEHart2006a.
+}
+}
+
+@unpublished{PaulEMcKenney2007QRCUspin
+,Author="Paul E. McKenney"
+,Title="Using Promela and Spin to verify parallel algorithms"
+,month="August"
+,day="1"
+,year="2007"
+,note="Available:
+\url{http://lwn.net/Articles/243851/}
+[Viewed September 8, 2007]"
+,annotation="
+	LWN article describing Promela and spin, and also using Oleg
+	Nesterov's QRCU as an example (with Paul McKenney's fastpath).
+"
+}
+

Documentation/RCU/rcu.txt

@@ -36,6 +36,14 @@ o	How can the updater tell when a grace period has completed
 	executed in user mode, or executed in the idle loop, we can
 	safely free up that item.
 
+	Preemptible variants of RCU (CONFIG_PREEMPT_RCU) get the
+	same effect, but require that the readers manipulate CPU-local
+	counters.  These counters allow limited types of blocking
+	within RCU read-side critical sections.  SRCU also uses
+	CPU-local counters, and permits general blocking within
+	RCU read-side critical sections.  These two variants of
+	RCU detect grace periods by sampling these counters.
+
 o	If I am running on a uniprocessor kernel, which can only do one
 	thing at a time, why should I wait for a grace period?
 
@@ -46,7 +54,10 @@ o	How can I see where RCU is currently used in the Linux kernel?
 	Search for "rcu_read_lock", "rcu_read_unlock", "call_rcu",
 	"rcu_read_lock_bh", "rcu_read_unlock_bh", "call_rcu_bh",
 	"srcu_read_lock", "srcu_read_unlock", "synchronize_rcu",
-	"synchronize_net", and "synchronize_srcu".
+	"synchronize_net", "synchronize_srcu", and the other RCU
+	primitives.  Or grab one of the cscope databases from:
+
+	http://www.rdrop.com/users/paulmck/RCU/linuxusage/rculocktab.html
 
 o	What guidelines should I follow when writing code that uses RCU?
 
@@ -67,7 +78,11 @@ o	I hear that RCU is patented?  What is with that?
 
 o	I hear that RCU needs work in order to support realtime kernels?
 
-	Yes, work in progress.
+	This work is largely completed.  Realtime-friendly RCU can be
+	enabled via the CONFIG_PREEMPT_RCU kernel configuration parameter.
+	However, work is in progress for enabling priority boosting of
+	preempted RCU read-side critical sections.This is needed if you
+	have CPU-bound realtime threads.
 
 o	Where can I find more information on RCU?
 

Documentation/RCU/torture.txt

@@ -46,12 +46,13 @@ stat_interval	The number of seconds between output of torture
 
 shuffle_interval
 		The number of seconds to keep the test threads affinitied
-		to a particular subset of the CPUs.  Used in conjunction
-		with test_no_idle_hz.
+		to a particular subset of the CPUs, defaults to 5 seconds.
+		Used in conjunction with test_no_idle_hz.
 
 test_no_idle_hz	Whether or not to test the ability of RCU to operate in
 		a kernel that disables the scheduling-clock interrupt to
 		idle CPUs.  Boolean parameter, "1" to test, "0" otherwise.
+		Defaults to omitting this test.
 
 torture_type	The type of RCU to test: "rcu" for the rcu_read_lock() API,
 		"rcu_sync" for rcu_read_lock() with synchronous reclamation,
@@ -82,8 +83,6 @@ be evident.  ;-)
 
 The entries are as follows:
 
-o	"ggp": The number of counter flips (or batches) since boot.
-
 o	"rtc": The hexadecimal address of the structure currently visible
 	to readers.
 
@@ -117,8 +116,8 @@ o	"Reader Pipe": Histogram of "ages" of structures seen by readers.
 o	"Reader Batch": Another histogram of "ages" of structures seen
 	by readers, but in terms of counter flips (or batches) rather
 	than in terms of grace periods.  The legal number of non-zero
-	entries is again two.  The reason for this separate view is
-	that it is easier to get the third entry to show up in the
+	entries is again two.  The reason for this separate view is that
+	it is sometimes easier to get the third entry to show up in the
 	"Reader Batch" list than in the "Reader Pipe" list.
 
 o	"Free-Block Circulation": Shows the number of torture structures

Documentation/cpu-hotplug.txt

@@ -109,12 +109,13 @@ Never use anything other than cpumask_t to represent bitmap of CPUs.
 	for_each_cpu_mask(x,mask) - Iterate over some random collection of cpu mask.
 
 	#include <linux/cpu.h>
-	lock_cpu_hotplug() and unlock_cpu_hotplug():
+	get_online_cpus() and put_online_cpus():
 
-The above calls are used to inhibit cpu hotplug operations. While holding the
-cpucontrol mutex, cpu_online_map will not change. If you merely need to avoid
-cpus going away, you could also use preempt_disable() and preempt_enable()
-for those sections. Just remember the critical section cannot call any
+The above calls are used to inhibit cpu hotplug operations. While the
+cpu_hotplug.refcount is non zero, the cpu_online_map will not change.
+If you merely need to avoid cpus going away, you could also use
+preempt_disable() and preempt_enable() for those sections.
+Just remember the critical section cannot call any
 function that can sleep or schedule this process away. The preempt_disable()
 will work as long as stop_machine_run() is used to take a cpu down.
 

arch/arm/kernel/time.c

@@ -79,17 +79,6 @@ static unsigned long dummy_gettimeoffset(void)
 }
 #endif
 
-/*
- * An implementation of printk_clock() independent from
- * sched_clock().  This avoids non-bootable kernels when
- * printk_clock is enabled.
- */
-unsigned long long printk_clock(void)
-{
-	return (unsigned long long)(jiffies - INITIAL_JIFFIES) *
-			(1000000000 / HZ);
-}
-
 static unsigned long next_rtc_update;
 
 /*

arch/ia64/kernel/setup.c

@@ -71,8 +71,6 @@ unsigned long __per_cpu_offset[NR_CPUS];
 EXPORT_SYMBOL(__per_cpu_offset);
 #endif
 
-extern void ia64_setup_printk_clock(void);
-
 DEFINE_PER_CPU(struct cpuinfo_ia64, cpu_info);
 DEFINE_PER_CPU(unsigned long, local_per_cpu_offset);
 unsigned long ia64_cycles_per_usec;
@@ -507,8 +505,6 @@ setup_arch (char **cmdline_p)
 	/* process SAL system table: */
 	ia64_sal_init(__va(efi.sal_systab));
 
-	ia64_setup_printk_clock();
-
 #ifdef CONFIG_SMP
 	cpu_physical_id(0) = hard_smp_processor_id();
 #endif

arch/ia64/kernel/time.c

@@ -344,33 +344,6 @@ udelay (unsigned long usecs)
 }
 EXPORT_SYMBOL(udelay);
 
-static unsigned long long ia64_itc_printk_clock(void)
-{
-	if (ia64_get_kr(IA64_KR_PER_CPU_DATA))
-		return sched_clock();
-	return 0;
-}
-
-static unsigned long long ia64_default_printk_clock(void)
-{
-	return (unsigned long long)(jiffies_64 - INITIAL_JIFFIES) *
-		(1000000000/HZ);
-}
-
-unsigned long long (*ia64_printk_clock)(void) = &ia64_default_printk_clock;
-
-unsigned long long printk_clock(void)
-{
-	return ia64_printk_clock();
-}
-
-void __init
-ia64_setup_printk_clock(void)
-{
-	if (!(sal_platform_features & IA64_SAL_PLATFORM_FEATURE_ITC_DRIFT))
-		ia64_printk_clock = ia64_itc_printk_clock;
-}
-
 /* IA64 doesn't cache the timezone */
 void update_vsyscall_tz(void)
 {

arch/ia64/sn/kernel/setup.c

@@ -64,7 +64,6 @@ extern void sn_timer_init(void);
 extern unsigned long last_time_offset;
 extern void (*ia64_mark_idle) (int);
 extern void snidle(int);
-extern unsigned long long (*ia64_printk_clock)(void);
 
 unsigned long sn_rtc_cycles_per_second;
 EXPORT_SYMBOL(sn_rtc_cycles_per_second);
@@ -360,14 +359,6 @@ sn_scan_pcdp(void)
 
 static unsigned long sn2_rtc_initial;
 
-static unsigned long long ia64_sn2_printk_clock(void)
-{
-	unsigned long rtc_now = rtc_time();
-
-	return (rtc_now - sn2_rtc_initial) *
-		(1000000000 / sn_rtc_cycles_per_second);
-}
-
 /**
  * sn_setup - SN platform setup routine
  * @cmdline_p: kernel command line
@@ -468,8 +459,6 @@ void __init sn_setup(char **cmdline_p)
 
 	platform_intr_list[ACPI_INTERRUPT_CPEI] = IA64_CPE_VECTOR;
 
-	ia64_printk_clock = ia64_sn2_printk_clock;
-
 	printk("SGI SAL version %x.%02x\n", version >> 8, version & 0x00FF);
 
 	/*

arch/mips/kernel/mips-mt-fpaff.c

@@ -58,13 +58,13 @@ asmlinkage long mipsmt_sys_sched_setaffinity(pid_t pid, unsigned int len,
 	if (copy_from_user(&new_mask, user_mask_ptr, sizeof(new_mask)))
 		return -EFAULT;
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	read_lock(&tasklist_lock);
 
 	p = find_process_by_pid(pid);
 	if (!p) {
 		read_unlock(&tasklist_lock);
-		unlock_cpu_hotplug();
+		put_online_cpus();
 		return -ESRCH;
 	}
 
@@ -106,7 +106,7 @@ asmlinkage long mipsmt_sys_sched_setaffinity(pid_t pid, unsigned int len,
 
 out_unlock:
 	put_task_struct(p);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 	return retval;
 }
 
@@ -125,7 +125,7 @@ asmlinkage long mipsmt_sys_sched_getaffinity(pid_t pid, unsigned int len,
 	if (len < real_len)
 		return -EINVAL;
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	read_lock(&tasklist_lock);
 
 	retval = -ESRCH;
@@ -140,7 +140,7 @@ asmlinkage long mipsmt_sys_sched_getaffinity(pid_t pid, unsigned int len,
 
 out_unlock:
 	read_unlock(&tasklist_lock);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 	if (retval)
 		return retval;
 	if (copy_to_user(user_mask_ptr, &mask, real_len))

arch/powerpc/platforms/pseries/hotplug-cpu.c

@@ -153,7 +153,7 @@ static int pseries_add_processor(struct device_node *np)
 	for (i = 0; i < nthreads; i++)
 		cpu_set(i, tmp);
 
-	lock_cpu_hotplug();
+	cpu_maps_update_begin();
 
 	BUG_ON(!cpus_subset(cpu_present_map, cpu_possible_map));
 
@@ -190,7 +190,7 @@ static int pseries_add_processor(struct device_node *np)
 	}
 	err = 0;
 out_unlock:
-	unlock_cpu_hotplug();
+	cpu_maps_update_done();
 	return err;
 }
 
@@ -211,7 +211,7 @@ static void pseries_remove_processor(struct device_node *np)
 
 	nthreads = len / sizeof(u32);
 
-	lock_cpu_hotplug();
+	cpu_maps_update_begin();
 	for (i = 0; i < nthreads; i++) {
 		for_each_present_cpu(cpu) {
 			if (get_hard_smp_processor_id(cpu) != intserv[i])
@@ -225,7 +225,7 @@ static void pseries_remove_processor(struct device_node *np)
 			printk(KERN_WARNING "Could not find cpu to remove "
 			       "with physical id 0x%x\n", intserv[i]);
 	}
-	unlock_cpu_hotplug();
+	cpu_maps_update_done();
 }
 
 static int pseries_smp_notifier(struct notifier_block *nb,

arch/powerpc/platforms/pseries/rtasd.c

@@ -382,7 +382,7 @@ static void do_event_scan_all_cpus(long delay)
 {
 	int cpu;
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	cpu = first_cpu(cpu_online_map);
 	for (;;) {
 		set_cpus_allowed(current, cpumask_of_cpu(cpu));
@@ -390,15 +390,15 @@ static void do_event_scan_all_cpus(long delay)
 		set_cpus_allowed(current, CPU_MASK_ALL);
 
 		/* Drop hotplug lock, and sleep for the specified delay */
-		unlock_cpu_hotplug();
+		put_online_cpus();
 		msleep_interruptible(delay);
-		lock_cpu_hotplug();
+		get_online_cpus();
 
 		cpu = next_cpu(cpu, cpu_online_map);
 		if (cpu == NR_CPUS)
 			break;
 	}
-	unlock_cpu_hotplug();
+	put_online_cpus();
 }
 
 static int rtasd(void *unused)

arch/x86/kernel/cpu/mtrr/main.c

@@ -349,7 +349,7 @@ int mtrr_add_page(unsigned long base, unsigned long size,
 	replace = -1;
 
 	/* No CPU hotplug when we change MTRR entries */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	/*  Search for existing MTRR  */
 	mutex_lock(&mtrr_mutex);
 	for (i = 0; i < num_var_ranges; ++i) {
@@ -405,7 +405,7 @@ int mtrr_add_page(unsigned long base, unsigned long size,
 	error = i;
  out:
 	mutex_unlock(&mtrr_mutex);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 	return error;
 }
 
@@ -495,7 +495,7 @@ int mtrr_del_page(int reg, unsigned long base, unsigned long size)
 
 	max = num_var_ranges;
 	/* No CPU hotplug when we change MTRR entries */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	mutex_lock(&mtrr_mutex);
 	if (reg < 0) {
 		/*  Search for existing MTRR  */
@@ -536,7 +536,7 @@ int mtrr_del_page(int reg, unsigned long base, unsigned long size)
 	error = reg;
  out:
 	mutex_unlock(&mtrr_mutex);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 	return error;
 }
 /**

arch/x86/kernel/entry_64.S

@@ -283,7 +283,7 @@ sysret_careful:
 sysret_signal:
 	TRACE_IRQS_ON
 	sti
-	testl $(_TIF_SIGPENDING|_TIF_SINGLESTEP|_TIF_MCE_NOTIFY),%edx
+	testl $_TIF_DO_NOTIFY_MASK,%edx
 	jz    1f
 
 	/* Really a signal */
@@ -377,7 +377,7 @@ int_very_careful:
 	jmp int_restore_rest
 	
 int_signal:
-	testl $(_TIF_SIGPENDING|_TIF_SINGLESTEP|_TIF_MCE_NOTIFY),%edx
+	testl $_TIF_DO_NOTIFY_MASK,%edx
 	jz 1f
 	movq %rsp,%rdi		# &ptregs -> arg1
 	xorl %esi,%esi		# oldset -> arg2
@@ -603,7 +603,7 @@ retint_careful:
 	jmp retint_check
 	
 retint_signal:
-	testl $(_TIF_SIGPENDING|_TIF_SINGLESTEP|_TIF_MCE_NOTIFY),%edx
+	testl $_TIF_DO_NOTIFY_MASK,%edx
 	jz    retint_swapgs
 	TRACE_IRQS_ON
 	sti

arch/x86/kernel/microcode.c

@@ -436,7 +436,7 @@ static ssize_t microcode_write (struct file *file, const char __user *buf, size_
 		return -EINVAL;
 	}
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	mutex_lock(&microcode_mutex);
 
 	user_buffer = (void __user *) buf;
@@ -447,7 +447,7 @@ static ssize_t microcode_write (struct file *file, const char __user *buf, size_
 		ret = (ssize_t)len;
 
 	mutex_unlock(&microcode_mutex);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 
 	return ret;
 }
@@ -658,14 +658,14 @@ static ssize_t reload_store(struct sys_device *dev, const char *buf, size_t sz)
 
 		old = current->cpus_allowed;
 
-		lock_cpu_hotplug();
+		get_online_cpus();
 		set_cpus_allowed(current, cpumask_of_cpu(cpu));
 
 		mutex_lock(&microcode_mutex);
 		if (uci->valid)
 			err = cpu_request_microcode(cpu);
 		mutex_unlock(&microcode_mutex);
-		unlock_cpu_hotplug();
+		put_online_cpus();
 		set_cpus_allowed(current, old);
 	}
 	if (err)
@@ -817,9 +817,9 @@ static int __init microcode_init (void)
 		return PTR_ERR(microcode_pdev);
 	}
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	error = sysdev_driver_register(&cpu_sysdev_class, &mc_sysdev_driver);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 	if (error) {
 		microcode_dev_exit();
 		platform_device_unregister(microcode_pdev);
@@ -839,9 +839,9 @@ static void __exit microcode_exit (void)
 
 	unregister_hotcpu_notifier(&mc_cpu_notifier);
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 	sysdev_driver_unregister(&cpu_sysdev_class, &mc_sysdev_driver);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 
 	platform_device_unregister(microcode_pdev);
 }

arch/x86/kernel/signal_32.c

@@ -658,6 +658,9 @@ void do_notify_resume(struct pt_regs *regs, void *_unused,
 	/* deal with pending signal delivery */
 	if (thread_info_flags & (_TIF_SIGPENDING | _TIF_RESTORE_SIGMASK))
 		do_signal(regs);
+
+	if (thread_info_flags & _TIF_HRTICK_RESCHED)
+		hrtick_resched();
 	
 	clear_thread_flag(TIF_IRET);
 }

arch/x86/kernel/signal_64.c

@@ -480,6 +480,9 @@ do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags)
 	/* deal with pending signal delivery */
 	if (thread_info_flags & (_TIF_SIGPENDING|_TIF_RESTORE_SIGMASK))
 		do_signal(regs);
+
+	if (thread_info_flags & _TIF_HRTICK_RESCHED)
+		hrtick_resched();
 }
 
 void signal_fault(struct pt_regs *regs, void __user *frame, char *where)

arch/x86/kernel/stacktrace.c

@@ -33,6 +33,19 @@ static void save_stack_address(void *data, unsigned long addr)
 		trace->entries[trace->nr_entries++] = addr;
 }
 
+static void save_stack_address_nosched(void *data, unsigned long addr)
+{
+	struct stack_trace *trace = (struct stack_trace *)data;
+	if (in_sched_functions(addr))
+		return;
+	if (trace->skip > 0) {
+		trace->skip--;
+		return;
+	}
+	if (trace->nr_entries < trace->max_entries)
+		trace->entries[trace->nr_entries++] = addr;
+}
+
 static const struct stacktrace_ops save_stack_ops = {
 	.warning = save_stack_warning,
 	.warning_symbol = save_stack_warning_symbol,
@@ -40,6 +53,13 @@ static const struct stacktrace_ops save_stack_ops = {
 	.address = save_stack_address,
 };
 
+static const struct stacktrace_ops save_stack_ops_nosched = {
+	.warning = save_stack_warning,
+	.warning_symbol = save_stack_warning_symbol,
+	.stack = save_stack_stack,
+	.address = save_stack_address_nosched,
+};
+
 /*
  * Save stack-backtrace addresses into a stack_trace buffer.
  */
@@ -50,3 +70,10 @@ void save_stack_trace(struct stack_trace *trace)
 		trace->entries[trace->nr_entries++] = ULONG_MAX;
 }
 EXPORT_SYMBOL(save_stack_trace);
+
+void save_stack_trace_tsk(struct task_struct *tsk, struct stack_trace *trace)
+{
+	dump_trace(tsk, NULL, NULL, &save_stack_ops_nosched, trace);
+	if (trace->nr_entries < trace->max_entries)
+		trace->entries[trace->nr_entries++] = ULONG_MAX;
+}

drivers/lguest/x86/core.c

@@ -459,7 +459,7 @@ void __init lguest_arch_host_init(void)
 
 	/* We don't need the complexity of CPUs coming and going while we're
 	 * doing this. */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	if (cpu_has_pge) { /* We have a broader idea of "global". */
 		/* Remember that this was originally set (for cleanup). */
 		cpu_had_pge = 1;
@@ -469,20 +469,20 @@ void __init lguest_arch_host_init(void)
 		/* Turn off the feature in the global feature set. */
 		clear_bit(X86_FEATURE_PGE, boot_cpu_data.x86_capability);
 	}
-	unlock_cpu_hotplug();
+	put_online_cpus();
 };
 /*:*/
 
 void __exit lguest_arch_host_fini(void)
 {
 	/* If we had PGE before we started, turn it back on now. */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	if (cpu_had_pge) {
 		set_bit(X86_FEATURE_PGE, boot_cpu_data.x86_capability);
 		/* adjust_pge's argument "1" means set PGE. */
 		on_each_cpu(adjust_pge, (void *)1, 0, 1);
 	}
-	unlock_cpu_hotplug();
+	put_online_cpus();
 }
 
 

drivers/s390/char/sclp_config.c

@@ -29,12 +29,12 @@ static void sclp_cpu_capability_notify(struct work_struct *work)
 	struct sys_device *sysdev;
 
 	printk(KERN_WARNING TAG "cpu capability changed.\n");
-	lock_cpu_hotplug();
+	get_online_cpus();
 	for_each_online_cpu(cpu) {
 		sysdev = get_cpu_sysdev(cpu);
 		kobject_uevent(&sysdev->kobj, KOBJ_CHANGE);
 	}
-	unlock_cpu_hotplug();
+	put_online_cpus();
 }
 
 static void sclp_conf_receiver_fn(struct evbuf_header *evbuf)

fs/Kconfig

@@ -2130,4 +2130,3 @@ source "fs/nls/Kconfig"
 source "fs/dlm/Kconfig"
 
 endmenu
-

fs/proc/base.c

@@ -310,6 +310,77 @@ static int proc_pid_schedstat(struct task_struct *task, char *buffer)
 }
 #endif
 
+#ifdef CONFIG_LATENCYTOP
+static int lstats_show_proc(struct seq_file *m, void *v)
+{
+	int i;
+	struct task_struct *task = m->private;
+	seq_puts(m, "Latency Top version : v0.1\n");
+
+	for (i = 0; i < 32; i++) {
+		if (task->latency_record[i].backtrace[0]) {
+			int q;
+			seq_printf(m, "%i %li %li ",
+				task->latency_record[i].count,
+				task->latency_record[i].time,
+				task->latency_record[i].max);
+			for (q = 0; q < LT_BACKTRACEDEPTH; q++) {
+				char sym[KSYM_NAME_LEN];
+				char *c;
+				if (!task->latency_record[i].backtrace[q])
+					break;
+				if (task->latency_record[i].backtrace[q] == ULONG_MAX)
+					break;
+				sprint_symbol(sym, task->latency_record[i].backtrace[q]);
+				c = strchr(sym, '+');
+				if (c)
+					*c = 0;
+				seq_printf(m, "%s ", sym);
+			}
+			seq_printf(m, "\n");
+		}
+
+	}
+	return 0;
+}
+
+static int lstats_open(struct inode *inode, struct file *file)
+{
+	int ret;
+	struct seq_file *m;
+	struct task_struct *task = get_proc_task(inode);
+
+	ret = single_open(file, lstats_show_proc, NULL);
+	if (!ret) {
+		m = file->private_data;
+		m->private = task;
+	}
+	return ret;
+}
+
+static ssize_t lstats_write(struct file *file, const char __user *buf,
+			    size_t count, loff_t *offs)
+{
+	struct seq_file *m;
+	struct task_struct *task;
+
+	m = file->private_data;
+	task = m->private;
+	clear_all_latency_tracing(task);
+
+	return count;
+}
+
+static const struct file_operations proc_lstats_operations = {
+	.open		= lstats_open,
+	.read		= seq_read,
+	.write		= lstats_write,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+#endif
+
 /* The badness from the OOM killer */
 unsigned long badness(struct task_struct *p, unsigned long uptime);
 static int proc_oom_score(struct task_struct *task, char *buffer)
@@ -1020,6 +1091,7 @@ static const struct file_operations proc_fault_inject_operations = {
 };
 #endif
 
+
 #ifdef CONFIG_SCHED_DEBUG
 /*
  * Print out various scheduling related per-task fields:
@@ -2230,6 +2302,9 @@ static const struct pid_entry tgid_base_stuff[] = {
 #ifdef CONFIG_SCHEDSTATS
 	INF("schedstat",  S_IRUGO, pid_schedstat),
 #endif
+#ifdef CONFIG_LATENCYTOP
+	REG("latency",  S_IRUGO, lstats),
+#endif
 #ifdef CONFIG_PROC_PID_CPUSET
 	REG("cpuset",     S_IRUGO, cpuset),
 #endif
@@ -2555,6 +2630,9 @@ static const struct pid_entry tid_base_stuff[] = {
 #ifdef CONFIG_SCHEDSTATS
 	INF("schedstat", S_IRUGO, pid_schedstat),
 #endif
+#ifdef CONFIG_LATENCYTOP
+	REG("latency",  S_IRUGO, lstats),
+#endif
 #ifdef CONFIG_PROC_PID_CPUSET
 	REG("cpuset",    S_IRUGO, cpuset),
 #endif

include/asm-generic/resource.h

@@ -44,8 +44,8 @@
 #define RLIMIT_NICE		13	/* max nice prio allowed to raise to
 					   0-39 for nice level 19 .. -20 */
 #define RLIMIT_RTPRIO		14	/* maximum realtime priority */
-
-#define RLIM_NLIMITS		15
+#define RLIMIT_RTTIME		15	/* timeout for RT tasks in us */
+#define RLIM_NLIMITS		16
 
 /*
  * SuS says limits have to be unsigned.
@@ -86,6 +86,7 @@
 	[RLIMIT_MSGQUEUE]	= {   MQ_BYTES_MAX,   MQ_BYTES_MAX },	\
 	[RLIMIT_NICE]		= { 0, 0 },				\
 	[RLIMIT_RTPRIO]		= { 0, 0 },				\
+	[RLIMIT_RTTIME]		= {  RLIM_INFINITY,  RLIM_INFINITY },	\
 }
 
 #endif	/* __KERNEL__ */

include/asm-x86/thread_info_32.h

@@ -132,6 +132,7 @@ static inline struct thread_info *current_thread_info(void)
 #define TIF_SYSCALL_AUDIT	6	/* syscall auditing active */
 #define TIF_SECCOMP		7	/* secure computing */
 #define TIF_RESTORE_SIGMASK	8	/* restore signal mask in do_signal() */
+#define TIF_HRTICK_RESCHED	9	/* reprogram hrtick timer */
 #define TIF_MEMDIE		16
 #define TIF_DEBUG		17	/* uses debug registers */
 #define TIF_IO_BITMAP		18	/* uses I/O bitmap */
@@ -147,6 +148,7 @@ static inline struct thread_info *current_thread_info(void)
 #define _TIF_SYSCALL_AUDIT	(1<<TIF_SYSCALL_AUDIT)
 #define _TIF_SECCOMP		(1<<TIF_SECCOMP)
 #define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
+#define _TIF_HRTICK_RESCHED	(1<<TIF_HRTICK_RESCHED)
 #define _TIF_DEBUG		(1<<TIF_DEBUG)
 #define _TIF_IO_BITMAP		(1<<TIF_IO_BITMAP)
 #define _TIF_FREEZE		(1<<TIF_FREEZE)

include/asm-x86/thread_info_64.h

@@ -115,6 +115,7 @@ static inline struct thread_info *stack_thread_info(void)
 #define TIF_SECCOMP		8	/* secure computing */
 #define TIF_RESTORE_SIGMASK	9	/* restore signal mask in do_signal */
 #define TIF_MCE_NOTIFY		10	/* notify userspace of an MCE */
+#define TIF_HRTICK_RESCHED	11	/* reprogram hrtick timer */
 /* 16 free */
 #define TIF_IA32		17	/* 32bit process */ 
 #define TIF_FORK		18	/* ret_from_fork */
@@ -133,6 +134,7 @@ static inline struct thread_info *stack_thread_info(void)
 #define _TIF_SECCOMP		(1<<TIF_SECCOMP)
 #define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
 #define _TIF_MCE_NOTIFY		(1<<TIF_MCE_NOTIFY)
+#define _TIF_HRTICK_RESCHED	(1<<TIF_HRTICK_RESCHED)
 #define _TIF_IA32		(1<<TIF_IA32)
 #define _TIF_FORK		(1<<TIF_FORK)
 #define _TIF_ABI_PENDING	(1<<TIF_ABI_PENDING)
@@ -146,6 +148,9 @@ static inline struct thread_info *stack_thread_info(void)
 /* work to do on any return to user space */
 #define _TIF_ALLWORK_MASK (0x0000FFFF & ~_TIF_SECCOMP)
 
+#define _TIF_DO_NOTIFY_MASK \
+	(_TIF_SIGPENDING|_TIF_SINGLESTEP|_TIF_MCE_NOTIFY|_TIF_HRTICK_RESCHED)
+
 /* flags to check in __switch_to() */
 #define _TIF_WORK_CTXSW (_TIF_DEBUG|_TIF_IO_BITMAP)
 

include/linux/cpu.h

@@ -71,18 +71,27 @@ static inline void unregister_cpu_notifier(struct notifier_block *nb)
 
 int cpu_up(unsigned int cpu);
 
+extern void cpu_hotplug_init(void);
+
 #else
 
 static inline int register_cpu_notifier(struct notifier_block *nb)
 {
 	return 0;
 }
+
 static inline void unregister_cpu_notifier(struct notifier_block *nb)
 {
 }
 
+static inline void cpu_hotplug_init(void)
+{
+}
+
 #endif /* CONFIG_SMP */
 extern struct sysdev_class cpu_sysdev_class;
+extern void cpu_maps_update_begin(void);
+extern void cpu_maps_update_done(void);
 
 #ifdef CONFIG_HOTPLUG_CPU
 /* Stop CPUs going up and down. */
@@ -97,8 +106,8 @@ static inline void cpuhotplug_mutex_unlock(struct mutex *cpu_hp_mutex)
 	mutex_unlock(cpu_hp_mutex);
 }
 
-extern void lock_cpu_hotplug(void);
-extern void unlock_cpu_hotplug(void);
+extern void get_online_cpus(void);
+extern void put_online_cpus(void);
 #define hotcpu_notifier(fn, pri) {				\
 	static struct notifier_block fn##_nb =			\
 		{ .notifier_call = fn, .priority = pri };	\
@@ -115,8 +124,8 @@ static inline void cpuhotplug_mutex_lock(struct mutex *cpu_hp_mutex)
 static inline void cpuhotplug_mutex_unlock(struct mutex *cpu_hp_mutex)
 { }
 
-#define lock_cpu_hotplug()	do { } while (0)
-#define unlock_cpu_hotplug()	do { } while (0)
+#define get_online_cpus()	do { } while (0)
+#define put_online_cpus()	do { } while (0)
 #define hotcpu_notifier(fn, pri)	do { (void)(fn); } while (0)
 /* These aren't inline functions due to a GCC bug. */
 #define register_hotcpu_notifier(nb)	({ (void)(nb); 0; })

include/linux/debug_locks.h

@@ -47,6 +47,7 @@ struct task_struct;
 
 #ifdef CONFIG_LOCKDEP
 extern void debug_show_all_locks(void);
+extern void __debug_show_held_locks(struct task_struct *task);
 extern void debug_show_held_locks(struct task_struct *task);
 extern void debug_check_no_locks_freed(const void *from, unsigned long len);
 extern void debug_check_no_locks_held(struct task_struct *task);
@@ -55,6 +56,10 @@ static inline void debug_show_all_locks(void)
 {
 }
 
+static inline void __debug_show_held_locks(struct task_struct *task)
+{
+}
+
 static inline void debug_show_held_locks(struct task_struct *task)
 {
 }

include/linux/futex.h

@@ -1,8 +1,12 @@
 #ifndef _LINUX_FUTEX_H
 #define _LINUX_FUTEX_H
 
-#include <linux/sched.h>
+#include <linux/compiler.h>
+#include <linux/types.h>
 
+struct inode;
+struct mm_struct;
+struct task_struct;
 union ktime;
 
 /* Second argument to futex syscall */

include/linux/hardirq.h

@@ -72,11 +72,7 @@
 #define in_softirq()		(softirq_count())
 #define in_interrupt()		(irq_count())
 
-#if defined(CONFIG_PREEMPT) && !defined(CONFIG_PREEMPT_BKL)
-# define in_atomic()	((preempt_count() & ~PREEMPT_ACTIVE) != kernel_locked())
-#else
-# define in_atomic()	((preempt_count() & ~PREEMPT_ACTIVE) != 0)
-#endif
+#define in_atomic()		((preempt_count() & ~PREEMPT_ACTIVE) != 0)
 
 #ifdef CONFIG_PREEMPT
 # define PREEMPT_CHECK_OFFSET 1

include/linux/hrtimer.h

@@ -115,10 +115,8 @@ struct hrtimer {
 	enum hrtimer_restart		(*function)(struct hrtimer *);
 	struct hrtimer_clock_base	*base;
 	unsigned long			state;
-#ifdef CONFIG_HIGH_RES_TIMERS
 	enum hrtimer_cb_mode		cb_mode;
 	struct list_head		cb_entry;
-#endif
 #ifdef CONFIG_TIMER_STATS
 	void				*start_site;
 	char				start_comm[16];
@@ -194,10 +192,10 @@ struct hrtimer_cpu_base {
 	spinlock_t			lock;
 	struct lock_class_key		lock_key;
 	struct hrtimer_clock_base	clock_base[HRTIMER_MAX_CLOCK_BASES];
+	struct list_head		cb_pending;
 #ifdef CONFIG_HIGH_RES_TIMERS
 	ktime_t				expires_next;
 	int				hres_active;
-	struct list_head		cb_pending;
 	unsigned long			nr_events;
 #endif
 };
@@ -217,6 +215,11 @@ static inline ktime_t hrtimer_cb_get_time(struct hrtimer *timer)
 	return timer->base->get_time();
 }
 
+static inline int hrtimer_is_hres_active(struct hrtimer *timer)
+{
+	return timer->base->cpu_base->hres_active;
+}
+
 /*
  * The resolution of the clocks. The resolution value is returned in
  * the clock_getres() system call to give application programmers an
@@ -248,6 +251,10 @@ static inline ktime_t hrtimer_cb_get_time(struct hrtimer *timer)
 	return timer->base->softirq_time;
 }
 
+static inline int hrtimer_is_hres_active(struct hrtimer *timer)
+{
+	return 0;
+}
 #endif
 
 extern ktime_t ktime_get(void);
@@ -310,6 +317,7 @@ extern void hrtimer_init_sleeper(struct hrtimer_sleeper *sl,
 
 /* Soft interrupt function to run the hrtimer queues: */
 extern void hrtimer_run_queues(void);
+extern void hrtimer_run_pending(void);
 
 /* Bootup initialization: */
 extern void __init hrtimers_init(void);

include/linux/init_task.h

@@ -132,9 +132,12 @@ extern struct group_info init_groups;
 	.cpus_allowed	= CPU_MASK_ALL,					\
 	.mm		= NULL,						\
 	.active_mm	= &init_mm,					\
-	.run_list	= LIST_HEAD_INIT(tsk.run_list),			\
+	.rt		= {						\
+		.run_list	= LIST_HEAD_INIT(tsk.rt.run_list),	\
+		.time_slice	= HZ, 					\
+		.nr_cpus_allowed = NR_CPUS,				\
+	},								\
 	.ioprio		= 0,						\
-	.time_slice	= HZ,						\
 	.tasks		= LIST_HEAD_INIT(tsk.tasks),			\
 	.ptrace_children= LIST_HEAD_INIT(tsk.ptrace_children),		\
 	.ptrace_list	= LIST_HEAD_INIT(tsk.ptrace_list),		\

include/linux/interrupt.h

@@ -256,6 +256,7 @@ enum
 #ifdef CONFIG_HIGH_RES_TIMERS
 	HRTIMER_SOFTIRQ,
 #endif
+	RCU_SOFTIRQ, 	/* Preferable RCU should always be the last softirq */
 };
 
 /* softirq mask and active fields moved to irq_cpustat_t in

include/linux/jiffies.h

@@ -29,6 +29,12 @@
 # define SHIFT_HZ	9
 #elif HZ >= 768 && HZ < 1536
 # define SHIFT_HZ	10
+#elif HZ >= 1536 && HZ < 3072
+# define SHIFT_HZ	11
+#elif HZ >= 3072 && HZ < 6144
+# define SHIFT_HZ	12
+#elif HZ >= 6144 && HZ < 12288
+# define SHIFT_HZ	13
 #else
 # error You lose.
 #endif

include/linux/kernel.h

@@ -105,8 +105,8 @@ struct user;
  * supposed to.
  */
 #ifdef CONFIG_PREEMPT_VOLUNTARY
-extern int cond_resched(void);
-# define might_resched() cond_resched()
+extern int _cond_resched(void);
+# define might_resched() _cond_resched()
 #else
 # define might_resched() do { } while (0)
 #endif

include/linux/latencytop.h

@@ -0,0 +1,44 @@
+/*
+ * latencytop.h: Infrastructure for displaying latency
+ *
+ * (C) Copyright 2008 Intel Corporation
+ * Author: Arjan van de Ven <arjan@linux.intel.com>
+ *
+ */
+
+#ifndef _INCLUDE_GUARD_LATENCYTOP_H_
+#define _INCLUDE_GUARD_LATENCYTOP_H_
+
+#ifdef CONFIG_LATENCYTOP
+
+#define LT_SAVECOUNT		32
+#define LT_BACKTRACEDEPTH	12
+
+struct latency_record {
+	unsigned long	backtrace[LT_BACKTRACEDEPTH];
+	unsigned int	count;
+	unsigned long	time;
+	unsigned long	max;
+};
+
+
+struct task_struct;
+
+void account_scheduler_latency(struct task_struct *task, int usecs, int inter);
+
+void clear_all_latency_tracing(struct task_struct *p);
+
+#else
+
+static inline void
+account_scheduler_latency(struct task_struct *task, int usecs, int inter)
+{
+}
+
+static inline void clear_all_latency_tracing(struct task_struct *p)
+{
+}
+
+#endif
+
+#endif

include/linux/notifier.h

@@ -207,9 +207,7 @@ static inline int notifier_to_errno(int ret)
 #define CPU_DOWN_PREPARE	0x0005 /* CPU (unsigned)v going down */
 #define CPU_DOWN_FAILED		0x0006 /* CPU (unsigned)v NOT going down */
 #define CPU_DEAD		0x0007 /* CPU (unsigned)v dead */
-#define CPU_LOCK_ACQUIRE	0x0008 /* Acquire all hotcpu locks */
-#define CPU_LOCK_RELEASE	0x0009 /* Release all hotcpu locks */
-#define CPU_DYING		0x000A /* CPU (unsigned)v not running any task,
+#define CPU_DYING		0x0008 /* CPU (unsigned)v not running any task,
 				        * not handling interrupts, soon dead */
 
 /* Used for CPU hotplug events occuring while tasks are frozen due to a suspend

include/linux/rcuclassic.h

@@ -0,0 +1,164 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion (classic version)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright IBM Corporation, 2001
+ *
+ * Author: Dipankar Sarma <dipankar@in.ibm.com>
+ *
+ * Based on the original work by Paul McKenney <paulmck@us.ibm.com>
+ * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
+ * Papers:
+ * http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf
+ * http://lse.sourceforge.net/locking/rclock_OLS.2001.05.01c.sc.pdf (OLS2001)
+ *
+ * For detailed explanation of Read-Copy Update mechanism see -
+ * 		Documentation/RCU
+ *
+ */
+
+#ifndef __LINUX_RCUCLASSIC_H
+#define __LINUX_RCUCLASSIC_H
+
+#ifdef __KERNEL__
+
+#include <linux/cache.h>
+#include <linux/spinlock.h>
+#include <linux/threads.h>
+#include <linux/percpu.h>
+#include <linux/cpumask.h>
+#include <linux/seqlock.h>
+
+
+/* Global control variables for rcupdate callback mechanism. */
+struct rcu_ctrlblk {
+	long	cur;		/* Current batch number.                      */
+	long	completed;	/* Number of the last completed batch         */
+	int	next_pending;	/* Is the next batch already waiting?         */
+
+	int	signaled;
+
+	spinlock_t	lock	____cacheline_internodealigned_in_smp;
+	cpumask_t	cpumask; /* CPUs that need to switch in order    */
+				 /* for current batch to proceed.        */
+} ____cacheline_internodealigned_in_smp;
+
+/* Is batch a before batch b ? */
+static inline int rcu_batch_before(long a, long b)
+{
+	return (a - b) < 0;
+}
+
+/* Is batch a after batch b ? */
+static inline int rcu_batch_after(long a, long b)
+{
+	return (a - b) > 0;
+}
+
+/*
+ * Per-CPU data for Read-Copy UPdate.
+ * nxtlist - new callbacks are added here
+ * curlist - current batch for which quiescent cycle started if any
+ */
+struct rcu_data {
+	/* 1) quiescent state handling : */
+	long		quiescbatch;     /* Batch # for grace period */
+	int		passed_quiesc;	 /* User-mode/idle loop etc. */
+	int		qs_pending;	 /* core waits for quiesc state */
+
+	/* 2) batch handling */
+	long  	       	batch;           /* Batch # for current RCU batch */
+	struct rcu_head *nxtlist;
+	struct rcu_head **nxttail;
+	long            qlen; 	 	 /* # of queued callbacks */
+	struct rcu_head *curlist;
+	struct rcu_head **curtail;
+	struct rcu_head *donelist;
+	struct rcu_head **donetail;
+	long		blimit;		 /* Upper limit on a processed batch */
+	int cpu;
+	struct rcu_head barrier;
+};
+
+DECLARE_PER_CPU(struct rcu_data, rcu_data);
+DECLARE_PER_CPU(struct rcu_data, rcu_bh_data);
+
+/*
+ * Increment the quiescent state counter.
+ * The counter is a bit degenerated: We do not need to know
+ * how many quiescent states passed, just if there was at least
+ * one since the start of the grace period. Thus just a flag.
+ */
+static inline void rcu_qsctr_inc(int cpu)
+{
+	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
+	rdp->passed_quiesc = 1;
+}
+static inline void rcu_bh_qsctr_inc(int cpu)
+{
+	struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu);
+	rdp->passed_quiesc = 1;
+}
+
+extern int rcu_pending(int cpu);
+extern int rcu_needs_cpu(int cpu);
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+extern struct lockdep_map rcu_lock_map;
+# define rcu_read_acquire()	\
+			lock_acquire(&rcu_lock_map, 0, 0, 2, 1, _THIS_IP_)
+# define rcu_read_release()	lock_release(&rcu_lock_map, 1, _THIS_IP_)
+#else
+# define rcu_read_acquire()	do { } while (0)
+# define rcu_read_release()	do { } while (0)
+#endif
+
+#define __rcu_read_lock() \
+	do { \
+		preempt_disable(); \
+		__acquire(RCU); \
+		rcu_read_acquire(); \
+	} while (0)
+#define __rcu_read_unlock() \
+	do { \
+		rcu_read_release(); \
+		__release(RCU); \
+		preempt_enable(); \
+	} while (0)
+#define __rcu_read_lock_bh() \
+	do { \
+		local_bh_disable(); \
+		__acquire(RCU_BH); \
+		rcu_read_acquire(); \
+	} while (0)
+#define __rcu_read_unlock_bh() \
+	do { \
+		rcu_read_release(); \
+		__release(RCU_BH); \
+		local_bh_enable(); \
+	} while (0)
+
+#define __synchronize_sched() synchronize_rcu()
+
+extern void __rcu_init(void);
+extern void rcu_check_callbacks(int cpu, int user);
+extern void rcu_restart_cpu(int cpu);
+
+extern long rcu_batches_completed(void);
+extern long rcu_batches_completed_bh(void);
+
+#endif /* __KERNEL__ */
+#endif /* __LINUX_RCUCLASSIC_H */

include/linux/rcupdate.h

@@ -15,7 +15,7 @@
  * along with this program; if not, write to the Free Software
  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  *
- * Copyright (C) IBM Corporation, 2001
+ * Copyright IBM Corporation, 2001
  *
  * Author: Dipankar Sarma <dipankar@in.ibm.com>
  * 
@@ -53,96 +53,18 @@ struct rcu_head {
 	void (*func)(struct rcu_head *head);
 };
 
+#ifdef CONFIG_CLASSIC_RCU
+#include <linux/rcuclassic.h>
+#else /* #ifdef CONFIG_CLASSIC_RCU */
+#include <linux/rcupreempt.h>
+#endif /* #else #ifdef CONFIG_CLASSIC_RCU */
+
 #define RCU_HEAD_INIT 	{ .next = NULL, .func = NULL }
 #define RCU_HEAD(head) struct rcu_head head = RCU_HEAD_INIT
 #define INIT_RCU_HEAD(ptr) do { \
        (ptr)->next = NULL; (ptr)->func = NULL; \
 } while (0)
 
-
-
-/* Global control variables for rcupdate callback mechanism. */
-struct rcu_ctrlblk {
-	long	cur;		/* Current batch number.                      */
-	long	completed;	/* Number of the last completed batch         */
-	int	next_pending;	/* Is the next batch already waiting?         */
-
-	int	signaled;
-
-	spinlock_t	lock	____cacheline_internodealigned_in_smp;
-	cpumask_t	cpumask; /* CPUs that need to switch in order    */
-	                         /* for current batch to proceed.        */
-} ____cacheline_internodealigned_in_smp;
-
-/* Is batch a before batch b ? */
-static inline int rcu_batch_before(long a, long b)
-{
-        return (a - b) < 0;
-}
-
-/* Is batch a after batch b ? */
-static inline int rcu_batch_after(long a, long b)
-{
-        return (a - b) > 0;
-}
-
-/*
- * Per-CPU data for Read-Copy UPdate.
- * nxtlist - new callbacks are added here
- * curlist - current batch for which quiescent cycle started if any
- */
-struct rcu_data {
-	/* 1) quiescent state handling : */
-	long		quiescbatch;     /* Batch # for grace period */
-	int		passed_quiesc;	 /* User-mode/idle loop etc. */
-	int		qs_pending;	 /* core waits for quiesc state */
-
-	/* 2) batch handling */
-	long  	       	batch;           /* Batch # for current RCU batch */
-	struct rcu_head *nxtlist;
-	struct rcu_head **nxttail;
-	long            qlen; 	 	 /* # of queued callbacks */
-	struct rcu_head *curlist;
-	struct rcu_head **curtail;
-	struct rcu_head *donelist;
-	struct rcu_head **donetail;
-	long		blimit;		 /* Upper limit on a processed batch */
-	int cpu;
-	struct rcu_head barrier;
-};
-
-DECLARE_PER_CPU(struct rcu_data, rcu_data);
-DECLARE_PER_CPU(struct rcu_data, rcu_bh_data);
-
-/*
- * Increment the quiescent state counter.
- * The counter is a bit degenerated: We do not need to know
- * how many quiescent states passed, just if there was at least
- * one since the start of the grace period. Thus just a flag.
- */
-static inline void rcu_qsctr_inc(int cpu)
-{
-	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
-	rdp->passed_quiesc = 1;
-}
-static inline void rcu_bh_qsctr_inc(int cpu)
-{
-	struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu);
-	rdp->passed_quiesc = 1;
-}
-
-extern int rcu_pending(int cpu);
-extern int rcu_needs_cpu(int cpu);
-
-#ifdef CONFIG_DEBUG_LOCK_ALLOC
-extern struct lockdep_map rcu_lock_map;
-# define rcu_read_acquire()	lock_acquire(&rcu_lock_map, 0, 0, 2, 1, _THIS_IP_)
-# define rcu_read_release()	lock_release(&rcu_lock_map, 1, _THIS_IP_)
-#else
-# define rcu_read_acquire()	do { } while (0)
-# define rcu_read_release()	do { } while (0)
-#endif
-
 /**
  * rcu_read_lock - mark the beginning of an RCU read-side critical section.
  *
@@ -172,24 +94,13 @@ extern struct lockdep_map rcu_lock_map;
  *
  * It is illegal to block while in an RCU read-side critical section.
  */
-#define rcu_read_lock() \
-	do { \
-		preempt_disable(); \
-		__acquire(RCU); \
-		rcu_read_acquire(); \
-	} while(0)
+#define rcu_read_lock() __rcu_read_lock()
 
 /**
  * rcu_read_unlock - marks the end of an RCU read-side critical section.
  *
  * See rcu_read_lock() for more information.
  */
-#define rcu_read_unlock() \
-	do { \
-		rcu_read_release(); \
-		__release(RCU); \
-		preempt_enable(); \
-	} while(0)
 
 /*
  * So where is rcu_write_lock()?  It does not exist, as there is no
@@ -200,6 +111,7 @@ extern struct lockdep_map rcu_lock_map;
  * used as well.  RCU does not care how the writers keep out of each
  * others' way, as long as they do so.
  */
+#define rcu_read_unlock() __rcu_read_unlock()
 
 /**
  * rcu_read_lock_bh - mark the beginning of a softirq-only RCU critical section
@@ -212,24 +124,14 @@ extern struct lockdep_map rcu_lock_map;
  * can use just rcu_read_lock().
  *
  */
-#define rcu_read_lock_bh() \
-	do { \
-		local_bh_disable(); \
-		__acquire(RCU_BH); \
-		rcu_read_acquire(); \
-	} while(0)
+#define rcu_read_lock_bh() __rcu_read_lock_bh()
 
 /*
  * rcu_read_unlock_bh - marks the end of a softirq-only RCU critical section
  *
  * See rcu_read_lock_bh() for more information.
  */
-#define rcu_read_unlock_bh() \
-	do { \
-		rcu_read_release(); \
-		__release(RCU_BH); \
-		local_bh_enable(); \
-	} while(0)
+#define rcu_read_unlock_bh() __rcu_read_unlock_bh()
 
 /*
  * Prevent the compiler from merging or refetching accesses.  The compiler
@@ -293,21 +195,52 @@ extern struct lockdep_map rcu_lock_map;
  * In "classic RCU", these two guarantees happen to be one and
  * the same, but can differ in realtime RCU implementations.
  */
-#define synchronize_sched() synchronize_rcu()
+#define synchronize_sched() __synchronize_sched()
 
-extern void rcu_init(void);
-extern void rcu_check_callbacks(int cpu, int user);
-extern void rcu_restart_cpu(int cpu);
-extern long rcu_batches_completed(void);
-extern long rcu_batches_completed_bh(void);
+/**
+ * call_rcu - Queue an RCU callback for invocation after a grace period.
+ * @head: structure to be used for queueing the RCU updates.
+ * @func: actual update function to be invoked after the grace period
+ *
+ * The update function will be invoked some time after a full grace
+ * period elapses, in other words after all currently executing RCU
+ * read-side critical sections have completed.  RCU read-side critical
+ * sections are delimited by rcu_read_lock() and rcu_read_unlock(),
+ * and may be nested.
+ */
+extern void call_rcu(struct rcu_head *head,
+			      void (*func)(struct rcu_head *head));
 
-/* Exported interfaces */
-extern void FASTCALL(call_rcu(struct rcu_head *head, 
-				void (*func)(struct rcu_head *head)));
-extern void FASTCALL(call_rcu_bh(struct rcu_head *head,
-				void (*func)(struct rcu_head *head)));
+/**
+ * call_rcu_bh - Queue an RCU for invocation after a quicker grace period.
+ * @head: structure to be used for queueing the RCU updates.
+ * @func: actual update function to be invoked after the grace period
+ *
+ * The update function will be invoked some time after a full grace
+ * period elapses, in other words after all currently executing RCU
+ * read-side critical sections have completed. call_rcu_bh() assumes
+ * that the read-side critical sections end on completion of a softirq
+ * handler. This means that read-side critical sections in process
+ * context must not be interrupted by softirqs. This interface is to be
+ * used when most of the read-side critical sections are in softirq context.
+ * RCU read-side critical sections are delimited by :
+ *  - rcu_read_lock() and  rcu_read_unlock(), if in interrupt context.
+ *  OR
+ *  - rcu_read_lock_bh() and rcu_read_unlock_bh(), if in process context.
+ *  These may be nested.
+ */
+extern void call_rcu_bh(struct rcu_head *head,
+			void (*func)(struct rcu_head *head));
+
+/* Exported common interfaces */
 extern void synchronize_rcu(void);
 extern void rcu_barrier(void);
+extern long rcu_batches_completed(void);
+extern long rcu_batches_completed_bh(void);
+
+/* Internal to kernel */
+extern void rcu_init(void);
+extern int rcu_needs_cpu(int cpu);
 
 #endif /* __KERNEL__ */
 #endif /* __LINUX_RCUPDATE_H */

include/linux/rcupreempt.h

@@ -0,0 +1,86 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion (RT implementation)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2006
+ *
+ * Author:  Paul McKenney <paulmck@us.ibm.com>
+ *
+ * Based on the original work by Paul McKenney <paul.mckenney@us.ibm.com>
+ * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
+ * Papers:
+ * http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf
+ * http://lse.sourceforge.net/locking/rclock_OLS.2001.05.01c.sc.pdf (OLS2001)
+ *
+ * For detailed explanation of Read-Copy Update mechanism see -
+ * 		Documentation/RCU
+ *
+ */
+
+#ifndef __LINUX_RCUPREEMPT_H
+#define __LINUX_RCUPREEMPT_H
+
+#ifdef __KERNEL__
+
+#include <linux/cache.h>
+#include <linux/spinlock.h>
+#include <linux/threads.h>
+#include <linux/percpu.h>
+#include <linux/cpumask.h>
+#include <linux/seqlock.h>
+
+#define rcu_qsctr_inc(cpu)
+#define rcu_bh_qsctr_inc(cpu)
+#define call_rcu_bh(head, rcu) call_rcu(head, rcu)
+
+extern void __rcu_read_lock(void);
+extern void __rcu_read_unlock(void);
+extern int rcu_pending(int cpu);
+extern int rcu_needs_cpu(int cpu);
+
+#define __rcu_read_lock_bh()	{ rcu_read_lock(); local_bh_disable(); }
+#define __rcu_read_unlock_bh()	{ local_bh_enable(); rcu_read_unlock(); }
+
+extern void __synchronize_sched(void);
+
+extern void __rcu_init(void);
+extern void rcu_check_callbacks(int cpu, int user);
+extern void rcu_restart_cpu(int cpu);
+extern long rcu_batches_completed(void);
+
+/*
+ * Return the number of RCU batches processed thus far. Useful for debug
+ * and statistic. The _bh variant is identifcal to straight RCU
+ */
+static inline long rcu_batches_completed_bh(void)
+{
+	return rcu_batches_completed();
+}
+
+#ifdef CONFIG_RCU_TRACE
+struct rcupreempt_trace;
+extern long *rcupreempt_flipctr(int cpu);
+extern long rcupreempt_data_completed(void);
+extern int rcupreempt_flip_flag(int cpu);
+extern int rcupreempt_mb_flag(int cpu);
+extern char *rcupreempt_try_flip_state_name(void);
+extern struct rcupreempt_trace *rcupreempt_trace_cpu(int cpu);
+#endif
+
+struct softirq_action;
+
+#endif /* __KERNEL__ */
+#endif /* __LINUX_RCUPREEMPT_H */

include/linux/rcupreempt_trace.h

@@ -0,0 +1,99 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion (RT implementation)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2006
+ *
+ * Author:  Paul McKenney <paulmck@us.ibm.com>
+ *
+ * Based on the original work by Paul McKenney <paul.mckenney@us.ibm.com>
+ * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
+ * Papers:
+ * http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf
+ * http://lse.sourceforge.net/locking/rclock_OLS.2001.05.01c.sc.pdf (OLS2001)
+ *
+ * For detailed explanation of the Preemptible Read-Copy Update mechanism see -
+ * 		 http://lwn.net/Articles/253651/
+ */
+
+#ifndef __LINUX_RCUPREEMPT_TRACE_H
+#define __LINUX_RCUPREEMPT_TRACE_H
+
+#ifdef __KERNEL__
+#include <linux/types.h>
+#include <linux/kernel.h>
+
+#include <asm/atomic.h>
+
+/*
+ * PREEMPT_RCU data structures.
+ */
+
+struct rcupreempt_trace {
+	long		next_length;
+	long		next_add;
+	long		wait_length;
+	long		wait_add;
+	long		done_length;
+	long		done_add;
+	long		done_remove;
+	atomic_t	done_invoked;
+	long		rcu_check_callbacks;
+	atomic_t	rcu_try_flip_1;
+	atomic_t	rcu_try_flip_e1;
+	long		rcu_try_flip_i1;
+	long		rcu_try_flip_ie1;
+	long		rcu_try_flip_g1;
+	long		rcu_try_flip_a1;
+	long		rcu_try_flip_ae1;
+	long		rcu_try_flip_a2;
+	long		rcu_try_flip_z1;
+	long		rcu_try_flip_ze1;
+	long		rcu_try_flip_z2;
+	long		rcu_try_flip_m1;
+	long		rcu_try_flip_me1;
+	long		rcu_try_flip_m2;
+};
+
+#ifdef CONFIG_RCU_TRACE
+#define RCU_TRACE(fn, arg) 	fn(arg);
+#else
+#define RCU_TRACE(fn, arg)
+#endif
+
+extern void rcupreempt_trace_move2done(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_move2wait(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_e1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_i1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_ie1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_g1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_a1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_ae1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_a2(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_z1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_ze1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_z2(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_m1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_me1(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_try_flip_m2(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_check_callbacks(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_done_remove(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_invoke(struct rcupreempt_trace *trace);
+extern void rcupreempt_trace_next_add(struct rcupreempt_trace *trace);
+
+#endif /* __KERNEL__ */
+#endif /* __LINUX_RCUPREEMPT_TRACE_H */

include/linux/sched.h

@@ -78,7 +78,6 @@ struct sched_param {
 #include <linux/proportions.h>
 #include <linux/seccomp.h>
 #include <linux/rcupdate.h>
-#include <linux/futex.h>
 #include <linux/rtmutex.h>
 
 #include <linux/time.h>
@@ -88,11 +87,13 @@ struct sched_param {
 #include <linux/hrtimer.h>
 #include <linux/task_io_accounting.h>
 #include <linux/kobject.h>
+#include <linux/latencytop.h>
 
 #include <asm/processor.h>
 
 struct exec_domain;
 struct futex_pi_state;
+struct robust_list_head;
 struct bio;
 
 /*
@@ -230,6 +231,8 @@ static inline int select_nohz_load_balancer(int cpu)
 }
 #endif
 
+extern unsigned long rt_needs_cpu(int cpu);
+
 /*
  * Only dump TASK_* tasks. (0 for all tasks)
  */
@@ -257,13 +260,19 @@ extern void trap_init(void);
 extern void account_process_tick(struct task_struct *task, int user);
 extern void update_process_times(int user);
 extern void scheduler_tick(void);
+extern void hrtick_resched(void);
+
+extern void sched_show_task(struct task_struct *p);
 
 #ifdef CONFIG_DETECT_SOFTLOCKUP
 extern void softlockup_tick(void);
 extern void spawn_softlockup_task(void);
 extern void touch_softlockup_watchdog(void);
 extern void touch_all_softlockup_watchdogs(void);
-extern int softlockup_thresh;
+extern unsigned long  softlockup_thresh;
+extern unsigned long sysctl_hung_task_check_count;
+extern unsigned long sysctl_hung_task_timeout_secs;
+extern unsigned long sysctl_hung_task_warnings;
 #else
 static inline void softlockup_tick(void)
 {
@@ -822,6 +831,7 @@ struct sched_class {
 	void (*enqueue_task) (struct rq *rq, struct task_struct *p, int wakeup);
 	void (*dequeue_task) (struct rq *rq, struct task_struct *p, int sleep);
 	void (*yield_task) (struct rq *rq);
+	int  (*select_task_rq)(struct task_struct *p, int sync);
 
 	void (*check_preempt_curr) (struct rq *rq, struct task_struct *p);
 
@@ -837,11 +847,25 @@ struct sched_class {
 	int (*move_one_task) (struct rq *this_rq, int this_cpu,
 			      struct rq *busiest, struct sched_domain *sd,
 			      enum cpu_idle_type idle);
+	void (*pre_schedule) (struct rq *this_rq, struct task_struct *task);
+	void (*post_schedule) (struct rq *this_rq);
+	void (*task_wake_up) (struct rq *this_rq, struct task_struct *task);
 #endif
 
 	void (*set_curr_task) (struct rq *rq);
-	void (*task_tick) (struct rq *rq, struct task_struct *p);
+	void (*task_tick) (struct rq *rq, struct task_struct *p, int queued);
 	void (*task_new) (struct rq *rq, struct task_struct *p);
+	void (*set_cpus_allowed)(struct task_struct *p, cpumask_t *newmask);
+
+	void (*join_domain)(struct rq *rq);
+	void (*leave_domain)(struct rq *rq);
+
+	void (*switched_from) (struct rq *this_rq, struct task_struct *task,
+			       int running);
+	void (*switched_to) (struct rq *this_rq, struct task_struct *task,
+			     int running);
+	void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
+			     int oldprio, int running);
 };
 
 struct load_weight {
@@ -871,6 +895,8 @@ struct sched_entity {
 #ifdef CONFIG_SCHEDSTATS
 	u64			wait_start;
 	u64			wait_max;
+	u64			wait_count;
+	u64			wait_sum;
 
 	u64			sleep_start;
 	u64			sleep_max;
@@ -909,6 +935,21 @@ struct sched_entity {
 #endif
 };
 
+struct sched_rt_entity {
+	struct list_head run_list;
+	unsigned int time_slice;
+	unsigned long timeout;
+	int nr_cpus_allowed;
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	struct sched_rt_entity	*parent;
+	/* rq on which this entity is (to be) queued: */
+	struct rt_rq		*rt_rq;
+	/* rq "owned" by this entity/group: */
+	struct rt_rq		*my_q;
+#endif
+};
+
 struct task_struct {
 	volatile long state;	/* -1 unrunnable, 0 runnable, >0 stopped */
 	void *stack;
@@ -925,9 +966,9 @@ struct task_struct {
 #endif
 
 	int prio, static_prio, normal_prio;
-	struct list_head run_list;
 	const struct sched_class *sched_class;
 	struct sched_entity se;
+	struct sched_rt_entity rt;
 
 #ifdef CONFIG_PREEMPT_NOTIFIERS
 	/* list of struct preempt_notifier: */
@@ -951,7 +992,11 @@ struct task_struct {
 
 	unsigned int policy;
 	cpumask_t cpus_allowed;
-	unsigned int time_slice;
+
+#ifdef CONFIG_PREEMPT_RCU
+	int rcu_read_lock_nesting;
+	int rcu_flipctr_idx;
+#endif /* #ifdef CONFIG_PREEMPT_RCU */
 
 #if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
 	struct sched_info sched_info;
@@ -1041,6 +1086,11 @@ struct task_struct {
 /* ipc stuff */
 	struct sysv_sem sysvsem;
 #endif
+#ifdef CONFIG_DETECT_SOFTLOCKUP
+/* hung task detection */
+	unsigned long last_switch_timestamp;
+	unsigned long last_switch_count;
+#endif
 /* CPU-specific state of this task */
 	struct thread_struct thread;
 /* filesystem information */
@@ -1173,6 +1223,10 @@ struct task_struct {
 	int make_it_fail;
 #endif
 	struct prop_local_single dirties;
+#ifdef CONFIG_LATENCYTOP
+	int latency_record_count;
+	struct latency_record latency_record[LT_SAVECOUNT];
+#endif
 };
 
 /*
@@ -1453,6 +1507,12 @@ extern unsigned int sysctl_sched_child_runs_first;
 extern unsigned int sysctl_sched_features;
 extern unsigned int sysctl_sched_migration_cost;
 extern unsigned int sysctl_sched_nr_migrate;
+extern unsigned int sysctl_sched_rt_period;
+extern unsigned int sysctl_sched_rt_ratio;
+#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP)
+extern unsigned int sysctl_sched_min_bal_int_shares;
+extern unsigned int sysctl_sched_max_bal_int_shares;
+#endif
 
 int sched_nr_latency_handler(struct ctl_table *table, int write,
 		struct file *file, void __user *buffer, size_t *length,
@@ -1845,7 +1905,18 @@ static inline int need_resched(void)
  * cond_resched_lock() will drop the spinlock before scheduling,
  * cond_resched_softirq() will enable bhs before scheduling.
  */
-extern int cond_resched(void);
+#ifdef CONFIG_PREEMPT
+static inline int cond_resched(void)
+{
+	return 0;
+}
+#else
+extern int _cond_resched(void);
+static inline int cond_resched(void)
+{
+	return _cond_resched();
+}
+#endif
 extern int cond_resched_lock(spinlock_t * lock);
 extern int cond_resched_softirq(void);
 

include/linux/smp_lock.h

@@ -17,22 +17,10 @@ extern void __lockfunc __release_kernel_lock(void);
 		__release_kernel_lock();	\
 } while (0)
 
-/*
- * Non-SMP kernels will never block on the kernel lock,
- * so we are better off returning a constant zero from
- * reacquire_kernel_lock() so that the compiler can see
- * it at compile-time.
- */
-#if defined(CONFIG_SMP) && !defined(CONFIG_PREEMPT_BKL)
-# define return_value_on_smp return
-#else
-# define return_value_on_smp
-#endif
-
 static inline int reacquire_kernel_lock(struct task_struct *task)
 {
 	if (unlikely(task->lock_depth >= 0))
-		return_value_on_smp __reacquire_kernel_lock();
+		return __reacquire_kernel_lock();
 	return 0;
 }
 

include/linux/stacktrace.h

@@ -9,10 +9,13 @@ struct stack_trace {
 };
 
 extern void save_stack_trace(struct stack_trace *trace);
+extern void save_stack_trace_tsk(struct task_struct *tsk,
+				struct stack_trace *trace);
 
 extern void print_stack_trace(struct stack_trace *trace, int spaces);
 #else
 # define save_stack_trace(trace)			do { } while (0)
+# define save_stack_trace_tsk(tsk, trace)		do { } while (0)
 # define print_stack_trace(trace, spaces)		do { } while (0)
 #endif
 

include/linux/topology.h

@@ -5,7 +5,7 @@
  *
  * Copyright (C) 2002, IBM Corp.
  *
- * All rights reserved.          
+ * All rights reserved.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -103,6 +103,7 @@
 	.forkexec_idx		= 0,			\
 	.flags			= SD_LOAD_BALANCE	\
 				| SD_BALANCE_NEWIDLE	\
+				| SD_BALANCE_FORK	\
 				| SD_BALANCE_EXEC	\
 				| SD_WAKE_AFFINE	\
 				| SD_WAKE_IDLE		\
@@ -134,6 +135,7 @@
 	.forkexec_idx		= 1,			\
 	.flags			= SD_LOAD_BALANCE	\
 				| SD_BALANCE_NEWIDLE	\
+				| SD_BALANCE_FORK	\
 				| SD_BALANCE_EXEC	\
 				| SD_WAKE_AFFINE	\
 				| SD_WAKE_IDLE		\
@@ -165,6 +167,7 @@
 	.forkexec_idx		= 1,			\
 	.flags			= SD_LOAD_BALANCE	\
 				| SD_BALANCE_NEWIDLE	\
+				| SD_BALANCE_FORK	\
 				| SD_BALANCE_EXEC	\
 				| SD_WAKE_AFFINE	\
 				| BALANCE_FOR_PKG_POWER,\

init/Kconfig

@@ -763,3 +763,31 @@ source "block/Kconfig"
 
 config PREEMPT_NOTIFIERS
 	bool
+
+choice
+	prompt "RCU implementation type:"
+	default CLASSIC_RCU
+
+config CLASSIC_RCU
+	bool "Classic RCU"
+	help
+	  This option selects the classic RCU implementation that is
+	  designed for best read-side performance on non-realtime
+	  systems.
+
+	  Say Y if you are unsure.
+
+config PREEMPT_RCU
+	bool "Preemptible RCU"
+	depends on PREEMPT
+	help
+	  This option reduces the latency of the kernel by making certain
+	  RCU sections preemptible. Normally RCU code is non-preemptible, if
+	  this option is selected then read-only RCU sections become
+	  preemptible. This helps latency, but may expose bugs due to
+	  now-naive assumptions about each RCU read-side critical section
+	  remaining on a given CPU through its execution.
+
+	  Say N if you are unsure.
+
+endchoice

init/main.c

@@ -607,6 +607,7 @@ asmlinkage void __init start_kernel(void)
 	vfs_caches_init_early();
 	cpuset_init_early();
 	mem_init();
+	cpu_hotplug_init();
 	kmem_cache_init();
 	setup_per_cpu_pageset();
 	numa_policy_init();

kernel/Kconfig.hz

@@ -54,3 +54,5 @@ config HZ
 	default 300 if HZ_300
 	default 1000 if HZ_1000
 
+config SCHED_HRTICK
+	def_bool HIGH_RES_TIMERS && X86

kernel/Kconfig.preempt

@@ -52,14 +52,13 @@ config PREEMPT
 
 endchoice
 
-config PREEMPT_BKL
-	bool "Preempt The Big Kernel Lock"
-	depends on SMP || PREEMPT
+config RCU_TRACE
+	bool "Enable tracing for RCU - currently stats in debugfs"
+	select DEBUG_FS
 	default y
 	help
-	  This option reduces the latency of the kernel by making the
-	  big kernel lock preemptible.
+	  This option provides tracing in RCU which presents stats
+	  in debugfs for debugging RCU implementation.
 
-	  Say Y here if you are building a kernel for a desktop system.
+	  Say Y here if you want to enable RCU tracing
 	  Say N if you are unsure.
-

kernel/Makefile

@@ -52,11 +52,17 @@ obj-$(CONFIG_DETECT_SOFTLOCKUP) += softlockup.o
 obj-$(CONFIG_GENERIC_HARDIRQS) += irq/
 obj-$(CONFIG_SECCOMP) += seccomp.o
 obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o
+obj-$(CONFIG_CLASSIC_RCU) += rcuclassic.o
+obj-$(CONFIG_PREEMPT_RCU) += rcupreempt.o
+ifeq ($(CONFIG_PREEMPT_RCU),y)
+obj-$(CONFIG_RCU_TRACE) += rcupreempt_trace.o
+endif
 obj-$(CONFIG_RELAY) += relay.o
 obj-$(CONFIG_SYSCTL) += utsname_sysctl.o
 obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o
 obj-$(CONFIG_TASKSTATS) += taskstats.o tsacct.o
 obj-$(CONFIG_MARKERS) += marker.o
+obj-$(CONFIG_LATENCYTOP) += latencytop.o
 
 ifneq ($(CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER),y)
 # According to Alan Modra <alan@linuxcare.com.au>, the -fno-omit-frame-pointer is

kernel/cpu.c

@@ -15,9 +15,8 @@
 #include <linux/stop_machine.h>
 #include <linux/mutex.h>
 
-/* This protects CPUs going up and down... */
+/* Serializes the updates to cpu_online_map, cpu_present_map */
 static DEFINE_MUTEX(cpu_add_remove_lock);
-static DEFINE_MUTEX(cpu_bitmask_lock);
 
 static __cpuinitdata RAW_NOTIFIER_HEAD(cpu_chain);
 
@@ -26,52 +25,123 @@ static __cpuinitdata RAW_NOTIFIER_HEAD(cpu_chain);
  */
 static int cpu_hotplug_disabled;
 
-#ifdef CONFIG_HOTPLUG_CPU
+static struct {
+	struct task_struct *active_writer;
+	struct mutex lock; /* Synchronizes accesses to refcount, */
+	/*
+	 * Also blocks the new readers during
+	 * an ongoing cpu hotplug operation.
+	 */
+	int refcount;
+	wait_queue_head_t writer_queue;
+} cpu_hotplug;
 
-/* Crappy recursive lock-takers in cpufreq! Complain loudly about idiots */
-static struct task_struct *recursive;
-static int recursive_depth;
+#define writer_exists() (cpu_hotplug.active_writer != NULL)
 
-void lock_cpu_hotplug(void)
+void __init cpu_hotplug_init(void)
 {
-	struct task_struct *tsk = current;
-
-	if (tsk == recursive) {
-		static int warnings = 10;
-		if (warnings) {
-			printk(KERN_ERR "Lukewarm IQ detected in hotplug locking\n");
-			WARN_ON(1);
-			warnings--;
-		}
-		recursive_depth++;
+	cpu_hotplug.active_writer = NULL;
+	mutex_init(&cpu_hotplug.lock);
+	cpu_hotplug.refcount = 0;
+	init_waitqueue_head(&cpu_hotplug.writer_queue);
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+void get_online_cpus(void)
+{
+	might_sleep();
+	if (cpu_hotplug.active_writer == current)
 		return;
-	}
-	mutex_lock(&cpu_bitmask_lock);
-	recursive = tsk;
+	mutex_lock(&cpu_hotplug.lock);
+	cpu_hotplug.refcount++;
+	mutex_unlock(&cpu_hotplug.lock);
+
 }
-EXPORT_SYMBOL_GPL(lock_cpu_hotplug);
+EXPORT_SYMBOL_GPL(get_online_cpus);
 
-void unlock_cpu_hotplug(void)
+void put_online_cpus(void)
 {
-	WARN_ON(recursive != current);
-	if (recursive_depth) {
-		recursive_depth--;
+	if (cpu_hotplug.active_writer == current)
 		return;
-	}
-	recursive = NULL;
-	mutex_unlock(&cpu_bitmask_lock);
+	mutex_lock(&cpu_hotplug.lock);
+	cpu_hotplug.refcount--;
+
+	if (unlikely(writer_exists()) && !cpu_hotplug.refcount)
+		wake_up(&cpu_hotplug.writer_queue);
+
+	mutex_unlock(&cpu_hotplug.lock);
+
 }
-EXPORT_SYMBOL_GPL(unlock_cpu_hotplug);
+EXPORT_SYMBOL_GPL(put_online_cpus);
 
 #endif	/* CONFIG_HOTPLUG_CPU */
 
+/*
+ * The following two API's must be used when attempting
+ * to serialize the updates to cpu_online_map, cpu_present_map.
+ */
+void cpu_maps_update_begin(void)
+{
+	mutex_lock(&cpu_add_remove_lock);
+}
+
+void cpu_maps_update_done(void)
+{
+	mutex_unlock(&cpu_add_remove_lock);
+}
+
+/*
+ * This ensures that the hotplug operation can begin only when the
+ * refcount goes to zero.
+ *
+ * Note that during a cpu-hotplug operation, the new readers, if any,
+ * will be blocked by the cpu_hotplug.lock
+ *
+ * Since cpu_maps_update_begin is always called after invoking
+ * cpu_maps_update_begin, we can be sure that only one writer is active.
+ *
+ * Note that theoretically, there is a possibility of a livelock:
+ * - Refcount goes to zero, last reader wakes up the sleeping
+ *   writer.
+ * - Last reader unlocks the cpu_hotplug.lock.
+ * - A new reader arrives at this moment, bumps up the refcount.
+ * - The writer acquires the cpu_hotplug.lock finds the refcount
+ *   non zero and goes to sleep again.
+ *
+ * However, this is very difficult to achieve in practice since
+ * get_online_cpus() not an api which is called all that often.
+ *
+ */
+static void cpu_hotplug_begin(void)
+{
+	DECLARE_WAITQUEUE(wait, current);
+
+	mutex_lock(&cpu_hotplug.lock);
+
+	cpu_hotplug.active_writer = current;
+	add_wait_queue_exclusive(&cpu_hotplug.writer_queue, &wait);
+	while (cpu_hotplug.refcount) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		mutex_unlock(&cpu_hotplug.lock);
+		schedule();
+		mutex_lock(&cpu_hotplug.lock);
+	}
+	remove_wait_queue_locked(&cpu_hotplug.writer_queue, &wait);
+}
+
+static void cpu_hotplug_done(void)
+{
+	cpu_hotplug.active_writer = NULL;
+	mutex_unlock(&cpu_hotplug.lock);
+}
 /* Need to know about CPUs going up/down? */
 int __cpuinit register_cpu_notifier(struct notifier_block *nb)
 {
 	int ret;
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	ret = raw_notifier_chain_register(&cpu_chain, nb);
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 	return ret;
 }
 
@@ -81,9 +151,9 @@ EXPORT_SYMBOL(register_cpu_notifier);
 
 void unregister_cpu_notifier(struct notifier_block *nb)
 {
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	raw_notifier_chain_unregister(&cpu_chain, nb);
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 }
 EXPORT_SYMBOL(unregister_cpu_notifier);
 
@@ -147,7 +217,7 @@ static int _cpu_down(unsigned int cpu, int tasks_frozen)
 	if (!cpu_online(cpu))
 		return -EINVAL;
 
-	raw_notifier_call_chain(&cpu_chain, CPU_LOCK_ACQUIRE, hcpu);
+	cpu_hotplug_begin();
 	err = __raw_notifier_call_chain(&cpu_chain, CPU_DOWN_PREPARE | mod,
 					hcpu, -1, &nr_calls);
 	if (err == NOTIFY_BAD) {
@@ -166,9 +236,7 @@ static int _cpu_down(unsigned int cpu, int tasks_frozen)
 	cpu_clear(cpu, tmp);
 	set_cpus_allowed(current, tmp);
 
-	mutex_lock(&cpu_bitmask_lock);
 	p = __stop_machine_run(take_cpu_down, &tcd_param, cpu);
-	mutex_unlock(&cpu_bitmask_lock);
 
 	if (IS_ERR(p) || cpu_online(cpu)) {
 		/* CPU didn't die: tell everyone.  Can't complain. */
@@ -202,7 +270,7 @@ out_thread:
 out_allowed:
 	set_cpus_allowed(current, old_allowed);
 out_release:
-	raw_notifier_call_chain(&cpu_chain, CPU_LOCK_RELEASE, hcpu);
+	cpu_hotplug_done();
 	return err;
 }
 
@@ -210,13 +278,13 @@ int cpu_down(unsigned int cpu)
 {
 	int err = 0;
 
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	if (cpu_hotplug_disabled)
 		err = -EBUSY;
 	else
 		err = _cpu_down(cpu, 0);
 
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 	return err;
 }
 #endif /*CONFIG_HOTPLUG_CPU*/
@@ -231,7 +299,7 @@ static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen)
 	if (cpu_online(cpu) || !cpu_present(cpu))
 		return -EINVAL;
 
-	raw_notifier_call_chain(&cpu_chain, CPU_LOCK_ACQUIRE, hcpu);
+	cpu_hotplug_begin();
 	ret = __raw_notifier_call_chain(&cpu_chain, CPU_UP_PREPARE | mod, hcpu,
 							-1, &nr_calls);
 	if (ret == NOTIFY_BAD) {
@@ -243,9 +311,7 @@ static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen)
 	}
 
 	/* Arch-specific enabling code. */
-	mutex_lock(&cpu_bitmask_lock);
 	ret = __cpu_up(cpu);
-	mutex_unlock(&cpu_bitmask_lock);
 	if (ret != 0)
 		goto out_notify;
 	BUG_ON(!cpu_online(cpu));
@@ -257,7 +323,7 @@ out_notify:
 	if (ret != 0)
 		__raw_notifier_call_chain(&cpu_chain,
 				CPU_UP_CANCELED | mod, hcpu, nr_calls, NULL);
-	raw_notifier_call_chain(&cpu_chain, CPU_LOCK_RELEASE, hcpu);
+	cpu_hotplug_done();
 
 	return ret;
 }
@@ -275,13 +341,13 @@ int __cpuinit cpu_up(unsigned int cpu)
 		return -EINVAL;
 	}
 
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	if (cpu_hotplug_disabled)
 		err = -EBUSY;
 	else
 		err = _cpu_up(cpu, 0);
 
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 	return err;
 }
 
@@ -292,7 +358,7 @@ int disable_nonboot_cpus(void)
 {
 	int cpu, first_cpu, error = 0;
 
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	first_cpu = first_cpu(cpu_online_map);
 	/* We take down all of the non-boot CPUs in one shot to avoid races
 	 * with the userspace trying to use the CPU hotplug at the same time
@@ -319,7 +385,7 @@ int disable_nonboot_cpus(void)
 	} else {
 		printk(KERN_ERR "Non-boot CPUs are not disabled\n");
 	}
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 	return error;
 }
 
@@ -328,7 +394,7 @@ void enable_nonboot_cpus(void)
 	int cpu, error;
 
 	/* Allow everyone to use the CPU hotplug again */
-	mutex_lock(&cpu_add_remove_lock);
+	cpu_maps_update_begin();
 	cpu_hotplug_disabled = 0;
 	if (cpus_empty(frozen_cpus))
 		goto out;
@@ -344,6 +410,6 @@ void enable_nonboot_cpus(void)
 	}
 	cpus_clear(frozen_cpus);
 out:
-	mutex_unlock(&cpu_add_remove_lock);
+	cpu_maps_update_done();
 }
 #endif /* CONFIG_PM_SLEEP_SMP */

kernel/cpuset.c

@@ -537,10 +537,10 @@ static int cpusets_overlap(struct cpuset *a, struct cpuset *b)
  *
  * Call with cgroup_mutex held.  May take callback_mutex during
  * call due to the kfifo_alloc() and kmalloc() calls.  May nest
- * a call to the lock_cpu_hotplug()/unlock_cpu_hotplug() pair.
+ * a call to the get_online_cpus()/put_online_cpus() pair.
  * Must not be called holding callback_mutex, because we must not
- * call lock_cpu_hotplug() while holding callback_mutex.  Elsewhere
- * the kernel nests callback_mutex inside lock_cpu_hotplug() calls.
+ * call get_online_cpus() while holding callback_mutex.  Elsewhere
+ * the kernel nests callback_mutex inside get_online_cpus() calls.
  * So the reverse nesting would risk an ABBA deadlock.
  *
  * The three key local variables below are:
@@ -691,9 +691,9 @@ restart:
 
 rebuild:
 	/* Have scheduler rebuild sched domains */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	partition_sched_domains(ndoms, doms);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 
 done:
 	if (q && !IS_ERR(q))
@@ -1617,10 +1617,10 @@ static struct cgroup_subsys_state *cpuset_create(
  *
  * If the cpuset being removed has its flag 'sched_load_balance'
  * enabled, then simulate turning sched_load_balance off, which
- * will call rebuild_sched_domains().  The lock_cpu_hotplug()
+ * will call rebuild_sched_domains().  The get_online_cpus()
  * call in rebuild_sched_domains() must not be made while holding
  * callback_mutex.  Elsewhere the kernel nests callback_mutex inside
- * lock_cpu_hotplug() calls.  So the reverse nesting would risk an
+ * get_online_cpus() calls.  So the reverse nesting would risk an
  * ABBA deadlock.
  */
 

kernel/fork.c

@@ -1045,6 +1045,10 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	copy_flags(clone_flags, p);
 	INIT_LIST_HEAD(&p->children);
 	INIT_LIST_HEAD(&p->sibling);
+#ifdef CONFIG_PREEMPT_RCU
+	p->rcu_read_lock_nesting = 0;
+	p->rcu_flipctr_idx = 0;
+#endif /* #ifdef CONFIG_PREEMPT_RCU */
 	p->vfork_done = NULL;
 	spin_lock_init(&p->alloc_lock);
 
@@ -1059,6 +1063,11 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	p->prev_utime = cputime_zero;
 	p->prev_stime = cputime_zero;
 
+#ifdef CONFIG_DETECT_SOFTLOCKUP
+	p->last_switch_count = 0;
+	p->last_switch_timestamp = 0;
+#endif
+
 #ifdef CONFIG_TASK_XACCT
 	p->rchar = 0;		/* I/O counter: bytes read */
 	p->wchar = 0;		/* I/O counter: bytes written */
@@ -1196,6 +1205,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 #ifdef TIF_SYSCALL_EMU
 	clear_tsk_thread_flag(p, TIF_SYSCALL_EMU);
 #endif
+	clear_all_latency_tracing(p);
 
 	/* Our parent execution domain becomes current domain
 	   These must match for thread signalling to apply */
@@ -1237,6 +1247,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	 * parent's CPU). This avoids alot of nasty races.
 	 */
 	p->cpus_allowed = current->cpus_allowed;
+	p->rt.nr_cpus_allowed = current->rt.nr_cpus_allowed;
 	if (unlikely(!cpu_isset(task_cpu(p), p->cpus_allowed) ||
 			!cpu_online(task_cpu(p))))
 		set_task_cpu(p, smp_processor_id());

kernel/hrtimer.c

@@ -325,6 +325,22 @@ unsigned long ktime_divns(const ktime_t kt, s64 div)
 }
 #endif /* BITS_PER_LONG >= 64 */
 
+/*
+ * Check, whether the timer is on the callback pending list
+ */
+static inline int hrtimer_cb_pending(const struct hrtimer *timer)
+{
+	return timer->state & HRTIMER_STATE_PENDING;
+}
+
+/*
+ * Remove a timer from the callback pending list
+ */
+static inline void hrtimer_remove_cb_pending(struct hrtimer *timer)
+{
+	list_del_init(&timer->cb_entry);
+}
+
 /* High resolution timer related functions */
 #ifdef CONFIG_HIGH_RES_TIMERS
 
@@ -493,22 +509,6 @@ void hres_timers_resume(void)
 	retrigger_next_event(NULL);
 }
 
-/*
- * Check, whether the timer is on the callback pending list
- */
-static inline int hrtimer_cb_pending(const struct hrtimer *timer)
-{
-	return timer->state & HRTIMER_STATE_PENDING;
-}
-
-/*
- * Remove a timer from the callback pending list
- */
-static inline void hrtimer_remove_cb_pending(struct hrtimer *timer)
-{
-	list_del_init(&timer->cb_entry);
-}
-
 /*
  * Initialize the high resolution related parts of cpu_base
  */
@@ -516,7 +516,6 @@ static inline void hrtimer_init_hres(struct hrtimer_cpu_base *base)
 {
 	base->expires_next.tv64 = KTIME_MAX;
 	base->hres_active = 0;
-	INIT_LIST_HEAD(&base->cb_pending);
 }
 
 /*
@@ -524,7 +523,6 @@ static inline void hrtimer_init_hres(struct hrtimer_cpu_base *base)
  */
 static inline void hrtimer_init_timer_hres(struct hrtimer *timer)
 {
-	INIT_LIST_HEAD(&timer->cb_entry);
 }
 
 /*
@@ -618,10 +616,13 @@ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer,
 {
 	return 0;
 }
-static inline int hrtimer_cb_pending(struct hrtimer *timer) { return 0; }
-static inline void hrtimer_remove_cb_pending(struct hrtimer *timer) { }
 static inline void hrtimer_init_hres(struct hrtimer_cpu_base *base) { }
 static inline void hrtimer_init_timer_hres(struct hrtimer *timer) { }
+static inline int hrtimer_reprogram(struct hrtimer *timer,
+				    struct hrtimer_clock_base *base)
+{
+	return 0;
+}
 
 #endif /* CONFIG_HIGH_RES_TIMERS */
 
@@ -1001,6 +1002,7 @@ void hrtimer_init(struct hrtimer *timer, clockid_t clock_id,
 		clock_id = CLOCK_MONOTONIC;
 
 	timer->base = &cpu_base->clock_base[clock_id];
+	INIT_LIST_HEAD(&timer->cb_entry);
 	hrtimer_init_timer_hres(timer);
 
 #ifdef CONFIG_TIMER_STATS
@@ -1030,6 +1032,85 @@ int hrtimer_get_res(const clockid_t which_clock, struct timespec *tp)
 }
 EXPORT_SYMBOL_GPL(hrtimer_get_res);
 
+static void run_hrtimer_pending(struct hrtimer_cpu_base *cpu_base)
+{
+	spin_lock_irq(&cpu_base->lock);
+
+	while (!list_empty(&cpu_base->cb_pending)) {
+		enum hrtimer_restart (*fn)(struct hrtimer *);
+		struct hrtimer *timer;
+		int restart;
+
+		timer = list_entry(cpu_base->cb_pending.next,
+				   struct hrtimer, cb_entry);
+
+		timer_stats_account_hrtimer(timer);
+
+		fn = timer->function;
+		__remove_hrtimer(timer, timer->base, HRTIMER_STATE_CALLBACK, 0);
+		spin_unlock_irq(&cpu_base->lock);
+
+		restart = fn(timer);
+
+		spin_lock_irq(&cpu_base->lock);
+
+		timer->state &= ~HRTIMER_STATE_CALLBACK;
+		if (restart == HRTIMER_RESTART) {
+			BUG_ON(hrtimer_active(timer));
+			/*
+			 * Enqueue the timer, allow reprogramming of the event
+			 * device
+			 */
+			enqueue_hrtimer(timer, timer->base, 1);
+		} else if (hrtimer_active(timer)) {
+			/*
+			 * If the timer was rearmed on another CPU, reprogram
+			 * the event device.
+			 */
+			if (timer->base->first == &timer->node)
+				hrtimer_reprogram(timer, timer->base);
+		}
+	}
+	spin_unlock_irq(&cpu_base->lock);
+}
+
+static void __run_hrtimer(struct hrtimer *timer)
+{
+	struct hrtimer_clock_base *base = timer->base;
+	struct hrtimer_cpu_base *cpu_base = base->cpu_base;
+	enum hrtimer_restart (*fn)(struct hrtimer *);
+	int restart;
+
+	__remove_hrtimer(timer, base, HRTIMER_STATE_CALLBACK, 0);
+	timer_stats_account_hrtimer(timer);
+
+	fn = timer->function;
+	if (timer->cb_mode == HRTIMER_CB_IRQSAFE_NO_SOFTIRQ) {
+		/*
+		 * Used for scheduler timers, avoid lock inversion with
+		 * rq->lock and tasklist_lock.
+		 *
+		 * These timers are required to deal with enqueue expiry
+		 * themselves and are not allowed to migrate.
+		 */
+		spin_unlock(&cpu_base->lock);
+		restart = fn(timer);
+		spin_lock(&cpu_base->lock);
+	} else
+		restart = fn(timer);
+
+	/*
+	 * Note: We clear the CALLBACK bit after enqueue_hrtimer to avoid
+	 * reprogramming of the event hardware. This happens at the end of this
+	 * function anyway.
+	 */
+	if (restart != HRTIMER_NORESTART) {
+		BUG_ON(timer->state != HRTIMER_STATE_CALLBACK);
+		enqueue_hrtimer(timer, base, 0);
+	}
+	timer->state &= ~HRTIMER_STATE_CALLBACK;
+}
+
 #ifdef CONFIG_HIGH_RES_TIMERS
 
 /*
@@ -1087,21 +1168,7 @@ void hrtimer_interrupt(struct clock_event_device *dev)
 				continue;
 			}
 
-			__remove_hrtimer(timer, base,
-					 HRTIMER_STATE_CALLBACK, 0);
-			timer_stats_account_hrtimer(timer);
-
-			/*
-			 * Note: We clear the CALLBACK bit after
-			 * enqueue_hrtimer to avoid reprogramming of
-			 * the event hardware. This happens at the end
-			 * of this function anyway.
-			 */
-			if (timer->function(timer) != HRTIMER_NORESTART) {
-				BUG_ON(timer->state != HRTIMER_STATE_CALLBACK);
-				enqueue_hrtimer(timer, base, 0);
-			}
-			timer->state &= ~HRTIMER_STATE_CALLBACK;
+			__run_hrtimer(timer);
 		}
 		spin_unlock(&cpu_base->lock);
 		base++;
@@ -1122,52 +1189,41 @@ void hrtimer_interrupt(struct clock_event_device *dev)
 
 static void run_hrtimer_softirq(struct softirq_action *h)
 {
-	struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
-
-	spin_lock_irq(&cpu_base->lock);
-
-	while (!list_empty(&cpu_base->cb_pending)) {
-		enum hrtimer_restart (*fn)(struct hrtimer *);
-		struct hrtimer *timer;
-		int restart;
-
-		timer = list_entry(cpu_base->cb_pending.next,
-				   struct hrtimer, cb_entry);
+	run_hrtimer_pending(&__get_cpu_var(hrtimer_bases));
+}
 
-		timer_stats_account_hrtimer(timer);
+#endif	/* CONFIG_HIGH_RES_TIMERS */
 
-		fn = timer->function;
-		__remove_hrtimer(timer, timer->base, HRTIMER_STATE_CALLBACK, 0);
-		spin_unlock_irq(&cpu_base->lock);
+/*
+ * Called from timer softirq every jiffy, expire hrtimers:
+ *
+ * For HRT its the fall back code to run the softirq in the timer
+ * softirq context in case the hrtimer initialization failed or has
+ * not been done yet.
+ */
+void hrtimer_run_pending(void)
+{
+	struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
 
-		restart = fn(timer);
+	if (hrtimer_hres_active())
+		return;
 
-		spin_lock_irq(&cpu_base->lock);
+	/*
+	 * This _is_ ugly: We have to check in the softirq context,
+	 * whether we can switch to highres and / or nohz mode. The
+	 * clocksource switch happens in the timer interrupt with
+	 * xtime_lock held. Notification from there only sets the
+	 * check bit in the tick_oneshot code, otherwise we might
+	 * deadlock vs. xtime_lock.
+	 */
+	if (tick_check_oneshot_change(!hrtimer_is_hres_enabled()))
+		hrtimer_switch_to_hres();
 
-		timer->state &= ~HRTIMER_STATE_CALLBACK;
-		if (restart == HRTIMER_RESTART) {
-			BUG_ON(hrtimer_active(timer));
-			/*
-			 * Enqueue the timer, allow reprogramming of the event
-			 * device
-			 */
-			enqueue_hrtimer(timer, timer->base, 1);
-		} else if (hrtimer_active(timer)) {
-			/*
-			 * If the timer was rearmed on another CPU, reprogram
-			 * the event device.
-			 */
-			if (timer->base->first == &timer->node)
-				hrtimer_reprogram(timer, timer->base);
-		}
-	}
-	spin_unlock_irq(&cpu_base->lock);
+	run_hrtimer_pending(cpu_base);
 }
 
-#endif	/* CONFIG_HIGH_RES_TIMERS */
-
 /*
- * Expire the per base hrtimer-queue:
+ * Called from hardirq context every jiffy
  */
 static inline void run_hrtimer_queue(struct hrtimer_cpu_base *cpu_base,
 				     int index)
@@ -1181,46 +1237,27 @@ static inline void run_hrtimer_queue(struct hrtimer_cpu_base *cpu_base,
 	if (base->get_softirq_time)
 		base->softirq_time = base->get_softirq_time();
 
-	spin_lock_irq(&cpu_base->lock);
+	spin_lock(&cpu_base->lock);
 
 	while ((node = base->first)) {
 		struct hrtimer *timer;
-		enum hrtimer_restart (*fn)(struct hrtimer *);
-		int restart;
 
 		timer = rb_entry(node, struct hrtimer, node);
 		if (base->softirq_time.tv64 <= timer->expires.tv64)
 			break;
 
-#ifdef CONFIG_HIGH_RES_TIMERS
-		WARN_ON_ONCE(timer->cb_mode == HRTIMER_CB_IRQSAFE_NO_SOFTIRQ);
-#endif
-		timer_stats_account_hrtimer(timer);
-
-		fn = timer->function;
-		__remove_hrtimer(timer, base, HRTIMER_STATE_CALLBACK, 0);
-		spin_unlock_irq(&cpu_base->lock);
-
-		restart = fn(timer);
-
-		spin_lock_irq(&cpu_base->lock);
-
-		timer->state &= ~HRTIMER_STATE_CALLBACK;
-		if (restart != HRTIMER_NORESTART) {
-			BUG_ON(hrtimer_active(timer));
-			enqueue_hrtimer(timer, base, 0);
+		if (timer->cb_mode == HRTIMER_CB_SOFTIRQ) {
+			__remove_hrtimer(timer, base, HRTIMER_STATE_PENDING, 0);
+			list_add_tail(&timer->cb_entry,
+					&base->cpu_base->cb_pending);
+			continue;
 		}
+
+		__run_hrtimer(timer);
 	}
-	spin_unlock_irq(&cpu_base->lock);
+	spin_unlock(&cpu_base->lock);
 }
 
-/*
- * Called from timer softirq every jiffy, expire hrtimers:
- *
- * For HRT its the fall back code to run the softirq in the timer
- * softirq context in case the hrtimer initialization failed or has
- * not been done yet.
- */
 void hrtimer_run_queues(void)
 {
 	struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases);
@@ -1229,18 +1266,6 @@ void hrtimer_run_queues(void)
 	if (hrtimer_hres_active())
 		return;
 
-	/*
-	 * This _is_ ugly: We have to check in the softirq context,
-	 * whether we can switch to highres and / or nohz mode. The
-	 * clocksource switch happens in the timer interrupt with
-	 * xtime_lock held. Notification from there only sets the
-	 * check bit in the tick_oneshot code, otherwise we might
-	 * deadlock vs. xtime_lock.
-	 */
-	if (tick_check_oneshot_change(!hrtimer_is_hres_enabled()))
-		if (hrtimer_switch_to_hres())
-			return;
-
 	hrtimer_get_softirq_time(cpu_base);
 
 	for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++)
@@ -1268,7 +1293,7 @@ void hrtimer_init_sleeper(struct hrtimer_sleeper *sl, struct task_struct *task)
 	sl->timer.function = hrtimer_wakeup;
 	sl->task = task;
 #ifdef CONFIG_HIGH_RES_TIMERS
-	sl->timer.cb_mode = HRTIMER_CB_IRQSAFE_NO_RESTART;
+	sl->timer.cb_mode = HRTIMER_CB_IRQSAFE_NO_SOFTIRQ;
 #endif
 }
 
@@ -1279,6 +1304,8 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod
 	do {
 		set_current_state(TASK_INTERRUPTIBLE);
 		hrtimer_start(&t->timer, t->timer.expires, mode);
+		if (!hrtimer_active(&t->timer))
+			t->task = NULL;
 
 		if (likely(t->task))
 			schedule();
@@ -1389,6 +1416,7 @@ static void __cpuinit init_hrtimers_cpu(int cpu)
 	for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++)
 		cpu_base->clock_base[i].cpu_base = cpu_base;
 
+	INIT_LIST_HEAD(&cpu_base->cb_pending);
 	hrtimer_init_hres(cpu_base);
 }
 

kernel/kthread.c

@@ -15,6 +15,8 @@
 #include <linux/mutex.h>
 #include <asm/semaphore.h>
 
+#define KTHREAD_NICE_LEVEL (-5)
+
 static DEFINE_SPINLOCK(kthread_create_lock);
 static LIST_HEAD(kthread_create_list);
 struct task_struct *kthreadd_task;
@@ -94,10 +96,18 @@ static void create_kthread(struct kthread_create_info *create)
 	if (pid < 0) {
 		create->result = ERR_PTR(pid);
 	} else {
+		struct sched_param param = { .sched_priority = 0 };
 		wait_for_completion(&create->started);
 		read_lock(&tasklist_lock);
 		create->result = find_task_by_pid(pid);
 		read_unlock(&tasklist_lock);
+		/*
+		 * root may have changed our (kthreadd's) priority or CPU mask.
+		 * The kernel thread should not inherit these properties.
+		 */
+		sched_setscheduler(create->result, SCHED_NORMAL, &param);
+		set_user_nice(create->result, KTHREAD_NICE_LEVEL);
+		set_cpus_allowed(create->result, CPU_MASK_ALL);
 	}
 	complete(&create->done);
 }
@@ -221,7 +231,7 @@ int kthreadd(void *unused)
 	/* Setup a clean context for our children to inherit. */
 	set_task_comm(tsk, "kthreadd");
 	ignore_signals(tsk);
-	set_user_nice(tsk, -5);
+	set_user_nice(tsk, KTHREAD_NICE_LEVEL);
 	set_cpus_allowed(tsk, CPU_MASK_ALL);
 
 	current->flags |= PF_NOFREEZE;

kernel/latencytop.c

@@ -0,0 +1,239 @@
+/*
+ * latencytop.c: Latency display infrastructure
+ *
+ * (C) Copyright 2008 Intel Corporation
+ * Author: Arjan van de Ven <arjan@linux.intel.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; version 2
+ * of the License.
+ */
+#include <linux/latencytop.h>
+#include <linux/kallsyms.h>
+#include <linux/seq_file.h>
+#include <linux/notifier.h>
+#include <linux/spinlock.h>
+#include <linux/proc_fs.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/stacktrace.h>
+
+static DEFINE_SPINLOCK(latency_lock);
+
+#define MAXLR 128
+static struct latency_record latency_record[MAXLR];
+
+int latencytop_enabled;
+
+void clear_all_latency_tracing(struct task_struct *p)
+{
+	unsigned long flags;
+
+	if (!latencytop_enabled)
+		return;
+
+	spin_lock_irqsave(&latency_lock, flags);
+	memset(&p->latency_record, 0, sizeof(p->latency_record));
+	p->latency_record_count = 0;
+	spin_unlock_irqrestore(&latency_lock, flags);
+}
+
+static void clear_global_latency_tracing(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&latency_lock, flags);
+	memset(&latency_record, 0, sizeof(latency_record));
+	spin_unlock_irqrestore(&latency_lock, flags);
+}
+
+static void __sched
+account_global_scheduler_latency(struct task_struct *tsk, struct latency_record *lat)
+{
+	int firstnonnull = MAXLR + 1;
+	int i;
+
+	if (!latencytop_enabled)
+		return;
+
+	/* skip kernel threads for now */
+	if (!tsk->mm)
+		return;
+
+	for (i = 0; i < MAXLR; i++) {
+		int q;
+		int same = 1;
+		/* Nothing stored: */
+		if (!latency_record[i].backtrace[0]) {
+			if (firstnonnull > i)
+				firstnonnull = i;
+			continue;
+		}
+		for (q = 0 ; q < LT_BACKTRACEDEPTH ; q++) {
+			if (latency_record[i].backtrace[q] !=
+				lat->backtrace[q])
+				same = 0;
+			if (same && lat->backtrace[q] == 0)
+				break;
+			if (same && lat->backtrace[q] == ULONG_MAX)
+				break;
+		}
+		if (same) {
+			latency_record[i].count++;
+			latency_record[i].time += lat->time;
+			if (lat->time > latency_record[i].max)
+				latency_record[i].max = lat->time;
+			return;
+		}
+	}
+
+	i = firstnonnull;
+	if (i >= MAXLR - 1)
+		return;
+
+	/* Allocted a new one: */
+	memcpy(&latency_record[i], lat, sizeof(struct latency_record));
+}
+
+static inline void store_stacktrace(struct task_struct *tsk, struct latency_record *lat)
+{
+	struct stack_trace trace;
+
+	memset(&trace, 0, sizeof(trace));
+	trace.max_entries = LT_BACKTRACEDEPTH;
+	trace.entries = &lat->backtrace[0];
+	trace.skip = 0;
+	save_stack_trace_tsk(tsk, &trace);
+}
+
+void __sched
+account_scheduler_latency(struct task_struct *tsk, int usecs, int inter)
+{
+	unsigned long flags;
+	int i, q;
+	struct latency_record lat;
+
+	if (!latencytop_enabled)
+		return;
+
+	/* Long interruptible waits are generally user requested... */
+	if (inter && usecs > 5000)
+		return;
+
+	memset(&lat, 0, sizeof(lat));
+	lat.count = 1;
+	lat.time = usecs;
+	lat.max = usecs;
+	store_stacktrace(tsk, &lat);
+
+	spin_lock_irqsave(&latency_lock, flags);
+
+	account_global_scheduler_latency(tsk, &lat);
+
+	/*
+	 * short term hack; if we're > 32 we stop; future we recycle:
+	 */
+	tsk->latency_record_count++;
+	if (tsk->latency_record_count >= LT_SAVECOUNT)
+		goto out_unlock;
+
+	for (i = 0; i < LT_SAVECOUNT ; i++) {
+		struct latency_record *mylat;
+		int same = 1;
+		mylat = &tsk->latency_record[i];
+		for (q = 0 ; q < LT_BACKTRACEDEPTH ; q++) {
+			if (mylat->backtrace[q] !=
+				lat.backtrace[q])
+				same = 0;
+			if (same && lat.backtrace[q] == 0)
+				break;
+			if (same && lat.backtrace[q] == ULONG_MAX)
+				break;
+		}
+		if (same) {
+			mylat->count++;
+			mylat->time += lat.time;
+			if (lat.time > mylat->max)
+				mylat->max = lat.time;
+			goto out_unlock;
+		}
+	}
+
+	/* Allocated a new one: */
+	i = tsk->latency_record_count;
+	memcpy(&tsk->latency_record[i], &lat, sizeof(struct latency_record));
+
+out_unlock:
+	spin_unlock_irqrestore(&latency_lock, flags);
+}
+
+static int lstats_show(struct seq_file *m, void *v)
+{
+	int i;
+
+	seq_puts(m, "Latency Top version : v0.1\n");
+
+	for (i = 0; i < MAXLR; i++) {
+		if (latency_record[i].backtrace[0]) {
+			int q;
+			seq_printf(m, "%i %li %li ",
+				latency_record[i].count,
+				latency_record[i].time,
+				latency_record[i].max);
+			for (q = 0; q < LT_BACKTRACEDEPTH; q++) {
+				char sym[KSYM_NAME_LEN];
+				char *c;
+				if (!latency_record[i].backtrace[q])
+					break;
+				if (latency_record[i].backtrace[q] == ULONG_MAX)
+					break;
+				sprint_symbol(sym, latency_record[i].backtrace[q]);
+				c = strchr(sym, '+');
+				if (c)
+					*c = 0;
+				seq_printf(m, "%s ", sym);
+			}
+			seq_printf(m, "\n");
+		}
+	}
+	return 0;
+}
+
+static ssize_t
+lstats_write(struct file *file, const char __user *buf, size_t count,
+	     loff_t *offs)
+{
+	clear_global_latency_tracing();
+
+	return count;
+}
+
+static int lstats_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, lstats_show, NULL);
+}
+
+static struct file_operations lstats_fops = {
+	.open		= lstats_open,
+	.read		= seq_read,
+	.write		= lstats_write,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static int __init init_lstats_procfs(void)
+{
+	struct proc_dir_entry *pe;
+
+	pe = create_proc_entry("latency_stats", 0644, NULL);
+	if (!pe)
+		return -ENOMEM;
+
+	pe->proc_fops = &lstats_fops;
+
+	return 0;
+}
+__initcall(init_lstats_procfs);

kernel/lockdep.c

@@ -3206,7 +3206,11 @@ retry:
 
 EXPORT_SYMBOL_GPL(debug_show_all_locks);
 
-void debug_show_held_locks(struct task_struct *task)
+/*
+ * Careful: only use this function if you are sure that
+ * the task cannot run in parallel!
+ */
+void __debug_show_held_locks(struct task_struct *task)
 {
 	if (unlikely(!debug_locks)) {
 		printk("INFO: lockdep is turned off.\n");
@@ -3214,6 +3218,12 @@ void debug_show_held_locks(struct task_struct *task)
 	}
 	lockdep_print_held_locks(task);
 }
+EXPORT_SYMBOL_GPL(__debug_show_held_locks);
+
+void debug_show_held_locks(struct task_struct *task)
+{
+		__debug_show_held_locks(task);
+}
 
 EXPORT_SYMBOL_GPL(debug_show_held_locks);
 

kernel/module.c

@@ -496,6 +496,8 @@ static struct module_attribute modinfo_##field = {                    \
 MODINFO_ATTR(version);
 MODINFO_ATTR(srcversion);
 
+static char last_unloaded_module[MODULE_NAME_LEN+1];
+
 #ifdef CONFIG_MODULE_UNLOAD
 /* Init the unload section of the module. */
 static void module_unload_init(struct module *mod)
@@ -719,6 +721,8 @@ sys_delete_module(const char __user *name_user, unsigned int flags)
 		mod->exit();
 		mutex_lock(&module_mutex);
 	}
+	/* Store the name of the last unloaded module for diagnostic purposes */
+	sprintf(last_unloaded_module, mod->name);
 	free_module(mod);
 
  out:
@@ -2357,21 +2361,30 @@ static void m_stop(struct seq_file *m, void *p)
 	mutex_unlock(&module_mutex);
 }
 
-static char *taint_flags(unsigned int taints, char *buf)
+static char *module_flags(struct module *mod, char *buf)
 {
 	int bx = 0;
 
-	if (taints) {
+	if (mod->taints ||
+	    mod->state == MODULE_STATE_GOING ||
+	    mod->state == MODULE_STATE_COMING) {
 		buf[bx++] = '(';
-		if (taints & TAINT_PROPRIETARY_MODULE)
+		if (mod->taints & TAINT_PROPRIETARY_MODULE)
 			buf[bx++] = 'P';
-		if (taints & TAINT_FORCED_MODULE)
+		if (mod->taints & TAINT_FORCED_MODULE)
 			buf[bx++] = 'F';
 		/*
 		 * TAINT_FORCED_RMMOD: could be added.
 		 * TAINT_UNSAFE_SMP, TAINT_MACHINE_CHECK, TAINT_BAD_PAGE don't
 		 * apply to modules.
 		 */
+
+		/* Show a - for module-is-being-unloaded */
+		if (mod->state == MODULE_STATE_GOING)
+			buf[bx++] = '-';
+		/* Show a + for module-is-being-loaded */
+		if (mod->state == MODULE_STATE_COMING)
+			buf[bx++] = '+';
 		buf[bx++] = ')';
 	}
 	buf[bx] = '\0';
@@ -2398,7 +2411,7 @@ static int m_show(struct seq_file *m, void *p)
 
 	/* Taints info */
 	if (mod->taints)
-		seq_printf(m, " %s", taint_flags(mod->taints, buf));
+		seq_printf(m, " %s", module_flags(mod, buf));
 
 	seq_printf(m, "\n");
 	return 0;
@@ -2493,7 +2506,9 @@ void print_modules(void)
 
 	printk("Modules linked in:");
 	list_for_each_entry(mod, &modules, list)
-		printk(" %s%s", mod->name, taint_flags(mod->taints, buf));
+		printk(" %s%s", mod->name, module_flags(mod, buf));
+	if (last_unloaded_module[0])
+		printk(" [last unloaded: %s]", last_unloaded_module);
 	printk("\n");
 }
 

kernel/posix-cpu-timers.c

@@ -967,6 +967,7 @@ static void check_thread_timers(struct task_struct *tsk,
 {
 	int maxfire;
 	struct list_head *timers = tsk->cpu_timers;
+	struct signal_struct *const sig = tsk->signal;
 
 	maxfire = 20;
 	tsk->it_prof_expires = cputime_zero;
@@ -1011,6 +1012,35 @@ static void check_thread_timers(struct task_struct *tsk,
 		t->firing = 1;
 		list_move_tail(&t->entry, firing);
 	}
+
+	/*
+	 * Check for the special case thread timers.
+	 */
+	if (sig->rlim[RLIMIT_RTTIME].rlim_cur != RLIM_INFINITY) {
+		unsigned long hard = sig->rlim[RLIMIT_RTTIME].rlim_max;
+		unsigned long *soft = &sig->rlim[RLIMIT_RTTIME].rlim_cur;
+
+		if (hard != RLIM_INFINITY &&
+		    tsk->rt.timeout > DIV_ROUND_UP(hard, USEC_PER_SEC/HZ)) {
+			/*
+			 * At the hard limit, we just die.
+			 * No need to calculate anything else now.
+			 */
+			__group_send_sig_info(SIGKILL, SEND_SIG_PRIV, tsk);
+			return;
+		}
+		if (tsk->rt.timeout > DIV_ROUND_UP(*soft, USEC_PER_SEC/HZ)) {
+			/*
+			 * At the soft limit, send a SIGXCPU every second.
+			 */
+			if (sig->rlim[RLIMIT_RTTIME].rlim_cur
+			    < sig->rlim[RLIMIT_RTTIME].rlim_max) {
+				sig->rlim[RLIMIT_RTTIME].rlim_cur +=
+								USEC_PER_SEC;
+			}
+			__group_send_sig_info(SIGXCPU, SEND_SIG_PRIV, tsk);
+		}
+	}
 }
 
 /*

kernel/printk.c

@@ -573,11 +573,6 @@ static int __init printk_time_setup(char *str)
 
 __setup("time", printk_time_setup);
 
-__attribute__((weak)) unsigned long long printk_clock(void)
-{
-	return sched_clock();
-}
-
 /* Check if we have any console registered that can be called early in boot. */
 static int have_callable_console(void)
 {
@@ -628,30 +623,57 @@ asmlinkage int printk(const char *fmt, ...)
 /* cpu currently holding logbuf_lock */
 static volatile unsigned int printk_cpu = UINT_MAX;
 
+const char printk_recursion_bug_msg [] =
+			KERN_CRIT "BUG: recent printk recursion!\n";
+static int printk_recursion_bug;
+
 asmlinkage int vprintk(const char *fmt, va_list args)
 {
+	static int log_level_unknown = 1;
+	static char printk_buf[1024];
+
 	unsigned long flags;
-	int printed_len;
+	int printed_len = 0;
+	int this_cpu;
 	char *p;
-	static char printk_buf[1024];
-	static int log_level_unknown = 1;
 
 	boot_delay_msec();
 
 	preempt_disable();
-	if (unlikely(oops_in_progress) && printk_cpu == smp_processor_id())
-		/* If a crash is occurring during printk() on this CPU,
-		 * make sure we can't deadlock */
-		zap_locks();
-
 	/* This stops the holder of console_sem just where we want him */
 	raw_local_irq_save(flags);
+	this_cpu = smp_processor_id();
+
+	/*
+	 * Ouch, printk recursed into itself!
+	 */
+	if (unlikely(printk_cpu == this_cpu)) {
+		/*
+		 * If a crash is occurring during printk() on this CPU,
+		 * then try to get the crash message out but make sure
+		 * we can't deadlock. Otherwise just return to avoid the
+		 * recursion and return - but flag the recursion so that
+		 * it can be printed at the next appropriate moment:
+		 */
+		if (!oops_in_progress) {
+			printk_recursion_bug = 1;
+			goto out_restore_irqs;
+		}
+		zap_locks();
+	}
+
 	lockdep_off();
 	spin_lock(&logbuf_lock);
-	printk_cpu = smp_processor_id();
+	printk_cpu = this_cpu;
 
+	if (printk_recursion_bug) {
+		printk_recursion_bug = 0;
+		strcpy(printk_buf, printk_recursion_bug_msg);
+		printed_len = sizeof(printk_recursion_bug_msg);
+	}
 	/* Emit the output into the temporary buffer */
-	printed_len = vscnprintf(printk_buf, sizeof(printk_buf), fmt, args);
+	printed_len += vscnprintf(printk_buf + printed_len,
+				  sizeof(printk_buf), fmt, args);
 
 	/*
 	 * Copy the output into log_buf.  If the caller didn't provide
@@ -680,7 +702,9 @@ asmlinkage int vprintk(const char *fmt, va_list args)
 					loglev_char = default_message_loglevel
 						+ '0';
 				}
-				t = printk_clock();
+				t = 0;
+				if (system_state != SYSTEM_BOOTING)
+					t = ktime_to_ns(ktime_get());
 				nanosec_rem = do_div(t, 1000000000);
 				tlen = sprintf(tbuf,
 						"<%c>[%5lu.%06lu] ",
@@ -744,6 +768,7 @@ asmlinkage int vprintk(const char *fmt, va_list args)
 		printk_cpu = UINT_MAX;
 		spin_unlock(&logbuf_lock);
 		lockdep_on();
+out_restore_irqs:
 		raw_local_irq_restore(flags);
 	}
 

kernel/profile.c

@@ -52,7 +52,7 @@ static DEFINE_PER_CPU(int, cpu_profile_flip);
 static DEFINE_MUTEX(profile_flip_mutex);
 #endif /* CONFIG_SMP */
 
-static int __init profile_setup(char * str)
+static int __init profile_setup(char *str)
 {
 	static char __initdata schedstr[] = "schedule";
 	static char __initdata sleepstr[] = "sleep";
@@ -104,28 +104,28 @@ __setup("profile=", profile_setup);
 
 void __init profile_init(void)
 {
-	if (!prof_on) 
+	if (!prof_on)
 		return;
- 
+
 	/* only text is profiled */
 	prof_len = (_etext - _stext) >> prof_shift;
 	prof_buffer = alloc_bootmem(prof_len*sizeof(atomic_t));
 }
 
 /* Profile event notifications */
- 
+
 #ifdef CONFIG_PROFILING
- 
+
 static BLOCKING_NOTIFIER_HEAD(task_exit_notifier);
 static ATOMIC_NOTIFIER_HEAD(task_free_notifier);
 static BLOCKING_NOTIFIER_HEAD(munmap_notifier);
- 
-void profile_task_exit(struct task_struct * task)
+
+void profile_task_exit(struct task_struct *task)
 {
 	blocking_notifier_call_chain(&task_exit_notifier, 0, task);
 }
- 
-int profile_handoff_task(struct task_struct * task)
+
+int profile_handoff_task(struct task_struct *task)
 {
 	int ret;
 	ret = atomic_notifier_call_chain(&task_free_notifier, 0, task);
@@ -137,52 +137,55 @@ void profile_munmap(unsigned long addr)
 	blocking_notifier_call_chain(&munmap_notifier, 0, (void *)addr);
 }
 
-int task_handoff_register(struct notifier_block * n)
+int task_handoff_register(struct notifier_block *n)
 {
 	return atomic_notifier_chain_register(&task_free_notifier, n);
 }
+EXPORT_SYMBOL_GPL(task_handoff_register);
 
-int task_handoff_unregister(struct notifier_block * n)
+int task_handoff_unregister(struct notifier_block *n)
 {
 	return atomic_notifier_chain_unregister(&task_free_notifier, n);
 }
+EXPORT_SYMBOL_GPL(task_handoff_unregister);
 
-int profile_event_register(enum profile_type type, struct notifier_block * n)
+int profile_event_register(enum profile_type type, struct notifier_block *n)
 {
 	int err = -EINVAL;
- 
+
 	switch (type) {
-		case PROFILE_TASK_EXIT:
-			err = blocking_notifier_chain_register(
-					&task_exit_notifier, n);
-			break;
-		case PROFILE_MUNMAP:
-			err = blocking_notifier_chain_register(
-					&munmap_notifier, n);
-			break;
+	case PROFILE_TASK_EXIT:
+		err = blocking_notifier_chain_register(
+				&task_exit_notifier, n);
+		break;
+	case PROFILE_MUNMAP:
+		err = blocking_notifier_chain_register(
+				&munmap_notifier, n);
+		break;
 	}
- 
+
 	return err;
 }
+EXPORT_SYMBOL_GPL(profile_event_register);
 
- 
-int profile_event_unregister(enum profile_type type, struct notifier_block * n)
+int profile_event_unregister(enum profile_type type, struct notifier_block *n)
 {
 	int err = -EINVAL;
- 
+
 	switch (type) {
-		case PROFILE_TASK_EXIT:
-			err = blocking_notifier_chain_unregister(
-					&task_exit_notifier, n);
-			break;
-		case PROFILE_MUNMAP:
-			err = blocking_notifier_chain_unregister(
-					&munmap_notifier, n);
-			break;
+	case PROFILE_TASK_EXIT:
+		err = blocking_notifier_chain_unregister(
+				&task_exit_notifier, n);
+		break;
+	case PROFILE_MUNMAP:
+		err = blocking_notifier_chain_unregister(
+				&munmap_notifier, n);
+		break;
 	}
 
 	return err;
 }
+EXPORT_SYMBOL_GPL(profile_event_unregister);
 
 int register_timer_hook(int (*hook)(struct pt_regs *))
 {
@@ -191,6 +194,7 @@ int register_timer_hook(int (*hook)(struct pt_regs *))
 	timer_hook = hook;
 	return 0;
 }
+EXPORT_SYMBOL_GPL(register_timer_hook);
 
 void unregister_timer_hook(int (*hook)(struct pt_regs *))
 {
@@ -199,13 +203,7 @@ void unregister_timer_hook(int (*hook)(struct pt_regs *))
 	/* make sure all CPUs see the NULL hook */
 	synchronize_sched();  /* Allow ongoing interrupts to complete. */
 }
-
-EXPORT_SYMBOL_GPL(register_timer_hook);
 EXPORT_SYMBOL_GPL(unregister_timer_hook);
-EXPORT_SYMBOL_GPL(task_handoff_register);
-EXPORT_SYMBOL_GPL(task_handoff_unregister);
-EXPORT_SYMBOL_GPL(profile_event_register);
-EXPORT_SYMBOL_GPL(profile_event_unregister);
 
 #endif /* CONFIG_PROFILING */
 
@@ -366,7 +364,7 @@ static int __devinit profile_cpu_callback(struct notifier_block *info,
 			per_cpu(cpu_profile_hits, cpu)[0] = page_address(page);
 		}
 		break;
-	out_free:
+out_free:
 		page = virt_to_page(per_cpu(cpu_profile_hits, cpu)[1]);
 		per_cpu(cpu_profile_hits, cpu)[1] = NULL;
 		__free_page(page);
@@ -409,7 +407,6 @@ void profile_hits(int type, void *__pc, unsigned int nr_hits)
 	atomic_add(nr_hits, &prof_buffer[min(pc, prof_len - 1)]);
 }
 #endif /* !CONFIG_SMP */
-
 EXPORT_SYMBOL_GPL(profile_hits);
 
 void profile_tick(int type)
@@ -427,7 +424,7 @@ void profile_tick(int type)
 #include <asm/uaccess.h>
 #include <asm/ptrace.h>
 
-static int prof_cpu_mask_read_proc (char *page, char **start, off_t off,
+static int prof_cpu_mask_read_proc(char *page, char **start, off_t off,
 			int count, int *eof, void *data)
 {
 	int len = cpumask_scnprintf(page, count, *(cpumask_t *)data);
@@ -437,8 +434,8 @@ static int prof_cpu_mask_read_proc (char *page, char **start, off_t off,
 	return len;
 }
 
-static int prof_cpu_mask_write_proc (struct file *file, const char __user *buffer,
-					unsigned long count, void *data)
+static int prof_cpu_mask_write_proc(struct file *file,
+	const char __user *buffer,  unsigned long count, void *data)
 {
 	cpumask_t *mask = (cpumask_t *)data;
 	unsigned long full_count = count, err;
@@ -457,7 +454,8 @@ void create_prof_cpu_mask(struct proc_dir_entry *root_irq_dir)
 	struct proc_dir_entry *entry;
 
 	/* create /proc/irq/prof_cpu_mask */
-	if (!(entry = create_proc_entry("prof_cpu_mask", 0600, root_irq_dir)))
+	entry = create_proc_entry("prof_cpu_mask", 0600, root_irq_dir);
+	if (!entry)
 		return;
 	entry->data = (void *)&prof_cpu_mask;
 	entry->read_proc = prof_cpu_mask_read_proc;
@@ -475,7 +473,7 @@ read_profile(struct file *file, char __user *buf, size_t count, loff_t *ppos)
 {
 	unsigned long p = *ppos;
 	ssize_t read;
-	char * pnt;
+	char *pnt;
 	unsigned int sample_step = 1 << prof_shift;
 
 	profile_flip_buffers();
@@ -486,12 +484,12 @@ read_profile(struct file *file, char __user *buf, size_t count, loff_t *ppos)
 	read = 0;
 
 	while (p < sizeof(unsigned int) && count > 0) {
-		if (put_user(*((char *)(&sample_step)+p),buf))
+		if (put_user(*((char *)(&sample_step)+p), buf))
 			return -EFAULT;
 		buf++; p++; count--; read++;
 	}
 	pnt = (char *)prof_buffer + p - sizeof(atomic_t);
-	if (copy_to_user(buf,(void *)pnt,count))
+	if (copy_to_user(buf, (void *)pnt, count))
 		return -EFAULT;
 	read += count;
 	*ppos += read;
@@ -508,7 +506,7 @@ static ssize_t write_profile(struct file *file, const char __user *buf,
 			     size_t count, loff_t *ppos)
 {
 #ifdef CONFIG_SMP
-	extern int setup_profiling_timer (unsigned int multiplier);
+	extern int setup_profiling_timer(unsigned int multiplier);
 
 	if (count == sizeof(int)) {
 		unsigned int multiplier;
@@ -591,7 +589,8 @@ static int __init create_proc_profile(void)
 		return 0;
 	if (create_hash_tables())
 		return -1;
-	if (!(entry = create_proc_entry("profile", S_IWUSR | S_IRUGO, NULL)))
+	entry = create_proc_entry("profile", S_IWUSR | S_IRUGO, NULL);
+	if (!entry)
 		return 0;
 	entry->proc_fops = &proc_profile_operations;
 	entry->size = (1+prof_len) * sizeof(atomic_t);

kernel/rcuclassic.c

@@ -0,0 +1,575 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright IBM Corporation, 2001
+ *
+ * Authors: Dipankar Sarma <dipankar@in.ibm.com>
+ *	    Manfred Spraul <manfred@colorfullife.com>
+ *
+ * Based on the original work by Paul McKenney <paulmck@us.ibm.com>
+ * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
+ * Papers:
+ * http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf
+ * http://lse.sourceforge.net/locking/rclock_OLS.2001.05.01c.sc.pdf (OLS2001)
+ *
+ * For detailed explanation of Read-Copy Update mechanism see -
+ * 		Documentation/RCU
+ *
+ */
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/spinlock.h>
+#include <linux/smp.h>
+#include <linux/rcupdate.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <asm/atomic.h>
+#include <linux/bitops.h>
+#include <linux/module.h>
+#include <linux/completion.h>
+#include <linux/moduleparam.h>
+#include <linux/percpu.h>
+#include <linux/notifier.h>
+#include <linux/cpu.h>
+#include <linux/mutex.h>
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+static struct lock_class_key rcu_lock_key;
+struct lockdep_map rcu_lock_map =
+	STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key);
+EXPORT_SYMBOL_GPL(rcu_lock_map);
+#endif
+
+
+/* Definition for rcupdate control block. */
+static struct rcu_ctrlblk rcu_ctrlblk = {
+	.cur = -300,
+	.completed = -300,
+	.lock = __SPIN_LOCK_UNLOCKED(&rcu_ctrlblk.lock),
+	.cpumask = CPU_MASK_NONE,
+};
+static struct rcu_ctrlblk rcu_bh_ctrlblk = {
+	.cur = -300,
+	.completed = -300,
+	.lock = __SPIN_LOCK_UNLOCKED(&rcu_bh_ctrlblk.lock),
+	.cpumask = CPU_MASK_NONE,
+};
+
+DEFINE_PER_CPU(struct rcu_data, rcu_data) = { 0L };
+DEFINE_PER_CPU(struct rcu_data, rcu_bh_data) = { 0L };
+
+static int blimit = 10;
+static int qhimark = 10000;
+static int qlowmark = 100;
+
+#ifdef CONFIG_SMP
+static void force_quiescent_state(struct rcu_data *rdp,
+			struct rcu_ctrlblk *rcp)
+{
+	int cpu;
+	cpumask_t cpumask;
+	set_need_resched();
+	if (unlikely(!rcp->signaled)) {
+		rcp->signaled = 1;
+		/*
+		 * Don't send IPI to itself. With irqs disabled,
+		 * rdp->cpu is the current cpu.
+		 */
+		cpumask = rcp->cpumask;
+		cpu_clear(rdp->cpu, cpumask);
+		for_each_cpu_mask(cpu, cpumask)
+			smp_send_reschedule(cpu);
+	}
+}
+#else
+static inline void force_quiescent_state(struct rcu_data *rdp,
+			struct rcu_ctrlblk *rcp)
+{
+	set_need_resched();
+}
+#endif
+
+/**
+ * call_rcu - Queue an RCU callback for invocation after a grace period.
+ * @head: structure to be used for queueing the RCU updates.
+ * @func: actual update function to be invoked after the grace period
+ *
+ * The update function will be invoked some time after a full grace
+ * period elapses, in other words after all currently executing RCU
+ * read-side critical sections have completed.  RCU read-side critical
+ * sections are delimited by rcu_read_lock() and rcu_read_unlock(),
+ * and may be nested.
+ */
+void call_rcu(struct rcu_head *head,
+				void (*func)(struct rcu_head *rcu))
+{
+	unsigned long flags;
+	struct rcu_data *rdp;
+
+	head->func = func;
+	head->next = NULL;
+	local_irq_save(flags);
+	rdp = &__get_cpu_var(rcu_data);
+	*rdp->nxttail = head;
+	rdp->nxttail = &head->next;
+	if (unlikely(++rdp->qlen > qhimark)) {
+		rdp->blimit = INT_MAX;
+		force_quiescent_state(rdp, &rcu_ctrlblk);
+	}
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(call_rcu);
+
+/**
+ * call_rcu_bh - Queue an RCU for invocation after a quicker grace period.
+ * @head: structure to be used for queueing the RCU updates.
+ * @func: actual update function to be invoked after the grace period
+ *
+ * The update function will be invoked some time after a full grace
+ * period elapses, in other words after all currently executing RCU
+ * read-side critical sections have completed. call_rcu_bh() assumes
+ * that the read-side critical sections end on completion of a softirq
+ * handler. This means that read-side critical sections in process
+ * context must not be interrupted by softirqs. This interface is to be
+ * used when most of the read-side critical sections are in softirq context.
+ * RCU read-side critical sections are delimited by rcu_read_lock() and
+ * rcu_read_unlock(), * if in interrupt context or rcu_read_lock_bh()
+ * and rcu_read_unlock_bh(), if in process context. These may be nested.
+ */
+void call_rcu_bh(struct rcu_head *head,
+				void (*func)(struct rcu_head *rcu))
+{
+	unsigned long flags;
+	struct rcu_data *rdp;
+
+	head->func = func;
+	head->next = NULL;
+	local_irq_save(flags);
+	rdp = &__get_cpu_var(rcu_bh_data);
+	*rdp->nxttail = head;
+	rdp->nxttail = &head->next;
+
+	if (unlikely(++rdp->qlen > qhimark)) {
+		rdp->blimit = INT_MAX;
+		force_quiescent_state(rdp, &rcu_bh_ctrlblk);
+	}
+
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(call_rcu_bh);
+
+/*
+ * Return the number of RCU batches processed thus far.  Useful
+ * for debug and statistics.
+ */
+long rcu_batches_completed(void)
+{
+	return rcu_ctrlblk.completed;
+}
+EXPORT_SYMBOL_GPL(rcu_batches_completed);
+
+/*
+ * Return the number of RCU batches processed thus far.  Useful
+ * for debug and statistics.
+ */
+long rcu_batches_completed_bh(void)
+{
+	return rcu_bh_ctrlblk.completed;
+}
+EXPORT_SYMBOL_GPL(rcu_batches_completed_bh);
+
+/* Raises the softirq for processing rcu_callbacks. */
+static inline void raise_rcu_softirq(void)
+{
+	raise_softirq(RCU_SOFTIRQ);
+	/*
+	 * The smp_mb() here is required to ensure that this cpu's
+	 * __rcu_process_callbacks() reads the most recently updated
+	 * value of rcu->cur.
+	 */
+	smp_mb();
+}
+
+/*
+ * Invoke the completed RCU callbacks. They are expected to be in
+ * a per-cpu list.
+ */
+static void rcu_do_batch(struct rcu_data *rdp)
+{
+	struct rcu_head *next, *list;
+	int count = 0;
+
+	list = rdp->donelist;
+	while (list) {
+		next = list->next;
+		prefetch(next);
+		list->func(list);
+		list = next;
+		if (++count >= rdp->blimit)
+			break;
+	}
+	rdp->donelist = list;
+
+	local_irq_disable();
+	rdp->qlen -= count;
+	local_irq_enable();
+	if (rdp->blimit == INT_MAX && rdp->qlen <= qlowmark)
+		rdp->blimit = blimit;
+
+	if (!rdp->donelist)
+		rdp->donetail = &rdp->donelist;
+	else
+		raise_rcu_softirq();
+}
+
+/*
+ * Grace period handling:
+ * The grace period handling consists out of two steps:
+ * - A new grace period is started.
+ *   This is done by rcu_start_batch. The start is not broadcasted to
+ *   all cpus, they must pick this up by comparing rcp->cur with
+ *   rdp->quiescbatch. All cpus are recorded  in the
+ *   rcu_ctrlblk.cpumask bitmap.
+ * - All cpus must go through a quiescent state.
+ *   Since the start of the grace period is not broadcasted, at least two
+ *   calls to rcu_check_quiescent_state are required:
+ *   The first call just notices that a new grace period is running. The
+ *   following calls check if there was a quiescent state since the beginning
+ *   of the grace period. If so, it updates rcu_ctrlblk.cpumask. If
+ *   the bitmap is empty, then the grace period is completed.
+ *   rcu_check_quiescent_state calls rcu_start_batch(0) to start the next grace
+ *   period (if necessary).
+ */
+/*
+ * Register a new batch of callbacks, and start it up if there is currently no
+ * active batch and the batch to be registered has not already occurred.
+ * Caller must hold rcu_ctrlblk.lock.
+ */
+static void rcu_start_batch(struct rcu_ctrlblk *rcp)
+{
+	if (rcp->next_pending &&
+			rcp->completed == rcp->cur) {
+		rcp->next_pending = 0;
+		/*
+		 * next_pending == 0 must be visible in
+		 * __rcu_process_callbacks() before it can see new value of cur.
+		 */
+		smp_wmb();
+		rcp->cur++;
+
+		/*
+		 * Accessing nohz_cpu_mask before incrementing rcp->cur needs a
+		 * Barrier  Otherwise it can cause tickless idle CPUs to be
+		 * included in rcp->cpumask, which will extend graceperiods
+		 * unnecessarily.
+		 */
+		smp_mb();
+		cpus_andnot(rcp->cpumask, cpu_online_map, nohz_cpu_mask);
+
+		rcp->signaled = 0;
+	}
+}
+
+/*
+ * cpu went through a quiescent state since the beginning of the grace period.
+ * Clear it from the cpu mask and complete the grace period if it was the last
+ * cpu. Start another grace period if someone has further entries pending
+ */
+static void cpu_quiet(int cpu, struct rcu_ctrlblk *rcp)
+{
+	cpu_clear(cpu, rcp->cpumask);
+	if (cpus_empty(rcp->cpumask)) {
+		/* batch completed ! */
+		rcp->completed = rcp->cur;
+		rcu_start_batch(rcp);
+	}
+}
+
+/*
+ * Check if the cpu has gone through a quiescent state (say context
+ * switch). If so and if it already hasn't done so in this RCU
+ * quiescent cycle, then indicate that it has done so.
+ */
+static void rcu_check_quiescent_state(struct rcu_ctrlblk *rcp,
+					struct rcu_data *rdp)
+{
+	if (rdp->quiescbatch != rcp->cur) {
+		/* start new grace period: */
+		rdp->qs_pending = 1;
+		rdp->passed_quiesc = 0;
+		rdp->quiescbatch = rcp->cur;
+		return;
+	}
+
+	/* Grace period already completed for this cpu?
+	 * qs_pending is checked instead of the actual bitmap to avoid
+	 * cacheline trashing.
+	 */
+	if (!rdp->qs_pending)
+		return;
+
+	/*
+	 * Was there a quiescent state since the beginning of the grace
+	 * period? If no, then exit and wait for the next call.
+	 */
+	if (!rdp->passed_quiesc)
+		return;
+	rdp->qs_pending = 0;
+
+	spin_lock(&rcp->lock);
+	/*
+	 * rdp->quiescbatch/rcp->cur and the cpu bitmap can come out of sync
+	 * during cpu startup. Ignore the quiescent state.
+	 */
+	if (likely(rdp->quiescbatch == rcp->cur))
+		cpu_quiet(rdp->cpu, rcp);
+
+	spin_unlock(&rcp->lock);
+}
+
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+/* warning! helper for rcu_offline_cpu. do not use elsewhere without reviewing
+ * locking requirements, the list it's pulling from has to belong to a cpu
+ * which is dead and hence not processing interrupts.
+ */
+static void rcu_move_batch(struct rcu_data *this_rdp, struct rcu_head *list,
+				struct rcu_head **tail)
+{
+	local_irq_disable();
+	*this_rdp->nxttail = list;
+	if (list)
+		this_rdp->nxttail = tail;
+	local_irq_enable();
+}
+
+static void __rcu_offline_cpu(struct rcu_data *this_rdp,
+				struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
+{
+	/* if the cpu going offline owns the grace period
+	 * we can block indefinitely waiting for it, so flush
+	 * it here
+	 */
+	spin_lock_bh(&rcp->lock);
+	if (rcp->cur != rcp->completed)
+		cpu_quiet(rdp->cpu, rcp);
+	spin_unlock_bh(&rcp->lock);
+	rcu_move_batch(this_rdp, rdp->donelist, rdp->donetail);
+	rcu_move_batch(this_rdp, rdp->curlist, rdp->curtail);
+	rcu_move_batch(this_rdp, rdp->nxtlist, rdp->nxttail);
+}
+
+static void rcu_offline_cpu(int cpu)
+{
+	struct rcu_data *this_rdp = &get_cpu_var(rcu_data);
+	struct rcu_data *this_bh_rdp = &get_cpu_var(rcu_bh_data);
+
+	__rcu_offline_cpu(this_rdp, &rcu_ctrlblk,
+					&per_cpu(rcu_data, cpu));
+	__rcu_offline_cpu(this_bh_rdp, &rcu_bh_ctrlblk,
+					&per_cpu(rcu_bh_data, cpu));
+	put_cpu_var(rcu_data);
+	put_cpu_var(rcu_bh_data);
+}
+
+#else
+
+static void rcu_offline_cpu(int cpu)
+{
+}
+
+#endif
+
+/*
+ * This does the RCU processing work from softirq context.
+ */
+static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp,
+					struct rcu_data *rdp)
+{
+	if (rdp->curlist && !rcu_batch_before(rcp->completed, rdp->batch)) {
+		*rdp->donetail = rdp->curlist;
+		rdp->donetail = rdp->curtail;
+		rdp->curlist = NULL;
+		rdp->curtail = &rdp->curlist;
+	}
+
+	if (rdp->nxtlist && !rdp->curlist) {
+		local_irq_disable();
+		rdp->curlist = rdp->nxtlist;
+		rdp->curtail = rdp->nxttail;
+		rdp->nxtlist = NULL;
+		rdp->nxttail = &rdp->nxtlist;
+		local_irq_enable();
+
+		/*
+		 * start the next batch of callbacks
+		 */
+
+		/* determine batch number */
+		rdp->batch = rcp->cur + 1;
+		/* see the comment and corresponding wmb() in
+		 * the rcu_start_batch()
+		 */
+		smp_rmb();
+
+		if (!rcp->next_pending) {
+			/* and start it/schedule start if it's a new batch */
+			spin_lock(&rcp->lock);
+			rcp->next_pending = 1;
+			rcu_start_batch(rcp);
+			spin_unlock(&rcp->lock);
+		}
+	}
+
+	rcu_check_quiescent_state(rcp, rdp);
+	if (rdp->donelist)
+		rcu_do_batch(rdp);
+}
+
+static void rcu_process_callbacks(struct softirq_action *unused)
+{
+	__rcu_process_callbacks(&rcu_ctrlblk, &__get_cpu_var(rcu_data));
+	__rcu_process_callbacks(&rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data));
+}
+
+static int __rcu_pending(struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
+{
+	/* This cpu has pending rcu entries and the grace period
+	 * for them has completed.
+	 */
+	if (rdp->curlist && !rcu_batch_before(rcp->completed, rdp->batch))
+		return 1;
+
+	/* This cpu has no pending entries, but there are new entries */
+	if (!rdp->curlist && rdp->nxtlist)
+		return 1;
+
+	/* This cpu has finished callbacks to invoke */
+	if (rdp->donelist)
+		return 1;
+
+	/* The rcu core waits for a quiescent state from the cpu */
+	if (rdp->quiescbatch != rcp->cur || rdp->qs_pending)
+		return 1;
+
+	/* nothing to do */
+	return 0;
+}
+
+/*
+ * Check to see if there is any immediate RCU-related work to be done
+ * by the current CPU, returning 1 if so.  This function is part of the
+ * RCU implementation; it is -not- an exported member of the RCU API.
+ */
+int rcu_pending(int cpu)
+{
+	return __rcu_pending(&rcu_ctrlblk, &per_cpu(rcu_data, cpu)) ||
+		__rcu_pending(&rcu_bh_ctrlblk, &per_cpu(rcu_bh_data, cpu));
+}
+
+/*
+ * Check to see if any future RCU-related work will need to be done
+ * by the current CPU, even if none need be done immediately, returning
+ * 1 if so.  This function is part of the RCU implementation; it is -not-
+ * an exported member of the RCU API.
+ */
+int rcu_needs_cpu(int cpu)
+{
+	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
+	struct rcu_data *rdp_bh = &per_cpu(rcu_bh_data, cpu);
+
+	return (!!rdp->curlist || !!rdp_bh->curlist || rcu_pending(cpu));
+}
+
+void rcu_check_callbacks(int cpu, int user)
+{
+	if (user ||
+	    (idle_cpu(cpu) && !in_softirq() &&
+				hardirq_count() <= (1 << HARDIRQ_SHIFT))) {
+		rcu_qsctr_inc(cpu);
+		rcu_bh_qsctr_inc(cpu);
+	} else if (!in_softirq())
+		rcu_bh_qsctr_inc(cpu);
+	raise_rcu_softirq();
+}
+
+static void rcu_init_percpu_data(int cpu, struct rcu_ctrlblk *rcp,
+						struct rcu_data *rdp)
+{
+	memset(rdp, 0, sizeof(*rdp));
+	rdp->curtail = &rdp->curlist;
+	rdp->nxttail = &rdp->nxtlist;
+	rdp->donetail = &rdp->donelist;
+	rdp->quiescbatch = rcp->completed;
+	rdp->qs_pending = 0;
+	rdp->cpu = cpu;
+	rdp->blimit = blimit;
+}
+
+static void __cpuinit rcu_online_cpu(int cpu)
+{
+	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
+	struct rcu_data *bh_rdp = &per_cpu(rcu_bh_data, cpu);
+
+	rcu_init_percpu_data(cpu, &rcu_ctrlblk, rdp);
+	rcu_init_percpu_data(cpu, &rcu_bh_ctrlblk, bh_rdp);
+	open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL);
+}
+
+static int __cpuinit rcu_cpu_notify(struct notifier_block *self,
+				unsigned long action, void *hcpu)
+{
+	long cpu = (long)hcpu;
+
+	switch (action) {
+	case CPU_UP_PREPARE:
+	case CPU_UP_PREPARE_FROZEN:
+		rcu_online_cpu(cpu);
+		break;
+	case CPU_DEAD:
+	case CPU_DEAD_FROZEN:
+		rcu_offline_cpu(cpu);
+		break;
+	default:
+		break;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block __cpuinitdata rcu_nb = {
+	.notifier_call	= rcu_cpu_notify,
+};
+
+/*
+ * Initializes rcu mechanism.  Assumed to be called early.
+ * That is before local timer(SMP) or jiffie timer (uniproc) is setup.
+ * Note that rcu_qsctr and friends are implicitly
+ * initialized due to the choice of ``0'' for RCU_CTR_INVALID.
+ */
+void __init __rcu_init(void)
+{
+	rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE,
+			(void *)(long)smp_processor_id());
+	/* Register notifier for non-boot CPUs */
+	register_cpu_notifier(&rcu_nb);
+}
+
+module_param(blimit, int, 0);
+module_param(qhimark, int, 0);
+module_param(qlowmark, int, 0);

kernel/rcupdate.c

@@ -15,7 +15,7 @@
  * along with this program; if not, write to the Free Software
  * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  *
- * Copyright (C) IBM Corporation, 2001
+ * Copyright IBM Corporation, 2001
  *
  * Authors: Dipankar Sarma <dipankar@in.ibm.com>
  *	    Manfred Spraul <manfred@colorfullife.com>
@@ -35,165 +35,57 @@
 #include <linux/init.h>
 #include <linux/spinlock.h>
 #include <linux/smp.h>
-#include <linux/rcupdate.h>
 #include <linux/interrupt.h>
 #include <linux/sched.h>
 #include <asm/atomic.h>
 #include <linux/bitops.h>
-#include <linux/module.h>
 #include <linux/completion.h>
-#include <linux/moduleparam.h>
 #include <linux/percpu.h>
 #include <linux/notifier.h>
 #include <linux/cpu.h>
 #include <linux/mutex.h>
+#include <linux/module.h>
 
-#ifdef CONFIG_DEBUG_LOCK_ALLOC
-static struct lock_class_key rcu_lock_key;
-struct lockdep_map rcu_lock_map =
-	STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key);
-
-EXPORT_SYMBOL_GPL(rcu_lock_map);
-#endif
-
-/* Definition for rcupdate control block. */
-static struct rcu_ctrlblk rcu_ctrlblk = {
-	.cur = -300,
-	.completed = -300,
-	.lock = __SPIN_LOCK_UNLOCKED(&rcu_ctrlblk.lock),
-	.cpumask = CPU_MASK_NONE,
-};
-static struct rcu_ctrlblk rcu_bh_ctrlblk = {
-	.cur = -300,
-	.completed = -300,
-	.lock = __SPIN_LOCK_UNLOCKED(&rcu_bh_ctrlblk.lock),
-	.cpumask = CPU_MASK_NONE,
+struct rcu_synchronize {
+	struct rcu_head head;
+	struct completion completion;
 };
 
-DEFINE_PER_CPU(struct rcu_data, rcu_data) = { 0L };
-DEFINE_PER_CPU(struct rcu_data, rcu_bh_data) = { 0L };
-
-/* Fake initialization required by compiler */
-static DEFINE_PER_CPU(struct tasklet_struct, rcu_tasklet) = {NULL};
-static int blimit = 10;
-static int qhimark = 10000;
-static int qlowmark = 100;
-
+static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL};
 static atomic_t rcu_barrier_cpu_count;
 static DEFINE_MUTEX(rcu_barrier_mutex);
 static struct completion rcu_barrier_completion;
 
-#ifdef CONFIG_SMP
-static void force_quiescent_state(struct rcu_data *rdp,
-			struct rcu_ctrlblk *rcp)
-{
-	int cpu;
-	cpumask_t cpumask;
-	set_need_resched();
-	if (unlikely(!rcp->signaled)) {
-		rcp->signaled = 1;
-		/*
-		 * Don't send IPI to itself. With irqs disabled,
-		 * rdp->cpu is the current cpu.
-		 */
-		cpumask = rcp->cpumask;
-		cpu_clear(rdp->cpu, cpumask);
-		for_each_cpu_mask(cpu, cpumask)
-			smp_send_reschedule(cpu);
-	}
-}
-#else
-static inline void force_quiescent_state(struct rcu_data *rdp,
-			struct rcu_ctrlblk *rcp)
+/* Because of FASTCALL declaration of complete, we use this wrapper */
+static void wakeme_after_rcu(struct rcu_head  *head)
 {
-	set_need_resched();
+	struct rcu_synchronize *rcu;
+
+	rcu = container_of(head, struct rcu_synchronize, head);
+	complete(&rcu->completion);
 }
-#endif
 
 /**
- * call_rcu - Queue an RCU callback for invocation after a grace period.
- * @head: structure to be used for queueing the RCU updates.
- * @func: actual update function to be invoked after the grace period
+ * synchronize_rcu - wait until a grace period has elapsed.
  *
- * The update function will be invoked some time after a full grace
- * period elapses, in other words after all currently executing RCU
+ * Control will return to the caller some time after a full grace
+ * period has elapsed, in other words after all currently executing RCU
  * read-side critical sections have completed.  RCU read-side critical
  * sections are delimited by rcu_read_lock() and rcu_read_unlock(),
  * and may be nested.
  */
-void fastcall call_rcu(struct rcu_head *head,
-				void (*func)(struct rcu_head *rcu))
-{
-	unsigned long flags;
-	struct rcu_data *rdp;
-
-	head->func = func;
-	head->next = NULL;
-	local_irq_save(flags);
-	rdp = &__get_cpu_var(rcu_data);
-	*rdp->nxttail = head;
-	rdp->nxttail = &head->next;
-	if (unlikely(++rdp->qlen > qhimark)) {
-		rdp->blimit = INT_MAX;
-		force_quiescent_state(rdp, &rcu_ctrlblk);
-	}
-	local_irq_restore(flags);
-}
-
-/**
- * call_rcu_bh - Queue an RCU for invocation after a quicker grace period.
- * @head: structure to be used for queueing the RCU updates.
- * @func: actual update function to be invoked after the grace period
- *
- * The update function will be invoked some time after a full grace
- * period elapses, in other words after all currently executing RCU
- * read-side critical sections have completed. call_rcu_bh() assumes
- * that the read-side critical sections end on completion of a softirq
- * handler. This means that read-side critical sections in process
- * context must not be interrupted by softirqs. This interface is to be
- * used when most of the read-side critical sections are in softirq context.
- * RCU read-side critical sections are delimited by rcu_read_lock() and
- * rcu_read_unlock(), * if in interrupt context or rcu_read_lock_bh()
- * and rcu_read_unlock_bh(), if in process context. These may be nested.
- */
-void fastcall call_rcu_bh(struct rcu_head *head,
-				void (*func)(struct rcu_head *rcu))
+void synchronize_rcu(void)
 {
-	unsigned long flags;
-	struct rcu_data *rdp;
-
-	head->func = func;
-	head->next = NULL;
-	local_irq_save(flags);
-	rdp = &__get_cpu_var(rcu_bh_data);
-	*rdp->nxttail = head;
-	rdp->nxttail = &head->next;
-
-	if (unlikely(++rdp->qlen > qhimark)) {
-		rdp->blimit = INT_MAX;
-		force_quiescent_state(rdp, &rcu_bh_ctrlblk);
-	}
-
-	local_irq_restore(flags);
-}
+	struct rcu_synchronize rcu;
 
-/*
- * Return the number of RCU batches processed thus far.  Useful
- * for debug and statistics.
- */
-long rcu_batches_completed(void)
-{
-	return rcu_ctrlblk.completed;
-}
+	init_completion(&rcu.completion);
+	/* Will wake me after RCU finished */
+	call_rcu(&rcu.head, wakeme_after_rcu);
 
-/*
- * Return the number of RCU batches processed thus far.  Useful
- * for debug and statistics.
- */
-long rcu_batches_completed_bh(void)
-{
-	return rcu_bh_ctrlblk.completed;
+	/* Wait for it */
+	wait_for_completion(&rcu.completion);
 }
+EXPORT_SYMBOL_GPL(synchronize_rcu);
 
 static void rcu_barrier_callback(struct rcu_head *notused)
 {
@@ -207,10 +99,8 @@ static void rcu_barrier_callback(struct rcu_head *notused)
 static void rcu_barrier_func(void *notused)
 {
 	int cpu = smp_processor_id();
-	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
-	struct rcu_head *head;
+	struct rcu_head *head = &per_cpu(rcu_barrier_head, cpu);
 
-	head = &rdp->barrier;
 	atomic_inc(&rcu_barrier_cpu_count);
 	call_rcu(head, rcu_barrier_callback);
 }
@@ -225,420 +115,24 @@ void rcu_barrier(void)
 	mutex_lock(&rcu_barrier_mutex);
 	init_completion(&rcu_barrier_completion);
 	atomic_set(&rcu_barrier_cpu_count, 0);
+	/*
+	 * The queueing of callbacks in all CPUs must be atomic with
+	 * respect to RCU, otherwise one CPU may queue a callback,
+	 * wait for a grace period, decrement barrier count and call
+	 * complete(), while other CPUs have not yet queued anything.
+	 * So, we need to make sure that grace periods cannot complete
+	 * until all the callbacks are queued.
+	 */
+	rcu_read_lock();
 	on_each_cpu(rcu_barrier_func, NULL, 0, 1);
+	rcu_read_unlock();
 	wait_for_completion(&rcu_barrier_completion);
 	mutex_unlock(&rcu_barrier_mutex);
 }
 EXPORT_SYMBOL_GPL(rcu_barrier);
 
-/*
- * Invoke the completed RCU callbacks. They are expected to be in
- * a per-cpu list.
- */
-static void rcu_do_batch(struct rcu_data *rdp)
-{
-	struct rcu_head *next, *list;
-	int count = 0;
-
-	list = rdp->donelist;
-	while (list) {
-		next = list->next;
-		prefetch(next);
-		list->func(list);
-		list = next;
-		if (++count >= rdp->blimit)
-			break;
-	}
-	rdp->donelist = list;
-
-	local_irq_disable();
-	rdp->qlen -= count;
-	local_irq_enable();
-	if (rdp->blimit == INT_MAX && rdp->qlen <= qlowmark)
-		rdp->blimit = blimit;
-
-	if (!rdp->donelist)
-		rdp->donetail = &rdp->donelist;
-	else
-		tasklet_schedule(&per_cpu(rcu_tasklet, rdp->cpu));
-}
-
-/*
- * Grace period handling:
- * The grace period handling consists out of two steps:
- * - A new grace period is started.
- *   This is done by rcu_start_batch. The start is not broadcasted to
- *   all cpus, they must pick this up by comparing rcp->cur with
- *   rdp->quiescbatch. All cpus are recorded  in the
- *   rcu_ctrlblk.cpumask bitmap.
- * - All cpus must go through a quiescent state.
- *   Since the start of the grace period is not broadcasted, at least two
- *   calls to rcu_check_quiescent_state are required:
- *   The first call just notices that a new grace period is running. The
- *   following calls check if there was a quiescent state since the beginning
- *   of the grace period. If so, it updates rcu_ctrlblk.cpumask. If
- *   the bitmap is empty, then the grace period is completed.
- *   rcu_check_quiescent_state calls rcu_start_batch(0) to start the next grace
- *   period (if necessary).
- */
-/*
- * Register a new batch of callbacks, and start it up if there is currently no
- * active batch and the batch to be registered has not already occurred.
- * Caller must hold rcu_ctrlblk.lock.
- */
-static void rcu_start_batch(struct rcu_ctrlblk *rcp)
-{
-	if (rcp->next_pending &&
-			rcp->completed == rcp->cur) {
-		rcp->next_pending = 0;
-		/*
-		 * next_pending == 0 must be visible in
-		 * __rcu_process_callbacks() before it can see new value of cur.
-		 */
-		smp_wmb();
-		rcp->cur++;
-
-		/*
-		 * Accessing nohz_cpu_mask before incrementing rcp->cur needs a
-		 * Barrier  Otherwise it can cause tickless idle CPUs to be
-		 * included in rcp->cpumask, which will extend graceperiods
-		 * unnecessarily.
-		 */
-		smp_mb();
-		cpus_andnot(rcp->cpumask, cpu_online_map, nohz_cpu_mask);
-
-		rcp->signaled = 0;
-	}
-}
-
-/*
- * cpu went through a quiescent state since the beginning of the grace period.
- * Clear it from the cpu mask and complete the grace period if it was the last
- * cpu. Start another grace period if someone has further entries pending
- */
-static void cpu_quiet(int cpu, struct rcu_ctrlblk *rcp)
-{
-	cpu_clear(cpu, rcp->cpumask);
-	if (cpus_empty(rcp->cpumask)) {
-		/* batch completed ! */
-		rcp->completed = rcp->cur;
-		rcu_start_batch(rcp);
-	}
-}
-
-/*
- * Check if the cpu has gone through a quiescent state (say context
- * switch). If so and if it already hasn't done so in this RCU
- * quiescent cycle, then indicate that it has done so.
- */
-static void rcu_check_quiescent_state(struct rcu_ctrlblk *rcp,
-					struct rcu_data *rdp)
-{
-	if (rdp->quiescbatch != rcp->cur) {
-		/* start new grace period: */
-		rdp->qs_pending = 1;
-		rdp->passed_quiesc = 0;
-		rdp->quiescbatch = rcp->cur;
-		return;
-	}
-
-	/* Grace period already completed for this cpu?
-	 * qs_pending is checked instead of the actual bitmap to avoid
-	 * cacheline trashing.
-	 */
-	if (!rdp->qs_pending)
-		return;
-
-	/* 
-	 * Was there a quiescent state since the beginning of the grace
-	 * period? If no, then exit and wait for the next call.
-	 */
-	if (!rdp->passed_quiesc)
-		return;
-	rdp->qs_pending = 0;
-
-	spin_lock(&rcp->lock);
-	/*
-	 * rdp->quiescbatch/rcp->cur and the cpu bitmap can come out of sync
-	 * during cpu startup. Ignore the quiescent state.
-	 */
-	if (likely(rdp->quiescbatch == rcp->cur))
-		cpu_quiet(rdp->cpu, rcp);
-
-	spin_unlock(&rcp->lock);
-}
-
-
-#ifdef CONFIG_HOTPLUG_CPU
-
-/* warning! helper for rcu_offline_cpu. do not use elsewhere without reviewing
- * locking requirements, the list it's pulling from has to belong to a cpu
- * which is dead and hence not processing interrupts.
- */
-static void rcu_move_batch(struct rcu_data *this_rdp, struct rcu_head *list,
-				struct rcu_head **tail)
-{
-	local_irq_disable();
-	*this_rdp->nxttail = list;
-	if (list)
-		this_rdp->nxttail = tail;
-	local_irq_enable();
-}
-
-static void __rcu_offline_cpu(struct rcu_data *this_rdp,
-				struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
-{
-	/* if the cpu going offline owns the grace period
-	 * we can block indefinitely waiting for it, so flush
-	 * it here
-	 */
-	spin_lock_bh(&rcp->lock);
-	if (rcp->cur != rcp->completed)
-		cpu_quiet(rdp->cpu, rcp);
-	spin_unlock_bh(&rcp->lock);
-	rcu_move_batch(this_rdp, rdp->curlist, rdp->curtail);
-	rcu_move_batch(this_rdp, rdp->nxtlist, rdp->nxttail);
-	rcu_move_batch(this_rdp, rdp->donelist, rdp->donetail);
-}
-
-static void rcu_offline_cpu(int cpu)
-{
-	struct rcu_data *this_rdp = &get_cpu_var(rcu_data);
-	struct rcu_data *this_bh_rdp = &get_cpu_var(rcu_bh_data);
-
-	__rcu_offline_cpu(this_rdp, &rcu_ctrlblk,
-					&per_cpu(rcu_data, cpu));
-	__rcu_offline_cpu(this_bh_rdp, &rcu_bh_ctrlblk,
-					&per_cpu(rcu_bh_data, cpu));
-	put_cpu_var(rcu_data);
-	put_cpu_var(rcu_bh_data);
-	tasklet_kill_immediate(&per_cpu(rcu_tasklet, cpu), cpu);
-}
-
-#else
-
-static void rcu_offline_cpu(int cpu)
-{
-}
-
-#endif
-
-/*
- * This does the RCU processing work from tasklet context. 
- */
-static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp,
-					struct rcu_data *rdp)
-{
-	if (rdp->curlist && !rcu_batch_before(rcp->completed, rdp->batch)) {
-		*rdp->donetail = rdp->curlist;
-		rdp->donetail = rdp->curtail;
-		rdp->curlist = NULL;
-		rdp->curtail = &rdp->curlist;
-	}
-
-	if (rdp->nxtlist && !rdp->curlist) {
-		local_irq_disable();
-		rdp->curlist = rdp->nxtlist;
-		rdp->curtail = rdp->nxttail;
-		rdp->nxtlist = NULL;
-		rdp->nxttail = &rdp->nxtlist;
-		local_irq_enable();
-
-		/*
-		 * start the next batch of callbacks
-		 */
-
-		/* determine batch number */
-		rdp->batch = rcp->cur + 1;
-		/* see the comment and corresponding wmb() in
-		 * the rcu_start_batch()
-		 */
-		smp_rmb();
-
-		if (!rcp->next_pending) {
-			/* and start it/schedule start if it's a new batch */
-			spin_lock(&rcp->lock);
-			rcp->next_pending = 1;
-			rcu_start_batch(rcp);
-			spin_unlock(&rcp->lock);
-		}
-	}
-
-	rcu_check_quiescent_state(rcp, rdp);
-	if (rdp->donelist)
-		rcu_do_batch(rdp);
-}
-
-static void rcu_process_callbacks(unsigned long unused)
-{
-	__rcu_process_callbacks(&rcu_ctrlblk, &__get_cpu_var(rcu_data));
-	__rcu_process_callbacks(&rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data));
-}
-
-static int __rcu_pending(struct rcu_ctrlblk *rcp, struct rcu_data *rdp)
-{
-	/* This cpu has pending rcu entries and the grace period
-	 * for them has completed.
-	 */
-	if (rdp->curlist && !rcu_batch_before(rcp->completed, rdp->batch))
-		return 1;
-
-	/* This cpu has no pending entries, but there are new entries */
-	if (!rdp->curlist && rdp->nxtlist)
-		return 1;
-
-	/* This cpu has finished callbacks to invoke */
-	if (rdp->donelist)
-		return 1;
-
-	/* The rcu core waits for a quiescent state from the cpu */
-	if (rdp->quiescbatch != rcp->cur || rdp->qs_pending)
-		return 1;
-
-	/* nothing to do */
-	return 0;
-}
-
-/*
- * Check to see if there is any immediate RCU-related work to be done
- * by the current CPU, returning 1 if so.  This function is part of the
- * RCU implementation; it is -not- an exported member of the RCU API.
- */
-int rcu_pending(int cpu)
-{
-	return __rcu_pending(&rcu_ctrlblk, &per_cpu(rcu_data, cpu)) ||
-		__rcu_pending(&rcu_bh_ctrlblk, &per_cpu(rcu_bh_data, cpu));
-}
-
-/*
- * Check to see if any future RCU-related work will need to be done
- * by the current CPU, even if none need be done immediately, returning
- * 1 if so.  This function is part of the RCU implementation; it is -not-
- * an exported member of the RCU API.
- */
-int rcu_needs_cpu(int cpu)
-{
-	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
-	struct rcu_data *rdp_bh = &per_cpu(rcu_bh_data, cpu);
-
-	return (!!rdp->curlist || !!rdp_bh->curlist || rcu_pending(cpu));
-}
-
-void rcu_check_callbacks(int cpu, int user)
-{
-	if (user || 
-	    (idle_cpu(cpu) && !in_softirq() && 
-				hardirq_count() <= (1 << HARDIRQ_SHIFT))) {
-		rcu_qsctr_inc(cpu);
-		rcu_bh_qsctr_inc(cpu);
-	} else if (!in_softirq())
-		rcu_bh_qsctr_inc(cpu);
-	tasklet_schedule(&per_cpu(rcu_tasklet, cpu));
-}
-
-static void rcu_init_percpu_data(int cpu, struct rcu_ctrlblk *rcp,
-						struct rcu_data *rdp)
-{
-	memset(rdp, 0, sizeof(*rdp));
-	rdp->curtail = &rdp->curlist;
-	rdp->nxttail = &rdp->nxtlist;
-	rdp->donetail = &rdp->donelist;
-	rdp->quiescbatch = rcp->completed;
-	rdp->qs_pending = 0;
-	rdp->cpu = cpu;
-	rdp->blimit = blimit;
-}
-
-static void __cpuinit rcu_online_cpu(int cpu)
-{
-	struct rcu_data *rdp = &per_cpu(rcu_data, cpu);
-	struct rcu_data *bh_rdp = &per_cpu(rcu_bh_data, cpu);
-
-	rcu_init_percpu_data(cpu, &rcu_ctrlblk, rdp);
-	rcu_init_percpu_data(cpu, &rcu_bh_ctrlblk, bh_rdp);
-	tasklet_init(&per_cpu(rcu_tasklet, cpu), rcu_process_callbacks, 0UL);
-}
-
-static int __cpuinit rcu_cpu_notify(struct notifier_block *self,
-				unsigned long action, void *hcpu)
-{
-	long cpu = (long)hcpu;
-	switch (action) {
-	case CPU_UP_PREPARE:
-	case CPU_UP_PREPARE_FROZEN:
-		rcu_online_cpu(cpu);
-		break;
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		rcu_offline_cpu(cpu);
-		break;
-	default:
-		break;
-	}
-	return NOTIFY_OK;
-}
-
-static struct notifier_block __cpuinitdata rcu_nb = {
-	.notifier_call	= rcu_cpu_notify,
-};
-
-/*
- * Initializes rcu mechanism.  Assumed to be called early.
- * That is before local timer(SMP) or jiffie timer (uniproc) is setup.
- * Note that rcu_qsctr and friends are implicitly
- * initialized due to the choice of ``0'' for RCU_CTR_INVALID.
- */
 void __init rcu_init(void)
 {
-	rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE,
-			(void *)(long)smp_processor_id());
-	/* Register notifier for non-boot CPUs */
-	register_cpu_notifier(&rcu_nb);
-}
-
-struct rcu_synchronize {
-	struct rcu_head head;
-	struct completion completion;
-};
-
-/* Because of FASTCALL declaration of complete, we use this wrapper */
-static void wakeme_after_rcu(struct rcu_head  *head)
-{
-	struct rcu_synchronize *rcu;
-
-	rcu = container_of(head, struct rcu_synchronize, head);
-	complete(&rcu->completion);
+	__rcu_init();
 }
 
-/**
- * synchronize_rcu - wait until a grace period has elapsed.
- *
- * Control will return to the caller some time after a full grace
- * period has elapsed, in other words after all currently executing RCU
- * read-side critical sections have completed.  RCU read-side critical
- * sections are delimited by rcu_read_lock() and rcu_read_unlock(),
- * and may be nested.
- *
- * If your read-side code is not protected by rcu_read_lock(), do -not-
- * use synchronize_rcu().
- */
-void synchronize_rcu(void)
-{
-	struct rcu_synchronize rcu;
-
-	init_completion(&rcu.completion);
-	/* Will wake me after RCU finished */
-	call_rcu(&rcu.head, wakeme_after_rcu);
-
-	/* Wait for it */
-	wait_for_completion(&rcu.completion);
-}
-
-module_param(blimit, int, 0);
-module_param(qhimark, int, 0);
-module_param(qlowmark, int, 0);
-EXPORT_SYMBOL_GPL(rcu_batches_completed);
-EXPORT_SYMBOL_GPL(rcu_batches_completed_bh);
-EXPORT_SYMBOL_GPL(call_rcu);
-EXPORT_SYMBOL_GPL(call_rcu_bh);
-EXPORT_SYMBOL_GPL(synchronize_rcu);

kernel/rcupreempt.c

@@ -0,0 +1,953 @@
+/*
+ * Read-Copy Update mechanism for mutual exclusion, realtime implementation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright IBM Corporation, 2006
+ *
+ * Authors: Paul E. McKenney <paulmck@us.ibm.com>
+ *		With thanks to Esben Nielsen, Bill Huey, and Ingo Molnar
+ *		for pushing me away from locks and towards counters, and
+ *		to Suparna Bhattacharya for pushing me completely away
+ *		from atomic instructions on the read side.
+ *
+ * Papers:  http://www.rdrop.com/users/paulmck/RCU
+ *
+ * Design Document: http://lwn.net/Articles/253651/
+ *
+ * For detailed explanation of Read-Copy Update mechanism see -
+ * 		Documentation/RCU/ *.txt
+ *
+ */
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/spinlock.h>
+#include <linux/smp.h>
+#include <linux/rcupdate.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <asm/atomic.h>
+#include <linux/bitops.h>
+#include <linux/module.h>
+#include <linux/completion.h>
+#include <linux/moduleparam.h>
+#include <linux/percpu.h>
+#include <linux/notifier.h>
+#include <linux/rcupdate.h>
+#include <linux/cpu.h>
+#include <linux/random.h>
+#include <linux/delay.h>
+#include <linux/byteorder/swabb.h>
+#include <linux/cpumask.h>
+#include <linux/rcupreempt_trace.h>
+
+/*
+ * Macro that prevents the compiler from reordering accesses, but does
+ * absolutely -nothing- to prevent CPUs from reordering.  This is used
+ * only to mediate communication between mainline code and hardware
+ * interrupt and NMI handlers.
+ */
+#define ACCESS_ONCE(x) (*(volatile typeof(x) *)&(x))
+
+/*
+ * PREEMPT_RCU data structures.
+ */
+
+/*
+ * GP_STAGES specifies the number of times the state machine has
+ * to go through the all the rcu_try_flip_states (see below)
+ * in a single Grace Period.
+ *
+ * GP in GP_STAGES stands for Grace Period ;)
+ */
+#define GP_STAGES    2
+struct rcu_data {
+	spinlock_t	lock;		/* Protect rcu_data fields. */
+	long		completed;	/* Number of last completed batch. */
+	int		waitlistcount;
+	struct tasklet_struct rcu_tasklet;
+	struct rcu_head *nextlist;
+	struct rcu_head **nexttail;
+	struct rcu_head *waitlist[GP_STAGES];
+	struct rcu_head **waittail[GP_STAGES];
+	struct rcu_head *donelist;
+	struct rcu_head **donetail;
+	long rcu_flipctr[2];
+#ifdef CONFIG_RCU_TRACE
+	struct rcupreempt_trace trace;
+#endif /* #ifdef CONFIG_RCU_TRACE */
+};
+
+/*
+ * States for rcu_try_flip() and friends.
+ */
+
+enum rcu_try_flip_states {
+
+	/*
+	 * Stay here if nothing is happening. Flip the counter if somthing
+	 * starts happening. Denoted by "I"
+	 */
+	rcu_try_flip_idle_state,
+
+	/*
+	 * Wait here for all CPUs to notice that the counter has flipped. This
+	 * prevents the old set of counters from ever being incremented once
+	 * we leave this state, which in turn is necessary because we cannot
+	 * test any individual counter for zero -- we can only check the sum.
+	 * Denoted by "A".
+	 */
+	rcu_try_flip_waitack_state,
+
+	/*
+	 * Wait here for the sum of the old per-CPU counters to reach zero.
+	 * Denoted by "Z".
+	 */
+	rcu_try_flip_waitzero_state,
+
+	/*
+	 * Wait here for each of the other CPUs to execute a memory barrier.
+	 * This is necessary to ensure that these other CPUs really have
+	 * completed executing their RCU read-side critical sections, despite
+	 * their CPUs wildly reordering memory. Denoted by "M".
+	 */
+	rcu_try_flip_waitmb_state,
+};
+
+struct rcu_ctrlblk {
+	spinlock_t	fliplock;	/* Protect state-machine transitions. */
+	long		completed;	/* Number of last completed batch. */
+	enum rcu_try_flip_states rcu_try_flip_state; /* The current state of
+							the rcu state machine */
+};
+
+static DEFINE_PER_CPU(struct rcu_data, rcu_data);
+static struct rcu_ctrlblk rcu_ctrlblk = {
+	.fliplock = __SPIN_LOCK_UNLOCKED(rcu_ctrlblk.fliplock),
+	.completed = 0,
+	.rcu_try_flip_state = rcu_try_flip_idle_state,
+};
+
+
+#ifdef CONFIG_RCU_TRACE
+static char *rcu_try_flip_state_names[] =
+	{ "idle", "waitack", "waitzero", "waitmb" };
+#endif /* #ifdef CONFIG_RCU_TRACE */
+
+static cpumask_t rcu_cpu_online_map __read_mostly = CPU_MASK_NONE;
+
+/*
+ * Enum and per-CPU flag to determine when each CPU has seen
+ * the most recent counter flip.
+ */
+
+enum rcu_flip_flag_values {
+	rcu_flip_seen,		/* Steady/initial state, last flip seen. */
+				/* Only GP detector can update. */
+	rcu_flipped		/* Flip just completed, need confirmation. */
+				/* Only corresponding CPU can update. */
+};
+static DEFINE_PER_CPU_SHARED_ALIGNED(enum rcu_flip_flag_values, rcu_flip_flag)
+								= rcu_flip_seen;
+
+/*
+ * Enum and per-CPU flag to determine when each CPU has executed the
+ * needed memory barrier to fence in memory references from its last RCU
+ * read-side critical section in the just-completed grace period.
+ */
+
+enum rcu_mb_flag_values {
+	rcu_mb_done,		/* Steady/initial state, no mb()s required. */
+				/* Only GP detector can update. */
+	rcu_mb_needed		/* Flip just completed, need an mb(). */
+				/* Only corresponding CPU can update. */
+};
+static DEFINE_PER_CPU_SHARED_ALIGNED(enum rcu_mb_flag_values, rcu_mb_flag)
+								= rcu_mb_done;
+
+/*
+ * RCU_DATA_ME: find the current CPU's rcu_data structure.
+ * RCU_DATA_CPU: find the specified CPU's rcu_data structure.
+ */
+#define RCU_DATA_ME()		(&__get_cpu_var(rcu_data))
+#define RCU_DATA_CPU(cpu)	(&per_cpu(rcu_data, cpu))
+
+/*
+ * Helper macro for tracing when the appropriate rcu_data is not
+ * cached in a local variable, but where the CPU number is so cached.
+ */
+#define RCU_TRACE_CPU(f, cpu) RCU_TRACE(f, &(RCU_DATA_CPU(cpu)->trace));
+
+/*
+ * Helper macro for tracing when the appropriate rcu_data is not
+ * cached in a local variable.
+ */
+#define RCU_TRACE_ME(f) RCU_TRACE(f, &(RCU_DATA_ME()->trace));
+
+/*
+ * Helper macro for tracing when the appropriate rcu_data is pointed
+ * to by a local variable.
+ */
+#define RCU_TRACE_RDP(f, rdp) RCU_TRACE(f, &((rdp)->trace));
+
+/*
+ * Return the number of RCU batches processed thus far.  Useful
+ * for debug and statistics.
+ */
+long rcu_batches_completed(void)
+{
+	return rcu_ctrlblk.completed;
+}
+EXPORT_SYMBOL_GPL(rcu_batches_completed);
+
+EXPORT_SYMBOL_GPL(rcu_batches_completed_bh);
+
+void __rcu_read_lock(void)
+{
+	int idx;
+	struct task_struct *t = current;
+	int nesting;
+
+	nesting = ACCESS_ONCE(t->rcu_read_lock_nesting);
+	if (nesting != 0) {
+
+		/* An earlier rcu_read_lock() covers us, just count it. */
+
+		t->rcu_read_lock_nesting = nesting + 1;
+
+	} else {
+		unsigned long flags;
+
+		/*
+		 * We disable interrupts for the following reasons:
+		 * - If we get scheduling clock interrupt here, and we
+		 *   end up acking the counter flip, it's like a promise
+		 *   that we will never increment the old counter again.
+		 *   Thus we will break that promise if that
+		 *   scheduling clock interrupt happens between the time
+		 *   we pick the .completed field and the time that we
+		 *   increment our counter.
+		 *
+		 * - We don't want to be preempted out here.
+		 *
+		 * NMIs can still occur, of course, and might themselves
+		 * contain rcu_read_lock().
+		 */
+
+		local_irq_save(flags);
+
+		/*
+		 * Outermost nesting of rcu_read_lock(), so increment
+		 * the current counter for the current CPU.  Use volatile
+		 * casts to prevent the compiler from reordering.
+		 */
+
+		idx = ACCESS_ONCE(rcu_ctrlblk.completed) & 0x1;
+		ACCESS_ONCE(RCU_DATA_ME()->rcu_flipctr[idx])++;
+
+		/*
+		 * Now that the per-CPU counter has been incremented, we
+		 * are protected from races with rcu_read_lock() invoked
+		 * from NMI handlers on this CPU.  We can therefore safely
+		 * increment the nesting counter, relieving further NMIs
+		 * of the need to increment the per-CPU counter.
+		 */
+
+		ACCESS_ONCE(t->rcu_read_lock_nesting) = nesting + 1;
+
+		/*
+		 * Now that we have preventing any NMIs from storing
+		 * to the ->rcu_flipctr_idx, we can safely use it to
+		 * remember which counter to decrement in the matching
+		 * rcu_read_unlock().
+		 */
+
+		ACCESS_ONCE(t->rcu_flipctr_idx) = idx;
+		local_irq_restore(flags);
+	}
+}
+EXPORT_SYMBOL_GPL(__rcu_read_lock);
+
+void __rcu_read_unlock(void)
+{
+	int idx;
+	struct task_struct *t = current;
+	int nesting;
+
+	nesting = ACCESS_ONCE(t->rcu_read_lock_nesting);
+	if (nesting > 1) {
+
+		/*
+		 * We are still protected by the enclosing rcu_read_lock(),
+		 * so simply decrement the counter.
+		 */
+
+		t->rcu_read_lock_nesting = nesting - 1;
+
+	} else {
+		unsigned long flags;
+
+		/*
+		 * Disable local interrupts to prevent the grace-period
+		 * detection state machine from seeing us half-done.
+		 * NMIs can still occur, of course, and might themselves
+		 * contain rcu_read_lock() and rcu_read_unlock().
+		 */
+
+		local_irq_save(flags);
+
+		/*
+		 * Outermost nesting of rcu_read_unlock(), so we must
+		 * decrement the current counter for the current CPU.
+		 * This must be done carefully, because NMIs can
+		 * occur at any point in this code, and any rcu_read_lock()
+		 * and rcu_read_unlock() pairs in the NMI handlers
+		 * must interact non-destructively with this code.
+		 * Lots of volatile casts, and -very- careful ordering.
+		 *
+		 * Changes to this code, including this one, must be
+		 * inspected, validated, and tested extremely carefully!!!
+		 */
+
+		/*
+		 * First, pick up the index.
+		 */
+
+		idx = ACCESS_ONCE(t->rcu_flipctr_idx);
+
+		/*
+		 * Now that we have fetched the counter index, it is
+		 * safe to decrement the per-task RCU nesting counter.
+		 * After this, any interrupts or NMIs will increment and
+		 * decrement the per-CPU counters.
+		 */
+		ACCESS_ONCE(t->rcu_read_lock_nesting) = nesting - 1;
+
+		/*
+		 * It is now safe to decrement this task's nesting count.
+		 * NMIs that occur after this statement will route their
+		 * rcu_read_lock() calls through this "else" clause, and
+		 * will thus start incrementing the per-CPU counter on
+		 * their own.  They will also clobber ->rcu_flipctr_idx,
+		 * but that is OK, since we have already fetched it.
+		 */
+
+		ACCESS_ONCE(RCU_DATA_ME()->rcu_flipctr[idx])--;
+		local_irq_restore(flags);
+	}
+}
+EXPORT_SYMBOL_GPL(__rcu_read_unlock);
+
+/*
+ * If a global counter flip has occurred since the last time that we
+ * advanced callbacks, advance them.  Hardware interrupts must be
+ * disabled when calling this function.
+ */
+static void __rcu_advance_callbacks(struct rcu_data *rdp)
+{
+	int cpu;
+	int i;
+	int wlc = 0;
+
+	if (rdp->completed != rcu_ctrlblk.completed) {
+		if (rdp->waitlist[GP_STAGES - 1] != NULL) {
+			*rdp->donetail = rdp->waitlist[GP_STAGES - 1];
+			rdp->donetail = rdp->waittail[GP_STAGES - 1];
+			RCU_TRACE_RDP(rcupreempt_trace_move2done, rdp);
+		}
+		for (i = GP_STAGES - 2; i >= 0; i--) {
+			if (rdp->waitlist[i] != NULL) {
+				rdp->waitlist[i + 1] = rdp->waitlist[i];
+				rdp->waittail[i + 1] = rdp->waittail[i];
+				wlc++;
+			} else {
+				rdp->waitlist[i + 1] = NULL;
+				rdp->waittail[i + 1] =
+					&rdp->waitlist[i + 1];
+			}
+		}
+		if (rdp->nextlist != NULL) {
+			rdp->waitlist[0] = rdp->nextlist;
+			rdp->waittail[0] = rdp->nexttail;
+			wlc++;
+			rdp->nextlist = NULL;
+			rdp->nexttail = &rdp->nextlist;
+			RCU_TRACE_RDP(rcupreempt_trace_move2wait, rdp);
+		} else {
+			rdp->waitlist[0] = NULL;
+			rdp->waittail[0] = &rdp->waitlist[0];
+		}
+		rdp->waitlistcount = wlc;
+		rdp->completed = rcu_ctrlblk.completed;
+	}
+
+	/*
+	 * Check to see if this CPU needs to report that it has seen
+	 * the most recent counter flip, thereby declaring that all
+	 * subsequent rcu_read_lock() invocations will respect this flip.
+	 */
+
+	cpu = raw_smp_processor_id();
+	if (per_cpu(rcu_flip_flag, cpu) == rcu_flipped) {
+		smp_mb();  /* Subsequent counter accesses must see new value */
+		per_cpu(rcu_flip_flag, cpu) = rcu_flip_seen;
+		smp_mb();  /* Subsequent RCU read-side critical sections */
+			   /*  seen -after- acknowledgement. */
+	}
+}
+
+/*
+ * Get here when RCU is idle.  Decide whether we need to
+ * move out of idle state, and return non-zero if so.
+ * "Straightforward" approach for the moment, might later
+ * use callback-list lengths, grace-period duration, or
+ * some such to determine when to exit idle state.
+ * Might also need a pre-idle test that does not acquire
+ * the lock, but let's get the simple case working first...
+ */
+
+static int
+rcu_try_flip_idle(void)
+{
+	int cpu;
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_i1);
+	if (!rcu_pending(smp_processor_id())) {
+		RCU_TRACE_ME(rcupreempt_trace_try_flip_ie1);
+		return 0;
+	}
+
+	/*
+	 * Do the flip.
+	 */
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_g1);
+	rcu_ctrlblk.completed++;  /* stands in for rcu_try_flip_g2 */
+
+	/*
+	 * Need a memory barrier so that other CPUs see the new
+	 * counter value before they see the subsequent change of all
+	 * the rcu_flip_flag instances to rcu_flipped.
+	 */
+
+	smp_mb();	/* see above block comment. */
+
+	/* Now ask each CPU for acknowledgement of the flip. */
+
+	for_each_cpu_mask(cpu, rcu_cpu_online_map)
+		per_cpu(rcu_flip_flag, cpu) = rcu_flipped;
+
+	return 1;
+}
+
+/*
+ * Wait for CPUs to acknowledge the flip.
+ */
+
+static int
+rcu_try_flip_waitack(void)
+{
+	int cpu;
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_a1);
+	for_each_cpu_mask(cpu, rcu_cpu_online_map)
+		if (per_cpu(rcu_flip_flag, cpu) != rcu_flip_seen) {
+			RCU_TRACE_ME(rcupreempt_trace_try_flip_ae1);
+			return 0;
+		}
+
+	/*
+	 * Make sure our checks above don't bleed into subsequent
+	 * waiting for the sum of the counters to reach zero.
+	 */
+
+	smp_mb();	/* see above block comment. */
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_a2);
+	return 1;
+}
+
+/*
+ * Wait for collective ``last'' counter to reach zero,
+ * then tell all CPUs to do an end-of-grace-period memory barrier.
+ */
+
+static int
+rcu_try_flip_waitzero(void)
+{
+	int cpu;
+	int lastidx = !(rcu_ctrlblk.completed & 0x1);
+	int sum = 0;
+
+	/* Check to see if the sum of the "last" counters is zero. */
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_z1);
+	for_each_cpu_mask(cpu, rcu_cpu_online_map)
+		sum += RCU_DATA_CPU(cpu)->rcu_flipctr[lastidx];
+	if (sum != 0) {
+		RCU_TRACE_ME(rcupreempt_trace_try_flip_ze1);
+		return 0;
+	}
+
+	/*
+	 * This ensures that the other CPUs see the call for
+	 * memory barriers -after- the sum to zero has been
+	 * detected here
+	 */
+	smp_mb();  /*  ^^^^^^^^^^^^ */
+
+	/* Call for a memory barrier from each CPU. */
+	for_each_cpu_mask(cpu, rcu_cpu_online_map)
+		per_cpu(rcu_mb_flag, cpu) = rcu_mb_needed;
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_z2);
+	return 1;
+}
+
+/*
+ * Wait for all CPUs to do their end-of-grace-period memory barrier.
+ * Return 0 once all CPUs have done so.
+ */
+
+static int
+rcu_try_flip_waitmb(void)
+{
+	int cpu;
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_m1);
+	for_each_cpu_mask(cpu, rcu_cpu_online_map)
+		if (per_cpu(rcu_mb_flag, cpu) != rcu_mb_done) {
+			RCU_TRACE_ME(rcupreempt_trace_try_flip_me1);
+			return 0;
+		}
+
+	smp_mb(); /* Ensure that the above checks precede any following flip. */
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_m2);
+	return 1;
+}
+
+/*
+ * Attempt a single flip of the counters.  Remember, a single flip does
+ * -not- constitute a grace period.  Instead, the interval between
+ * at least GP_STAGES consecutive flips is a grace period.
+ *
+ * If anyone is nuts enough to run this CONFIG_PREEMPT_RCU implementation
+ * on a large SMP, they might want to use a hierarchical organization of
+ * the per-CPU-counter pairs.
+ */
+static void rcu_try_flip(void)
+{
+	unsigned long flags;
+
+	RCU_TRACE_ME(rcupreempt_trace_try_flip_1);
+	if (unlikely(!spin_trylock_irqsave(&rcu_ctrlblk.fliplock, flags))) {
+		RCU_TRACE_ME(rcupreempt_trace_try_flip_e1);
+		return;
+	}
+
+	/*
+	 * Take the next transition(s) through the RCU grace-period
+	 * flip-counter state machine.
+	 */
+
+	switch (rcu_ctrlblk.rcu_try_flip_state) {
+	case rcu_try_flip_idle_state:
+		if (rcu_try_flip_idle())
+			rcu_ctrlblk.rcu_try_flip_state =
+				rcu_try_flip_waitack_state;
+		break;
+	case rcu_try_flip_waitack_state:
+		if (rcu_try_flip_waitack())
+			rcu_ctrlblk.rcu_try_flip_state =
+				rcu_try_flip_waitzero_state;
+		break;
+	case rcu_try_flip_waitzero_state:
+		if (rcu_try_flip_waitzero())
+			rcu_ctrlblk.rcu_try_flip_state =
+				rcu_try_flip_waitmb_state;
+		break;
+	case rcu_try_flip_waitmb_state:
+		if (rcu_try_flip_waitmb())
+			rcu_ctrlblk.rcu_try_flip_state =
+				rcu_try_flip_idle_state;
+	}
+	spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags);
+}
+
+/*
+ * Check to see if this CPU needs to do a memory barrier in order to
+ * ensure that any prior RCU read-side critical sections have committed
+ * their counter manipulations and critical-section memory references
+ * before declaring the grace period to be completed.
+ */
+static void rcu_check_mb(int cpu)
+{
+	if (per_cpu(rcu_mb_flag, cpu) == rcu_mb_needed) {
+		smp_mb();  /* Ensure RCU read-side accesses are visible. */
+		per_cpu(rcu_mb_flag, cpu) = rcu_mb_done;
+	}
+}
+
+void rcu_check_callbacks(int cpu, int user)
+{
+	unsigned long flags;
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+
+	rcu_check_mb(cpu);
+	if (rcu_ctrlblk.completed == rdp->completed)
+		rcu_try_flip();
+	spin_lock_irqsave(&rdp->lock, flags);
+	RCU_TRACE_RDP(rcupreempt_trace_check_callbacks, rdp);
+	__rcu_advance_callbacks(rdp);
+	if (rdp->donelist == NULL) {
+		spin_unlock_irqrestore(&rdp->lock, flags);
+	} else {
+		spin_unlock_irqrestore(&rdp->lock, flags);
+		raise_softirq(RCU_SOFTIRQ);
+	}
+}
+
+/*
+ * Needed by dynticks, to make sure all RCU processing has finished
+ * when we go idle:
+ */
+void rcu_advance_callbacks(int cpu, int user)
+{
+	unsigned long flags;
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+
+	if (rcu_ctrlblk.completed == rdp->completed) {
+		rcu_try_flip();
+		if (rcu_ctrlblk.completed == rdp->completed)
+			return;
+	}
+	spin_lock_irqsave(&rdp->lock, flags);
+	RCU_TRACE_RDP(rcupreempt_trace_check_callbacks, rdp);
+	__rcu_advance_callbacks(rdp);
+	spin_unlock_irqrestore(&rdp->lock, flags);
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+#define rcu_offline_cpu_enqueue(srclist, srctail, dstlist, dsttail) do { \
+		*dsttail = srclist; \
+		if (srclist != NULL) { \
+			dsttail = srctail; \
+			srclist = NULL; \
+			srctail = &srclist;\
+		} \
+	} while (0)
+
+void rcu_offline_cpu(int cpu)
+{
+	int i;
+	struct rcu_head *list = NULL;
+	unsigned long flags;
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+	struct rcu_head **tail = &list;
+
+	/*
+	 * Remove all callbacks from the newly dead CPU, retaining order.
+	 * Otherwise rcu_barrier() will fail
+	 */
+
+	spin_lock_irqsave(&rdp->lock, flags);
+	rcu_offline_cpu_enqueue(rdp->donelist, rdp->donetail, list, tail);
+	for (i = GP_STAGES - 1; i >= 0; i--)
+		rcu_offline_cpu_enqueue(rdp->waitlist[i], rdp->waittail[i],
+						list, tail);
+	rcu_offline_cpu_enqueue(rdp->nextlist, rdp->nexttail, list, tail);
+	spin_unlock_irqrestore(&rdp->lock, flags);
+	rdp->waitlistcount = 0;
+
+	/* Disengage the newly dead CPU from the grace-period computation. */
+
+	spin_lock_irqsave(&rcu_ctrlblk.fliplock, flags);
+	rcu_check_mb(cpu);
+	if (per_cpu(rcu_flip_flag, cpu) == rcu_flipped) {
+		smp_mb();  /* Subsequent counter accesses must see new value */
+		per_cpu(rcu_flip_flag, cpu) = rcu_flip_seen;
+		smp_mb();  /* Subsequent RCU read-side critical sections */
+			   /*  seen -after- acknowledgement. */
+	}
+
+	RCU_DATA_ME()->rcu_flipctr[0] += RCU_DATA_CPU(cpu)->rcu_flipctr[0];
+	RCU_DATA_ME()->rcu_flipctr[1] += RCU_DATA_CPU(cpu)->rcu_flipctr[1];
+
+	RCU_DATA_CPU(cpu)->rcu_flipctr[0] = 0;
+	RCU_DATA_CPU(cpu)->rcu_flipctr[1] = 0;
+
+	cpu_clear(cpu, rcu_cpu_online_map);
+
+	spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags);
+
+	/*
+	 * Place the removed callbacks on the current CPU's queue.
+	 * Make them all start a new grace period: simple approach,
+	 * in theory could starve a given set of callbacks, but
+	 * you would need to be doing some serious CPU hotplugging
+	 * to make this happen.  If this becomes a problem, adding
+	 * a synchronize_rcu() to the hotplug path would be a simple
+	 * fix.
+	 */
+
+	rdp = RCU_DATA_ME();
+	spin_lock_irqsave(&rdp->lock, flags);
+	*rdp->nexttail = list;
+	if (list)
+		rdp->nexttail = tail;
+	spin_unlock_irqrestore(&rdp->lock, flags);
+}
+
+void __devinit rcu_online_cpu(int cpu)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&rcu_ctrlblk.fliplock, flags);
+	cpu_set(cpu, rcu_cpu_online_map);
+	spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags);
+}
+
+#else /* #ifdef CONFIG_HOTPLUG_CPU */
+
+void rcu_offline_cpu(int cpu)
+{
+}
+
+void __devinit rcu_online_cpu(int cpu)
+{
+}
+
+#endif /* #else #ifdef CONFIG_HOTPLUG_CPU */
+
+static void rcu_process_callbacks(struct softirq_action *unused)
+{
+	unsigned long flags;
+	struct rcu_head *next, *list;
+	struct rcu_data *rdp = RCU_DATA_ME();
+
+	spin_lock_irqsave(&rdp->lock, flags);
+	list = rdp->donelist;
+	if (list == NULL) {
+		spin_unlock_irqrestore(&rdp->lock, flags);
+		return;
+	}
+	rdp->donelist = NULL;
+	rdp->donetail = &rdp->donelist;
+	RCU_TRACE_RDP(rcupreempt_trace_done_remove, rdp);
+	spin_unlock_irqrestore(&rdp->lock, flags);
+	while (list) {
+		next = list->next;
+		list->func(list);
+		list = next;
+		RCU_TRACE_ME(rcupreempt_trace_invoke);
+	}
+}
+
+void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
+{
+	unsigned long flags;
+	struct rcu_data *rdp;
+
+	head->func = func;
+	head->next = NULL;
+	local_irq_save(flags);
+	rdp = RCU_DATA_ME();
+	spin_lock(&rdp->lock);
+	__rcu_advance_callbacks(rdp);
+	*rdp->nexttail = head;
+	rdp->nexttail = &head->next;
+	RCU_TRACE_RDP(rcupreempt_trace_next_add, rdp);
+	spin_unlock(&rdp->lock);
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(call_rcu);
+
+/*
+ * Wait until all currently running preempt_disable() code segments
+ * (including hardware-irq-disable segments) complete.  Note that
+ * in -rt this does -not- necessarily result in all currently executing
+ * interrupt -handlers- having completed.
+ */
+void __synchronize_sched(void)
+{
+	cpumask_t oldmask;
+	int cpu;
+
+	if (sched_getaffinity(0, &oldmask) < 0)
+		oldmask = cpu_possible_map;
+	for_each_online_cpu(cpu) {
+		sched_setaffinity(0, cpumask_of_cpu(cpu));
+		schedule();
+	}
+	sched_setaffinity(0, oldmask);
+}
+EXPORT_SYMBOL_GPL(__synchronize_sched);
+
+/*
+ * Check to see if any future RCU-related work will need to be done
+ * by the current CPU, even if none need be done immediately, returning
+ * 1 if so.  Assumes that notifiers would take care of handling any
+ * outstanding requests from the RCU core.
+ *
+ * This function is part of the RCU implementation; it is -not-
+ * an exported member of the RCU API.
+ */
+int rcu_needs_cpu(int cpu)
+{
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+
+	return (rdp->donelist != NULL ||
+		!!rdp->waitlistcount ||
+		rdp->nextlist != NULL);
+}
+
+int rcu_pending(int cpu)
+{
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+
+	/* The CPU has at least one callback queued somewhere. */
+
+	if (rdp->donelist != NULL ||
+	    !!rdp->waitlistcount ||
+	    rdp->nextlist != NULL)
+		return 1;
+
+	/* The RCU core needs an acknowledgement from this CPU. */
+
+	if ((per_cpu(rcu_flip_flag, cpu) == rcu_flipped) ||
+	    (per_cpu(rcu_mb_flag, cpu) == rcu_mb_needed))
+		return 1;
+
+	/* This CPU has fallen behind the global grace-period number. */
+
+	if (rdp->completed != rcu_ctrlblk.completed)
+		return 1;
+
+	/* Nothing needed from this CPU. */
+
+	return 0;
+}
+
+static int __cpuinit rcu_cpu_notify(struct notifier_block *self,
+				unsigned long action, void *hcpu)
+{
+	long cpu = (long)hcpu;
+
+	switch (action) {
+	case CPU_UP_PREPARE:
+	case CPU_UP_PREPARE_FROZEN:
+		rcu_online_cpu(cpu);
+		break;
+	case CPU_UP_CANCELED:
+	case CPU_UP_CANCELED_FROZEN:
+	case CPU_DEAD:
+	case CPU_DEAD_FROZEN:
+		rcu_offline_cpu(cpu);
+		break;
+	default:
+		break;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block __cpuinitdata rcu_nb = {
+	.notifier_call = rcu_cpu_notify,
+};
+
+void __init __rcu_init(void)
+{
+	int cpu;
+	int i;
+	struct rcu_data *rdp;
+
+	printk(KERN_NOTICE "Preemptible RCU implementation.\n");
+	for_each_possible_cpu(cpu) {
+		rdp = RCU_DATA_CPU(cpu);
+		spin_lock_init(&rdp->lock);
+		rdp->completed = 0;
+		rdp->waitlistcount = 0;
+		rdp->nextlist = NULL;
+		rdp->nexttail = &rdp->nextlist;
+		for (i = 0; i < GP_STAGES; i++) {
+			rdp->waitlist[i] = NULL;
+			rdp->waittail[i] = &rdp->waitlist[i];
+		}
+		rdp->donelist = NULL;
+		rdp->donetail = &rdp->donelist;
+		rdp->rcu_flipctr[0] = 0;
+		rdp->rcu_flipctr[1] = 0;
+	}
+	register_cpu_notifier(&rcu_nb);
+
+	/*
+	 * We don't need protection against CPU-Hotplug here
+	 * since
+	 * a) If a CPU comes online while we are iterating over the
+	 *    cpu_online_map below, we would only end up making a
+	 *    duplicate call to rcu_online_cpu() which sets the corresponding
+	 *    CPU's mask in the rcu_cpu_online_map.
+	 *
+	 * b) A CPU cannot go offline at this point in time since the user
+	 *    does not have access to the sysfs interface, nor do we
+	 *    suspend the system.
+	 */
+	for_each_online_cpu(cpu)
+		rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE,	(void *)(long) cpu);
+
+	open_softirq(RCU_SOFTIRQ, rcu_process_callbacks, NULL);
+}
+
+/*
+ * Deprecated, use synchronize_rcu() or synchronize_sched() instead.
+ */
+void synchronize_kernel(void)
+{
+	synchronize_rcu();
+}
+
+#ifdef CONFIG_RCU_TRACE
+long *rcupreempt_flipctr(int cpu)
+{
+	return &RCU_DATA_CPU(cpu)->rcu_flipctr[0];
+}
+EXPORT_SYMBOL_GPL(rcupreempt_flipctr);
+
+int rcupreempt_flip_flag(int cpu)
+{
+	return per_cpu(rcu_flip_flag, cpu);
+}
+EXPORT_SYMBOL_GPL(rcupreempt_flip_flag);
+
+int rcupreempt_mb_flag(int cpu)
+{
+	return per_cpu(rcu_mb_flag, cpu);
+}
+EXPORT_SYMBOL_GPL(rcupreempt_mb_flag);
+
+char *rcupreempt_try_flip_state_name(void)
+{
+	return rcu_try_flip_state_names[rcu_ctrlblk.rcu_try_flip_state];
+}
+EXPORT_SYMBOL_GPL(rcupreempt_try_flip_state_name);
+
+struct rcupreempt_trace *rcupreempt_trace_cpu(int cpu)
+{
+	struct rcu_data *rdp = RCU_DATA_CPU(cpu);
+
+	return &rdp->trace;
+}
+EXPORT_SYMBOL_GPL(rcupreempt_trace_cpu);
+
+#endif /* #ifdef RCU_TRACE */

kernel/rcupreempt_trace.c

@@ -0,0 +1,330 @@
+/*
+ * Read-Copy Update tracing for realtime implementation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright IBM Corporation, 2006
+ *
+ * Papers:  http://www.rdrop.com/users/paulmck/RCU
+ *
+ * For detailed explanation of Read-Copy Update mechanism see -
+ * 		Documentation/RCU/ *.txt
+ *
+ */
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/spinlock.h>
+#include <linux/smp.h>
+#include <linux/rcupdate.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <asm/atomic.h>
+#include <linux/bitops.h>
+#include <linux/module.h>
+#include <linux/completion.h>
+#include <linux/moduleparam.h>
+#include <linux/percpu.h>
+#include <linux/notifier.h>
+#include <linux/rcupdate.h>
+#include <linux/cpu.h>
+#include <linux/mutex.h>
+#include <linux/rcupreempt_trace.h>
+#include <linux/debugfs.h>
+
+static struct mutex rcupreempt_trace_mutex;
+static char *rcupreempt_trace_buf;
+#define RCUPREEMPT_TRACE_BUF_SIZE 4096
+
+void rcupreempt_trace_move2done(struct rcupreempt_trace *trace)
+{
+	trace->done_length += trace->wait_length;
+	trace->done_add += trace->wait_length;
+	trace->wait_length = 0;
+}
+void rcupreempt_trace_move2wait(struct rcupreempt_trace *trace)
+{
+	trace->wait_length += trace->next_length;
+	trace->wait_add += trace->next_length;
+	trace->next_length = 0;
+}
+void rcupreempt_trace_try_flip_1(struct rcupreempt_trace *trace)
+{
+	atomic_inc(&trace->rcu_try_flip_1);
+}
+void rcupreempt_trace_try_flip_e1(struct rcupreempt_trace *trace)
+{
+	atomic_inc(&trace->rcu_try_flip_e1);
+}
+void rcupreempt_trace_try_flip_i1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_i1++;
+}
+void rcupreempt_trace_try_flip_ie1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_ie1++;
+}
+void rcupreempt_trace_try_flip_g1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_g1++;
+}
+void rcupreempt_trace_try_flip_a1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_a1++;
+}
+void rcupreempt_trace_try_flip_ae1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_ae1++;
+}
+void rcupreempt_trace_try_flip_a2(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_a2++;
+}
+void rcupreempt_trace_try_flip_z1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_z1++;
+}
+void rcupreempt_trace_try_flip_ze1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_ze1++;
+}
+void rcupreempt_trace_try_flip_z2(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_z2++;
+}
+void rcupreempt_trace_try_flip_m1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_m1++;
+}
+void rcupreempt_trace_try_flip_me1(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_me1++;
+}
+void rcupreempt_trace_try_flip_m2(struct rcupreempt_trace *trace)
+{
+	trace->rcu_try_flip_m2++;
+}
+void rcupreempt_trace_check_callbacks(struct rcupreempt_trace *trace)
+{
+	trace->rcu_check_callbacks++;
+}
+void rcupreempt_trace_done_remove(struct rcupreempt_trace *trace)
+{
+	trace->done_remove += trace->done_length;
+	trace->done_length = 0;
+}
+void rcupreempt_trace_invoke(struct rcupreempt_trace *trace)
+{
+	atomic_inc(&trace->done_invoked);
+}
+void rcupreempt_trace_next_add(struct rcupreempt_trace *trace)
+{
+	trace->next_add++;
+	trace->next_length++;
+}
+
+static void rcupreempt_trace_sum(struct rcupreempt_trace *sp)
+{
+	struct rcupreempt_trace *cp;
+	int cpu;
+
+	memset(sp, 0, sizeof(*sp));
+	for_each_possible_cpu(cpu) {
+		cp = rcupreempt_trace_cpu(cpu);
+		sp->next_length += cp->next_length;
+		sp->next_add += cp->next_add;
+		sp->wait_length += cp->wait_length;
+		sp->wait_add += cp->wait_add;
+		sp->done_length += cp->done_length;
+		sp->done_add += cp->done_add;
+		sp->done_remove += cp->done_remove;
+		atomic_set(&sp->done_invoked, atomic_read(&cp->done_invoked));
+		sp->rcu_check_callbacks += cp->rcu_check_callbacks;
+		atomic_set(&sp->rcu_try_flip_1,
+			   atomic_read(&cp->rcu_try_flip_1));
+		atomic_set(&sp->rcu_try_flip_e1,
+			   atomic_read(&cp->rcu_try_flip_e1));
+		sp->rcu_try_flip_i1 += cp->rcu_try_flip_i1;
+		sp->rcu_try_flip_ie1 += cp->rcu_try_flip_ie1;
+		sp->rcu_try_flip_g1 += cp->rcu_try_flip_g1;
+		sp->rcu_try_flip_a1 += cp->rcu_try_flip_a1;
+		sp->rcu_try_flip_ae1 += cp->rcu_try_flip_ae1;
+		sp->rcu_try_flip_a2 += cp->rcu_try_flip_a2;
+		sp->rcu_try_flip_z1 += cp->rcu_try_flip_z1;
+		sp->rcu_try_flip_ze1 += cp->rcu_try_flip_ze1;
+		sp->rcu_try_flip_z2 += cp->rcu_try_flip_z2;
+		sp->rcu_try_flip_m1 += cp->rcu_try_flip_m1;
+		sp->rcu_try_flip_me1 += cp->rcu_try_flip_me1;
+		sp->rcu_try_flip_m2 += cp->rcu_try_flip_m2;
+	}
+}
+
+static ssize_t rcustats_read(struct file *filp, char __user *buffer,
+				size_t count, loff_t *ppos)
+{
+	struct rcupreempt_trace trace;
+	ssize_t bcount;
+	int cnt = 0;
+
+	rcupreempt_trace_sum(&trace);
+	mutex_lock(&rcupreempt_trace_mutex);
+	snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE - cnt,
+		 "ggp=%ld rcc=%ld\n",
+		 rcu_batches_completed(),
+		 trace.rcu_check_callbacks);
+	snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE - cnt,
+		 "na=%ld nl=%ld wa=%ld wl=%ld da=%ld dl=%ld dr=%ld di=%d\n"
+		 "1=%d e1=%d i1=%ld ie1=%ld g1=%ld a1=%ld ae1=%ld a2=%ld\n"
+		 "z1=%ld ze1=%ld z2=%ld m1=%ld me1=%ld m2=%ld\n",
+
+		 trace.next_add, trace.next_length,
+		 trace.wait_add, trace.wait_length,
+		 trace.done_add, trace.done_length,
+		 trace.done_remove, atomic_read(&trace.done_invoked),
+		 atomic_read(&trace.rcu_try_flip_1),
+		 atomic_read(&trace.rcu_try_flip_e1),
+		 trace.rcu_try_flip_i1, trace.rcu_try_flip_ie1,
+		 trace.rcu_try_flip_g1,
+		 trace.rcu_try_flip_a1, trace.rcu_try_flip_ae1,
+			 trace.rcu_try_flip_a2,
+		 trace.rcu_try_flip_z1, trace.rcu_try_flip_ze1,
+			 trace.rcu_try_flip_z2,
+		 trace.rcu_try_flip_m1, trace.rcu_try_flip_me1,
+			trace.rcu_try_flip_m2);
+	bcount = simple_read_from_buffer(buffer, count, ppos,
+			rcupreempt_trace_buf, strlen(rcupreempt_trace_buf));
+	mutex_unlock(&rcupreempt_trace_mutex);
+	return bcount;
+}
+
+static ssize_t rcugp_read(struct file *filp, char __user *buffer,
+				size_t count, loff_t *ppos)
+{
+	long oldgp = rcu_batches_completed();
+	ssize_t bcount;
+
+	mutex_lock(&rcupreempt_trace_mutex);
+	synchronize_rcu();
+	snprintf(rcupreempt_trace_buf, RCUPREEMPT_TRACE_BUF_SIZE,
+		"oldggp=%ld  newggp=%ld\n", oldgp, rcu_batches_completed());
+	bcount = simple_read_from_buffer(buffer, count, ppos,
+			rcupreempt_trace_buf, strlen(rcupreempt_trace_buf));
+	mutex_unlock(&rcupreempt_trace_mutex);
+	return bcount;
+}
+
+static ssize_t rcuctrs_read(struct file *filp, char __user *buffer,
+				size_t count, loff_t *ppos)
+{
+	int cnt = 0;
+	int cpu;
+	int f = rcu_batches_completed() & 0x1;
+	ssize_t bcount;
+
+	mutex_lock(&rcupreempt_trace_mutex);
+
+	cnt += snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE,
+				"CPU last cur F M\n");
+	for_each_online_cpu(cpu) {
+		long *flipctr = rcupreempt_flipctr(cpu);
+		cnt += snprintf(&rcupreempt_trace_buf[cnt],
+				RCUPREEMPT_TRACE_BUF_SIZE - cnt,
+					"%3d %4ld %3ld %d %d\n",
+			       cpu,
+			       flipctr[!f],
+			       flipctr[f],
+			       rcupreempt_flip_flag(cpu),
+			       rcupreempt_mb_flag(cpu));
+	}
+	cnt += snprintf(&rcupreempt_trace_buf[cnt],
+			RCUPREEMPT_TRACE_BUF_SIZE - cnt,
+			"ggp = %ld, state = %s\n",
+			rcu_batches_completed(),
+			rcupreempt_try_flip_state_name());
+	cnt += snprintf(&rcupreempt_trace_buf[cnt],
+			RCUPREEMPT_TRACE_BUF_SIZE - cnt,
+			"\n");
+	bcount = simple_read_from_buffer(buffer, count, ppos,
+			rcupreempt_trace_buf, strlen(rcupreempt_trace_buf));
+	mutex_unlock(&rcupreempt_trace_mutex);
+	return bcount;
+}
+
+static struct file_operations rcustats_fops = {
+	.owner = THIS_MODULE,
+	.read = rcustats_read,
+};
+
+static struct file_operations rcugp_fops = {
+	.owner = THIS_MODULE,
+	.read = rcugp_read,
+};
+
+static struct file_operations rcuctrs_fops = {
+	.owner = THIS_MODULE,
+	.read = rcuctrs_read,
+};
+
+static struct dentry *rcudir, *statdir, *ctrsdir, *gpdir;
+static int rcupreempt_debugfs_init(void)
+{
+	rcudir = debugfs_create_dir("rcu", NULL);
+	if (!rcudir)
+		goto out;
+	statdir = debugfs_create_file("rcustats", 0444, rcudir,
+						NULL, &rcustats_fops);
+	if (!statdir)
+		goto free_out;
+
+	gpdir = debugfs_create_file("rcugp", 0444, rcudir, NULL, &rcugp_fops);
+	if (!gpdir)
+		goto free_out;
+
+	ctrsdir = debugfs_create_file("rcuctrs", 0444, rcudir,
+						NULL, &rcuctrs_fops);
+	if (!ctrsdir)
+		goto free_out;
+	return 0;
+free_out:
+	if (statdir)
+		debugfs_remove(statdir);
+	if (gpdir)
+		debugfs_remove(gpdir);
+	debugfs_remove(rcudir);
+out:
+	return 1;
+}
+
+static int __init rcupreempt_trace_init(void)
+{
+	mutex_init(&rcupreempt_trace_mutex);
+	rcupreempt_trace_buf = kmalloc(RCUPREEMPT_TRACE_BUF_SIZE, GFP_KERNEL);
+	if (!rcupreempt_trace_buf)
+		return 1;
+	return rcupreempt_debugfs_init();
+}
+
+static void __exit rcupreempt_trace_cleanup(void)
+{
+	debugfs_remove(statdir);
+	debugfs_remove(gpdir);
+	debugfs_remove(ctrsdir);
+	debugfs_remove(rcudir);
+	kfree(rcupreempt_trace_buf);
+}
+
+
+module_init(rcupreempt_trace_init);
+module_exit(rcupreempt_trace_cleanup);

kernel/rcutorture.c

@@ -726,11 +726,11 @@ static void rcu_torture_shuffle_tasks(void)
 	cpumask_t tmp_mask = CPU_MASK_ALL;
 	int i;
 
-	lock_cpu_hotplug();
+	get_online_cpus();
 
 	/* No point in shuffling if there is only one online CPU (ex: UP) */
 	if (num_online_cpus() == 1) {
-		unlock_cpu_hotplug();
+		put_online_cpus();
 		return;
 	}
 
@@ -762,7 +762,7 @@ static void rcu_torture_shuffle_tasks(void)
 	else
 		rcu_idle_cpu--;
 
-	unlock_cpu_hotplug();
+	put_online_cpus();
 }
 
 /* Shuffle tasks across CPUs, with the intent of allowing each CPU in the

kernel/sched.c

@@ -22,6 +22,8 @@
  *              by Peter Williams
  *  2007-05-06  Interactivity improvements to CFS by Mike Galbraith
  *  2007-07-01  Group scheduling enhancements by Srivatsa Vaddagiri
+ *  2007-11-29  RT balancing improvements by Steven Rostedt, Gregory Haskins,
+ *              Thomas Gleixner, Mike Kravetz
  */
 
 #include <linux/mm.h>
@@ -63,6 +65,7 @@
 #include <linux/reciprocal_div.h>
 #include <linux/unistd.h>
 #include <linux/pagemap.h>
+#include <linux/hrtimer.h>
 
 #include <asm/tlb.h>
 #include <asm/irq_regs.h>
@@ -96,10 +99,9 @@ unsigned long long __attribute__((weak)) sched_clock(void)
 #define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))
 
 /*
- * Some helpers for converting nanosecond timing to jiffy resolution
+ * Helpers for converting nanosecond timing to jiffy resolution
  */
 #define NS_TO_JIFFIES(TIME)	((unsigned long)(TIME) / (NSEC_PER_SEC / HZ))
-#define JIFFIES_TO_NS(TIME)	((TIME) * (NSEC_PER_SEC / HZ))
 
 #define NICE_0_LOAD		SCHED_LOAD_SCALE
 #define NICE_0_SHIFT		SCHED_LOAD_SHIFT
@@ -159,6 +161,8 @@ struct rt_prio_array {
 
 struct cfs_rq;
 
+static LIST_HEAD(task_groups);
+
 /* task group related information */
 struct task_group {
 #ifdef CONFIG_FAIR_CGROUP_SCHED
@@ -168,10 +172,50 @@ struct task_group {
 	struct sched_entity **se;
 	/* runqueue "owned" by this group on each cpu */
 	struct cfs_rq **cfs_rq;
+
+	struct sched_rt_entity **rt_se;
+	struct rt_rq **rt_rq;
+
+	unsigned int rt_ratio;
+
+	/*
+	 * shares assigned to a task group governs how much of cpu bandwidth
+	 * is allocated to the group. The more shares a group has, the more is
+	 * the cpu bandwidth allocated to it.
+	 *
+	 * For ex, lets say that there are three task groups, A, B and C which
+	 * have been assigned shares 1000, 2000 and 3000 respectively. Then,
+	 * cpu bandwidth allocated by the scheduler to task groups A, B and C
+	 * should be:
+	 *
+	 *	Bw(A) = 1000/(1000+2000+3000) * 100 = 16.66%
+	 *	Bw(B) = 2000/(1000+2000+3000) * 100 = 33.33%
+	 *	Bw(C) = 3000/(1000+2000+3000) * 100 = 50%
+	 *
+	 * The weight assigned to a task group's schedulable entities on every
+	 * cpu (task_group.se[a_cpu]->load.weight) is derived from the task
+	 * group's shares. For ex: lets say that task group A has been
+	 * assigned shares of 1000 and there are two CPUs in a system. Then,
+	 *
+	 *  tg_A->se[0]->load.weight = tg_A->se[1]->load.weight = 1000;
+	 *
+	 * Note: It's not necessary that each of a task's group schedulable
+	 *	 entity have the same weight on all CPUs. If the group
+	 *	 has 2 of its tasks on CPU0 and 1 task on CPU1, then a
+	 *	 better distribution of weight could be:
+	 *
+	 *	tg_A->se[0]->load.weight = 2/3 * 2000 = 1333
+	 *	tg_A->se[1]->load.weight = 1/2 * 2000 =  667
+	 *
+	 * rebalance_shares() is responsible for distributing the shares of a
+	 * task groups like this among the group's schedulable entities across
+	 * cpus.
+	 *
+	 */
 	unsigned long shares;
-	/* spinlock to serialize modification to shares */
-	spinlock_t lock;
+
 	struct rcu_head rcu;
+	struct list_head list;
 };
 
 /* Default task group's sched entity on each cpu */
@@ -179,24 +223,51 @@ static DEFINE_PER_CPU(struct sched_entity, init_sched_entity);
 /* Default task group's cfs_rq on each cpu */
 static DEFINE_PER_CPU(struct cfs_rq, init_cfs_rq) ____cacheline_aligned_in_smp;
 
+static DEFINE_PER_CPU(struct sched_rt_entity, init_sched_rt_entity);
+static DEFINE_PER_CPU(struct rt_rq, init_rt_rq) ____cacheline_aligned_in_smp;
+
 static struct sched_entity *init_sched_entity_p[NR_CPUS];
 static struct cfs_rq *init_cfs_rq_p[NR_CPUS];
 
+static struct sched_rt_entity *init_sched_rt_entity_p[NR_CPUS];
+static struct rt_rq *init_rt_rq_p[NR_CPUS];
+
+/* task_group_mutex serializes add/remove of task groups and also changes to
+ * a task group's cpu shares.
+ */
+static DEFINE_MUTEX(task_group_mutex);
+
+/* doms_cur_mutex serializes access to doms_cur[] array */
+static DEFINE_MUTEX(doms_cur_mutex);
+
+#ifdef CONFIG_SMP
+/* kernel thread that runs rebalance_shares() periodically */
+static struct task_struct *lb_monitor_task;
+static int load_balance_monitor(void *unused);
+#endif
+
+static void set_se_shares(struct sched_entity *se, unsigned long shares);
+
 /* Default task group.
  *	Every task in system belong to this group at bootup.
  */
 struct task_group init_task_group = {
-	.se     = init_sched_entity_p,
+	.se	= init_sched_entity_p,
 	.cfs_rq = init_cfs_rq_p,
+
+	.rt_se	= init_sched_rt_entity_p,
+	.rt_rq	= init_rt_rq_p,
 };
 
 #ifdef CONFIG_FAIR_USER_SCHED
-# define INIT_TASK_GRP_LOAD	2*NICE_0_LOAD
+# define INIT_TASK_GROUP_LOAD	(2*NICE_0_LOAD)
 #else
-# define INIT_TASK_GRP_LOAD	NICE_0_LOAD
+# define INIT_TASK_GROUP_LOAD	NICE_0_LOAD
 #endif
 
-static int init_task_group_load = INIT_TASK_GRP_LOAD;
+#define MIN_GROUP_SHARES	2
+
+static int init_task_group_load = INIT_TASK_GROUP_LOAD;
 
 /* return group to which a task belongs */
 static inline struct task_group *task_group(struct task_struct *p)
@@ -215,15 +286,42 @@ static inline struct task_group *task_group(struct task_struct *p)
 }
 
 /* Change a task's cfs_rq and parent entity if it moves across CPUs/groups */
-static inline void set_task_cfs_rq(struct task_struct *p, unsigned int cpu)
+static inline void set_task_rq(struct task_struct *p, unsigned int cpu)
 {
 	p->se.cfs_rq = task_group(p)->cfs_rq[cpu];
 	p->se.parent = task_group(p)->se[cpu];
+
+	p->rt.rt_rq  = task_group(p)->rt_rq[cpu];
+	p->rt.parent = task_group(p)->rt_se[cpu];
+}
+
+static inline void lock_task_group_list(void)
+{
+	mutex_lock(&task_group_mutex);
+}
+
+static inline void unlock_task_group_list(void)
+{
+	mutex_unlock(&task_group_mutex);
+}
+
+static inline void lock_doms_cur(void)
+{
+	mutex_lock(&doms_cur_mutex);
+}
+
+static inline void unlock_doms_cur(void)
+{
+	mutex_unlock(&doms_cur_mutex);
 }
 
 #else
 
-static inline void set_task_cfs_rq(struct task_struct *p, unsigned int cpu) { }
+static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { }
+static inline void lock_task_group_list(void) { }
+static inline void unlock_task_group_list(void) { }
+static inline void lock_doms_cur(void) { }
+static inline void unlock_doms_cur(void) { }
 
 #endif	/* CONFIG_FAIR_GROUP_SCHED */
 
@@ -264,10 +362,56 @@ struct cfs_rq {
 /* Real-Time classes' related field in a runqueue: */
 struct rt_rq {
 	struct rt_prio_array active;
-	int rt_load_balance_idx;
-	struct list_head *rt_load_balance_head, *rt_load_balance_curr;
+	unsigned long rt_nr_running;
+#if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
+	int highest_prio; /* highest queued rt task prio */
+#endif
+#ifdef CONFIG_SMP
+	unsigned long rt_nr_migratory;
+	int overloaded;
+#endif
+	int rt_throttled;
+	u64 rt_time;
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	struct rq *rq;
+	struct list_head leaf_rt_rq_list;
+	struct task_group *tg;
+	struct sched_rt_entity *rt_se;
+#endif
 };
 
+#ifdef CONFIG_SMP
+
+/*
+ * We add the notion of a root-domain which will be used to define per-domain
+ * variables. Each exclusive cpuset essentially defines an island domain by
+ * fully partitioning the member cpus from any other cpuset. Whenever a new
+ * exclusive cpuset is created, we also create and attach a new root-domain
+ * object.
+ *
+ */
+struct root_domain {
+	atomic_t refcount;
+	cpumask_t span;
+	cpumask_t online;
+
+	/*
+	 * The "RT overload" flag: it gets set if a CPU has more than
+	 * one runnable RT task.
+	 */
+	cpumask_t rto_mask;
+	atomic_t rto_count;
+};
+
+/*
+ * By default the system creates a single root-domain with all cpus as
+ * members (mimicking the global state we have today).
+ */
+static struct root_domain def_root_domain;
+
+#endif
+
 /*
  * This is the main, per-CPU runqueue data structure.
  *
@@ -296,11 +440,15 @@ struct rq {
 	u64 nr_switches;
 
 	struct cfs_rq cfs;
+	struct rt_rq rt;
+	u64 rt_period_expire;
+	int rt_throttled;
+
 #ifdef CONFIG_FAIR_GROUP_SCHED
 	/* list of leaf cfs_rq on this cpu: */
 	struct list_head leaf_cfs_rq_list;
+	struct list_head leaf_rt_rq_list;
 #endif
-	struct rt_rq rt;
 
 	/*
 	 * This is part of a global counter where only the total sum
@@ -317,7 +465,7 @@ struct rq {
 	u64 clock, prev_clock_raw;
 	s64 clock_max_delta;
 
-	unsigned int clock_warps, clock_overflows;
+	unsigned int clock_warps, clock_overflows, clock_underflows;
 	u64 idle_clock;
 	unsigned int clock_deep_idle_events;
 	u64 tick_timestamp;
@@ -325,6 +473,7 @@ struct rq {
 	atomic_t nr_iowait;
 
 #ifdef CONFIG_SMP
+	struct root_domain *rd;
 	struct sched_domain *sd;
 
 	/* For active balancing */
@@ -337,6 +486,12 @@ struct rq {
 	struct list_head migration_queue;
 #endif
 
+#ifdef CONFIG_SCHED_HRTICK
+	unsigned long hrtick_flags;
+	ktime_t hrtick_expire;
+	struct hrtimer hrtick_timer;
+#endif
+
 #ifdef CONFIG_SCHEDSTATS
 	/* latency stats */
 	struct sched_info rq_sched_info;
@@ -363,7 +518,6 @@ struct rq {
 };
 
 static DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
-static DEFINE_MUTEX(sched_hotcpu_mutex);
 
 static inline void check_preempt_curr(struct rq *rq, struct task_struct *p)
 {
@@ -441,6 +595,23 @@ static void update_rq_clock(struct rq *rq)
 #define task_rq(p)		cpu_rq(task_cpu(p))
 #define cpu_curr(cpu)		(cpu_rq(cpu)->curr)
 
+unsigned long rt_needs_cpu(int cpu)
+{
+	struct rq *rq = cpu_rq(cpu);
+	u64 delta;
+
+	if (!rq->rt_throttled)
+		return 0;
+
+	if (rq->clock > rq->rt_period_expire)
+		return 1;
+
+	delta = rq->rt_period_expire - rq->clock;
+	do_div(delta, NSEC_PER_SEC / HZ);
+
+	return (unsigned long)delta;
+}
+
 /*
  * Tunables that become constants when CONFIG_SCHED_DEBUG is off:
  */
@@ -459,6 +630,8 @@ enum {
 	SCHED_FEAT_START_DEBIT		= 4,
 	SCHED_FEAT_TREE_AVG		= 8,
 	SCHED_FEAT_APPROX_AVG		= 16,
+	SCHED_FEAT_HRTICK		= 32,
+	SCHED_FEAT_DOUBLE_TICK		= 64,
 };
 
 const_debug unsigned int sysctl_sched_features =
@@ -466,7 +639,9 @@ const_debug unsigned int sysctl_sched_features =
 		SCHED_FEAT_WAKEUP_PREEMPT	* 1 |
 		SCHED_FEAT_START_DEBIT		* 1 |
 		SCHED_FEAT_TREE_AVG		* 0 |
-		SCHED_FEAT_APPROX_AVG		* 0;
+		SCHED_FEAT_APPROX_AVG		* 0 |
+		SCHED_FEAT_HRTICK		* 1 |
+		SCHED_FEAT_DOUBLE_TICK		* 0;
 
 #define sched_feat(x) (sysctl_sched_features & SCHED_FEAT_##x)
 
@@ -476,6 +651,21 @@ const_debug unsigned int sysctl_sched_features =
  */
 const_debug unsigned int sysctl_sched_nr_migrate = 32;
 
+/*
+ * period over which we measure -rt task cpu usage in ms.
+ * default: 1s
+ */
+const_debug unsigned int sysctl_sched_rt_period = 1000;
+
+#define SCHED_RT_FRAC_SHIFT	16
+#define SCHED_RT_FRAC		(1UL << SCHED_RT_FRAC_SHIFT)
+
+/*
+ * ratio of time -rt tasks may consume.
+ * default: 95%
+ */
+const_debug unsigned int sysctl_sched_rt_ratio = 62259;
+
 /*
  * For kernel-internal use: high-speed (but slightly incorrect) per-cpu
  * clock constructed from sched_clock():
@@ -668,7 +858,6 @@ void sched_clock_idle_wakeup_event(u64 delta_ns)
 	struct rq *rq = cpu_rq(smp_processor_id());
 	u64 now = sched_clock();
 
-	touch_softlockup_watchdog();
 	rq->idle_clock += delta_ns;
 	/*
 	 * Override the previous timestamp and ignore all
@@ -680,9 +869,177 @@ void sched_clock_idle_wakeup_event(u64 delta_ns)
 	rq->prev_clock_raw = now;
 	rq->clock += delta_ns;
 	spin_unlock(&rq->lock);
+	touch_softlockup_watchdog();
 }
 EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event);
 
+static void __resched_task(struct task_struct *p, int tif_bit);
+
+static inline void resched_task(struct task_struct *p)
+{
+	__resched_task(p, TIF_NEED_RESCHED);
+}
+
+#ifdef CONFIG_SCHED_HRTICK
+/*
+ * Use HR-timers to deliver accurate preemption points.
+ *
+ * Its all a bit involved since we cannot program an hrt while holding the
+ * rq->lock. So what we do is store a state in in rq->hrtick_* and ask for a
+ * reschedule event.
+ *
+ * When we get rescheduled we reprogram the hrtick_timer outside of the
+ * rq->lock.
+ */
+static inline void resched_hrt(struct task_struct *p)
+{
+	__resched_task(p, TIF_HRTICK_RESCHED);
+}
+
+static inline void resched_rq(struct rq *rq)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&rq->lock, flags);
+	resched_task(rq->curr);
+	spin_unlock_irqrestore(&rq->lock, flags);
+}
+
+enum {
+	HRTICK_SET,		/* re-programm hrtick_timer */
+	HRTICK_RESET,		/* not a new slice */
+};
+
+/*
+ * Use hrtick when:
+ *  - enabled by features
+ *  - hrtimer is actually high res
+ */
+static inline int hrtick_enabled(struct rq *rq)
+{
+	if (!sched_feat(HRTICK))
+		return 0;
+	return hrtimer_is_hres_active(&rq->hrtick_timer);
+}
+
+/*
+ * Called to set the hrtick timer state.
+ *
+ * called with rq->lock held and irqs disabled
+ */
+static void hrtick_start(struct rq *rq, u64 delay, int reset)
+{
+	assert_spin_locked(&rq->lock);
+
+	/*
+	 * preempt at: now + delay
+	 */
+	rq->hrtick_expire =
+		ktime_add_ns(rq->hrtick_timer.base->get_time(), delay);
+	/*
+	 * indicate we need to program the timer
+	 */
+	__set_bit(HRTICK_SET, &rq->hrtick_flags);
+	if (reset)
+		__set_bit(HRTICK_RESET, &rq->hrtick_flags);
+
+	/*
+	 * New slices are called from the schedule path and don't need a
+	 * forced reschedule.
+	 */
+	if (reset)
+		resched_hrt(rq->curr);
+}
+
+static void hrtick_clear(struct rq *rq)
+{
+	if (hrtimer_active(&rq->hrtick_timer))
+		hrtimer_cancel(&rq->hrtick_timer);
+}
+
+/*
+ * Update the timer from the possible pending state.
+ */
+static void hrtick_set(struct rq *rq)
+{
+	ktime_t time;
+	int set, reset;
+	unsigned long flags;
+
+	WARN_ON_ONCE(cpu_of(rq) != smp_processor_id());
+
+	spin_lock_irqsave(&rq->lock, flags);
+	set = __test_and_clear_bit(HRTICK_SET, &rq->hrtick_flags);
+	reset = __test_and_clear_bit(HRTICK_RESET, &rq->hrtick_flags);
+	time = rq->hrtick_expire;
+	clear_thread_flag(TIF_HRTICK_RESCHED);
+	spin_unlock_irqrestore(&rq->lock, flags);
+
+	if (set) {
+		hrtimer_start(&rq->hrtick_timer, time, HRTIMER_MODE_ABS);
+		if (reset && !hrtimer_active(&rq->hrtick_timer))
+			resched_rq(rq);
+	} else
+		hrtick_clear(rq);
+}
+
+/*
+ * High-resolution timer tick.
+ * Runs from hardirq context with interrupts disabled.
+ */
+static enum hrtimer_restart hrtick(struct hrtimer *timer)
+{
+	struct rq *rq = container_of(timer, struct rq, hrtick_timer);
+
+	WARN_ON_ONCE(cpu_of(rq) != smp_processor_id());
+
+	spin_lock(&rq->lock);
+	__update_rq_clock(rq);
+	rq->curr->sched_class->task_tick(rq, rq->curr, 1);
+	spin_unlock(&rq->lock);
+
+	return HRTIMER_NORESTART;
+}
+
+static inline void init_rq_hrtick(struct rq *rq)
+{
+	rq->hrtick_flags = 0;
+	hrtimer_init(&rq->hrtick_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	rq->hrtick_timer.function = hrtick;
+	rq->hrtick_timer.cb_mode = HRTIMER_CB_IRQSAFE_NO_SOFTIRQ;
+}
+
+void hrtick_resched(void)
+{
+	struct rq *rq;
+	unsigned long flags;
+
+	if (!test_thread_flag(TIF_HRTICK_RESCHED))
+		return;
+
+	local_irq_save(flags);
+	rq = cpu_rq(smp_processor_id());
+	hrtick_set(rq);
+	local_irq_restore(flags);
+}
+#else
+static inline void hrtick_clear(struct rq *rq)
+{
+}
+
+static inline void hrtick_set(struct rq *rq)
+{
+}
+
+static inline void init_rq_hrtick(struct rq *rq)
+{
+}
+
+void hrtick_resched(void)
+{
+}
+#endif
+
 /*
  * resched_task - mark a task 'to be rescheduled now'.
  *
@@ -696,16 +1053,16 @@ EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event);
 #define tsk_is_polling(t) test_tsk_thread_flag(t, TIF_POLLING_NRFLAG)
 #endif
 
-static void resched_task(struct task_struct *p)
+static void __resched_task(struct task_struct *p, int tif_bit)
 {
 	int cpu;
 
 	assert_spin_locked(&task_rq(p)->lock);
 
-	if (unlikely(test_tsk_thread_flag(p, TIF_NEED_RESCHED)))
+	if (unlikely(test_tsk_thread_flag(p, tif_bit)))
 		return;
 
-	set_tsk_thread_flag(p, TIF_NEED_RESCHED);
+	set_tsk_thread_flag(p, tif_bit);
 
 	cpu = task_cpu(p);
 	if (cpu == smp_processor_id())
@@ -728,10 +1085,10 @@ static void resched_cpu(int cpu)
 	spin_unlock_irqrestore(&rq->lock, flags);
 }
 #else
-static inline void resched_task(struct task_struct *p)
+static void __resched_task(struct task_struct *p, int tif_bit)
 {
 	assert_spin_locked(&task_rq(p)->lock);
-	set_tsk_need_resched(p);
+	set_tsk_thread_flag(p, tif_bit);
 }
 #endif
 
@@ -871,6 +1228,23 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime);
 static inline void cpuacct_charge(struct task_struct *tsk, u64 cputime) {}
 #endif
 
+static inline void inc_cpu_load(struct rq *rq, unsigned long load)
+{
+	update_load_add(&rq->load, load);
+}
+
+static inline void dec_cpu_load(struct rq *rq, unsigned long load)
+{
+	update_load_sub(&rq->load, load);
+}
+
+#ifdef CONFIG_SMP
+static unsigned long source_load(int cpu, int type);
+static unsigned long target_load(int cpu, int type);
+static unsigned long cpu_avg_load_per_task(int cpu);
+static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd);
+#endif /* CONFIG_SMP */
+
 #include "sched_stats.h"
 #include "sched_idletask.c"
 #include "sched_fair.c"
@@ -881,41 +1255,14 @@ static inline void cpuacct_charge(struct task_struct *tsk, u64 cputime) {}
 
 #define sched_class_highest (&rt_sched_class)
 
-/*
- * Update delta_exec, delta_fair fields for rq.
- *
- * delta_fair clock advances at a rate inversely proportional to
- * total load (rq->load.weight) on the runqueue, while
- * delta_exec advances at the same rate as wall-clock (provided
- * cpu is not idle).
- *
- * delta_exec / delta_fair is a measure of the (smoothened) load on this
- * runqueue over any given interval. This (smoothened) load is used
- * during load balance.
- *
- * This function is called /before/ updating rq->load
- * and when switching tasks.
- */
-static inline void inc_load(struct rq *rq, const struct task_struct *p)
-{
-	update_load_add(&rq->load, p->se.load.weight);
-}
-
-static inline void dec_load(struct rq *rq, const struct task_struct *p)
-{
-	update_load_sub(&rq->load, p->se.load.weight);
-}
-
 static void inc_nr_running(struct task_struct *p, struct rq *rq)
 {
 	rq->nr_running++;
-	inc_load(rq, p);
 }
 
 static void dec_nr_running(struct task_struct *p, struct rq *rq)
 {
 	rq->nr_running--;
-	dec_load(rq, p);
 }
 
 static void set_load_weight(struct task_struct *p)
@@ -1039,7 +1386,7 @@ unsigned long weighted_cpuload(const int cpu)
 
 static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu)
 {
-	set_task_cfs_rq(p, cpu);
+	set_task_rq(p, cpu);
 #ifdef CONFIG_SMP
 	/*
 	 * After ->cpu is set up to a new value, task_rq_lock(p, ...) can be
@@ -1051,12 +1398,24 @@ static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu)
 #endif
 }
 
+static inline void check_class_changed(struct rq *rq, struct task_struct *p,
+				       const struct sched_class *prev_class,
+				       int oldprio, int running)
+{
+	if (prev_class != p->sched_class) {
+		if (prev_class->switched_from)
+			prev_class->switched_from(rq, p, running);
+		p->sched_class->switched_to(rq, p, running);
+	} else
+		p->sched_class->prio_changed(rq, p, oldprio, running);
+}
+
 #ifdef CONFIG_SMP
 
 /*
  * Is this task likely cache-hot:
  */
-static inline int
+static int
 task_hot(struct task_struct *p, u64 now, struct sched_domain *sd)
 {
 	s64 delta;
@@ -1281,7 +1640,7 @@ static unsigned long target_load(int cpu, int type)
 /*
  * Return the average load per task on the cpu's run queue
  */
-static inline unsigned long cpu_avg_load_per_task(int cpu)
+static unsigned long cpu_avg_load_per_task(int cpu)
 {
 	struct rq *rq = cpu_rq(cpu);
 	unsigned long total = weighted_cpuload(cpu);
@@ -1438,58 +1797,6 @@ static int sched_balance_self(int cpu, int flag)
 
 #endif /* CONFIG_SMP */
 
-/*
- * wake_idle() will wake a task on an idle cpu if task->cpu is
- * not idle and an idle cpu is available.  The span of cpus to
- * search starts with cpus closest then further out as needed,
- * so we always favor a closer, idle cpu.
- *
- * Returns the CPU we should wake onto.
- */
-#if defined(ARCH_HAS_SCHED_WAKE_IDLE)
-static int wake_idle(int cpu, struct task_struct *p)
-{
-	cpumask_t tmp;
-	struct sched_domain *sd;
-	int i;
-
-	/*
-	 * If it is idle, then it is the best cpu to run this task.
-	 *
-	 * This cpu is also the best, if it has more than one task already.
-	 * Siblings must be also busy(in most cases) as they didn't already
-	 * pickup the extra load from this cpu and hence we need not check
-	 * sibling runqueue info. This will avoid the checks and cache miss
-	 * penalities associated with that.
-	 */
-	if (idle_cpu(cpu) || cpu_rq(cpu)->nr_running > 1)
-		return cpu;
-
-	for_each_domain(cpu, sd) {
-		if (sd->flags & SD_WAKE_IDLE) {
-			cpus_and(tmp, sd->span, p->cpus_allowed);
-			for_each_cpu_mask(i, tmp) {
-				if (idle_cpu(i)) {
-					if (i != task_cpu(p)) {
-						schedstat_inc(p,
-							se.nr_wakeups_idle);
-					}
-					return i;
-				}
-			}
-		} else {
-			break;
-		}
-	}
-	return cpu;
-}
-#else
-static inline int wake_idle(int cpu, struct task_struct *p)
-{
-	return cpu;
-}
-#endif
-
 /***
  * try_to_wake_up - wake up a thread
  * @p: the to-be-woken-up thread
@@ -1510,11 +1817,6 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync)
 	unsigned long flags;
 	long old_state;
 	struct rq *rq;
-#ifdef CONFIG_SMP
-	struct sched_domain *sd, *this_sd = NULL;
-	unsigned long load, this_load;
-	int new_cpu;
-#endif
 
 	rq = task_rq_lock(p, &flags);
 	old_state = p->state;
@@ -1532,92 +1834,9 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync)
 	if (unlikely(task_running(rq, p)))
 		goto out_activate;
 
-	new_cpu = cpu;
-
-	schedstat_inc(rq, ttwu_count);
-	if (cpu == this_cpu) {
-		schedstat_inc(rq, ttwu_local);
-		goto out_set_cpu;
-	}
-
-	for_each_domain(this_cpu, sd) {
-		if (cpu_isset(cpu, sd->span)) {
-			schedstat_inc(sd, ttwu_wake_remote);
-			this_sd = sd;
-			break;
-		}
-	}
-
-	if (unlikely(!cpu_isset(this_cpu, p->cpus_allowed)))
-		goto out_set_cpu;
-
-	/*
-	 * Check for affine wakeup and passive balancing possibilities.
-	 */
-	if (this_sd) {
-		int idx = this_sd->wake_idx;
-		unsigned int imbalance;
-
-		imbalance = 100 + (this_sd->imbalance_pct - 100) / 2;
-
-		load = source_load(cpu, idx);
-		this_load = target_load(this_cpu, idx);
-
-		new_cpu = this_cpu; /* Wake to this CPU if we can */
-
-		if (this_sd->flags & SD_WAKE_AFFINE) {
-			unsigned long tl = this_load;
-			unsigned long tl_per_task;
-
-			/*
-			 * Attract cache-cold tasks on sync wakeups:
-			 */
-			if (sync && !task_hot(p, rq->clock, this_sd))
-				goto out_set_cpu;
-
-			schedstat_inc(p, se.nr_wakeups_affine_attempts);
-			tl_per_task = cpu_avg_load_per_task(this_cpu);
-
-			/*
-			 * If sync wakeup then subtract the (maximum possible)
-			 * effect of the currently running task from the load
-			 * of the current CPU:
-			 */
-			if (sync)
-				tl -= current->se.load.weight;
-
-			if ((tl <= load &&
-				tl + target_load(cpu, idx) <= tl_per_task) ||
-			       100*(tl + p->se.load.weight) <= imbalance*load) {
-				/*
-				 * This domain has SD_WAKE_AFFINE and
-				 * p is cache cold in this domain, and
-				 * there is no bad imbalance.
-				 */
-				schedstat_inc(this_sd, ttwu_move_affine);
-				schedstat_inc(p, se.nr_wakeups_affine);
-				goto out_set_cpu;
-			}
-		}
-
-		/*
-		 * Start passive balancing when half the imbalance_pct
-		 * limit is reached.
-		 */
-		if (this_sd->flags & SD_WAKE_BALANCE) {
-			if (imbalance*this_load <= 100*load) {
-				schedstat_inc(this_sd, ttwu_move_balance);
-				schedstat_inc(p, se.nr_wakeups_passive);
-				goto out_set_cpu;
-			}
-		}
-	}
-
-	new_cpu = cpu; /* Could not wake to this_cpu. Wake to cpu instead */
-out_set_cpu:
-	new_cpu = wake_idle(new_cpu, p);
-	if (new_cpu != cpu) {
-		set_task_cpu(p, new_cpu);
+	cpu = p->sched_class->select_task_rq(p, sync);
+	if (cpu != orig_cpu) {
+		set_task_cpu(p, cpu);
 		task_rq_unlock(rq, &flags);
 		/* might preempt at this point */
 		rq = task_rq_lock(p, &flags);
@@ -1631,6 +1850,21 @@ out_set_cpu:
 		cpu = task_cpu(p);
 	}
 
+#ifdef CONFIG_SCHEDSTATS
+	schedstat_inc(rq, ttwu_count);
+	if (cpu == this_cpu)
+		schedstat_inc(rq, ttwu_local);
+	else {
+		struct sched_domain *sd;
+		for_each_domain(this_cpu, sd) {
+			if (cpu_isset(cpu, sd->span)) {
+				schedstat_inc(sd, ttwu_wake_remote);
+				break;
+			}
+		}
+	}
+#endif
+
 out_activate:
 #endif /* CONFIG_SMP */
 	schedstat_inc(p, se.nr_wakeups);
@@ -1649,6 +1883,10 @@ out_activate:
 
 out_running:
 	p->state = TASK_RUNNING;
+#ifdef CONFIG_SMP
+	if (p->sched_class->task_wake_up)
+		p->sched_class->task_wake_up(rq, p);
+#endif
 out:
 	task_rq_unlock(rq, &flags);
 
@@ -1691,7 +1929,7 @@ static void __sched_fork(struct task_struct *p)
 	p->se.wait_max			= 0;
 #endif
 
-	INIT_LIST_HEAD(&p->run_list);
+	INIT_LIST_HEAD(&p->rt.run_list);
 	p->se.on_rq = 0;
 
 #ifdef CONFIG_PREEMPT_NOTIFIERS
@@ -1771,6 +2009,10 @@ void fastcall wake_up_new_task(struct task_struct *p, unsigned long clone_flags)
 		inc_nr_running(p, rq);
 	}
 	check_preempt_curr(rq, p);
+#ifdef CONFIG_SMP
+	if (p->sched_class->task_wake_up)
+		p->sched_class->task_wake_up(rq, p);
+#endif
 	task_rq_unlock(rq, &flags);
 }
 
@@ -1891,6 +2133,11 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev)
 	prev_state = prev->state;
 	finish_arch_switch(prev);
 	finish_lock_switch(rq, prev);
+#ifdef CONFIG_SMP
+	if (current->sched_class->post_schedule)
+		current->sched_class->post_schedule(rq);
+#endif
+
 	fire_sched_in_preempt_notifiers(current);
 	if (mm)
 		mmdrop(mm);
@@ -2124,11 +2371,13 @@ static void double_rq_unlock(struct rq *rq1, struct rq *rq2)
 /*
  * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
  */
-static void double_lock_balance(struct rq *this_rq, struct rq *busiest)
+static int double_lock_balance(struct rq *this_rq, struct rq *busiest)
 	__releases(this_rq->lock)
 	__acquires(busiest->lock)
 	__acquires(this_rq->lock)
 {
+	int ret = 0;
+
 	if (unlikely(!irqs_disabled())) {
 		/* printk() doesn't work good under rq->lock */
 		spin_unlock(&this_rq->lock);
@@ -2139,9 +2388,11 @@ static void double_lock_balance(struct rq *this_rq, struct rq *busiest)
 			spin_unlock(&this_rq->lock);
 			spin_lock(&busiest->lock);
 			spin_lock(&this_rq->lock);
+			ret = 1;
 		} else
 			spin_lock(&busiest->lock);
 	}
+	return ret;
 }
 
 /*
@@ -3485,12 +3736,14 @@ void scheduler_tick(void)
 	/*
 	 * Let rq->clock advance by at least TICK_NSEC:
 	 */
-	if (unlikely(rq->clock < next_tick))
+	if (unlikely(rq->clock < next_tick)) {
 		rq->clock = next_tick;
+		rq->clock_underflows++;
+	}
 	rq->tick_timestamp = rq->clock;
 	update_cpu_load(rq);
-	if (curr != rq->idle) /* FIXME: needed? */
-		curr->sched_class->task_tick(rq, curr);
+	curr->sched_class->task_tick(rq, curr, 0);
+	update_sched_rt_period(rq);
 	spin_unlock(&rq->lock);
 
 #ifdef CONFIG_SMP
@@ -3636,6 +3889,8 @@ need_resched_nonpreemptible:
 
 	schedule_debug(prev);
 
+	hrtick_clear(rq);
+
 	/*
 	 * Do the rq-clock update outside the rq lock:
 	 */
@@ -3654,6 +3909,11 @@ need_resched_nonpreemptible:
 		switch_count = &prev->nvcsw;
 	}
 
+#ifdef CONFIG_SMP
+	if (prev->sched_class->pre_schedule)
+		prev->sched_class->pre_schedule(rq, prev);
+#endif
+
 	if (unlikely(!rq->nr_running))
 		idle_balance(cpu, rq);
 
@@ -3668,14 +3928,20 @@ need_resched_nonpreemptible:
 		++*switch_count;
 
 		context_switch(rq, prev, next); /* unlocks the rq */
+		/*
+		 * the context switch might have flipped the stack from under
+		 * us, hence refresh the local variables.
+		 */
+		cpu = smp_processor_id();
+		rq = cpu_rq(cpu);
 	} else
 		spin_unlock_irq(&rq->lock);
 
-	if (unlikely(reacquire_kernel_lock(current) < 0)) {
-		cpu = smp_processor_id();
-		rq = cpu_rq(cpu);
+	hrtick_set(rq);
+
+	if (unlikely(reacquire_kernel_lock(current) < 0))
 		goto need_resched_nonpreemptible;
-	}
+
 	preempt_enable_no_resched();
 	if (unlikely(test_thread_flag(TIF_NEED_RESCHED)))
 		goto need_resched;
@@ -3691,10 +3957,9 @@ EXPORT_SYMBOL(schedule);
 asmlinkage void __sched preempt_schedule(void)
 {
 	struct thread_info *ti = current_thread_info();
-#ifdef CONFIG_PREEMPT_BKL
 	struct task_struct *task = current;
 	int saved_lock_depth;
-#endif
+
 	/*
 	 * If there is a non-zero preempt_count or interrupts are disabled,
 	 * we do not want to preempt the current task. Just return..
@@ -3710,14 +3975,10 @@ asmlinkage void __sched preempt_schedule(void)
 		 * clear ->lock_depth so that schedule() doesnt
 		 * auto-release the semaphore:
 		 */
-#ifdef CONFIG_PREEMPT_BKL
 		saved_lock_depth = task->lock_depth;
 		task->lock_depth = -1;
-#endif
 		schedule();
-#ifdef CONFIG_PREEMPT_BKL
 		task->lock_depth = saved_lock_depth;
-#endif
 		sub_preempt_count(PREEMPT_ACTIVE);
 
 		/*
@@ -3738,10 +3999,9 @@ EXPORT_SYMBOL(preempt_schedule);
 asmlinkage void __sched preempt_schedule_irq(void)
 {
 	struct thread_info *ti = current_thread_info();
-#ifdef CONFIG_PREEMPT_BKL
 	struct task_struct *task = current;
 	int saved_lock_depth;
-#endif
+
 	/* Catch callers which need to be fixed */
 	BUG_ON(ti->preempt_count || !irqs_disabled());
 
@@ -3753,16 +4013,12 @@ asmlinkage void __sched preempt_schedule_irq(void)
 		 * clear ->lock_depth so that schedule() doesnt
 		 * auto-release the semaphore:
 		 */
-#ifdef CONFIG_PREEMPT_BKL
 		saved_lock_depth = task->lock_depth;
 		task->lock_depth = -1;
-#endif
 		local_irq_enable();
 		schedule();
 		local_irq_disable();
-#ifdef CONFIG_PREEMPT_BKL
 		task->lock_depth = saved_lock_depth;
-#endif
 		sub_preempt_count(PREEMPT_ACTIVE);
 
 		/*
@@ -4019,6 +4275,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio)
 	unsigned long flags;
 	int oldprio, on_rq, running;
 	struct rq *rq;
+	const struct sched_class *prev_class = p->sched_class;
 
 	BUG_ON(prio < 0 || prio > MAX_PRIO);
 
@@ -4044,18 +4301,10 @@ void rt_mutex_setprio(struct task_struct *p, int prio)
 	if (on_rq) {
 		if (running)
 			p->sched_class->set_curr_task(rq);
+
 		enqueue_task(rq, p, 0);
-		/*
-		 * Reschedule if we are currently running on this runqueue and
-		 * our priority decreased, or if we are not currently running on
-		 * this runqueue and our priority is higher than the current's
-		 */
-		if (running) {
-			if (p->prio > oldprio)
-				resched_task(rq->curr);
-		} else {
-			check_preempt_curr(rq, p);
-		}
+
+		check_class_changed(rq, p, prev_class, oldprio, running);
 	}
 	task_rq_unlock(rq, &flags);
 }
@@ -4087,10 +4336,8 @@ void set_user_nice(struct task_struct *p, long nice)
 		goto out_unlock;
 	}
 	on_rq = p->se.on_rq;
-	if (on_rq) {
+	if (on_rq)
 		dequeue_task(rq, p, 0);
-		dec_load(rq, p);
-	}
 
 	p->static_prio = NICE_TO_PRIO(nice);
 	set_load_weight(p);
@@ -4100,7 +4347,6 @@ void set_user_nice(struct task_struct *p, long nice)
 
 	if (on_rq) {
 		enqueue_task(rq, p, 0);
-		inc_load(rq, p);
 		/*
 		 * If the task increased its priority or is running and
 		 * lowered its priority, then reschedule its CPU:
@@ -4258,6 +4504,7 @@ int sched_setscheduler(struct task_struct *p, int policy,
 {
 	int retval, oldprio, oldpolicy = -1, on_rq, running;
 	unsigned long flags;
+	const struct sched_class *prev_class = p->sched_class;
 	struct rq *rq;
 
 	/* may grab non-irq protected spin_locks */
@@ -4351,18 +4598,10 @@ recheck:
 	if (on_rq) {
 		if (running)
 			p->sched_class->set_curr_task(rq);
+
 		activate_task(rq, p, 0);
-		/*
-		 * Reschedule if we are currently running on this runqueue and
-		 * our priority decreased, or if we are not currently running on
-		 * this runqueue and our priority is higher than the current's
-		 */
-		if (running) {
-			if (p->prio > oldprio)
-				resched_task(rq->curr);
-		} else {
-			check_preempt_curr(rq, p);
-		}
+
+		check_class_changed(rq, p, prev_class, oldprio, running);
 	}
 	__task_rq_unlock(rq);
 	spin_unlock_irqrestore(&p->pi_lock, flags);
@@ -4490,13 +4729,13 @@ long sched_setaffinity(pid_t pid, cpumask_t new_mask)
 	struct task_struct *p;
 	int retval;
 
-	mutex_lock(&sched_hotcpu_mutex);
+	get_online_cpus();
 	read_lock(&tasklist_lock);
 
 	p = find_process_by_pid(pid);
 	if (!p) {
 		read_unlock(&tasklist_lock);
-		mutex_unlock(&sched_hotcpu_mutex);
+		put_online_cpus();
 		return -ESRCH;
 	}
 
@@ -4536,7 +4775,7 @@ long sched_setaffinity(pid_t pid, cpumask_t new_mask)
 	}
 out_unlock:
 	put_task_struct(p);
-	mutex_unlock(&sched_hotcpu_mutex);
+	put_online_cpus();
 	return retval;
 }
 
@@ -4593,7 +4832,7 @@ long sched_getaffinity(pid_t pid, cpumask_t *mask)
 	struct task_struct *p;
 	int retval;
 
-	mutex_lock(&sched_hotcpu_mutex);
+	get_online_cpus();
 	read_lock(&tasklist_lock);
 
 	retval = -ESRCH;
@@ -4609,7 +4848,7 @@ long sched_getaffinity(pid_t pid, cpumask_t *mask)
 
 out_unlock:
 	read_unlock(&tasklist_lock);
-	mutex_unlock(&sched_hotcpu_mutex);
+	put_online_cpus();
 
 	return retval;
 }
@@ -4683,7 +4922,8 @@ static void __cond_resched(void)
 	} while (need_resched());
 }
 
-int __sched cond_resched(void)
+#if !defined(CONFIG_PREEMPT) || defined(CONFIG_PREEMPT_VOLUNTARY)
+int __sched _cond_resched(void)
 {
 	if (need_resched() && !(preempt_count() & PREEMPT_ACTIVE) &&
 					system_state == SYSTEM_RUNNING) {
@@ -4692,7 +4932,8 @@ int __sched cond_resched(void)
 	}
 	return 0;
 }
-EXPORT_SYMBOL(cond_resched);
+EXPORT_SYMBOL(_cond_resched);
+#endif
 
 /*
  * cond_resched_lock() - if a reschedule is pending, drop the given lock,
@@ -4890,7 +5131,7 @@ out_unlock:
 
 static const char stat_nam[] = "RSDTtZX";
 
-static void show_task(struct task_struct *p)
+void sched_show_task(struct task_struct *p)
 {
 	unsigned long free = 0;
 	unsigned state;
@@ -4920,8 +5161,7 @@ static void show_task(struct task_struct *p)
 	printk(KERN_CONT "%5lu %5d %6d\n", free,
 		task_pid_nr(p), task_pid_nr(p->real_parent));
 
-	if (state != TASK_RUNNING)
-		show_stack(p, NULL);
+	show_stack(p, NULL);
 }
 
 void show_state_filter(unsigned long state_filter)
@@ -4943,7 +5183,7 @@ void show_state_filter(unsigned long state_filter)
 		 */
 		touch_nmi_watchdog();
 		if (!state_filter || (p->state & state_filter))
-			show_task(p);
+			sched_show_task(p);
 	} while_each_thread(g, p);
 
 	touch_all_softlockup_watchdogs();
@@ -4992,11 +5232,8 @@ void __cpuinit init_idle(struct task_struct *idle, int cpu)
 	spin_unlock_irqrestore(&rq->lock, flags);
 
 	/* Set the preempt count _outside_ the spinlocks! */
-#if defined(CONFIG_PREEMPT) && !defined(CONFIG_PREEMPT_BKL)
-	task_thread_info(idle)->preempt_count = (idle->lock_depth >= 0);
-#else
 	task_thread_info(idle)->preempt_count = 0;
-#endif
+
 	/*
 	 * The idle tasks have their own, simple scheduling class:
 	 */
@@ -5077,7 +5314,13 @@ int set_cpus_allowed(struct task_struct *p, cpumask_t new_mask)
 		goto out;
 	}
 
-	p->cpus_allowed = new_mask;
+	if (p->sched_class->set_cpus_allowed)
+		p->sched_class->set_cpus_allowed(p, &new_mask);
+	else {
+		p->cpus_allowed = new_mask;
+		p->rt.nr_cpus_allowed = cpus_weight(new_mask);
+	}
+
 	/* Can the task run on the task's current CPU? If so, we're done */
 	if (cpu_isset(task_cpu(p), new_mask))
 		goto out;
@@ -5569,9 +5812,6 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu)
 	struct rq *rq;
 
 	switch (action) {
-	case CPU_LOCK_ACQUIRE:
-		mutex_lock(&sched_hotcpu_mutex);
-		break;
 
 	case CPU_UP_PREPARE:
 	case CPU_UP_PREPARE_FROZEN:
@@ -5590,6 +5830,15 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu)
 	case CPU_ONLINE_FROZEN:
 		/* Strictly unnecessary, as first user will wake it. */
 		wake_up_process(cpu_rq(cpu)->migration_thread);
+
+		/* Update our root-domain */
+		rq = cpu_rq(cpu);
+		spin_lock_irqsave(&rq->lock, flags);
+		if (rq->rd) {
+			BUG_ON(!cpu_isset(cpu, rq->rd->span));
+			cpu_set(cpu, rq->rd->online);
+		}
+		spin_unlock_irqrestore(&rq->lock, flags);
 		break;
 
 #ifdef CONFIG_HOTPLUG_CPU
@@ -5640,10 +5889,18 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu)
 		}
 		spin_unlock_irq(&rq->lock);
 		break;
-#endif
-	case CPU_LOCK_RELEASE:
-		mutex_unlock(&sched_hotcpu_mutex);
+
+	case CPU_DOWN_PREPARE:
+		/* Update our root-domain */
+		rq = cpu_rq(cpu);
+		spin_lock_irqsave(&rq->lock, flags);
+		if (rq->rd) {
+			BUG_ON(!cpu_isset(cpu, rq->rd->span));
+			cpu_clear(cpu, rq->rd->online);
+		}
+		spin_unlock_irqrestore(&rq->lock, flags);
 		break;
+#endif
 	}
 	return NOTIFY_OK;
 }
@@ -5831,11 +6088,76 @@ sd_parent_degenerate(struct sched_domain *sd, struct sched_domain *parent)
 	return 1;
 }
 
+static void rq_attach_root(struct rq *rq, struct root_domain *rd)
+{
+	unsigned long flags;
+	const struct sched_class *class;
+
+	spin_lock_irqsave(&rq->lock, flags);
+
+	if (rq->rd) {
+		struct root_domain *old_rd = rq->rd;
+
+		for (class = sched_class_highest; class; class = class->next) {
+			if (class->leave_domain)
+				class->leave_domain(rq);
+		}
+
+		cpu_clear(rq->cpu, old_rd->span);
+		cpu_clear(rq->cpu, old_rd->online);
+
+		if (atomic_dec_and_test(&old_rd->refcount))
+			kfree(old_rd);
+	}
+
+	atomic_inc(&rd->refcount);
+	rq->rd = rd;
+
+	cpu_set(rq->cpu, rd->span);
+	if (cpu_isset(rq->cpu, cpu_online_map))
+		cpu_set(rq->cpu, rd->online);
+
+	for (class = sched_class_highest; class; class = class->next) {
+		if (class->join_domain)
+			class->join_domain(rq);
+	}
+
+	spin_unlock_irqrestore(&rq->lock, flags);
+}
+
+static void init_rootdomain(struct root_domain *rd)
+{
+	memset(rd, 0, sizeof(*rd));
+
+	cpus_clear(rd->span);
+	cpus_clear(rd->online);
+}
+
+static void init_defrootdomain(void)
+{
+	init_rootdomain(&def_root_domain);
+	atomic_set(&def_root_domain.refcount, 1);
+}
+
+static struct root_domain *alloc_rootdomain(void)
+{
+	struct root_domain *rd;
+
+	rd = kmalloc(sizeof(*rd), GFP_KERNEL);
+	if (!rd)
+		return NULL;
+
+	init_rootdomain(rd);
+
+	return rd;
+}
+
 /*
- * Attach the domain 'sd' to 'cpu' as its base domain.  Callers must
+ * Attach the domain 'sd' to 'cpu' as its base domain. Callers must
  * hold the hotplug lock.
  */
-static void cpu_attach_domain(struct sched_domain *sd, int cpu)
+static void
+cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
 {
 	struct rq *rq = cpu_rq(cpu);
 	struct sched_domain *tmp;
@@ -5860,6 +6182,7 @@ static void cpu_attach_domain(struct sched_domain *sd, int cpu)
 
 	sched_domain_debug(sd, cpu);
 
+	rq_attach_root(rq, rd);
 	rcu_assign_pointer(rq->sd, sd);
 }
 
@@ -6228,6 +6551,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd)
 static int build_sched_domains(const cpumask_t *cpu_map)
 {
 	int i;
+	struct root_domain *rd;
 #ifdef CONFIG_NUMA
 	struct sched_group **sched_group_nodes = NULL;
 	int sd_allnodes = 0;
@@ -6244,6 +6568,12 @@ static int build_sched_domains(const cpumask_t *cpu_map)
 	sched_group_nodes_bycpu[first_cpu(*cpu_map)] = sched_group_nodes;
 #endif
 
+	rd = alloc_rootdomain();
+	if (!rd) {
+		printk(KERN_WARNING "Cannot alloc root domain\n");
+		return -ENOMEM;
+	}
+
 	/*
 	 * Set up domains for cpus specified by the cpu_map.
 	 */
@@ -6460,7 +6790,7 @@ static int build_sched_domains(const cpumask_t *cpu_map)
 #else
 		sd = &per_cpu(phys_domains, i);
 #endif
-		cpu_attach_domain(sd, i);
+		cpu_attach_domain(sd, rd, i);
 	}
 
 	return 0;
@@ -6518,7 +6848,7 @@ static void detach_destroy_domains(const cpumask_t *cpu_map)
 	unregister_sched_domain_sysctl();
 
 	for_each_cpu_mask(i, *cpu_map)
-		cpu_attach_domain(NULL, i);
+		cpu_attach_domain(NULL, &def_root_domain, i);
 	synchronize_sched();
 	arch_destroy_sched_domains(cpu_map);
 }
@@ -6548,6 +6878,8 @@ void partition_sched_domains(int ndoms_new, cpumask_t *doms_new)
 {
 	int i, j;
 
+	lock_doms_cur();
+
 	/* always unregister in case we don't destroy any domains */
 	unregister_sched_domain_sysctl();
 
@@ -6588,6 +6920,8 @@ match2:
 	ndoms_cur = ndoms_new;
 
 	register_sched_domain_sysctl();
+
+	unlock_doms_cur();
 }
 
 #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT)
@@ -6595,10 +6929,10 @@ static int arch_reinit_sched_domains(void)
 {
 	int err;
 
-	mutex_lock(&sched_hotcpu_mutex);
+	get_online_cpus();
 	detach_destroy_domains(&cpu_online_map);
 	err = arch_init_sched_domains(&cpu_online_map);
-	mutex_unlock(&sched_hotcpu_mutex);
+	put_online_cpus();
 
 	return err;
 }
@@ -6709,12 +7043,12 @@ void __init sched_init_smp(void)
 {
 	cpumask_t non_isolated_cpus;
 
-	mutex_lock(&sched_hotcpu_mutex);
+	get_online_cpus();
 	arch_init_sched_domains(&cpu_online_map);
 	cpus_andnot(non_isolated_cpus, cpu_possible_map, cpu_isolated_map);
 	if (cpus_empty(non_isolated_cpus))
 		cpu_set(smp_processor_id(), non_isolated_cpus);
-	mutex_unlock(&sched_hotcpu_mutex);
+	put_online_cpus();
 	/* XXX: Theoretical race here - CPU may be hotplugged now */
 	hotcpu_notifier(update_sched_domains, 0);
 
@@ -6722,6 +7056,21 @@ void __init sched_init_smp(void)
 	if (set_cpus_allowed(current, non_isolated_cpus) < 0)
 		BUG();
 	sched_init_granularity();
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	if (nr_cpu_ids == 1)
+		return;
+
+	lb_monitor_task = kthread_create(load_balance_monitor, NULL,
+					 "group_balance");
+	if (!IS_ERR(lb_monitor_task)) {
+		lb_monitor_task->flags |= PF_NOFREEZE;
+		wake_up_process(lb_monitor_task);
+	} else {
+		printk(KERN_ERR "Could not create load balance monitor thread"
+			"(error = %ld) \n", PTR_ERR(lb_monitor_task));
+	}
+#endif
 }
 #else
 void __init sched_init_smp(void)
@@ -6746,13 +7095,87 @@ static void init_cfs_rq(struct cfs_rq *cfs_rq, struct rq *rq)
 	cfs_rq->min_vruntime = (u64)(-(1LL << 20));
 }
 
+static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq)
+{
+	struct rt_prio_array *array;
+	int i;
+
+	array = &rt_rq->active;
+	for (i = 0; i < MAX_RT_PRIO; i++) {
+		INIT_LIST_HEAD(array->queue + i);
+		__clear_bit(i, array->bitmap);
+	}
+	/* delimiter for bitsearch: */
+	__set_bit(MAX_RT_PRIO, array->bitmap);
+
+#if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
+	rt_rq->highest_prio = MAX_RT_PRIO;
+#endif
+#ifdef CONFIG_SMP
+	rt_rq->rt_nr_migratory = 0;
+	rt_rq->overloaded = 0;
+#endif
+
+	rt_rq->rt_time = 0;
+	rt_rq->rt_throttled = 0;
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	rt_rq->rq = rq;
+#endif
+}
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+static void init_tg_cfs_entry(struct rq *rq, struct task_group *tg,
+		struct cfs_rq *cfs_rq, struct sched_entity *se,
+		int cpu, int add)
+{
+	tg->cfs_rq[cpu] = cfs_rq;
+	init_cfs_rq(cfs_rq, rq);
+	cfs_rq->tg = tg;
+	if (add)
+		list_add(&cfs_rq->leaf_cfs_rq_list, &rq->leaf_cfs_rq_list);
+
+	tg->se[cpu] = se;
+	se->cfs_rq = &rq->cfs;
+	se->my_q = cfs_rq;
+	se->load.weight = tg->shares;
+	se->load.inv_weight = div64_64(1ULL<<32, se->load.weight);
+	se->parent = NULL;
+}
+
+static void init_tg_rt_entry(struct rq *rq, struct task_group *tg,
+		struct rt_rq *rt_rq, struct sched_rt_entity *rt_se,
+		int cpu, int add)
+{
+	tg->rt_rq[cpu] = rt_rq;
+	init_rt_rq(rt_rq, rq);
+	rt_rq->tg = tg;
+	rt_rq->rt_se = rt_se;
+	if (add)
+		list_add(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list);
+
+	tg->rt_se[cpu] = rt_se;
+	rt_se->rt_rq = &rq->rt;
+	rt_se->my_q = rt_rq;
+	rt_se->parent = NULL;
+	INIT_LIST_HEAD(&rt_se->run_list);
+}
+#endif
+
 void __init sched_init(void)
 {
 	int highest_cpu = 0;
 	int i, j;
 
+#ifdef CONFIG_SMP
+	init_defrootdomain();
+#endif
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	list_add(&init_task_group.list, &task_groups);
+#endif
+
 	for_each_possible_cpu(i) {
-		struct rt_prio_array *array;
 		struct rq *rq;
 
 		rq = cpu_rq(i);
@@ -6761,52 +7184,39 @@ void __init sched_init(void)
 		rq->nr_running = 0;
 		rq->clock = 1;
 		init_cfs_rq(&rq->cfs, rq);
+		init_rt_rq(&rq->rt, rq);
 #ifdef CONFIG_FAIR_GROUP_SCHED
-		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
-		{
-			struct cfs_rq *cfs_rq = &per_cpu(init_cfs_rq, i);
-			struct sched_entity *se =
-					 &per_cpu(init_sched_entity, i);
-
-			init_cfs_rq_p[i] = cfs_rq;
-			init_cfs_rq(cfs_rq, rq);
-			cfs_rq->tg = &init_task_group;
-			list_add(&cfs_rq->leaf_cfs_rq_list,
-							 &rq->leaf_cfs_rq_list);
-
-			init_sched_entity_p[i] = se;
-			se->cfs_rq = &rq->cfs;
-			se->my_q = cfs_rq;
-			se->load.weight = init_task_group_load;
-			se->load.inv_weight =
-				 div64_64(1ULL<<32, init_task_group_load);
-			se->parent = NULL;
-		}
 		init_task_group.shares = init_task_group_load;
-		spin_lock_init(&init_task_group.lock);
+		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
+		init_tg_cfs_entry(rq, &init_task_group,
+				&per_cpu(init_cfs_rq, i),
+				&per_cpu(init_sched_entity, i), i, 1);
+
+		init_task_group.rt_ratio = sysctl_sched_rt_ratio; /* XXX */
+		INIT_LIST_HEAD(&rq->leaf_rt_rq_list);
+		init_tg_rt_entry(rq, &init_task_group,
+				&per_cpu(init_rt_rq, i),
+				&per_cpu(init_sched_rt_entity, i), i, 1);
 #endif
+		rq->rt_period_expire = 0;
+		rq->rt_throttled = 0;
 
 		for (j = 0; j < CPU_LOAD_IDX_MAX; j++)
 			rq->cpu_load[j] = 0;
 #ifdef CONFIG_SMP
 		rq->sd = NULL;
+		rq->rd = NULL;
 		rq->active_balance = 0;
 		rq->next_balance = jiffies;
 		rq->push_cpu = 0;
 		rq->cpu = i;
 		rq->migration_thread = NULL;
 		INIT_LIST_HEAD(&rq->migration_queue);
+		rq_attach_root(rq, &def_root_domain);
 #endif
+		init_rq_hrtick(rq);
 		atomic_set(&rq->nr_iowait, 0);
-
-		array = &rq->rt.active;
-		for (j = 0; j < MAX_RT_PRIO; j++) {
-			INIT_LIST_HEAD(array->queue + j);
-			__clear_bit(j, array->bitmap);
-		}
 		highest_cpu = i;
-		/* delimiter for bitsearch: */
-		__set_bit(MAX_RT_PRIO, array->bitmap);
 	}
 
 	set_load_weight(&init_task);
@@ -6975,12 +7385,187 @@ void set_curr_task(int cpu, struct task_struct *p)
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 
+#ifdef CONFIG_SMP
+/*
+ * distribute shares of all task groups among their schedulable entities,
+ * to reflect load distribution across cpus.
+ */
+static int rebalance_shares(struct sched_domain *sd, int this_cpu)
+{
+	struct cfs_rq *cfs_rq;
+	struct rq *rq = cpu_rq(this_cpu);
+	cpumask_t sdspan = sd->span;
+	int balanced = 1;
+
+	/* Walk thr' all the task groups that we have */
+	for_each_leaf_cfs_rq(rq, cfs_rq) {
+		int i;
+		unsigned long total_load = 0, total_shares;
+		struct task_group *tg = cfs_rq->tg;
+
+		/* Gather total task load of this group across cpus */
+		for_each_cpu_mask(i, sdspan)
+			total_load += tg->cfs_rq[i]->load.weight;
+
+		/* Nothing to do if this group has no load */
+		if (!total_load)
+			continue;
+
+		/*
+		 * tg->shares represents the number of cpu shares the task group
+		 * is eligible to hold on a single cpu. On N cpus, it is
+		 * eligible to hold (N * tg->shares) number of cpu shares.
+		 */
+		total_shares = tg->shares * cpus_weight(sdspan);
+
+		/*
+		 * redistribute total_shares across cpus as per the task load
+		 * distribution.
+		 */
+		for_each_cpu_mask(i, sdspan) {
+			unsigned long local_load, local_shares;
+
+			local_load = tg->cfs_rq[i]->load.weight;
+			local_shares = (local_load * total_shares) / total_load;
+			if (!local_shares)
+				local_shares = MIN_GROUP_SHARES;
+			if (local_shares == tg->se[i]->load.weight)
+				continue;
+
+			spin_lock_irq(&cpu_rq(i)->lock);
+			set_se_shares(tg->se[i], local_shares);
+			spin_unlock_irq(&cpu_rq(i)->lock);
+			balanced = 0;
+		}
+	}
+
+	return balanced;
+}
+
+/*
+ * How frequently should we rebalance_shares() across cpus?
+ *
+ * The more frequently we rebalance shares, the more accurate is the fairness
+ * of cpu bandwidth distribution between task groups. However higher frequency
+ * also implies increased scheduling overhead.
+ *
+ * sysctl_sched_min_bal_int_shares represents the minimum interval between
+ * consecutive calls to rebalance_shares() in the same sched domain.
+ *
+ * sysctl_sched_max_bal_int_shares represents the maximum interval between
+ * consecutive calls to rebalance_shares() in the same sched domain.
+ *
+ * These settings allows for the appropriate trade-off between accuracy of
+ * fairness and the associated overhead.
+ *
+ */
+
+/* default: 8ms, units: milliseconds */
+const_debug unsigned int sysctl_sched_min_bal_int_shares = 8;
+
+/* default: 128ms, units: milliseconds */
+const_debug unsigned int sysctl_sched_max_bal_int_shares = 128;
+
+/* kernel thread that runs rebalance_shares() periodically */
+static int load_balance_monitor(void *unused)
+{
+	unsigned int timeout = sysctl_sched_min_bal_int_shares;
+	struct sched_param schedparm;
+	int ret;
+
+	/*
+	 * We don't want this thread's execution to be limited by the shares
+	 * assigned to default group (init_task_group). Hence make it run
+	 * as a SCHED_RR RT task at the lowest priority.
+	 */
+	schedparm.sched_priority = 1;
+	ret = sched_setscheduler(current, SCHED_RR, &schedparm);
+	if (ret)
+		printk(KERN_ERR "Couldn't set SCHED_RR policy for load balance"
+				" monitor thread (error = %d) \n", ret);
+
+	while (!kthread_should_stop()) {
+		int i, cpu, balanced = 1;
+
+		/* Prevent cpus going down or coming up */
+		get_online_cpus();
+		/* lockout changes to doms_cur[] array */
+		lock_doms_cur();
+		/*
+		 * Enter a rcu read-side critical section to safely walk rq->sd
+		 * chain on various cpus and to walk task group list
+		 * (rq->leaf_cfs_rq_list) in rebalance_shares().
+		 */
+		rcu_read_lock();
+
+		for (i = 0; i < ndoms_cur; i++) {
+			cpumask_t cpumap = doms_cur[i];
+			struct sched_domain *sd = NULL, *sd_prev = NULL;
+
+			cpu = first_cpu(cpumap);
+
+			/* Find the highest domain at which to balance shares */
+			for_each_domain(cpu, sd) {
+				if (!(sd->flags & SD_LOAD_BALANCE))
+					continue;
+				sd_prev = sd;
+			}
+
+			sd = sd_prev;
+			/* sd == NULL? No load balance reqd in this domain */
+			if (!sd)
+				continue;
+
+			balanced &= rebalance_shares(sd, cpu);
+		}
+
+		rcu_read_unlock();
+
+		unlock_doms_cur();
+		put_online_cpus();
+
+		if (!balanced)
+			timeout = sysctl_sched_min_bal_int_shares;
+		else if (timeout < sysctl_sched_max_bal_int_shares)
+			timeout *= 2;
+
+		msleep_interruptible(timeout);
+	}
+
+	return 0;
+}
+#endif	/* CONFIG_SMP */
+
+static void free_sched_group(struct task_group *tg)
+{
+	int i;
+
+	for_each_possible_cpu(i) {
+		if (tg->cfs_rq)
+			kfree(tg->cfs_rq[i]);
+		if (tg->se)
+			kfree(tg->se[i]);
+		if (tg->rt_rq)
+			kfree(tg->rt_rq[i]);
+		if (tg->rt_se)
+			kfree(tg->rt_se[i]);
+	}
+
+	kfree(tg->cfs_rq);
+	kfree(tg->se);
+	kfree(tg->rt_rq);
+	kfree(tg->rt_se);
+	kfree(tg);
+}
+
 /* allocate runqueue etc for a new task group */
 struct task_group *sched_create_group(void)
 {
 	struct task_group *tg;
 	struct cfs_rq *cfs_rq;
 	struct sched_entity *se;
+	struct rt_rq *rt_rq;
+	struct sched_rt_entity *rt_se;
 	struct rq *rq;
 	int i;
 
@@ -6994,97 +7579,89 @@ struct task_group *sched_create_group(void)
 	tg->se = kzalloc(sizeof(se) * NR_CPUS, GFP_KERNEL);
 	if (!tg->se)
 		goto err;
+	tg->rt_rq = kzalloc(sizeof(rt_rq) * NR_CPUS, GFP_KERNEL);
+	if (!tg->rt_rq)
+		goto err;
+	tg->rt_se = kzalloc(sizeof(rt_se) * NR_CPUS, GFP_KERNEL);
+	if (!tg->rt_se)
+		goto err;
+
+	tg->shares = NICE_0_LOAD;
+	tg->rt_ratio = 0; /* XXX */
 
 	for_each_possible_cpu(i) {
 		rq = cpu_rq(i);
 
-		cfs_rq = kmalloc_node(sizeof(struct cfs_rq), GFP_KERNEL,
-							 cpu_to_node(i));
+		cfs_rq = kmalloc_node(sizeof(struct cfs_rq),
+				GFP_KERNEL|__GFP_ZERO, cpu_to_node(i));
 		if (!cfs_rq)
 			goto err;
 
-		se = kmalloc_node(sizeof(struct sched_entity), GFP_KERNEL,
-							cpu_to_node(i));
+		se = kmalloc_node(sizeof(struct sched_entity),
+				GFP_KERNEL|__GFP_ZERO, cpu_to_node(i));
 		if (!se)
 			goto err;
 
-		memset(cfs_rq, 0, sizeof(struct cfs_rq));
-		memset(se, 0, sizeof(struct sched_entity));
+		rt_rq = kmalloc_node(sizeof(struct rt_rq),
+				GFP_KERNEL|__GFP_ZERO, cpu_to_node(i));
+		if (!rt_rq)
+			goto err;
 
-		tg->cfs_rq[i] = cfs_rq;
-		init_cfs_rq(cfs_rq, rq);
-		cfs_rq->tg = tg;
+		rt_se = kmalloc_node(sizeof(struct sched_rt_entity),
+				GFP_KERNEL|__GFP_ZERO, cpu_to_node(i));
+		if (!rt_se)
+			goto err;
 
-		tg->se[i] = se;
-		se->cfs_rq = &rq->cfs;
-		se->my_q = cfs_rq;
-		se->load.weight = NICE_0_LOAD;
-		se->load.inv_weight = div64_64(1ULL<<32, NICE_0_LOAD);
-		se->parent = NULL;
+		init_tg_cfs_entry(rq, tg, cfs_rq, se, i, 0);
+		init_tg_rt_entry(rq, tg, rt_rq, rt_se, i, 0);
 	}
 
+	lock_task_group_list();
 	for_each_possible_cpu(i) {
 		rq = cpu_rq(i);
 		cfs_rq = tg->cfs_rq[i];
 		list_add_rcu(&cfs_rq->leaf_cfs_rq_list, &rq->leaf_cfs_rq_list);
+		rt_rq = tg->rt_rq[i];
+		list_add_rcu(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list);
 	}
-
-	tg->shares = NICE_0_LOAD;
-	spin_lock_init(&tg->lock);
+	list_add_rcu(&tg->list, &task_groups);
+	unlock_task_group_list();
 
 	return tg;
 
 err:
-	for_each_possible_cpu(i) {
-		if (tg->cfs_rq)
-			kfree(tg->cfs_rq[i]);
-		if (tg->se)
-			kfree(tg->se[i]);
-	}
-	kfree(tg->cfs_rq);
-	kfree(tg->se);
-	kfree(tg);
-
+	free_sched_group(tg);
 	return ERR_PTR(-ENOMEM);
 }
 
 /* rcu callback to free various structures associated with a task group */
-static void free_sched_group(struct rcu_head *rhp)
+static void free_sched_group_rcu(struct rcu_head *rhp)
 {
-	struct task_group *tg = container_of(rhp, struct task_group, rcu);
-	struct cfs_rq *cfs_rq;
-	struct sched_entity *se;
-	int i;
-
 	/* now it should be safe to free those cfs_rqs */
-	for_each_possible_cpu(i) {
-		cfs_rq = tg->cfs_rq[i];
-		kfree(cfs_rq);
-
-		se = tg->se[i];
-		kfree(se);
-	}
-
-	kfree(tg->cfs_rq);
-	kfree(tg->se);
-	kfree(tg);
+	free_sched_group(container_of(rhp, struct task_group, rcu));
 }
 
 /* Destroy runqueue etc associated with a task group */
 void sched_destroy_group(struct task_group *tg)
 {
 	struct cfs_rq *cfs_rq = NULL;
+	struct rt_rq *rt_rq = NULL;
 	int i;
 
+	lock_task_group_list();
 	for_each_possible_cpu(i) {
 		cfs_rq = tg->cfs_rq[i];
 		list_del_rcu(&cfs_rq->leaf_cfs_rq_list);
+		rt_rq = tg->rt_rq[i];
+		list_del_rcu(&rt_rq->leaf_rt_rq_list);
 	}
+	list_del_rcu(&tg->list);
+	unlock_task_group_list();
 
 	BUG_ON(!cfs_rq);
 
 	/* wait for possible concurrent references to cfs_rqs complete */
-	call_rcu(&tg->rcu, free_sched_group);
+	call_rcu(&tg->rcu, free_sched_group_rcu);
 }
 
 /* change task's runqueue when it moves between groups.
@@ -7100,11 +7677,6 @@ void sched_move_task(struct task_struct *tsk)
 
 	rq = task_rq_lock(tsk, &flags);
 
-	if (tsk->sched_class != &fair_sched_class) {
-		set_task_cfs_rq(tsk, task_cpu(tsk));
-		goto done;
-	}
-
 	update_rq_clock(rq);
 
 	running = task_current(rq, tsk);
@@ -7116,7 +7688,7 @@ void sched_move_task(struct task_struct *tsk)
 			tsk->sched_class->put_prev_task(rq, tsk);
 	}
 
-	set_task_cfs_rq(tsk, task_cpu(tsk));
+	set_task_rq(tsk, task_cpu(tsk));
 
 	if (on_rq) {
 		if (unlikely(running))
@@ -7124,53 +7696,82 @@ void sched_move_task(struct task_struct *tsk)
 		enqueue_task(rq, tsk, 0);
 	}
 
-done:
 	task_rq_unlock(rq, &flags);
 }
 
+/* rq->lock to be locked by caller */
 static void set_se_shares(struct sched_entity *se, unsigned long shares)
 {
 	struct cfs_rq *cfs_rq = se->cfs_rq;
 	struct rq *rq = cfs_rq->rq;
 	int on_rq;
 
-	spin_lock_irq(&rq->lock);
+	if (!shares)
+		shares = MIN_GROUP_SHARES;
 
 	on_rq = se->on_rq;
-	if (on_rq)
+	if (on_rq) {
 		dequeue_entity(cfs_rq, se, 0);
+		dec_cpu_load(rq, se->load.weight);
+	}
 
 	se->load.weight = shares;
 	se->load.inv_weight = div64_64((1ULL<<32), shares);
 
-	if (on_rq)
+	if (on_rq) {
 		enqueue_entity(cfs_rq, se, 0);
-
-	spin_unlock_irq(&rq->lock);
+		inc_cpu_load(rq, se->load.weight);
+	}
 }
 
 int sched_group_set_shares(struct task_group *tg, unsigned long shares)
 {
 	int i;
+	struct cfs_rq *cfs_rq;
+	struct rq *rq;
+
+	lock_task_group_list();
+	if (tg->shares == shares)
+		goto done;
+
+	if (shares < MIN_GROUP_SHARES)
+		shares = MIN_GROUP_SHARES;
 
 	/*
-	 * A weight of 0 or 1 can cause arithmetics problems.
-	 * (The default weight is 1024 - so there's no practical
-	 *  limitation from this.)
+	 * Prevent any load balance activity (rebalance_shares,
+	 * load_balance_fair) from referring to this group first,
+	 * by taking it off the rq->leaf_cfs_rq_list on each cpu.
 	 */
-	if (shares < 2)
-		shares = 2;
+	for_each_possible_cpu(i) {
+		cfs_rq = tg->cfs_rq[i];
+		list_del_rcu(&cfs_rq->leaf_cfs_rq_list);
+	}
 
-	spin_lock(&tg->lock);
-	if (tg->shares == shares)
-		goto done;
+	/* wait for any ongoing reference to this group to finish */
+	synchronize_sched();
 
+	/*
+	 * Now we are free to modify the group's share on each cpu
+	 * w/o tripping rebalance_share or load_balance_fair.
+	 */
 	tg->shares = shares;
-	for_each_possible_cpu(i)
+	for_each_possible_cpu(i) {
+		spin_lock_irq(&cpu_rq(i)->lock);
 		set_se_shares(tg->se[i], shares);
+		spin_unlock_irq(&cpu_rq(i)->lock);
+	}
 
+	/*
+	 * Enable load balance activity on this group, by inserting it back on
+	 * each cpu's rq->leaf_cfs_rq_list.
+	 */
+	for_each_possible_cpu(i) {
+		rq = cpu_rq(i);
+		cfs_rq = tg->cfs_rq[i];
+		list_add_rcu(&cfs_rq->leaf_cfs_rq_list, &rq->leaf_cfs_rq_list);
+	}
 done:
-	spin_unlock(&tg->lock);
+	unlock_task_group_list();
 	return 0;
 }
 
@@ -7179,6 +7780,31 @@ unsigned long sched_group_shares(struct task_group *tg)
 	return tg->shares;
 }
 
+/*
+ * Ensure the total rt_ratio <= sysctl_sched_rt_ratio
+ */
+int sched_group_set_rt_ratio(struct task_group *tg, unsigned long rt_ratio)
+{
+	struct task_group *tgi;
+	unsigned long total = 0;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(tgi, &task_groups, list)
+		total += tgi->rt_ratio;
+	rcu_read_unlock();
+
+	if (total + rt_ratio - tg->rt_ratio > sysctl_sched_rt_ratio)
+		return -EINVAL;
+
+	tg->rt_ratio = rt_ratio;
+	return 0;
+}
+
+unsigned long sched_group_rt_ratio(struct task_group *tg)
+{
+	return tg->rt_ratio;
+}
+
 #endif	/* CONFIG_FAIR_GROUP_SCHED */
 
 #ifdef CONFIG_FAIR_CGROUP_SCHED
@@ -7254,12 +7880,30 @@ static u64 cpu_shares_read_uint(struct cgroup *cgrp, struct cftype *cft)
 	return (u64) tg->shares;
 }
 
+static int cpu_rt_ratio_write_uint(struct cgroup *cgrp, struct cftype *cftype,
+		u64 rt_ratio_val)
+{
+	return sched_group_set_rt_ratio(cgroup_tg(cgrp), rt_ratio_val);
+}
+
+static u64 cpu_rt_ratio_read_uint(struct cgroup *cgrp, struct cftype *cft)
+{
+	struct task_group *tg = cgroup_tg(cgrp);
+
+	return (u64) tg->rt_ratio;
+}
+
 static struct cftype cpu_files[] = {
 	{
 		.name = "shares",
 		.read_uint = cpu_shares_read_uint,
 		.write_uint = cpu_shares_write_uint,
 	},
+	{
+		.name = "rt_ratio",
+		.read_uint = cpu_rt_ratio_read_uint,
+		.write_uint = cpu_rt_ratio_write_uint,
+	},
 };
 
 static int cpu_cgroup_populate(struct cgroup_subsys *ss, struct cgroup *cont)

kernel/sched_debug.c

@@ -179,6 +179,7 @@ static void print_cpu(struct seq_file *m, int cpu)
 	PN(prev_clock_raw);
 	P(clock_warps);
 	P(clock_overflows);
+	P(clock_underflows);
 	P(clock_deep_idle_events);
 	PN(clock_max_delta);
 	P(cpu_load[0]);
@@ -299,6 +300,8 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
 	PN(se.exec_max);
 	PN(se.slice_max);
 	PN(se.wait_max);
+	PN(se.wait_sum);
+	P(se.wait_count);
 	P(sched_info.bkl_count);
 	P(se.nr_migrations);
 	P(se.nr_migrations_cold);
@@ -366,6 +369,8 @@ void proc_sched_set_task(struct task_struct *p)
 {
 #ifdef CONFIG_SCHEDSTATS
 	p->se.wait_max				= 0;
+	p->se.wait_sum				= 0;
+	p->se.wait_count			= 0;
 	p->se.sleep_max				= 0;
 	p->se.sum_sleep_runtime			= 0;
 	p->se.block_max				= 0;

kernel/sched_fair.c

@@ -20,6 +20,8 @@
  *  Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
  */
 
+#include <linux/latencytop.h>
+
 /*
  * Targeted preemption latency for CPU-bound tasks:
  * (default: 20ms * (1 + ilog(ncpus)), units: nanoseconds)
@@ -248,8 +250,8 @@ static u64 __sched_period(unsigned long nr_running)
 	unsigned long nr_latency = sched_nr_latency;
 
 	if (unlikely(nr_running > nr_latency)) {
+		period = sysctl_sched_min_granularity;
 		period *= nr_running;
-		do_div(period, nr_latency);
 	}
 
 	return period;
@@ -383,6 +385,9 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
 	schedstat_set(se->wait_max, max(se->wait_max,
 			rq_of(cfs_rq)->clock - se->wait_start));
+	schedstat_set(se->wait_count, se->wait_count + 1);
+	schedstat_set(se->wait_sum, se->wait_sum +
+			rq_of(cfs_rq)->clock - se->wait_start);
 	schedstat_set(se->wait_start, 0);
 }
 
@@ -434,6 +439,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se)
 #ifdef CONFIG_SCHEDSTATS
 	if (se->sleep_start) {
 		u64 delta = rq_of(cfs_rq)->clock - se->sleep_start;
+		struct task_struct *tsk = task_of(se);
 
 		if ((s64)delta < 0)
 			delta = 0;
@@ -443,9 +449,12 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se)
 
 		se->sleep_start = 0;
 		se->sum_sleep_runtime += delta;
+
+		account_scheduler_latency(tsk, delta >> 10, 1);
 	}
 	if (se->block_start) {
 		u64 delta = rq_of(cfs_rq)->clock - se->block_start;
+		struct task_struct *tsk = task_of(se);
 
 		if ((s64)delta < 0)
 			delta = 0;
@@ -462,11 +471,11 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se)
 		 * time that the task spent sleeping:
 		 */
 		if (unlikely(prof_on == SLEEP_PROFILING)) {
-			struct task_struct *tsk = task_of(se);
 
 			profile_hits(SLEEP_PROFILING, (void *)get_wchan(tsk),
 				     delta >> 20);
 		}
+		account_scheduler_latency(tsk, delta >> 10, 0);
 	}
 #endif
 }
@@ -642,13 +651,29 @@ static void put_prev_entity(struct cfs_rq *cfs_rq, struct sched_entity *prev)
 	cfs_rq->curr = NULL;
 }
 
-static void entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr)
+static void
+entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr, int queued)
 {
 	/*
 	 * Update run-time statistics of the 'current'.
 	 */
 	update_curr(cfs_rq);
 
+#ifdef CONFIG_SCHED_HRTICK
+	/*
+	 * queued ticks are scheduled to match the slice, so don't bother
+	 * validating it and just reschedule.
+	 */
+	if (queued)
+		return resched_task(rq_of(cfs_rq)->curr);
+	/*
+	 * don't let the period tick interfere with the hrtick preemption
+	 */
+	if (!sched_feat(DOUBLE_TICK) &&
+			hrtimer_active(&rq_of(cfs_rq)->hrtick_timer))
+		return;
+#endif
+
 	if (cfs_rq->nr_running > 1 || !sched_feat(WAKEUP_PREEMPT))
 		check_preempt_tick(cfs_rq, curr);
 }
@@ -690,7 +715,7 @@ static inline struct cfs_rq *cpu_cfs_rq(struct cfs_rq *cfs_rq, int this_cpu)
 
 /* Iterate thr' all leaf cfs_rq's on a runqueue */
 #define for_each_leaf_cfs_rq(rq, cfs_rq) \
-	list_for_each_entry(cfs_rq, &rq->leaf_cfs_rq_list, leaf_cfs_rq_list)
+	list_for_each_entry_rcu(cfs_rq, &rq->leaf_cfs_rq_list, leaf_cfs_rq_list)
 
 /* Do the two (enqueued) entities belong to the same group ? */
 static inline int
@@ -707,6 +732,8 @@ static inline struct sched_entity *parent_entity(struct sched_entity *se)
 	return se->parent;
 }
 
+#define GROUP_IMBALANCE_PCT	20
+
 #else	/* CONFIG_FAIR_GROUP_SCHED */
 
 #define for_each_sched_entity(se) \
@@ -752,6 +779,43 @@ static inline struct sched_entity *parent_entity(struct sched_entity *se)
 
 #endif	/* CONFIG_FAIR_GROUP_SCHED */
 
+#ifdef CONFIG_SCHED_HRTICK
+static void hrtick_start_fair(struct rq *rq, struct task_struct *p)
+{
+	int requeue = rq->curr == p;
+	struct sched_entity *se = &p->se;
+	struct cfs_rq *cfs_rq = cfs_rq_of(se);
+
+	WARN_ON(task_rq(p) != rq);
+
+	if (hrtick_enabled(rq) && cfs_rq->nr_running > 1) {
+		u64 slice = sched_slice(cfs_rq, se);
+		u64 ran = se->sum_exec_runtime - se->prev_sum_exec_runtime;
+		s64 delta = slice - ran;
+
+		if (delta < 0) {
+			if (rq->curr == p)
+				resched_task(p);
+			return;
+		}
+
+		/*
+		 * Don't schedule slices shorter than 10000ns, that just
+		 * doesn't make sense. Rely on vruntime for fairness.
+		 */
+		if (!requeue)
+			delta = max(10000LL, delta);
+
+		hrtick_start(rq, delta, requeue);
+	}
+}
+#else
+static inline void
+hrtick_start_fair(struct rq *rq, struct task_struct *p)
+{
+}
+#endif
+
 /*
  * The enqueue_task method is called before nr_running is
  * increased. Here we update the fair scheduling stats and
@@ -760,15 +824,28 @@ static inline struct sched_entity *parent_entity(struct sched_entity *se)
 static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup)
 {
 	struct cfs_rq *cfs_rq;
-	struct sched_entity *se = &p->se;
+	struct sched_entity *se = &p->se,
+			    *topse = NULL;	/* Highest schedulable entity */
+	int incload = 1;
 
 	for_each_sched_entity(se) {
-		if (se->on_rq)
+		topse = se;
+		if (se->on_rq) {
+			incload = 0;
 			break;
+		}
 		cfs_rq = cfs_rq_of(se);
 		enqueue_entity(cfs_rq, se, wakeup);
 		wakeup = 1;
 	}
+	/* Increment cpu load if we just enqueued the first task of a group on
+	 * 'rq->cpu'. 'topse' represents the group to which task 'p' belongs
+	 * at the highest grouping level.
+	 */
+	if (incload)
+		inc_cpu_load(rq, topse->load.weight);
+
+	hrtick_start_fair(rq, rq->curr);
 }
 
 /*
@@ -779,16 +856,30 @@ static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup)
 static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int sleep)
 {
 	struct cfs_rq *cfs_rq;
-	struct sched_entity *se = &p->se;
+	struct sched_entity *se = &p->se,
+			    *topse = NULL; 	/* Highest schedulable entity */
+	int decload = 1;
 
 	for_each_sched_entity(se) {
+		topse = se;
 		cfs_rq = cfs_rq_of(se);
 		dequeue_entity(cfs_rq, se, sleep);
 		/* Don't dequeue parent if it has other entities besides us */
-		if (cfs_rq->load.weight)
+		if (cfs_rq->load.weight) {
+			if (parent_entity(se))
+				decload = 0;
 			break;
+		}
 		sleep = 1;
 	}
+	/* Decrement cpu load if we just dequeued the last task of a group on
+	 * 'rq->cpu'. 'topse' represents the group to which task 'p' belongs
+	 * at the highest grouping level.
+	 */
+	if (decload)
+		dec_cpu_load(rq, topse->load.weight);
+
+	hrtick_start_fair(rq, rq->curr);
 }
 
 /*
@@ -835,6 +926,154 @@ static void yield_task_fair(struct rq *rq)
 	se->vruntime = rightmost->vruntime + 1;
 }
 
+/*
+ * wake_idle() will wake a task on an idle cpu if task->cpu is
+ * not idle and an idle cpu is available.  The span of cpus to
+ * search starts with cpus closest then further out as needed,
+ * so we always favor a closer, idle cpu.
+ *
+ * Returns the CPU we should wake onto.
+ */
+#if defined(ARCH_HAS_SCHED_WAKE_IDLE)
+static int wake_idle(int cpu, struct task_struct *p)
+{
+	cpumask_t tmp;
+	struct sched_domain *sd;
+	int i;
+
+	/*
+	 * If it is idle, then it is the best cpu to run this task.
+	 *
+	 * This cpu is also the best, if it has more than one task already.
+	 * Siblings must be also busy(in most cases) as they didn't already
+	 * pickup the extra load from this cpu and hence we need not check
+	 * sibling runqueue info. This will avoid the checks and cache miss
+	 * penalities associated with that.
+	 */
+	if (idle_cpu(cpu) || cpu_rq(cpu)->nr_running > 1)
+		return cpu;
+
+	for_each_domain(cpu, sd) {
+		if (sd->flags & SD_WAKE_IDLE) {
+			cpus_and(tmp, sd->span, p->cpus_allowed);
+			for_each_cpu_mask(i, tmp) {
+				if (idle_cpu(i)) {
+					if (i != task_cpu(p)) {
+						schedstat_inc(p,
+						       se.nr_wakeups_idle);
+					}
+					return i;
+				}
+			}
+		} else {
+			break;
+		}
+	}
+	return cpu;
+}
+#else
+static inline int wake_idle(int cpu, struct task_struct *p)
+{
+	return cpu;
+}
+#endif
+
+#ifdef CONFIG_SMP
+static int select_task_rq_fair(struct task_struct *p, int sync)
+{
+	int cpu, this_cpu;
+	struct rq *rq;
+	struct sched_domain *sd, *this_sd = NULL;
+	int new_cpu;
+
+	cpu      = task_cpu(p);
+	rq       = task_rq(p);
+	this_cpu = smp_processor_id();
+	new_cpu  = cpu;
+
+	if (cpu == this_cpu)
+		goto out_set_cpu;
+
+	for_each_domain(this_cpu, sd) {
+		if (cpu_isset(cpu, sd->span)) {
+			this_sd = sd;
+			break;
+		}
+	}
+
+	if (unlikely(!cpu_isset(this_cpu, p->cpus_allowed)))
+		goto out_set_cpu;
+
+	/*
+	 * Check for affine wakeup and passive balancing possibilities.
+	 */
+	if (this_sd) {
+		int idx = this_sd->wake_idx;
+		unsigned int imbalance;
+		unsigned long load, this_load;
+
+		imbalance = 100 + (this_sd->imbalance_pct - 100) / 2;
+
+		load = source_load(cpu, idx);
+		this_load = target_load(this_cpu, idx);
+
+		new_cpu = this_cpu; /* Wake to this CPU if we can */
+
+		if (this_sd->flags & SD_WAKE_AFFINE) {
+			unsigned long tl = this_load;
+			unsigned long tl_per_task;
+
+			/*
+			 * Attract cache-cold tasks on sync wakeups:
+			 */
+			if (sync && !task_hot(p, rq->clock, this_sd))
+				goto out_set_cpu;
+
+			schedstat_inc(p, se.nr_wakeups_affine_attempts);
+			tl_per_task = cpu_avg_load_per_task(this_cpu);
+
+			/*
+			 * If sync wakeup then subtract the (maximum possible)
+			 * effect of the currently running task from the load
+			 * of the current CPU:
+			 */
+			if (sync)
+				tl -= current->se.load.weight;
+
+			if ((tl <= load &&
+				tl + target_load(cpu, idx) <= tl_per_task) ||
+			       100*(tl + p->se.load.weight) <= imbalance*load) {
+				/*
+				 * This domain has SD_WAKE_AFFINE and
+				 * p is cache cold in this domain, and
+				 * there is no bad imbalance.
+				 */
+				schedstat_inc(this_sd, ttwu_move_affine);
+				schedstat_inc(p, se.nr_wakeups_affine);
+				goto out_set_cpu;
+			}
+		}
+
+		/*
+		 * Start passive balancing when half the imbalance_pct
+		 * limit is reached.
+		 */
+		if (this_sd->flags & SD_WAKE_BALANCE) {
+			if (imbalance*this_load <= 100*load) {
+				schedstat_inc(this_sd, ttwu_move_balance);
+				schedstat_inc(p, se.nr_wakeups_passive);
+				goto out_set_cpu;
+			}
+		}
+	}
+
+	new_cpu = cpu; /* Could not wake to this_cpu. Wake to cpu instead */
+out_set_cpu:
+	return wake_idle(new_cpu, p);
+}
+#endif /* CONFIG_SMP */
+
+
 /*
  * Preempt the current task with a newly woken task if needed:
  */
@@ -876,6 +1115,7 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
 
 static struct task_struct *pick_next_task_fair(struct rq *rq)
 {
+	struct task_struct *p;
 	struct cfs_rq *cfs_rq = &rq->cfs;
 	struct sched_entity *se;
 
@@ -887,7 +1127,10 @@ static struct task_struct *pick_next_task_fair(struct rq *rq)
 		cfs_rq = group_cfs_rq(se);
 	} while (cfs_rq);
 
-	return task_of(se);
+	p = task_of(se);
+	hrtick_start_fair(rq, p);
+
+	return p;
 }
 
 /*
@@ -944,25 +1187,6 @@ static struct task_struct *load_balance_next_fair(void *arg)
 	return __load_balance_iterator(cfs_rq, cfs_rq->rb_load_balance_curr);
 }
 
-#ifdef CONFIG_FAIR_GROUP_SCHED
-static int cfs_rq_best_prio(struct cfs_rq *cfs_rq)
-{
-	struct sched_entity *curr;
-	struct task_struct *p;
-
-	if (!cfs_rq->nr_running)
-		return MAX_PRIO;
-
-	curr = cfs_rq->curr;
-	if (!curr)
-		curr = __pick_next_entity(cfs_rq);
-
-	p = task_of(curr);
-
-	return p->prio;
-}
-#endif
-
 static unsigned long
 load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
 		  unsigned long max_load_move,
@@ -972,28 +1196,45 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
 	struct cfs_rq *busy_cfs_rq;
 	long rem_load_move = max_load_move;
 	struct rq_iterator cfs_rq_iterator;
+	unsigned long load_moved;
 
 	cfs_rq_iterator.start = load_balance_start_fair;
 	cfs_rq_iterator.next = load_balance_next_fair;
 
 	for_each_leaf_cfs_rq(busiest, busy_cfs_rq) {
 #ifdef CONFIG_FAIR_GROUP_SCHED
-		struct cfs_rq *this_cfs_rq;
-		long imbalance;
-		unsigned long maxload;
+		struct cfs_rq *this_cfs_rq = busy_cfs_rq->tg->cfs_rq[this_cpu];
+		unsigned long maxload, task_load, group_weight;
+		unsigned long thisload, per_task_load;
+		struct sched_entity *se = busy_cfs_rq->tg->se[busiest->cpu];
+
+		task_load = busy_cfs_rq->load.weight;
+		group_weight = se->load.weight;
 
-		this_cfs_rq = cpu_cfs_rq(busy_cfs_rq, this_cpu);
+		/*
+		 * 'group_weight' is contributed by tasks of total weight
+		 * 'task_load'. To move 'rem_load_move' worth of weight only,
+		 * we need to move a maximum task load of:
+		 *
+		 * 	maxload = (remload / group_weight) * task_load;
+		 */
+		maxload = (rem_load_move * task_load) / group_weight;
 
-		imbalance = busy_cfs_rq->load.weight - this_cfs_rq->load.weight;
-		/* Don't pull if this_cfs_rq has more load than busy_cfs_rq */
-		if (imbalance <= 0)
+		if (!maxload || !task_load)
 			continue;
 
-		/* Don't pull more than imbalance/2 */
-		imbalance /= 2;
-		maxload = min(rem_load_move, imbalance);
+		per_task_load = task_load / busy_cfs_rq->nr_running;
+		/*
+		 * balance_tasks will try to forcibly move atleast one task if
+		 * possible (because of SCHED_LOAD_SCALE_FUZZ). Avoid that if
+		 * maxload is less than GROUP_IMBALANCE_FUZZ% the per_task_load.
+		 */
+		 if (100 * maxload < GROUP_IMBALANCE_PCT * per_task_load)
+			continue;
 
-		*this_best_prio = cfs_rq_best_prio(this_cfs_rq);
+		/* Disable priority-based load balance */
+		*this_best_prio = 0;
+		thisload = this_cfs_rq->load.weight;
 #else
 # define maxload rem_load_move
 #endif
@@ -1002,11 +1243,33 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
 		 * load_balance_[start|next]_fair iterators
 		 */
 		cfs_rq_iterator.arg = busy_cfs_rq;
-		rem_load_move -= balance_tasks(this_rq, this_cpu, busiest,
+		load_moved = balance_tasks(this_rq, this_cpu, busiest,
 					       maxload, sd, idle, all_pinned,
 					       this_best_prio,
 					       &cfs_rq_iterator);
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
+		/*
+		 * load_moved holds the task load that was moved. The
+		 * effective (group) weight moved would be:
+		 * 	load_moved_eff = load_moved/task_load * group_weight;
+		 */
+		load_moved = (group_weight * load_moved) / task_load;
+
+		/* Adjust shares on both cpus to reflect load_moved */
+		group_weight -= load_moved;
+		set_se_shares(se, group_weight);
+
+		se = busy_cfs_rq->tg->se[this_cpu];
+		if (!thisload)
+			group_weight = load_moved;
+		else
+			group_weight = se->load.weight + load_moved;
+		set_se_shares(se, group_weight);
+#endif
+
+		rem_load_move -= load_moved;
+
 		if (rem_load_move <= 0)
 			break;
 	}
@@ -1042,14 +1305,14 @@ move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest,
 /*
  * scheduler tick hitting a task of our scheduling class:
  */
-static void task_tick_fair(struct rq *rq, struct task_struct *curr)
+static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued)
 {
 	struct cfs_rq *cfs_rq;
 	struct sched_entity *se = &curr->se;
 
 	for_each_sched_entity(se) {
 		cfs_rq = cfs_rq_of(se);
-		entity_tick(cfs_rq, se);
+		entity_tick(cfs_rq, se, queued);
 	}
 }
 
@@ -1087,6 +1350,42 @@ static void task_new_fair(struct rq *rq, struct task_struct *p)
 	resched_task(rq->curr);
 }
 
+/*
+ * Priority of the task has changed. Check to see if we preempt
+ * the current task.
+ */
+static void prio_changed_fair(struct rq *rq, struct task_struct *p,
+			      int oldprio, int running)
+{
+	/*
+	 * Reschedule if we are currently running on this runqueue and
+	 * our priority decreased, or if we are not currently running on
+	 * this runqueue and our priority is higher than the current's
+	 */
+	if (running) {
+		if (p->prio > oldprio)
+			resched_task(rq->curr);
+	} else
+		check_preempt_curr(rq, p);
+}
+
+/*
+ * We switched to the sched_fair class.
+ */
+static void switched_to_fair(struct rq *rq, struct task_struct *p,
+			     int running)
+{
+	/*
+	 * We were most likely switched from sched_rt, so
+	 * kick off the schedule if running, otherwise just see
+	 * if we can still preempt the current task.
+	 */
+	if (running)
+		resched_task(rq->curr);
+	else
+		check_preempt_curr(rq, p);
+}
+
 /* Account for a task changing its policy or group.
  *
  * This routine is mostly called to set cfs_rq->curr field when a task
@@ -1108,6 +1407,9 @@ static const struct sched_class fair_sched_class = {
 	.enqueue_task		= enqueue_task_fair,
 	.dequeue_task		= dequeue_task_fair,
 	.yield_task		= yield_task_fair,
+#ifdef CONFIG_SMP
+	.select_task_rq		= select_task_rq_fair,
+#endif /* CONFIG_SMP */
 
 	.check_preempt_curr	= check_preempt_wakeup,
 
@@ -1122,6 +1424,9 @@ static const struct sched_class fair_sched_class = {
 	.set_curr_task          = set_curr_task_fair,
 	.task_tick		= task_tick_fair,
 	.task_new		= task_new_fair,
+
+	.prio_changed		= prio_changed_fair,
+	.switched_to		= switched_to_fair,
 };
 
 #ifdef CONFIG_SCHED_DEBUG
@@ -1132,7 +1437,9 @@ static void print_cfs_stats(struct seq_file *m, int cpu)
 #ifdef CONFIG_FAIR_GROUP_SCHED
 	print_cfs_rq(m, cpu, &cpu_rq(cpu)->cfs);
 #endif
+	rcu_read_lock();
 	for_each_leaf_cfs_rq(cpu_rq(cpu), cfs_rq)
 		print_cfs_rq(m, cpu, cfs_rq);
+	rcu_read_unlock();
 }
 #endif

kernel/sched_idletask.c

@@ -5,6 +5,12 @@
  *  handled in sched_fair.c)
  */
 
+#ifdef CONFIG_SMP
+static int select_task_rq_idle(struct task_struct *p, int sync)
+{
+	return task_cpu(p); /* IDLE tasks as never migrated */
+}
+#endif /* CONFIG_SMP */
 /*
  * Idle tasks are unconditionally rescheduled:
  */
@@ -55,7 +61,7 @@ move_one_task_idle(struct rq *this_rq, int this_cpu, struct rq *busiest,
 }
 #endif
 
-static void task_tick_idle(struct rq *rq, struct task_struct *curr)
+static void task_tick_idle(struct rq *rq, struct task_struct *curr, int queued)
 {
 }
 
@@ -63,6 +69,33 @@ static void set_curr_task_idle(struct rq *rq)
 {
 }
 
+static void switched_to_idle(struct rq *rq, struct task_struct *p,
+			     int running)
+{
+	/* Can this actually happen?? */
+	if (running)
+		resched_task(rq->curr);
+	else
+		check_preempt_curr(rq, p);
+}
+
+static void prio_changed_idle(struct rq *rq, struct task_struct *p,
+			      int oldprio, int running)
+{
+	/* This can happen for hot plug CPUS */
+
+	/*
+	 * Reschedule if we are currently running on this runqueue and
+	 * our priority decreased, or if we are not currently running on
+	 * this runqueue and our priority is higher than the current's
+	 */
+	if (running) {
+		if (p->prio > oldprio)
+			resched_task(rq->curr);
+	} else
+		check_preempt_curr(rq, p);
+}
+
 /*
  * Simple, special scheduling class for the per-CPU idle tasks:
  */
@@ -72,6 +105,9 @@ const struct sched_class idle_sched_class = {
 
 	/* dequeue is not valid, we print a debug message there: */
 	.dequeue_task		= dequeue_task_idle,
+#ifdef CONFIG_SMP
+	.select_task_rq		= select_task_rq_idle,
+#endif /* CONFIG_SMP */
 
 	.check_preempt_curr	= check_preempt_curr_idle,
 
@@ -85,5 +121,9 @@ const struct sched_class idle_sched_class = {
 
 	.set_curr_task          = set_curr_task_idle,
 	.task_tick		= task_tick_idle,
+
+	.prio_changed		= prio_changed_idle,
+	.switched_to		= switched_to_idle,
+
 	/* no .task_new for idle tasks */
 };

kernel/sched_rt.c

@@ -3,6 +3,217 @@
  * policies)
  */
 
+#ifdef CONFIG_SMP
+
+static inline int rt_overloaded(struct rq *rq)
+{
+	return atomic_read(&rq->rd->rto_count);
+}
+
+static inline void rt_set_overload(struct rq *rq)
+{
+	cpu_set(rq->cpu, rq->rd->rto_mask);
+	/*
+	 * Make sure the mask is visible before we set
+	 * the overload count. That is checked to determine
+	 * if we should look at the mask. It would be a shame
+	 * if we looked at the mask, but the mask was not
+	 * updated yet.
+	 */
+	wmb();
+	atomic_inc(&rq->rd->rto_count);
+}
+
+static inline void rt_clear_overload(struct rq *rq)
+{
+	/* the order here really doesn't matter */
+	atomic_dec(&rq->rd->rto_count);
+	cpu_clear(rq->cpu, rq->rd->rto_mask);
+}
+
+static void update_rt_migration(struct rq *rq)
+{
+	if (rq->rt.rt_nr_migratory && (rq->rt.rt_nr_running > 1)) {
+		if (!rq->rt.overloaded) {
+			rt_set_overload(rq);
+			rq->rt.overloaded = 1;
+		}
+	} else if (rq->rt.overloaded) {
+		rt_clear_overload(rq);
+		rq->rt.overloaded = 0;
+	}
+}
+#endif /* CONFIG_SMP */
+
+static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
+{
+	return container_of(rt_se, struct task_struct, rt);
+}
+
+static inline int on_rt_rq(struct sched_rt_entity *rt_se)
+{
+	return !list_empty(&rt_se->run_list);
+}
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+
+static inline unsigned int sched_rt_ratio(struct rt_rq *rt_rq)
+{
+	if (!rt_rq->tg)
+		return SCHED_RT_FRAC;
+
+	return rt_rq->tg->rt_ratio;
+}
+
+#define for_each_leaf_rt_rq(rt_rq, rq) \
+	list_for_each_entry(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
+
+static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
+{
+	return rt_rq->rq;
+}
+
+static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
+{
+	return rt_se->rt_rq;
+}
+
+#define for_each_sched_rt_entity(rt_se) \
+	for (; rt_se; rt_se = rt_se->parent)
+
+static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
+{
+	return rt_se->my_q;
+}
+
+static void enqueue_rt_entity(struct sched_rt_entity *rt_se);
+static void dequeue_rt_entity(struct sched_rt_entity *rt_se);
+
+static void sched_rt_ratio_enqueue(struct rt_rq *rt_rq)
+{
+	struct sched_rt_entity *rt_se = rt_rq->rt_se;
+
+	if (rt_se && !on_rt_rq(rt_se) && rt_rq->rt_nr_running) {
+		struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr;
+
+		enqueue_rt_entity(rt_se);
+		if (rt_rq->highest_prio < curr->prio)
+			resched_task(curr);
+	}
+}
+
+static void sched_rt_ratio_dequeue(struct rt_rq *rt_rq)
+{
+	struct sched_rt_entity *rt_se = rt_rq->rt_se;
+
+	if (rt_se && on_rt_rq(rt_se))
+		dequeue_rt_entity(rt_se);
+}
+
+#else
+
+static inline unsigned int sched_rt_ratio(struct rt_rq *rt_rq)
+{
+	return sysctl_sched_rt_ratio;
+}
+
+#define for_each_leaf_rt_rq(rt_rq, rq) \
+	for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
+
+static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
+{
+	return container_of(rt_rq, struct rq, rt);
+}
+
+static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
+{
+	struct task_struct *p = rt_task_of(rt_se);
+	struct rq *rq = task_rq(p);
+
+	return &rq->rt;
+}
+
+#define for_each_sched_rt_entity(rt_se) \
+	for (; rt_se; rt_se = NULL)
+
+static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
+{
+	return NULL;
+}
+
+static inline void sched_rt_ratio_enqueue(struct rt_rq *rt_rq)
+{
+}
+
+static inline void sched_rt_ratio_dequeue(struct rt_rq *rt_rq)
+{
+}
+
+#endif
+
+static inline int rt_se_prio(struct sched_rt_entity *rt_se)
+{
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	struct rt_rq *rt_rq = group_rt_rq(rt_se);
+
+	if (rt_rq)
+		return rt_rq->highest_prio;
+#endif
+
+	return rt_task_of(rt_se)->prio;
+}
+
+static int sched_rt_ratio_exceeded(struct rt_rq *rt_rq)
+{
+	unsigned int rt_ratio = sched_rt_ratio(rt_rq);
+	u64 period, ratio;
+
+	if (rt_ratio == SCHED_RT_FRAC)
+		return 0;
+
+	if (rt_rq->rt_throttled)
+		return 1;
+
+	period = (u64)sysctl_sched_rt_period * NSEC_PER_MSEC;
+	ratio = (period * rt_ratio) >> SCHED_RT_FRAC_SHIFT;
+
+	if (rt_rq->rt_time > ratio) {
+		struct rq *rq = rq_of_rt_rq(rt_rq);
+
+		rq->rt_throttled = 1;
+		rt_rq->rt_throttled = 1;
+
+		sched_rt_ratio_dequeue(rt_rq);
+		return 1;
+	}
+
+	return 0;
+}
+
+static void update_sched_rt_period(struct rq *rq)
+{
+	struct rt_rq *rt_rq;
+	u64 period;
+
+	while (rq->clock > rq->rt_period_expire) {
+		period = (u64)sysctl_sched_rt_period * NSEC_PER_MSEC;
+		rq->rt_period_expire += period;
+
+		for_each_leaf_rt_rq(rt_rq, rq) {
+			unsigned long rt_ratio = sched_rt_ratio(rt_rq);
+			u64 ratio = (period * rt_ratio) >> SCHED_RT_FRAC_SHIFT;
+
+			rt_rq->rt_time -= min(rt_rq->rt_time, ratio);
+			if (rt_rq->rt_throttled) {
+				rt_rq->rt_throttled = 0;
+				sched_rt_ratio_enqueue(rt_rq);
+			}
+		}
+
+		rq->rt_throttled = 0;
+	}
+}
+
 /*
  * Update the current task's runtime statistics. Skip current tasks that
  * are not in our scheduling class.
@@ -10,6 +221,8 @@
 static void update_curr_rt(struct rq *rq)
 {
 	struct task_struct *curr = rq->curr;
+	struct sched_rt_entity *rt_se = &curr->rt;
+	struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
 	u64 delta_exec;
 
 	if (!task_has_rt_policy(curr))
@@ -24,47 +237,228 @@ static void update_curr_rt(struct rq *rq)
 	curr->se.sum_exec_runtime += delta_exec;
 	curr->se.exec_start = rq->clock;
 	cpuacct_charge(curr, delta_exec);
+
+	rt_rq->rt_time += delta_exec;
+	/*
+	 * might make it a tad more accurate:
+	 *
+	 * update_sched_rt_period(rq);
+	 */
+	if (sched_rt_ratio_exceeded(rt_rq))
+		resched_task(curr);
 }
 
-static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
+static inline
+void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+	WARN_ON(!rt_prio(rt_se_prio(rt_se)));
+	rt_rq->rt_nr_running++;
+#if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
+	if (rt_se_prio(rt_se) < rt_rq->highest_prio)
+		rt_rq->highest_prio = rt_se_prio(rt_se);
+#endif
+#ifdef CONFIG_SMP
+	if (rt_se->nr_cpus_allowed > 1) {
+		struct rq *rq = rq_of_rt_rq(rt_rq);
+		rq->rt.rt_nr_migratory++;
+	}
+
+	update_rt_migration(rq_of_rt_rq(rt_rq));
+#endif
+}
+
+static inline
+void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+	WARN_ON(!rt_prio(rt_se_prio(rt_se)));
+	WARN_ON(!rt_rq->rt_nr_running);
+	rt_rq->rt_nr_running--;
+#if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
+	if (rt_rq->rt_nr_running) {
+		struct rt_prio_array *array;
+
+		WARN_ON(rt_se_prio(rt_se) < rt_rq->highest_prio);
+		if (rt_se_prio(rt_se) == rt_rq->highest_prio) {
+			/* recalculate */
+			array = &rt_rq->active;
+			rt_rq->highest_prio =
+				sched_find_first_bit(array->bitmap);
+		} /* otherwise leave rq->highest prio alone */
+	} else
+		rt_rq->highest_prio = MAX_RT_PRIO;
+#endif
+#ifdef CONFIG_SMP
+	if (rt_se->nr_cpus_allowed > 1) {
+		struct rq *rq = rq_of_rt_rq(rt_rq);
+		rq->rt.rt_nr_migratory--;
+	}
+
+	update_rt_migration(rq_of_rt_rq(rt_rq));
+#endif /* CONFIG_SMP */
+}
+
+static void enqueue_rt_entity(struct sched_rt_entity *rt_se)
+{
+	struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
+	struct rt_prio_array *array = &rt_rq->active;
+	struct rt_rq *group_rq = group_rt_rq(rt_se);
+
+	if (group_rq && group_rq->rt_throttled)
+		return;
+
+	list_add_tail(&rt_se->run_list, array->queue + rt_se_prio(rt_se));
+	__set_bit(rt_se_prio(rt_se), array->bitmap);
+
+	inc_rt_tasks(rt_se, rt_rq);
+}
+
+static void dequeue_rt_entity(struct sched_rt_entity *rt_se)
 {
-	struct rt_prio_array *array = &rq->rt.active;
+	struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
+	struct rt_prio_array *array = &rt_rq->active;
+
+	list_del_init(&rt_se->run_list);
+	if (list_empty(array->queue + rt_se_prio(rt_se)))
+		__clear_bit(rt_se_prio(rt_se), array->bitmap);
 
-	list_add_tail(&p->run_list, array->queue + p->prio);
-	__set_bit(p->prio, array->bitmap);
+	dec_rt_tasks(rt_se, rt_rq);
+}
+
+/*
+ * Because the prio of an upper entry depends on the lower
+ * entries, we must remove entries top - down.
+ *
+ * XXX: O(1/2 h^2) because we can only walk up, not down the chain.
+ *      doesn't matter much for now, as h=2 for GROUP_SCHED.
+ */
+static void dequeue_rt_stack(struct task_struct *p)
+{
+	struct sched_rt_entity *rt_se, *top_se;
+
+	/*
+	 * dequeue all, top - down.
+	 */
+	do {
+		rt_se = &p->rt;
+		top_se = NULL;
+		for_each_sched_rt_entity(rt_se) {
+			if (on_rt_rq(rt_se))
+				top_se = rt_se;
+		}
+		if (top_se)
+			dequeue_rt_entity(top_se);
+	} while (top_se);
 }
 
 /*
  * Adding/removing a task to/from a priority array:
  */
+static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
+{
+	struct sched_rt_entity *rt_se = &p->rt;
+
+	if (wakeup)
+		rt_se->timeout = 0;
+
+	dequeue_rt_stack(p);
+
+	/*
+	 * enqueue everybody, bottom - up.
+	 */
+	for_each_sched_rt_entity(rt_se)
+		enqueue_rt_entity(rt_se);
+
+	inc_cpu_load(rq, p->se.load.weight);
+}
+
 static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
 {
-	struct rt_prio_array *array = &rq->rt.active;
+	struct sched_rt_entity *rt_se = &p->rt;
+	struct rt_rq *rt_rq;
 
 	update_curr_rt(rq);
 
-	list_del(&p->run_list);
-	if (list_empty(array->queue + p->prio))
-		__clear_bit(p->prio, array->bitmap);
+	dequeue_rt_stack(p);
+
+	/*
+	 * re-enqueue all non-empty rt_rq entities.
+	 */
+	for_each_sched_rt_entity(rt_se) {
+		rt_rq = group_rt_rq(rt_se);
+		if (rt_rq && rt_rq->rt_nr_running)
+			enqueue_rt_entity(rt_se);
+	}
+
+	dec_cpu_load(rq, p->se.load.weight);
 }
 
 /*
  * Put task to the end of the run list without the overhead of dequeue
  * followed by enqueue.
  */
+static
+void requeue_rt_entity(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se)
+{
+	struct rt_prio_array *array = &rt_rq->active;
+
+	list_move_tail(&rt_se->run_list, array->queue + rt_se_prio(rt_se));
+}
+
 static void requeue_task_rt(struct rq *rq, struct task_struct *p)
 {
-	struct rt_prio_array *array = &rq->rt.active;
+	struct sched_rt_entity *rt_se = &p->rt;
+	struct rt_rq *rt_rq;
 
-	list_move_tail(&p->run_list, array->queue + p->prio);
+	for_each_sched_rt_entity(rt_se) {
+		rt_rq = rt_rq_of_se(rt_se);
+		requeue_rt_entity(rt_rq, rt_se);
+	}
 }
 
-static void
-yield_task_rt(struct rq *rq)
+static void yield_task_rt(struct rq *rq)
 {
 	requeue_task_rt(rq, rq->curr);
 }
 
+#ifdef CONFIG_SMP
+static int find_lowest_rq(struct task_struct *task);
+
+static int select_task_rq_rt(struct task_struct *p, int sync)
+{
+	struct rq *rq = task_rq(p);
+
+	/*
+	 * If the current task is an RT task, then
+	 * try to see if we can wake this RT task up on another
+	 * runqueue. Otherwise simply start this RT task
+	 * on its current runqueue.
+	 *
+	 * We want to avoid overloading runqueues. Even if
+	 * the RT task is of higher priority than the current RT task.
+	 * RT tasks behave differently than other tasks. If
+	 * one gets preempted, we try to push it off to another queue.
+	 * So trying to keep a preempting RT task on the same
+	 * cache hot CPU will force the running RT task to
+	 * a cold CPU. So we waste all the cache for the lower
+	 * RT task in hopes of saving some of a RT task
+	 * that is just being woken and probably will have
+	 * cold cache anyway.
+	 */
+	if (unlikely(rt_task(rq->curr)) &&
+	    (p->rt.nr_cpus_allowed > 1)) {
+		int cpu = find_lowest_rq(p);
+
+		return (cpu == -1) ? task_cpu(p) : cpu;
+	}
+
+	/*
+	 * Otherwise, just let it ride on the affined RQ and the
+	 * post-schedule router will push the preempted task away
+	 */
+	return task_cpu(p);
+}
+#endif /* CONFIG_SMP */
+
 /*
  * Preempt the current task with a newly woken task if needed:
  */
@@ -74,25 +468,48 @@ static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p)
 		resched_task(rq->curr);
 }
 
-static struct task_struct *pick_next_task_rt(struct rq *rq)
+static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
+						   struct rt_rq *rt_rq)
 {
-	struct rt_prio_array *array = &rq->rt.active;
-	struct task_struct *next;
+	struct rt_prio_array *array = &rt_rq->active;
+	struct sched_rt_entity *next = NULL;
 	struct list_head *queue;
 	int idx;
 
 	idx = sched_find_first_bit(array->bitmap);
-	if (idx >= MAX_RT_PRIO)
-		return NULL;
+	BUG_ON(idx >= MAX_RT_PRIO);
 
 	queue = array->queue + idx;
-	next = list_entry(queue->next, struct task_struct, run_list);
-
-	next->se.exec_start = rq->clock;
+	next = list_entry(queue->next, struct sched_rt_entity, run_list);
 
 	return next;
 }
 
+static struct task_struct *pick_next_task_rt(struct rq *rq)
+{
+	struct sched_rt_entity *rt_se;
+	struct task_struct *p;
+	struct rt_rq *rt_rq;
+
+	rt_rq = &rq->rt;
+
+	if (unlikely(!rt_rq->rt_nr_running))
+		return NULL;
+
+	if (sched_rt_ratio_exceeded(rt_rq))
+		return NULL;
+
+	do {
+		rt_se = pick_next_rt_entity(rq, rt_rq);
+		BUG_ON(!rt_se);
+		rt_rq = group_rt_rq(rt_se);
+	} while (rt_rq);
+
+	p = rt_task_of(rt_se);
+	p->se.exec_start = rq->clock;
+	return p;
+}
+
 static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
 {
 	update_curr_rt(rq);
@@ -100,76 +517,448 @@ static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
 }
 
 #ifdef CONFIG_SMP
-/*
- * Load-balancing iterator. Note: while the runqueue stays locked
- * during the whole iteration, the current task might be
- * dequeued so the iterator has to be dequeue-safe. Here we
- * achieve that by always pre-iterating before returning
- * the current task:
- */
-static struct task_struct *load_balance_start_rt(void *arg)
+
+/* Only try algorithms three times */
+#define RT_MAX_TRIES 3
+
+static int double_lock_balance(struct rq *this_rq, struct rq *busiest);
+static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep);
+
+static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu)
 {
-	struct rq *rq = arg;
-	struct rt_prio_array *array = &rq->rt.active;
-	struct list_head *head, *curr;
-	struct task_struct *p;
+	if (!task_running(rq, p) &&
+	    (cpu < 0 || cpu_isset(cpu, p->cpus_allowed)) &&
+	    (p->rt.nr_cpus_allowed > 1))
+		return 1;
+	return 0;
+}
+
+/* Return the second highest RT task, NULL otherwise */
+static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu)
+{
+	struct task_struct *next = NULL;
+	struct sched_rt_entity *rt_se;
+	struct rt_prio_array *array;
+	struct rt_rq *rt_rq;
 	int idx;
 
-	idx = sched_find_first_bit(array->bitmap);
-	if (idx >= MAX_RT_PRIO)
-		return NULL;
+	for_each_leaf_rt_rq(rt_rq, rq) {
+		array = &rt_rq->active;
+		idx = sched_find_first_bit(array->bitmap);
+ next_idx:
+		if (idx >= MAX_RT_PRIO)
+			continue;
+		if (next && next->prio < idx)
+			continue;
+		list_for_each_entry(rt_se, array->queue + idx, run_list) {
+			struct task_struct *p = rt_task_of(rt_se);
+			if (pick_rt_task(rq, p, cpu)) {
+				next = p;
+				break;
+			}
+		}
+		if (!next) {
+			idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1);
+			goto next_idx;
+		}
+	}
 
-	head = array->queue + idx;
-	curr = head->prev;
+	return next;
+}
 
-	p = list_entry(curr, struct task_struct, run_list);
+static DEFINE_PER_CPU(cpumask_t, local_cpu_mask);
 
-	curr = curr->prev;
+static int find_lowest_cpus(struct task_struct *task, cpumask_t *lowest_mask)
+{
+	int       lowest_prio = -1;
+	int       lowest_cpu  = -1;
+	int       count       = 0;
+	int       cpu;
 
-	rq->rt.rt_load_balance_idx = idx;
-	rq->rt.rt_load_balance_head = head;
-	rq->rt.rt_load_balance_curr = curr;
+	cpus_and(*lowest_mask, task_rq(task)->rd->online, task->cpus_allowed);
 
-	return p;
+	/*
+	 * Scan each rq for the lowest prio.
+	 */
+	for_each_cpu_mask(cpu, *lowest_mask) {
+		struct rq *rq = cpu_rq(cpu);
+
+		/* We look for lowest RT prio or non-rt CPU */
+		if (rq->rt.highest_prio >= MAX_RT_PRIO) {
+			/*
+			 * if we already found a low RT queue
+			 * and now we found this non-rt queue
+			 * clear the mask and set our bit.
+			 * Otherwise just return the queue as is
+			 * and the count==1 will cause the algorithm
+			 * to use the first bit found.
+			 */
+			if (lowest_cpu != -1) {
+				cpus_clear(*lowest_mask);
+				cpu_set(rq->cpu, *lowest_mask);
+			}
+			return 1;
+		}
+
+		/* no locking for now */
+		if ((rq->rt.highest_prio > task->prio)
+		    && (rq->rt.highest_prio >= lowest_prio)) {
+			if (rq->rt.highest_prio > lowest_prio) {
+				/* new low - clear old data */
+				lowest_prio = rq->rt.highest_prio;
+				lowest_cpu = cpu;
+				count = 0;
+			}
+			count++;
+		} else
+			cpu_clear(cpu, *lowest_mask);
+	}
+
+	/*
+	 * Clear out all the set bits that represent
+	 * runqueues that were of higher prio than
+	 * the lowest_prio.
+	 */
+	if (lowest_cpu > 0) {
+		/*
+		 * Perhaps we could add another cpumask op to
+		 * zero out bits. Like cpu_zero_bits(cpumask, nrbits);
+		 * Then that could be optimized to use memset and such.
+		 */
+		for_each_cpu_mask(cpu, *lowest_mask) {
+			if (cpu >= lowest_cpu)
+				break;
+			cpu_clear(cpu, *lowest_mask);
+		}
+	}
+
+	return count;
 }
 
-static struct task_struct *load_balance_next_rt(void *arg)
+static inline int pick_optimal_cpu(int this_cpu, cpumask_t *mask)
 {
-	struct rq *rq = arg;
-	struct rt_prio_array *array = &rq->rt.active;
-	struct list_head *head, *curr;
-	struct task_struct *p;
-	int idx;
+	int first;
+
+	/* "this_cpu" is cheaper to preempt than a remote processor */
+	if ((this_cpu != -1) && cpu_isset(this_cpu, *mask))
+		return this_cpu;
+
+	first = first_cpu(*mask);
+	if (first != NR_CPUS)
+		return first;
+
+	return -1;
+}
+
+static int find_lowest_rq(struct task_struct *task)
+{
+	struct sched_domain *sd;
+	cpumask_t *lowest_mask = &__get_cpu_var(local_cpu_mask);
+	int this_cpu = smp_processor_id();
+	int cpu      = task_cpu(task);
+	int count    = find_lowest_cpus(task, lowest_mask);
 
-	idx = rq->rt.rt_load_balance_idx;
-	head = rq->rt.rt_load_balance_head;
-	curr = rq->rt.rt_load_balance_curr;
+	if (!count)
+		return -1; /* No targets found */
 
 	/*
-	 * If we arrived back to the head again then
-	 * iterate to the next queue (if any):
+	 * There is no sense in performing an optimal search if only one
+	 * target is found.
 	 */
-	if (unlikely(head == curr)) {
-		int next_idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1);
+	if (count == 1)
+		return first_cpu(*lowest_mask);
 
-		if (next_idx >= MAX_RT_PRIO)
-			return NULL;
+	/*
+	 * At this point we have built a mask of cpus representing the
+	 * lowest priority tasks in the system.  Now we want to elect
+	 * the best one based on our affinity and topology.
+	 *
+	 * We prioritize the last cpu that the task executed on since
+	 * it is most likely cache-hot in that location.
+	 */
+	if (cpu_isset(cpu, *lowest_mask))
+		return cpu;
+
+	/*
+	 * Otherwise, we consult the sched_domains span maps to figure
+	 * out which cpu is logically closest to our hot cache data.
+	 */
+	if (this_cpu == cpu)
+		this_cpu = -1; /* Skip this_cpu opt if the same */
+
+	for_each_domain(cpu, sd) {
+		if (sd->flags & SD_WAKE_AFFINE) {
+			cpumask_t domain_mask;
+			int       best_cpu;
 
-		idx = next_idx;
-		head = array->queue + idx;
-		curr = head->prev;
+			cpus_and(domain_mask, sd->span, *lowest_mask);
 
-		rq->rt.rt_load_balance_idx = idx;
-		rq->rt.rt_load_balance_head = head;
+			best_cpu = pick_optimal_cpu(this_cpu,
+						    &domain_mask);
+			if (best_cpu != -1)
+				return best_cpu;
+		}
 	}
 
-	p = list_entry(curr, struct task_struct, run_list);
+	/*
+	 * And finally, if there were no matches within the domains
+	 * just give the caller *something* to work with from the compatible
+	 * locations.
+	 */
+	return pick_optimal_cpu(this_cpu, lowest_mask);
+}
 
-	curr = curr->prev;
+/* Will lock the rq it finds */
+static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
+{
+	struct rq *lowest_rq = NULL;
+	int tries;
+	int cpu;
 
-	rq->rt.rt_load_balance_curr = curr;
+	for (tries = 0; tries < RT_MAX_TRIES; tries++) {
+		cpu = find_lowest_rq(task);
 
-	return p;
+		if ((cpu == -1) || (cpu == rq->cpu))
+			break;
+
+		lowest_rq = cpu_rq(cpu);
+
+		/* if the prio of this runqueue changed, try again */
+		if (double_lock_balance(rq, lowest_rq)) {
+			/*
+			 * We had to unlock the run queue. In
+			 * the mean time, task could have
+			 * migrated already or had its affinity changed.
+			 * Also make sure that it wasn't scheduled on its rq.
+			 */
+			if (unlikely(task_rq(task) != rq ||
+				     !cpu_isset(lowest_rq->cpu,
+						task->cpus_allowed) ||
+				     task_running(rq, task) ||
+				     !task->se.on_rq)) {
+
+				spin_unlock(&lowest_rq->lock);
+				lowest_rq = NULL;
+				break;
+			}
+		}
+
+		/* If this rq is still suitable use it. */
+		if (lowest_rq->rt.highest_prio > task->prio)
+			break;
+
+		/* try again */
+		spin_unlock(&lowest_rq->lock);
+		lowest_rq = NULL;
+	}
+
+	return lowest_rq;
+}
+
+/*
+ * If the current CPU has more than one RT task, see if the non
+ * running task can migrate over to a CPU that is running a task
+ * of lesser priority.
+ */
+static int push_rt_task(struct rq *rq)
+{
+	struct task_struct *next_task;
+	struct rq *lowest_rq;
+	int ret = 0;
+	int paranoid = RT_MAX_TRIES;
+
+	if (!rq->rt.overloaded)
+		return 0;
+
+	next_task = pick_next_highest_task_rt(rq, -1);
+	if (!next_task)
+		return 0;
+
+ retry:
+	if (unlikely(next_task == rq->curr)) {
+		WARN_ON(1);
+		return 0;
+	}
+
+	/*
+	 * It's possible that the next_task slipped in of
+	 * higher priority than current. If that's the case
+	 * just reschedule current.
+	 */
+	if (unlikely(next_task->prio < rq->curr->prio)) {
+		resched_task(rq->curr);
+		return 0;
+	}
+
+	/* We might release rq lock */
+	get_task_struct(next_task);
+
+	/* find_lock_lowest_rq locks the rq if found */
+	lowest_rq = find_lock_lowest_rq(next_task, rq);
+	if (!lowest_rq) {
+		struct task_struct *task;
+		/*
+		 * find lock_lowest_rq releases rq->lock
+		 * so it is possible that next_task has changed.
+		 * If it has, then try again.
+		 */
+		task = pick_next_highest_task_rt(rq, -1);
+		if (unlikely(task != next_task) && task && paranoid--) {
+			put_task_struct(next_task);
+			next_task = task;
+			goto retry;
+		}
+		goto out;
+	}
+
+	deactivate_task(rq, next_task, 0);
+	set_task_cpu(next_task, lowest_rq->cpu);
+	activate_task(lowest_rq, next_task, 0);
+
+	resched_task(lowest_rq->curr);
+
+	spin_unlock(&lowest_rq->lock);
+
+	ret = 1;
+out:
+	put_task_struct(next_task);
+
+	return ret;
+}
+
+/*
+ * TODO: Currently we just use the second highest prio task on
+ *       the queue, and stop when it can't migrate (or there's
+ *       no more RT tasks).  There may be a case where a lower
+ *       priority RT task has a different affinity than the
+ *       higher RT task. In this case the lower RT task could
+ *       possibly be able to migrate where as the higher priority
+ *       RT task could not.  We currently ignore this issue.
+ *       Enhancements are welcome!
+ */
+static void push_rt_tasks(struct rq *rq)
+{
+	/* push_rt_task will return true if it moved an RT */
+	while (push_rt_task(rq))
+		;
+}
+
+static int pull_rt_task(struct rq *this_rq)
+{
+	int this_cpu = this_rq->cpu, ret = 0, cpu;
+	struct task_struct *p, *next;
+	struct rq *src_rq;
+
+	if (likely(!rt_overloaded(this_rq)))
+		return 0;
+
+	next = pick_next_task_rt(this_rq);
+
+	for_each_cpu_mask(cpu, this_rq->rd->rto_mask) {
+		if (this_cpu == cpu)
+			continue;
+
+		src_rq = cpu_rq(cpu);
+		/*
+		 * We can potentially drop this_rq's lock in
+		 * double_lock_balance, and another CPU could
+		 * steal our next task - hence we must cause
+		 * the caller to recalculate the next task
+		 * in that case:
+		 */
+		if (double_lock_balance(this_rq, src_rq)) {
+			struct task_struct *old_next = next;
+
+			next = pick_next_task_rt(this_rq);
+			if (next != old_next)
+				ret = 1;
+		}
+
+		/*
+		 * Are there still pullable RT tasks?
+		 */
+		if (src_rq->rt.rt_nr_running <= 1)
+			goto skip;
+
+		p = pick_next_highest_task_rt(src_rq, this_cpu);
+
+		/*
+		 * Do we have an RT task that preempts
+		 * the to-be-scheduled task?
+		 */
+		if (p && (!next || (p->prio < next->prio))) {
+			WARN_ON(p == src_rq->curr);
+			WARN_ON(!p->se.on_rq);
+
+			/*
+			 * There's a chance that p is higher in priority
+			 * than what's currently running on its cpu.
+			 * This is just that p is wakeing up and hasn't
+			 * had a chance to schedule. We only pull
+			 * p if it is lower in priority than the
+			 * current task on the run queue or
+			 * this_rq next task is lower in prio than
+			 * the current task on that rq.
+			 */
+			if (p->prio < src_rq->curr->prio ||
+			    (next && next->prio < src_rq->curr->prio))
+				goto skip;
+
+			ret = 1;
+
+			deactivate_task(src_rq, p, 0);
+			set_task_cpu(p, this_cpu);
+			activate_task(this_rq, p, 0);
+			/*
+			 * We continue with the search, just in
+			 * case there's an even higher prio task
+			 * in another runqueue. (low likelyhood
+			 * but possible)
+			 *
+			 * Update next so that we won't pick a task
+			 * on another cpu with a priority lower (or equal)
+			 * than the one we just picked.
+			 */
+			next = p;
+
+		}
+ skip:
+		spin_unlock(&src_rq->lock);
+	}
+
+	return ret;
+}
+
+static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
+{
+	/* Try to pull RT tasks here if we lower this rq's prio */
+	if (unlikely(rt_task(prev)) && rq->rt.highest_prio > prev->prio)
+		pull_rt_task(rq);
+}
+
+static void post_schedule_rt(struct rq *rq)
+{
+	/*
+	 * If we have more than one rt_task queued, then
+	 * see if we can push the other rt_tasks off to other CPUS.
+	 * Note we may release the rq lock, and since
+	 * the lock was owned by prev, we need to release it
+	 * first via finish_lock_switch and then reaquire it here.
+	 */
+	if (unlikely(rq->rt.overloaded)) {
+		spin_lock_irq(&rq->lock);
+		push_rt_tasks(rq);
+		spin_unlock_irq(&rq->lock);
+	}
+}
+
+
+static void task_wake_up_rt(struct rq *rq, struct task_struct *p)
+{
+	if (!task_running(rq, p) &&
+	    (p->prio >= rq->rt.highest_prio) &&
+	    rq->rt.overloaded)
+		push_rt_tasks(rq);
 }
 
 static unsigned long
@@ -178,38 +967,170 @@ load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
 		struct sched_domain *sd, enum cpu_idle_type idle,
 		int *all_pinned, int *this_best_prio)
 {
-	struct rq_iterator rt_rq_iterator;
-
-	rt_rq_iterator.start = load_balance_start_rt;
-	rt_rq_iterator.next = load_balance_next_rt;
-	/* pass 'busiest' rq argument into
-	 * load_balance_[start|next]_rt iterators
-	 */
-	rt_rq_iterator.arg = busiest;
-
-	return balance_tasks(this_rq, this_cpu, busiest, max_load_move, sd,
-			     idle, all_pinned, this_best_prio, &rt_rq_iterator);
+	/* don't touch RT tasks */
+	return 0;
 }
 
 static int
 move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
 		 struct sched_domain *sd, enum cpu_idle_type idle)
 {
-	struct rq_iterator rt_rq_iterator;
+	/* don't touch RT tasks */
+	return 0;
+}
+
+static void set_cpus_allowed_rt(struct task_struct *p, cpumask_t *new_mask)
+{
+	int weight = cpus_weight(*new_mask);
+
+	BUG_ON(!rt_task(p));
 
-	rt_rq_iterator.start = load_balance_start_rt;
-	rt_rq_iterator.next = load_balance_next_rt;
-	rt_rq_iterator.arg = busiest;
+	/*
+	 * Update the migration status of the RQ if we have an RT task
+	 * which is running AND changing its weight value.
+	 */
+	if (p->se.on_rq && (weight != p->rt.nr_cpus_allowed)) {
+		struct rq *rq = task_rq(p);
+
+		if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
+			rq->rt.rt_nr_migratory++;
+		} else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
+			BUG_ON(!rq->rt.rt_nr_migratory);
+			rq->rt.rt_nr_migratory--;
+		}
+
+		update_rt_migration(rq);
+	}
 
-	return iter_move_one_task(this_rq, this_cpu, busiest, sd, idle,
-				  &rt_rq_iterator);
+	p->cpus_allowed    = *new_mask;
+	p->rt.nr_cpus_allowed = weight;
 }
-#endif
 
-static void task_tick_rt(struct rq *rq, struct task_struct *p)
+/* Assumes rq->lock is held */
+static void join_domain_rt(struct rq *rq)
+{
+	if (rq->rt.overloaded)
+		rt_set_overload(rq);
+}
+
+/* Assumes rq->lock is held */
+static void leave_domain_rt(struct rq *rq)
+{
+	if (rq->rt.overloaded)
+		rt_clear_overload(rq);
+}
+
+/*
+ * When switch from the rt queue, we bring ourselves to a position
+ * that we might want to pull RT tasks from other runqueues.
+ */
+static void switched_from_rt(struct rq *rq, struct task_struct *p,
+			   int running)
+{
+	/*
+	 * If there are other RT tasks then we will reschedule
+	 * and the scheduling of the other RT tasks will handle
+	 * the balancing. But if we are the last RT task
+	 * we may need to handle the pulling of RT tasks
+	 * now.
+	 */
+	if (!rq->rt.rt_nr_running)
+		pull_rt_task(rq);
+}
+#endif /* CONFIG_SMP */
+
+/*
+ * When switching a task to RT, we may overload the runqueue
+ * with RT tasks. In this case we try to push them off to
+ * other runqueues.
+ */
+static void switched_to_rt(struct rq *rq, struct task_struct *p,
+			   int running)
+{
+	int check_resched = 1;
+
+	/*
+	 * If we are already running, then there's nothing
+	 * that needs to be done. But if we are not running
+	 * we may need to preempt the current running task.
+	 * If that current running task is also an RT task
+	 * then see if we can move to another run queue.
+	 */
+	if (!running) {
+#ifdef CONFIG_SMP
+		if (rq->rt.overloaded && push_rt_task(rq) &&
+		    /* Don't resched if we changed runqueues */
+		    rq != task_rq(p))
+			check_resched = 0;
+#endif /* CONFIG_SMP */
+		if (check_resched && p->prio < rq->curr->prio)
+			resched_task(rq->curr);
+	}
+}
+
+/*
+ * Priority of the task has changed. This may cause
+ * us to initiate a push or pull.
+ */
+static void prio_changed_rt(struct rq *rq, struct task_struct *p,
+			    int oldprio, int running)
+{
+	if (running) {
+#ifdef CONFIG_SMP
+		/*
+		 * If our priority decreases while running, we
+		 * may need to pull tasks to this runqueue.
+		 */
+		if (oldprio < p->prio)
+			pull_rt_task(rq);
+		/*
+		 * If there's a higher priority task waiting to run
+		 * then reschedule.
+		 */
+		if (p->prio > rq->rt.highest_prio)
+			resched_task(p);
+#else
+		/* For UP simply resched on drop of prio */
+		if (oldprio < p->prio)
+			resched_task(p);
+#endif /* CONFIG_SMP */
+	} else {
+		/*
+		 * This task is not running, but if it is
+		 * greater than the current running task
+		 * then reschedule.
+		 */
+		if (p->prio < rq->curr->prio)
+			resched_task(rq->curr);
+	}
+}
+
+static void watchdog(struct rq *rq, struct task_struct *p)
+{
+	unsigned long soft, hard;
+
+	if (!p->signal)
+		return;
+
+	soft = p->signal->rlim[RLIMIT_RTTIME].rlim_cur;
+	hard = p->signal->rlim[RLIMIT_RTTIME].rlim_max;
+
+	if (soft != RLIM_INFINITY) {
+		unsigned long next;
+
+		p->rt.timeout++;
+		next = DIV_ROUND_UP(min(soft, hard), USEC_PER_SEC/HZ);
+		if (p->rt.timeout > next)
+			p->it_sched_expires = p->se.sum_exec_runtime;
+	}
+}
+
+static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued)
 {
 	update_curr_rt(rq);
 
+	watchdog(rq, p);
+
 	/*
 	 * RR tasks need a special form of timeslice management.
 	 * FIFO tasks have no timeslices.
@@ -217,16 +1138,16 @@ static void task_tick_rt(struct rq *rq, struct task_struct *p)
 	if (p->policy != SCHED_RR)
 		return;
 
-	if (--p->time_slice)
+	if (--p->rt.time_slice)
 		return;
 
-	p->time_slice = DEF_TIMESLICE;
+	p->rt.time_slice = DEF_TIMESLICE;
 
 	/*
 	 * Requeue to the end of queue if we are not the only element
 	 * on the queue:
 	 */
-	if (p->run_list.prev != p->run_list.next) {
+	if (p->rt.run_list.prev != p->rt.run_list.next) {
 		requeue_task_rt(rq, p);
 		set_tsk_need_resched(p);
 	}
@@ -244,6 +1165,9 @@ const struct sched_class rt_sched_class = {
 	.enqueue_task		= enqueue_task_rt,
 	.dequeue_task		= dequeue_task_rt,
 	.yield_task		= yield_task_rt,
+#ifdef CONFIG_SMP
+	.select_task_rq		= select_task_rq_rt,
+#endif /* CONFIG_SMP */
 
 	.check_preempt_curr	= check_preempt_curr_rt,
 
@@ -253,8 +1177,18 @@ const struct sched_class rt_sched_class = {
 #ifdef CONFIG_SMP
 	.load_balance		= load_balance_rt,
 	.move_one_task		= move_one_task_rt,
+	.set_cpus_allowed       = set_cpus_allowed_rt,
+	.join_domain            = join_domain_rt,
+	.leave_domain           = leave_domain_rt,
+	.pre_schedule		= pre_schedule_rt,
+	.post_schedule		= post_schedule_rt,
+	.task_wake_up		= task_wake_up_rt,
+	.switched_from		= switched_from_rt,
 #endif
 
 	.set_curr_task          = set_curr_task_rt,
 	.task_tick		= task_tick_rt,
+
+	.prio_changed		= prio_changed_rt,
+	.switched_to		= switched_to_rt,
 };

kernel/softlockup.c

@@ -8,6 +8,7 @@
  */
 #include <linux/mm.h>
 #include <linux/cpu.h>
+#include <linux/nmi.h>
 #include <linux/init.h>
 #include <linux/delay.h>
 #include <linux/freezer.h>
@@ -23,8 +24,8 @@ static DEFINE_PER_CPU(unsigned long, touch_timestamp);
 static DEFINE_PER_CPU(unsigned long, print_timestamp);
 static DEFINE_PER_CPU(struct task_struct *, watchdog_task);
 
-static int did_panic;
-int softlockup_thresh = 10;
+static int __read_mostly did_panic;
+unsigned long __read_mostly softlockup_thresh = 60;
 
 static int
 softlock_panic(struct notifier_block *this, unsigned long event, void *ptr)
@@ -45,7 +46,7 @@ static struct notifier_block panic_block = {
  */
 static unsigned long get_timestamp(int this_cpu)
 {
-	return cpu_clock(this_cpu) >> 30;  /* 2^30 ~= 10^9 */
+	return cpu_clock(this_cpu) >> 30LL;  /* 2^30 ~= 10^9 */
 }
 
 void touch_softlockup_watchdog(void)
@@ -100,11 +101,7 @@ void softlockup_tick(void)
 
 	now = get_timestamp(this_cpu);
 
-	/* Wake up the high-prio watchdog task every second: */
-	if (now > (touch_timestamp + 1))
-		wake_up_process(per_cpu(watchdog_task, this_cpu));
-
-	/* Warn about unreasonable 10+ seconds delays: */
+	/* Warn about unreasonable delays: */
 	if (now <= (touch_timestamp + softlockup_thresh))
 		return;
 
@@ -121,12 +118,94 @@ void softlockup_tick(void)
 	spin_unlock(&print_lock);
 }
 
+/*
+ * Have a reasonable limit on the number of tasks checked:
+ */
+unsigned long __read_mostly sysctl_hung_task_check_count = 1024;
+
+/*
+ * Zero means infinite timeout - no checking done:
+ */
+unsigned long __read_mostly sysctl_hung_task_timeout_secs = 120;
+
+unsigned long __read_mostly sysctl_hung_task_warnings = 10;
+
+/*
+ * Only do the hung-tasks check on one CPU:
+ */
+static int check_cpu __read_mostly = -1;
+
+static void check_hung_task(struct task_struct *t, unsigned long now)
+{
+	unsigned long switch_count = t->nvcsw + t->nivcsw;
+
+	if (t->flags & PF_FROZEN)
+		return;
+
+	if (switch_count != t->last_switch_count || !t->last_switch_timestamp) {
+		t->last_switch_count = switch_count;
+		t->last_switch_timestamp = now;
+		return;
+	}
+	if ((long)(now - t->last_switch_timestamp) <
+					sysctl_hung_task_timeout_secs)
+		return;
+	if (sysctl_hung_task_warnings < 0)
+		return;
+	sysctl_hung_task_warnings--;
+
+	/*
+	 * Ok, the task did not get scheduled for more than 2 minutes,
+	 * complain:
+	 */
+	printk(KERN_ERR "INFO: task %s:%d blocked for more than "
+			"%ld seconds.\n", t->comm, t->pid,
+			sysctl_hung_task_timeout_secs);
+	printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
+			" disables this message.\n");
+	sched_show_task(t);
+	__debug_show_held_locks(t);
+
+	t->last_switch_timestamp = now;
+	touch_nmi_watchdog();
+}
+
+/*
+ * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
+ * a really long time (120 seconds). If that happens, print out
+ * a warning.
+ */
+static void check_hung_uninterruptible_tasks(int this_cpu)
+{
+	int max_count = sysctl_hung_task_check_count;
+	unsigned long now = get_timestamp(this_cpu);
+	struct task_struct *g, *t;
+
+	/*
+	 * If the system crashed already then all bets are off,
+	 * do not report extra hung tasks:
+	 */
+	if ((tainted & TAINT_DIE) || did_panic)
+		return;
+
+	read_lock(&tasklist_lock);
+	do_each_thread(g, t) {
+		if (!--max_count)
+			break;
+		if (t->state & TASK_UNINTERRUPTIBLE)
+			check_hung_task(t, now);
+	} while_each_thread(g, t);
+
+	read_unlock(&tasklist_lock);
+}
+
 /*
  * The watchdog thread - runs every second and touches the timestamp.
  */
 static int watchdog(void *__bind_cpu)
 {
 	struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
+	int this_cpu = (long)__bind_cpu;
 
 	sched_setscheduler(current, SCHED_FIFO, &param);
 
@@ -135,13 +214,18 @@ static int watchdog(void *__bind_cpu)
 
 	/*
 	 * Run briefly once per second to reset the softlockup timestamp.
-	 * If this gets delayed for more than 10 seconds then the
+	 * If this gets delayed for more than 60 seconds then the
 	 * debug-printout triggers in softlockup_tick().
 	 */
 	while (!kthread_should_stop()) {
-		set_current_state(TASK_INTERRUPTIBLE);
 		touch_softlockup_watchdog();
-		schedule();
+		msleep_interruptible(10000);
+
+		if (this_cpu != check_cpu)
+			continue;
+
+		if (sysctl_hung_task_timeout_secs)
+			check_hung_uninterruptible_tasks(this_cpu);
 	}
 
 	return 0;
@@ -171,6 +255,7 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
 		break;
 	case CPU_ONLINE:
 	case CPU_ONLINE_FROZEN:
+		check_cpu = any_online_cpu(cpu_online_map);
 		wake_up_process(per_cpu(watchdog_task, hotcpu));
 		break;
 #ifdef CONFIG_HOTPLUG_CPU
@@ -181,6 +266,15 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
 		/* Unbind so it can run.  Fall thru. */
 		kthread_bind(per_cpu(watchdog_task, hotcpu),
 			     any_online_cpu(cpu_online_map));
+	case CPU_DOWN_PREPARE:
+	case CPU_DOWN_PREPARE_FROZEN:
+		if (hotcpu == check_cpu) {
+			cpumask_t temp_cpu_online_map = cpu_online_map;
+
+			cpu_clear(hotcpu, temp_cpu_online_map);
+			check_cpu = any_online_cpu(temp_cpu_online_map);
+		}
+		break;
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
 		p = per_cpu(watchdog_task, hotcpu);

kernel/stop_machine.c

@@ -203,13 +203,13 @@ int stop_machine_run(int (*fn)(void *), void *data, unsigned int cpu)
 	int ret;
 
 	/* No CPUs can come up or down during this. */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	p = __stop_machine_run(fn, data, cpu);
 	if (!IS_ERR(p))
 		ret = kthread_stop(p);
 	else
 		ret = PTR_ERR(p);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 
 	return ret;
 }

kernel/sysctl.c

@@ -81,6 +81,7 @@ extern int compat_log;
 extern int maps_protect;
 extern int sysctl_stat_interval;
 extern int audit_argv_kb;
+extern int latencytop_enabled;
 
 /* Constants used for minimum and  maximum */
 #ifdef CONFIG_DETECT_SOFTLOCKUP
@@ -306,9 +307,43 @@ static struct ctl_table kern_table[] = {
 		.procname	= "sched_nr_migrate",
 		.data		= &sysctl_sched_nr_migrate,
 		.maxlen		= sizeof(unsigned int),
-		.mode		= 644,
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "sched_rt_period_ms",
+		.data		= &sysctl_sched_rt_period,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
 		.proc_handler	= &proc_dointvec,
 	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "sched_rt_ratio",
+		.data		= &sysctl_sched_rt_ratio,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
+#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP)
+	{
+		.ctl_name       = CTL_UNNUMBERED,
+		.procname       = "sched_min_bal_int_shares",
+		.data           = &sysctl_sched_min_bal_int_shares,
+		.maxlen         = sizeof(unsigned int),
+		.mode           = 0644,
+		.proc_handler   = &proc_dointvec,
+	},
+	{
+		.ctl_name       = CTL_UNNUMBERED,
+		.procname       = "sched_max_bal_int_shares",
+		.data           = &sysctl_sched_max_bal_int_shares,
+		.maxlen         = sizeof(unsigned int),
+		.mode           = 0644,
+		.proc_handler   = &proc_dointvec,
+	},
+#endif
 #endif
 	{
 		.ctl_name	= CTL_UNNUMBERED,
@@ -382,6 +417,15 @@ static struct ctl_table kern_table[] = {
 		.proc_handler	= &proc_dointvec_taint,
 	},
 #endif
+#ifdef CONFIG_LATENCYTOP
+	{
+		.procname	= "latencytop",
+		.data		= &latencytop_enabled,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
+	},
+#endif
 #ifdef CONFIG_SECURITY_CAPABILITIES
 	{
 		.procname	= "cap-bound",
@@ -728,13 +772,40 @@ static struct ctl_table kern_table[] = {
 		.ctl_name	= CTL_UNNUMBERED,
 		.procname	= "softlockup_thresh",
 		.data		= &softlockup_thresh,
-		.maxlen		= sizeof(int),
+		.maxlen		= sizeof(unsigned long),
 		.mode		= 0644,
-		.proc_handler	= &proc_dointvec_minmax,
+		.proc_handler	= &proc_doulongvec_minmax,
 		.strategy	= &sysctl_intvec,
 		.extra1		= &one,
 		.extra2		= &sixty,
 	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "hung_task_check_count",
+		.data		= &sysctl_hung_task_check_count,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= &proc_doulongvec_minmax,
+		.strategy	= &sysctl_intvec,
+	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "hung_task_timeout_secs",
+		.data		= &sysctl_hung_task_timeout_secs,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= &proc_doulongvec_minmax,
+		.strategy	= &sysctl_intvec,
+	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "hung_task_warnings",
+		.data		= &sysctl_hung_task_warnings,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= &proc_doulongvec_minmax,
+		.strategy	= &sysctl_intvec,
+	},
 #endif
 #ifdef CONFIG_COMPAT
 	{

kernel/time/tick-sched.c

@@ -153,6 +153,7 @@ void tick_nohz_update_jiffies(void)
 void tick_nohz_stop_sched_tick(void)
 {
 	unsigned long seq, last_jiffies, next_jiffies, delta_jiffies, flags;
+	unsigned long rt_jiffies;
 	struct tick_sched *ts;
 	ktime_t last_update, expires, now, delta;
 	struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev;
@@ -216,6 +217,10 @@ void tick_nohz_stop_sched_tick(void)
 	next_jiffies = get_next_timer_interrupt(last_jiffies);
 	delta_jiffies = next_jiffies - last_jiffies;
 
+	rt_jiffies = rt_needs_cpu(cpu);
+	if (rt_jiffies && rt_jiffies < delta_jiffies)
+		delta_jiffies = rt_jiffies;
+
 	if (rcu_needs_cpu(cpu))
 		delta_jiffies = 1;
 	/*
@@ -509,7 +514,6 @@ static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer)
 {
 	struct tick_sched *ts =
 		container_of(timer, struct tick_sched, sched_timer);
-	struct hrtimer_cpu_base *base = timer->base->cpu_base;
 	struct pt_regs *regs = get_irq_regs();
 	ktime_t now = ktime_get();
 	int cpu = smp_processor_id();
@@ -547,15 +551,8 @@ static enum hrtimer_restart tick_sched_timer(struct hrtimer *timer)
 			touch_softlockup_watchdog();
 			ts->idle_jiffies++;
 		}
-		/*
-		 * update_process_times() might take tasklist_lock, hence
-		 * drop the base lock. sched-tick hrtimers are per-CPU and
-		 * never accessible by userspace APIs, so this is safe to do.
-		 */
-		spin_unlock(&base->lock);
 		update_process_times(user_mode(regs));
 		profile_tick(CPU_PROFILING);
-		spin_lock(&base->lock);
 	}
 
 	/* Do not restart, when we are in the idle loop */

kernel/timer.c

@@ -896,7 +896,7 @@ static void run_timer_softirq(struct softirq_action *h)
 {
 	tvec_base_t *base = __get_cpu_var(tvec_bases);
 
-	hrtimer_run_queues();
+	hrtimer_run_pending();
 
 	if (time_after_eq(jiffies, base->timer_jiffies))
 		__run_timers(base);
@@ -907,6 +907,7 @@ static void run_timer_softirq(struct softirq_action *h)
  */
 void run_local_timers(void)
 {
+	hrtimer_run_queues();
 	raise_softirq(TIMER_SOFTIRQ);
 	softlockup_tick();
 }

kernel/user.c

@@ -319,7 +319,7 @@ void free_uid(struct user_struct *up)
 struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 {
 	struct hlist_head *hashent = uidhashentry(ns, uid);
-	struct user_struct *up;
+	struct user_struct *up, *new;
 
 	/* Make uid_hash_find() + uids_user_create() + uid_hash_insert()
 	 * atomic.
@@ -331,13 +331,9 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 	spin_unlock_irq(&uidhash_lock);
 
 	if (!up) {
-		struct user_struct *new;
-
 		new = kmem_cache_alloc(uid_cachep, GFP_KERNEL);
-		if (!new) {
-			uids_mutex_unlock();
-			return NULL;
-		}
+		if (!new)
+			goto out_unlock;
 
 		new->uid = uid;
 		atomic_set(&new->__count, 1);
@@ -353,28 +349,14 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 #endif
 		new->locked_shm = 0;
 
-		if (alloc_uid_keyring(new, current) < 0) {
-			kmem_cache_free(uid_cachep, new);
-			uids_mutex_unlock();
-			return NULL;
-		}
+		if (alloc_uid_keyring(new, current) < 0)
+			goto out_free_user;
 
-		if (sched_create_user(new) < 0) {
-			key_put(new->uid_keyring);
-			key_put(new->session_keyring);
-			kmem_cache_free(uid_cachep, new);
-			uids_mutex_unlock();
-			return NULL;
-		}
+		if (sched_create_user(new) < 0)
+			goto out_put_keys;
 
-		if (uids_user_create(new)) {
-			sched_destroy_user(new);
-			key_put(new->uid_keyring);
-			key_put(new->session_keyring);
-			kmem_cache_free(uid_cachep, new);
-			uids_mutex_unlock();
-			return NULL;
-		}
+		if (uids_user_create(new))
+			goto out_destoy_sched;
 
 		/*
 		 * Before adding this, check whether we raced
@@ -402,6 +384,17 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 	uids_mutex_unlock();
 
 	return up;
+
+out_destoy_sched:
+	sched_destroy_user(new);
+out_put_keys:
+	key_put(new->uid_keyring);
+	key_put(new->session_keyring);
+out_free_user:
+	kmem_cache_free(uid_cachep, new);
+out_unlock:
+	uids_mutex_unlock();
+	return NULL;
 }
 
 void switch_uid(struct user_struct *new_user)

kernel/workqueue.c

@@ -67,9 +67,8 @@ struct workqueue_struct {
 #endif
 };
 
-/* All the per-cpu workqueues on the system, for hotplug cpu to add/remove
-   threads to each one as cpus come/go. */
-static DEFINE_MUTEX(workqueue_mutex);
+/* Serializes the accesses to the list of workqueues. */
+static DEFINE_SPINLOCK(workqueue_lock);
 static LIST_HEAD(workqueues);
 
 static int singlethread_cpu __read_mostly;
@@ -592,8 +591,6 @@ EXPORT_SYMBOL(schedule_delayed_work_on);
  * Returns zero on success.
  * Returns -ve errno on failure.
  *
- * Appears to be racy against CPU hotplug.
- *
  * schedule_on_each_cpu() is very slow.
  */
 int schedule_on_each_cpu(work_func_t func)
@@ -605,7 +602,7 @@ int schedule_on_each_cpu(work_func_t func)
 	if (!works)
 		return -ENOMEM;
 
-	preempt_disable();		/* CPU hotplug */
+	get_online_cpus();
 	for_each_online_cpu(cpu) {
 		struct work_struct *work = per_cpu_ptr(works, cpu);
 
@@ -613,8 +610,8 @@ int schedule_on_each_cpu(work_func_t func)
 		set_bit(WORK_STRUCT_PENDING, work_data_bits(work));
 		__queue_work(per_cpu_ptr(keventd_wq->cpu_wq, cpu), work);
 	}
-	preempt_enable();
 	flush_workqueue(keventd_wq);
+	put_online_cpus();
 	free_percpu(works);
 	return 0;
 }
@@ -750,8 +747,10 @@ struct workqueue_struct *__create_workqueue_key(const char *name,
 		err = create_workqueue_thread(cwq, singlethread_cpu);
 		start_workqueue_thread(cwq, -1);
 	} else {
-		mutex_lock(&workqueue_mutex);
+		get_online_cpus();
+		spin_lock(&workqueue_lock);
 		list_add(&wq->list, &workqueues);
+		spin_unlock(&workqueue_lock);
 
 		for_each_possible_cpu(cpu) {
 			cwq = init_cpu_workqueue(wq, cpu);
@@ -760,7 +759,7 @@ struct workqueue_struct *__create_workqueue_key(const char *name,
 			err = create_workqueue_thread(cwq, cpu);
 			start_workqueue_thread(cwq, cpu);
 		}
-		mutex_unlock(&workqueue_mutex);
+		put_online_cpus();
 	}
 
 	if (err) {
@@ -775,7 +774,7 @@ static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
 {
 	/*
 	 * Our caller is either destroy_workqueue() or CPU_DEAD,
-	 * workqueue_mutex protects cwq->thread
+	 * get_online_cpus() protects cwq->thread.
 	 */
 	if (cwq->thread == NULL)
 		return;
@@ -810,9 +809,11 @@ void destroy_workqueue(struct workqueue_struct *wq)
 	struct cpu_workqueue_struct *cwq;
 	int cpu;
 
-	mutex_lock(&workqueue_mutex);
+	get_online_cpus();
+	spin_lock(&workqueue_lock);
 	list_del(&wq->list);
-	mutex_unlock(&workqueue_mutex);
+	spin_unlock(&workqueue_lock);
+	put_online_cpus();
 
 	for_each_cpu_mask(cpu, *cpu_map) {
 		cwq = per_cpu_ptr(wq->cpu_wq, cpu);
@@ -835,13 +836,6 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb,
 	action &= ~CPU_TASKS_FROZEN;
 
 	switch (action) {
-	case CPU_LOCK_ACQUIRE:
-		mutex_lock(&workqueue_mutex);
-		return NOTIFY_OK;
-
-	case CPU_LOCK_RELEASE:
-		mutex_unlock(&workqueue_mutex);
-		return NOTIFY_OK;
 
 	case CPU_UP_PREPARE:
 		cpu_set(cpu, cpu_populated_map);
@@ -854,7 +848,8 @@ static int __devinit workqueue_cpu_callback(struct notifier_block *nfb,
 		case CPU_UP_PREPARE:
 			if (!create_workqueue_thread(cwq, cpu))
 				break;
-			printk(KERN_ERR "workqueue for %i failed\n", cpu);
+			printk(KERN_ERR "workqueue [%s] for %i failed\n",
+				wq->name, cpu);
 			return NOTIFY_BAD;
 
 		case CPU_ONLINE:

lib/Kconfig.debug

@@ -517,4 +517,18 @@ config FAULT_INJECTION_STACKTRACE_FILTER
 	help
 	  Provide stacktrace filter for fault-injection capabilities
 
+config LATENCYTOP
+	bool "Latency measuring infrastructure"
+	select FRAME_POINTER if !MIPS
+	select KALLSYMS
+	select KALLSYMS_ALL
+	select STACKTRACE
+	select SCHEDSTATS
+	select SCHED_DEBUG
+	depends on X86 || X86_64
+	help
+	  Enable this option if you want to use the LatencyTOP tool
+	  to find out which userspace is blocking on what kernel operations.
+
+
 source "samples/Kconfig"

lib/kernel_lock.c

@@ -9,7 +9,6 @@
 #include <linux/module.h>
 #include <linux/kallsyms.h>
 
-#ifdef CONFIG_PREEMPT_BKL
 /*
  * The 'big kernel semaphore'
  *
@@ -86,128 +85,6 @@ void __lockfunc unlock_kernel(void)
 		up(&kernel_sem);
 }
 
-#else
-
-/*
- * The 'big kernel lock'
- *
- * This spinlock is taken and released recursively by lock_kernel()
- * and unlock_kernel().  It is transparently dropped and reacquired
- * over schedule().  It is used to protect legacy code that hasn't
- * been migrated to a proper locking design yet.
- *
- * Don't use in new code.
- */
-static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(kernel_flag);
-
-
-/*
- * Acquire/release the underlying lock from the scheduler.
- *
- * This is called with preemption disabled, and should
- * return an error value if it cannot get the lock and
- * TIF_NEED_RESCHED gets set.
- *
- * If it successfully gets the lock, it should increment
- * the preemption count like any spinlock does.
- *
- * (This works on UP too - _raw_spin_trylock will never
- * return false in that case)
- */
-int __lockfunc __reacquire_kernel_lock(void)
-{
-	while (!_raw_spin_trylock(&kernel_flag)) {
-		if (test_thread_flag(TIF_NEED_RESCHED))
-			return -EAGAIN;
-		cpu_relax();
-	}
-	preempt_disable();
-	return 0;
-}
-
-void __lockfunc __release_kernel_lock(void)
-{
-	_raw_spin_unlock(&kernel_flag);
-	preempt_enable_no_resched();
-}
-
-/*
- * These are the BKL spinlocks - we try to be polite about preemption. 
- * If SMP is not on (ie UP preemption), this all goes away because the
- * _raw_spin_trylock() will always succeed.
- */
-#ifdef CONFIG_PREEMPT
-static inline void __lock_kernel(void)
-{
-	preempt_disable();
-	if (unlikely(!_raw_spin_trylock(&kernel_flag))) {
-		/*
-		 * If preemption was disabled even before this
-		 * was called, there's nothing we can be polite
-		 * about - just spin.
-		 */
-		if (preempt_count() > 1) {
-			_raw_spin_lock(&kernel_flag);
-			return;
-		}
-
-		/*
-		 * Otherwise, let's wait for the kernel lock
-		 * with preemption enabled..
-		 */
-		do {
-			preempt_enable();
-			while (spin_is_locked(&kernel_flag))
-				cpu_relax();
-			preempt_disable();
-		} while (!_raw_spin_trylock(&kernel_flag));
-	}
-}
-
-#else
-
-/*
- * Non-preemption case - just get the spinlock
- */
-static inline void __lock_kernel(void)
-{
-	_raw_spin_lock(&kernel_flag);
-}
-#endif
-
-static inline void __unlock_kernel(void)
-{
-	/*
-	 * the BKL is not covered by lockdep, so we open-code the
-	 * unlocking sequence (and thus avoid the dep-chain ops):
-	 */
-	_raw_spin_unlock(&kernel_flag);
-	preempt_enable();
-}
-
-/*
- * Getting the big kernel lock.
- *
- * This cannot happen asynchronously, so we only need to
- * worry about other CPU's.
- */
-void __lockfunc lock_kernel(void)
-{
-	int depth = current->lock_depth+1;
-	if (likely(!depth))
-		__lock_kernel();
-	current->lock_depth = depth;
-}
-
-void __lockfunc unlock_kernel(void)
-{
-	BUG_ON(current->lock_depth < 0);
-	if (likely(--current->lock_depth < 0))
-		__unlock_kernel();
-}
-
-#endif
-
 EXPORT_SYMBOL(lock_kernel);
 EXPORT_SYMBOL(unlock_kernel);
 

mm/oom_kill.c

@@ -286,7 +286,7 @@ static void __oom_kill_task(struct task_struct *p, int verbose)
 	 * all the memory it needs. That way it should be able to
 	 * exit() and clear out its resources quickly...
 	 */
-	p->time_slice = HZ;
+	p->rt.time_slice = HZ;
 	set_tsk_thread_flag(p, TIF_MEMDIE);
 
 	force_sig(SIGKILL, p);

mm/slab.c

@@ -730,8 +730,7 @@ static inline void init_lock_keys(void)
 #endif
 
 /*
- * 1. Guard access to the cache-chain.
- * 2. Protect sanity of cpu_online_map against cpu hotplug events
+ * Guard access to the cache-chain.
  */
 static DEFINE_MUTEX(cache_chain_mutex);
 static struct list_head cache_chain;
@@ -1331,12 +1330,11 @@ static int __cpuinit cpuup_callback(struct notifier_block *nfb,
 	int err = 0;
 
 	switch (action) {
-	case CPU_LOCK_ACQUIRE:
-		mutex_lock(&cache_chain_mutex);
-		break;
 	case CPU_UP_PREPARE:
 	case CPU_UP_PREPARE_FROZEN:
+		mutex_lock(&cache_chain_mutex);
 		err = cpuup_prepare(cpu);
+		mutex_unlock(&cache_chain_mutex);
 		break;
 	case CPU_ONLINE:
 	case CPU_ONLINE_FROZEN:
@@ -1373,9 +1371,8 @@ static int __cpuinit cpuup_callback(struct notifier_block *nfb,
 #endif
 	case CPU_UP_CANCELED:
 	case CPU_UP_CANCELED_FROZEN:
+		mutex_lock(&cache_chain_mutex);
 		cpuup_canceled(cpu);
-		break;
-	case CPU_LOCK_RELEASE:
 		mutex_unlock(&cache_chain_mutex);
 		break;
 	}
@@ -2170,6 +2167,7 @@ kmem_cache_create (const char *name, size_t size, size_t align,
 	 * We use cache_chain_mutex to ensure a consistent view of
 	 * cpu_online_map as well.  Please see cpuup_callback
 	 */
+	get_online_cpus();
 	mutex_lock(&cache_chain_mutex);
 
 	list_for_each_entry(pc, &cache_chain, next) {
@@ -2396,6 +2394,7 @@ oops:
 		panic("kmem_cache_create(): failed to create slab `%s'\n",
 		      name);
 	mutex_unlock(&cache_chain_mutex);
+	put_online_cpus();
 	return cachep;
 }
 EXPORT_SYMBOL(kmem_cache_create);
@@ -2547,9 +2546,11 @@ int kmem_cache_shrink(struct kmem_cache *cachep)
 	int ret;
 	BUG_ON(!cachep || in_interrupt());
 
+	get_online_cpus();
 	mutex_lock(&cache_chain_mutex);
 	ret = __cache_shrink(cachep);
 	mutex_unlock(&cache_chain_mutex);
+	put_online_cpus();
 	return ret;
 }
 EXPORT_SYMBOL(kmem_cache_shrink);
@@ -2575,6 +2576,7 @@ void kmem_cache_destroy(struct kmem_cache *cachep)
 	BUG_ON(!cachep || in_interrupt());
 
 	/* Find the cache in the chain of caches. */
+	get_online_cpus();
 	mutex_lock(&cache_chain_mutex);
 	/*
 	 * the chain is never empty, cache_cache is never destroyed
@@ -2584,6 +2586,7 @@ void kmem_cache_destroy(struct kmem_cache *cachep)
 		slab_error(cachep, "Can't free all objects");
 		list_add(&cachep->next, &cache_chain);
 		mutex_unlock(&cache_chain_mutex);
+		put_online_cpus();
 		return;
 	}
 
@@ -2592,6 +2595,7 @@ void kmem_cache_destroy(struct kmem_cache *cachep)
 
 	__kmem_cache_destroy(cachep);
 	mutex_unlock(&cache_chain_mutex);
+	put_online_cpus();
 }
 EXPORT_SYMBOL(kmem_cache_destroy);
 

net/core/flow.c

@@ -293,7 +293,7 @@ void flow_cache_flush(void)
 	static DEFINE_MUTEX(flow_flush_sem);
 
 	/* Don't want cpus going down or up during this. */
-	lock_cpu_hotplug();
+	get_online_cpus();
 	mutex_lock(&flow_flush_sem);
 	atomic_set(&info.cpuleft, num_online_cpus());
 	init_completion(&info.completion);
@@ -305,7 +305,7 @@ void flow_cache_flush(void)
 
 	wait_for_completion(&info.completion);
 	mutex_unlock(&flow_flush_sem);
-	unlock_cpu_hotplug();
+	put_online_cpus();
 }
 
 static void __devinit flow_cache_cpu_prepare(int cpu)