From 7fd7d83d49914f03aefffba6aee09032fcd54cce Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 17 Feb 2009 23:24:03 -0800 Subject: x86/pvops: replace arch_enter_lazy_cpu_mode with arch_start_context_switch Impact: simplification, prepare for later changes Make lazy cpu mode more specific to context switching, so that it makes sense to do more context-switch specific things in the callbacks. Signed-off-by: Jeremy Fitzhardinge Acked-by: Peter Zijlstra --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 5757e03..7530fdd 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2746,7 +2746,7 @@ context_switch(struct rq *rq, struct task_struct *prev, * combine the page table reload and the switch backend into * one hypercall. */ - arch_enter_lazy_cpu_mode(); + arch_start_context_switch(); if (unlikely(!mm)) { next->active_mm = oldmm; -- cgit v1.1 From 224101ed69d3fbb486868e0f6e0f9fa37302efb4 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Wed, 18 Feb 2009 11:18:57 -0800 Subject: x86/paravirt: finish change from lazy cpu to context switch start/end Impact: fix lazy context switch API Pass the previous and next tasks into the context switch start end calls, so that the called functions can properly access the task state (esp in end_context_switch, in which the next task is not yet completely current). Signed-off-by: Jeremy Fitzhardinge Acked-by: Peter Zijlstra --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 7530fdd..133762a 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2746,7 +2746,7 @@ context_switch(struct rq *rq, struct task_struct *prev, * combine the page table reload and the switch backend into * one hypercall. */ - arch_start_context_switch(); + arch_start_context_switch(prev); if (unlikely(!mm)) { next->active_mm = oldmm; -- cgit v1.1 From a26b89f05d194413c7238e0bea071054f6b5d3c8 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Fri, 3 Apr 2009 16:43:34 +0200 Subject: sched, hw-branch-tracer: add wait_task_context_switch() function to sched.h Add a function to wait until some other task has been switched out at least once. This differs from wait_task_inactive() subtly, in that the latter will wait until the task has left the CPU. Signed-off-by: Markus Metzger Cc: markus.t.metzger@gmail.com Cc: roland@redhat.com Cc: eranian@googlemail.com Cc: oleg@redhat.com Cc: juan.villacis@intel.com Cc: ak@linux.jf.intel.com LKML-Reference: <20090403144549.794157000@intel.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 43 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 43 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 6cc1fd5..f91bc81 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2003,6 +2003,49 @@ migrate_task(struct task_struct *p, int dest_cpu, struct migration_req *req) } /* + * wait_task_context_switch - wait for a thread to complete at least one + * context switch. + * + * @p must not be current. + */ +void wait_task_context_switch(struct task_struct *p) +{ + unsigned long nvcsw, nivcsw, flags; + int running; + struct rq *rq; + + nvcsw = p->nvcsw; + nivcsw = p->nivcsw; + for (;;) { + /* + * The runqueue is assigned before the actual context + * switch. We need to take the runqueue lock. + * + * We could check initially without the lock but it is + * very likely that we need to take the lock in every + * iteration. + */ + rq = task_rq_lock(p, &flags); + running = task_running(rq, p); + task_rq_unlock(rq, &flags); + + if (likely(!running)) + break; + /* + * The switch count is incremented before the actual + * context switch. We thus wait for two switches to be + * sure at least one completed. + */ + if ((p->nvcsw - nvcsw) > 1) + break; + if ((p->nivcsw - nivcsw) > 1) + break; + + cpu_relax(); + } +} + +/* * wait_task_inactive - wait for a thread to unschedule. * * If @match_state is nonzero, it's the @p->state value just checked and -- cgit v1.1 From f711f6090a81cbd396b63de90f415d33f563af9b Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Tue, 14 Apr 2009 10:25:30 +0530 Subject: sched: Nominate idle load balancer from a semi-idle package. Currently the nomination of idle-load balancer is done by choosing the first idle cpu in the nohz.cpu_mask. This may not be power-efficient, since such an idle cpu could come from a completely idle core/package thereby preventing the whole core/package from being in a low-power state. For eg, consider a quad-core dual package system. The cpu numbering need not be sequential and can something like [0, 2, 4, 6] and [1, 3, 5, 7]. With sched_mc/smt_power_savings and the power-aware IRQ balance, we try to keep as fewer Packages/Cores active. But the current idle load balancer logic goes against this by choosing the first_cpu in the nohz.cpu_mask and not taking the system topology into consideration. Improve the algorithm to nominate the idle load balancer from a semi idle cores/packages thereby increasing the probability of the cores/packages being in deeper sleep states for longer duration. The algorithm is activated only when sched_mc/smt_power_savings != 0. Signed-off-by: Gautham R Shenoy Acked-by: Peter Zijlstra LKML-Reference: <20090414045530.7645.12175.stgit@sofia.in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 127 +++++++++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 118 insertions(+), 9 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 5724508..b0fefa3 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4240,10 +4240,126 @@ static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) static struct { atomic_t load_balancer; cpumask_var_t cpu_mask; + cpumask_var_t ilb_grp_nohz_mask; } nohz ____cacheline_aligned = { .load_balancer = ATOMIC_INIT(-1), }; +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * lowest_flag_domain - Return lowest sched_domain containing flag. + * @cpu: The cpu whose lowest level of sched domain is to + * be returned. + * @flag: The flag to check for the lowest sched_domain + * for the given cpu. + * + * Returns the lowest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) + if (sd && (sd->flags & flag)) + break; + + return sd; +} + +/** + * for_each_flag_domain - Iterates over sched_domains containing the flag. + * @cpu: The cpu whose domains we're iterating over. + * @sd: variable holding the value of the power_savings_sd + * for cpu. + * @flag: The flag to filter the sched_domains to be iterated. + * + * Iterates over all the scheduler domains for a given cpu that has the 'flag' + * set, starting from the lowest sched_domain to the highest. + */ +#define for_each_flag_domain(cpu, sd, flag) \ + for (sd = lowest_flag_domain(cpu, flag); \ + (sd && (sd->flags & flag)); sd = sd->parent) + +/** + * is_semi_idle_group - Checks if the given sched_group is semi-idle. + * @ilb_group: group to be checked for semi-idleness + * + * Returns: 1 if the group is semi-idle. 0 otherwise. + * + * We define a sched_group to be semi idle if it has atleast one idle-CPU + * and atleast one non-idle CPU. This helper function checks if the given + * sched_group is semi-idle or not. + */ +static inline int is_semi_idle_group(struct sched_group *ilb_group) +{ + cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, + sched_group_cpus(ilb_group)); + + /* + * A sched_group is semi-idle when it has atleast one busy cpu + * and atleast one idle cpu. + */ + if (cpumask_empty(nohz.ilb_grp_nohz_mask)) + return 0; + + if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) + return 0; + + return 1; +} +/** + * find_new_ilb - Finds the optimum idle load balancer for nomination. + * @cpu: The cpu which is nominating a new idle_load_balancer. + * + * Returns: Returns the id of the idle load balancer if it exists, + * Else, returns >= nr_cpu_ids. + * + * This algorithm picks the idle load balancer such that it belongs to a + * semi-idle powersavings sched_domain. The idea is to try and avoid + * completely idle packages/cores just for the purpose of idle load balancing + * when there are other idle cpu's which are better suited for that job. + */ +static int find_new_ilb(int cpu) +{ + struct sched_domain *sd; + struct sched_group *ilb_group; + + /* + * Have idle load balancer selection from semi-idle packages only + * when power-aware load balancing is enabled + */ + if (!(sched_smt_power_savings || sched_mc_power_savings)) + goto out_done; + + /* + * Optimize for the case when we have no idle CPUs or only one + * idle CPU. Don't walk the sched_domain hierarchy in such cases + */ + if (cpumask_weight(nohz.cpu_mask) < 2) + goto out_done; + + for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { + ilb_group = sd->groups; + + do { + if (is_semi_idle_group(ilb_group)) + return cpumask_first(nohz.ilb_grp_nohz_mask); + + ilb_group = ilb_group->next; + + } while (ilb_group != sd->groups); + } + +out_done: + return cpumask_first(nohz.cpu_mask); +} +#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ +static inline int find_new_ilb(int call_cpu) +{ + return first_cpu(nohz.cpu_mask); +} +#endif + /* * This routine will try to nominate the ilb (idle load balancing) * owner among the cpus whose ticks are stopped. ilb owner will do the idle @@ -4468,15 +4584,7 @@ static inline void trigger_load_balance(struct rq *rq, int cpu) } if (atomic_read(&nohz.load_balancer) == -1) { - /* - * simple selection for now: Nominate the - * first cpu in the nohz list to be the next - * ilb owner. - * - * TBD: Traverse the sched domains and nominate - * the nearest cpu in the nohz.cpu_mask. - */ - int ilb = cpumask_first(nohz.cpu_mask); + int ilb = find_new_ilb(cpu); if (ilb < nr_cpu_ids) resched_cpu(ilb); @@ -9051,6 +9159,7 @@ void __init sched_init(void) #ifdef CONFIG_SMP #ifdef CONFIG_NO_HZ alloc_bootmem_cpumask_var(&nohz.cpu_mask); + alloc_bootmem_cpumask_var(&nohz.ilb_grp_nohz_mask); #endif alloc_bootmem_cpumask_var(&cpu_isolated_map); #endif /* SMP */ -- cgit v1.1 From e790fb0ba64bfec158e1219d899cb588275d12ab Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Tue, 14 Apr 2009 10:25:35 +0530 Subject: sched: Nominate a power-efficient ilb in select_nohz_balancer() The CPU that first goes idle becomes the idle-load-balancer and remains that until either it picks up a task or till all the CPUs of the system goes idle. Optimize this further to allow it to relinquish it's post once all it's siblings in the power-aware sched_domain go idle, thereby allowing the whole package-core to go idle. While relinquising the post, nominate another an idle-load balancer from a semi-idle core/package. Signed-off-by: Gautham R Shenoy Acked-by: Peter Zijlstra LKML-Reference: <20090414045535.7645.31641.stgit@sofia.in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index b0fefa3..36d213b 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4414,8 +4414,24 @@ int select_nohz_load_balancer(int stop_tick) /* make me the ilb owner */ if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) return 1; - } else if (atomic_read(&nohz.load_balancer) == cpu) + } else if (atomic_read(&nohz.load_balancer) == cpu) { + int new_ilb; + + if (!(sched_smt_power_savings || + sched_mc_power_savings)) + return 1; + /* + * Check to see if there is a more power-efficient + * ilb. + */ + new_ilb = find_new_ilb(cpu); + if (new_ilb < nr_cpu_ids && new_ilb != cpu) { + atomic_set(&nohz.load_balancer, -1); + resched_cpu(new_ilb); + return 0; + } return 1; + } } else { if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) return 0; -- cgit v1.1 From 78ddb08feb7d4fbe3c0a9931804c51ee58be4023 Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Tue, 14 Apr 2009 16:53:05 +0200 Subject: wait: don't use __wake_up_common() '777c6c5 wait: prevent exclusive waiter starvation' made __wake_up_common() global to be used from abort_exclusive_wait(). It was needed to do a wake-up with the waitqueue lock held while passing down a key to the wake-up function. Since '4ede816 epoll keyed wakeups: add __wake_up_locked_key() and __wake_up_sync_key()' there is an appropriate wrapper for this case: __wake_up_locked_key(). Use it here and make __wake_up_common() private to the scheduler again. Signed-off-by: Johannes Weiner Cc: Andrew Morton Cc: Peter Zijlstra LKML-Reference: <1239720785-19661-1-git-send-email-hannes@cmpxchg.org> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 36d213b..92b4b56 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5345,7 +5345,7 @@ EXPORT_SYMBOL(default_wake_function); * started to run but is not in state TASK_RUNNING. try_to_wake_up() returns * zero in this (rare) case, and we handle it by continuing to scan the queue. */ -void __wake_up_common(wait_queue_head_t *q, unsigned int mode, +static void __wake_up_common(wait_queue_head_t *q, unsigned int mode, int nr_exclusive, int sync, void *key) { wait_queue_t *curr, *next; -- cgit v1.1 From a8d154b009168337494fbf345671bab74d3e4b8b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 10 Apr 2009 09:36:00 -0400 Subject: tracing: create automated trace defines This patch lowers the number of places a developer must modify to add new tracepoints. The current method to add a new tracepoint into an existing system is to write the trace point macro in the trace header with one of the macros TRACE_EVENT, TRACE_FORMAT or DECLARE_TRACE, then they must add the same named item into the C file with the macro DEFINE_TRACE(name) and then add the trace point. This change cuts out the needing to add the DEFINE_TRACE(name). Every file that uses the tracepoint must still include the trace/.h file, but the one C file must also add a define before the including of that file. #define CREATE_TRACE_POINTS #include This will cause the trace/mytrace.h file to also produce the C code necessary to implement the trace point. Note, if more than one trace/.h is used to create the C code it is best to list them all together. #define CREATE_TRACE_POINTS #include #include #include Thanks to Mathieu Desnoyers and Christoph Hellwig for coming up with the cleaner solution of the define above the includes over my first design to have the C code include a "special" header. This patch converts sched, irq and lockdep and skb to use this new method. Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Neil Horman Cc: Zhao Lei Cc: Eduard - Gabriel Munteanu Cc: Pekka Enberg Signed-off-by: Steven Rostedt --- kernel/sched.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 5724508..e6d4518 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -72,13 +72,15 @@ #include #include #include -#include #include #include #include "sched_cpupri.h" +#define CREATE_TRACE_POINTS +#include + /* * Convert user-nice values [ -20 ... 0 ... 19 ] * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ], @@ -118,12 +120,6 @@ */ #define RUNTIME_INF ((u64)~0ULL) -DEFINE_TRACE(sched_wait_task); -DEFINE_TRACE(sched_wakeup); -DEFINE_TRACE(sched_wakeup_new); -DEFINE_TRACE(sched_switch); -DEFINE_TRACE(sched_migrate_task); - #ifdef CONFIG_SMP static void double_rq_lock(struct rq *rq1, struct rq *rq2); -- cgit v1.1 From ad8d75fff811a6a230f7f43b05a6483099349533 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 14 Apr 2009 19:39:12 -0400 Subject: tracing/events: move trace point headers into include/trace/events Impact: clean up Create a sub directory in include/trace called events to keep the trace point headers in their own separate directory. Only headers that declare trace points should be defined in this directory. Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Neil Horman Cc: Zhao Lei Cc: Eduard - Gabriel Munteanu Cc: Pekka Enberg Signed-off-by: Steven Rostedt --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index e6d4518..9f7ffd0 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -79,7 +79,7 @@ #include "sched_cpupri.h" #define CREATE_TRACE_POINTS -#include +#include /* * Convert user-nice values [ -20 ... 0 ... 19 ] -- cgit v1.1 From 13318a7186d8e0ae08c996ea4111a945e7789772 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 15 Apr 2009 09:59:10 +0800 Subject: sched: use group_first_cpu() instead of cpumask_first(sched_group_cpus()) Impact: cleanup This patch changes cpumask_first(sched_group_cpus()) to group_first_cpu() for maintainability. Signed-off-by: Miao Xie Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- kernel/sched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 92b4b56..7601cee 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7995,7 +7995,7 @@ static void init_numa_sched_groups_power(struct sched_group *group_head) struct sched_domain *sd; sd = &per_cpu(phys_domains, j).sd; - if (j != cpumask_first(sched_group_cpus(sd->groups))) { + if (j != group_first_cpu(sd->groups)) { /* * Only add "power" once for each * physical package. @@ -8073,7 +8073,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) WARN_ON(!sd || !sd->groups); - if (cpu != cpumask_first(sched_group_cpus(sd->groups))) + if (cpu != group_first_cpu(sd->groups)) return; child = sd->child; -- cgit v1.1 From ff743345bf7685a207868048a70e23164c4785e5 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 13 Mar 2009 12:21:26 +0100 Subject: sched: remove extra call overhead for schedule() Lai Jiangshan's patch reminded me that I promised Nick to remove that extra call overhead in schedule(). Signed-off-by: Peter Zijlstra LKML-Reference: <20090313112300.927414207@chello.nl> Signed-off-by: Ingo Molnar --- kernel/sched.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 7601cee..797f6fd 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5131,13 +5131,15 @@ pick_next_task(struct rq *rq) /* * schedule() is the main scheduler function. */ -asmlinkage void __sched __schedule(void) +asmlinkage void __sched schedule(void) { struct task_struct *prev, *next; unsigned long *switch_count; struct rq *rq; int cpu; +need_resched: + preempt_disable(); cpu = smp_processor_id(); rq = cpu_rq(cpu); rcu_qsctr_inc(cpu); @@ -5194,15 +5196,9 @@ need_resched_nonpreemptible: if (unlikely(reacquire_kernel_lock(current) < 0)) goto need_resched_nonpreemptible; -} -asmlinkage void __sched schedule(void) -{ -need_resched: - preempt_disable(); - __schedule(); preempt_enable_no_resched(); - if (unlikely(test_thread_flag(TIF_NEED_RESCHED))) + if (need_resched()) goto need_resched; } EXPORT_SYMBOL(schedule); -- cgit v1.1 From 6e29ec5701e9d44fa02b96c1c5c45f7516182b65 Mon Sep 17 00:00:00 2001 From: Gautham R Shenoy Date: Tue, 21 Apr 2009 08:40:49 +0530 Subject: sched: Replace first_cpu() with cpumask_first() in ILB nomination code Stephen Rothwell reported this build warning: > kernel/sched.c: In function 'find_new_ilb': > kernel/sched.c:4355: warning: passing argument 1 of '__first_cpu' from incompatible pointer type > > Possibly caused by commit f711f6090a81cbd396b63de90f415d33f563af9b > ("sched: Nominate idle load balancer from a semi-idle package") from > the sched tree. Should this call to first_cpu be cpumask_first? For !(CONFIG_SCHED_MC || CONFIG_SCHED_SMT), find_new_ilb() nominates the Idle load balancer as the first cpu from the nohz.cpu_mask. This code uses the older API first_cpu(). Replace it with cpumask_first(), which is the correct API here. [ Impact: cleanup, address build warning ] Reported-by: Stephen Rothwell Signed-off-by: Gautham R Shenoy Cc: Rusty Russell LKML-Reference: <20090421031049.GA4140@in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 797f6fd..54d67b9 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4356,7 +4356,7 @@ out_done: #else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ static inline int find_new_ilb(int call_cpu) { - return first_cpu(nohz.cpu_mask); + return cpumask_first(nohz.cpu_mask); } #endif -- cgit v1.1 From 50fa610a3b6ba7cf91d7a92229177dfaff2b81a1 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 28 Apr 2009 15:01:38 +0100 Subject: sched: Document memory barriers implied by sleep/wake-up primitives Add a section to the memory barriers document to note the implied memory barriers of sleep primitives (set_current_state() and wrappers) and wake-up primitives (wake_up() and co.). Also extend the in-code comments on the wake_up() functions to note these implied barriers. [ Impact: add documentation ] Signed-off-by: David Howells Cc: Oleg Nesterov Cc: Linus Torvalds Cc: Andrew Morton LKML-Reference: <20090428140138.1192.94723.stgit@warthog.procyon.org.uk> Signed-off-by: Ingo Molnar --- kernel/sched.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index b902e58..fd0c2ce 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2458,6 +2458,17 @@ out: return success; } +/** + * wake_up_process - Wake up a specific process + * @p: The process to be woken up. + * + * Attempt to wake up the nominated process and move it to the set of runnable + * processes. Returns 1 if the process was woken up, 0 if it was already + * running. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. + */ int wake_up_process(struct task_struct *p) { return try_to_wake_up(p, TASK_ALL, 0); @@ -5241,6 +5252,9 @@ void __wake_up_common(wait_queue_head_t *q, unsigned int mode, * @mode: which threads * @nr_exclusive: how many wake-one or wake-many threads to wake up * @key: is directly passed to the wakeup function + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void __wake_up(wait_queue_head_t *q, unsigned int mode, int nr_exclusive, void *key) @@ -5279,6 +5293,9 @@ void __wake_up_locked_key(wait_queue_head_t *q, unsigned int mode, void *key) * with each other. This can prevent needless bouncing between CPUs. * * On UP it can prevent extra preemption. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void __wake_up_sync_key(wait_queue_head_t *q, unsigned int mode, int nr_exclusive, void *key) @@ -5315,6 +5332,9 @@ EXPORT_SYMBOL_GPL(__wake_up_sync); /* For internal use only */ * awakened in the same order in which they were queued. * * See also complete_all(), wait_for_completion() and related routines. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void complete(struct completion *x) { @@ -5332,6 +5352,9 @@ EXPORT_SYMBOL(complete); * @x: holds the state of this particular completion * * This will wake up all threads waiting on this particular completion event. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void complete_all(struct completion *x) { -- cgit v1.1 From 60aa605dfce2976e54fa76e805ab0f221372d4d9 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 5 May 2009 17:50:21 +0200 Subject: sched: rt: document the risk of small values in the bandwidth settings Thomas noted that we should disallow sysctl_sched_rt_runtime == 0 for (!RT_GROUP) since the root group always has some RT tasks in it. Further, update the documentation to inspire clue. [ Impact: exclude corner-case sysctl_sched_rt_runtime value ] Reported-by: Thomas Gleixner Signed-off-by: Peter Zijlstra LKML-Reference: <20090505155436.863098054@chello.nl> Signed-off-by: Ingo Molnar --- kernel/sched.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 54d67b9..2a43a58 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -9917,6 +9917,13 @@ static int sched_rt_global_constraints(void) if (sysctl_sched_rt_period <= 0) return -EINVAL; + /* + * There's always some RT tasks in the root group + * -- migration, kstopmachine etc.. + */ + if (sysctl_sched_rt_runtime == 0) + return -EBUSY; + spin_lock_irqsave(&def_rt_bandwidth.rt_runtime_lock, flags); for_each_possible_cpu(i) { struct rt_rq *rt_rq = &cpu_rq(i)->rt; -- cgit v1.1 From de1d7286060430e79a1d50ad6e5fee8fe863c5f6 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Tue, 5 May 2009 16:49:59 +0800 Subject: tracepoint: trace_sched_migrate_task(): remove parameter The orig_cpu parameter in trace_sched_migrate_task() is not necessary, it can be got by using task_cpu(p) in the probe. [ Impact: micro-optimization ] Signed-off-by: Mathieu Desnoyers [ modified from Mathieu's patch. The original patch is at: http://marc.info/?l=linux-kernel&m=123791201716239&w=2 ] Signed-off-by: Xiao Guangrong Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org Cc: Li Zefan Cc: zhaolei@cn.fujitsu.com Cc: laijs@cn.fujitsu.com LKML-Reference: <49FFFDB7.1050402@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 9f7ffd0..9cdedbd 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -1954,7 +1954,7 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu) clock_offset = old_rq->clock - new_rq->clock; - trace_sched_migrate_task(p, task_cpu(p), new_cpu); + trace_sched_migrate_task(p, new_cpu); #ifdef CONFIG_SCHEDSTATS if (p->se.wait_start) -- cgit v1.1 From aa47b7e0f89b9998dad4d1667447e8cb7703ff4e Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 4 May 2009 01:38:05 -0700 Subject: sched: emit thread info flags with stack trace When a thread is oom killed and fails to exit, it's helpful to know which threads have access to memory reserves if the machine livelocks. This is done by testing for the TIF_MEMDIE thread info flag and should be displayed alongside stack traces to identify tasks that have access to such reserves but are still stuck allocating pages, for instance. It would probably be helpful in other cases as well, so all thread info flags are emitted when showing a task. ( v2: fix warning reported by Stephen Rothwell ) [ Impact: extend debug printout info ] Signed-off-by: David Rientjes Cc: Peter Zijlstra Cc: Stephen Rothwell LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 2a43a58..5aa63f5 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -6610,8 +6610,9 @@ void sched_show_task(struct task_struct *p) #ifdef CONFIG_DEBUG_STACK_USAGE free = stack_not_used(p); #endif - printk(KERN_CONT "%5lu %5d %6d\n", free, - task_pid_nr(p), task_pid_nr(p->real_parent)); + printk(KERN_CONT "%5lu %5d %6d 0x%08lx\n", free, + task_pid_nr(p), task_pid_nr(p->real_parent), + (unsigned long)task_thread_info(p)->flags); show_stack(p, NULL); } -- cgit v1.1 From dce48a84adf1806676319f6f480e30a6daa012f9 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 11 Apr 2009 10:43:41 +0200 Subject: sched, timers: move calc_load() to scheduler Dimitri Sivanich noticed that xtime_lock is held write locked across calc_load() which iterates over all online CPUs. That can cause long latencies for xtime_lock readers on large SMP systems. The load average calculation is an rough estimate anyway so there is no real need to protect the readers vs. the update. It's not a problem when the avenrun array is updated while a reader copies the values. Instead of iterating over all online CPUs let the scheduler_tick code update the number of active tasks shortly before the avenrun update happens. The avenrun update itself is handled by the CPU which calls do_timer(). [ Impact: reduce xtime_lock write locked section ] Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra --- kernel/sched.c | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 74 insertions(+), 10 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 8908d19..f4eb881 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -630,6 +630,10 @@ struct rq { struct list_head migration_queue; #endif + /* calc_load related fields */ + unsigned long calc_load_update; + long calc_load_active; + #ifdef CONFIG_SCHED_HRTICK #ifdef CONFIG_SMP int hrtick_csd_pending; @@ -1728,6 +1732,8 @@ static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) } #endif +static void calc_load_account_active(struct rq *this_rq); + #include "sched_stats.h" #include "sched_idletask.c" #include "sched_fair.c" @@ -2856,19 +2862,57 @@ unsigned long nr_iowait(void) return sum; } -unsigned long nr_active(void) +/* Variables and functions for calc_load */ +static atomic_long_t calc_load_tasks; +static unsigned long calc_load_update; +unsigned long avenrun[3]; +EXPORT_SYMBOL(avenrun); + +static unsigned long +calc_load(unsigned long load, unsigned long exp, unsigned long active) { - unsigned long i, running = 0, uninterruptible = 0; + load *= exp; + load += active * (FIXED_1 - exp); + return load >> FSHIFT; +} - for_each_online_cpu(i) { - running += cpu_rq(i)->nr_running; - uninterruptible += cpu_rq(i)->nr_uninterruptible; - } +/* + * calc_load - update the avenrun load estimates 10 ticks after the + * CPUs have updated calc_load_tasks. + */ +void calc_global_load(void) +{ + unsigned long upd = calc_load_update + 10; + long active; + + if (time_before(jiffies, upd)) + return; - if (unlikely((long)uninterruptible < 0)) - uninterruptible = 0; + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; - return running + uninterruptible; + avenrun[0] = calc_load(avenrun[0], EXP_1, active); + avenrun[1] = calc_load(avenrun[1], EXP_5, active); + avenrun[2] = calc_load(avenrun[2], EXP_15, active); + + calc_load_update += LOAD_FREQ; +} + +/* + * Either called from update_cpu_load() or from a cpu going idle + */ +static void calc_load_account_active(struct rq *this_rq) +{ + long nr_active, delta; + + nr_active = this_rq->nr_running; + nr_active += (long) this_rq->nr_uninterruptible; + + if (nr_active != this_rq->calc_load_active) { + delta = nr_active - this_rq->calc_load_active; + this_rq->calc_load_active = nr_active; + atomic_long_add(delta, &calc_load_tasks); + } } /* @@ -2899,6 +2943,11 @@ static void update_cpu_load(struct rq *this_rq) new_load += scale-1; this_rq->cpu_load[i] = (old_load*(scale-1) + new_load) >> i; } + + if (time_after_eq(jiffies, this_rq->calc_load_update)) { + this_rq->calc_load_update += LOAD_FREQ; + calc_load_account_active(this_rq); + } } #ifdef CONFIG_SMP @@ -7091,6 +7140,14 @@ static void migrate_dead_tasks(unsigned int dead_cpu) } } + +/* + * remove the tasks which were accounted by rq from calc_load_tasks. + */ +static void calc_global_load_remove(struct rq *rq) +{ + atomic_long_sub(rq->calc_load_active, &calc_load_tasks); +} #endif /* CONFIG_HOTPLUG_CPU */ #if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL) @@ -7325,6 +7382,8 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) /* Update our root-domain */ rq = cpu_rq(cpu); spin_lock_irqsave(&rq->lock, flags); + rq->calc_load_update = calc_load_update; + rq->calc_load_active = 0; if (rq->rd) { BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span)); @@ -7364,7 +7423,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) cpuset_unlock(); migrate_nr_uninterruptible(rq); BUG_ON(rq->nr_running != 0); - + calc_global_load_remove(rq); /* * No need to migrate the tasks: it was best-effort if * they didn't take sched_hotcpu_mutex. Just wake up @@ -9059,6 +9118,8 @@ void __init sched_init(void) rq = cpu_rq(i); spin_lock_init(&rq->lock); rq->nr_running = 0; + rq->calc_load_active = 0; + rq->calc_load_update = jiffies + LOAD_FREQ; init_cfs_rq(&rq->cfs, rq); init_rt_rq(&rq->rt, rq); #ifdef CONFIG_FAIR_GROUP_SCHED @@ -9166,6 +9227,9 @@ void __init sched_init(void) * when this runqueue becomes "idle". */ init_idle(current, smp_processor_id()); + + calc_load_update = jiffies + LOAD_FREQ; + /* * During early bootup we pretend to be a normal task: */ -- cgit v1.1 From 2d02494f5a90f2e4b3c4c6acc85ec94674cdc431 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 2 May 2009 20:08:52 +0200 Subject: sched, timers: cleanup avenrun users avenrun is an rough estimate so we don't have to worry about consistency of the three avenrun values. Remove the xtime lock dependency and provide a function to scale the values. Cleanup the users. [ Impact: cleanup ] Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra --- kernel/sched.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index f4eb881..497c09b 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2868,6 +2868,21 @@ static unsigned long calc_load_update; unsigned long avenrun[3]; EXPORT_SYMBOL(avenrun); +/** + * get_avenrun - get the load average array + * @loads: pointer to dest load array + * @offset: offset to add + * @shift: shift count to shift the result left + * + * These values are estimates at best, so no need for locking. + */ +void get_avenrun(unsigned long *loads, unsigned long offset, int shift) +{ + loads[0] = (avenrun[0] + offset) << shift; + loads[1] = (avenrun[1] + offset) << shift; + loads[2] = (avenrun[2] + offset) << shift; +} + static unsigned long calc_load(unsigned long load, unsigned long exp, unsigned long active) { -- cgit v1.1 From 4200efd9acda4accf24640f1e77d24fdcdb524df Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 19 May 2009 09:22:19 +0200 Subject: sched: properly define the sched_group::cpumask and sched_domain::span fields Properly document the variable-size structure tricks we are doing wrt. struct sched_group and sched_domain, and use the field[0] GCC extension instead of defining a vla array. Dont use unions for this, as pointed out by Linus. [ Impact: cleanup, un-confuse Sparse and LLVM ] Reported-by: Jeff Garzik Acked-by: Linus Torvalds LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/sched.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'kernel/sched.c') diff --git a/kernel/sched.c b/kernel/sched.c index 497c09b..228acae 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -7948,8 +7948,9 @@ int sched_smt_power_savings = 0, sched_mc_power_savings = 0; /* * The cpus mask in sched_group and sched_domain hangs off the end. - * FIXME: use cpumask_var_t or dynamic percpu alloc to avoid wasting space - * for nr_cpu_ids < CONFIG_NR_CPUS. + * + * ( See the the comments in include/linux/sched.h:struct sched_group + * and struct sched_domain. ) */ struct static_sched_group { struct sched_group sg; -- cgit v1.1