diff options
Diffstat (limited to 'kernel')
71 files changed, 4117 insertions, 3879 deletions
diff --git a/kernel/Kconfig.locks b/kernel/Kconfig.locks index 44511d100eaa..d2b32ac27a39 100644 --- a/kernel/Kconfig.locks +++ b/kernel/Kconfig.locks @@ -138,7 +138,7 @@ config INLINE_SPIN_UNLOCK_BH config INLINE_SPIN_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_SPIN_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_SPIN_UNLOCK_IRQ config INLINE_SPIN_UNLOCK_IRQRESTORE def_bool y @@ -175,7 +175,7 @@ config INLINE_READ_UNLOCK_BH config INLINE_READ_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_READ_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_READ_UNLOCK_IRQ config INLINE_READ_UNLOCK_IRQRESTORE def_bool y @@ -212,7 +212,7 @@ config INLINE_WRITE_UNLOCK_BH config INLINE_WRITE_UNLOCK_IRQ def_bool y - depends on !PREEMPT || ARCH_INLINE_WRITE_UNLOCK_BH + depends on !PREEMPT || ARCH_INLINE_WRITE_UNLOCK_IRQ config INLINE_WRITE_UNLOCK_IRQRESTORE def_bool y diff --git a/kernel/audit.c b/kernel/audit.c index 21c7fa615bd3..91e53d04b6a9 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -1056,7 +1056,7 @@ static inline void audit_get_stamp(struct audit_context *ctx, static void wait_for_auditd(unsigned long sleep_time) { DECLARE_WAITQUEUE(wait, current); - set_current_state(TASK_INTERRUPTIBLE); + set_current_state(TASK_UNINTERRUPTIBLE); add_wait_queue(&audit_backlog_wait, &wait); if (audit_backlog_limit && diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index a291aa23fb3f..43c307dc9453 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -658,6 +658,7 @@ int audit_add_tree_rule(struct audit_krule *rule) struct vfsmount *mnt; int err; + rule->tree = NULL; list_for_each_entry(tree, &tree_list, list) { if (!strcmp(seed->pathname, tree->pathname)) { put_tree(seed); diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 2a9926275f80..e5583d10a325 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -63,9 +63,6 @@ #include <linux/atomic.h> -/* css deactivation bias, makes css->refcnt negative to deny new trygets */ -#define CSS_DEACT_BIAS INT_MIN - /* * cgroup_mutex is the master lock. Any modification to cgroup or its * hierarchy must be performed while holding it. @@ -99,16 +96,19 @@ static DEFINE_MUTEX(cgroup_root_mutex); */ #define SUBSYS(_x) [_x ## _subsys_id] = &_x ## _subsys, #define IS_SUBSYS_ENABLED(option) IS_BUILTIN(option) -static struct cgroup_subsys *subsys[CGROUP_SUBSYS_COUNT] = { +static struct cgroup_subsys *cgroup_subsys[CGROUP_SUBSYS_COUNT] = { #include <linux/cgroup_subsys.h> }; /* - * The "rootnode" hierarchy is the "dummy hierarchy", reserved for the - * subsystems that are otherwise unattached - it never has more than a - * single cgroup, and all tasks are part of that cgroup. + * The dummy hierarchy, reserved for the subsystems that are otherwise + * unattached - it never has more than a single cgroup, and all tasks are + * part of that cgroup. */ -static struct cgroupfs_root rootnode; +static struct cgroupfs_root cgroup_dummy_root; + +/* dummy_top is a shorthand for the dummy hierarchy's top cgroup */ +static struct cgroup * const cgroup_dummy_top = &cgroup_dummy_root.top_cgroup; /* * cgroupfs file entry, pointed to from leaf dentry->d_fsdata. @@ -186,18 +186,28 @@ struct cgroup_event { /* The list of hierarchy roots */ -static LIST_HEAD(roots); -static int root_count; +static LIST_HEAD(cgroup_roots); +static int cgroup_root_count; -static DEFINE_IDA(hierarchy_ida); -static int next_hierarchy_id; -static DEFINE_SPINLOCK(hierarchy_id_lock); - -/* dummytop is a shorthand for the dummy hierarchy's top cgroup */ -#define dummytop (&rootnode.top_cgroup) +/* + * Hierarchy ID allocation and mapping. It follows the same exclusion + * rules as other root ops - both cgroup_mutex and cgroup_root_mutex for + * writes, either for reads. + */ +static DEFINE_IDR(cgroup_hierarchy_idr); static struct cgroup_name root_cgroup_name = { .name = "/" }; +/* + * Assign a monotonically increasing serial number to cgroups. It + * guarantees cgroups with bigger numbers are newer than those with smaller + * numbers. Also, as cgroups are always appended to the parent's + * ->children list, it guarantees that sibling cgroups are always sorted in + * the ascending serial number order on the list. Protected by + * cgroup_mutex. + */ +static u64 cgroup_serial_nr_next = 1; + /* This flag indicates whether tasks in the fork and exit paths should * check for fork/exit handlers to call. This avoids us having to do * extra work in the fork/exit path if none of the subsystems need to @@ -205,27 +215,15 @@ static struct cgroup_name root_cgroup_name = { .name = "/" }; */ static int need_forkexit_callback __read_mostly; +static void cgroup_offline_fn(struct work_struct *work); static int cgroup_destroy_locked(struct cgroup *cgrp); static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, struct cftype cfts[], bool is_add); -static int css_unbias_refcnt(int refcnt) -{ - return refcnt >= 0 ? refcnt : refcnt - CSS_DEACT_BIAS; -} - -/* the current nr of refs, always >= 0 whether @css is deactivated or not */ -static int css_refcnt(struct cgroup_subsys_state *css) -{ - int v = atomic_read(&css->refcnt); - - return css_unbias_refcnt(v); -} - /* convenient tests for these bits */ -inline int cgroup_is_removed(const struct cgroup *cgrp) +static inline bool cgroup_is_dead(const struct cgroup *cgrp) { - return test_bit(CGRP_REMOVED, &cgrp->flags); + return test_bit(CGRP_DEAD, &cgrp->flags); } /** @@ -261,16 +259,38 @@ static int notify_on_release(const struct cgroup *cgrp) return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); } -/* - * for_each_subsys() allows you to iterate on each subsystem attached to - * an active hierarchy +/** + * for_each_subsys - iterate all loaded cgroup subsystems + * @ss: the iteration cursor + * @i: the index of @ss, CGROUP_SUBSYS_COUNT after reaching the end + * + * Should be called under cgroup_mutex. */ -#define for_each_subsys(_root, _ss) \ -list_for_each_entry(_ss, &_root->subsys_list, sibling) +#define for_each_subsys(ss, i) \ + for ((i) = 0; (i) < CGROUP_SUBSYS_COUNT; (i)++) \ + if (({ lockdep_assert_held(&cgroup_mutex); \ + !((ss) = cgroup_subsys[i]); })) { } \ + else -/* for_each_active_root() allows you to iterate across the active hierarchies */ -#define for_each_active_root(_root) \ -list_for_each_entry(_root, &roots, root_list) +/** + * for_each_builtin_subsys - iterate all built-in cgroup subsystems + * @ss: the iteration cursor + * @i: the index of @ss, CGROUP_BUILTIN_SUBSYS_COUNT after reaching the end + * + * Bulit-in subsystems are always present and iteration itself doesn't + * require any synchronization. + */ +#define for_each_builtin_subsys(ss, i) \ + for ((i) = 0; (i) < CGROUP_BUILTIN_SUBSYS_COUNT && \ + (((ss) = cgroup_subsys[i]) || true); (i)++) + +/* iterate each subsystem attached to a hierarchy */ +#define for_each_root_subsys(root, ss) \ + list_for_each_entry((ss), &(root)->subsys_list, sibling) + +/* iterate across the active hierarchies */ +#define for_each_active_root(root) \ + list_for_each_entry((root), &cgroup_roots, root_list) static inline struct cgroup *__d_cgrp(struct dentry *dentry) { @@ -297,7 +317,7 @@ static inline struct cftype *__d_cft(struct dentry *dentry) static bool cgroup_lock_live_group(struct cgroup *cgrp) { mutex_lock(&cgroup_mutex); - if (cgroup_is_removed(cgrp)) { + if (cgroup_is_dead(cgrp)) { mutex_unlock(&cgroup_mutex); return false; } @@ -312,20 +332,24 @@ static void cgroup_release_agent(struct work_struct *work); static DECLARE_WORK(release_agent_work, cgroup_release_agent); static void check_for_release(struct cgroup *cgrp); -/* Link structure for associating css_set objects with cgroups */ -struct cg_cgroup_link { - /* - * List running through cg_cgroup_links associated with a - * cgroup, anchored on cgroup->css_sets - */ - struct list_head cgrp_link_list; - struct cgroup *cgrp; - /* - * List running through cg_cgroup_links pointing at a - * single css_set object, anchored on css_set->cg_links - */ - struct list_head cg_link_list; - struct css_set *cg; +/* + * A cgroup can be associated with multiple css_sets as different tasks may + * belong to different cgroups on different hierarchies. In the other + * direction, a css_set is naturally associated with multiple cgroups. + * This M:N relationship is represented by the following link structure + * which exists for each association and allows traversing the associations + * from both sides. + */ +struct cgrp_cset_link { + /* the cgroup and css_set this link associates */ + struct cgroup *cgrp; + struct css_set *cset; + + /* list of cgrp_cset_links anchored at cgrp->cset_links */ + struct list_head cset_link; + + /* list of cgrp_cset_links anchored at css_set->cgrp_links */ + struct list_head cgrp_link; }; /* The default css_set - used by init and its children prior to any @@ -336,7 +360,7 @@ struct cg_cgroup_link { */ static struct css_set init_css_set; -static struct cg_cgroup_link init_css_set_link; +static struct cgrp_cset_link init_cgrp_cset_link; static int cgroup_init_idr(struct cgroup_subsys *ss, struct cgroup_subsys_state *css); @@ -357,10 +381,11 @@ static DEFINE_HASHTABLE(css_set_table, CSS_SET_HASH_BITS); static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) { - int i; unsigned long key = 0UL; + struct cgroup_subsys *ss; + int i; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) + for_each_subsys(ss, i) key += (unsigned long)css[i]; key = (key >> 16) ^ key; @@ -373,90 +398,83 @@ static unsigned long css_set_hash(struct cgroup_subsys_state *css[]) * compiled into their kernel but not actually in use */ static int use_task_css_set_links __read_mostly; -static void __put_css_set(struct css_set *cg, int taskexit) +static void __put_css_set(struct css_set *cset, int taskexit) { - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; + struct cgrp_cset_link *link, *tmp_link; + /* * Ensure that the refcount doesn't hit zero while any readers * can see it. Similar to atomic_dec_and_lock(), but for an * rwlock */ - if (atomic_add_unless(&cg->refcount, -1, 1)) + if (atomic_add_unless(&cset->refcount, -1, 1)) return; write_lock(&css_set_lock); - if (!atomic_dec_and_test(&cg->refcount)) { + if (!atomic_dec_and_test(&cset->refcount)) { write_unlock(&css_set_lock); return; } /* This css_set is dead. unlink it and release cgroup refcounts */ - hash_del(&cg->hlist); + hash_del(&cset->hlist); css_set_count--; - list_for_each_entry_safe(link, saved_link, &cg->cg_links, - cg_link_list) { + list_for_each_entry_safe(link, tmp_link, &cset->cgrp_links, cgrp_link) { struct cgroup *cgrp = link->cgrp; - list_del(&link->cg_link_list); - list_del(&link->cgrp_link_list); - /* - * We may not be holding cgroup_mutex, and if cgrp->count is - * dropped to 0 the cgroup can be destroyed at any time, hence - * rcu_read_lock is used to keep it alive. - */ - rcu_read_lock(); - if (atomic_dec_and_test(&cgrp->count) && - notify_on_release(cgrp)) { + list_del(&link->cset_link); + list_del(&link->cgrp_link); + + /* @cgrp can't go away while we're holding css_set_lock */ + if (list_empty(&cgrp->cset_links) && notify_on_release(cgrp)) { if (taskexit) set_bit(CGRP_RELEASABLE, &cgrp->flags); check_for_release(cgrp); } - rcu_read_unlock(); kfree(link); } write_unlock(&css_set_lock); - kfree_rcu(cg, rcu_head); + kfree_rcu(cset, rcu_head); } /* * refcounted get/put for css_set objects */ -static inline void get_css_set(struct css_set *cg) +static inline void get_css_set(struct css_set *cset) { - atomic_inc(&cg->refcount); + atomic_inc(&cset->refcount); } -static inline void put_css_set(struct css_set *cg) +static inline void put_css_set(struct css_set *cset) { - __put_css_set(cg, 0); + __put_css_set(cset, 0); } -static inline void put_css_set_taskexit(struct css_set *cg) +static inline void put_css_set_taskexit(struct css_set *cset) { - __put_css_set(cg, 1); + __put_css_set(cset, 1); } -/* +/** * compare_css_sets - helper function for find_existing_css_set(). - * @cg: candidate css_set being tested - * @old_cg: existing css_set for a task + * @cset: candidate css_set being tested + * @old_cset: existing css_set for a task * @new_cgrp: cgroup that's being entered by the task * @template: desired set of css pointers in css_set (pre-calculated) * * Returns true if "cg" matches "old_cg" except for the hierarchy * which "new_cgrp" belongs to, for which it should match "new_cgrp". */ -static bool compare_css_sets(struct css_set *cg, - struct css_set *old_cg, +static bool compare_css_sets(struct css_set *cset, + struct css_set *old_cset, struct cgroup *new_cgrp, struct cgroup_subsys_state *template[]) { struct list_head *l1, *l2; - if (memcmp(template, cg->subsys, sizeof(cg->subsys))) { + if (memcmp(template, cset->subsys, sizeof(cset->subsys))) { /* Not all subsystems matched */ return false; } @@ -470,28 +488,28 @@ static bool compare_css_sets(struct css_set *cg, * candidates. */ - l1 = &cg->cg_links; - l2 = &old_cg->cg_links; + l1 = &cset->cgrp_links; + l2 = &old_cset->cgrp_links; while (1) { - struct cg_cgroup_link *cgl1, *cgl2; - struct cgroup *cg1, *cg2; + struct cgrp_cset_link *link1, *link2; + struct cgroup *cgrp1, *cgrp2; l1 = l1->next; l2 = l2->next; /* See if we reached the end - both lists are equal length. */ - if (l1 == &cg->cg_links) { - BUG_ON(l2 != &old_cg->cg_links); + if (l1 == &cset->cgrp_links) { + BUG_ON(l2 != &old_cset->cgrp_links); break; } else { - BUG_ON(l2 == &old_cg->cg_links); + BUG_ON(l2 == &old_cset->cgrp_links); } /* Locate the cgroups associated with these links. */ - cgl1 = list_entry(l1, struct cg_cgroup_link, cg_link_list); - cgl2 = list_entry(l2, struct cg_cgroup_link, cg_link_list); - cg1 = cgl1->cgrp; - cg2 = cgl2->cgrp; + link1 = list_entry(l1, struct cgrp_cset_link, cgrp_link); + link2 = list_entry(l2, struct cgrp_cset_link, cgrp_link); + cgrp1 = link1->cgrp; + cgrp2 = link2->cgrp; /* Hierarchies should be linked in the same order. */ - BUG_ON(cg1->root != cg2->root); + BUG_ON(cgrp1->root != cgrp2->root); /* * If this hierarchy is the hierarchy of the cgroup @@ -500,46 +518,39 @@ static bool compare_css_sets(struct css_set *cg, * hierarchy, then this css_set should point to the * same cgroup as the old css_set. */ - if (cg1->root == new_cgrp->root) { - if (cg1 != new_cgrp) + if (cgrp1->root == new_cgrp->root) { + if (cgrp1 != new_cgrp) return false; } else { - if (cg1 != cg2) + if (cgrp1 != cgrp2) return false; } } return true; } -/* - * find_existing_css_set() is a helper for - * find_css_set(), and checks to see whether an existing - * css_set is suitable. - * - * oldcg: the cgroup group that we're using before the cgroup - * transition - * - * cgrp: the cgroup that we're moving into - * - * template: location in which to build the desired set of subsystem - * state objects for the new cgroup group +/** + * find_existing_css_set - init css array and find the matching css_set + * @old_cset: the css_set that we're using before the cgroup transition + * @cgrp: the cgroup that we're moving into + * @template: out param for the new set of csses, should be clear on entry */ -static struct css_set *find_existing_css_set( - struct css_set *oldcg, - struct cgroup *cgrp, - struct cgroup_subsys_state *template[]) +static struct css_set *find_existing_css_set(struct css_set *old_cset, + struct cgroup *cgrp, + struct cgroup_subsys_state *template[]) { - int i; struct cgroupfs_root *root = cgrp->root; - struct css_set *cg; + struct cgroup_subsys *ss; + struct css_set *cset; unsigned long key; + int i; /* * Build the set of subsystem state objects that we want to see in the * new css_set. while subsystems can change globally, the entries here * won't change, so no need for locking. */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + for_each_subsys(ss, i) { if (root->subsys_mask & (1UL << i)) { /* Subsystem is in this hierarchy. So we want * the subsystem state from the new @@ -548,148 +559,152 @@ static struct css_set *find_existing_css_set( } else { /* Subsystem is not in this hierarchy, so we * don't want to change the subsystem state */ - template[i] = oldcg->subsys[i]; + template[i] = old_cset->subsys[i]; } } key = css_set_hash(template); - hash_for_each_possible(css_set_table, cg, hlist, key) { - if (!compare_css_sets(cg, oldcg, cgrp, template)) + hash_for_each_possible(css_set_table, cset, hlist, key) { + if (!compare_css_sets(cset, old_cset, cgrp, template)) continue; /* This css_set matches what we need */ - return cg; + return cset; } /* No existing cgroup group matched */ return NULL; } -static void free_cg_links(struct list_head *tmp) +static void free_cgrp_cset_links(struct list_head *links_to_free) { - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; + struct cgrp_cset_link *link, *tmp_link; - list_for_each_entry_safe(link, saved_link, tmp, cgrp_link_list) { - list_del(&link->cgrp_link_list); + list_for_each_entry_safe(link, tmp_link, links_to_free, cset_link) { + list_del(&link->cset_link); kfree(link); } } -/* - * allocate_cg_links() allocates "count" cg_cgroup_link structures - * and chains them on tmp through their cgrp_link_list fields. Returns 0 on - * success or a negative error +/** + * allocate_cgrp_cset_links - allocate cgrp_cset_links + * @count: the number of links to allocate + * @tmp_links: list_head the allocated links are put on + * + * Allocate @count cgrp_cset_link structures and chain them on @tmp_links + * through ->cset_link. Returns 0 on success or -errno. */ -static int allocate_cg_links(int count, struct list_head *tmp) +static int allocate_cgrp_cset_links(int count, struct list_head *tmp_links) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; int i; - INIT_LIST_HEAD(tmp); + + INIT_LIST_HEAD(tmp_links); + for (i = 0; i < count; i++) { - link = kmalloc(sizeof(*link), GFP_KERNEL); + link = kzalloc(sizeof(*link), GFP_KERNEL); if (!link) { - free_cg_links(tmp); + free_cgrp_cset_links(tmp_links); return -ENOMEM; } - list_add(&link->cgrp_link_list, tmp); + list_add(&link->cset_link, tmp_links); } return 0; } /** * link_css_set - a helper function to link a css_set to a cgroup - * @tmp_cg_links: cg_cgroup_link objects allocated by allocate_cg_links() - * @cg: the css_set to be linked + * @tmp_links: cgrp_cset_link objects allocated by allocate_cgrp_cset_links() + * @cset: the css_set to be linked * @cgrp: the destination cgroup */ -static void link_css_set(struct list_head *tmp_cg_links, - struct css_set *cg, struct cgroup *cgrp) +static void link_css_set(struct list_head *tmp_links, struct css_set *cset, + struct cgroup *cgrp) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; - BUG_ON(list_empty(tmp_cg_links)); - link = list_first_entry(tmp_cg_links, struct cg_cgroup_link, - cgrp_link_list); - link->cg = cg; + BUG_ON(list_empty(tmp_links)); + link = list_first_entry(tmp_links, struct cgrp_cset_link, cset_link); + link->cset = cset; link->cgrp = cgrp; - atomic_inc(&cgrp->count); - list_move(&link->cgrp_link_list, &cgrp->css_sets); + list_move(&link->cset_link, &cgrp->cset_links); /* * Always add links to the tail of the list so that the list * is sorted by order of hierarchy creation */ - list_add_tail(&link->cg_link_list, &cg->cg_links); + list_add_tail(&link->cgrp_link, &cset->cgrp_links); } -/* - * find_css_set() takes an existing cgroup group and a - * cgroup object, and returns a css_set object that's - * equivalent to the old group, but with the given cgroup - * substituted into the appropriate hierarchy. Must be called with - * cgroup_mutex held +/** + * find_css_set - return a new css_set with one cgroup updated + * @old_cset: the baseline css_set + * @cgrp: the cgroup to be updated + * + * Return a new css_set that's equivalent to @old_cset, but with @cgrp + * substituted into the appropriate hierarchy. */ -static struct css_set *find_css_set( - struct css_set *oldcg, struct cgroup *cgrp) +static struct css_set *find_css_set(struct css_set *old_cset, + struct cgroup *cgrp) { - struct css_set *res; - struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT]; - - struct list_head tmp_cg_links; - - struct cg_cgroup_link *link; + struct cgroup_subsys_state *template[CGROUP_SUBSYS_COUNT] = { }; + struct css_set *cset; + struct list_head tmp_links; + struct cgrp_cset_link *link; unsigned long key; + lockdep_assert_held(&cgroup_mutex); + /* First see if we already have a cgroup group that matches * the desired set */ read_lock(&css_set_lock); - res = find_existing_css_set(oldcg, cgrp, template); - if (res) - get_css_set(res); + cset = find_existing_css_set(old_cset, cgrp, template); + if (cset) + get_css_set(cset); read_unlock(&css_set_lock); - if (res) - return res; + if (cset) + return cset; - res = kmalloc(sizeof(*res), GFP_KERNEL); - if (!res) + cset = kzalloc(sizeof(*cset), GFP_KERNEL); + if (!cset) return NULL; - /* Allocate all the cg_cgroup_link objects that we'll need */ - if (allocate_cg_links(root_count, &tmp_cg_links) < 0) { - kfree(res); + /* Allocate all the cgrp_cset_link objects that we'll need */ + if (allocate_cgrp_cset_links(cgroup_root_count, &tmp_links) < 0) { + kfree(cset); return NULL; } - atomic_set(&res->refcount, 1); - INIT_LIST_HEAD(&res->cg_links); - INIT_LIST_HEAD(&res->tasks); - INIT_HLIST_NODE(&res->hlist); + atomic_set(&cset->refcount, 1); + INIT_LIST_HEAD(&cset->cgrp_links); + INIT_LIST_HEAD(&cset->tasks); + INIT_HLIST_NODE(&cset->hlist); /* Copy the set of subsystem state objects generated in * find_existing_css_set() */ - memcpy(res->subsys, template, sizeof(res->subsys)); + memcpy(cset->subsys, template, sizeof(cset->subsys)); write_lock(&css_set_lock); /* Add reference counts and links from the new css_set. */ - list_for_each_entry(link, &oldcg->cg_links, cg_link_list) { + list_for_each_entry(link, &old_cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; + if (c->root == cgrp->root) c = cgrp; - link_css_set(&tmp_cg_links, res, c); + link_css_set(&tmp_links, cset, c); } - BUG_ON(!list_empty(&tmp_cg_links)); + BUG_ON(!list_empty(&tmp_links)); css_set_count++; /* Add this cgroup group to the hash table */ - key = css_set_hash(res->subsys); - hash_add(css_set_table, &res->hlist, key); + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); write_unlock(&css_set_lock); - return res; + return cset; } /* @@ -699,7 +714,7 @@ static struct css_set *find_css_set( static struct cgroup *task_cgroup_from_root(struct task_struct *task, struct cgroupfs_root *root) { - struct css_set *css; + struct css_set *cset; struct cgroup *res = NULL; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -709,13 +724,15 @@ static struct cgroup *task_cgroup_from_root(struct task_struct *task, * task can't change groups, so the only thing that can happen * is that it exits and its css is set back to init_css_set. */ - css = task->cgroups; - if (css == &init_css_set) { + cset = task_css_set(task); + if (cset == &init_css_set) { res = &root->top_cgroup; } else { - struct cg_cgroup_link *link; - list_for_each_entry(link, &css->cg_links, cg_link_list) { + struct cgrp_cset_link *link; + + list_for_each_entry(link, &cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; + if (c->root == root) { res = c; break; @@ -828,14 +845,14 @@ static struct cgroup_name *cgroup_alloc_name(struct dentry *dentry) static void cgroup_free_fn(struct work_struct *work) { - struct cgroup *cgrp = container_of(work, struct cgroup, free_work); + struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); struct cgroup_subsys *ss; mutex_lock(&cgroup_mutex); /* * Release the subsystem state objects. */ - for_each_subsys(cgrp->root, ss) + for_each_root_subsys(cgrp->root, ss) ss->css_free(cgrp); cgrp->root->number_of_cgroups--; @@ -873,7 +890,8 @@ static void cgroup_free_rcu(struct rcu_head *head) { struct cgroup *cgrp = container_of(head, struct cgroup, rcu_head); - schedule_work(&cgrp->free_work); + INIT_WORK(&cgrp->destroy_work, cgroup_free_fn); + schedule_work(&cgrp->destroy_work); } static void cgroup_diput(struct dentry *dentry, struct inode *inode) @@ -882,7 +900,7 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode) if (S_ISDIR(inode->i_mode)) { struct cgroup *cgrp = dentry->d_fsdata; - BUG_ON(!(cgroup_is_removed(cgrp))); + BUG_ON(!(cgroup_is_dead(cgrp))); call_rcu(&cgrp->rcu_head, cgroup_free_rcu); } else { struct cfent *cfe = __d_cfe(dentry); @@ -950,7 +968,7 @@ static void cgroup_clear_directory(struct dentry *dir, bool base_files, struct cgroup *cgrp = __d_cgrp(dir); struct cgroup_subsys *ss; - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cftype_set *set; if (!test_bit(ss->subsys_id, &subsys_mask)) continue; @@ -988,30 +1006,23 @@ static void cgroup_d_remove_dir(struct dentry *dentry) * returns an error, no reference counts are touched. */ static int rebind_subsystems(struct cgroupfs_root *root, - unsigned long final_subsys_mask) + unsigned long added_mask, unsigned removed_mask) { - unsigned long added_mask, removed_mask; struct cgroup *cgrp = &root->top_cgroup; + struct cgroup_subsys *ss; int i; BUG_ON(!mutex_is_locked(&cgroup_mutex)); BUG_ON(!mutex_is_locked(&cgroup_root_mutex)); - removed_mask = root->actual_subsys_mask & ~final_subsys_mask; - added_mask = final_subsys_mask & ~root->actual_subsys_mask; /* Check that any added subsystems are currently free */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + for_each_subsys(ss, i) { unsigned long bit = 1UL << i; - struct cgroup_subsys *ss = subsys[i]; + if (!(bit & added_mask)) continue; - /* - * Nobody should tell us to do a subsys that doesn't exist: - * parse_cgroupfs_options should catch that case and refcounts - * ensure that subsystems won't disappear once selected. - */ - BUG_ON(ss == NULL); - if (ss->root != &rootnode) { + + if (ss->root != &cgroup_dummy_root) { /* Subsystem isn't free */ return -EBUSY; } @@ -1025,38 +1036,41 @@ static int rebind_subsystems(struct cgroupfs_root *root, return -EBUSY; /* Process each subsystem */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; + for_each_subsys(ss, i) { unsigned long bit = 1UL << i; + if (bit & added_mask) { /* We're binding this subsystem to this hierarchy */ - BUG_ON(ss == NULL); BUG_ON(cgrp->subsys[i]); - BUG_ON(!dummytop->subsys[i]); - BUG_ON(dummytop->subsys[i]->cgroup != dummytop); - cgrp->subsys[i] = dummytop->subsys[i]; + BUG_ON(!cgroup_dummy_top->subsys[i]); + BUG_ON(cgroup_dummy_top->subsys[i]->cgroup != cgroup_dummy_top); + + cgrp->subsys[i] = cgroup_dummy_top->subsys[i]; cgrp->subsys[i]->cgroup = cgrp; list_move(&ss->sibling, &root->subsys_list); ss->root = root; if (ss->bind) ss->bind(cgrp); + /* refcount was already taken, and we're keeping it */ + root->subsys_mask |= bit; } else if (bit & removed_mask) { /* We're removing this subsystem */ - BUG_ON(ss == NULL); - BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); + BUG_ON(cgrp->subsys[i] != cgroup_dummy_top->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); + if (ss->bind) - ss->bind(dummytop); - dummytop->subsys[i]->cgroup = dummytop; + ss->bind(cgroup_dummy_top); + cgroup_dummy_top->subsys[i]->cgroup = cgroup_dummy_top; cgrp->subsys[i] = NULL; - subsys[i]->root = &rootnode; - list_move(&ss->sibling, &rootnode.subsys_list); + cgroup_subsys[i]->root = &cgroup_dummy_root; + list_move(&ss->sibling, &cgroup_dummy_root.subsys_list); + /* subsystem is now free - drop reference on module */ module_put(ss->module); - } else if (bit & final_subsys_mask) { + root->subsys_mask &= ~bit; + } else if (bit & root->subsys_mask) { /* Subsystem state should already exist */ - BUG_ON(ss == NULL); BUG_ON(!cgrp->subsys[i]); /* * a refcount was taken, but we already had one, so @@ -1071,7 +1085,12 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(cgrp->subsys[i]); } } - root->subsys_mask = root->actual_subsys_mask = final_subsys_mask; + + /* + * Mark @root has finished binding subsystems. @root->subsys_mask + * now matches the bound subsystems. + */ + root->flags |= CGRP_ROOT_SUBSYS_BOUND; return 0; } @@ -1082,7 +1101,7 @@ static int cgroup_show_options(struct seq_file *seq, struct dentry *dentry) struct cgroup_subsys *ss; mutex_lock(&cgroup_root_mutex); - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) seq_printf(seq, ",%s", ss->name); if (root->flags & CGRP_ROOT_SANE_BEHAVIOR) seq_puts(seq, ",sane_behavior"); @@ -1114,18 +1133,19 @@ struct cgroup_sb_opts { }; /* - * Convert a hierarchy specifier into a bitmask of subsystems and flags. Call - * with cgroup_mutex held to protect the subsys[] array. This function takes - * refcounts on subsystems to be used, unless it returns error, in which case - * no refcounts are taken. + * Convert a hierarchy specifier into a bitmask of subsystems and + * flags. Call with cgroup_mutex held to protect the cgroup_subsys[] + * array. This function takes refcounts on subsystems to be used, unless it + * returns error, in which case no refcounts are taken. */ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) { char *token, *o = data; bool all_ss = false, one_ss = false; unsigned long mask = (unsigned long)-1; - int i; bool module_pin_failed = false; + struct cgroup_subsys *ss; + int i; BUG_ON(!mutex_is_locked(&cgroup_mutex)); @@ -1202,10 +1222,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) continue; } - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - if (ss == NULL) - continue; + for_each_subsys(ss, i) { if (strcmp(token, ss->name)) continue; if (ss->disabled) @@ -1228,16 +1245,10 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) * otherwise if 'none', 'name=' and a subsystem name options * were not specified, let's default to 'all' */ - if (all_ss || (!one_ss && !opts->none && !opts->name)) { - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - if (ss == NULL) - continue; - if (ss->disabled) - continue; - set_bit(i, &opts->subsys_mask); - } - } + if (all_ss || (!one_ss && !opts->none && !opts->name)) + for_each_subsys(ss, i) + if (!ss->disabled) + set_bit(i, &opts->subsys_mask); /* Consistency checks */ @@ -1281,12 +1292,10 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) * take duplicate reference counts on a subsystem that's already used, * but rebind_subsystems handles this case. */ - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - unsigned long bit = 1UL << i; - - if (!(bit & opts->subsys_mask)) + for_each_subsys(ss, i) { + if (!(opts->subsys_mask & (1UL << i))) continue; - if (!try_module_get(subsys[i]->module)) { + if (!try_module_get(cgroup_subsys[i]->module)) { module_pin_failed = true; break; } @@ -1303,7 +1312,7 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) if (!(bit & opts->subsys_mask)) continue; - module_put(subsys[i]->module); + module_put(cgroup_subsys[i]->module); } return -ENOENT; } @@ -1313,14 +1322,14 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) static void drop_parsed_module_refcounts(unsigned long subsys_mask) { + struct cgroup_subsys *ss; int i; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - unsigned long bit = 1UL << i; - if (!(bit & subsys_mask)) - continue; - module_put(subsys[i]->module); - } + mutex_lock(&cgroup_mutex); + for_each_subsys(ss, i) + if (subsys_mask & (1UL << i)) + module_put(cgroup_subsys[i]->module); + mutex_unlock(&cgroup_mutex); } static int cgroup_remount(struct super_block *sb, int *flags, char *data) @@ -1345,7 +1354,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) if (ret) goto out_unlock; - if (opts.subsys_mask != root->actual_subsys_mask || opts.release_agent) + if (opts.subsys_mask != root->subsys_mask || opts.release_agent) pr_warning("cgroup: option changes via remount are deprecated (pid=%d comm=%s)\n", task_tgid_nr(current), current->comm); @@ -1353,10 +1362,12 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) removed_mask = root->subsys_mask & ~opts.subsys_mask; /* Don't allow flags or name to change at remount */ - if (opts.flags != root->flags || + if (((opts.flags ^ root->flags) & CGRP_ROOT_OPTION_MASK) || (opts.name && strcmp(opts.name, root->name))) { + pr_err("cgroup: option or name mismatch, new: 0x%lx \"%s\", old: 0x%lx \"%s\"\n", + opts.flags & CGRP_ROOT_OPTION_MASK, opts.name ?: "", + root->flags & CGRP_ROOT_OPTION_MASK, root->name); ret = -EINVAL; - drop_parsed_module_refcounts(opts.subsys_mask); goto out_unlock; } @@ -1367,11 +1378,10 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) */ cgroup_clear_directory(cgrp->dentry, false, removed_mask); - ret = rebind_subsystems(root, opts.subsys_mask); + ret = rebind_subsystems(root, added_mask, removed_mask); if (ret) { /* rebind_subsystems failed, re-populate the removed files */ cgroup_populate_dir(cgrp, false, removed_mask); - drop_parsed_module_refcounts(opts.subsys_mask); goto out_unlock; } @@ -1386,6 +1396,8 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&cgrp->dentry->d_inode->i_mutex); + if (ret) + drop_parsed_module_refcounts(opts.subsys_mask); return ret; } @@ -1401,11 +1413,9 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->sibling); INIT_LIST_HEAD(&cgrp->children); INIT_LIST_HEAD(&cgrp->files); - INIT_LIST_HEAD(&cgrp->css_sets); - INIT_LIST_HEAD(&cgrp->allcg_node); + INIT_LIST_HEAD(&cgrp->cset_links); INIT_LIST_HEAD(&cgrp->release_list); INIT_LIST_HEAD(&cgrp->pidlists); - INIT_WORK(&cgrp->free_work, cgroup_free_fn); mutex_init(&cgrp->pidlist_mutex); INIT_LIST_HEAD(&cgrp->event_list); spin_lock_init(&cgrp->event_list_lock); @@ -1418,37 +1428,37 @@ static void init_cgroup_root(struct cgroupfs_root *root) INIT_LIST_HEAD(&root->subsys_list); INIT_LIST_HEAD(&root->root_list); - INIT_LIST_HEAD(&root->allcg_list); root->number_of_cgroups = 1; cgrp->root = root; - cgrp->name = &root_cgroup_name; + RCU_INIT_POINTER(cgrp->name, &root_cgroup_name); init_cgroup_housekeeping(cgrp); - list_add_tail(&cgrp->allcg_node, &root->allcg_list); } -static bool init_root_id(struct cgroupfs_root *root) +static int cgroup_init_root_id(struct cgroupfs_root *root, int start, int end) { - int ret = 0; + int id; - do { - if (!ida_pre_get(&hierarchy_ida, GFP_KERNEL)) - return false; - spin_lock(&hierarchy_id_lock); - /* Try to allocate the next unused ID */ - ret = ida_get_new_above(&hierarchy_ida, next_hierarchy_id, - &root->hierarchy_id); - if (ret == -ENOSPC) - /* Try again starting from 0 */ - ret = ida_get_new(&hierarchy_ida, &root->hierarchy_id); - if (!ret) { - next_hierarchy_id = root->hierarchy_id + 1; - } else if (ret != -EAGAIN) { - /* Can only get here if the 31-bit IDR is full ... */ - BUG_ON(ret); - } - spin_unlock(&hierarchy_id_lock); - } while (ret); - return true; + lockdep_assert_held(&cgroup_mutex); + lockdep_assert_held(&cgroup_root_mutex); + + id = idr_alloc_cyclic(&cgroup_hierarchy_idr, root, start, end, + GFP_KERNEL); + if (id < 0) + return id; + + root->hierarchy_id = id; + return 0; +} + +static void cgroup_exit_root_id(struct cgroupfs_root *root) +{ + lockdep_assert_held(&cgroup_mutex); + lockdep_assert_held(&cgroup_root_mutex); + + if (root->hierarchy_id) { + idr_remove(&cgroup_hierarchy_idr, root->hierarchy_id); + root->hierarchy_id = 0; + } } static int cgroup_test_super(struct super_block *sb, void *data) @@ -1482,12 +1492,16 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) if (!root) return ERR_PTR(-ENOMEM); - if (!init_root_id(root)) { - kfree(root); - return ERR_PTR(-ENOMEM); - } init_cgroup_root(root); + /* + * We need to set @root->subsys_mask now so that @root can be + * matched by cgroup_test_super() before it finishes + * initialization; otherwise, competing mounts with the same + * options may try to bind the same subsystems instead of waiting + * for the first one leading to unexpected mount errors. + * SUBSYS_BOUND will be set once actual binding is complete. + */ root->subsys_mask = opts->subsys_mask; root->flags = opts->flags; ida_init(&root->cgroup_ida); @@ -1500,17 +1514,15 @@ static struct cgroupfs_root *cgroup_root_from_opts(struct cgroup_sb_opts *opts) return root; } -static void cgroup_drop_root(struct cgroupfs_root *root) +static void cgroup_free_root(struct cgroupfs_root *root) { - if (!root) - return; + if (root) { + /* hierarhcy ID shoulid already have been released */ + WARN_ON_ONCE(root->hierarchy_id); - BUG_ON(!root->hierarchy_id); - spin_lock(&hierarchy_id_lock); - ida_remove(&hierarchy_ida, root->hierarchy_id); - spin_unlock(&hierarchy_id_lock); - ida_destroy(&root->cgroup_ida); - kfree(root); + ida_destroy(&root->cgroup_ida); + kfree(root); + } } static int cgroup_set_super(struct super_block *sb, void *data) @@ -1597,7 +1609,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, sb = sget(fs_type, cgroup_test_super, cgroup_set_super, 0, &opts); if (IS_ERR(sb)) { ret = PTR_ERR(sb); - cgroup_drop_root(opts.new_root); + cgroup_free_root(opts.new_root); goto drop_modules; } @@ -1605,12 +1617,12 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, BUG_ON(!root); if (root == opts.new_root) { /* We used the new root structure, so this is a new hierarchy */ - struct list_head tmp_cg_links; + struct list_head tmp_links; struct cgroup *root_cgrp = &root->top_cgroup; struct cgroupfs_root *existing_root; const struct cred *cred; int i; - struct css_set *cg; + struct css_set *cset; BUG_ON(sb->s_root != NULL); @@ -1637,13 +1649,18 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, * that's us. The worst that can happen is that we * have some link structures left over */ - ret = allocate_cg_links(css_set_count, &tmp_cg_links); + ret = allocate_cgrp_cset_links(css_set_count, &tmp_links); if (ret) goto unlock_drop; - ret = rebind_subsystems(root, root->subsys_mask); + /* ID 0 is reserved for dummy root, 1 for unified hierarchy */ + ret = cgroup_init_root_id(root, 2, 0); + if (ret) + goto unlock_drop; + + ret = rebind_subsystems(root, root->subsys_mask, 0); if (ret == -EBUSY) { - free_cg_links(&tmp_cg_links); + free_cgrp_cset_links(&tmp_links); goto unlock_drop; } /* @@ -1655,8 +1672,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, /* EBUSY should be the only error here */ BUG_ON(ret); - list_add(&root->root_list, &roots); - root_count++; + list_add(&root->root_list, &cgroup_roots); + cgroup_root_count++; sb->s_root->d_fsdata = root_cgrp; root->top_cgroup.dentry = sb->s_root; @@ -1664,11 +1681,11 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, /* Link the top cgroup in this hierarchy into all * the css_set objects */ write_lock(&css_set_lock); - hash_for_each(css_set_table, i, cg, hlist) - link_css_set(&tmp_cg_links, cg, root_cgrp); + hash_for_each(css_set_table, i, cset, hlist) + link_css_set(&tmp_links, cset, root_cgrp); write_unlock(&css_set_lock); - free_cg_links(&tmp_cg_links); + free_cgrp_cset_links(&tmp_links); BUG_ON(!list_empty(&root_cgrp->children)); BUG_ON(root->number_of_cgroups != 1); @@ -1684,13 +1701,16 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, * We re-used an existing hierarchy - the new root (if * any) is not needed */ - cgroup_drop_root(opts.new_root); + cgroup_free_root(opts.new_root); - if (((root->flags | opts.flags) & CGRP_ROOT_SANE_BEHAVIOR) && - root->flags != opts.flags) { - pr_err("cgroup: sane_behavior: new mount options should match the existing superblock\n"); - ret = -EINVAL; - goto drop_new_super; + if ((root->flags ^ opts.flags) & CGRP_ROOT_OPTION_MASK) { + if ((root->flags | opts.flags) & CGRP_ROOT_SANE_BEHAVIOR) { + pr_err("cgroup: sane_behavior: new mount options should match the existing superblock\n"); + ret = -EINVAL; + goto drop_new_super; + } else { + pr_warning("cgroup: new mount options do not match the existing superblock, will be ignored\n"); + } } /* no subsys rebinding, so refcounts don't change */ @@ -1702,6 +1722,7 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, return dget(sb->s_root); unlock_drop: + cgroup_exit_root_id(root); mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); @@ -1718,9 +1739,8 @@ static struct dentry *cgroup_mount(struct file_system_type *fs_type, static void cgroup_kill_sb(struct super_block *sb) { struct cgroupfs_root *root = sb->s_fs_info; struct cgroup *cgrp = &root->top_cgroup; + struct cgrp_cset_link *link, *tmp_link; int ret; - struct cg_cgroup_link *link; - struct cg_cgroup_link *saved_link; BUG_ON(!root); @@ -1731,36 +1751,39 @@ static void cgroup_kill_sb(struct super_block *sb) { mutex_lock(&cgroup_root_mutex); /* Rebind all subsystems back to the default hierarchy */ - ret = rebind_subsystems(root, 0); - /* Shouldn't be able to fail ... */ - BUG_ON(ret); + if (root->flags & CGRP_ROOT_SUBSYS_BOUND) { + ret = rebind_subsystems(root, 0, root->subsys_mask); + /* Shouldn't be able to fail ... */ + BUG_ON(ret); + } /* - * Release all the links from css_sets to this hierarchy's + * Release all the links from cset_links to this hierarchy's * root cgroup */ write_lock(&css_set_lock); - list_for_each_entry_safe(link, saved_link, &cgrp->css_sets, - cgrp_link_list) { - list_del(&link->cg_link_list); - list_del(&link->cgrp_link_list); + list_for_each_entry_safe(link, tmp_link, &cgrp->cset_links, cset_link) { + list_del(&link->cset_link); + list_del(&link->cgrp_link); kfree(link); } write_unlock(&css_set_lock); if (!list_empty(&root->root_list)) { list_del(&root->root_list); - root_count--; + cgroup_root_count--; } + cgroup_exit_root_id(root); + mutex_unlock(&cgroup_root_mutex); mutex_unlock(&cgroup_mutex); simple_xattrs_free(&cgrp->xattrs); kill_litter_super(sb); - cgroup_drop_root(root); + cgroup_free_root(root); } static struct file_system_type cgroup_fs_type = { @@ -1822,6 +1845,38 @@ out: } EXPORT_SYMBOL_GPL(cgroup_path); +/** + * task_cgroup_path_from_hierarchy - cgroup path of a task on a hierarchy + * @task: target task + * @hierarchy_id: the hierarchy to look up @task's cgroup from + * @buf: the buffer to write the path into + * @buflen: the length of the buffer + * + * Determine @task's cgroup on the hierarchy specified by @hierarchy_id and + * copy its path into @buf. This function grabs cgroup_mutex and shouldn't + * be used inside locks used by cgroup controller callbacks. + */ +int task_cgroup_path_from_hierarchy(struct task_struct *task, int hierarchy_id, + char *buf, size_t buflen) +{ + struct cgroupfs_root *root; + struct cgroup *cgrp = NULL; + int ret = -ENOENT; + + mutex_lock(&cgroup_mutex); + + root = idr_find(&cgroup_hierarchy_idr, hierarchy_id); + if (root) { + cgrp = task_cgroup_from_root(task, root); + ret = cgroup_path(cgrp, buf, buflen); + } + + mutex_unlock(&cgroup_mutex); + + return ret; +} +EXPORT_SYMBOL_GPL(task_cgroup_path_from_hierarchy); + /* * Control Group taskset */ @@ -1907,10 +1962,11 @@ EXPORT_SYMBOL_GPL(cgroup_taskset_size); * * Must be called with cgroup_mutex and threadgroup locked. */ -static void cgroup_task_migrate(struct cgroup *oldcgrp, - struct task_struct *tsk, struct css_set *newcg) +static void cgroup_task_migrate(struct cgroup *old_cgrp, + struct task_struct *tsk, + struct css_set *new_cset) { - struct css_set *oldcg; + struct css_set *old_cset; /* * We are synchronized through threadgroup_lock() against PF_EXITING @@ -1918,25 +1974,25 @@ static void cgroup_task_migrate(struct cgroup *oldcgrp, * css_set to init_css_set and dropping the old one. */ WARN_ON_ONCE(tsk->flags & PF_EXITING); - oldcg = tsk->cgroups; + old_cset = task_css_set(tsk); task_lock(tsk); - rcu_assign_pointer(tsk->cgroups, newcg); + rcu_assign_pointer(tsk->cgroups, new_cset); task_unlock(tsk); /* Update the css_set linked lists if we're using them */ write_lock(&css_set_lock); if (!list_empty(&tsk->cg_list)) - list_move(&tsk->cg_list, &newcg->tasks); + list_move(&tsk->cg_list, &new_cset->tasks); write_unlock(&css_set_lock); /* - * We just gained a reference on oldcg by taking it from the task. As - * trading it for newcg is protected by cgroup_mutex, we're safe to drop - * it here; it will be freed under RCU. + * We just gained a reference on old_cset by taking it from the + * task. As trading it for new_cset is protected by cgroup_mutex, + * we're safe to drop it here; it will be freed under RCU. */ - set_bit(CGRP_RELEASABLE, &oldcgrp->flags); - put_css_set(oldcg); + set_bit(CGRP_RELEASABLE, &old_cgrp->flags); + put_css_set(old_cset); } /** @@ -2026,7 +2082,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, /* * step 1: check that we can legitimately attach to the cgroup. */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss->can_attach) { retval = ss->can_attach(cgrp, &tset); if (retval) { @@ -2041,8 +2097,11 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, * we use find_css_set, which allocates a new one if necessary. */ for (i = 0; i < group_size; i++) { + struct css_set *old_cset; + tc = flex_array_get(group, i); - tc->cg = find_css_set(tc->task->cgroups, cgrp); + old_cset = task_css_set(tc->task); + tc->cg = find_css_set(old_cset, cgrp); if (!tc->cg) { retval = -ENOMEM; goto out_put_css_set_refs; @@ -2063,7 +2122,7 @@ static int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk, /* * step 4: do subsystem attach callbacks. */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss->attach) ss->attach(cgrp, &tset); } @@ -2083,7 +2142,7 @@ out_put_css_set_refs: } out_cancel_attach: if (retval) { - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { if (ss == failed_ss) break; if (ss->cancel_attach) @@ -2320,7 +2379,7 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_removed(cgrp)) + if (cgroup_is_dead(cgrp)) return -ENODEV; if (cft->write) return cft->write(cgrp, cft, file, buf, nbytes, ppos); @@ -2365,7 +2424,7 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, struct cftype *cft = __d_cft(file->f_dentry); struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - if (cgroup_is_removed(cgrp)) + if (cgroup_is_dead(cgrp)) return -ENODEV; if (cft->read) @@ -2432,10 +2491,12 @@ static int cgroup_file_open(struct inode *inode, struct file *file) cft = __d_cft(file->f_dentry); if (cft->read_map || cft->read_seq_string) { - struct cgroup_seqfile_state *state = - kzalloc(sizeof(*state), GFP_USER); + struct cgroup_seqfile_state *state; + + state = kzalloc(sizeof(*state), GFP_USER); if (!state) return -ENOMEM; + state->cft = cft; state->cgroup = __d_cgrp(file->f_dentry->d_parent); file->f_op = &cgroup_seqfile_operations; @@ -2483,6 +2544,13 @@ static int cgroup_rename(struct inode *old_dir, struct dentry *old_dentry, cgrp = __d_cgrp(old_dentry); + /* + * This isn't a proper migration and its usefulness is very + * limited. Disallow if sane_behavior. + */ + if (cgroup_sane_behavior(cgrp)) + return -EPERM; + name = cgroup_alloc_name(new_dentry); if (!name) return -ENOMEM; @@ -2493,7 +2561,7 @@ static int cgroup_rename(struct inode *old_dir, struct dentry *old_dentry, return ret; } - old_name = cgrp->name; + old_name = rcu_dereference_protected(cgrp->name, true); rcu_assign_pointer(cgrp->name, name); kfree_rcu(old_name, rcu_head); @@ -2699,13 +2767,14 @@ static int cgroup_add_file(struct cgroup *cgrp, struct cgroup_subsys *subsys, goto out; } + cfe->type = (void *)cft; + cfe->dentry = dentry; + dentry->d_fsdata = cfe; + simple_xattrs_init(&cfe->xattrs); + mode = cgroup_file_mode(cft); error = cgroup_create_file(dentry, mode | S_IFREG, cgrp->root->sb); if (!error) { - cfe->type = (void *)cft; - cfe->dentry = dentry; - dentry->d_fsdata = cfe; - simple_xattrs_init(&cfe->xattrs); list_add_tail(&cfe->node, &parent->files); cfe = NULL; } @@ -2743,58 +2812,78 @@ static int cgroup_addrm_files(struct cgroup *cgrp, struct cgroup_subsys *subsys, return ret; } -static DEFINE_MUTEX(cgroup_cft_mutex); - static void cgroup_cfts_prepare(void) - __acquires(&cgroup_cft_mutex) __acquires(&cgroup_mutex) + __acquires(&cgroup_mutex) { /* * Thanks to the entanglement with vfs inode locking, we can't walk * the existing cgroups under cgroup_mutex and create files. - * Instead, we increment reference on all cgroups and build list of - * them using @cgrp->cft_q_node. Grab cgroup_cft_mutex to ensure - * exclusive access to the field. + * Instead, we use cgroup_for_each_descendant_pre() and drop RCU + * read lock before calling cgroup_addrm_files(). */ - mutex_lock(&cgroup_cft_mutex); mutex_lock(&cgroup_mutex); } static void cgroup_cfts_commit(struct cgroup_subsys *ss, struct cftype *cfts, bool is_add) - __releases(&cgroup_mutex) __releases(&cgroup_cft_mutex) + __releases(&cgroup_mutex) { LIST_HEAD(pending); - struct cgroup *cgrp, *n; + struct cgroup *cgrp, *root = &ss->root->top_cgroup; + struct super_block *sb = ss->root->sb; + struct dentry *prev = NULL; + struct inode *inode; + u64 update_before; /* %NULL @cfts indicates abort and don't bother if @ss isn't attached */ - if (cfts && ss->root != &rootnode) { - list_for_each_entry(cgrp, &ss->root->allcg_list, allcg_node) { - dget(cgrp->dentry); - list_add_tail(&cgrp->cft_q_node, &pending); - } + if (!cfts || ss->root == &cgroup_dummy_root || + !atomic_inc_not_zero(&sb->s_active)) { + mutex_unlock(&cgroup_mutex); + return; } - mutex_unlock(&cgroup_mutex); - /* - * All new cgroups will see @cfts update on @ss->cftsets. Add/rm - * files for all cgroups which were created before. + * All cgroups which are created after we drop cgroup_mutex will + * have the updated set of files, so we only need to update the + * cgroups created before the current @cgroup_serial_nr_next. */ - list_for_each_entry_safe(cgrp, n, &pending, cft_q_node) { - struct inode *inode = cgrp->dentry->d_inode; + update_before = cgroup_serial_nr_next; + + mutex_unlock(&cgroup_mutex); + + /* @root always needs to be updated */ + inode = root->dentry->d_inode; + mutex_lock(&inode->i_mutex); + mutex_lock(&cgroup_mutex); + cgroup_addrm_files(root, ss, cfts, is_add); + mutex_unlock(&cgroup_mutex); + mutex_unlock(&inode->i_mutex); + + /* add/rm files for all cgroups created before */ + rcu_read_lock(); + cgroup_for_each_descendant_pre(cgrp, root) { + if (cgroup_is_dead(cgrp)) + continue; + + inode = cgrp->dentry->d_inode; + dget(cgrp->dentry); + rcu_read_unlock(); + + dput(prev); + prev = cgrp->dentry; mutex_lock(&inode->i_mutex); mutex_lock(&cgroup_mutex); - if (!cgroup_is_removed(cgrp)) + if (cgrp->serial_nr < update_before && !cgroup_is_dead(cgrp)) cgroup_addrm_files(cgrp, ss, cfts, is_add); mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); - list_del_init(&cgrp->cft_q_node); - dput(cgrp->dentry); + rcu_read_lock(); } - - mutex_unlock(&cgroup_cft_mutex); + rcu_read_unlock(); + dput(prev); + deactivate_super(sb); } /** @@ -2849,7 +2938,8 @@ int cgroup_rm_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) list_for_each_entry(set, &ss->cftsets, node) { if (set->cfts == cfts) { - list_del_init(&set->node); + list_del(&set->node); + kfree(set); cgroup_cfts_commit(ss, cfts, false); return 0; } @@ -2868,12 +2958,11 @@ int cgroup_rm_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) int cgroup_task_count(const struct cgroup *cgrp) { int count = 0; - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cgrp->css_sets, cgrp_link_list) { - count += atomic_read(&link->cg->refcount); - } + list_for_each_entry(link, &cgrp->cset_links, cset_link) + count += atomic_read(&link->cset->refcount); read_unlock(&css_set_lock); return count; } @@ -2882,25 +2971,24 @@ int cgroup_task_count(const struct cgroup *cgrp) * Advance a list_head iterator. The iterator should be positioned at * the start of a css_set */ -static void cgroup_advance_iter(struct cgroup *cgrp, - struct cgroup_iter *it) +static void cgroup_advance_iter(struct cgroup *cgrp, struct cgroup_iter *it) { - struct list_head *l = it->cg_link; - struct cg_cgroup_link *link; - struct css_set *cg; + struct list_head *l = it->cset_link; + struct cgrp_cset_link *link; + struct css_set *cset; /* Advance to the next non-empty css_set */ do { l = l->next; - if (l == &cgrp->css_sets) { - it->cg_link = NULL; + if (l == &cgrp->cset_links) { + it->cset_link = NULL; return; } - link = list_entry(l, struct cg_cgroup_link, cgrp_link_list); - cg = link->cg; - } while (list_empty(&cg->tasks)); - it->cg_link = l; - it->task = cg->tasks.next; + link = list_entry(l, struct cgrp_cset_link, cset_link); + cset = link->cset; + } while (list_empty(&cset->tasks)); + it->cset_link = l; + it->task = cset->tasks.next; } /* @@ -2930,7 +3018,7 @@ static void cgroup_enable_task_cg_lists(void) * entry won't be deleted though the process has exited. */ if (!(p->flags & PF_EXITING) && list_empty(&p->cg_list)) - list_add(&p->cg_list, &p->cgroups->tasks); + list_add(&p->cg_list, &task_css_set(p)->tasks); task_unlock(p); } while_each_thread(g, p); read_unlock(&tasklist_lock); @@ -2938,12 +3026,67 @@ static void cgroup_enable_task_cg_lists(void) } /** + * cgroup_next_sibling - find the next sibling of a given cgroup + * @pos: the current cgroup + * + * This function returns the next sibling of @pos and should be called + * under RCU read lock. The only requirement is that @pos is accessible. + * The next sibling is guaranteed to be returned regardless of @pos's + * state. + */ +struct cgroup *cgroup_next_sibling(struct cgroup *pos) +{ + struct cgroup *next; + + WARN_ON_ONCE(!rcu_read_lock_held()); + + /* + * @pos could already have been removed. Once a cgroup is removed, + * its ->sibling.next is no longer updated when its next sibling + * changes. As CGRP_DEAD assertion is serialized and happens + * before the cgroup is taken off the ->sibling list, if we see it + * unasserted, it's guaranteed that the next sibling hasn't + * finished its grace period even if it's already removed, and thus + * safe to dereference from this RCU critical section. If + * ->sibling.next is inaccessible, cgroup_is_dead() is guaranteed + * to be visible as %true here. + */ + if (likely(!cgroup_is_dead(pos))) { + next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); + if (&next->sibling != &pos->parent->children) + return next; + return NULL; + } + + /* + * Can't dereference the next pointer. Each cgroup is given a + * monotonically increasing unique serial number and always + * appended to the sibling list, so the next one can be found by + * walking the parent's children until we see a cgroup with higher + * serial number than @pos's. + * + * While this path can be slow, it's taken only when either the + * current cgroup is removed or iteration and removal race. + */ + list_for_each_entry_rcu(next, &pos->parent->children, sibling) + if (next->serial_nr > pos->serial_nr) + return next; + return NULL; +} +EXPORT_SYMBOL_GPL(cgroup_next_sibling); + +/** * cgroup_next_descendant_pre - find the next descendant for pre-order walk * @pos: the current position (%NULL to initiate traversal) * @cgroup: cgroup whose descendants to walk * * To be used by cgroup_for_each_descendant_pre(). Find the next * descendant to visit for pre-order traversal of @cgroup's descendants. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct next descendant as long as both @pos + * and @cgroup are accessible and @pos is a descendant of @cgroup. */ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, struct cgroup *cgroup) @@ -2953,11 +3096,8 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, WARN_ON_ONCE(!rcu_read_lock_held()); /* if first iteration, pretend we just visited @cgroup */ - if (!pos) { - if (list_empty(&cgroup->children)) - return NULL; + if (!pos) pos = cgroup; - } /* visit the first child if exists */ next = list_first_or_null_rcu(&pos->children, struct cgroup, sibling); @@ -2965,14 +3105,12 @@ struct cgroup *cgroup_next_descendant_pre(struct cgroup *pos, return next; /* no child, visit my or the closest ancestor's next sibling */ - do { - next = list_entry_rcu(pos->sibling.next, struct cgroup, - sibling); - if (&next->sibling != &pos->parent->children) + while (pos != cgroup) { + next = cgroup_next_sibling(pos); + if (next) return next; - pos = pos->parent; - } while (pos != cgroup); + } return NULL; } @@ -2985,6 +3123,11 @@ EXPORT_SYMBOL_GPL(cgroup_next_descendant_pre); * Return the rightmost descendant of @pos. If there's no descendant, * @pos is returned. This can be used during pre-order traversal to skip * subtree of @pos. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct rightmost descendant as long as @pos is + * accessible. */ struct cgroup *cgroup_rightmost_descendant(struct cgroup *pos) { @@ -3024,6 +3167,11 @@ static struct cgroup *cgroup_leftmost_descendant(struct cgroup *pos) * * To be used by cgroup_for_each_descendant_post(). Find the next * descendant to visit for post-order traversal of @cgroup's descendants. + * + * While this function requires RCU read locking, it doesn't require the + * whole traversal to be contained in a single RCU critical section. This + * function will return the correct next descendant as long as both @pos + * and @cgroup are accessible and @pos is a descendant of @cgroup. */ struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, struct cgroup *cgroup) @@ -3039,8 +3187,8 @@ struct cgroup *cgroup_next_descendant_post(struct cgroup *pos, } /* if there's an unvisited sibling, visit its leftmost descendant */ - next = list_entry_rcu(pos->sibling.next, struct cgroup, sibling); - if (&next->sibling != &pos->parent->children) + next = cgroup_next_sibling(pos); + if (next) return cgroup_leftmost_descendant(next); /* no sibling left, visit parent */ @@ -3061,7 +3209,7 @@ void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) cgroup_enable_task_cg_lists(); read_lock(&css_set_lock); - it->cg_link = &cgrp->css_sets; + it->cset_link = &cgrp->cset_links; cgroup_advance_iter(cgrp, it); } @@ -3070,16 +3218,16 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, { struct task_struct *res; struct list_head *l = it->task; - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; /* If the iterator cg is NULL, we have no tasks */ - if (!it->cg_link) + if (!it->cset_link) return NULL; res = list_entry(l, struct task_struct, cg_list); /* Advance iterator to find next entry */ l = l->next; - link = list_entry(it->cg_link, struct cg_cgroup_link, cgrp_link_list); - if (l == &link->cg->tasks) { + link = list_entry(it->cset_link, struct cgrp_cset_link, cset_link); + if (l == &link->cset->tasks) { /* We reached the end of this task list - move on to * the next cg_cgroup_link */ cgroup_advance_iter(cgrp, it); @@ -3410,7 +3558,7 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, } } /* entry not found; create a new one */ - l = kmalloc(sizeof(struct cgroup_pidlist), GFP_KERNEL); + l = kzalloc(sizeof(struct cgroup_pidlist), GFP_KERNEL); if (!l) { mutex_unlock(&cgrp->pidlist_mutex); return l; @@ -3419,8 +3567,6 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, down_write(&l->mutex); l->key.type = type; l->key.ns = get_pid_ns(ns); - l->use_count = 0; /* don't increment here */ - l->list = NULL; l->owner = cgrp; list_add(&l->links, &cgrp->pidlists); mutex_unlock(&cgrp->pidlist_mutex); @@ -3726,6 +3872,23 @@ static int cgroup_write_notify_on_release(struct cgroup *cgrp, } /* + * When dput() is called asynchronously, if umount has been done and + * then deactivate_super() in cgroup_free_fn() kills the superblock, + * there's a small window that vfs will see the root dentry with non-zero + * refcnt and trigger BUG(). + * + * That's why we hold a reference before dput() and drop it right after. + */ +static void cgroup_dput(struct cgroup *cgrp) +{ + struct super_block *sb = cgrp->root->sb; + + atomic_inc(&sb->s_active); + dput(cgrp->dentry); + deactivate_super(sb); +} + +/* * Unregister event and free resources. * * Gets called from workqueue. @@ -3745,7 +3908,7 @@ static void cgroup_event_remove(struct work_struct *work) eventfd_ctx_put(event->eventfd); kfree(event); - dput(cgrp->dentry); + cgroup_dput(cgrp); } /* @@ -3932,33 +4095,16 @@ static int cgroup_clone_children_write(struct cgroup *cgrp, return 0; } -/* - * for the common functions, 'private' gives the type of file - */ -/* for hysterical raisins, we can't put this on the older files */ -#define CGROUP_FILE_GENERIC_PREFIX "cgroup." -static struct cftype files[] = { +static struct cftype cgroup_base_files[] = { { - .name = "tasks", - .open = cgroup_tasks_open, - .write_u64 = cgroup_tasks_write, - .release = cgroup_pidlist_release, - .mode = S_IRUGO | S_IWUSR, - }, - { - .name = CGROUP_FILE_GENERIC_PREFIX "procs", + .name = "cgroup.procs", .open = cgroup_procs_open, .write_u64 = cgroup_procs_write, .release = cgroup_pidlist_release, .mode = S_IRUGO | S_IWUSR, }, { - .name = "notify_on_release", - .read_u64 = cgroup_read_notify_on_release, - .write_u64 = cgroup_write_notify_on_release, - }, - { - .name = CGROUP_FILE_GENERIC_PREFIX "event_control", + .name = "cgroup.event_control", .write_string = cgroup_write_event_control, .mode = S_IWUGO, }, @@ -3973,9 +4119,29 @@ static struct cftype files[] = { .flags = CFTYPE_ONLY_ON_ROOT, .read_seq_string = cgroup_sane_behavior_show, }, + + /* + * Historical crazy stuff. These don't have "cgroup." prefix and + * don't exist if sane_behavior. If you're depending on these, be + * prepared to be burned. + */ + { + .name = "tasks", + .flags = CFTYPE_INSANE, /* use "procs" instead */ + .open = cgroup_tasks_open, + .write_u64 = cgroup_tasks_write, + .release = cgroup_pidlist_release, + .mode = S_IRUGO | S_IWUSR, + }, + { + .name = "notify_on_release", + .flags = CFTYPE_INSANE, + .read_u64 = cgroup_read_notify_on_release, + .write_u64 = cgroup_write_notify_on_release, + }, { .name = "release_agent", - .flags = CFTYPE_ONLY_ON_ROOT, + .flags = CFTYPE_INSANE | CFTYPE_ONLY_ON_ROOT, .read_seq_string = cgroup_release_agent_show, .write_string = cgroup_release_agent_write, .max_write_len = PATH_MAX, @@ -3996,13 +4162,13 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, struct cgroup_subsys *ss; if (base_files) { - err = cgroup_addrm_files(cgrp, NULL, files, true); + err = cgroup_addrm_files(cgrp, NULL, cgroup_base_files, true); if (err < 0) return err; } /* process cftsets of each subsystem */ - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cftype_set *set; if (!test_bit(ss->subsys_id, &subsys_mask)) continue; @@ -4012,15 +4178,17 @@ static int cgroup_populate_dir(struct cgroup *cgrp, bool base_files, } /* This cgroup is ready now */ - for_each_subsys(cgrp->root, ss) { + for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + struct css_id *id = rcu_dereference_protected(css->id, true); + /* * Update id->css pointer and make this css visible from * CSS ID functions. This pointer will be dereferened * from RCU-read-side without locks. */ - if (css->id) - rcu_assign_pointer(css->id->css, css); + if (id) + rcu_assign_pointer(id->css, css); } return 0; @@ -4030,12 +4198,16 @@ static void css_dput_fn(struct work_struct *work) { struct cgroup_subsys_state *css = container_of(work, struct cgroup_subsys_state, dput_work); - struct dentry *dentry = css->cgroup->dentry; - struct super_block *sb = dentry->d_sb; - atomic_inc(&sb->s_active); - dput(dentry); - deactivate_super(sb); + cgroup_dput(css->cgroup); +} + +static void css_release(struct percpu_ref *ref) +{ + struct cgroup_subsys_state *css = + container_of(ref, struct cgroup_subsys_state, refcnt); + + schedule_work(&css->dput_work); } static void init_cgroup_css(struct cgroup_subsys_state *css, @@ -4043,10 +4215,9 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, struct cgroup *cgrp) { css->cgroup = cgrp; - atomic_set(&css->refcnt, 1); css->flags = 0; css->id = NULL; - if (cgrp == dummytop) + if (cgrp == cgroup_dummy_top) css->flags |= CSS_ROOT; BUG_ON(cgrp->subsys[ss->subsys_id]); cgrp->subsys[ss->subsys_id] = css; @@ -4156,7 +4327,7 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, if (test_bit(CGRP_CPUSET_CLONE_CHILDREN, &parent->flags)) set_bit(CGRP_CPUSET_CLONE_CHILDREN, &cgrp->flags); - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { struct cgroup_subsys_state *css; css = ss->css_alloc(cgrp); @@ -4164,7 +4335,13 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err = PTR_ERR(css); goto err_free_all; } + + err = percpu_ref_init(&css->refcnt, css_release); + if (err) + goto err_free_all; + init_cgroup_css(css, ss, cgrp); + if (ss->use_id) { err = alloc_css_id(ss, parent, cgrp); if (err) @@ -4182,20 +4359,21 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_free_all; lockdep_assert_held(&dentry->d_inode->i_mutex); + cgrp->serial_nr = cgroup_serial_nr_next++; + /* allocation complete, commit to creation */ - list_add_tail(&cgrp->allcg_node, &root->allcg_list); list_add_tail_rcu(&cgrp->sibling, &cgrp->parent->children); root->number_of_cgroups++; /* each css holds a ref to the cgroup's dentry */ - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) dget(dentry); /* hold a ref to the parent's dentry */ dget(parent->dentry); /* creation succeeded, notify subsystems */ - for_each_subsys(root, ss) { + for_each_root_subsys(root, ss) { err = online_css(ss, cgrp); if (err) goto err_destroy; @@ -4220,9 +4398,13 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, return 0; err_free_all: - for_each_subsys(root, ss) { - if (cgrp->subsys[ss->subsys_id]) + for_each_root_subsys(root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + + if (css) { + percpu_ref_cancel_init(&css->refcnt); ss->css_free(cgrp); + } } mutex_unlock(&cgroup_mutex); /* Release the reference count that we took on the superblock */ @@ -4250,63 +4432,120 @@ static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) return cgroup_create(c_parent, dentry, mode | S_IFDIR); } +static void cgroup_css_killed(struct cgroup *cgrp) +{ + if (!atomic_dec_and_test(&cgrp->css_kill_cnt)) + return; + + /* percpu ref's of all css's are killed, kick off the next step */ + INIT_WORK(&cgrp->destroy_work, cgroup_offline_fn); + schedule_work(&cgrp->destroy_work); +} + +static void css_ref_killed_fn(struct percpu_ref *ref) +{ + struct cgroup_subsys_state *css = + container_of(ref, struct cgroup_subsys_state, refcnt); + + cgroup_css_killed(css->cgroup); +} + +/** + * cgroup_destroy_locked - the first stage of cgroup destruction + * @cgrp: cgroup to be destroyed + * + * css's make use of percpu refcnts whose killing latency shouldn't be + * exposed to userland and are RCU protected. Also, cgroup core needs to + * guarantee that css_tryget() won't succeed by the time ->css_offline() is + * invoked. To satisfy all the requirements, destruction is implemented in + * the following two steps. + * + * s1. Verify @cgrp can be destroyed and mark it dying. Remove all + * userland visible parts and start killing the percpu refcnts of + * css's. Set up so that the next stage will be kicked off once all + * the percpu refcnts are confirmed to be killed. + * + * s2. Invoke ->css_offline(), mark the cgroup dead and proceed with the + * rest of destruction. Once all cgroup references are gone, the + * cgroup is RCU-freed. + * + * This function implements s1. After this step, @cgrp is gone as far as + * the userland is concerned and a new cgroup with the same name may be + * created. As cgroup doesn't care about the names internally, this + * doesn't cause any problem. + */ static int cgroup_destroy_locked(struct cgroup *cgrp) __releases(&cgroup_mutex) __acquires(&cgroup_mutex) { struct dentry *d = cgrp->dentry; - struct cgroup *parent = cgrp->parent; struct cgroup_event *event, *tmp; struct cgroup_subsys *ss; + bool empty; lockdep_assert_held(&d->d_inode->i_mutex); lockdep_assert_held(&cgroup_mutex); - if (atomic_read(&cgrp->count) || !list_empty(&cgrp->children)) + /* + * css_set_lock synchronizes access to ->cset_links and prevents + * @cgrp from being removed while __put_css_set() is in progress. + */ + read_lock(&css_set_lock); + empty = list_empty(&cgrp->cset_links) && list_empty(&cgrp->children); + read_unlock(&css_set_lock); + if (!empty) return -EBUSY; /* - * Block new css_tryget() by deactivating refcnt and mark @cgrp - * removed. This makes future css_tryget() and child creation - * attempts fail thus maintaining the removal conditions verified - * above. + * Block new css_tryget() by killing css refcnts. cgroup core + * guarantees that, by the time ->css_offline() is invoked, no new + * css reference will be given out via css_tryget(). We can't + * simply call percpu_ref_kill() and proceed to offlining css's + * because percpu_ref_kill() doesn't guarantee that the ref is seen + * as killed on all CPUs on return. + * + * Use percpu_ref_kill_and_confirm() to get notifications as each + * css is confirmed to be seen as killed on all CPUs. The + * notification callback keeps track of the number of css's to be + * killed and schedules cgroup_offline_fn() to perform the rest of + * destruction once the percpu refs of all css's are confirmed to + * be killed. */ - for_each_subsys(cgrp->root, ss) { + atomic_set(&cgrp->css_kill_cnt, 1); + for_each_root_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; - WARN_ON(atomic_read(&css->refcnt) < 0); - atomic_add(CSS_DEACT_BIAS, &css->refcnt); - } - set_bit(CGRP_REMOVED, &cgrp->flags); + /* + * Killing would put the base ref, but we need to keep it + * alive until after ->css_offline. + */ + percpu_ref_get(&css->refcnt); - /* tell subsystems to initate destruction */ - for_each_subsys(cgrp->root, ss) - offline_css(ss, cgrp); + atomic_inc(&cgrp->css_kill_cnt); + percpu_ref_kill_and_confirm(&css->refcnt, css_ref_killed_fn); + } + cgroup_css_killed(cgrp); /* - * Put all the base refs. Each css holds an extra reference to the - * cgroup's dentry and cgroup removal proceeds regardless of css - * refs. On the last put of each css, whenever that may be, the - * extra dentry ref is put so that dentry destruction happens only - * after all css's are released. + * Mark @cgrp dead. This prevents further task migration and child + * creation by disabling cgroup_lock_live_group(). Note that + * CGRP_DEAD assertion is depended upon by cgroup_next_sibling() to + * resume iteration after dropping RCU read lock. See + * cgroup_next_sibling() for details. */ - for_each_subsys(cgrp->root, ss) - css_put(cgrp->subsys[ss->subsys_id]); + set_bit(CGRP_DEAD, &cgrp->flags); + /* CGRP_DEAD is set, remove from ->release_list for the last time */ raw_spin_lock(&release_list_lock); if (!list_empty(&cgrp->release_list)) list_del_init(&cgrp->release_list); raw_spin_unlock(&release_list_lock); - /* delete this cgroup from parent->children */ - list_del_rcu(&cgrp->sibling); - list_del_init(&cgrp->allcg_node); - + /* + * Remove @cgrp directory. The removal puts the base ref but we + * aren't quite done with @cgrp yet, so hold onto it. + */ dget(d); cgroup_d_remove_dir(d); - dput(d); - - set_bit(CGRP_RELEASABLE, &parent->flags); - check_for_release(parent); /* * Unregister events and notify userspace. @@ -4321,6 +4560,53 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) spin_unlock(&cgrp->event_list_lock); return 0; +}; + +/** + * cgroup_offline_fn - the second step of cgroup destruction + * @work: cgroup->destroy_free_work + * + * This function is invoked from a work item for a cgroup which is being + * destroyed after the percpu refcnts of all css's are guaranteed to be + * seen as killed on all CPUs, and performs the rest of destruction. This + * is the second step of destruction described in the comment above + * cgroup_destroy_locked(). + */ +static void cgroup_offline_fn(struct work_struct *work) +{ + struct cgroup *cgrp = container_of(work, struct cgroup, destroy_work); + struct cgroup *parent = cgrp->parent; + struct dentry *d = cgrp->dentry; + struct cgroup_subsys *ss; + + mutex_lock(&cgroup_mutex); + + /* + * css_tryget() is guaranteed to fail now. Tell subsystems to + * initate destruction. + */ + for_each_root_subsys(cgrp->root, ss) + offline_css(ss, cgrp); + + /* + * Put the css refs from cgroup_destroy_locked(). Each css holds + * an extra reference to the cgroup's dentry and cgroup removal + * proceeds regardless of css refs. On the last put of each css, + * whenever that may be, the extra dentry ref is put so that dentry + * destruction happens only after all css's are released. + */ + for_each_root_subsys(cgrp->root, ss) + css_put(cgrp->subsys[ss->subsys_id]); + + /* delete this cgroup from parent->children */ + list_del_rcu(&cgrp->sibling); + + dput(d); + + set_bit(CGRP_RELEASABLE, &parent->flags); + check_for_release(parent); + + mutex_unlock(&cgroup_mutex); } static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) @@ -4360,12 +4646,12 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) cgroup_init_cftsets(ss); /* Create the top cgroup state for this subsystem */ - list_add(&ss->sibling, &rootnode.subsys_list); - ss->root = &rootnode; - css = ss->css_alloc(dummytop); + list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); + ss->root = &cgroup_dummy_root; + css = ss->css_alloc(cgroup_dummy_top); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); - init_cgroup_css(css, ss, dummytop); + init_cgroup_css(css, ss, cgroup_dummy_top); /* Update the init_css_set to contain a subsys * pointer to this state - since the subsystem is @@ -4380,7 +4666,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) * need to invoke fork callbacks here. */ BUG_ON(!list_empty(&init_task.tasks)); - BUG_ON(online_css(ss, dummytop)); + BUG_ON(online_css(ss, cgroup_dummy_top)); mutex_unlock(&cgroup_mutex); @@ -4403,7 +4689,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) struct cgroup_subsys_state *css; int i, ret; struct hlist_node *tmp; - struct css_set *cg; + struct css_set *cset; unsigned long key; /* check name and function validity */ @@ -4426,7 +4712,7 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) */ if (ss->module == NULL) { /* a sanity check */ - BUG_ON(subsys[ss->subsys_id] != ss); + BUG_ON(cgroup_subsys[ss->subsys_id] != ss); return 0; } @@ -4434,26 +4720,26 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) cgroup_init_cftsets(ss); mutex_lock(&cgroup_mutex); - subsys[ss->subsys_id] = ss; + cgroup_subsys[ss->subsys_id] = ss; /* * no ss->css_alloc seems to need anything important in the ss - * struct, so this can happen first (i.e. before the rootnode + * struct, so this can happen first (i.e. before the dummy root * attachment). */ - css = ss->css_alloc(dummytop); + css = ss->css_alloc(cgroup_dummy_top); if (IS_ERR(css)) { - /* failure case - need to deassign the subsys[] slot. */ - subsys[ss->subsys_id] = NULL; + /* failure case - need to deassign the cgroup_subsys[] slot. */ + cgroup_subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); return PTR_ERR(css); } - list_add(&ss->sibling, &rootnode.subsys_list); - ss->root = &rootnode; + list_add(&ss->sibling, &cgroup_dummy_root.subsys_list); + ss->root = &cgroup_dummy_root; /* our new subsystem will be attached to the dummy hierarchy. */ - init_cgroup_css(css, ss, dummytop); + init_cgroup_css(css, ss, cgroup_dummy_top); /* init_idr must be after init_cgroup_css because it sets css->id. */ if (ss->use_id) { ret = cgroup_init_idr(ss, css); @@ -4470,21 +4756,21 @@ int __init_or_module cgroup_load_subsys(struct cgroup_subsys *ss) * this is all done under the css_set_lock. */ write_lock(&css_set_lock); - hash_for_each_safe(css_set_table, i, tmp, cg, hlist) { + hash_for_each_safe(css_set_table, i, tmp, cset, hlist) { /* skip entries that we already rehashed */ - if (cg->subsys[ss->subsys_id]) + if (cset->subsys[ss->subsys_id]) continue; /* remove existing entry */ - hash_del(&cg->hlist); + hash_del(&cset->hlist); /* set new value */ - cg->subsys[ss->subsys_id] = css; + cset->subsys[ss->subsys_id] = css; /* recompute hash and restore entry */ - key = css_set_hash(cg->subsys); - hash_add(css_set_table, &cg->hlist, key); + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); } write_unlock(&css_set_lock); - ret = online_css(ss, dummytop); + ret = online_css(ss, cgroup_dummy_top); if (ret) goto err_unload; @@ -4510,7 +4796,7 @@ EXPORT_SYMBOL_GPL(cgroup_load_subsys); */ void cgroup_unload_subsys(struct cgroup_subsys *ss) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; BUG_ON(ss->module == NULL); @@ -4519,45 +4805,46 @@ void cgroup_unload_subsys(struct cgroup_subsys *ss) * try_module_get in parse_cgroupfs_options should ensure that it * doesn't start being used while we're killing it off. */ - BUG_ON(ss->root != &rootnode); + BUG_ON(ss->root != &cgroup_dummy_root); mutex_lock(&cgroup_mutex); - offline_css(ss, dummytop); + offline_css(ss, cgroup_dummy_top); if (ss->use_id) idr_destroy(&ss->idr); /* deassign the subsys_id */ - subsys[ss->subsys_id] = NULL; + cgroup_subsys[ss->subsys_id] = NULL; - /* remove subsystem from rootnode's list of subsystems */ + /* remove subsystem from the dummy root's list of subsystems */ list_del_init(&ss->sibling); /* - * disentangle the css from all css_sets attached to the dummytop. as - * in loading, we need to pay our respects to the hashtable gods. + * disentangle the css from all css_sets attached to the dummy + * top. as in loading, we need to pay our respects to the hashtable + * gods. */ write_lock(&css_set_lock); - list_for_each_entry(link, &dummytop->css_sets, cgrp_link_list) { - struct css_set *cg = link->cg; + list_for_each_entry(link, &cgroup_dummy_top->cset_links, cset_link) { + struct css_set *cset = link->cset; unsigned long key; - hash_del(&cg->hlist); - cg->subsys[ss->subsys_id] = NULL; - key = css_set_hash(cg->subsys); - hash_add(css_set_table, &cg->hlist, key); + hash_del(&cset->hlist); + cset->subsys[ss->subsys_id] = NULL; + key = css_set_hash(cset->subsys); + hash_add(css_set_table, &cset->hlist, key); } write_unlock(&css_set_lock); /* - * remove subsystem's css from the dummytop and free it - need to - * free before marking as null because ss->css_free needs the - * cgrp->subsys pointer to find their state. note that this also - * takes care of freeing the css_id. + * remove subsystem's css from the cgroup_dummy_top and free it - + * need to free before marking as null because ss->css_free needs + * the cgrp->subsys pointer to find their state. note that this + * also takes care of freeing the css_id. */ - ss->css_free(dummytop); - dummytop->subsys[ss->subsys_id] = NULL; + ss->css_free(cgroup_dummy_top); + cgroup_dummy_top->subsys[ss->subsys_id] = NULL; mutex_unlock(&cgroup_mutex); } @@ -4571,30 +4858,25 @@ EXPORT_SYMBOL_GPL(cgroup_unload_subsys); */ int __init cgroup_init_early(void) { + struct cgroup_subsys *ss; int i; + atomic_set(&init_css_set.refcount, 1); - INIT_LIST_HEAD(&init_css_set.cg_links); + INIT_LIST_HEAD(&init_css_set.cgrp_links); INIT_LIST_HEAD(&init_css_set.tasks); INIT_HLIST_NODE(&init_css_set.hlist); css_set_count = 1; - init_cgroup_root(&rootnode); - root_count = 1; - init_task.cgroups = &init_css_set; - - init_css_set_link.cg = &init_css_set; - init_css_set_link.cgrp = dummytop; - list_add(&init_css_set_link.cgrp_link_list, - &rootnode.top_cgroup.css_sets); - list_add(&init_css_set_link.cg_link_list, - &init_css_set.cg_links); - - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - - /* at bootup time, we don't worry about modular subsystems */ - if (!ss || ss->module) - continue; + init_cgroup_root(&cgroup_dummy_root); + cgroup_root_count = 1; + RCU_INIT_POINTER(init_task.cgroups, &init_css_set); + init_cgrp_cset_link.cset = &init_css_set; + init_cgrp_cset_link.cgrp = cgroup_dummy_top; + list_add(&init_cgrp_cset_link.cset_link, &cgroup_dummy_top->cset_links); + list_add(&init_cgrp_cset_link.cgrp_link, &init_css_set.cgrp_links); + + /* at bootup time, we don't worry about modular subsystems */ + for_each_builtin_subsys(ss, i) { BUG_ON(!ss->name); BUG_ON(strlen(ss->name) > MAX_CGROUP_TYPE_NAMELEN); BUG_ON(!ss->css_alloc); @@ -4619,30 +4901,33 @@ int __init cgroup_init_early(void) */ int __init cgroup_init(void) { - int err; - int i; + struct cgroup_subsys *ss; unsigned long key; + int i, err; err = bdi_init(&cgroup_backing_dev_info); if (err) return err; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - - /* at bootup time, we don't worry about modular subsystems */ - if (!ss || ss->module) - continue; + for_each_builtin_subsys(ss, i) { if (!ss->early_init) cgroup_init_subsys(ss); if (ss->use_id) cgroup_init_idr(ss, init_css_set.subsys[ss->subsys_id]); } + /* allocate id for the dummy hierarchy */ + mutex_lock(&cgroup_mutex); + mutex_lock(&cgroup_root_mutex); + /* Add init_css_set to the hash table */ key = css_set_hash(init_css_set.subsys); hash_add(css_set_table, &init_css_set.hlist, key); - BUG_ON(!init_root_id(&rootnode)); + + BUG_ON(cgroup_init_root_id(&cgroup_dummy_root, 0, 1)); + + mutex_unlock(&cgroup_root_mutex); + mutex_unlock(&cgroup_mutex); cgroup_kobj = kobject_create_and_add("cgroup", fs_kobj); if (!cgroup_kobj) { @@ -4707,7 +4992,7 @@ int proc_cgroup_show(struct seq_file *m, void *v) int count = 0; seq_printf(m, "%d:", root->hierarchy_id); - for_each_subsys(root, ss) + for_each_root_subsys(root, ss) seq_printf(m, "%s%s", count++ ? "," : "", ss->name); if (strlen(root->name)) seq_printf(m, "%sname=%s", count ? "," : "", @@ -4733,6 +5018,7 @@ out: /* Display information about each subsystem and each hierarchy */ static int proc_cgroupstats_show(struct seq_file *m, void *v) { + struct cgroup_subsys *ss; int i; seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\tenabled\n"); @@ -4742,14 +5028,12 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) * subsys/hierarchy state. */ mutex_lock(&cgroup_mutex); - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - if (ss == NULL) - continue; + + for_each_subsys(ss, i) seq_printf(m, "%s\t%d\t%d\t%d\n", ss->name, ss->root->hierarchy_id, ss->root->number_of_cgroups, !ss->disabled); - } + mutex_unlock(&cgroup_mutex); return 0; } @@ -4785,8 +5069,8 @@ static const struct file_operations proc_cgroupstats_operations = { void cgroup_fork(struct task_struct *child) { task_lock(current); + get_css_set(task_css_set(current)); child->cgroups = current->cgroups; - get_css_set(child->cgroups); task_unlock(current); INIT_LIST_HEAD(&child->cg_list); } @@ -4803,6 +5087,7 @@ void cgroup_fork(struct task_struct *child) */ void cgroup_post_fork(struct task_struct *child) { + struct cgroup_subsys *ss; int i; /* @@ -4820,7 +5105,7 @@ void cgroup_post_fork(struct task_struct *child) write_lock(&css_set_lock); task_lock(child); if (list_empty(&child->cg_list)) - list_add(&child->cg_list, &child->cgroups->tasks); + list_add(&child->cg_list, &task_css_set(child)->tasks); task_unlock(child); write_unlock(&css_set_lock); } @@ -4839,12 +5124,9 @@ void cgroup_post_fork(struct task_struct *child) * of the array can be freed at module unload, so we * can't touch that. */ - for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - + for_each_builtin_subsys(ss, i) if (ss->fork) ss->fork(child); - } } } @@ -4885,7 +5167,8 @@ void cgroup_post_fork(struct task_struct *child) */ void cgroup_exit(struct task_struct *tsk, int run_callbacks) { - struct css_set *cg; + struct cgroup_subsys *ss; + struct css_set *cset; int i; /* @@ -4902,36 +5185,32 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) /* Reassign the task to the init_css_set. */ task_lock(tsk); - cg = tsk->cgroups; - tsk->cgroups = &init_css_set; + cset = task_css_set(tsk); + RCU_INIT_POINTER(tsk->cgroups, &init_css_set); if (run_callbacks && need_forkexit_callback) { /* * fork/exit callbacks are supported only for builtin * subsystems, see cgroup_post_fork() for details. */ - for (i = 0; i < CGROUP_BUILTIN_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - + for_each_builtin_subsys(ss, i) { if (ss->exit) { - struct cgroup *old_cgrp = - rcu_dereference_raw(cg->subsys[i])->cgroup; + struct cgroup *old_cgrp = cset->subsys[i]->cgroup; struct cgroup *cgrp = task_cgroup(tsk, i); + ss->exit(cgrp, old_cgrp, tsk); } } } task_unlock(tsk); - put_css_set_taskexit(cg); + put_css_set_taskexit(cset); } static void check_for_release(struct cgroup *cgrp) { - /* All of these checks rely on RCU to keep the cgroup - * structure alive */ if (cgroup_is_releasable(cgrp) && - !atomic_read(&cgrp->count) && list_empty(&cgrp->children)) { + list_empty(&cgrp->cset_links) && list_empty(&cgrp->children)) { /* * Control Group is currently removeable. If it's not * already queued for a userspace notification, queue @@ -4940,7 +5219,7 @@ static void check_for_release(struct cgroup *cgrp) int need_schedule_work = 0; raw_spin_lock(&release_list_lock); - if (!cgroup_is_removed(cgrp) && + if (!cgroup_is_dead(cgrp) && list_empty(&cgrp->release_list)) { list_add(&cgrp->release_list, &release_list); need_schedule_work = 1; @@ -4951,34 +5230,6 @@ static void check_for_release(struct cgroup *cgrp) } } -/* Caller must verify that the css is not for root cgroup */ -bool __css_tryget(struct cgroup_subsys_state *css) -{ - while (true) { - int t, v; - - v = css_refcnt(css); - t = atomic_cmpxchg(&css->refcnt, v, v + 1); - if (likely(t == v)) - return true; - else if (t < 0) - return false; - cpu_relax(); - } -} -EXPORT_SYMBOL_GPL(__css_tryget); - -/* Caller must verify that the css is not for root cgroup */ -void __css_put(struct cgroup_subsys_state *css) -{ - int v; - - v = css_unbias_refcnt(atomic_dec_return(&css->refcnt)); - if (v == 0) - schedule_work(&css->dput_work); -} -EXPORT_SYMBOL_GPL(__css_put); - /* * Notify userspace when a cgroup is released, by running the * configured release agent with the name of the cgroup (path @@ -5053,23 +5304,19 @@ static void cgroup_release_agent(struct work_struct *work) static int __init cgroup_disable(char *str) { - int i; + struct cgroup_subsys *ss; char *token; + int i; while ((token = strsep(&str, ",")) != NULL) { if (!*token) continue; - for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup_subsys *ss = subsys[i]; - - /* - * cgroup_disable, being at boot time, can't - * know about module subsystems, so we don't - * worry about them. - */ - if (!ss || ss->module) - continue; + /* + * cgroup_disable, being at boot time, can't know about + * module subsystems, so we don't worry about them. + */ + for_each_builtin_subsys(ss, i) { if (!strcmp(token, ss->name)) { ss->disabled = 1; printk(KERN_INFO "Disabling %s control group" @@ -5086,9 +5333,7 @@ __setup("cgroup_disable=", cgroup_disable); * Functons for CSS ID. */ -/* - *To get ID other than 0, this should be called when !cgroup_is_removed(). - */ +/* to get ID other than 0, this should be called when !cgroup_is_dead() */ unsigned short css_id(struct cgroup_subsys_state *css) { struct css_id *cssid; @@ -5098,7 +5343,7 @@ unsigned short css_id(struct cgroup_subsys_state *css) * on this or this is under rcu_read_lock(). Once css->id is allocated, * it's unchanged until freed. */ - cssid = rcu_dereference_check(css->id, css_refcnt(css)); + cssid = rcu_dereference_raw(css->id); if (cssid) return cssid->id; @@ -5106,18 +5351,6 @@ unsigned short css_id(struct cgroup_subsys_state *css) } EXPORT_SYMBOL_GPL(css_id); -unsigned short css_depth(struct cgroup_subsys_state *css) -{ - struct css_id *cssid; - - cssid = rcu_dereference_check(css->id, css_refcnt(css)); - - if (cssid) - return cssid->depth; - return 0; -} -EXPORT_SYMBOL_GPL(css_depth); - /** * css_is_ancestor - test "root" css is an ancestor of "child" * @child: the css to be tested. @@ -5152,7 +5385,8 @@ bool css_is_ancestor(struct cgroup_subsys_state *child, void free_css_id(struct cgroup_subsys *ss, struct cgroup_subsys_state *css) { - struct css_id *id = css->id; + struct css_id *id = rcu_dereference_protected(css->id, true); + /* When this is called before css_id initialization, id can be NULL */ if (!id) return; @@ -5218,8 +5452,8 @@ static int __init_or_module cgroup_init_idr(struct cgroup_subsys *ss, return PTR_ERR(newid); newid->stack[0] = newid->id; - newid->css = rootcss; - rootcss->id = newid; + RCU_INIT_POINTER(newid->css, rootcss); + RCU_INIT_POINTER(rootcss->id, newid); return 0; } @@ -5233,7 +5467,7 @@ static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, subsys_id = ss->subsys_id; parent_css = parent->subsys[subsys_id]; child_css = child->subsys[subsys_id]; - parent_id = parent_css->id; + parent_id = rcu_dereference_protected(parent_css->id, true); depth = parent_id->depth + 1; child_id = get_new_cssid(ss, depth); @@ -5298,7 +5532,7 @@ struct cgroup_subsys_state *cgroup_css_from_dir(struct file *f, int id) } #ifdef CONFIG_CGROUP_DEBUG -static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cont) +static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cgrp) { struct cgroup_subsys_state *css = kzalloc(sizeof(*css), GFP_KERNEL); @@ -5308,48 +5542,43 @@ static struct cgroup_subsys_state *debug_css_alloc(struct cgroup *cont) return css; } -static void debug_css_free(struct cgroup *cont) -{ - kfree(cont->subsys[debug_subsys_id]); -} - -static u64 cgroup_refcount_read(struct cgroup *cont, struct cftype *cft) +static void debug_css_free(struct cgroup *cgrp) { - return atomic_read(&cont->count); + kfree(cgrp->subsys[debug_subsys_id]); } -static u64 debug_taskcount_read(struct cgroup *cont, struct cftype *cft) +static u64 debug_taskcount_read(struct cgroup *cgrp, struct cftype *cft) { - return cgroup_task_count(cont); + return cgroup_task_count(cgrp); } -static u64 current_css_set_read(struct cgroup *cont, struct cftype *cft) +static u64 current_css_set_read(struct cgroup *cgrp, struct cftype *cft) { return (u64)(unsigned long)current->cgroups; } -static u64 current_css_set_refcount_read(struct cgroup *cont, - struct cftype *cft) +static u64 current_css_set_refcount_read(struct cgroup *cgrp, + struct cftype *cft) { u64 count; rcu_read_lock(); - count = atomic_read(¤t->cgroups->refcount); + count = atomic_read(&task_css_set(current)->refcount); rcu_read_unlock(); return count; } -static int current_css_set_cg_links_read(struct cgroup *cont, +static int current_css_set_cg_links_read(struct cgroup *cgrp, struct cftype *cft, struct seq_file *seq) { - struct cg_cgroup_link *link; - struct css_set *cg; + struct cgrp_cset_link *link; + struct css_set *cset; read_lock(&css_set_lock); rcu_read_lock(); - cg = rcu_dereference(current->cgroups); - list_for_each_entry(link, &cg->cg_links, cg_link_list) { + cset = rcu_dereference(current->cgroups); + list_for_each_entry(link, &cset->cgrp_links, cgrp_link) { struct cgroup *c = link->cgrp; const char *name; @@ -5366,19 +5595,19 @@ static int current_css_set_cg_links_read(struct cgroup *cont, } #define MAX_TASKS_SHOWN_PER_CSS 25 -static int cgroup_css_links_read(struct cgroup *cont, +static int cgroup_css_links_read(struct cgroup *cgrp, struct cftype *cft, struct seq_file *seq) { - struct cg_cgroup_link *link; + struct cgrp_cset_link *link; read_lock(&css_set_lock); - list_for_each_entry(link, &cont->css_sets, cgrp_link_list) { - struct css_set *cg = link->cg; + list_for_each_entry(link, &cgrp->cset_links, cset_link) { + struct css_set *cset = link->cset; struct task_struct *task; int count = 0; - seq_printf(seq, "css_set %p\n", cg); - list_for_each_entry(task, &cg->tasks, cg_list) { + seq_printf(seq, "css_set %p\n", cset); + list_for_each_entry(task, &cset->tasks, cg_list) { if (count++ > MAX_TASKS_SHOWN_PER_CSS) { seq_puts(seq, " ...\n"); break; @@ -5399,10 +5628,6 @@ static u64 releasable_read(struct cgroup *cgrp, struct cftype *cft) static struct cftype debug_files[] = { { - .name = "cgroup_refcount", - .read_u64 = cgroup_refcount_read, - }, - { .name = "taskcount", .read_u64 = debug_taskcount_read, }, diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c index 65349f07b878..383f8231e436 100644 --- a/kernel/context_tracking.c +++ b/kernel/context_tracking.c @@ -15,7 +15,6 @@ */ #include <linux/context_tracking.h> -#include <linux/kvm_host.h> #include <linux/rcupdate.h> #include <linux/sched.h> #include <linux/hardirq.h> @@ -71,6 +70,46 @@ void user_enter(void) local_irq_restore(flags); } +#ifdef CONFIG_PREEMPT +/** + * preempt_schedule_context - preempt_schedule called by tracing + * + * The tracing infrastructure uses preempt_enable_notrace to prevent + * recursion and tracing preempt enabling caused by the tracing + * infrastructure itself. But as tracing can happen in areas coming + * from userspace or just about to enter userspace, a preempt enable + * can occur before user_exit() is called. This will cause the scheduler + * to be called when the system is still in usermode. + * + * To prevent this, the preempt_enable_notrace will use this function + * instead of preempt_schedule() to exit user context if needed before + * calling the scheduler. + */ +void __sched notrace preempt_schedule_context(void) +{ + struct thread_info *ti = current_thread_info(); + enum ctx_state prev_ctx; + + if (likely(ti->preempt_count || irqs_disabled())) + return; + + /* + * Need to disable preemption in case user_exit() is traced + * and the tracer calls preempt_enable_notrace() causing + * an infinite recursion. + */ + preempt_disable_notrace(); + prev_ctx = exception_enter(); + preempt_enable_no_resched_notrace(); + + preempt_schedule(); + + preempt_disable_notrace(); + exception_exit(prev_ctx); + preempt_enable_notrace(); +} +EXPORT_SYMBOL_GPL(preempt_schedule_context); +#endif /* CONFIG_PREEMPT */ /** * user_exit - Inform the context tracking that the CPU is diff --git a/kernel/cpu.c b/kernel/cpu.c index b5e4ab2d427e..198a38883e64 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -133,6 +133,27 @@ static void cpu_hotplug_done(void) mutex_unlock(&cpu_hotplug.lock); } +/* + * Wait for currently running CPU hotplug operations to complete (if any) and + * disable future CPU hotplug (from sysfs). The 'cpu_add_remove_lock' protects + * the 'cpu_hotplug_disabled' flag. The same lock is also acquired by the + * hotplug path before performing hotplug operations. So acquiring that lock + * guarantees mutual exclusion from any currently running hotplug operations. + */ +void cpu_hotplug_disable(void) +{ + cpu_maps_update_begin(); + cpu_hotplug_disabled = 1; + cpu_maps_update_done(); +} + +void cpu_hotplug_enable(void) +{ + cpu_maps_update_begin(); + cpu_hotplug_disabled = 0; + cpu_maps_update_done(); +} + #else /* #if CONFIG_HOTPLUG_CPU */ static void cpu_hotplug_begin(void) {} static void cpu_hotplug_done(void) {} @@ -541,36 +562,6 @@ static int __init alloc_frozen_cpus(void) core_initcall(alloc_frozen_cpus); /* - * Prevent regular CPU hotplug from racing with the freezer, by disabling CPU - * hotplug when tasks are about to be frozen. Also, don't allow the freezer - * to continue until any currently running CPU hotplug operation gets - * completed. - * To modify the 'cpu_hotplug_disabled' flag, we need to acquire the - * 'cpu_add_remove_lock'. And this same lock is also taken by the regular - * CPU hotplug path and released only after it is complete. Thus, we - * (and hence the freezer) will block here until any currently running CPU - * hotplug operation gets completed. - */ -void cpu_hotplug_disable_before_freeze(void) -{ - cpu_maps_update_begin(); - cpu_hotplug_disabled = 1; - cpu_maps_update_done(); -} - - -/* - * When tasks have been thawed, re-enable regular CPU hotplug (which had been - * disabled while beginning to freeze tasks). - */ -void cpu_hotplug_enable_after_thaw(void) -{ - cpu_maps_update_begin(); - cpu_hotplug_disabled = 0; - cpu_maps_update_done(); -} - -/* * When callbacks for CPU hotplug notifications are being executed, we must * ensure that the state of the system with respect to the tasks being frozen * or not, as reported by the notification, remains unchanged *throughout the @@ -589,12 +580,12 @@ cpu_hotplug_pm_callback(struct notifier_block *nb, case PM_SUSPEND_PREPARE: case PM_HIBERNATION_PREPARE: - cpu_hotplug_disable_before_freeze(); + cpu_hotplug_disable(); break; case PM_POST_SUSPEND: case PM_POST_HIBERNATION: - cpu_hotplug_enable_after_thaw(); + cpu_hotplug_enable(); break; default: diff --git a/kernel/cpu/idle.c b/kernel/cpu/idle.c index d5585f5e038e..e695c0a0bcb5 100644 --- a/kernel/cpu/idle.c +++ b/kernel/cpu/idle.c @@ -5,6 +5,7 @@ #include <linux/cpu.h> #include <linux/tick.h> #include <linux/mm.h> +#include <linux/stackprotector.h> #include <asm/tlb.h> @@ -58,6 +59,7 @@ void __weak arch_cpu_idle_dead(void) { } void __weak arch_cpu_idle(void) { cpu_idle_force_poll = 1; + local_irq_enable(); } /* @@ -112,6 +114,21 @@ static void cpu_idle_loop(void) void cpu_startup_entry(enum cpuhp_state state) { + /* + * This #ifdef needs to die, but it's too late in the cycle to + * make this generic (arm and sh have never invoked the canary + * init for the non boot cpus!). Will be fixed in 3.11 + */ +#ifdef CONFIG_X86 + /* + * If we're the non-boot CPU, nothing set the stack canary up + * for us. The boot CPU already has it initialized but no harm + * in doing it again. This is a good place for updating it, as + * we wont ever return from this function (so the invalid + * canaries already on the stack wont ever trigger). + */ + boot_init_stack_canary(); +#endif current_set_polling(); arch_cpu_idle_prepare(); cpu_idle_loop(); diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 64b3f791bbe5..e5657788fedd 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -59,6 +59,7 @@ #include <linux/mutex.h> #include <linux/workqueue.h> #include <linux/cgroup.h> +#include <linux/wait.h> /* * Tracks how many cpusets are currently defined in system. @@ -87,6 +88,18 @@ struct cpuset { cpumask_var_t cpus_allowed; /* CPUs allowed to tasks in cpuset */ nodemask_t mems_allowed; /* Memory Nodes allowed to tasks */ + /* + * This is old Memory Nodes tasks took on. + * + * - top_cpuset.old_mems_allowed is initialized to mems_allowed. + * - A new cpuset's old_mems_allowed is initialized when some + * task is moved into it. + * - old_mems_allowed is used in cpuset_migrate_mm() when we change + * cpuset.mems_allowed and have tasks' nodemask updated, and + * then old_mems_allowed is updated to mems_allowed. + */ + nodemask_t old_mems_allowed; + struct fmeter fmeter; /* memory_pressure filter */ /* @@ -100,14 +113,12 @@ struct cpuset { /* for custom sched domain */ int relax_domain_level; - - struct work_struct hotplug_work; }; /* Retrieve the cpuset for a cgroup */ -static inline struct cpuset *cgroup_cs(struct cgroup *cont) +static inline struct cpuset *cgroup_cs(struct cgroup *cgrp) { - return container_of(cgroup_subsys_state(cont, cpuset_subsys_id), + return container_of(cgroup_subsys_state(cgrp, cpuset_subsys_id), struct cpuset, css); } @@ -267,14 +278,11 @@ static DEFINE_MUTEX(callback_mutex); /* * CPU / memory hotplug is handled asynchronously. */ -static struct workqueue_struct *cpuset_propagate_hotplug_wq; - static void cpuset_hotplug_workfn(struct work_struct *work); -static void cpuset_propagate_hotplug_workfn(struct work_struct *work); -static void schedule_cpuset_propagate_hotplug(struct cpuset *cs); - static DECLARE_WORK(cpuset_hotplug_work, cpuset_hotplug_workfn); +static DECLARE_WAIT_QUEUE_HEAD(cpuset_attach_wq); + /* * This is ugly, but preserves the userspace API for existing cpuset * users. If someone tries to mount the "cpuset" filesystem, we @@ -304,53 +312,38 @@ static struct file_system_type cpuset_fs_type = { /* * Return in pmask the portion of a cpusets's cpus_allowed that * are online. If none are online, walk up the cpuset hierarchy - * until we find one that does have some online cpus. If we get - * all the way to the top and still haven't found any online cpus, - * return cpu_online_mask. Or if passed a NULL cs from an exit'ing - * task, return cpu_online_mask. + * until we find one that does have some online cpus. The top + * cpuset always has some cpus online. * * One way or another, we guarantee to return some non-empty subset * of cpu_online_mask. * * Call with callback_mutex held. */ - static void guarantee_online_cpus(const struct cpuset *cs, struct cpumask *pmask) { - while (cs && !cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) + while (!cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) cs = parent_cs(cs); - if (cs) - cpumask_and(pmask, cs->cpus_allowed, cpu_online_mask); - else - cpumask_copy(pmask, cpu_online_mask); - BUG_ON(!cpumask_intersects(pmask, cpu_online_mask)); + cpumask_and(pmask, cs->cpus_allowed, cpu_online_mask); } /* * Return in *pmask the portion of a cpusets's mems_allowed that * are online, with memory. If none are online with memory, walk * up the cpuset hierarchy until we find one that does have some - * online mems. If we get all the way to the top and still haven't - * found any online mems, return node_states[N_MEMORY]. + * online mems. The top cpuset always has some mems online. * * One way or another, we guarantee to return some non-empty subset * of node_states[N_MEMORY]. * * Call with callback_mutex held. */ - static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) { - while (cs && !nodes_intersects(cs->mems_allowed, - node_states[N_MEMORY])) + while (!nodes_intersects(cs->mems_allowed, node_states[N_MEMORY])) cs = parent_cs(cs); - if (cs) - nodes_and(*pmask, cs->mems_allowed, - node_states[N_MEMORY]); - else - *pmask = node_states[N_MEMORY]; - BUG_ON(!nodes_intersects(*pmask, node_states[N_MEMORY])); + nodes_and(*pmask, cs->mems_allowed, node_states[N_MEMORY]); } /* @@ -440,7 +433,7 @@ static void free_trial_cpuset(struct cpuset *trial) static int validate_change(const struct cpuset *cur, const struct cpuset *trial) { - struct cgroup *cont; + struct cgroup *cgrp; struct cpuset *c, *par; int ret; @@ -448,7 +441,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) /* Each of our child cpusets must be a subset of us */ ret = -EBUSY; - cpuset_for_each_child(c, cont, cur) + cpuset_for_each_child(c, cgrp, cur) if (!is_cpuset_subset(c, trial)) goto out; @@ -469,7 +462,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) * overlap */ ret = -EINVAL; - cpuset_for_each_child(c, cont, par) { + cpuset_for_each_child(c, cgrp, par) { if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && c != cur && cpumask_intersects(trial->cpus_allowed, c->cpus_allowed)) @@ -486,7 +479,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) */ ret = -ENOSPC; if ((cgroup_task_count(cur->css.cgroup) || cur->attach_in_progress) && - (cpumask_empty(trial->cpus_allowed) || + (cpumask_empty(trial->cpus_allowed) && nodes_empty(trial->mems_allowed))) goto out; @@ -540,7 +533,7 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, * This function builds a partial partition of the systems CPUs * A 'partial partition' is a set of non-overlapping subsets whose * union is a subset of that set. - * The output of this function needs to be passed to kernel/sched.c + * The output of this function needs to be passed to kernel/sched/core.c * partition_sched_domains() routine, which will rebuild the scheduler's * load balancing domains (sched domains) as specified by that partial * partition. @@ -569,7 +562,7 @@ static void update_domain_attr_tree(struct sched_domain_attr *dattr, * is a subset of one of these domains, while there are as * many such domains as possible, each as small as possible. * doms - Conversion of 'csa' to an array of cpumasks, for passing to - * the kernel/sched.c routine partition_sched_domains() in a + * the kernel/sched/core.c routine partition_sched_domains() in a * convenient format, that can be easily compared to the prior * value to determine what partition elements (sched domains) * were changed (added or removed.) @@ -798,21 +791,43 @@ void rebuild_sched_domains(void) mutex_unlock(&cpuset_mutex); } -/** - * cpuset_test_cpumask - test a task's cpus_allowed versus its cpuset's - * @tsk: task to test - * @scan: struct cgroup_scanner contained in its struct cpuset_hotplug_scanner +/* + * effective_cpumask_cpuset - return nearest ancestor with non-empty cpus + * @cs: the cpuset in interest * - * Call with cpuset_mutex held. May take callback_mutex during call. - * Called for each task in a cgroup by cgroup_scan_tasks(). - * Return nonzero if this tasks's cpus_allowed mask should be changed (in other - * words, if its mask is not equal to its cpuset's mask). + * A cpuset's effective cpumask is the cpumask of the nearest ancestor + * with non-empty cpus. We use effective cpumask whenever: + * - we update tasks' cpus_allowed. (they take on the ancestor's cpumask + * if the cpuset they reside in has no cpus) + * - we want to retrieve task_cs(tsk)'s cpus_allowed. + * + * Called with cpuset_mutex held. cpuset_cpus_allowed_fallback() is an + * exception. See comments there. */ -static int cpuset_test_cpumask(struct task_struct *tsk, - struct cgroup_scanner *scan) +static struct cpuset *effective_cpumask_cpuset(struct cpuset *cs) { - return !cpumask_equal(&tsk->cpus_allowed, - (cgroup_cs(scan->cg))->cpus_allowed); + while (cpumask_empty(cs->cpus_allowed)) + cs = parent_cs(cs); + return cs; +} + +/* + * effective_nodemask_cpuset - return nearest ancestor with non-empty mems + * @cs: the cpuset in interest + * + * A cpuset's effective nodemask is the nodemask of the nearest ancestor + * with non-empty memss. We use effective nodemask whenever: + * - we update tasks' mems_allowed. (they take on the ancestor's nodemask + * if the cpuset they reside in has no mems) + * - we want to retrieve task_cs(tsk)'s mems_allowed. + * + * Called with cpuset_mutex held. + */ +static struct cpuset *effective_nodemask_cpuset(struct cpuset *cs) +{ + while (nodes_empty(cs->mems_allowed)) + cs = parent_cs(cs); + return cs; } /** @@ -829,7 +844,10 @@ static int cpuset_test_cpumask(struct task_struct *tsk, static void cpuset_change_cpumask(struct task_struct *tsk, struct cgroup_scanner *scan) { - set_cpus_allowed_ptr(tsk, ((cgroup_cs(scan->cg))->cpus_allowed)); + struct cpuset *cpus_cs; + + cpus_cs = effective_cpumask_cpuset(cgroup_cs(scan->cg)); + set_cpus_allowed_ptr(tsk, cpus_cs->cpus_allowed); } /** @@ -850,12 +868,51 @@ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) struct cgroup_scanner scan; scan.cg = cs->css.cgroup; - scan.test_task = cpuset_test_cpumask; + scan.test_task = NULL; scan.process_task = cpuset_change_cpumask; scan.heap = heap; cgroup_scan_tasks(&scan); } +/* + * update_tasks_cpumask_hier - Update the cpumasks of tasks in the hierarchy. + * @root_cs: the root cpuset of the hierarchy + * @update_root: update root cpuset or not? + * @heap: the heap used by cgroup_scan_tasks() + * + * This will update cpumasks of tasks in @root_cs and all other empty cpusets + * which take on cpumask of @root_cs. + * + * Called with cpuset_mutex held + */ +static void update_tasks_cpumask_hier(struct cpuset *root_cs, + bool update_root, struct ptr_heap *heap) +{ + struct cpuset *cp; + struct cgroup *pos_cgrp; + + if (update_root) + update_tasks_cpumask(root_cs, heap); + + rcu_read_lock(); + cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + /* skip the whole subtree if @cp have some CPU */ + if (!cpumask_empty(cp->cpus_allowed)) { + pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + continue; + } + if (!css_tryget(&cp->css)) + continue; + rcu_read_unlock(); + + update_tasks_cpumask(cp, heap); + + rcu_read_lock(); + css_put(&cp->css); + } + rcu_read_unlock(); +} + /** * update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it * @cs: the cpuset to consider @@ -888,14 +945,15 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, if (!cpumask_subset(trialcs->cpus_allowed, cpu_active_mask)) return -EINVAL; } - retval = validate_change(cs, trialcs); - if (retval < 0) - return retval; /* Nothing to do if the cpus didn't change */ if (cpumask_equal(cs->cpus_allowed, trialcs->cpus_allowed)) return 0; + retval = validate_change(cs, trialcs); + if (retval < 0) + return retval; + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); if (retval) return retval; @@ -906,11 +964,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed); mutex_unlock(&callback_mutex); - /* - * Scan tasks in the cpuset, and update the cpumasks of any - * that need an update. - */ - update_tasks_cpumask(cs, &heap); + update_tasks_cpumask_hier(cs, true, &heap); heap_free(&heap); @@ -943,12 +997,14 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, const nodemask_t *to) { struct task_struct *tsk = current; + struct cpuset *mems_cs; tsk->mems_allowed = *to; do_migrate_pages(mm, from, to, MPOL_MF_MOVE_ALL); - guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed); + mems_cs = effective_nodemask_cpuset(task_cs(tsk)); + guarantee_online_mems(mems_cs, &tsk->mems_allowed); } /* @@ -1007,16 +1063,12 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, static void cpuset_change_nodemask(struct task_struct *p, struct cgroup_scanner *scan) { + struct cpuset *cs = cgroup_cs(scan->cg); struct mm_struct *mm; - struct cpuset *cs; int migrate; - const nodemask_t *oldmem = scan->data; - static nodemask_t newmems; /* protected by cpuset_mutex */ - - cs = cgroup_cs(scan->cg); - guarantee_online_mems(cs, &newmems); + nodemask_t *newmems = scan->data; - cpuset_change_task_nodemask(p, &newmems); + cpuset_change_task_nodemask(p, newmems); mm = get_task_mm(p); if (!mm) @@ -1026,7 +1078,7 @@ static void cpuset_change_nodemask(struct task_struct *p, mpol_rebind_mm(mm, &cs->mems_allowed); if (migrate) - cpuset_migrate_mm(mm, oldmem, &cs->mems_allowed); + cpuset_migrate_mm(mm, &cs->old_mems_allowed, newmems); mmput(mm); } @@ -1035,25 +1087,27 @@ static void *cpuset_being_rebound; /** * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed - * @oldmem: old mems_allowed of cpuset cs * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() * * Called with cpuset_mutex held * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 * if @heap != NULL. */ -static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, - struct ptr_heap *heap) +static void update_tasks_nodemask(struct cpuset *cs, struct ptr_heap *heap) { + static nodemask_t newmems; /* protected by cpuset_mutex */ struct cgroup_scanner scan; + struct cpuset *mems_cs = effective_nodemask_cpuset(cs); cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ + guarantee_online_mems(mems_cs, &newmems); + scan.cg = cs->css.cgroup; scan.test_task = NULL; scan.process_task = cpuset_change_nodemask; scan.heap = heap; - scan.data = (nodemask_t *)oldmem; + scan.data = &newmems; /* * The mpol_rebind_mm() call takes mmap_sem, which we couldn't @@ -1067,11 +1121,56 @@ static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, */ cgroup_scan_tasks(&scan); + /* + * All the tasks' nodemasks have been updated, update + * cs->old_mems_allowed. + */ + cs->old_mems_allowed = newmems; + /* We're done rebinding vmas to this cpuset's new mems_allowed. */ cpuset_being_rebound = NULL; } /* + * update_tasks_nodemask_hier - Update the nodemasks of tasks in the hierarchy. + * @cs: the root cpuset of the hierarchy + * @update_root: update the root cpuset or not? + * @heap: the heap used by cgroup_scan_tasks() + * + * This will update nodemasks of tasks in @root_cs and all other empty cpusets + * which take on nodemask of @root_cs. + * + * Called with cpuset_mutex held + */ +static void update_tasks_nodemask_hier(struct cpuset *root_cs, + bool update_root, struct ptr_heap *heap) +{ + struct cpuset *cp; + struct cgroup *pos_cgrp; + + if (update_root) + update_tasks_nodemask(root_cs, heap); + + rcu_read_lock(); + cpuset_for_each_descendant_pre(cp, pos_cgrp, root_cs) { + /* skip the whole subtree if @cp have some CPU */ + if (!nodes_empty(cp->mems_allowed)) { + pos_cgrp = cgroup_rightmost_descendant(pos_cgrp); + continue; + } + if (!css_tryget(&cp->css)) + continue; + rcu_read_unlock(); + + update_tasks_nodemask(cp, heap); + + rcu_read_lock(); + css_put(&cp->css); + } + rcu_read_unlock(); +} + +/* * Handle user request to change the 'mems' memory placement * of a cpuset. Needs to validate the request, update the * cpusets mems_allowed, and for each task in the cpuset, @@ -1087,13 +1186,9 @@ static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, const char *buf) { - NODEMASK_ALLOC(nodemask_t, oldmem, GFP_KERNEL); int retval; struct ptr_heap heap; - if (!oldmem) - return -ENOMEM; - /* * top_cpuset.mems_allowed tracks node_stats[N_MEMORY]; * it's read-only @@ -1122,8 +1217,8 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, goto done; } } - *oldmem = cs->mems_allowed; - if (nodes_equal(*oldmem, trialcs->mems_allowed)) { + + if (nodes_equal(cs->mems_allowed, trialcs->mems_allowed)) { retval = 0; /* Too easy - nothing to do */ goto done; } @@ -1139,11 +1234,10 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, cs->mems_allowed = trialcs->mems_allowed; mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, oldmem, &heap); + update_tasks_nodemask_hier(cs, true, &heap); heap_free(&heap); done: - NODEMASK_FREE(oldmem); return retval; } @@ -1372,8 +1466,13 @@ static int cpuset_can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) mutex_lock(&cpuset_mutex); + /* + * We allow to move tasks into an empty cpuset if sane_behavior + * flag is set. + */ ret = -ENOSPC; - if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) + if (!cgroup_sane_behavior(cgrp) && + (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))) goto out_unlock; cgroup_taskset_for_each(task, cgrp, tset) { @@ -1422,8 +1521,7 @@ static cpumask_var_t cpus_attach; static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) { - /* static bufs protected by cpuset_mutex */ - static nodemask_t cpuset_attach_nodemask_from; + /* static buf protected by cpuset_mutex */ static nodemask_t cpuset_attach_nodemask_to; struct mm_struct *mm; struct task_struct *task; @@ -1431,6 +1529,8 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) struct cgroup *oldcgrp = cgroup_taskset_cur_cgroup(tset); struct cpuset *cs = cgroup_cs(cgrp); struct cpuset *oldcs = cgroup_cs(oldcgrp); + struct cpuset *cpus_cs = effective_cpumask_cpuset(cs); + struct cpuset *mems_cs = effective_nodemask_cpuset(cs); mutex_lock(&cpuset_mutex); @@ -1438,9 +1538,9 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) if (cs == &top_cpuset) cpumask_copy(cpus_attach, cpu_possible_mask); else - guarantee_online_cpus(cs, cpus_attach); + guarantee_online_cpus(cpus_cs, cpus_attach); - guarantee_online_mems(cs, &cpuset_attach_nodemask_to); + guarantee_online_mems(mems_cs, &cpuset_attach_nodemask_to); cgroup_taskset_for_each(task, cgrp, tset) { /* @@ -1457,26 +1557,32 @@ static void cpuset_attach(struct cgroup *cgrp, struct cgroup_taskset *tset) * Change mm, possibly for multiple threads in a threadgroup. This is * expensive and may sleep. */ - cpuset_attach_nodemask_from = oldcs->mems_allowed; cpuset_attach_nodemask_to = cs->mems_allowed; mm = get_task_mm(leader); if (mm) { + struct cpuset *mems_oldcs = effective_nodemask_cpuset(oldcs); + mpol_rebind_mm(mm, &cpuset_attach_nodemask_to); - if (is_memory_migrate(cs)) - cpuset_migrate_mm(mm, &cpuset_attach_nodemask_from, + + /* + * old_mems_allowed is the same with mems_allowed here, except + * if this task is being moved automatically due to hotplug. + * In that case @mems_allowed has been updated and is empty, + * so @old_mems_allowed is the right nodesets that we migrate + * mm from. + */ + if (is_memory_migrate(cs)) { + cpuset_migrate_mm(mm, &mems_oldcs->old_mems_allowed, &cpuset_attach_nodemask_to); + } mmput(mm); } - cs->attach_in_progress--; + cs->old_mems_allowed = cpuset_attach_nodemask_to; - /* - * We may have raced with CPU/memory hotunplug. Trigger hotplug - * propagation if @cs doesn't have any CPU or memory. It will move - * the newly added tasks to the nearest parent which can execute. - */ - if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) - schedule_cpuset_propagate_hotplug(cs); + cs->attach_in_progress--; + if (!cs->attach_in_progress) + wake_up(&cpuset_attach_wq); mutex_unlock(&cpuset_mutex); } @@ -1588,13 +1694,8 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, * resources, wait for the previously scheduled operations before * proceeding, so that we don't end up keep removing tasks added * after execution capability is restored. - * - * Flushing cpuset_hotplug_work is enough to synchronize against - * hotplug hanlding; however, cpuset_attach() may schedule - * propagation work directly. Flush the workqueue too. */ flush_work(&cpuset_hotplug_work); - flush_workqueue(cpuset_propagate_hotplug_wq); mutex_lock(&cpuset_mutex); if (!is_cpuset_online(cs)) @@ -1658,13 +1759,13 @@ static size_t cpuset_sprintf_memlist(char *page, struct cpuset *cs) return count; } -static ssize_t cpuset_common_file_read(struct cgroup *cont, +static ssize_t cpuset_common_file_read(struct cgroup *cgrp, struct cftype *cft, struct file *file, char __user *buf, size_t nbytes, loff_t *ppos) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; char *page; ssize_t retval = 0; @@ -1694,9 +1795,9 @@ out: return retval; } -static u64 cpuset_read_u64(struct cgroup *cont, struct cftype *cft) +static u64 cpuset_read_u64(struct cgroup *cgrp, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; switch (type) { case FILE_CPU_EXCLUSIVE: @@ -1725,9 +1826,9 @@ static u64 cpuset_read_u64(struct cgroup *cont, struct cftype *cft) return 0; } -static s64 cpuset_read_s64(struct cgroup *cont, struct cftype *cft) +static s64 cpuset_read_s64(struct cgroup *cgrp, struct cftype *cft) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); cpuset_filetype_t type = cft->private; switch (type) { case FILE_SCHED_RELAX_DOMAIN_LEVEL: @@ -1839,14 +1940,14 @@ static struct cftype files[] = { /* * cpuset_css_alloc - allocate a cpuset css - * cont: control group that the new cpuset will be part of + * cgrp: control group that the new cpuset will be part of */ -static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont) +static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cgrp) { struct cpuset *cs; - if (!cont->parent) + if (!cgrp->parent) return &top_cpuset.css; cs = kzalloc(sizeof(*cs), GFP_KERNEL); @@ -1861,7 +1962,6 @@ static struct cgroup_subsys_state *cpuset_css_alloc(struct cgroup *cont) cpumask_clear(cs->cpus_allowed); nodes_clear(cs->mems_allowed); fmeter_init(&cs->fmeter); - INIT_WORK(&cs->hotplug_work, cpuset_propagate_hotplug_workfn); cs->relax_domain_level = -1; return &cs->css; @@ -1942,9 +2042,9 @@ static void cpuset_css_offline(struct cgroup *cgrp) * will call rebuild_sched_domains_locked(). */ -static void cpuset_css_free(struct cgroup *cont) +static void cpuset_css_free(struct cgroup *cgrp) { - struct cpuset *cs = cgroup_cs(cont); + struct cpuset *cs = cgroup_cs(cgrp); free_cpumask_var(cs->cpus_allowed); kfree(cs); @@ -2024,41 +2124,64 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs) } /** - * cpuset_propagate_hotplug_workfn - propagate CPU/memory hotplug to a cpuset + * cpuset_hotplug_update_tasks - update tasks in a cpuset for hotunplug * @cs: cpuset in interest * * Compare @cs's cpu and mem masks against top_cpuset and if some have gone * offline, update @cs accordingly. If @cs ends up with no CPU or memory, * all its tasks are moved to the nearest ancestor with both resources. */ -static void cpuset_propagate_hotplug_workfn(struct work_struct *work) +static void cpuset_hotplug_update_tasks(struct cpuset *cs) { static cpumask_t off_cpus; - static nodemask_t off_mems, tmp_mems; - struct cpuset *cs = container_of(work, struct cpuset, hotplug_work); + static nodemask_t off_mems; bool is_empty; + bool sane = cgroup_sane_behavior(cs->css.cgroup); + +retry: + wait_event(cpuset_attach_wq, cs->attach_in_progress == 0); mutex_lock(&cpuset_mutex); + /* + * We have raced with task attaching. We wait until attaching + * is finished, so we won't attach a task to an empty cpuset. + */ + if (cs->attach_in_progress) { + mutex_unlock(&cpuset_mutex); + goto retry; + } + cpumask_andnot(&off_cpus, cs->cpus_allowed, top_cpuset.cpus_allowed); nodes_andnot(off_mems, cs->mems_allowed, top_cpuset.mems_allowed); - /* remove offline cpus from @cs */ - if (!cpumask_empty(&off_cpus)) { - mutex_lock(&callback_mutex); - cpumask_andnot(cs->cpus_allowed, cs->cpus_allowed, &off_cpus); - mutex_unlock(&callback_mutex); + mutex_lock(&callback_mutex); + cpumask_andnot(cs->cpus_allowed, cs->cpus_allowed, &off_cpus); + mutex_unlock(&callback_mutex); + + /* + * If sane_behavior flag is set, we need to update tasks' cpumask + * for empty cpuset to take on ancestor's cpumask. Otherwise, don't + * call update_tasks_cpumask() if the cpuset becomes empty, as + * the tasks in it will be migrated to an ancestor. + */ + if ((sane && cpumask_empty(cs->cpus_allowed)) || + (!cpumask_empty(&off_cpus) && !cpumask_empty(cs->cpus_allowed))) update_tasks_cpumask(cs, NULL); - } - /* remove offline mems from @cs */ - if (!nodes_empty(off_mems)) { - tmp_mems = cs->mems_allowed; - mutex_lock(&callback_mutex); - nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems); - mutex_unlock(&callback_mutex); - update_tasks_nodemask(cs, &tmp_mems, NULL); - } + mutex_lock(&callback_mutex); + nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems); + mutex_unlock(&callback_mutex); + + /* + * If sane_behavior flag is set, we need to update tasks' nodemask + * for empty cpuset to take on ancestor's nodemask. Otherwise, don't + * call update_tasks_nodemask() if the cpuset becomes empty, as + * the tasks in it will be migratd to an ancestor. + */ + if ((sane && nodes_empty(cs->mems_allowed)) || + (!nodes_empty(off_mems) && !nodes_empty(cs->mems_allowed))) + update_tasks_nodemask(cs, NULL); is_empty = cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed); @@ -2066,40 +2189,14 @@ static void cpuset_propagate_hotplug_workfn(struct work_struct *work) mutex_unlock(&cpuset_mutex); /* - * If @cs became empty, move tasks to the nearest ancestor with - * execution resources. This is full cgroup operation which will + * If sane_behavior flag is set, we'll keep tasks in empty cpusets. + * + * Otherwise move tasks to the nearest ancestor with execution + * resources. This is full cgroup operation which will * also call back into cpuset. Should be done outside any lock. */ - if (is_empty) + if (!sane && is_empty) remove_tasks_in_empty_cpuset(cs); - - /* the following may free @cs, should be the last operation */ - css_put(&cs->css); -} - -/** - * schedule_cpuset_propagate_hotplug - schedule hotplug propagation to a cpuset - * @cs: cpuset of interest - * - * Schedule cpuset_propagate_hotplug_workfn() which will update CPU and - * memory masks according to top_cpuset. - */ -static void schedule_cpuset_propagate_hotplug(struct cpuset *cs) -{ - /* - * Pin @cs. The refcnt will be released when the work item - * finishes executing. - */ - if (!css_tryget(&cs->css)) - return; - - /* - * Queue @cs->hotplug_work. If already pending, lose the css ref. - * cpuset_propagate_hotplug_wq is ordered and propagation will - * happen in the order this function is called. - */ - if (!queue_work(cpuset_propagate_hotplug_wq, &cs->hotplug_work)) - css_put(&cs->css); } /** @@ -2112,18 +2209,17 @@ static void schedule_cpuset_propagate_hotplug(struct cpuset *cs) * actively using CPU hotplug but making no active use of cpusets. * * Non-root cpusets are only affected by offlining. If any CPUs or memory - * nodes have been taken down, cpuset_propagate_hotplug() is invoked on all - * descendants. + * nodes have been taken down, cpuset_hotplug_update_tasks() is invoked on + * all descendants. * * Note that CPU offlining during suspend is ignored. We don't modify * cpusets across suspend/resume cycles at all. */ static void cpuset_hotplug_workfn(struct work_struct *work) { - static cpumask_t new_cpus, tmp_cpus; - static nodemask_t new_mems, tmp_mems; + static cpumask_t new_cpus; + static nodemask_t new_mems; bool cpus_updated, mems_updated; - bool cpus_offlined, mems_offlined; mutex_lock(&cpuset_mutex); @@ -2132,12 +2228,7 @@ static void cpuset_hotplug_workfn(struct work_struct *work) new_mems = node_states[N_MEMORY]; cpus_updated = !cpumask_equal(top_cpuset.cpus_allowed, &new_cpus); - cpus_offlined = cpumask_andnot(&tmp_cpus, top_cpuset.cpus_allowed, - &new_cpus); - mems_updated = !nodes_equal(top_cpuset.mems_allowed, new_mems); - nodes_andnot(tmp_mems, top_cpuset.mems_allowed, new_mems); - mems_offlined = !nodes_empty(tmp_mems); /* synchronize cpus_allowed to cpu_active_mask */ if (cpus_updated) { @@ -2149,28 +2240,32 @@ static void cpuset_hotplug_workfn(struct work_struct *work) /* synchronize mems_allowed to N_MEMORY */ if (mems_updated) { - tmp_mems = top_cpuset.mems_allowed; mutex_lock(&callback_mutex); top_cpuset.mems_allowed = new_mems; mutex_unlock(&callback_mutex); - update_tasks_nodemask(&top_cpuset, &tmp_mems, NULL); + update_tasks_nodemask(&top_cpuset, NULL); } - /* if cpus or mems went down, we need to propagate to descendants */ - if (cpus_offlined || mems_offlined) { + mutex_unlock(&cpuset_mutex); + + /* if cpus or mems changed, we need to propagate to descendants */ + if (cpus_updated || mems_updated) { struct cpuset *cs; struct cgroup *pos_cgrp; rcu_read_lock(); - cpuset_for_each_descendant_pre(cs, pos_cgrp, &top_cpuset) - schedule_cpuset_propagate_hotplug(cs); - rcu_read_unlock(); - } + cpuset_for_each_descendant_pre(cs, pos_cgrp, &top_cpuset) { + if (!css_tryget(&cs->css)) + continue; + rcu_read_unlock(); - mutex_unlock(&cpuset_mutex); + cpuset_hotplug_update_tasks(cs); - /* wait for propagations to finish */ - flush_workqueue(cpuset_propagate_hotplug_wq); + rcu_read_lock(); + css_put(&cs->css); + } + rcu_read_unlock(); + } /* rebuild sched domains if cpus_allowed has changed */ if (cpus_updated) @@ -2219,12 +2314,9 @@ void __init cpuset_init_smp(void) { cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); top_cpuset.mems_allowed = node_states[N_MEMORY]; + top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; register_hotmemory_notifier(&cpuset_track_online_nodes_nb); - - cpuset_propagate_hotplug_wq = - alloc_ordered_workqueue("cpuset_hotplug", 0); - BUG_ON(!cpuset_propagate_hotplug_wq); } /** @@ -2240,21 +2332,23 @@ void __init cpuset_init_smp(void) void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) { + struct cpuset *cpus_cs; + mutex_lock(&callback_mutex); task_lock(tsk); - guarantee_online_cpus(task_cs(tsk), pmask); + cpus_cs = effective_cpumask_cpuset(task_cs(tsk)); + guarantee_online_cpus(cpus_cs, pmask); task_unlock(tsk); mutex_unlock(&callback_mutex); } void cpuset_cpus_allowed_fallback(struct task_struct *tsk) { - const struct cpuset *cs; + const struct cpuset *cpus_cs; rcu_read_lock(); - cs = task_cs(tsk); - if (cs) - do_set_cpus_allowed(tsk, cs->cpus_allowed); + cpus_cs = effective_cpumask_cpuset(task_cs(tsk)); + do_set_cpus_allowed(tsk, cpus_cs->cpus_allowed); rcu_read_unlock(); /* @@ -2293,11 +2387,13 @@ void cpuset_init_current_mems_allowed(void) nodemask_t cpuset_mems_allowed(struct task_struct *tsk) { + struct cpuset *mems_cs; nodemask_t mask; mutex_lock(&callback_mutex); task_lock(tsk); - guarantee_online_mems(task_cs(tsk), &mask); + mems_cs = effective_nodemask_cpuset(task_cs(tsk)); + guarantee_online_mems(mems_cs, &mask); task_unlock(tsk); mutex_unlock(&callback_mutex); diff --git a/kernel/events/core.c b/kernel/events/core.c index 9dc297faf7c0..1db3af933704 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -165,10 +165,28 @@ int sysctl_perf_event_mlock __read_mostly = 512 + (PAGE_SIZE / 1024); /* 'free' /* * max perf event sample rate */ -#define DEFAULT_MAX_SAMPLE_RATE 100000 -int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE; -static int max_samples_per_tick __read_mostly = - DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ); +#define DEFAULT_MAX_SAMPLE_RATE 100000 +#define DEFAULT_SAMPLE_PERIOD_NS (NSEC_PER_SEC / DEFAULT_MAX_SAMPLE_RATE) +#define DEFAULT_CPU_TIME_MAX_PERCENT 25 + +int sysctl_perf_event_sample_rate __read_mostly = DEFAULT_MAX_SAMPLE_RATE; + +static int max_samples_per_tick __read_mostly = DIV_ROUND_UP(DEFAULT_MAX_SAMPLE_RATE, HZ); +static int perf_sample_period_ns __read_mostly = DEFAULT_SAMPLE_PERIOD_NS; + +static atomic_t perf_sample_allowed_ns __read_mostly = + ATOMIC_INIT( DEFAULT_SAMPLE_PERIOD_NS * DEFAULT_CPU_TIME_MAX_PERCENT / 100); + +void update_perf_cpu_limits(void) +{ + u64 tmp = perf_sample_period_ns; + + tmp *= sysctl_perf_cpu_time_max_percent; + tmp = do_div(tmp, 100); + atomic_set(&perf_sample_allowed_ns, tmp); +} + +static int perf_rotate_context(struct perf_cpu_context *cpuctx); int perf_proc_update_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, @@ -180,10 +198,78 @@ int perf_proc_update_handler(struct ctl_table *table, int write, return ret; max_samples_per_tick = DIV_ROUND_UP(sysctl_perf_event_sample_rate, HZ); + perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate; + update_perf_cpu_limits(); return 0; } +int sysctl_perf_cpu_time_max_percent __read_mostly = DEFAULT_CPU_TIME_MAX_PERCENT; + +int perf_cpu_time_max_percent_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos) +{ + int ret = proc_dointvec(table, write, buffer, lenp, ppos); + + if (ret || !write) + return ret; + + update_perf_cpu_limits(); + + return 0; +} + +/* + * perf samples are done in some very critical code paths (NMIs). + * If they take too much CPU time, the system can lock up and not + * get any real work done. This will drop the sample rate when + * we detect that events are taking too long. + */ +#define NR_ACCUMULATED_SAMPLES 128 +DEFINE_PER_CPU(u64, running_sample_length); + +void perf_sample_event_took(u64 sample_len_ns) +{ + u64 avg_local_sample_len; + u64 local_samples_len = __get_cpu_var(running_sample_length); + + if (atomic_read(&perf_sample_allowed_ns) == 0) + return; + + /* decay the counter by 1 average sample */ + local_samples_len = __get_cpu_var(running_sample_length); + local_samples_len -= local_samples_len/NR_ACCUMULATED_SAMPLES; + local_samples_len += sample_len_ns; + __get_cpu_var(running_sample_length) = local_samples_len; + + /* + * note: this will be biased artifically low until we have + * seen NR_ACCUMULATED_SAMPLES. Doing it this way keeps us + * from having to maintain a count. + */ + avg_local_sample_len = local_samples_len/NR_ACCUMULATED_SAMPLES; + + if (avg_local_sample_len <= atomic_read(&perf_sample_allowed_ns)) + return; + + if (max_samples_per_tick <= 1) + return; + + max_samples_per_tick = DIV_ROUND_UP(max_samples_per_tick, 2); + sysctl_perf_event_sample_rate = max_samples_per_tick * HZ; + perf_sample_period_ns = NSEC_PER_SEC / sysctl_perf_event_sample_rate; + + printk_ratelimited(KERN_WARNING + "perf samples too long (%lld > %d), lowering " + "kernel.perf_event_max_sample_rate to %d\n", + avg_local_sample_len, + atomic_read(&perf_sample_allowed_ns), + sysctl_perf_event_sample_rate); + + update_perf_cpu_limits(); +} + static atomic64_t perf_event_id; static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, @@ -196,9 +282,6 @@ static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, static void update_context_time(struct perf_event_context *ctx); static u64 perf_event_time(struct perf_event *event); -static void ring_buffer_attach(struct perf_event *event, - struct ring_buffer *rb); - void __weak perf_event_print_debug(void) { } extern __weak const char *perf_pmu_name(void) @@ -658,6 +741,106 @@ perf_cgroup_mark_enabled(struct perf_event *event, } #endif +/* + * set default to be dependent on timer tick just + * like original code + */ +#define PERF_CPU_HRTIMER (1000 / HZ) +/* + * function must be called with interrupts disbled + */ +static enum hrtimer_restart perf_cpu_hrtimer_handler(struct hrtimer *hr) +{ + struct perf_cpu_context *cpuctx; + enum hrtimer_restart ret = HRTIMER_NORESTART; + int rotations = 0; + + WARN_ON(!irqs_disabled()); + + cpuctx = container_of(hr, struct perf_cpu_context, hrtimer); + + rotations = perf_rotate_context(cpuctx); + + /* + * arm timer if needed + */ + if (rotations) { + hrtimer_forward_now(hr, cpuctx->hrtimer_interval); + ret = HRTIMER_RESTART; + } + + return ret; +} + +/* CPU is going down */ +void perf_cpu_hrtimer_cancel(int cpu) +{ + struct perf_cpu_context *cpuctx; + struct pmu *pmu; + unsigned long flags; + + if (WARN_ON(cpu != smp_processor_id())) + return; + + local_irq_save(flags); + + rcu_read_lock(); + + list_for_each_entry_rcu(pmu, &pmus, entry) { + cpuctx = this_cpu_ptr(pmu->pmu_cpu_context); + + if (pmu->task_ctx_nr == perf_sw_context) + continue; + + hrtimer_cancel(&cpuctx->hrtimer); + } + + rcu_read_unlock(); + + local_irq_restore(flags); +} + +static void __perf_cpu_hrtimer_init(struct perf_cpu_context *cpuctx, int cpu) +{ + struct hrtimer *hr = &cpuctx->hrtimer; + struct pmu *pmu = cpuctx->ctx.pmu; + int timer; + + /* no multiplexing needed for SW PMU */ + if (pmu->task_ctx_nr == perf_sw_context) + return; + + /* + * check default is sane, if not set then force to + * default interval (1/tick) + */ + timer = pmu->hrtimer_interval_ms; + if (timer < 1) + timer = pmu->hrtimer_interval_ms = PERF_CPU_HRTIMER; + + cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer); + + hrtimer_init(hr, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); + hr->function = perf_cpu_hrtimer_handler; +} + +static void perf_cpu_hrtimer_restart(struct perf_cpu_context *cpuctx) +{ + struct hrtimer *hr = &cpuctx->hrtimer; + struct pmu *pmu = cpuctx->ctx.pmu; + + /* not for SW PMU */ + if (pmu->task_ctx_nr == perf_sw_context) + return; + + if (hrtimer_active(hr)) + return; + + if (!hrtimer_callback_running(hr)) + __hrtimer_start_range_ns(hr, cpuctx->hrtimer_interval, + 0, HRTIMER_MODE_REL_PINNED, 0); +} + void perf_pmu_disable(struct pmu *pmu) { int *count = this_cpu_ptr(pmu->pmu_disable_count); @@ -1506,6 +1689,7 @@ group_sched_in(struct perf_event *group_event, if (event_sched_in(group_event, cpuctx, ctx)) { pmu->cancel_txn(pmu); + perf_cpu_hrtimer_restart(cpuctx); return -EAGAIN; } @@ -1552,6 +1736,8 @@ group_error: pmu->cancel_txn(pmu); + perf_cpu_hrtimer_restart(cpuctx); + return -EAGAIN; } @@ -1807,8 +1993,10 @@ static int __perf_event_enable(void *info) * If this event can't go on and it's part of a * group, then the whole group has to come off. */ - if (leader != event) + if (leader != event) { group_sched_out(leader, cpuctx, ctx); + perf_cpu_hrtimer_restart(cpuctx); + } if (leader->attr.pinned) { update_group_times(leader); leader->state = PERF_EVENT_STATE_ERROR; @@ -2555,7 +2743,7 @@ static void rotate_ctx(struct perf_event_context *ctx) * because they're strictly cpu affine and rotate_start is called with IRQs * disabled, while rotate_context is called from IRQ context. */ -static void perf_rotate_context(struct perf_cpu_context *cpuctx) +static int perf_rotate_context(struct perf_cpu_context *cpuctx) { struct perf_event_context *ctx = NULL; int rotate = 0, remove = 1; @@ -2594,6 +2782,8 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx) done: if (remove) list_del_init(&cpuctx->rotation_list); + + return rotate; } #ifdef CONFIG_NO_HZ_FULL @@ -2625,10 +2815,6 @@ void perf_event_task_tick(void) ctx = cpuctx->task_ctx; if (ctx) perf_adjust_freq_unthr_context(ctx, throttled); - - if (cpuctx->jiffies_interval == 1 || - !(jiffies % cpuctx->jiffies_interval)) - perf_rotate_context(cpuctx); } } @@ -2918,6 +3104,7 @@ static void free_event_rcu(struct rcu_head *head) } static void ring_buffer_put(struct ring_buffer *rb); +static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb); static void free_event(struct perf_event *event) { @@ -2942,15 +3129,30 @@ static void free_event(struct perf_event *event) if (has_branch_stack(event)) { static_key_slow_dec_deferred(&perf_sched_events); /* is system-wide event */ - if (!(event->attach_state & PERF_ATTACH_TASK)) + if (!(event->attach_state & PERF_ATTACH_TASK)) { atomic_dec(&per_cpu(perf_branch_stack_events, event->cpu)); + } } } if (event->rb) { - ring_buffer_put(event->rb); - event->rb = NULL; + struct ring_buffer *rb; + + /* + * Can happen when we close an event with re-directed output. + * + * Since we have a 0 refcount, perf_mmap_close() will skip + * over us; possibly making our ring_buffer_put() the last. + */ + mutex_lock(&event->mmap_mutex); + rb = event->rb; + if (rb) { + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + ring_buffer_put(rb); /* could be last */ + } + mutex_unlock(&event->mmap_mutex); } if (is_cgroup_event(event)) @@ -3188,30 +3390,13 @@ static unsigned int perf_poll(struct file *file, poll_table *wait) unsigned int events = POLL_HUP; /* - * Race between perf_event_set_output() and perf_poll(): perf_poll() - * grabs the rb reference but perf_event_set_output() overrides it. - * Here is the timeline for two threads T1, T2: - * t0: T1, rb = rcu_dereference(event->rb) - * t1: T2, old_rb = event->rb - * t2: T2, event->rb = new rb - * t3: T2, ring_buffer_detach(old_rb) - * t4: T1, ring_buffer_attach(rb1) - * t5: T1, poll_wait(event->waitq) - * - * To avoid this problem, we grab mmap_mutex in perf_poll() - * thereby ensuring that the assignment of the new ring buffer - * and the detachment of the old buffer appear atomic to perf_poll() + * Pin the event->rb by taking event->mmap_mutex; otherwise + * perf_event_set_output() can swizzle our rb and make us miss wakeups. */ mutex_lock(&event->mmap_mutex); - - rcu_read_lock(); - rb = rcu_dereference(event->rb); - if (rb) { - ring_buffer_attach(event, rb); + rb = event->rb; + if (rb) events = atomic_xchg(&rb->poll, 0); - } - rcu_read_unlock(); - mutex_unlock(&event->mmap_mutex); poll_wait(file, &event->waitq, wait); @@ -3521,16 +3706,12 @@ static void ring_buffer_attach(struct perf_event *event, return; spin_lock_irqsave(&rb->event_lock, flags); - if (!list_empty(&event->rb_entry)) - goto unlock; - - list_add(&event->rb_entry, &rb->event_list); -unlock: + if (list_empty(&event->rb_entry)) + list_add(&event->rb_entry, &rb->event_list); spin_unlock_irqrestore(&rb->event_lock, flags); } -static void ring_buffer_detach(struct perf_event *event, - struct ring_buffer *rb) +static void ring_buffer_detach(struct perf_event *event, struct ring_buffer *rb) { unsigned long flags; @@ -3549,13 +3730,10 @@ static void ring_buffer_wakeup(struct perf_event *event) rcu_read_lock(); rb = rcu_dereference(event->rb); - if (!rb) - goto unlock; - - list_for_each_entry_rcu(event, &rb->event_list, rb_entry) - wake_up_all(&event->waitq); - -unlock: + if (rb) { + list_for_each_entry_rcu(event, &rb->event_list, rb_entry) + wake_up_all(&event->waitq); + } rcu_read_unlock(); } @@ -3584,18 +3762,10 @@ static struct ring_buffer *ring_buffer_get(struct perf_event *event) static void ring_buffer_put(struct ring_buffer *rb) { - struct perf_event *event, *n; - unsigned long flags; - if (!atomic_dec_and_test(&rb->refcount)) return; - spin_lock_irqsave(&rb->event_lock, flags); - list_for_each_entry_safe(event, n, &rb->event_list, rb_entry) { - list_del_init(&event->rb_entry); - wake_up_all(&event->waitq); - } - spin_unlock_irqrestore(&rb->event_lock, flags); + WARN_ON_ONCE(!list_empty(&rb->event_list)); call_rcu(&rb->rcu_head, rb_free_rcu); } @@ -3605,26 +3775,100 @@ static void perf_mmap_open(struct vm_area_struct *vma) struct perf_event *event = vma->vm_file->private_data; atomic_inc(&event->mmap_count); + atomic_inc(&event->rb->mmap_count); } +/* + * A buffer can be mmap()ed multiple times; either directly through the same + * event, or through other events by use of perf_event_set_output(). + * + * In order to undo the VM accounting done by perf_mmap() we need to destroy + * the buffer here, where we still have a VM context. This means we need + * to detach all events redirecting to us. + */ static void perf_mmap_close(struct vm_area_struct *vma) { struct perf_event *event = vma->vm_file->private_data; - if (atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) { - unsigned long size = perf_data_size(event->rb); - struct user_struct *user = event->mmap_user; - struct ring_buffer *rb = event->rb; + struct ring_buffer *rb = event->rb; + struct user_struct *mmap_user = rb->mmap_user; + int mmap_locked = rb->mmap_locked; + unsigned long size = perf_data_size(rb); + + atomic_dec(&rb->mmap_count); + + if (!atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) + return; - atomic_long_sub((size >> PAGE_SHIFT) + 1, &user->locked_vm); - vma->vm_mm->pinned_vm -= event->mmap_locked; - rcu_assign_pointer(event->rb, NULL); - ring_buffer_detach(event, rb); + /* Detach current event from the buffer. */ + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + mutex_unlock(&event->mmap_mutex); + + /* If there's still other mmap()s of this buffer, we're done. */ + if (atomic_read(&rb->mmap_count)) { + ring_buffer_put(rb); /* can't be last */ + return; + } + + /* + * No other mmap()s, detach from all other events that might redirect + * into the now unreachable buffer. Somewhat complicated by the + * fact that rb::event_lock otherwise nests inside mmap_mutex. + */ +again: + rcu_read_lock(); + list_for_each_entry_rcu(event, &rb->event_list, rb_entry) { + if (!atomic_long_inc_not_zero(&event->refcount)) { + /* + * This event is en-route to free_event() which will + * detach it and remove it from the list. + */ + continue; + } + rcu_read_unlock(); + + mutex_lock(&event->mmap_mutex); + /* + * Check we didn't race with perf_event_set_output() which can + * swizzle the rb from under us while we were waiting to + * acquire mmap_mutex. + * + * If we find a different rb; ignore this event, a next + * iteration will no longer find it on the list. We have to + * still restart the iteration to make sure we're not now + * iterating the wrong list. + */ + if (event->rb == rb) { + rcu_assign_pointer(event->rb, NULL); + ring_buffer_detach(event, rb); + ring_buffer_put(rb); /* can't be last, we still have one */ + } mutex_unlock(&event->mmap_mutex); + put_event(event); - ring_buffer_put(rb); - free_uid(user); + /* + * Restart the iteration; either we're on the wrong list or + * destroyed its integrity by doing a deletion. + */ + goto again; } + rcu_read_unlock(); + + /* + * It could be there's still a few 0-ref events on the list; they'll + * get cleaned up by free_event() -- they'll also still have their + * ref on the rb and will free it whenever they are done with it. + * + * Aside from that, this buffer is 'fully' detached and unmapped, + * undo the VM accounting. + */ + + atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm); + vma->vm_mm->pinned_vm -= mmap_locked; + free_uid(mmap_user); + + ring_buffer_put(rb); /* could be last */ } static const struct vm_operations_struct perf_mmap_vmops = { @@ -3674,12 +3918,24 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; WARN_ON_ONCE(event->ctx->parent_ctx); +again: mutex_lock(&event->mmap_mutex); if (event->rb) { - if (event->rb->nr_pages == nr_pages) - atomic_inc(&event->rb->refcount); - else + if (event->rb->nr_pages != nr_pages) { ret = -EINVAL; + goto unlock; + } + + if (!atomic_inc_not_zero(&event->rb->mmap_count)) { + /* + * Raced against perf_mmap_close() through + * perf_event_set_output(). Try again, hope for better + * luck. + */ + mutex_unlock(&event->mmap_mutex); + goto again; + } + goto unlock; } @@ -3720,12 +3976,16 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) ret = -ENOMEM; goto unlock; } - rcu_assign_pointer(event->rb, rb); + + atomic_set(&rb->mmap_count, 1); + rb->mmap_locked = extra; + rb->mmap_user = get_current_user(); atomic_long_add(user_extra, &user->locked_vm); - event->mmap_locked = extra; - event->mmap_user = get_current_user(); - vma->vm_mm->pinned_vm += event->mmap_locked; + vma->vm_mm->pinned_vm += extra; + + ring_buffer_attach(event, rb); + rcu_assign_pointer(event->rb, rb); perf_event_update_userpage(event); @@ -3734,7 +3994,11 @@ unlock: atomic_inc(&event->mmap_count); mutex_unlock(&event->mmap_mutex); - vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP; + /* + * Since pinned accounting is per vm we cannot allow fork() to copy our + * vma. + */ + vma->vm_flags |= VM_DONTCOPY | VM_DONTEXPAND | VM_DONTDUMP; vma->vm_ops = &perf_mmap_vmops; return ret; @@ -4961,7 +5225,7 @@ static DEFINE_PER_CPU(struct swevent_htable, swevent_htable); * sign as trigger. */ -static u64 perf_swevent_set_period(struct perf_event *event) +u64 perf_swevent_set_period(struct perf_event *event) { struct hw_perf_event *hwc = &event->hw; u64 period = hwc->last_period; @@ -5904,9 +6168,56 @@ type_show(struct device *dev, struct device_attribute *attr, char *page) return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->type); } +static ssize_t +perf_event_mux_interval_ms_show(struct device *dev, + struct device_attribute *attr, + char *page) +{ + struct pmu *pmu = dev_get_drvdata(dev); + + return snprintf(page, PAGE_SIZE-1, "%d\n", pmu->hrtimer_interval_ms); +} + +static ssize_t +perf_event_mux_interval_ms_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct pmu *pmu = dev_get_drvdata(dev); + int timer, cpu, ret; + + ret = kstrtoint(buf, 0, &timer); + if (ret) + return ret; + + if (timer < 1) + return -EINVAL; + + /* same value, noting to do */ + if (timer == pmu->hrtimer_interval_ms) + return count; + + pmu->hrtimer_interval_ms = timer; + + /* update all cpuctx for this PMU */ + for_each_possible_cpu(cpu) { + struct perf_cpu_context *cpuctx; + cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu); + cpuctx->hrtimer_interval = ns_to_ktime(NSEC_PER_MSEC * timer); + + if (hrtimer_active(&cpuctx->hrtimer)) + hrtimer_forward_now(&cpuctx->hrtimer, cpuctx->hrtimer_interval); + } + + return count; +} + +#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store) + static struct device_attribute pmu_dev_attrs[] = { - __ATTR_RO(type), - __ATTR_NULL, + __ATTR_RO(type), + __ATTR_RW(perf_event_mux_interval_ms), + __ATTR_NULL, }; static int pmu_bus_running; @@ -5952,7 +6263,7 @@ free_dev: static struct lock_class_key cpuctx_mutex; static struct lock_class_key cpuctx_lock; -int perf_pmu_register(struct pmu *pmu, char *name, int type) +int perf_pmu_register(struct pmu *pmu, const char *name, int type) { int cpu, ret; @@ -6001,7 +6312,9 @@ skip_type: lockdep_set_class(&cpuctx->ctx.lock, &cpuctx_lock); cpuctx->ctx.type = cpu_context; cpuctx->ctx.pmu = pmu; - cpuctx->jiffies_interval = 1; + + __perf_cpu_hrtimer_init(cpuctx, cpu); + INIT_LIST_HEAD(&cpuctx->rotation_list); cpuctx->unique_pmu = pmu; } @@ -6327,11 +6640,6 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, if (!(mask & ~PERF_SAMPLE_BRANCH_PLM_ALL)) return -EINVAL; - /* kernel level capture: check permissions */ - if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM) - && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) - return -EACCES; - /* propagate priv level, when not set for branch */ if (!(mask & PERF_SAMPLE_BRANCH_PLM_ALL)) { @@ -6349,6 +6657,10 @@ static int perf_copy_attr(struct perf_event_attr __user *uattr, */ attr->branch_sample_type = mask; } + /* privileged levels capture (kernel, hv): check permissions */ + if ((mask & PERF_SAMPLE_BRANCH_PERM_PLM) + && perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) + return -EACCES; } if (attr->sample_type & PERF_SAMPLE_REGS_USER) { @@ -6412,6 +6724,8 @@ set: if (atomic_read(&event->mmap_count)) goto unlock; + old_rb = event->rb; + if (output_event) { /* get the rb we want to redirect to */ rb = ring_buffer_get(output_event); @@ -6419,16 +6733,28 @@ set: goto unlock; } - old_rb = event->rb; - rcu_assign_pointer(event->rb, rb); if (old_rb) ring_buffer_detach(event, old_rb); + + if (rb) + ring_buffer_attach(event, rb); + + rcu_assign_pointer(event->rb, rb); + + if (old_rb) { + ring_buffer_put(old_rb); + /* + * Since we detached before setting the new rb, so that we + * could attach the new rb, we could have missed a wakeup. + * Provide it now. + */ + wake_up_all(&event->waitq); + } + ret = 0; unlock: mutex_unlock(&event->mmap_mutex); - if (old_rb) - ring_buffer_put(old_rb); out: return ret; } @@ -7387,7 +7713,6 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) case CPU_DOWN_PREPARE: perf_event_exit_cpu(cpu); break; - default: break; } diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index a64f8aeb5c1f..1559fb0b9296 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -46,23 +46,26 @@ #include <linux/smp.h> #include <linux/hw_breakpoint.h> - - /* * Constraints data */ +struct bp_cpuinfo { + /* Number of pinned cpu breakpoints in a cpu */ + unsigned int cpu_pinned; + /* tsk_pinned[n] is the number of tasks having n+1 breakpoints */ + unsigned int *tsk_pinned; + /* Number of non-pinned cpu/task breakpoints in a cpu */ + unsigned int flexible; /* XXX: placeholder, see fetch_this_slot() */ +}; -/* Number of pinned cpu breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int, nr_cpu_bp_pinned[TYPE_MAX]); - -/* Number of pinned task breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int *, nr_task_bp_pinned[TYPE_MAX]); - -/* Number of non-pinned cpu/task breakpoints in a cpu */ -static DEFINE_PER_CPU(unsigned int, nr_bp_flexible[TYPE_MAX]); - +static DEFINE_PER_CPU(struct bp_cpuinfo, bp_cpuinfo[TYPE_MAX]); static int nr_slots[TYPE_MAX]; +static struct bp_cpuinfo *get_bp_info(int cpu, enum bp_type_idx type) +{ + return per_cpu_ptr(bp_cpuinfo + type, cpu); +} + /* Keep track of the breakpoints attached to tasks */ static LIST_HEAD(bp_task_head); @@ -96,8 +99,8 @@ static inline enum bp_type_idx find_slot_idx(struct perf_event *bp) */ static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type) { + unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned; int i; - unsigned int *tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); for (i = nr_slots[type] - 1; i >= 0; i--) { if (tsk_pinned[i] > 0) @@ -120,13 +123,20 @@ static int task_bp_pinned(int cpu, struct perf_event *bp, enum bp_type_idx type) list_for_each_entry(iter, &bp_task_head, hw.bp_list) { if (iter->hw.bp_target == tsk && find_slot_idx(iter) == type && - cpu == iter->cpu) + (iter->cpu < 0 || cpu == iter->cpu)) count += hw_breakpoint_weight(iter); } return count; } +static const struct cpumask *cpumask_of_bp(struct perf_event *bp) +{ + if (bp->cpu >= 0) + return cpumask_of(bp->cpu); + return cpu_possible_mask; +} + /* * Report the number of pinned/un-pinned breakpoints we have in * a given cpu (cpu > -1) or in all of them (cpu = -1). @@ -135,25 +145,15 @@ static void fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, enum bp_type_idx type) { - int cpu = bp->cpu; - struct task_struct *tsk = bp->hw.bp_target; - - if (cpu >= 0) { - slots->pinned = per_cpu(nr_cpu_bp_pinned[type], cpu); - if (!tsk) - slots->pinned += max_task_bp_pinned(cpu, type); - else - slots->pinned += task_bp_pinned(cpu, bp, type); - slots->flexible = per_cpu(nr_bp_flexible[type], cpu); - - return; - } + const struct cpumask *cpumask = cpumask_of_bp(bp); + int cpu; - for_each_online_cpu(cpu) { - unsigned int nr; + for_each_cpu(cpu, cpumask) { + struct bp_cpuinfo *info = get_bp_info(cpu, type); + int nr; - nr = per_cpu(nr_cpu_bp_pinned[type], cpu); - if (!tsk) + nr = info->cpu_pinned; + if (!bp->hw.bp_target) nr += max_task_bp_pinned(cpu, type); else nr += task_bp_pinned(cpu, bp, type); @@ -161,8 +161,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, if (nr > slots->pinned) slots->pinned = nr; - nr = per_cpu(nr_bp_flexible[type], cpu); - + nr = info->flexible; if (nr > slots->flexible) slots->flexible = nr; } @@ -182,29 +181,19 @@ fetch_this_slot(struct bp_busy_slots *slots, int weight) /* * Add a pinned breakpoint for the given task in our constraint table */ -static void toggle_bp_task_slot(struct perf_event *bp, int cpu, bool enable, +static void toggle_bp_task_slot(struct perf_event *bp, int cpu, enum bp_type_idx type, int weight) { - unsigned int *tsk_pinned; - int old_count = 0; - int old_idx = 0; - int idx = 0; - - old_count = task_bp_pinned(cpu, bp, type); - old_idx = old_count - 1; - idx = old_idx + weight; - - /* tsk_pinned[n] is the number of tasks having n breakpoints */ - tsk_pinned = per_cpu(nr_task_bp_pinned[type], cpu); - if (enable) { - tsk_pinned[idx]++; - if (old_count > 0) - tsk_pinned[old_idx]--; - } else { - tsk_pinned[idx]--; - if (old_count > 0) - tsk_pinned[old_idx]++; - } + unsigned int *tsk_pinned = get_bp_info(cpu, type)->tsk_pinned; + int old_idx, new_idx; + + old_idx = task_bp_pinned(cpu, bp, type) - 1; + new_idx = old_idx + weight; + + if (old_idx >= 0) + tsk_pinned[old_idx]--; + if (new_idx >= 0) + tsk_pinned[new_idx]++; } /* @@ -214,33 +203,26 @@ static void toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, int weight) { - int cpu = bp->cpu; - struct task_struct *tsk = bp->hw.bp_target; + const struct cpumask *cpumask = cpumask_of_bp(bp); + int cpu; - /* Pinned counter cpu profiling */ - if (!tsk) { + if (!enable) + weight = -weight; - if (enable) - per_cpu(nr_cpu_bp_pinned[type], bp->cpu) += weight; - else - per_cpu(nr_cpu_bp_pinned[type], bp->cpu) -= weight; + /* Pinned counter cpu profiling */ + if (!bp->hw.bp_target) { + get_bp_info(bp->cpu, type)->cpu_pinned += weight; return; } /* Pinned counter task profiling */ - - if (!enable) - list_del(&bp->hw.bp_list); - - if (cpu >= 0) { - toggle_bp_task_slot(bp, cpu, enable, type, weight); - } else { - for_each_online_cpu(cpu) - toggle_bp_task_slot(bp, cpu, enable, type, weight); - } + for_each_cpu(cpu, cpumask) + toggle_bp_task_slot(bp, cpu, type, weight); if (enable) list_add_tail(&bp->hw.bp_list, &bp_task_head); + else + list_del(&bp->hw.bp_list); } /* @@ -261,8 +243,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to a single cpu, check: * - * (per_cpu(nr_bp_flexible, cpu) || (per_cpu(nr_cpu_bp_pinned, cpu) - * + max(per_cpu(nr_task_bp_pinned, cpu)))) < HBP_NUM + * (per_cpu(info->flexible, cpu) || (per_cpu(info->cpu_pinned, cpu) + * + max(per_cpu(info->tsk_pinned, cpu)))) < HBP_NUM * * -> If there are already non-pinned counters in this cpu, it means * there is already a free slot for them. @@ -272,8 +254,8 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to every cpus, check: * - * (per_cpu(nr_bp_flexible, *) || (max(per_cpu(nr_cpu_bp_pinned, *)) - * + max(per_cpu(nr_task_bp_pinned, *)))) < HBP_NUM + * (per_cpu(info->flexible, *) || (max(per_cpu(info->cpu_pinned, *)) + * + max(per_cpu(info->tsk_pinned, *)))) < HBP_NUM * * -> This is roughly the same, except we check the number of per cpu * bp for every cpu and we keep the max one. Same for the per tasks @@ -284,16 +266,16 @@ __weak void arch_unregister_hw_breakpoint(struct perf_event *bp) * * - If attached to a single cpu, check: * - * ((per_cpu(nr_bp_flexible, cpu) > 1) + per_cpu(nr_cpu_bp_pinned, cpu) - * + max(per_cpu(nr_task_bp_pinned, cpu))) < HBP_NUM + * ((per_cpu(info->flexible, cpu) > 1) + per_cpu(info->cpu_pinned, cpu) + * + max(per_cpu(info->tsk_pinned, cpu))) < HBP_NUM * - * -> Same checks as before. But now the nr_bp_flexible, if any, must keep + * -> Same checks as before. But now the info->flexible, if any, must keep * one register at least (or they will never be fed). * * - If attached to every cpus, check: * - * ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *)) - * + max(per_cpu(nr_task_bp_pinned, *))) < HBP_NUM + * ((per_cpu(info->flexible, *) > 1) + max(per_cpu(info->cpu_pinned, *)) + * + max(per_cpu(info->tsk_pinned, *))) < HBP_NUM */ static int __reserve_bp_slot(struct perf_event *bp) { @@ -518,8 +500,8 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, void *context) { - struct perf_event * __percpu *cpu_events, **pevent, *bp; - long err; + struct perf_event * __percpu *cpu_events, *bp; + long err = 0; int cpu; cpu_events = alloc_percpu(typeof(*cpu_events)); @@ -528,31 +510,21 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, get_online_cpus(); for_each_online_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); bp = perf_event_create_kernel_counter(attr, cpu, NULL, triggered, context); - - *pevent = bp; - if (IS_ERR(bp)) { err = PTR_ERR(bp); - goto fail; + break; } - } - put_online_cpus(); - return cpu_events; - -fail: - for_each_online_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); - if (IS_ERR(*pevent)) - break; - unregister_hw_breakpoint(*pevent); + per_cpu(*cpu_events, cpu) = bp; } put_online_cpus(); - free_percpu(cpu_events); + if (likely(!err)) + return cpu_events; + + unregister_wide_hw_breakpoint(cpu_events); return (void __percpu __force *)ERR_PTR(err); } EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); @@ -564,12 +536,10 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events) { int cpu; - struct perf_event **pevent; - for_each_possible_cpu(cpu) { - pevent = per_cpu_ptr(cpu_events, cpu); - unregister_hw_breakpoint(*pevent); - } + for_each_possible_cpu(cpu) + unregister_hw_breakpoint(per_cpu(*cpu_events, cpu)); + free_percpu(cpu_events); } EXPORT_SYMBOL_GPL(unregister_wide_hw_breakpoint); @@ -612,6 +582,11 @@ static int hw_breakpoint_add(struct perf_event *bp, int flags) if (!(flags & PERF_EF_START)) bp->hw.state = PERF_HES_STOPPED; + if (is_sampling_event(bp)) { + bp->hw.last_period = bp->hw.sample_period; + perf_swevent_set_period(bp); + } + return arch_install_hw_breakpoint(bp); } @@ -650,7 +625,6 @@ static struct pmu perf_breakpoint = { int __init init_hw_breakpoint(void) { - unsigned int **task_bp_pinned; int cpu, err_cpu; int i; @@ -659,10 +633,11 @@ int __init init_hw_breakpoint(void) for_each_possible_cpu(cpu) { for (i = 0; i < TYPE_MAX; i++) { - task_bp_pinned = &per_cpu(nr_task_bp_pinned[i], cpu); - *task_bp_pinned = kzalloc(sizeof(int) * nr_slots[i], - GFP_KERNEL); - if (!*task_bp_pinned) + struct bp_cpuinfo *info = get_bp_info(cpu, i); + + info->tsk_pinned = kcalloc(nr_slots[i], sizeof(int), + GFP_KERNEL); + if (!info->tsk_pinned) goto err_alloc; } } @@ -676,7 +651,7 @@ int __init init_hw_breakpoint(void) err_alloc: for_each_possible_cpu(err_cpu) { for (i = 0; i < TYPE_MAX; i++) - kfree(per_cpu(nr_task_bp_pinned[i], err_cpu)); + kfree(get_bp_info(err_cpu, i)->tsk_pinned); if (err_cpu == cpu) break; } diff --git a/kernel/events/internal.h b/kernel/events/internal.h index eb675c4d59df..ca6599723be5 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -31,6 +31,10 @@ struct ring_buffer { spinlock_t event_lock; struct list_head event_list; + atomic_t mmap_count; + unsigned long mmap_locked; + struct user_struct *mmap_user; + struct perf_event_mmap_page *user_page; void *data_pages[0]; }; diff --git a/kernel/exit.c b/kernel/exit.c index af2eb3cbd499..fafe75d9e6f6 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -312,17 +312,6 @@ kill_orphaned_pgrp(struct task_struct *tsk, struct task_struct *parent) } } -void __set_special_pids(struct pid *pid) -{ - struct task_struct *curr = current->group_leader; - - if (task_session(curr) != pid) - change_pid(curr, PIDTYPE_SID, pid); - - if (task_pgrp(curr) != pid) - change_pid(curr, PIDTYPE_PGID, pid); -} - /* * Let kernel threads use this to say that they allow a certain signal. * Must not be used if kthread was cloned with CLONE_SIGHAND. @@ -649,7 +638,6 @@ static void exit_notify(struct task_struct *tsk, int group_dead) * jobs, send them a SIGHUP and then a SIGCONT. (POSIX 3.2.2.2) */ forget_original_parent(tsk); - exit_task_namespaces(tsk); write_lock_irq(&tasklist_lock); if (group_dead) @@ -795,6 +783,7 @@ void do_exit(long code) exit_shm(tsk); exit_files(tsk); exit_fs(tsk); + exit_task_namespaces(tsk); exit_task_work(tsk); check_stack_usage(); exit_thread(); @@ -835,7 +824,7 @@ void do_exit(long code) /* * Make sure we are holding no locks: */ - debug_check_no_locks_held(tsk); + debug_check_no_locks_held(); /* * We can do this unlocked here. The futex code uses this flag * just to verify whether the pi state cleanup has been done diff --git a/kernel/fork.c b/kernel/fork.c index 987b28a1f01b..6e6a1c11b3e5 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1121,6 +1121,12 @@ static void posix_cpu_timers_init(struct task_struct *tsk) INIT_LIST_HEAD(&tsk->cpu_timers[2]); } +static inline void +init_task_pid(struct task_struct *task, enum pid_type type, struct pid *pid) +{ + task->pids[type].pid = pid; +} + /* * This creates a new process as a copy of the old one, * but does not actually start it yet. @@ -1199,8 +1205,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, retval = -EAGAIN; if (atomic_read(&p->real_cred->user->processes) >= task_rlimit(p, RLIMIT_NPROC)) { - if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) && - p->real_cred->user != INIT_USER) + if (p->real_cred->user != INIT_USER && + !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN)) goto bad_fork_free; } current->flags &= ~PF_NPROC_EXCEEDED; @@ -1354,11 +1360,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, goto bad_fork_cleanup_io; } - p->pid = pid_nr(pid); - p->tgid = p->pid; - if (clone_flags & CLONE_THREAD) - p->tgid = current->tgid; - p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL; /* * Clear TID on mm_release()? @@ -1394,12 +1395,19 @@ static struct task_struct *copy_process(unsigned long clone_flags, clear_all_latency_tracing(p); /* ok, now we should be set up.. */ - if (clone_flags & CLONE_THREAD) + p->pid = pid_nr(pid); + if (clone_flags & CLONE_THREAD) { p->exit_signal = -1; - else if (clone_flags & CLONE_PARENT) - p->exit_signal = current->group_leader->exit_signal; - else - p->exit_signal = (clone_flags & CSIGNAL); + p->group_leader = current->group_leader; + p->tgid = current->tgid; + } else { + if (clone_flags & CLONE_PARENT) + p->exit_signal = current->group_leader->exit_signal; + else + p->exit_signal = (clone_flags & CSIGNAL); + p->group_leader = p; + p->tgid = p->pid; + } p->pdeath_signal = 0; p->exit_state = 0; @@ -1408,15 +1416,13 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->nr_dirtied_pause = 128 >> (PAGE_SHIFT - 10); p->dirty_paused_when = 0; - /* - * Ok, make it visible to the rest of the system. - * We dont wake it up yet. - */ - p->group_leader = p; INIT_LIST_HEAD(&p->thread_group); p->task_works = NULL; - /* Need tasklist lock for parent etc handling! */ + /* + * Make it visible to the rest of the system, but dont wake it up yet. + * Need tasklist lock for parent etc handling! + */ write_lock_irq(&tasklist_lock); /* CLONE_PARENT re-uses the old parent */ @@ -1446,18 +1452,14 @@ static struct task_struct *copy_process(unsigned long clone_flags, goto bad_fork_free_pid; } - if (clone_flags & CLONE_THREAD) { - current->signal->nr_threads++; - atomic_inc(¤t->signal->live); - atomic_inc(¤t->signal->sigcnt); - p->group_leader = current->group_leader; - list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group); - } - if (likely(p->pid)) { ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace); + init_task_pid(p, PIDTYPE_PID, pid); if (thread_group_leader(p)) { + init_task_pid(p, PIDTYPE_PGID, task_pgrp(current)); + init_task_pid(p, PIDTYPE_SID, task_session(current)); + if (is_child_reaper(pid)) { ns_of_pid(pid)->child_reaper = p; p->signal->flags |= SIGNAL_UNKILLABLE; @@ -1465,13 +1467,19 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->signal->leader_pid = pid; p->signal->tty = tty_kref_get(current->signal->tty); - attach_pid(p, PIDTYPE_PGID, task_pgrp(current)); - attach_pid(p, PIDTYPE_SID, task_session(current)); list_add_tail(&p->sibling, &p->real_parent->children); list_add_tail_rcu(&p->tasks, &init_task.tasks); + attach_pid(p, PIDTYPE_PGID); + attach_pid(p, PIDTYPE_SID); __this_cpu_inc(process_counts); + } else { + current->signal->nr_threads++; + atomic_inc(¤t->signal->live); + atomic_inc(¤t->signal->sigcnt); + list_add_tail_rcu(&p->thread_group, + &p->group_leader->thread_group); } - attach_pid(p, PIDTYPE_PID, pid); + attach_pid(p, PIDTYPE_PID); nr_threads++; } diff --git a/kernel/freezer.c b/kernel/freezer.c index c38893b0efba..8b2afc1c9df0 100644 --- a/kernel/freezer.c +++ b/kernel/freezer.c @@ -110,6 +110,18 @@ bool freeze_task(struct task_struct *p) { unsigned long flags; + /* + * This check can race with freezer_do_not_count, but worst case that + * will result in an extra wakeup being sent to the task. It does not + * race with freezer_count(), the barriers in freezer_count() and + * freezer_should_skip() ensure that either freezer_count() sees + * freezing == true in try_to_freeze() and freezes, or + * freezer_should_skip() sees !PF_FREEZE_SKIP and freezes the task + * normally. + */ + if (freezer_should_skip(p)) + return false; + spin_lock_irqsave(&freezer_lock, flags); if (!freezing(p) || frozen(p)) { spin_unlock_irqrestore(&freezer_lock, flags); diff --git a/kernel/futex.c b/kernel/futex.c index b26dcfc02c94..c3a1a55a5214 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -61,6 +61,8 @@ #include <linux/nsproxy.h> #include <linux/ptrace.h> #include <linux/sched/rt.h> +#include <linux/hugetlb.h> +#include <linux/freezer.h> #include <asm/futex.h> @@ -365,7 +367,7 @@ again: } else { key->both.offset |= FUT_OFF_INODE; /* inode-based key */ key->shared.inode = page_head->mapping->host; - key->shared.pgoff = page_head->index; + key->shared.pgoff = basepage_index(page); } get_futex_key_refs(key); @@ -1807,7 +1809,7 @@ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, * is no timeout, or if it has yet to expire. */ if (!timeout || timeout->task) - schedule(); + freezable_schedule(); } __set_current_state(TASK_RUNNING); } diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index fd4b13b131f8..3ee4d06c6fc2 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -47,6 +47,7 @@ #include <linux/sched/sysctl.h> #include <linux/sched/rt.h> #include <linux/timer.h> +#include <linux/freezer.h> #include <asm/uaccess.h> @@ -1545,7 +1546,7 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod t->task = NULL; if (likely(t->task)) - schedule(); + freezable_schedule(); hrtimer_cancel(&t->timer); mode = HRTIMER_MODE_ABS; diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index cbd97ce0b000..a3bb14fbe5c6 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -213,6 +213,19 @@ void irq_enable(struct irq_desc *desc) irq_state_clr_masked(desc); } +/** + * irq_disable - Mark interupt disabled + * @desc: irq descriptor which should be disabled + * + * If the chip does not implement the irq_disable callback, we + * use a lazy disable approach. That means we mark the interrupt + * disabled, but leave the hardware unmasked. That's an + * optimization because we avoid the hardware access for the + * common case where no interrupt happens after we marked it + * disabled. If an interrupt happens, then the interrupt flow + * handler masks the line at the hardware level and marks it + * pending. + */ void irq_disable(struct irq_desc *desc) { irq_state_set_disabled(desc); diff --git a/kernel/irq/generic-chip.c b/kernel/irq/generic-chip.c index c89295a8f668..10e663ab1f4a 100644 --- a/kernel/irq/generic-chip.c +++ b/kernel/irq/generic-chip.c @@ -7,6 +7,7 @@ #include <linux/irq.h> #include <linux/slab.h> #include <linux/export.h> +#include <linux/irqdomain.h> #include <linux/interrupt.h> #include <linux/kernel_stat.h> #include <linux/syscore_ops.h> @@ -16,11 +17,6 @@ static LIST_HEAD(gc_list); static DEFINE_RAW_SPINLOCK(gc_lock); -static inline struct irq_chip_regs *cur_regs(struct irq_data *d) -{ - return &container_of(d->chip, struct irq_chip_type, chip)->regs; -} - /** * irq_gc_noop - NOOP function * @d: irq_data @@ -39,16 +35,17 @@ void irq_gc_noop(struct irq_data *d) void irq_gc_mask_disable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->disable); - gc->mask_cache &= ~mask; + irq_reg_writel(mask, gc->reg_base + ct->regs.disable); + *ct->mask_cache &= ~mask; irq_gc_unlock(gc); } /** - * irq_gc_mask_set_mask_bit - Mask chip via setting bit in mask register + * irq_gc_mask_set_bit - Mask chip via setting bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -57,16 +54,18 @@ void irq_gc_mask_disable_reg(struct irq_data *d) void irq_gc_mask_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - gc->mask_cache |= mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + cur_regs(d)->mask); + *ct->mask_cache |= mask; + irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_mask_set_bit); /** - * irq_gc_mask_set_mask_bit - Mask chip via clearing bit in mask register + * irq_gc_mask_clr_bit - Mask chip via clearing bit in mask register * @d: irq_data * * Chip has a single mask register. Values of this register are cached @@ -75,13 +74,15 @@ void irq_gc_mask_set_bit(struct irq_data *d) void irq_gc_mask_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - gc->mask_cache &= ~mask; - irq_reg_writel(gc->mask_cache, gc->reg_base + cur_regs(d)->mask); + *ct->mask_cache &= ~mask; + irq_reg_writel(*ct->mask_cache, gc->reg_base + ct->regs.mask); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_mask_clr_bit); /** * irq_gc_unmask_enable_reg - Unmask chip via enable register @@ -93,11 +94,12 @@ void irq_gc_mask_clr_bit(struct irq_data *d) void irq_gc_unmask_enable_reg(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->enable); - gc->mask_cache |= mask; + irq_reg_writel(mask, gc->reg_base + ct->regs.enable); + *ct->mask_cache |= mask; irq_gc_unlock(gc); } @@ -108,12 +110,14 @@ void irq_gc_unmask_enable_reg(struct irq_data *d) void irq_gc_ack_set_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } +EXPORT_SYMBOL_GPL(irq_gc_ack_set_bit); /** * irq_gc_ack_clr_bit - Ack pending interrupt via clearing bit @@ -122,25 +126,27 @@ void irq_gc_ack_set_bit(struct irq_data *d) void irq_gc_ack_clr_bit(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = ~(1 << (d->irq - gc->irq_base)); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = ~d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } /** - * irq_gc_mask_disable_reg_and_ack- Mask and ack pending interrupt + * irq_gc_mask_disable_reg_and_ack - Mask and ack pending interrupt * @d: irq_data */ void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->mask); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->ack); + irq_reg_writel(mask, gc->reg_base + ct->regs.mask); + irq_reg_writel(mask, gc->reg_base + ct->regs.ack); irq_gc_unlock(gc); } @@ -151,16 +157,18 @@ void irq_gc_mask_disable_reg_and_ack(struct irq_data *d) void irq_gc_eoi(struct irq_data *d) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + struct irq_chip_type *ct = irq_data_get_chip_type(d); + u32 mask = d->mask; irq_gc_lock(gc); - irq_reg_writel(mask, gc->reg_base + cur_regs(d)->eoi); + irq_reg_writel(mask, gc->reg_base + ct->regs.eoi); irq_gc_unlock(gc); } /** * irq_gc_set_wake - Set/clr wake bit for an interrupt - * @d: irq_data + * @d: irq_data + * @on: Indicates whether the wake bit should be set or cleared * * For chips where the wake from suspend functionality is not * configured in a separate register and the wakeup active state is @@ -169,7 +177,7 @@ void irq_gc_eoi(struct irq_data *d) int irq_gc_set_wake(struct irq_data *d, unsigned int on) { struct irq_chip_generic *gc = irq_data_get_irq_chip_data(d); - u32 mask = 1 << (d->irq - gc->irq_base); + u32 mask = d->mask; if (!(mask & gc->wake_enabled)) return -EINVAL; @@ -183,6 +191,19 @@ int irq_gc_set_wake(struct irq_data *d, unsigned int on) return 0; } +static void +irq_init_generic_chip(struct irq_chip_generic *gc, const char *name, + int num_ct, unsigned int irq_base, + void __iomem *reg_base, irq_flow_handler_t handler) +{ + raw_spin_lock_init(&gc->lock); + gc->num_ct = num_ct; + gc->irq_base = irq_base; + gc->reg_base = reg_base; + gc->chip_types->chip.name = name; + gc->chip_types->handler = handler; +} + /** * irq_alloc_generic_chip - Allocate a generic chip and initialize it * @name: Name of the irq chip @@ -203,23 +224,183 @@ irq_alloc_generic_chip(const char *name, int num_ct, unsigned int irq_base, gc = kzalloc(sz, GFP_KERNEL); if (gc) { - raw_spin_lock_init(&gc->lock); - gc->num_ct = num_ct; - gc->irq_base = irq_base; - gc->reg_base = reg_base; - gc->chip_types->chip.name = name; - gc->chip_types->handler = handler; + irq_init_generic_chip(gc, name, num_ct, irq_base, reg_base, + handler); } return gc; } EXPORT_SYMBOL_GPL(irq_alloc_generic_chip); +static void +irq_gc_init_mask_cache(struct irq_chip_generic *gc, enum irq_gc_flags flags) +{ + struct irq_chip_type *ct = gc->chip_types; + u32 *mskptr = &gc->mask_cache, mskreg = ct->regs.mask; + int i; + + for (i = 0; i < gc->num_ct; i++) { + if (flags & IRQ_GC_MASK_CACHE_PER_TYPE) { + mskptr = &ct[i].mask_cache_priv; + mskreg = ct[i].regs.mask; + } + ct[i].mask_cache = mskptr; + if (flags & IRQ_GC_INIT_MASK_CACHE) + *mskptr = irq_reg_readl(gc->reg_base + mskreg); + } +} + +/** + * irq_alloc_domain_generic_chip - Allocate generic chips for an irq domain + * @d: irq domain for which to allocate chips + * @irqs_per_chip: Number of interrupts each chip handles + * @num_ct: Number of irq_chip_type instances associated with this + * @name: Name of the irq chip + * @handler: Default flow handler associated with these chips + * @clr: IRQ_* bits to clear in the mapping function + * @set: IRQ_* bits to set in the mapping function + * @gcflags: Generic chip specific setup flags + */ +int irq_alloc_domain_generic_chips(struct irq_domain *d, int irqs_per_chip, + int num_ct, const char *name, + irq_flow_handler_t handler, + unsigned int clr, unsigned int set, + enum irq_gc_flags gcflags) +{ + struct irq_domain_chip_generic *dgc; + struct irq_chip_generic *gc; + int numchips, sz, i; + unsigned long flags; + void *tmp; + + if (d->gc) + return -EBUSY; + + numchips = d->revmap_size / irqs_per_chip; + if (!numchips) + return -EINVAL; + + /* Allocate a pointer, generic chip and chiptypes for each chip */ + sz = sizeof(*dgc) + numchips * sizeof(gc); + sz += numchips * (sizeof(*gc) + num_ct * sizeof(struct irq_chip_type)); + + tmp = dgc = kzalloc(sz, GFP_KERNEL); + if (!dgc) + return -ENOMEM; + dgc->irqs_per_chip = irqs_per_chip; + dgc->num_chips = numchips; + dgc->irq_flags_to_set = set; + dgc->irq_flags_to_clear = clr; + dgc->gc_flags = gcflags; + d->gc = dgc; + + /* Calc pointer to the first generic chip */ + tmp += sizeof(*dgc) + numchips * sizeof(gc); + for (i = 0; i < numchips; i++) { + /* Store the pointer to the generic chip */ + dgc->gc[i] = gc = tmp; + irq_init_generic_chip(gc, name, num_ct, i * irqs_per_chip, + NULL, handler); + gc->domain = d; + raw_spin_lock_irqsave(&gc_lock, flags); + list_add_tail(&gc->list, &gc_list); + raw_spin_unlock_irqrestore(&gc_lock, flags); + /* Calc pointer to the next generic chip */ + tmp += sizeof(*gc) + num_ct * sizeof(struct irq_chip_type); + } + d->name = name; + return 0; +} +EXPORT_SYMBOL_GPL(irq_alloc_domain_generic_chips); + +/** + * irq_get_domain_generic_chip - Get a pointer to the generic chip of a hw_irq + * @d: irq domain pointer + * @hw_irq: Hardware interrupt number + */ +struct irq_chip_generic * +irq_get_domain_generic_chip(struct irq_domain *d, unsigned int hw_irq) +{ + struct irq_domain_chip_generic *dgc = d->gc; + int idx; + + if (!dgc) + return NULL; + idx = hw_irq / dgc->irqs_per_chip; + if (idx >= dgc->num_chips) + return NULL; + return dgc->gc[idx]; +} +EXPORT_SYMBOL_GPL(irq_get_domain_generic_chip); + /* * Separate lockdep class for interrupt chip which can nest irq_desc * lock. */ static struct lock_class_key irq_nested_lock_class; +/* + * irq_map_generic_chip - Map a generic chip for an irq domain + */ +static int irq_map_generic_chip(struct irq_domain *d, unsigned int virq, + irq_hw_number_t hw_irq) +{ + struct irq_data *data = irq_get_irq_data(virq); + struct irq_domain_chip_generic *dgc = d->gc; + struct irq_chip_generic *gc; + struct irq_chip_type *ct; + struct irq_chip *chip; + unsigned long flags; + int idx; + + if (!d->gc) + return -ENODEV; + + idx = hw_irq / dgc->irqs_per_chip; + if (idx >= dgc->num_chips) + return -EINVAL; + gc = dgc->gc[idx]; + + idx = hw_irq % dgc->irqs_per_chip; + + if (test_bit(idx, &gc->unused)) + return -ENOTSUPP; + + if (test_bit(idx, &gc->installed)) + return -EBUSY; + + ct = gc->chip_types; + chip = &ct->chip; + + /* We only init the cache for the first mapping of a generic chip */ + if (!gc->installed) { + raw_spin_lock_irqsave(&gc->lock, flags); + irq_gc_init_mask_cache(gc, dgc->gc_flags); + raw_spin_unlock_irqrestore(&gc->lock, flags); + } + + /* Mark the interrupt as installed */ + set_bit(idx, &gc->installed); + + if (dgc->gc_flags & IRQ_GC_INIT_NESTED_LOCK) + irq_set_lockdep_class(virq, &irq_nested_lock_class); + + if (chip->irq_calc_mask) + chip->irq_calc_mask(data); + else + data->mask = 1 << idx; + + irq_set_chip_and_handler(virq, chip, ct->handler); + irq_set_chip_data(virq, gc); + irq_modify_status(virq, dgc->irq_flags_to_clear, dgc->irq_flags_to_set); + return 0; +} + +struct irq_domain_ops irq_generic_chip_ops = { + .map = irq_map_generic_chip, + .xlate = irq_domain_xlate_onetwocell, +}; +EXPORT_SYMBOL_GPL(irq_generic_chip_ops); + /** * irq_setup_generic_chip - Setup a range of interrupts with a generic chip * @gc: Generic irq chip holding all data @@ -237,15 +418,14 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, unsigned int set) { struct irq_chip_type *ct = gc->chip_types; + struct irq_chip *chip = &ct->chip; unsigned int i; raw_spin_lock(&gc_lock); list_add_tail(&gc->list, &gc_list); raw_spin_unlock(&gc_lock); - /* Init mask cache ? */ - if (flags & IRQ_GC_INIT_MASK_CACHE) - gc->mask_cache = irq_reg_readl(gc->reg_base + ct->regs.mask); + irq_gc_init_mask_cache(gc, flags); for (i = gc->irq_base; msk; msk >>= 1, i++) { if (!(msk & 0x01)) @@ -254,7 +434,15 @@ void irq_setup_generic_chip(struct irq_chip_generic *gc, u32 msk, if (flags & IRQ_GC_INIT_NESTED_LOCK) irq_set_lockdep_class(i, &irq_nested_lock_class); - irq_set_chip_and_handler(i, &ct->chip, ct->handler); + if (!(flags & IRQ_GC_NO_MASK)) { + struct irq_data *d = irq_get_irq_data(i); + + if (chip->irq_calc_mask) + chip->irq_calc_mask(d); + else + d->mask = 1 << (i - gc->irq_base); + } + irq_set_chip_and_handler(i, chip, ct->handler); irq_set_chip_data(i, gc); irq_modify_status(i, clr, set); } @@ -265,7 +453,7 @@ EXPORT_SYMBOL_GPL(irq_setup_generic_chip); /** * irq_setup_alt_chip - Switch to alternative chip * @d: irq_data for this interrupt - * @type Flow type to be initialized + * @type: Flow type to be initialized * * Only to be called from chip->irq_set_type() callbacks. */ @@ -317,6 +505,24 @@ void irq_remove_generic_chip(struct irq_chip_generic *gc, u32 msk, } EXPORT_SYMBOL_GPL(irq_remove_generic_chip); +static struct irq_data *irq_gc_get_irq_data(struct irq_chip_generic *gc) +{ + unsigned int virq; + + if (!gc->domain) + return irq_get_irq_data(gc->irq_base); + + /* + * We don't know which of the irqs has been actually + * installed. Use the first one. + */ + if (!gc->installed) + return NULL; + + virq = irq_find_mapping(gc->domain, gc->irq_base + __ffs(gc->installed)); + return virq ? irq_get_irq_data(virq) : NULL; +} + #ifdef CONFIG_PM static int irq_gc_suspend(void) { @@ -325,8 +531,12 @@ static int irq_gc_suspend(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_suspend) - ct->chip.irq_suspend(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_suspend) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_suspend(data); + } } return 0; } @@ -338,8 +548,12 @@ static void irq_gc_resume(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_resume) - ct->chip.irq_resume(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_resume) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_resume(data); + } } } #else @@ -354,8 +568,12 @@ static void irq_gc_shutdown(void) list_for_each_entry(gc, &gc_list, list) { struct irq_chip_type *ct = gc->chip_types; - if (ct->chip.irq_pm_shutdown) - ct->chip.irq_pm_shutdown(irq_get_irq_data(gc->irq_base)); + if (ct->chip.irq_pm_shutdown) { + struct irq_data *data = irq_gc_get_irq_data(gc); + + if (data) + ct->chip.irq_pm_shutdown(data); + } } } diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c index 5a83dde8ca0c..2d7cd3428365 100644 --- a/kernel/irq/irqdomain.c +++ b/kernel/irq/irqdomain.c @@ -16,12 +16,6 @@ #include <linux/smp.h> #include <linux/fs.h> -#define IRQ_DOMAIN_MAP_LEGACY 0 /* driver allocated fixed range of irqs. - * ie. legacy 8259, gets irqs 1..15 */ -#define IRQ_DOMAIN_MAP_NOMAP 1 /* no fast reverse mapping */ -#define IRQ_DOMAIN_MAP_LINEAR 2 /* linear map of interrupts */ -#define IRQ_DOMAIN_MAP_TREE 3 /* radix tree */ - static LIST_HEAD(irq_domain_list); static DEFINE_MUTEX(irq_domain_mutex); @@ -29,9 +23,11 @@ static DEFINE_MUTEX(revmap_trees_mutex); static struct irq_domain *irq_default_domain; /** - * irq_domain_alloc() - Allocate a new irq_domain data structure + * __irq_domain_add() - Allocate a new irq_domain data structure * @of_node: optional device-tree node of the interrupt controller - * @revmap_type: type of reverse mapping to use + * @size: Size of linear map; 0 for radix mapping only + * @direct_max: Maximum value of direct maps; Use ~0 for no limit; 0 for no + * direct mapping * @ops: map/unmap domain callbacks * @host_data: Controller private data pointer * @@ -39,41 +35,35 @@ static struct irq_domain *irq_default_domain; * register allocated irq_domain with irq_domain_register(). Returns pointer * to IRQ domain, or NULL on failure. */ -static struct irq_domain *irq_domain_alloc(struct device_node *of_node, - unsigned int revmap_type, - const struct irq_domain_ops *ops, - void *host_data) +struct irq_domain *__irq_domain_add(struct device_node *of_node, int size, + irq_hw_number_t hwirq_max, int direct_max, + const struct irq_domain_ops *ops, + void *host_data) { struct irq_domain *domain; - domain = kzalloc_node(sizeof(*domain), GFP_KERNEL, - of_node_to_nid(of_node)); + domain = kzalloc_node(sizeof(*domain) + (sizeof(unsigned int) * size), + GFP_KERNEL, of_node_to_nid(of_node)); if (WARN_ON(!domain)) return NULL; /* Fill structure */ - domain->revmap_type = revmap_type; + INIT_RADIX_TREE(&domain->revmap_tree, GFP_KERNEL); domain->ops = ops; domain->host_data = host_data; domain->of_node = of_node_get(of_node); + domain->hwirq_max = hwirq_max; + domain->revmap_size = size; + domain->revmap_direct_max_irq = direct_max; - return domain; -} - -static void irq_domain_free(struct irq_domain *domain) -{ - of_node_put(domain->of_node); - kfree(domain); -} - -static void irq_domain_add(struct irq_domain *domain) -{ mutex_lock(&irq_domain_mutex); list_add(&domain->link, &irq_domain_list); mutex_unlock(&irq_domain_mutex); - pr_debug("Allocated domain of type %d @0x%p\n", - domain->revmap_type, domain); + + pr_debug("Added domain %s\n", domain->name); + return domain; } +EXPORT_SYMBOL_GPL(__irq_domain_add); /** * irq_domain_remove() - Remove an irq domain. @@ -87,29 +77,12 @@ void irq_domain_remove(struct irq_domain *domain) { mutex_lock(&irq_domain_mutex); - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LEGACY: - /* - * Legacy domains don't manage their own irq_desc - * allocations, we expect the caller to handle irq_desc - * freeing on their own. - */ - break; - case IRQ_DOMAIN_MAP_TREE: - /* - * radix_tree_delete() takes care of destroying the root - * node when all entries are removed. Shout if there are - * any mappings left. - */ - WARN_ON(domain->revmap_data.tree.height); - break; - case IRQ_DOMAIN_MAP_LINEAR: - kfree(domain->revmap_data.linear.revmap); - domain->revmap_data.linear.size = 0; - break; - case IRQ_DOMAIN_MAP_NOMAP: - break; - } + /* + * radix_tree_delete() takes care of destroying the root + * node when all entries are removed. Shout if there are + * any mappings left. + */ + WARN_ON(domain->revmap_tree.height); list_del(&domain->link); @@ -121,41 +94,30 @@ void irq_domain_remove(struct irq_domain *domain) mutex_unlock(&irq_domain_mutex); - pr_debug("Removed domain of type %d @0x%p\n", - domain->revmap_type, domain); + pr_debug("Removed domain %s\n", domain->name); - irq_domain_free(domain); + of_node_put(domain->of_node); + kfree(domain); } EXPORT_SYMBOL_GPL(irq_domain_remove); -static unsigned int irq_domain_legacy_revmap(struct irq_domain *domain, - irq_hw_number_t hwirq) -{ - irq_hw_number_t first_hwirq = domain->revmap_data.legacy.first_hwirq; - int size = domain->revmap_data.legacy.size; - - if (WARN_ON(hwirq < first_hwirq || hwirq >= first_hwirq + size)) - return 0; - return hwirq - first_hwirq + domain->revmap_data.legacy.first_irq; -} - /** - * irq_domain_add_simple() - Allocate and register a simple irq_domain. + * irq_domain_add_simple() - Register an irq_domain and optionally map a range of irqs * @of_node: pointer to interrupt controller's device tree node. * @size: total number of irqs in mapping - * @first_irq: first number of irq block assigned to the domain + * @first_irq: first number of irq block assigned to the domain, + * pass zero to assign irqs on-the-fly. If first_irq is non-zero, then + * pre-map all of the irqs in the domain to virqs starting at first_irq. * @ops: map/unmap domain callbacks * @host_data: Controller private data pointer * - * Allocates a legacy irq_domain if irq_base is positive or a linear - * domain otherwise. For the legacy domain, IRQ descriptors will also - * be allocated. + * Allocates an irq_domain, and optionally if first_irq is positive then also + * allocate irq_descs and map all of the hwirqs to virqs starting at first_irq. * * This is intended to implement the expected behaviour for most - * interrupt controllers which is that a linear mapping should - * normally be used unless the system requires a legacy mapping in - * order to support supplying interrupt numbers during non-DT - * registration of devices. + * interrupt controllers. If device tree is used, then first_irq will be 0 and + * irqs get mapped dynamically on the fly. However, if the controller requires + * static virq assignments (non-DT boot) then it will set that up correctly. */ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, unsigned int size, @@ -163,34 +125,27 @@ struct irq_domain *irq_domain_add_simple(struct device_node *of_node, const struct irq_domain_ops *ops, void *host_data) { - if (first_irq > 0) { - int irq_base; + struct irq_domain *domain; + + domain = __irq_domain_add(of_node, size, size, 0, ops, host_data); + if (!domain) + return NULL; + if (first_irq > 0) { if (IS_ENABLED(CONFIG_SPARSE_IRQ)) { - /* - * Set the descriptor allocator to search for a - * 1-to-1 mapping, such as irq_alloc_desc_at(). - * Use of_node_to_nid() which is defined to - * numa_node_id() on platforms that have no custom - * implementation. - */ - irq_base = irq_alloc_descs(first_irq, first_irq, size, - of_node_to_nid(of_node)); - if (irq_base < 0) { + /* attempt to allocated irq_descs */ + int rc = irq_alloc_descs(first_irq, first_irq, size, + of_node_to_nid(of_node)); + if (rc < 0) pr_info("Cannot allocate irq_descs @ IRQ%d, assuming pre-allocated\n", first_irq); - irq_base = first_irq; - } - } else - irq_base = first_irq; - - return irq_domain_add_legacy(of_node, size, irq_base, 0, - ops, host_data); + } + irq_domain_associate_many(domain, first_irq, 0, size); } - /* A linear domain is the default */ - return irq_domain_add_linear(of_node, size, ops, host_data); + return domain; } +EXPORT_SYMBOL_GPL(irq_domain_add_simple); /** * irq_domain_add_legacy() - Allocate and register a legacy revmap irq_domain. @@ -215,131 +170,19 @@ struct irq_domain *irq_domain_add_legacy(struct device_node *of_node, void *host_data) { struct irq_domain *domain; - unsigned int i; - domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LEGACY, ops, host_data); + domain = __irq_domain_add(of_node, first_hwirq + size, + first_hwirq + size, 0, ops, host_data); if (!domain) return NULL; - domain->revmap_data.legacy.first_irq = first_irq; - domain->revmap_data.legacy.first_hwirq = first_hwirq; - domain->revmap_data.legacy.size = size; - - mutex_lock(&irq_domain_mutex); - /* Verify that all the irqs are available */ - for (i = 0; i < size; i++) { - int irq = first_irq + i; - struct irq_data *irq_data = irq_get_irq_data(irq); - - if (WARN_ON(!irq_data || irq_data->domain)) { - mutex_unlock(&irq_domain_mutex); - irq_domain_free(domain); - return NULL; - } - } - - /* Claim all of the irqs before registering a legacy domain */ - for (i = 0; i < size; i++) { - struct irq_data *irq_data = irq_get_irq_data(first_irq + i); - irq_data->hwirq = first_hwirq + i; - irq_data->domain = domain; - } - mutex_unlock(&irq_domain_mutex); - - for (i = 0; i < size; i++) { - int irq = first_irq + i; - int hwirq = first_hwirq + i; - - /* IRQ0 gets ignored */ - if (!irq) - continue; - - /* Legacy flags are left to default at this point, - * one can then use irq_create_mapping() to - * explicitly change them - */ - if (ops->map) - ops->map(domain, irq, hwirq); - - /* Clear norequest flags */ - irq_clear_status_flags(irq, IRQ_NOREQUEST); - } + irq_domain_associate_many(domain, first_irq, first_hwirq, size); - irq_domain_add(domain); return domain; } EXPORT_SYMBOL_GPL(irq_domain_add_legacy); /** - * irq_domain_add_linear() - Allocate and register a linear revmap irq_domain. - * @of_node: pointer to interrupt controller's device tree node. - * @size: Number of interrupts in the domain. - * @ops: map/unmap domain callbacks - * @host_data: Controller private data pointer - */ -struct irq_domain *irq_domain_add_linear(struct device_node *of_node, - unsigned int size, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain; - unsigned int *revmap; - - revmap = kzalloc_node(sizeof(*revmap) * size, GFP_KERNEL, - of_node_to_nid(of_node)); - if (WARN_ON(!revmap)) - return NULL; - - domain = irq_domain_alloc(of_node, IRQ_DOMAIN_MAP_LINEAR, ops, host_data); - if (!domain) { - kfree(revmap); - return NULL; - } - domain->revmap_data.linear.size = size; - domain->revmap_data.linear.revmap = revmap; - irq_domain_add(domain); - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_linear); - -struct irq_domain *irq_domain_add_nomap(struct device_node *of_node, - unsigned int max_irq, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain = irq_domain_alloc(of_node, - IRQ_DOMAIN_MAP_NOMAP, ops, host_data); - if (domain) { - domain->revmap_data.nomap.max_irq = max_irq ? max_irq : ~0; - irq_domain_add(domain); - } - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_nomap); - -/** - * irq_domain_add_tree() - * @of_node: pointer to interrupt controller's device tree node. - * @ops: map/unmap domain callbacks - * - * Note: The radix tree will be allocated later during boot automatically - * (the reverse mapping will use the slow path until that happens). - */ -struct irq_domain *irq_domain_add_tree(struct device_node *of_node, - const struct irq_domain_ops *ops, - void *host_data) -{ - struct irq_domain *domain = irq_domain_alloc(of_node, - IRQ_DOMAIN_MAP_TREE, ops, host_data); - if (domain) { - INIT_RADIX_TREE(&domain->revmap_data.tree, GFP_KERNEL); - irq_domain_add(domain); - } - return domain; -} -EXPORT_SYMBOL_GPL(irq_domain_add_tree); - -/** * irq_find_host() - Locates a domain for a given device node * @node: device-tree node of the interrupt controller */ @@ -387,124 +230,108 @@ void irq_set_default_host(struct irq_domain *domain) } EXPORT_SYMBOL_GPL(irq_set_default_host); -static void irq_domain_disassociate_many(struct irq_domain *domain, - unsigned int irq_base, int count) +static void irq_domain_disassociate(struct irq_domain *domain, unsigned int irq) { - /* - * disassociate in reverse order; - * not strictly necessary, but nice for unwinding - */ - while (count--) { - int irq = irq_base + count; - struct irq_data *irq_data = irq_get_irq_data(irq); - irq_hw_number_t hwirq = irq_data->hwirq; + struct irq_data *irq_data = irq_get_irq_data(irq); + irq_hw_number_t hwirq; - if (WARN_ON(!irq_data || irq_data->domain != domain)) - continue; + if (WARN(!irq_data || irq_data->domain != domain, + "virq%i doesn't exist; cannot disassociate\n", irq)) + return; - irq_set_status_flags(irq, IRQ_NOREQUEST); + hwirq = irq_data->hwirq; + irq_set_status_flags(irq, IRQ_NOREQUEST); - /* remove chip and handler */ - irq_set_chip_and_handler(irq, NULL, NULL); + /* remove chip and handler */ + irq_set_chip_and_handler(irq, NULL, NULL); - /* Make sure it's completed */ - synchronize_irq(irq); + /* Make sure it's completed */ + synchronize_irq(irq); - /* Tell the PIC about it */ - if (domain->ops->unmap) - domain->ops->unmap(domain, irq); - smp_mb(); + /* Tell the PIC about it */ + if (domain->ops->unmap) + domain->ops->unmap(domain, irq); + smp_mb(); - irq_data->domain = NULL; - irq_data->hwirq = 0; + irq_data->domain = NULL; + irq_data->hwirq = 0; - /* Clear reverse map */ - switch(domain->revmap_type) { - case IRQ_DOMAIN_MAP_LINEAR: - if (hwirq < domain->revmap_data.linear.size) - domain->revmap_data.linear.revmap[hwirq] = 0; - break; - case IRQ_DOMAIN_MAP_TREE: - mutex_lock(&revmap_trees_mutex); - radix_tree_delete(&domain->revmap_data.tree, hwirq); - mutex_unlock(&revmap_trees_mutex); - break; - } + /* Clear reverse map for this hwirq */ + if (hwirq < domain->revmap_size) { + domain->linear_revmap[hwirq] = 0; + } else { + mutex_lock(&revmap_trees_mutex); + radix_tree_delete(&domain->revmap_tree, hwirq); + mutex_unlock(&revmap_trees_mutex); } } -int irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, - irq_hw_number_t hwirq_base, int count) +int irq_domain_associate(struct irq_domain *domain, unsigned int virq, + irq_hw_number_t hwirq) { - unsigned int virq = irq_base; - irq_hw_number_t hwirq = hwirq_base; - int i, ret; + struct irq_data *irq_data = irq_get_irq_data(virq); + int ret; - pr_debug("%s(%s, irqbase=%i, hwbase=%i, count=%i)\n", __func__, - of_node_full_name(domain->of_node), irq_base, (int)hwirq_base, count); + if (WARN(hwirq >= domain->hwirq_max, + "error: hwirq 0x%x is too large for %s\n", (int)hwirq, domain->name)) + return -EINVAL; + if (WARN(!irq_data, "error: virq%i is not allocated", virq)) + return -EINVAL; + if (WARN(irq_data->domain, "error: virq%i is already associated", virq)) + return -EINVAL; - for (i = 0; i < count; i++) { - struct irq_data *irq_data = irq_get_irq_data(virq + i); - - if (WARN(!irq_data, "error: irq_desc not allocated; " - "irq=%i hwirq=0x%x\n", virq + i, (int)hwirq + i)) - return -EINVAL; - if (WARN(irq_data->domain, "error: irq_desc already associated; " - "irq=%i hwirq=0x%x\n", virq + i, (int)hwirq + i)) - return -EINVAL; - }; - - for (i = 0; i < count; i++, virq++, hwirq++) { - struct irq_data *irq_data = irq_get_irq_data(virq); - - irq_data->hwirq = hwirq; - irq_data->domain = domain; - if (domain->ops->map) { - ret = domain->ops->map(domain, virq, hwirq); - if (ret != 0) { - /* - * If map() returns -EPERM, this interrupt is protected - * by the firmware or some other service and shall not - * be mapped. - * - * Since on some platforms we blindly try to map everything - * we end up with a log full of backtraces. - * - * So instead, we silently fail on -EPERM, it is the - * responsibility of the PIC driver to display a relevant - * message if needed. - */ - if (ret != -EPERM) { - pr_err("irq-%i==>hwirq-0x%lx mapping failed: %d\n", - virq, hwirq, ret); - WARN_ON(1); - } - irq_data->domain = NULL; - irq_data->hwirq = 0; - goto err_unmap; + mutex_lock(&irq_domain_mutex); + irq_data->hwirq = hwirq; + irq_data->domain = domain; + if (domain->ops->map) { + ret = domain->ops->map(domain, virq, hwirq); + if (ret != 0) { + /* + * If map() returns -EPERM, this interrupt is protected + * by the firmware or some other service and shall not + * be mapped. Don't bother telling the user about it. + */ + if (ret != -EPERM) { + pr_info("%s didn't like hwirq-0x%lx to VIRQ%i mapping (rc=%d)\n", + domain->name, hwirq, virq, ret); } + irq_data->domain = NULL; + irq_data->hwirq = 0; + mutex_unlock(&irq_domain_mutex); + return ret; } - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LINEAR: - if (hwirq < domain->revmap_data.linear.size) - domain->revmap_data.linear.revmap[hwirq] = virq; - break; - case IRQ_DOMAIN_MAP_TREE: - mutex_lock(&revmap_trees_mutex); - radix_tree_insert(&domain->revmap_data.tree, hwirq, irq_data); - mutex_unlock(&revmap_trees_mutex); - break; - } + /* If not already assigned, give the domain the chip's name */ + if (!domain->name && irq_data->chip) + domain->name = irq_data->chip->name; + } - irq_clear_status_flags(virq, IRQ_NOREQUEST); + if (hwirq < domain->revmap_size) { + domain->linear_revmap[hwirq] = virq; + } else { + mutex_lock(&revmap_trees_mutex); + radix_tree_insert(&domain->revmap_tree, hwirq, irq_data); + mutex_unlock(&revmap_trees_mutex); } + mutex_unlock(&irq_domain_mutex); + + irq_clear_status_flags(virq, IRQ_NOREQUEST); return 0; +} +EXPORT_SYMBOL_GPL(irq_domain_associate); - err_unmap: - irq_domain_disassociate_many(domain, irq_base, i); - return -EINVAL; +void irq_domain_associate_many(struct irq_domain *domain, unsigned int irq_base, + irq_hw_number_t hwirq_base, int count) +{ + int i; + + pr_debug("%s(%s, irqbase=%i, hwbase=%i, count=%i)\n", __func__, + of_node_full_name(domain->of_node), irq_base, (int)hwirq_base, count); + + for (i = 0; i < count; i++) { + irq_domain_associate(domain, irq_base + i, hwirq_base + i); + } } EXPORT_SYMBOL_GPL(irq_domain_associate_many); @@ -514,7 +341,9 @@ EXPORT_SYMBOL_GPL(irq_domain_associate_many); * * This routine is used for irq controllers which can choose the hardware * interrupt numbers they generate. In such a case it's simplest to use - * the linux irq as the hardware interrupt number. + * the linux irq as the hardware interrupt number. It still uses the linear + * or radix tree to store the mapping, but the irq controller can optimize + * the revmap path by using the hwirq directly. */ unsigned int irq_create_direct_mapping(struct irq_domain *domain) { @@ -523,17 +352,14 @@ unsigned int irq_create_direct_mapping(struct irq_domain *domain) if (domain == NULL) domain = irq_default_domain; - if (WARN_ON(!domain || domain->revmap_type != IRQ_DOMAIN_MAP_NOMAP)) - return 0; - virq = irq_alloc_desc_from(1, of_node_to_nid(domain->of_node)); if (!virq) { pr_debug("create_direct virq allocation failed\n"); return 0; } - if (virq >= domain->revmap_data.nomap.max_irq) { + if (virq >= domain->revmap_direct_max_irq) { pr_err("ERROR: no free irqs available below %i maximum\n", - domain->revmap_data.nomap.max_irq); + domain->revmap_direct_max_irq); irq_free_desc(virq); return 0; } @@ -570,9 +396,7 @@ unsigned int irq_create_mapping(struct irq_domain *domain, if (domain == NULL) domain = irq_default_domain; if (domain == NULL) { - pr_warning("irq_create_mapping called for" - " NULL domain, hwirq=%lx\n", hwirq); - WARN_ON(1); + WARN(1, "%s(, %lx) called with NULL domain\n", __func__, hwirq); return 0; } pr_debug("-> using domain @%p\n", domain); @@ -584,10 +408,6 @@ unsigned int irq_create_mapping(struct irq_domain *domain, return virq; } - /* Get a virtual interrupt number */ - if (domain->revmap_type == IRQ_DOMAIN_MAP_LEGACY) - return irq_domain_legacy_revmap(domain, hwirq); - /* Allocate a virtual interrupt number */ hint = hwirq % nr_irqs; if (hint == 0) @@ -640,12 +460,7 @@ int irq_create_strict_mappings(struct irq_domain *domain, unsigned int irq_base, if (unlikely(ret < 0)) return ret; - ret = irq_domain_associate_many(domain, irq_base, hwirq_base, count); - if (unlikely(ret < 0)) { - irq_free_descs(irq_base, count); - return ret; - } - + irq_domain_associate_many(domain, irq_base, hwirq_base, count); return 0; } EXPORT_SYMBOL_GPL(irq_create_strict_mappings); @@ -672,8 +487,8 @@ unsigned int irq_create_of_mapping(struct device_node *controller, if (intsize > 0) return intspec[0]; #endif - pr_warning("no irq domain found for %s !\n", - of_node_full_name(controller)); + pr_warn("no irq domain found for %s !\n", + of_node_full_name(controller)); return 0; } @@ -693,7 +508,7 @@ unsigned int irq_create_of_mapping(struct device_node *controller, /* Set type if specified and different than the current one */ if (type != IRQ_TYPE_NONE && - type != (irqd_get_trigger_type(irq_get_irq_data(virq)))) + type != irq_get_trigger_type(virq)) irq_set_irq_type(virq, type); return virq; } @@ -715,11 +530,7 @@ void irq_dispose_mapping(unsigned int virq) if (WARN_ON(domain == NULL)) return; - /* Never unmap legacy interrupts */ - if (domain->revmap_type == IRQ_DOMAIN_MAP_LEGACY) - return; - - irq_domain_disassociate_many(domain, virq, 1); + irq_domain_disassociate(domain, virq); irq_free_desc(virq); } EXPORT_SYMBOL_GPL(irq_dispose_mapping); @@ -740,63 +551,51 @@ unsigned int irq_find_mapping(struct irq_domain *domain, if (domain == NULL) return 0; - switch (domain->revmap_type) { - case IRQ_DOMAIN_MAP_LEGACY: - return irq_domain_legacy_revmap(domain, hwirq); - case IRQ_DOMAIN_MAP_LINEAR: - return irq_linear_revmap(domain, hwirq); - case IRQ_DOMAIN_MAP_TREE: - rcu_read_lock(); - data = radix_tree_lookup(&domain->revmap_data.tree, hwirq); - rcu_read_unlock(); - if (data) - return data->irq; - break; - case IRQ_DOMAIN_MAP_NOMAP: + if (hwirq < domain->revmap_direct_max_irq) { data = irq_get_irq_data(hwirq); if (data && (data->domain == domain) && (data->hwirq == hwirq)) return hwirq; - break; } - return 0; -} -EXPORT_SYMBOL_GPL(irq_find_mapping); + /* Check if the hwirq is in the linear revmap. */ + if (hwirq < domain->revmap_size) + return domain->linear_revmap[hwirq]; -/** - * irq_linear_revmap() - Find a linux irq from a hw irq number. - * @domain: domain owning this hardware interrupt - * @hwirq: hardware irq number in that domain space - * - * This is a fast path that can be called directly by irq controller code to - * save a handful of instructions. - */ -unsigned int irq_linear_revmap(struct irq_domain *domain, - irq_hw_number_t hwirq) -{ - BUG_ON(domain->revmap_type != IRQ_DOMAIN_MAP_LINEAR); - - /* Check revmap bounds; complain if exceeded */ - if (WARN_ON(hwirq >= domain->revmap_data.linear.size)) - return 0; - - return domain->revmap_data.linear.revmap[hwirq]; + rcu_read_lock(); + data = radix_tree_lookup(&domain->revmap_tree, hwirq); + rcu_read_unlock(); + return data ? data->irq : 0; } -EXPORT_SYMBOL_GPL(irq_linear_revmap); +EXPORT_SYMBOL_GPL(irq_find_mapping); #ifdef CONFIG_IRQ_DOMAIN_DEBUG static int virq_debug_show(struct seq_file *m, void *private) { unsigned long flags; struct irq_desc *desc; - const char *p; - static const char none[] = "none"; - void *data; + struct irq_domain *domain; + struct radix_tree_iter iter; + void *data, **slot; int i; - seq_printf(m, "%-5s %-7s %-15s %-*s %s\n", "irq", "hwirq", + seq_printf(m, " %-16s %-6s %-10s %-10s %s\n", + "name", "mapped", "linear-max", "direct-max", "devtree-node"); + mutex_lock(&irq_domain_mutex); + list_for_each_entry(domain, &irq_domain_list, link) { + int count = 0; + radix_tree_for_each_slot(slot, &domain->revmap_tree, &iter, 0) + count++; + seq_printf(m, "%c%-16s %6u %10u %10u %s\n", + domain == irq_default_domain ? '*' : ' ', domain->name, + domain->revmap_size + count, domain->revmap_size, + domain->revmap_direct_max_irq, + domain->of_node ? of_node_full_name(domain->of_node) : ""); + } + mutex_unlock(&irq_domain_mutex); + + seq_printf(m, "%-5s %-7s %-15s %-*s %6s %-14s %s\n", "irq", "hwirq", "chip name", (int)(2 * sizeof(void *) + 2), "chip data", - "domain name"); + "active", "type", "domain"); for (i = 1; i < nr_irqs; i++) { desc = irq_to_desc(i); @@ -804,28 +603,28 @@ static int virq_debug_show(struct seq_file *m, void *private) continue; raw_spin_lock_irqsave(&desc->lock, flags); + domain = desc->irq_data.domain; - if (desc->action && desc->action->handler) { + if (domain) { struct irq_chip *chip; + int hwirq = desc->irq_data.hwirq; + bool direct; seq_printf(m, "%5d ", i); - seq_printf(m, "0x%05lx ", desc->irq_data.hwirq); + seq_printf(m, "0x%05x ", hwirq); chip = irq_desc_get_chip(desc); - if (chip && chip->name) - p = chip->name; - else - p = none; - seq_printf(m, "%-15s ", p); + seq_printf(m, "%-15s ", (chip && chip->name) ? chip->name : "none"); data = irq_desc_get_chip_data(desc); seq_printf(m, data ? "0x%p " : " %p ", data); - if (desc->irq_data.domain) - p = of_node_full_name(desc->irq_data.domain->of_node); - else - p = none; - seq_printf(m, "%s\n", p); + seq_printf(m, " %c ", (desc->action && desc->action->handler) ? '*' : ' '); + direct = (i == hwirq) && (i < domain->revmap_direct_max_irq); + seq_printf(m, "%6s%-8s ", + (hwirq < domain->revmap_size) ? "LINEAR" : "RADIX", + direct ? "(DIRECT)" : ""); + seq_printf(m, "%s\n", desc->irq_data.domain->name); } raw_spin_unlock_irqrestore(&desc->lock, flags); @@ -922,18 +721,3 @@ const struct irq_domain_ops irq_domain_simple_ops = { .xlate = irq_domain_xlate_onetwocell, }; EXPORT_SYMBOL_GPL(irq_domain_simple_ops); - -#ifdef CONFIG_OF_IRQ -void irq_domain_generate_simple(const struct of_device_id *match, - u64 phys_base, unsigned int irq_start) -{ - struct device_node *node; - pr_debug("looking for phys_base=%llx, irq_start=%i\n", - (unsigned long long) phys_base, (int) irq_start); - node = of_find_matching_node_by_address(NULL, match, phys_base); - if (node) - irq_domain_add_legacy(node, 32, irq_start, 0, - &irq_domain_simple_ops, NULL); -} -EXPORT_SYMBOL_GPL(irq_domain_generate_simple); -#endif diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index fa17855ca65a..514bcfd855a8 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -555,9 +555,9 @@ int can_request_irq(unsigned int irq, unsigned long irqflags) return 0; if (irq_settings_can_request(desc)) { - if (desc->action) - if (irqflags & desc->action->flags & IRQF_SHARED) - canrequest =1; + if (!desc->action || + irqflags & desc->action->flags & IRQF_SHARED) + canrequest = 1; } irq_put_desc_unlock(desc, flags); return canrequest; @@ -840,9 +840,6 @@ static void irq_thread_dtor(struct callback_head *unused) static int irq_thread(void *data) { struct callback_head on_exit_work; - static const struct sched_param param = { - .sched_priority = MAX_USER_RT_PRIO/2, - }; struct irqaction *action = data; struct irq_desc *desc = irq_to_desc(action->irq); irqreturn_t (*handler_fn)(struct irq_desc *desc, @@ -854,8 +851,6 @@ static int irq_thread(void *data) else handler_fn = irq_thread_fn; - sched_setscheduler(current, SCHED_FIFO, ¶m); - init_task_work(&on_exit_work, irq_thread_dtor); task_work_add(current, &on_exit_work, false); @@ -950,6 +945,9 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) */ if (new->thread_fn && !nested) { struct task_struct *t; + static const struct sched_param param = { + .sched_priority = MAX_USER_RT_PRIO/2, + }; t = kthread_create(irq_thread, new, "irq/%d-%s", irq, new->name); @@ -957,6 +955,9 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) ret = PTR_ERR(t); goto out_mput; } + + sched_setscheduler(t, SCHED_FIFO, ¶m); + /* * We keep the reference to the task struct even if * the thread dies to avoid that the interrupt code diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index 19ed5c425c3b..36f6ee181b0c 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -462,6 +462,8 @@ int show_interrupts(struct seq_file *p, void *v) } else { seq_printf(p, " %8s", "None"); } + if (desc->irq_data.domain) + seq_printf(p, " %*d", prec, (int) desc->irq_data.hwirq); #ifdef CONFIG_GENERIC_IRQ_SHOW_LEVEL seq_printf(p, " %-8s", irqd_is_level_type(&desc->irq_data) ? "Level" : "Edge"); #endif diff --git a/kernel/kmod.c b/kernel/kmod.c index 8241906c4b61..fb326365b694 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -147,6 +147,9 @@ int __request_module(bool wait, const char *fmt, ...) */ WARN_ON_ONCE(wait && current_is_async()); + if (!modprobe_path[0]) + return 0; + va_start(args, fmt); ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args); va_end(args); @@ -569,14 +572,6 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait) int retval = 0; helper_lock(); - if (!sub_info->path) { - retval = -EINVAL; - goto out; - } - - if (sub_info->path[0] == '\0') - goto out; - if (!khelper_wq || usermodehelper_disabled) { retval = -EBUSY; goto out; diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 3fed7f0cbcdf..6e33498d665c 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -467,6 +467,7 @@ static struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr) /* Optimization staging list, protected by kprobe_mutex */ static LIST_HEAD(optimizing_list); static LIST_HEAD(unoptimizing_list); +static LIST_HEAD(freeing_list); static void kprobe_optimizer(struct work_struct *work); static DECLARE_DELAYED_WORK(optimizing_work, kprobe_optimizer); @@ -504,7 +505,7 @@ static __kprobes void do_optimize_kprobes(void) * Unoptimize (replace a jump with a breakpoint and remove the breakpoint * if need) kprobes listed on unoptimizing_list. */ -static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) +static __kprobes void do_unoptimize_kprobes(void) { struct optimized_kprobe *op, *tmp; @@ -515,9 +516,9 @@ static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) /* Ditto to do_optimize_kprobes */ get_online_cpus(); mutex_lock(&text_mutex); - arch_unoptimize_kprobes(&unoptimizing_list, free_list); + arch_unoptimize_kprobes(&unoptimizing_list, &freeing_list); /* Loop free_list for disarming */ - list_for_each_entry_safe(op, tmp, free_list, list) { + list_for_each_entry_safe(op, tmp, &freeing_list, list) { /* Disarm probes if marked disabled */ if (kprobe_disabled(&op->kp)) arch_disarm_kprobe(&op->kp); @@ -536,11 +537,11 @@ static __kprobes void do_unoptimize_kprobes(struct list_head *free_list) } /* Reclaim all kprobes on the free_list */ -static __kprobes void do_free_cleaned_kprobes(struct list_head *free_list) +static __kprobes void do_free_cleaned_kprobes(void) { struct optimized_kprobe *op, *tmp; - list_for_each_entry_safe(op, tmp, free_list, list) { + list_for_each_entry_safe(op, tmp, &freeing_list, list) { BUG_ON(!kprobe_unused(&op->kp)); list_del_init(&op->list); free_aggr_kprobe(&op->kp); @@ -556,8 +557,6 @@ static __kprobes void kick_kprobe_optimizer(void) /* Kprobe jump optimizer */ static __kprobes void kprobe_optimizer(struct work_struct *work) { - LIST_HEAD(free_list); - mutex_lock(&kprobe_mutex); /* Lock modules while optimizing kprobes */ mutex_lock(&module_mutex); @@ -566,7 +565,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work) * Step 1: Unoptimize kprobes and collect cleaned (unused and disarmed) * kprobes before waiting for quiesence period. */ - do_unoptimize_kprobes(&free_list); + do_unoptimize_kprobes(); /* * Step 2: Wait for quiesence period to ensure all running interrupts @@ -581,7 +580,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work) do_optimize_kprobes(); /* Step 4: Free cleaned kprobes after quiesence period */ - do_free_cleaned_kprobes(&free_list); + do_free_cleaned_kprobes(); mutex_unlock(&module_mutex); mutex_unlock(&kprobe_mutex); @@ -723,8 +722,19 @@ static void __kprobes kill_optimized_kprobe(struct kprobe *p) if (!list_empty(&op->list)) /* Dequeue from the (un)optimization queue */ list_del_init(&op->list); - op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED; + + if (kprobe_unused(p)) { + /* Enqueue if it is unused */ + list_add(&op->list, &freeing_list); + /* + * Remove unused probes from the hash list. After waiting + * for synchronization, this probe is reclaimed. + * (reclaiming is done by do_free_cleaned_kprobes().) + */ + hlist_del_rcu(&op->kp.hlist); + } + /* Don't touch the code, because it is already freed. */ arch_remove_optimized_kprobe(op); } @@ -2322,6 +2332,7 @@ static ssize_t write_enabled_file_bool(struct file *file, if (copy_from_user(buf, user_buf, buf_size)) return -EFAULT; + buf[buf_size] = '\0'; switch (buf[0]) { case 'y': case 'Y': @@ -2333,6 +2344,8 @@ static ssize_t write_enabled_file_bool(struct file *file, case '0': disarm_all_kprobes(); break; + default: + return -EINVAL; } return count; diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 1f3186b37fd5..e16c45b9ee77 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -4090,7 +4090,7 @@ void debug_check_no_locks_freed(const void *mem_from, unsigned long mem_len) } EXPORT_SYMBOL_GPL(debug_check_no_locks_freed); -static void print_held_locks_bug(struct task_struct *curr) +static void print_held_locks_bug(void) { if (!debug_locks_off()) return; @@ -4099,22 +4099,21 @@ static void print_held_locks_bug(struct task_struct *curr) printk("\n"); printk("=====================================\n"); - printk("[ BUG: lock held at task exit time! ]\n"); + printk("[ BUG: %s/%d still has locks held! ]\n", + current->comm, task_pid_nr(current)); print_kernel_ident(); printk("-------------------------------------\n"); - printk("%s/%d is exiting with locks still held!\n", - curr->comm, task_pid_nr(curr)); - lockdep_print_held_locks(curr); - + lockdep_print_held_locks(current); printk("\nstack backtrace:\n"); dump_stack(); } -void debug_check_no_locks_held(struct task_struct *task) +void debug_check_no_locks_held(void) { - if (unlikely(task->lockdep_depth > 0)) - print_held_locks_bug(task); + if (unlikely(current->lockdep_depth > 0)) + print_held_locks_bug(); } +EXPORT_SYMBOL_GPL(debug_check_no_locks_held); void debug_show_all_locks(void) { diff --git a/kernel/mutex.c b/kernel/mutex.c index ad53a664f113..e581ada5faf4 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -254,16 +254,165 @@ void __sched mutex_unlock(struct mutex *lock) EXPORT_SYMBOL(mutex_unlock); +/** + * ww_mutex_unlock - release the w/w mutex + * @lock: the mutex to be released + * + * Unlock a mutex that has been locked by this task previously with any of the + * ww_mutex_lock* functions (with or without an acquire context). It is + * forbidden to release the locks after releasing the acquire context. + * + * This function must not be used in interrupt context. Unlocking + * of a unlocked mutex is not allowed. + */ +void __sched ww_mutex_unlock(struct ww_mutex *lock) +{ + /* + * The unlocking fastpath is the 0->1 transition from 'locked' + * into 'unlocked' state: + */ + if (lock->ctx) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(!lock->ctx->acquired); +#endif + if (lock->ctx->acquired > 0) + lock->ctx->acquired--; + lock->ctx = NULL; + } + +#ifndef CONFIG_DEBUG_MUTEXES + /* + * When debugging is enabled we must not clear the owner before time, + * the slow path will always be taken, and that clears the owner field + * after verifying that it was indeed current. + */ + mutex_clear_owner(&lock->base); +#endif + __mutex_fastpath_unlock(&lock->base.count, __mutex_unlock_slowpath); +} +EXPORT_SYMBOL(ww_mutex_unlock); + +static inline int __sched +__mutex_lock_check_stamp(struct mutex *lock, struct ww_acquire_ctx *ctx) +{ + struct ww_mutex *ww = container_of(lock, struct ww_mutex, base); + struct ww_acquire_ctx *hold_ctx = ACCESS_ONCE(ww->ctx); + + if (!hold_ctx) + return 0; + + if (unlikely(ctx == hold_ctx)) + return -EALREADY; + + if (ctx->stamp - hold_ctx->stamp <= LONG_MAX && + (ctx->stamp != hold_ctx->stamp || ctx > hold_ctx)) { +#ifdef CONFIG_DEBUG_MUTEXES + DEBUG_LOCKS_WARN_ON(ctx->contending_lock); + ctx->contending_lock = ww; +#endif + return -EDEADLK; + } + + return 0; +} + +static __always_inline void ww_mutex_lock_acquired(struct ww_mutex *ww, + struct ww_acquire_ctx *ww_ctx) +{ +#ifdef CONFIG_DEBUG_MUTEXES + /* + * If this WARN_ON triggers, you used ww_mutex_lock to acquire, + * but released with a normal mutex_unlock in this call. + * + * This should never happen, always use ww_mutex_unlock. + */ + DEBUG_LOCKS_WARN_ON(ww->ctx); + + /* + * Not quite done after calling ww_acquire_done() ? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->done_acquire); + + if (ww_ctx->contending_lock) { + /* + * After -EDEADLK you tried to + * acquire a different ww_mutex? Bad! + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->contending_lock != ww); + + /* + * You called ww_mutex_lock after receiving -EDEADLK, + * but 'forgot' to unlock everything else first? + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->acquired > 0); + ww_ctx->contending_lock = NULL; + } + + /* + * Naughty, using a different class will lead to undefined behavior! + */ + DEBUG_LOCKS_WARN_ON(ww_ctx->ww_class != ww->ww_class); +#endif + ww_ctx->acquired++; +} + +/* + * after acquiring lock with fastpath or when we lost out in contested + * slowpath, set ctx and wake up any waiters so they can recheck. + * + * This function is never called when CONFIG_DEBUG_LOCK_ALLOC is set, + * as the fastpath and opportunistic spinning are disabled in that case. + */ +static __always_inline void +ww_mutex_set_context_fastpath(struct ww_mutex *lock, + struct ww_acquire_ctx *ctx) +{ + unsigned long flags; + struct mutex_waiter *cur; + + ww_mutex_lock_acquired(lock, ctx); + + lock->ctx = ctx; + + /* + * The lock->ctx update should be visible on all cores before + * the atomic read is done, otherwise contended waiters might be + * missed. The contended waiters will either see ww_ctx == NULL + * and keep spinning, or it will acquire wait_lock, add itself + * to waiter list and sleep. + */ + smp_mb(); /* ^^^ */ + + /* + * Check if lock is contended, if not there is nobody to wake up + */ + if (likely(atomic_read(&lock->base.count) == 0)) + return; + + /* + * Uh oh, we raced in fastpath, wake up everyone in this case, + * so they can see the new lock->ctx. + */ + spin_lock_mutex(&lock->base.wait_lock, flags); + list_for_each_entry(cur, &lock->base.wait_list, list) { + debug_mutex_wake_waiter(&lock->base, cur); + wake_up_process(cur->task); + } + spin_unlock_mutex(&lock->base.wait_lock, flags); +} + /* * Lock a mutex (possibly interruptible), slowpath: */ -static inline int __sched +static __always_inline int __sched __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, - struct lockdep_map *nest_lock, unsigned long ip) + struct lockdep_map *nest_lock, unsigned long ip, + struct ww_acquire_ctx *ww_ctx) { struct task_struct *task = current; struct mutex_waiter waiter; unsigned long flags; + int ret; preempt_disable(); mutex_acquire_nest(&lock->dep_map, subclass, 0, nest_lock, ip); @@ -298,6 +447,22 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, struct task_struct *owner; struct mspin_node node; + if (!__builtin_constant_p(ww_ctx == NULL) && ww_ctx->acquired > 0) { + struct ww_mutex *ww; + + ww = container_of(lock, struct ww_mutex, base); + /* + * If ww->ctx is set the contents are undefined, only + * by acquiring wait_lock there is a guarantee that + * they are not invalid when reading. + * + * As such, when deadlock detection needs to be + * performed the optimistic spinning cannot be done. + */ + if (ACCESS_ONCE(ww->ctx)) + break; + } + /* * If there's an owner, wait for it to either * release the lock or go to sleep. @@ -312,6 +477,13 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, if ((atomic_read(&lock->count) == 1) && (atomic_cmpxchg(&lock->count, 1, 0) == 1)) { lock_acquired(&lock->dep_map, ip); + if (!__builtin_constant_p(ww_ctx == NULL)) { + struct ww_mutex *ww; + ww = container_of(lock, struct ww_mutex, base); + + ww_mutex_set_context_fastpath(ww, ww_ctx); + } + mutex_set_owner(lock); mspin_unlock(MLOCK(lock), &node); preempt_enable(); @@ -371,15 +543,16 @@ slowpath: * TASK_UNINTERRUPTIBLE case.) */ if (unlikely(signal_pending_state(state, task))) { - mutex_remove_waiter(lock, &waiter, - task_thread_info(task)); - mutex_release(&lock->dep_map, 1, ip); - spin_unlock_mutex(&lock->wait_lock, flags); + ret = -EINTR; + goto err; + } - debug_mutex_free_waiter(&waiter); - preempt_enable(); - return -EINTR; + if (!__builtin_constant_p(ww_ctx == NULL) && ww_ctx->acquired > 0) { + ret = __mutex_lock_check_stamp(lock, ww_ctx); + if (ret) + goto err; } + __set_task_state(task, state); /* didn't get the lock, go to sleep: */ @@ -394,6 +567,30 @@ done: mutex_remove_waiter(lock, &waiter, current_thread_info()); mutex_set_owner(lock); + if (!__builtin_constant_p(ww_ctx == NULL)) { + struct ww_mutex *ww = container_of(lock, + struct ww_mutex, + base); + struct mutex_waiter *cur; + + /* + * This branch gets optimized out for the common case, + * and is only important for ww_mutex_lock. + */ + + ww_mutex_lock_acquired(ww, ww_ctx); + ww->ctx = ww_ctx; + + /* + * Give any possible sleeping processes the chance to wake up, + * so they can recheck if they have to back off. + */ + list_for_each_entry(cur, &lock->wait_list, list) { + debug_mutex_wake_waiter(lock, cur); + wake_up_process(cur->task); + } + } + /* set it to 0 if there are no waiters left: */ if (likely(list_empty(&lock->wait_list))) atomic_set(&lock->count, 0); @@ -404,6 +601,14 @@ done: preempt_enable(); return 0; + +err: + mutex_remove_waiter(lock, &waiter, task_thread_info(task)); + spin_unlock_mutex(&lock->wait_lock, flags); + debug_mutex_free_waiter(&waiter); + mutex_release(&lock->dep_map, 1, ip); + preempt_enable(); + return ret; } #ifdef CONFIG_DEBUG_LOCK_ALLOC @@ -411,7 +616,8 @@ void __sched mutex_lock_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, subclass, NULL, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_nested); @@ -420,7 +626,8 @@ void __sched _mutex_lock_nest_lock(struct mutex *lock, struct lockdep_map *nest) { might_sleep(); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, nest, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, + 0, nest, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(_mutex_lock_nest_lock); @@ -429,7 +636,8 @@ int __sched mutex_lock_killable_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); - return __mutex_lock_common(lock, TASK_KILLABLE, subclass, NULL, _RET_IP_); + return __mutex_lock_common(lock, TASK_KILLABLE, + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_killable_nested); @@ -438,10 +646,68 @@ mutex_lock_interruptible_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, - subclass, NULL, _RET_IP_); + subclass, NULL, _RET_IP_, NULL); } EXPORT_SYMBOL_GPL(mutex_lock_interruptible_nested); + +static inline int +ww_mutex_deadlock_injection(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ +#ifdef CONFIG_DEBUG_WW_MUTEX_SLOWPATH + unsigned tmp; + + if (ctx->deadlock_inject_countdown-- == 0) { + tmp = ctx->deadlock_inject_interval; + if (tmp > UINT_MAX/4) + tmp = UINT_MAX; + else + tmp = tmp*2 + tmp + tmp/2; + + ctx->deadlock_inject_interval = tmp; + ctx->deadlock_inject_countdown = tmp; + ctx->contending_lock = lock; + + ww_mutex_unlock(lock); + + return -EDEADLK; + } +#endif + + return 0; +} + +int __sched +__ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + ret = __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, + 0, &ctx->dep_map, _RET_IP_, ctx); + if (!ret && ctx->acquired > 0) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; +} +EXPORT_SYMBOL_GPL(__ww_mutex_lock); + +int __sched +__ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + ret = __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, + 0, &ctx->dep_map, _RET_IP_, ctx); + + if (!ret && ctx->acquired > 0) + return ww_mutex_deadlock_injection(lock, ctx); + + return ret; +} +EXPORT_SYMBOL_GPL(__ww_mutex_lock_interruptible); + #endif /* @@ -494,10 +760,10 @@ __mutex_unlock_slowpath(atomic_t *lock_count) * mutex_lock_interruptible() and mutex_trylock(). */ static noinline int __sched -__mutex_lock_killable_slowpath(atomic_t *lock_count); +__mutex_lock_killable_slowpath(struct mutex *lock); static noinline int __sched -__mutex_lock_interruptible_slowpath(atomic_t *lock_count); +__mutex_lock_interruptible_slowpath(struct mutex *lock); /** * mutex_lock_interruptible - acquire the mutex, interruptible @@ -515,12 +781,12 @@ int __sched mutex_lock_interruptible(struct mutex *lock) int ret; might_sleep(); - ret = __mutex_fastpath_lock_retval - (&lock->count, __mutex_lock_interruptible_slowpath); - if (!ret) + ret = __mutex_fastpath_lock_retval(&lock->count); + if (likely(!ret)) { mutex_set_owner(lock); - - return ret; + return 0; + } else + return __mutex_lock_interruptible_slowpath(lock); } EXPORT_SYMBOL(mutex_lock_interruptible); @@ -530,12 +796,12 @@ int __sched mutex_lock_killable(struct mutex *lock) int ret; might_sleep(); - ret = __mutex_fastpath_lock_retval - (&lock->count, __mutex_lock_killable_slowpath); - if (!ret) + ret = __mutex_fastpath_lock_retval(&lock->count); + if (likely(!ret)) { mutex_set_owner(lock); - - return ret; + return 0; + } else + return __mutex_lock_killable_slowpath(lock); } EXPORT_SYMBOL(mutex_lock_killable); @@ -544,24 +810,39 @@ __mutex_lock_slowpath(atomic_t *lock_count) { struct mutex *lock = container_of(lock_count, struct mutex, count); - __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, NULL, _RET_IP_); + __mutex_lock_common(lock, TASK_UNINTERRUPTIBLE, 0, + NULL, _RET_IP_, NULL); } static noinline int __sched -__mutex_lock_killable_slowpath(atomic_t *lock_count) +__mutex_lock_killable_slowpath(struct mutex *lock) { - struct mutex *lock = container_of(lock_count, struct mutex, count); + return __mutex_lock_common(lock, TASK_KILLABLE, 0, + NULL, _RET_IP_, NULL); +} - return __mutex_lock_common(lock, TASK_KILLABLE, 0, NULL, _RET_IP_); +static noinline int __sched +__mutex_lock_interruptible_slowpath(struct mutex *lock) +{ + return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, + NULL, _RET_IP_, NULL); } static noinline int __sched -__mutex_lock_interruptible_slowpath(atomic_t *lock_count) +__ww_mutex_lock_slowpath(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) { - struct mutex *lock = container_of(lock_count, struct mutex, count); + return __mutex_lock_common(&lock->base, TASK_UNINTERRUPTIBLE, 0, + NULL, _RET_IP_, ctx); +} - return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, 0, NULL, _RET_IP_); +static noinline int __sched +__ww_mutex_lock_interruptible_slowpath(struct ww_mutex *lock, + struct ww_acquire_ctx *ctx) +{ + return __mutex_lock_common(&lock->base, TASK_INTERRUPTIBLE, 0, + NULL, _RET_IP_, ctx); } + #endif /* @@ -617,6 +898,45 @@ int __sched mutex_trylock(struct mutex *lock) } EXPORT_SYMBOL(mutex_trylock); +#ifndef CONFIG_DEBUG_LOCK_ALLOC +int __sched +__ww_mutex_lock(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + ret = __mutex_fastpath_lock_retval(&lock->base.count); + + if (likely(!ret)) { + ww_mutex_set_context_fastpath(lock, ctx); + mutex_set_owner(&lock->base); + } else + ret = __ww_mutex_lock_slowpath(lock, ctx); + return ret; +} +EXPORT_SYMBOL(__ww_mutex_lock); + +int __sched +__ww_mutex_lock_interruptible(struct ww_mutex *lock, struct ww_acquire_ctx *ctx) +{ + int ret; + + might_sleep(); + + ret = __mutex_fastpath_lock_retval(&lock->base.count); + + if (likely(!ret)) { + ww_mutex_set_context_fastpath(lock, ctx); + mutex_set_owner(&lock->base); + } else + ret = __ww_mutex_lock_interruptible_slowpath(lock, ctx); + return ret; +} +EXPORT_SYMBOL(__ww_mutex_lock_interruptible); + +#endif + /** * atomic_dec_and_mutex_lock - return holding mutex if we dec to 0 * @cnt: the atomic which we are to dec diff --git a/kernel/pid.c b/kernel/pid.c index 0db3e791a06d..66505c1dfc51 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -75,6 +75,7 @@ struct pid_namespace init_pid_ns = { [ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL } }, .last_pid = 0, + .nr_hashed = PIDNS_HASH_ADDING, .level = 0, .child_reaper = &init_task, .user_ns = &init_user_ns, @@ -373,14 +374,10 @@ EXPORT_SYMBOL_GPL(find_vpid); /* * attach_pid() must be called with the tasklist_lock write-held. */ -void attach_pid(struct task_struct *task, enum pid_type type, - struct pid *pid) +void attach_pid(struct task_struct *task, enum pid_type type) { - struct pid_link *link; - - link = &task->pids[type]; - link->pid = pid; - hlist_add_head_rcu(&link->node, &pid->tasks[type]); + struct pid_link *link = &task->pids[type]; + hlist_add_head_rcu(&link->node, &link->pid->tasks[type]); } static void __change_pid(struct task_struct *task, enum pid_type type, @@ -412,7 +409,7 @@ void change_pid(struct task_struct *task, enum pid_type type, struct pid *pid) { __change_pid(task, type, pid); - attach_pid(task, type, pid); + attach_pid(task, type); } /* transfer_pid is an optimization of attach_pid(new), detach_pid(old) */ @@ -594,7 +591,6 @@ void __init pidmap_init(void) /* Reserve PID 0. We never call free_pidmap(0) */ set_bit(0, init_pid_ns.pidmap[0].page); atomic_dec(&init_pid_ns.pidmap[0].nr_free); - init_pid_ns.nr_hashed = PIDNS_HASH_ADDING; init_pid_ns.pid_cachep = KMEM_CACHE(pid, SLAB_HWCACHE_ALIGN | SLAB_PANIC); diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9ea180b..d444c4e834f4 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -100,7 +100,6 @@ config PM_SLEEP_SMP depends on SMP depends on ARCH_SUSPEND_POSSIBLE || ARCH_HIBERNATION_POSSIBLE depends on PM_SLEEP - select HOTPLUG select HOTPLUG_CPU config PM_AUTOSLEEP @@ -263,6 +262,26 @@ config PM_GENERIC_DOMAINS bool depends on PM +config WQ_POWER_EFFICIENT_DEFAULT + bool "Enable workqueue power-efficient mode by default" + depends on PM + default n + help + Per-cpu workqueues are generally preferred because they show + better performance thanks to cache locality; unfortunately, + per-cpu workqueues tend to be more power hungry than unbound + workqueues. + + Enabling workqueue.power_efficient kernel parameter makes the + per-cpu workqueues which were observed to contribute + significantly to power consumption unbound, leading to measurably + lower power usage at the cost of small performance overhead. + + This config option determines whether workqueue.power_efficient + is enabled by default. + + If in doubt, say N. + config PM_GENERIC_DOMAINS_SLEEP def_bool y depends on PM_SLEEP && PM_GENERIC_DOMAINS diff --git a/kernel/power/main.c b/kernel/power/main.c index d77663bfedeb..1d1bf630e6e9 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -424,6 +424,8 @@ static ssize_t wakeup_count_store(struct kobject *kobj, if (sscanf(buf, "%u", &val) == 1) { if (pm_save_wakeup_count(val)) error = n; + else + pm_print_active_wakeup_sources(); } out: @@ -528,6 +530,10 @@ pm_trace_store(struct kobject *kobj, struct kobj_attribute *attr, if (sscanf(buf, "%d", &val) == 1) { pm_trace_enabled = !!val; + if (pm_trace_enabled) { + pr_warn("PM: Enabling pm_trace changes system date and time during resume.\n" + "PM: Correct system time has to be restored manually after resume.\n"); + } return n; } return -EINVAL; diff --git a/kernel/power/process.c b/kernel/power/process.c index 98088e0e71e8..fc0df8486449 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -30,9 +30,10 @@ static int try_to_freeze_tasks(bool user_only) unsigned int todo; bool wq_busy = false; struct timeval start, end; - u64 elapsed_csecs64; - unsigned int elapsed_csecs; + u64 elapsed_msecs64; + unsigned int elapsed_msecs; bool wakeup = false; + int sleep_usecs = USEC_PER_MSEC; do_gettimeofday(&start); @@ -68,22 +69,25 @@ static int try_to_freeze_tasks(bool user_only) /* * We need to retry, but first give the freezing tasks some - * time to enter the refrigerator. + * time to enter the refrigerator. Start with an initial + * 1 ms sleep followed by exponential backoff until 8 ms. */ - msleep(10); + usleep_range(sleep_usecs / 2, sleep_usecs); + if (sleep_usecs < 8 * USEC_PER_MSEC) + sleep_usecs *= 2; } do_gettimeofday(&end); - elapsed_csecs64 = timeval_to_ns(&end) - timeval_to_ns(&start); - do_div(elapsed_csecs64, NSEC_PER_SEC / 100); - elapsed_csecs = elapsed_csecs64; + elapsed_msecs64 = timeval_to_ns(&end) - timeval_to_ns(&start); + do_div(elapsed_msecs64, NSEC_PER_MSEC); + elapsed_msecs = elapsed_msecs64; if (todo) { printk("\n"); - printk(KERN_ERR "Freezing of tasks %s after %d.%02d seconds " + printk(KERN_ERR "Freezing of tasks %s after %d.%03d seconds " "(%d tasks refusing to freeze, wq_busy=%d):\n", wakeup ? "aborted" : "failed", - elapsed_csecs / 100, elapsed_csecs % 100, + elapsed_msecs / 1000, elapsed_msecs % 1000, todo - wq_busy, wq_busy); if (!wakeup) { @@ -96,8 +100,8 @@ static int try_to_freeze_tasks(bool user_only) read_unlock(&tasklist_lock); } } else { - printk("(elapsed %d.%02d seconds) ", elapsed_csecs / 100, - elapsed_csecs % 100); + printk("(elapsed %d.%03d seconds) ", elapsed_msecs / 1000, + elapsed_msecs % 1000); } return todo ? -EBUSY : 0; diff --git a/kernel/power/qos.c b/kernel/power/qos.c index 587dddeebf15..06fe28589e9c 100644 --- a/kernel/power/qos.c +++ b/kernel/power/qos.c @@ -44,6 +44,7 @@ #include <linux/uaccess.h> #include <linux/export.h> +#include <trace/events/power.h> /* * locking rule: all changes to constraints or notifiers lists @@ -202,6 +203,7 @@ int pm_qos_update_target(struct pm_qos_constraints *c, struct plist_node *node, spin_unlock_irqrestore(&pm_qos_lock, flags); + trace_pm_qos_update_target(action, prev_value, curr_value); if (prev_value != curr_value) { blocking_notifier_call_chain(c->notifiers, (unsigned long)curr_value, @@ -272,6 +274,7 @@ bool pm_qos_update_flags(struct pm_qos_flags *pqf, spin_unlock_irqrestore(&pm_qos_lock, irqflags); + trace_pm_qos_update_flags(action, prev_value, curr_value); return prev_value != curr_value; } @@ -333,6 +336,7 @@ void pm_qos_add_request(struct pm_qos_request *req, } req->pm_qos_class = pm_qos_class; INIT_DELAYED_WORK(&req->work, pm_qos_work_fn); + trace_pm_qos_add_request(pm_qos_class, value); pm_qos_update_target(pm_qos_array[pm_qos_class]->constraints, &req->node, PM_QOS_ADD_REQ, value); } @@ -361,6 +365,7 @@ void pm_qos_update_request(struct pm_qos_request *req, cancel_delayed_work_sync(&req->work); + trace_pm_qos_update_request(req->pm_qos_class, new_value); if (new_value != req->node.prio) pm_qos_update_target( pm_qos_array[req->pm_qos_class]->constraints, @@ -387,6 +392,8 @@ void pm_qos_update_request_timeout(struct pm_qos_request *req, s32 new_value, cancel_delayed_work_sync(&req->work); + trace_pm_qos_update_request_timeout(req->pm_qos_class, + new_value, timeout_us); if (new_value != req->node.prio) pm_qos_update_target( pm_qos_array[req->pm_qos_class]->constraints, @@ -416,6 +423,7 @@ void pm_qos_remove_request(struct pm_qos_request *req) cancel_delayed_work_sync(&req->work); + trace_pm_qos_remove_request(req->pm_qos_class, PM_QOS_DEFAULT_VALUE); pm_qos_update_target(pm_qos_array[req->pm_qos_class]->constraints, &req->node, PM_QOS_REMOVE_REQ, PM_QOS_DEFAULT_VALUE); @@ -477,7 +485,7 @@ static int find_pm_qos_object_by_minor(int minor) { int pm_qos_class; - for (pm_qos_class = 0; + for (pm_qos_class = PM_QOS_CPU_DMA_LATENCY; pm_qos_class < PM_QOS_NUM_CLASSES; pm_qos_class++) { if (minor == pm_qos_array[pm_qos_class]->pm_qos_power_miscdev.minor) @@ -491,7 +499,7 @@ static int pm_qos_power_open(struct inode *inode, struct file *filp) long pm_qos_class; pm_qos_class = find_pm_qos_object_by_minor(iminor(inode)); - if (pm_qos_class >= 0) { + if (pm_qos_class >= PM_QOS_CPU_DMA_LATENCY) { struct pm_qos_request *req = kzalloc(sizeof(*req), GFP_KERNEL); if (!req) return -ENOMEM; @@ -584,7 +592,7 @@ static int __init pm_qos_power_init(void) BUILD_BUG_ON(ARRAY_SIZE(pm_qos_array) != PM_QOS_NUM_CLASSES); - for (i = 1; i < PM_QOS_NUM_CLASSES; i++) { + for (i = PM_QOS_CPU_DMA_LATENCY; i < PM_QOS_NUM_CLASSES; i++) { ret = register_pm_qos_misc(pm_qos_array[i]); if (ret < 0) { printk(KERN_ERR "pm_qos_param: %s setup failed\n", diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 0de28576807d..349587bb03e1 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -642,8 +642,9 @@ __register_nosave_region(unsigned long start_pfn, unsigned long end_pfn, region->end_pfn = end_pfn; list_add_tail(®ion->list, &nosave_regions); Report: - printk(KERN_INFO "PM: Registered nosave memory: %016lx - %016lx\n", - start_pfn << PAGE_SHIFT, end_pfn << PAGE_SHIFT); + printk(KERN_INFO "PM: Registered nosave memory: [mem %#010llx-%#010llx]\n", + (unsigned long long) start_pfn << PAGE_SHIFT, + ((unsigned long long) end_pfn << PAGE_SHIFT) - 1); } /* @@ -1651,7 +1652,7 @@ unsigned long snapshot_get_image_size(void) static int init_header(struct swsusp_info *info) { memset(info, 0, sizeof(struct swsusp_info)); - info->num_physpages = num_physpages; + info->num_physpages = get_num_physpages(); info->image_pages = nr_copy_pages; info->pages = snapshot_get_image_size(); info->size = info->pages; @@ -1795,7 +1796,7 @@ static int check_header(struct swsusp_info *info) char *reason; reason = check_image_kernel(info); - if (!reason && info->num_physpages != num_physpages) + if (!reason && info->num_physpages != get_num_physpages()) reason = "memory size"; if (reason) { printk(KERN_ERR "PM: Image mismatch: %s\n", reason); diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c index bef86d121eb2..ece04223bb1e 100644 --- a/kernel/power/suspend.c +++ b/kernel/power/suspend.c @@ -269,7 +269,7 @@ int suspend_devices_and_enter(suspend_state_t state) suspend_test_start(); error = dpm_suspend_start(PMSG_SUSPEND); if (error) { - printk(KERN_ERR "PM: Some devices failed to suspend\n"); + pr_err("PM: Some devices failed to suspend, or early wake event detected\n"); goto Recover_platform; } suspend_test_finish("suspend devices"); diff --git a/kernel/printk.c b/kernel/printk.c index fa36e1494420..8212c1aef125 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -363,6 +363,53 @@ static void log_store(int facility, int level, log_next_seq++; } +#ifdef CONFIG_SECURITY_DMESG_RESTRICT +int dmesg_restrict = 1; +#else +int dmesg_restrict; +#endif + +static int syslog_action_restricted(int type) +{ + if (dmesg_restrict) + return 1; + /* + * Unless restricted, we allow "read all" and "get buffer size" + * for everybody. + */ + return type != SYSLOG_ACTION_READ_ALL && + type != SYSLOG_ACTION_SIZE_BUFFER; +} + +static int check_syslog_permissions(int type, bool from_file) +{ + /* + * If this is from /proc/kmsg and we've already opened it, then we've + * already done the capabilities checks at open time. + */ + if (from_file && type != SYSLOG_ACTION_OPEN) + return 0; + + if (syslog_action_restricted(type)) { + if (capable(CAP_SYSLOG)) + return 0; + /* + * For historical reasons, accept CAP_SYS_ADMIN too, with + * a warning. + */ + if (capable(CAP_SYS_ADMIN)) { + pr_warn_once("%s (%d): Attempt to access syslog with " + "CAP_SYS_ADMIN but no CAP_SYSLOG " + "(deprecated).\n", + current->comm, task_pid_nr(current)); + return 0; + } + return -EPERM; + } + return security_syslog(type); +} + + /* /dev/kmsg - userspace message inject/listen interface */ struct devkmsg_user { u64 seq; @@ -620,7 +667,8 @@ static int devkmsg_open(struct inode *inode, struct file *file) if ((file->f_flags & O_ACCMODE) == O_WRONLY) return 0; - err = security_syslog(SYSLOG_ACTION_READ_ALL); + err = check_syslog_permissions(SYSLOG_ACTION_READ_ALL, + SYSLOG_FROM_READER); if (err) return err; @@ -813,45 +861,6 @@ static inline void boot_delay_msec(int level) } #endif -#ifdef CONFIG_SECURITY_DMESG_RESTRICT -int dmesg_restrict = 1; -#else -int dmesg_restrict; -#endif - -static int syslog_action_restricted(int type) -{ - if (dmesg_restrict) - return 1; - /* Unless restricted, we allow "read all" and "get buffer size" for everybody */ - return type != SYSLOG_ACTION_READ_ALL && type != SYSLOG_ACTION_SIZE_BUFFER; -} - -static int check_syslog_permissions(int type, bool from_file) -{ - /* - * If this is from /proc/kmsg and we've already opened it, then we've - * already done the capabilities checks at open time. - */ - if (from_file && type != SYSLOG_ACTION_OPEN) - return 0; - - if (syslog_action_restricted(type)) { - if (capable(CAP_SYSLOG)) - return 0; - /* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */ - if (capable(CAP_SYS_ADMIN)) { - printk_once(KERN_WARNING "%s (%d): " - "Attempt to access syslog with CAP_SYS_ADMIN " - "but no CAP_SYSLOG (deprecated).\n", - current->comm, task_pid_nr(current)); - return 0; - } - return -EPERM; - } - return 0; -} - #if defined(CONFIG_PRINTK_TIME) static bool printk_time = 1; #else @@ -1249,7 +1258,7 @@ out: SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) { - return do_syslog(type, buf, len, SYSLOG_FROM_CALL); + return do_syslog(type, buf, len, SYSLOG_FROM_READER); } /* diff --git a/kernel/ptrace.c b/kernel/ptrace.c index aed981a3f69c..ba5e6cea181a 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -665,20 +665,22 @@ static int ptrace_peek_siginfo(struct task_struct *child, if (unlikely(is_compat_task())) { compat_siginfo_t __user *uinfo = compat_ptr(data); - ret = copy_siginfo_to_user32(uinfo, &info); - ret |= __put_user(info.si_code, &uinfo->si_code); + if (copy_siginfo_to_user32(uinfo, &info) || + __put_user(info.si_code, &uinfo->si_code)) { + ret = -EFAULT; + break; + } + } else #endif { siginfo_t __user *uinfo = (siginfo_t __user *) data; - ret = copy_siginfo_to_user(uinfo, &info); - ret |= __put_user(info.si_code, &uinfo->si_code); - } - - if (ret) { - ret = -EFAULT; - break; + if (copy_siginfo_to_user(uinfo, &info) || + __put_user(info.si_code, &uinfo->si_code)) { + ret = -EFAULT; + break; + } } data += sizeof(siginfo_t); @@ -842,6 +844,47 @@ int ptrace_request(struct task_struct *child, long request, ret = ptrace_setsiginfo(child, &siginfo); break; + case PTRACE_GETSIGMASK: + if (addr != sizeof(sigset_t)) { + ret = -EINVAL; + break; + } + + if (copy_to_user(datavp, &child->blocked, sizeof(sigset_t))) + ret = -EFAULT; + else + ret = 0; + + break; + + case PTRACE_SETSIGMASK: { + sigset_t new_set; + + if (addr != sizeof(sigset_t)) { + ret = -EINVAL; + break; + } + + if (copy_from_user(&new_set, datavp, sizeof(sigset_t))) { + ret = -EFAULT; + break; + } + + sigdelsetmask(&new_set, sigmask(SIGKILL)|sigmask(SIGSTOP)); + + /* + * Every thread does recalc_sigpending() after resume, so + * retarget_shared_pending() and recalc_sigpending() are not + * called here. + */ + spin_lock_irq(&child->sighand->siglock); + child->blocked = new_set; + spin_unlock_irq(&child->sighand->siglock); + + ret = 0; + break; + } + case PTRACE_INTERRUPT: /* * Stop tracee without any side-effect on signal or job @@ -946,8 +989,7 @@ int ptrace_request(struct task_struct *child, long request, #ifdef CONFIG_HAVE_ARCH_TRACEHOOK case PTRACE_GETREGSET: - case PTRACE_SETREGSET: - { + case PTRACE_SETREGSET: { struct iovec kiov; struct iovec __user *uiov = datavp; diff --git a/kernel/range.c b/kernel/range.c index eb911dbce267..322ea8e93e4b 100644 --- a/kernel/range.c +++ b/kernel/range.c @@ -4,7 +4,7 @@ #include <linux/kernel.h> #include <linux/init.h> #include <linux/sort.h> - +#include <linux/string.h> #include <linux/range.h> int add_range(struct range *range, int az, int nr_range, u64 start, u64 end) @@ -32,9 +32,8 @@ int add_range_with_merge(struct range *range, int az, int nr_range, if (start >= end) return nr_range; - /* Try to merge it with old one: */ + /* get new start/end: */ for (i = 0; i < nr_range; i++) { - u64 final_start, final_end; u64 common_start, common_end; if (!range[i].end) @@ -45,14 +44,16 @@ int add_range_with_merge(struct range *range, int az, int nr_range, if (common_start > common_end) continue; - final_start = min(range[i].start, start); - final_end = max(range[i].end, end); + /* new start/end, will add it back at last */ + start = min(range[i].start, start); + end = max(range[i].end, end); - /* clear it and add it back for further merge */ - range[i].start = 0; - range[i].end = 0; - return add_range_with_merge(range, az, nr_range, - final_start, final_end); + memmove(&range[i], &range[i + 1], + (nr_range - (i + 1)) * sizeof(range[i])); + range[nr_range - 1].start = 0; + range[nr_range - 1].end = 0; + nr_range--; + i--; } /* Need to add it: */ diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 48ab70384a4c..cce6ba8bbace 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -104,31 +104,7 @@ void __rcu_read_unlock(void) } EXPORT_SYMBOL_GPL(__rcu_read_unlock); -/* - * Check for a task exiting while in a preemptible-RCU read-side - * critical section, clean up if so. No need to issue warnings, - * as debug_check_no_locks_held() already does this if lockdep - * is enabled. - */ -void exit_rcu(void) -{ - struct task_struct *t = current; - - if (likely(list_empty(¤t->rcu_node_entry))) - return; - t->rcu_read_lock_nesting = 1; - barrier(); - t->rcu_read_unlock_special = RCU_READ_UNLOCK_BLOCKED; - __rcu_read_unlock(); -} - -#else /* #ifdef CONFIG_PREEMPT_RCU */ - -void exit_rcu(void) -{ -} - -#endif /* #else #ifdef CONFIG_PREEMPT_RCU */ +#endif /* #ifdef CONFIG_PREEMPT_RCU */ #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; @@ -145,9 +121,6 @@ static struct lock_class_key rcu_sched_lock_key; struct lockdep_map rcu_sched_lock_map = STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_sched", &rcu_sched_lock_key); EXPORT_SYMBOL_GPL(rcu_sched_lock_map); -#endif - -#ifdef CONFIG_DEBUG_LOCK_ALLOC int debug_lockdep_rcu_enabled(void) { diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index a0714a51b6d7..aa344111de3e 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -44,7 +44,6 @@ /* Forward declarations for rcutiny_plugin.h. */ struct rcu_ctrlblk; -static void invoke_rcu_callbacks(void); static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp); static void rcu_process_callbacks(struct softirq_action *unused); static void __call_rcu(struct rcu_head *head, @@ -205,7 +204,7 @@ static int rcu_is_cpu_rrupt_from_idle(void) */ static int rcu_qsctr_help(struct rcu_ctrlblk *rcp) { - reset_cpu_stall_ticks(rcp); + RCU_TRACE(reset_cpu_stall_ticks(rcp)); if (rcp->rcucblist != NULL && rcp->donetail != rcp->curtail) { rcp->donetail = rcp->curtail; @@ -227,7 +226,7 @@ void rcu_sched_qs(int cpu) local_irq_save(flags); if (rcu_qsctr_help(&rcu_sched_ctrlblk) + rcu_qsctr_help(&rcu_bh_ctrlblk)) - invoke_rcu_callbacks(); + raise_softirq(RCU_SOFTIRQ); local_irq_restore(flags); } @@ -240,7 +239,7 @@ void rcu_bh_qs(int cpu) local_irq_save(flags); if (rcu_qsctr_help(&rcu_bh_ctrlblk)) - invoke_rcu_callbacks(); + raise_softirq(RCU_SOFTIRQ); local_irq_restore(flags); } @@ -252,12 +251,11 @@ void rcu_bh_qs(int cpu) */ void rcu_check_callbacks(int cpu, int user) { - check_cpu_stalls(); + RCU_TRACE(check_cpu_stalls()); if (user || rcu_is_cpu_rrupt_from_idle()) rcu_sched_qs(cpu); else if (!in_softirq()) rcu_bh_qs(cpu); - rcu_preempt_check_callbacks(); } /* @@ -278,7 +276,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) ACCESS_ONCE(rcp->rcucblist), need_resched(), is_idle_task(current), - rcu_is_callbacks_kthread())); + false)); return; } @@ -290,7 +288,6 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) *rcp->donetail = NULL; if (rcp->curtail == rcp->donetail) rcp->curtail = &rcp->rcucblist; - rcu_preempt_remove_callbacks(rcp); rcp->donetail = &rcp->rcucblist; local_irq_restore(flags); @@ -309,14 +306,13 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) RCU_TRACE(rcu_trace_sub_qlen(rcp, cb_count)); RCU_TRACE(trace_rcu_batch_end(rcp->name, cb_count, 0, need_resched(), is_idle_task(current), - rcu_is_callbacks_kthread())); + false)); } static void rcu_process_callbacks(struct softirq_action *unused) { __rcu_process_callbacks(&rcu_sched_ctrlblk); __rcu_process_callbacks(&rcu_bh_ctrlblk); - rcu_preempt_process_callbacks(); } /* @@ -382,3 +378,8 @@ void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) __call_rcu(head, func, &rcu_bh_ctrlblk); } EXPORT_SYMBOL_GPL(call_rcu_bh); + +void rcu_init(void) +{ + open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); +} diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 8a233002faeb..0cd385acccfa 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -53,958 +53,10 @@ static struct rcu_ctrlblk rcu_bh_ctrlblk = { }; #ifdef CONFIG_DEBUG_LOCK_ALLOC +#include <linux/kernel_stat.h> + int rcu_scheduler_active __read_mostly; EXPORT_SYMBOL_GPL(rcu_scheduler_active); -#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ - -#ifdef CONFIG_RCU_TRACE - -static void check_cpu_stall(struct rcu_ctrlblk *rcp) -{ - unsigned long j; - unsigned long js; - - if (rcu_cpu_stall_suppress) - return; - rcp->ticks_this_gp++; - j = jiffies; - js = rcp->jiffies_stall; - if (*rcp->curtail && ULONG_CMP_GE(j, js)) { - pr_err("INFO: %s stall on CPU (%lu ticks this GP) idle=%llx (t=%lu jiffies q=%ld)\n", - rcp->name, rcp->ticks_this_gp, rcu_dynticks_nesting, - jiffies - rcp->gp_start, rcp->qlen); - dump_stack(); - } - if (*rcp->curtail && ULONG_CMP_GE(j, js)) - rcp->jiffies_stall = jiffies + - 3 * rcu_jiffies_till_stall_check() + 3; - else if (ULONG_CMP_GE(j, js)) - rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); -} - -static void check_cpu_stall_preempt(void); - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) -{ -#ifdef CONFIG_RCU_TRACE - rcp->ticks_this_gp = 0; - rcp->gp_start = jiffies; - rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); -#endif /* #ifdef CONFIG_RCU_TRACE */ -} - -static void check_cpu_stalls(void) -{ - RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); - RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); - RCU_TRACE(check_cpu_stall_preempt()); -} - -#ifdef CONFIG_TINY_PREEMPT_RCU - -#include <linux/delay.h> - -/* Global control variables for preemptible RCU. */ -struct rcu_preempt_ctrlblk { - struct rcu_ctrlblk rcb; /* curtail: ->next ptr of last CB for GP. */ - struct rcu_head **nexttail; - /* Tasks blocked in a preemptible RCU */ - /* read-side critical section while an */ - /* preemptible-RCU grace period is in */ - /* progress must wait for a later grace */ - /* period. This pointer points to the */ - /* ->next pointer of the last task that */ - /* must wait for a later grace period, or */ - /* to &->rcb.rcucblist if there is no */ - /* such task. */ - struct list_head blkd_tasks; - /* Tasks blocked in RCU read-side critical */ - /* section. Tasks are placed at the head */ - /* of this list and age towards the tail. */ - struct list_head *gp_tasks; - /* Pointer to the first task blocking the */ - /* current grace period, or NULL if there */ - /* is no such task. */ - struct list_head *exp_tasks; - /* Pointer to first task blocking the */ - /* current expedited grace period, or NULL */ - /* if there is no such task. If there */ - /* is no current expedited grace period, */ - /* then there cannot be any such task. */ -#ifdef CONFIG_RCU_BOOST - struct list_head *boost_tasks; - /* Pointer to first task that needs to be */ - /* priority-boosted, or NULL if no priority */ - /* boosting is needed. If there is no */ - /* current or expedited grace period, there */ - /* can be no such task. */ -#endif /* #ifdef CONFIG_RCU_BOOST */ - u8 gpnum; /* Current grace period. */ - u8 gpcpu; /* Last grace period blocked by the CPU. */ - u8 completed; /* Last grace period completed. */ - /* If all three are equal, RCU is idle. */ -#ifdef CONFIG_RCU_BOOST - unsigned long boost_time; /* When to start boosting (jiffies) */ -#endif /* #ifdef CONFIG_RCU_BOOST */ -#ifdef CONFIG_RCU_TRACE - unsigned long n_grace_periods; -#ifdef CONFIG_RCU_BOOST - unsigned long n_tasks_boosted; - /* Total number of tasks boosted. */ - unsigned long n_exp_boosts; - /* Number of tasks boosted for expedited GP. */ - unsigned long n_normal_boosts; - /* Number of tasks boosted for normal GP. */ - unsigned long n_balk_blkd_tasks; - /* Refused to boost: no blocked tasks. */ - unsigned long n_balk_exp_gp_tasks; - /* Refused to boost: nothing blocking GP. */ - unsigned long n_balk_boost_tasks; - /* Refused to boost: already boosting. */ - unsigned long n_balk_notyet; - /* Refused to boost: not yet time. */ - unsigned long n_balk_nos; - /* Refused to boost: not sure why, though. */ - /* This can happen due to race conditions. */ -#endif /* #ifdef CONFIG_RCU_BOOST */ -#endif /* #ifdef CONFIG_RCU_TRACE */ -}; - -static struct rcu_preempt_ctrlblk rcu_preempt_ctrlblk = { - .rcb.donetail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .rcb.curtail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .nexttail = &rcu_preempt_ctrlblk.rcb.rcucblist, - .blkd_tasks = LIST_HEAD_INIT(rcu_preempt_ctrlblk.blkd_tasks), - RCU_TRACE(.rcb.name = "rcu_preempt") -}; - -static int rcu_preempted_readers_exp(void); -static void rcu_report_exp_done(void); - -/* - * Return true if the CPU has not yet responded to the current grace period. - */ -static int rcu_cpu_blocking_cur_gp(void) -{ - return rcu_preempt_ctrlblk.gpcpu != rcu_preempt_ctrlblk.gpnum; -} - -/* - * Check for a running RCU reader. Because there is only one CPU, - * there can be but one running RCU reader at a time. ;-) - * - * Returns zero if there are no running readers. Returns a positive - * number if there is at least one reader within its RCU read-side - * critical section. Returns a negative number if an outermost reader - * is in the midst of exiting from its RCU read-side critical section - * - * Returns zero if there are no running readers. Returns a positive - * number if there is at least one reader within its RCU read-side - * critical section. Returns a negative number if an outermost reader - * is in the midst of exiting from its RCU read-side critical section. - */ -static int rcu_preempt_running_reader(void) -{ - return current->rcu_read_lock_nesting; -} - -/* - * Check for preempted RCU readers blocking any grace period. - * If the caller needs a reliable answer, it must disable hard irqs. - */ -static int rcu_preempt_blocked_readers_any(void) -{ - return !list_empty(&rcu_preempt_ctrlblk.blkd_tasks); -} - -/* - * Check for preempted RCU readers blocking the current grace period. - * If the caller needs a reliable answer, it must disable hard irqs. - */ -static int rcu_preempt_blocked_readers_cgp(void) -{ - return rcu_preempt_ctrlblk.gp_tasks != NULL; -} - -/* - * Return true if another preemptible-RCU grace period is needed. - */ -static int rcu_preempt_needs_another_gp(void) -{ - return *rcu_preempt_ctrlblk.rcb.curtail != NULL; -} - -/* - * Return true if a preemptible-RCU grace period is in progress. - * The caller must disable hardirqs. - */ -static int rcu_preempt_gp_in_progress(void) -{ - return rcu_preempt_ctrlblk.completed != rcu_preempt_ctrlblk.gpnum; -} - -/* - * Advance a ->blkd_tasks-list pointer to the next entry, instead - * returning NULL if at the end of the list. - */ -static struct list_head *rcu_next_node_entry(struct task_struct *t) -{ - struct list_head *np; - - np = t->rcu_node_entry.next; - if (np == &rcu_preempt_ctrlblk.blkd_tasks) - np = NULL; - return np; -} - -#ifdef CONFIG_RCU_TRACE - -#ifdef CONFIG_RCU_BOOST -static void rcu_initiate_boost_trace(void); -#endif /* #ifdef CONFIG_RCU_BOOST */ - -/* - * Dump additional statistice for TINY_PREEMPT_RCU. - */ -static void show_tiny_preempt_stats(struct seq_file *m) -{ - seq_printf(m, "rcu_preempt: qlen=%ld gp=%lu g%u/p%u/c%u tasks=%c%c%c\n", - rcu_preempt_ctrlblk.rcb.qlen, - rcu_preempt_ctrlblk.n_grace_periods, - rcu_preempt_ctrlblk.gpnum, - rcu_preempt_ctrlblk.gpcpu, - rcu_preempt_ctrlblk.completed, - "T."[list_empty(&rcu_preempt_ctrlblk.blkd_tasks)], - "N."[!rcu_preempt_ctrlblk.gp_tasks], - "E."[!rcu_preempt_ctrlblk.exp_tasks]); -#ifdef CONFIG_RCU_BOOST - seq_printf(m, "%sttb=%c ntb=%lu neb=%lu nnb=%lu j=%04x bt=%04x\n", - " ", - "B."[!rcu_preempt_ctrlblk.boost_tasks], - rcu_preempt_ctrlblk.n_tasks_boosted, - rcu_preempt_ctrlblk.n_exp_boosts, - rcu_preempt_ctrlblk.n_normal_boosts, - (int)(jiffies & 0xffff), - (int)(rcu_preempt_ctrlblk.boost_time & 0xffff)); - seq_printf(m, "%s: nt=%lu egt=%lu bt=%lu ny=%lu nos=%lu\n", - " balk", - rcu_preempt_ctrlblk.n_balk_blkd_tasks, - rcu_preempt_ctrlblk.n_balk_exp_gp_tasks, - rcu_preempt_ctrlblk.n_balk_boost_tasks, - rcu_preempt_ctrlblk.n_balk_notyet, - rcu_preempt_ctrlblk.n_balk_nos); -#endif /* #ifdef CONFIG_RCU_BOOST */ -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -#ifdef CONFIG_RCU_BOOST - -#include "rtmutex_common.h" - -#define RCU_BOOST_PRIO CONFIG_RCU_BOOST_PRIO - -/* Controls for rcu_kthread() kthread. */ -static struct task_struct *rcu_kthread_task; -static DECLARE_WAIT_QUEUE_HEAD(rcu_kthread_wq); -static unsigned long have_rcu_kthread_work; - -/* - * Carry out RCU priority boosting on the task indicated by ->boost_tasks, - * and advance ->boost_tasks to the next task in the ->blkd_tasks list. - */ -static int rcu_boost(void) -{ - unsigned long flags; - struct rt_mutex mtx; - struct task_struct *t; - struct list_head *tb; - - if (rcu_preempt_ctrlblk.boost_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) - return 0; /* Nothing to boost. */ - - local_irq_save(flags); - - /* - * Recheck with irqs disabled: all tasks in need of boosting - * might exit their RCU read-side critical sections on their own - * if we are preempted just before disabling irqs. - */ - if (rcu_preempt_ctrlblk.boost_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) { - local_irq_restore(flags); - return 0; - } - - /* - * Preferentially boost tasks blocking expedited grace periods. - * This cannot starve the normal grace periods because a second - * expedited grace period must boost all blocked tasks, including - * those blocking the pre-existing normal grace period. - */ - if (rcu_preempt_ctrlblk.exp_tasks != NULL) { - tb = rcu_preempt_ctrlblk.exp_tasks; - RCU_TRACE(rcu_preempt_ctrlblk.n_exp_boosts++); - } else { - tb = rcu_preempt_ctrlblk.boost_tasks; - RCU_TRACE(rcu_preempt_ctrlblk.n_normal_boosts++); - } - RCU_TRACE(rcu_preempt_ctrlblk.n_tasks_boosted++); - - /* - * We boost task t by manufacturing an rt_mutex that appears to - * be held by task t. We leave a pointer to that rt_mutex where - * task t can find it, and task t will release the mutex when it - * exits its outermost RCU read-side critical section. Then - * simply acquiring this artificial rt_mutex will boost task - * t's priority. (Thanks to tglx for suggesting this approach!) - */ - t = container_of(tb, struct task_struct, rcu_node_entry); - rt_mutex_init_proxy_locked(&mtx, t); - t->rcu_boost_mutex = &mtx; - local_irq_restore(flags); - rt_mutex_lock(&mtx); - rt_mutex_unlock(&mtx); /* Keep lockdep happy. */ - - return ACCESS_ONCE(rcu_preempt_ctrlblk.boost_tasks) != NULL || - ACCESS_ONCE(rcu_preempt_ctrlblk.exp_tasks) != NULL; -} - -/* - * Check to see if it is now time to start boosting RCU readers blocking - * the current grace period, and, if so, tell the rcu_kthread_task to - * start boosting them. If there is an expedited boost in progress, - * we wait for it to complete. - * - * If there are no blocked readers blocking the current grace period, - * return 0 to let the caller know, otherwise return 1. Note that this - * return value is independent of whether or not boosting was done. - */ -static int rcu_initiate_boost(void) -{ - if (!rcu_preempt_blocked_readers_cgp() && - rcu_preempt_ctrlblk.exp_tasks == NULL) { - RCU_TRACE(rcu_preempt_ctrlblk.n_balk_exp_gp_tasks++); - return 0; - } - if (rcu_preempt_ctrlblk.exp_tasks != NULL || - (rcu_preempt_ctrlblk.gp_tasks != NULL && - rcu_preempt_ctrlblk.boost_tasks == NULL && - ULONG_CMP_GE(jiffies, rcu_preempt_ctrlblk.boost_time))) { - if (rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_preempt_ctrlblk.boost_tasks = - rcu_preempt_ctrlblk.gp_tasks; - invoke_rcu_callbacks(); - } else { - RCU_TRACE(rcu_initiate_boost_trace()); - } - return 1; -} - -#define RCU_BOOST_DELAY_JIFFIES DIV_ROUND_UP(CONFIG_RCU_BOOST_DELAY * HZ, 1000) - -/* - * Do priority-boost accounting for the start of a new grace period. - */ -static void rcu_preempt_boost_start_gp(void) -{ - rcu_preempt_ctrlblk.boost_time = jiffies + RCU_BOOST_DELAY_JIFFIES; -} - -#else /* #ifdef CONFIG_RCU_BOOST */ - -/* - * If there is no RCU priority boosting, we don't initiate boosting, - * but we do indicate whether there are blocked readers blocking the - * current grace period. - */ -static int rcu_initiate_boost(void) -{ - return rcu_preempt_blocked_readers_cgp(); -} - -/* - * If there is no RCU priority boosting, nothing to do at grace-period start. - */ -static void rcu_preempt_boost_start_gp(void) -{ -} - -#endif /* else #ifdef CONFIG_RCU_BOOST */ - -/* - * Record a preemptible-RCU quiescent state for the specified CPU. Note - * that this just means that the task currently running on the CPU is - * in a quiescent state. There might be any number of tasks blocked - * while in an RCU read-side critical section. - * - * Unlike the other rcu_*_qs() functions, callers to this function - * must disable irqs in order to protect the assignment to - * ->rcu_read_unlock_special. - * - * Because this is a single-CPU implementation, the only way a grace - * period can end is if the CPU is in a quiescent state. The reason is - * that a blocked preemptible-RCU reader can exit its critical section - * only if the CPU is running it at the time. Therefore, when the - * last task blocking the current grace period exits its RCU read-side - * critical section, neither the CPU nor blocked tasks will be stopping - * the current grace period. (In contrast, SMP implementations - * might have CPUs running in RCU read-side critical sections that - * block later grace periods -- but this is not possible given only - * one CPU.) - */ -static void rcu_preempt_cpu_qs(void) -{ - /* Record both CPU and task as having responded to current GP. */ - rcu_preempt_ctrlblk.gpcpu = rcu_preempt_ctrlblk.gpnum; - current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS; - - /* If there is no GP then there is nothing more to do. */ - if (!rcu_preempt_gp_in_progress()) - return; - /* - * Check up on boosting. If there are readers blocking the - * current grace period, leave. - */ - if (rcu_initiate_boost()) - return; - - /* Advance callbacks. */ - rcu_preempt_ctrlblk.completed = rcu_preempt_ctrlblk.gpnum; - rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.rcb.curtail; - rcu_preempt_ctrlblk.rcb.curtail = rcu_preempt_ctrlblk.nexttail; - - /* If there are no blocked readers, next GP is done instantly. */ - if (!rcu_preempt_blocked_readers_any()) - rcu_preempt_ctrlblk.rcb.donetail = rcu_preempt_ctrlblk.nexttail; - - /* If there are done callbacks, cause them to be invoked. */ - if (*rcu_preempt_ctrlblk.rcb.donetail != NULL) - invoke_rcu_callbacks(); -} - -/* - * Start a new RCU grace period if warranted. Hard irqs must be disabled. - */ -static void rcu_preempt_start_gp(void) -{ - if (!rcu_preempt_gp_in_progress() && rcu_preempt_needs_another_gp()) { - - /* Official start of GP. */ - rcu_preempt_ctrlblk.gpnum++; - RCU_TRACE(rcu_preempt_ctrlblk.n_grace_periods++); - reset_cpu_stall_ticks(&rcu_preempt_ctrlblk.rcb); - - /* Any blocked RCU readers block new GP. */ - if (rcu_preempt_blocked_readers_any()) - rcu_preempt_ctrlblk.gp_tasks = - rcu_preempt_ctrlblk.blkd_tasks.next; - - /* Set up for RCU priority boosting. */ - rcu_preempt_boost_start_gp(); - - /* If there is no running reader, CPU is done with GP. */ - if (!rcu_preempt_running_reader()) - rcu_preempt_cpu_qs(); - } -} - -/* - * We have entered the scheduler, and the current task might soon be - * context-switched away from. If this task is in an RCU read-side - * critical section, we will no longer be able to rely on the CPU to - * record that fact, so we enqueue the task on the blkd_tasks list. - * If the task started after the current grace period began, as recorded - * by ->gpcpu, we enqueue at the beginning of the list. Otherwise - * before the element referenced by ->gp_tasks (or at the tail if - * ->gp_tasks is NULL) and point ->gp_tasks at the newly added element. - * The task will dequeue itself when it exits the outermost enclosing - * RCU read-side critical section. Therefore, the current grace period - * cannot be permitted to complete until the ->gp_tasks pointer becomes - * NULL. - * - * Caller must disable preemption. - */ -void rcu_preempt_note_context_switch(void) -{ - struct task_struct *t = current; - unsigned long flags; - - local_irq_save(flags); /* must exclude scheduler_tick(). */ - if (rcu_preempt_running_reader() > 0 && - (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { - - /* Possibly blocking in an RCU read-side critical section. */ - t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; - - /* - * If this CPU has already checked in, then this task - * will hold up the next grace period rather than the - * current grace period. Queue the task accordingly. - * If the task is queued for the current grace period - * (i.e., this CPU has not yet passed through a quiescent - * state for the current grace period), then as long - * as that task remains queued, the current grace period - * cannot end. - */ - list_add(&t->rcu_node_entry, &rcu_preempt_ctrlblk.blkd_tasks); - if (rcu_cpu_blocking_cur_gp()) - rcu_preempt_ctrlblk.gp_tasks = &t->rcu_node_entry; - } else if (rcu_preempt_running_reader() < 0 && - t->rcu_read_unlock_special) { - /* - * Complete exit from RCU read-side critical section on - * behalf of preempted instance of __rcu_read_unlock(). - */ - rcu_read_unlock_special(t); - } - - /* - * Either we were not in an RCU read-side critical section to - * begin with, or we have now recorded that critical section - * globally. Either way, we can now note a quiescent state - * for this CPU. Again, if we were in an RCU read-side critical - * section, and if that critical section was blocking the current - * grace period, then the fact that the task has been enqueued - * means that current grace period continues to be blocked. - */ - rcu_preempt_cpu_qs(); - local_irq_restore(flags); -} - -/* - * Handle special cases during rcu_read_unlock(), such as needing to - * notify RCU core processing or task having blocked during the RCU - * read-side critical section. - */ -void rcu_read_unlock_special(struct task_struct *t) -{ - int empty; - int empty_exp; - unsigned long flags; - struct list_head *np; -#ifdef CONFIG_RCU_BOOST - struct rt_mutex *rbmp = NULL; -#endif /* #ifdef CONFIG_RCU_BOOST */ - int special; - - /* - * NMI handlers cannot block and cannot safely manipulate state. - * They therefore cannot possibly be special, so just leave. - */ - if (in_nmi()) - return; - - local_irq_save(flags); - - /* - * If RCU core is waiting for this CPU to exit critical section, - * let it know that we have done so. - */ - special = t->rcu_read_unlock_special; - if (special & RCU_READ_UNLOCK_NEED_QS) - rcu_preempt_cpu_qs(); - - /* Hardware IRQ handlers cannot block. */ - if (in_irq() || in_serving_softirq()) { - local_irq_restore(flags); - return; - } - - /* Clean up if blocked during RCU read-side critical section. */ - if (special & RCU_READ_UNLOCK_BLOCKED) { - t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED; - - /* - * Remove this task from the ->blkd_tasks list and adjust - * any pointers that might have been referencing it. - */ - empty = !rcu_preempt_blocked_readers_cgp(); - empty_exp = rcu_preempt_ctrlblk.exp_tasks == NULL; - np = rcu_next_node_entry(t); - list_del_init(&t->rcu_node_entry); - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.gp_tasks) - rcu_preempt_ctrlblk.gp_tasks = np; - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.exp_tasks) - rcu_preempt_ctrlblk.exp_tasks = np; -#ifdef CONFIG_RCU_BOOST - if (&t->rcu_node_entry == rcu_preempt_ctrlblk.boost_tasks) - rcu_preempt_ctrlblk.boost_tasks = np; -#endif /* #ifdef CONFIG_RCU_BOOST */ - - /* - * If this was the last task on the current list, and if - * we aren't waiting on the CPU, report the quiescent state - * and start a new grace period if needed. - */ - if (!empty && !rcu_preempt_blocked_readers_cgp()) { - rcu_preempt_cpu_qs(); - rcu_preempt_start_gp(); - } - - /* - * If this was the last task on the expedited lists, - * then we need wake up the waiting task. - */ - if (!empty_exp && rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_report_exp_done(); - } -#ifdef CONFIG_RCU_BOOST - /* Unboost self if was boosted. */ - if (t->rcu_boost_mutex != NULL) { - rbmp = t->rcu_boost_mutex; - t->rcu_boost_mutex = NULL; - rt_mutex_unlock(rbmp); - } -#endif /* #ifdef CONFIG_RCU_BOOST */ - local_irq_restore(flags); -} - -/* - * Check for a quiescent state from the current CPU. When a task blocks, - * the task is recorded in the rcu_preempt_ctrlblk structure, which is - * checked elsewhere. This is called from the scheduling-clock interrupt. - * - * Caller must disable hard irqs. - */ -static void rcu_preempt_check_callbacks(void) -{ - struct task_struct *t = current; - - if (rcu_preempt_gp_in_progress() && - (!rcu_preempt_running_reader() || - !rcu_cpu_blocking_cur_gp())) - rcu_preempt_cpu_qs(); - if (&rcu_preempt_ctrlblk.rcb.rcucblist != - rcu_preempt_ctrlblk.rcb.donetail) - invoke_rcu_callbacks(); - if (rcu_preempt_gp_in_progress() && - rcu_cpu_blocking_cur_gp() && - rcu_preempt_running_reader() > 0) - t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; -} - -/* - * TINY_PREEMPT_RCU has an extra callback-list tail pointer to - * update, so this is invoked from rcu_process_callbacks() to - * handle that case. Of course, it is invoked for all flavors of - * RCU, but RCU callbacks can appear only on one of the lists, and - * neither ->nexttail nor ->donetail can possibly be NULL, so there - * is no need for an explicit check. - */ -static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) -{ - if (rcu_preempt_ctrlblk.nexttail == rcp->donetail) - rcu_preempt_ctrlblk.nexttail = &rcp->rcucblist; -} - -/* - * Process callbacks for preemptible RCU. - */ -static void rcu_preempt_process_callbacks(void) -{ - __rcu_process_callbacks(&rcu_preempt_ctrlblk.rcb); -} - -/* - * Queue a preemptible -RCU callback for invocation after a grace period. - */ -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - - debug_rcu_head_queue(head); - head->func = func; - head->next = NULL; - - local_irq_save(flags); - *rcu_preempt_ctrlblk.nexttail = head; - rcu_preempt_ctrlblk.nexttail = &head->next; - RCU_TRACE(rcu_preempt_ctrlblk.rcb.qlen++); - rcu_preempt_start_gp(); /* checks to see if GP needed. */ - local_irq_restore(flags); -} -EXPORT_SYMBOL_GPL(call_rcu); - -/* - * 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. - */ -void synchronize_rcu(void) -{ - rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && - !lock_is_held(&rcu_lock_map) && - !lock_is_held(&rcu_sched_lock_map), - "Illegal synchronize_rcu() in RCU read-side critical section"); - -#ifdef CONFIG_DEBUG_LOCK_ALLOC - if (!rcu_scheduler_active) - return; -#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ - - WARN_ON_ONCE(rcu_preempt_running_reader()); - if (!rcu_preempt_blocked_readers_any()) - return; - - /* Once we get past the fastpath checks, same code as rcu_barrier(). */ - if (rcu_expedited) - synchronize_rcu_expedited(); - else - rcu_barrier(); -} -EXPORT_SYMBOL_GPL(synchronize_rcu); - -static DECLARE_WAIT_QUEUE_HEAD(sync_rcu_preempt_exp_wq); -static unsigned long sync_rcu_preempt_exp_count; -static DEFINE_MUTEX(sync_rcu_preempt_exp_mutex); - -/* - * Return non-zero if there are any tasks in RCU read-side critical - * sections blocking the current preemptible-RCU expedited grace period. - * If there is no preemptible-RCU expedited grace period currently in - * progress, returns zero unconditionally. - */ -static int rcu_preempted_readers_exp(void) -{ - return rcu_preempt_ctrlblk.exp_tasks != NULL; -} - -/* - * Report the exit from RCU read-side critical section for the last task - * that queued itself during or before the current expedited preemptible-RCU - * grace period. - */ -static void rcu_report_exp_done(void) -{ - wake_up(&sync_rcu_preempt_exp_wq); -} - -/* - * Wait for an rcu-preempt grace period, but expedite it. The basic idea - * is to rely in the fact that there is but one CPU, and that it is - * illegal for a task to invoke synchronize_rcu_expedited() while in a - * preemptible-RCU read-side critical section. Therefore, any such - * critical sections must correspond to blocked tasks, which must therefore - * be on the ->blkd_tasks list. So just record the current head of the - * list in the ->exp_tasks pointer, and wait for all tasks including and - * after the task pointed to by ->exp_tasks to drain. - */ -void synchronize_rcu_expedited(void) -{ - unsigned long flags; - struct rcu_preempt_ctrlblk *rpcp = &rcu_preempt_ctrlblk; - unsigned long snap; - - barrier(); /* ensure prior action seen before grace period. */ - - WARN_ON_ONCE(rcu_preempt_running_reader()); - - /* - * Acquire lock so that there is only one preemptible RCU grace - * period in flight. Of course, if someone does the expedited - * grace period for us while we are acquiring the lock, just leave. - */ - snap = sync_rcu_preempt_exp_count + 1; - mutex_lock(&sync_rcu_preempt_exp_mutex); - if (ULONG_CMP_LT(snap, sync_rcu_preempt_exp_count)) - goto unlock_mb_ret; /* Others did our work for us. */ - - local_irq_save(flags); - - /* - * All RCU readers have to already be on blkd_tasks because - * we cannot legally be executing in an RCU read-side critical - * section. - */ - - /* Snapshot current head of ->blkd_tasks list. */ - rpcp->exp_tasks = rpcp->blkd_tasks.next; - if (rpcp->exp_tasks == &rpcp->blkd_tasks) - rpcp->exp_tasks = NULL; - - /* Wait for tail of ->blkd_tasks list to drain. */ - if (!rcu_preempted_readers_exp()) { - local_irq_restore(flags); - } else { - rcu_initiate_boost(); - local_irq_restore(flags); - wait_event(sync_rcu_preempt_exp_wq, - !rcu_preempted_readers_exp()); - } - - /* Clean up and exit. */ - barrier(); /* ensure expedited GP seen before counter increment. */ - sync_rcu_preempt_exp_count++; -unlock_mb_ret: - mutex_unlock(&sync_rcu_preempt_exp_mutex); - barrier(); /* ensure subsequent action seen after grace period. */ -} -EXPORT_SYMBOL_GPL(synchronize_rcu_expedited); - -/* - * Does preemptible RCU need the CPU to stay out of dynticks mode? - */ -int rcu_preempt_needs_cpu(void) -{ - return rcu_preempt_ctrlblk.rcb.rcucblist != NULL; -} - -#else /* #ifdef CONFIG_TINY_PREEMPT_RCU */ - -#ifdef CONFIG_RCU_TRACE - -/* - * Because preemptible RCU does not exist, it is not necessary to - * dump out its statistics. - */ -static void show_tiny_preempt_stats(struct seq_file *m) -{ -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to check. - */ -static void rcu_preempt_check_callbacks(void) -{ -} - -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to remove. - */ -static void rcu_preempt_remove_callbacks(struct rcu_ctrlblk *rcp) -{ -} - -/* - * Because preemptible RCU does not exist, it never has any callbacks - * to process. - */ -static void rcu_preempt_process_callbacks(void) -{ -} - -#endif /* #else #ifdef CONFIG_TINY_PREEMPT_RCU */ - -#ifdef CONFIG_RCU_BOOST - -/* - * Wake up rcu_kthread() to process callbacks now eligible for invocation - * or to boost readers. - */ -static void invoke_rcu_callbacks(void) -{ - have_rcu_kthread_work = 1; - if (rcu_kthread_task != NULL) - wake_up(&rcu_kthread_wq); -} - -#ifdef CONFIG_RCU_TRACE - -/* - * Is the current CPU running the RCU-callbacks kthread? - * Caller must have preemption disabled. - */ -static bool rcu_is_callbacks_kthread(void) -{ - return rcu_kthread_task == current; -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -/* - * This kthread invokes RCU callbacks whose grace periods have - * elapsed. It is awakened as needed, and takes the place of the - * RCU_SOFTIRQ that is used for this purpose when boosting is disabled. - * This is a kthread, but it is never stopped, at least not until - * the system goes down. - */ -static int rcu_kthread(void *arg) -{ - unsigned long work; - unsigned long morework; - unsigned long flags; - - for (;;) { - wait_event_interruptible(rcu_kthread_wq, - have_rcu_kthread_work != 0); - morework = rcu_boost(); - local_irq_save(flags); - work = have_rcu_kthread_work; - have_rcu_kthread_work = morework; - local_irq_restore(flags); - if (work) - rcu_process_callbacks(NULL); - schedule_timeout_interruptible(1); /* Leave CPU for others. */ - } - - return 0; /* Not reached, but needed to shut gcc up. */ -} - -/* - * Spawn the kthread that invokes RCU callbacks. - */ -static int __init rcu_spawn_kthreads(void) -{ - struct sched_param sp; - - rcu_kthread_task = kthread_run(rcu_kthread, NULL, "rcu_kthread"); - sp.sched_priority = RCU_BOOST_PRIO; - sched_setscheduler_nocheck(rcu_kthread_task, SCHED_FIFO, &sp); - return 0; -} -early_initcall(rcu_spawn_kthreads); - -#else /* #ifdef CONFIG_RCU_BOOST */ - -/* Hold off callback invocation until early_initcall() time. */ -static int rcu_scheduler_fully_active __read_mostly; - -/* - * Start up softirq processing of callbacks. - */ -void invoke_rcu_callbacks(void) -{ - if (rcu_scheduler_fully_active) - raise_softirq(RCU_SOFTIRQ); -} - -#ifdef CONFIG_RCU_TRACE - -/* - * There is no callback kthread, so this thread is never it. - */ -static bool rcu_is_callbacks_kthread(void) -{ - return false; -} - -#endif /* #ifdef CONFIG_RCU_TRACE */ - -static int __init rcu_scheduler_really_started(void) -{ - rcu_scheduler_fully_active = 1; - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); - raise_softirq(RCU_SOFTIRQ); /* Invoke any callbacks from early boot. */ - return 0; -} -early_initcall(rcu_scheduler_really_started); - -#endif /* #else #ifdef CONFIG_RCU_BOOST */ - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -#include <linux/kernel_stat.h> /* * During boot, we forgive RCU lockdep issues. After this function is @@ -1020,25 +72,6 @@ void __init rcu_scheduler_starting(void) #ifdef CONFIG_RCU_TRACE -#ifdef CONFIG_RCU_BOOST - -static void rcu_initiate_boost_trace(void) -{ - if (list_empty(&rcu_preempt_ctrlblk.blkd_tasks)) - rcu_preempt_ctrlblk.n_balk_blkd_tasks++; - else if (rcu_preempt_ctrlblk.gp_tasks == NULL && - rcu_preempt_ctrlblk.exp_tasks == NULL) - rcu_preempt_ctrlblk.n_balk_exp_gp_tasks++; - else if (rcu_preempt_ctrlblk.boost_tasks != NULL) - rcu_preempt_ctrlblk.n_balk_boost_tasks++; - else if (!ULONG_CMP_GE(jiffies, rcu_preempt_ctrlblk.boost_time)) - rcu_preempt_ctrlblk.n_balk_notyet++; - else - rcu_preempt_ctrlblk.n_balk_nos++; -} - -#endif /* #ifdef CONFIG_RCU_BOOST */ - static void rcu_trace_sub_qlen(struct rcu_ctrlblk *rcp, int n) { unsigned long flags; @@ -1053,7 +86,6 @@ static void rcu_trace_sub_qlen(struct rcu_ctrlblk *rcp, int n) */ static int show_tiny_stats(struct seq_file *m, void *unused) { - show_tiny_preempt_stats(m); seq_printf(m, "rcu_sched: qlen: %ld\n", rcu_sched_ctrlblk.qlen); seq_printf(m, "rcu_bh: qlen: %ld\n", rcu_bh_ctrlblk.qlen); return 0; @@ -1103,11 +135,40 @@ MODULE_AUTHOR("Paul E. McKenney"); MODULE_DESCRIPTION("Read-Copy Update tracing for tiny implementation"); MODULE_LICENSE("GPL"); -static void check_cpu_stall_preempt(void) +static void check_cpu_stall(struct rcu_ctrlblk *rcp) { -#ifdef CONFIG_TINY_PREEMPT_RCU - check_cpu_stall(&rcu_preempt_ctrlblk.rcb); -#endif /* #ifdef CONFIG_TINY_PREEMPT_RCU */ + unsigned long j; + unsigned long js; + + if (rcu_cpu_stall_suppress) + return; + rcp->ticks_this_gp++; + j = jiffies; + js = rcp->jiffies_stall; + if (*rcp->curtail && ULONG_CMP_GE(j, js)) { + pr_err("INFO: %s stall on CPU (%lu ticks this GP) idle=%llx (t=%lu jiffies q=%ld)\n", + rcp->name, rcp->ticks_this_gp, rcu_dynticks_nesting, + jiffies - rcp->gp_start, rcp->qlen); + dump_stack(); + } + if (*rcp->curtail && ULONG_CMP_GE(j, js)) + rcp->jiffies_stall = jiffies + + 3 * rcu_jiffies_till_stall_check() + 3; + else if (ULONG_CMP_GE(j, js)) + rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); +} + +static void reset_cpu_stall_ticks(struct rcu_ctrlblk *rcp) +{ + rcp->ticks_this_gp = 0; + rcp->gp_start = jiffies; + rcp->jiffies_stall = jiffies + rcu_jiffies_till_stall_check(); +} + +static void check_cpu_stalls(void) +{ + RCU_TRACE(check_cpu_stall(&rcu_bh_ctrlblk)); + RCU_TRACE(check_cpu_stall(&rcu_sched_ctrlblk)); } #endif /* #ifdef CONFIG_RCU_TRACE */ diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index e1f3a8c96724..b1fa5510388d 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -695,44 +695,6 @@ static struct rcu_torture_ops srcu_sync_ops = { .name = "srcu_sync" }; -static int srcu_torture_read_lock_raw(void) __acquires(&srcu_ctl) -{ - return srcu_read_lock_raw(&srcu_ctl); -} - -static void srcu_torture_read_unlock_raw(int idx) __releases(&srcu_ctl) -{ - srcu_read_unlock_raw(&srcu_ctl, idx); -} - -static struct rcu_torture_ops srcu_raw_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock_raw, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock_raw, - .completed = srcu_torture_completed, - .deferred_free = srcu_torture_deferred_free, - .sync = srcu_torture_synchronize, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_raw" -}; - -static struct rcu_torture_ops srcu_raw_sync_ops = { - .init = rcu_sync_torture_init, - .readlock = srcu_torture_read_lock_raw, - .read_delay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock_raw, - .completed = srcu_torture_completed, - .deferred_free = rcu_sync_torture_deferred_free, - .sync = srcu_torture_synchronize, - .call = NULL, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu_raw_sync" -}; - static void srcu_torture_synchronize_expedited(void) { synchronize_srcu_expedited(&srcu_ctl); @@ -1983,7 +1945,6 @@ rcu_torture_init(void) { &rcu_ops, &rcu_sync_ops, &rcu_expedited_ops, &rcu_bh_ops, &rcu_bh_sync_ops, &rcu_bh_expedited_ops, &srcu_ops, &srcu_sync_ops, &srcu_expedited_ops, - &srcu_raw_ops, &srcu_raw_sync_ops, &sched_ops, &sched_sync_ops, &sched_expedited_ops, }; mutex_lock(&fullstop_mutex); diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 16ea67925015..e08abb9461ac 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -218,8 +218,8 @@ module_param(blimit, long, 0444); module_param(qhimark, long, 0444); module_param(qlowmark, long, 0444); -static ulong jiffies_till_first_fqs = RCU_JIFFIES_TILL_FORCE_QS; -static ulong jiffies_till_next_fqs = RCU_JIFFIES_TILL_FORCE_QS; +static ulong jiffies_till_first_fqs = ULONG_MAX; +static ulong jiffies_till_next_fqs = ULONG_MAX; module_param(jiffies_till_first_fqs, ulong, 0644); module_param(jiffies_till_next_fqs, ulong, 0644); @@ -866,7 +866,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks:", + pr_err("INFO: %s detected stalls on CPUs/tasks:", rsp->name); print_cpu_stall_info_begin(); rcu_for_each_leaf_node(rsp, rnp) { @@ -899,7 +899,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) smp_processor_id(), (long)(jiffies - rsp->gp_start), rsp->gpnum, rsp->completed, totqlen); if (ndetected == 0) - printk(KERN_ERR "INFO: Stall ended before state dump start\n"); + pr_err("INFO: Stall ended before state dump start\n"); else if (!trigger_all_cpu_backtrace()) rcu_dump_cpu_stacks(rsp); @@ -922,7 +922,7 @@ static void print_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s self-detected stall on CPU", rsp->name); + pr_err("INFO: %s self-detected stall on CPU", rsp->name); print_cpu_stall_info_begin(); print_cpu_stall_info(rsp, smp_processor_id()); print_cpu_stall_info_end(); @@ -985,65 +985,6 @@ void rcu_cpu_stall_reset(void) } /* - * Update CPU-local rcu_data state to record the newly noticed grace period. - * This is used both when we started the grace period and when we notice - * that someone else started the grace period. The caller must hold the - * ->lock of the leaf rcu_node structure corresponding to the current CPU, - * and must have irqs disabled. - */ -static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) -{ - if (rdp->gpnum != rnp->gpnum) { - /* - * If the current grace period is waiting for this CPU, - * set up to detect a quiescent state, otherwise don't - * go looking for one. - */ - rdp->gpnum = rnp->gpnum; - trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); - rdp->passed_quiesce = 0; - rdp->qs_pending = !!(rnp->qsmask & rdp->grpmask); - zero_cpu_stall_ticks(rdp); - } -} - -static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - struct rcu_node *rnp; - - local_irq_save(flags); - rnp = rdp->mynode; - if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ - !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ - local_irq_restore(flags); - return; - } - __note_new_gpnum(rsp, rnp, rdp); - raw_spin_unlock_irqrestore(&rnp->lock, flags); -} - -/* - * Did someone else start a new RCU grace period start since we last - * checked? Update local state appropriately if so. Must be called - * on the CPU corresponding to rdp. - */ -static int -check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) -{ - unsigned long flags; - int ret = 0; - - local_irq_save(flags); - if (rdp->gpnum != rsp->gpnum) { - note_new_gpnum(rsp, rdp); - ret = 1; - } - local_irq_restore(flags); - return ret; -} - -/* * Initialize the specified rcu_data structure's callback list to empty. */ static void init_callback_list(struct rcu_data *rdp) @@ -1313,18 +1254,16 @@ static void rcu_advance_cbs(struct rcu_state *rsp, struct rcu_node *rnp, } /* - * Advance this CPU's callbacks, but only if the current grace period - * has ended. This may be called only from the CPU to whom the rdp - * belongs. In addition, the corresponding leaf rcu_node structure's - * ->lock must be held by the caller, with irqs disabled. + * Update CPU-local rcu_data state to record the beginnings and ends of + * grace periods. The caller must hold the ->lock of the leaf rcu_node + * structure corresponding to the current CPU, and must have irqs disabled. */ -static void -__rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) +static void __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { - /* Did another grace period end? */ + /* Handle the ends of any preceding grace periods first. */ if (rdp->completed == rnp->completed) { - /* No, so just accelerate recent callbacks. */ + /* No grace period end, so just accelerate recent callbacks. */ rcu_accelerate_cbs(rsp, rnp, rdp); } else { @@ -1335,68 +1274,40 @@ __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat /* Remember that we saw this grace-period completion. */ rdp->completed = rnp->completed; trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuend"); + } + if (rdp->gpnum != rnp->gpnum) { /* - * If we were in an extended quiescent state, we may have - * missed some grace periods that others CPUs handled on - * our behalf. Catch up with this state to avoid noting - * spurious new grace periods. If another grace period - * has started, then rnp->gpnum will have advanced, so - * we will detect this later on. Of course, any quiescent - * states we found for the old GP are now invalid. - */ - if (ULONG_CMP_LT(rdp->gpnum, rdp->completed)) { - rdp->gpnum = rdp->completed; - rdp->passed_quiesce = 0; - } - - /* - * If RCU does not need a quiescent state from this CPU, - * then make sure that this CPU doesn't go looking for one. + * If the current grace period is waiting for this CPU, + * set up to detect a quiescent state, otherwise don't + * go looking for one. */ - if ((rnp->qsmask & rdp->grpmask) == 0) - rdp->qs_pending = 0; + rdp->gpnum = rnp->gpnum; + trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); + rdp->passed_quiesce = 0; + rdp->qs_pending = !!(rnp->qsmask & rdp->grpmask); + zero_cpu_stall_ticks(rdp); } } -/* - * Advance this CPU's callbacks, but only if the current grace period - * has ended. This may be called only from the CPU to whom the rdp - * belongs. - */ -static void -rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) +static void note_gp_changes(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; struct rcu_node *rnp; local_irq_save(flags); rnp = rdp->mynode; - if (rdp->completed == ACCESS_ONCE(rnp->completed) || /* outside lock. */ + if ((rdp->gpnum == ACCESS_ONCE(rnp->gpnum) && + rdp->completed == ACCESS_ONCE(rnp->completed)) || /* w/out lock. */ !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } - __rcu_process_gp_end(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* - * Do per-CPU grace-period initialization for running CPU. The caller - * must hold the lock of the leaf rcu_node structure corresponding to - * this CPU. - */ -static void -rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) -{ - /* Prior grace period ended, so advance callbacks for current CPU. */ - __rcu_process_gp_end(rsp, rnp, rdp); - - /* Set state so that this CPU will detect the next quiescent state. */ - __note_new_gpnum(rsp, rnp, rdp); -} - -/* * Initialize a new grace period. */ static int rcu_gp_init(struct rcu_state *rsp) @@ -1444,16 +1355,16 @@ static int rcu_gp_init(struct rcu_state *rsp) WARN_ON_ONCE(rnp->completed != rsp->completed); ACCESS_ONCE(rnp->completed) = rsp->completed; if (rnp == rdp->mynode) - rcu_start_gp_per_cpu(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); rcu_preempt_boost_start_gp(rnp); trace_rcu_grace_period_init(rsp->name, rnp->gpnum, rnp->level, rnp->grplo, rnp->grphi, rnp->qsmask); raw_spin_unlock_irq(&rnp->lock); #ifdef CONFIG_PROVE_RCU_DELAY - if ((prandom_u32() % (rcu_num_nodes * 8)) == 0 && + if ((prandom_u32() % (rcu_num_nodes + 1)) == 0 && system_state == SYSTEM_RUNNING) - schedule_timeout_uninterruptible(2); + udelay(200); #endif /* #ifdef CONFIG_PROVE_RCU_DELAY */ cond_resched(); } @@ -1527,7 +1438,7 @@ static void rcu_gp_cleanup(struct rcu_state *rsp) ACCESS_ONCE(rnp->completed) = rsp->gpnum; rdp = this_cpu_ptr(rsp->rda); if (rnp == rdp->mynode) - __rcu_process_gp_end(rsp, rnp, rdp); + __note_gp_changes(rsp, rnp, rdp); nocb += rcu_future_gp_cleanup(rsp, rnp); raw_spin_unlock_irq(&rnp->lock); cond_resched(); @@ -1613,6 +1524,14 @@ static int __noreturn rcu_gp_kthread(void *arg) } } +static void rsp_wakeup(struct irq_work *work) +{ + struct rcu_state *rsp = container_of(work, struct rcu_state, wakeup_work); + + /* Wake up rcu_gp_kthread() to start the grace period. */ + wake_up(&rsp->gp_wq); +} + /* * Start a new RCU grace period if warranted, re-initializing the hierarchy * in preparation for detecting the next grace period. The caller must hold @@ -1637,8 +1556,12 @@ rcu_start_gp_advanced(struct rcu_state *rsp, struct rcu_node *rnp, } rsp->gp_flags = RCU_GP_FLAG_INIT; - /* Wake up rcu_gp_kthread() to start the grace period. */ - wake_up(&rsp->gp_wq); + /* + * We can't do wakeups while holding the rnp->lock, as that + * could cause possible deadlocks with the rq->lock. Deter + * the wakeup to interrupt context. + */ + irq_work_queue(&rsp->wakeup_work); } /* @@ -1793,9 +1716,8 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp) static void rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) { - /* If there is now a new grace period, record and return. */ - if (check_for_new_grace_period(rsp, rdp)) - return; + /* Check for grace-period ends and beginnings. */ + note_gp_changes(rsp, rdp); /* * Does this CPU still need to do its part for current grace period? @@ -2259,9 +2181,6 @@ __rcu_process_callbacks(struct rcu_state *rsp) WARN_ON_ONCE(rdp->beenonline == 0); - /* Handle the end of a grace period that some other CPU ended. */ - rcu_process_gp_end(rsp, rdp); - /* Update RCU state based on any recent quiescent states. */ rcu_check_quiescent_state(rsp, rdp); @@ -2346,8 +2265,7 @@ static void __call_rcu_core(struct rcu_state *rsp, struct rcu_data *rdp, if (unlikely(rdp->qlen > rdp->qlen_last_fqs_check + qhimark)) { /* Are we ignoring a completed grace period? */ - rcu_process_gp_end(rsp, rdp); - check_for_new_grace_period(rsp, rdp); + note_gp_changes(rsp, rdp); /* Start a new grace period if one not already started. */ if (!rcu_gp_in_progress(rsp)) { @@ -3108,7 +3026,7 @@ static int __init rcu_spawn_gp_kthread(void) struct task_struct *t; for_each_rcu_flavor(rsp) { - t = kthread_run(rcu_gp_kthread, rsp, rsp->name); + t = kthread_run(rcu_gp_kthread, rsp, "%s", rsp->name); BUG_ON(IS_ERR(t)); rnp = rcu_get_root(rsp); raw_spin_lock_irqsave(&rnp->lock, flags); @@ -3235,6 +3153,7 @@ static void __init rcu_init_one(struct rcu_state *rsp, rsp->rda = rda; init_waitqueue_head(&rsp->gp_wq); + init_irq_work(&rsp->wakeup_work, rsp_wakeup); rnp = rsp->level[rcu_num_lvls - 1]; for_each_possible_cpu(i) { while (i > rnp->grphi) @@ -3252,11 +3171,25 @@ static void __init rcu_init_one(struct rcu_state *rsp, */ static void __init rcu_init_geometry(void) { + ulong d; int i; int j; int n = nr_cpu_ids; int rcu_capacity[MAX_RCU_LVLS + 1]; + /* + * Initialize any unspecified boot parameters. + * The default values of jiffies_till_first_fqs and + * jiffies_till_next_fqs are set to the RCU_JIFFIES_TILL_FORCE_QS + * value, which is a function of HZ, then adding one for each + * RCU_JIFFIES_FQS_DIV CPUs that might be on the system. + */ + d = RCU_JIFFIES_TILL_FORCE_QS + nr_cpu_ids / RCU_JIFFIES_FQS_DIV; + if (jiffies_till_first_fqs == ULONG_MAX) + jiffies_till_first_fqs = d; + if (jiffies_till_next_fqs == ULONG_MAX) + jiffies_till_next_fqs = d; + /* If the compile-time values are accurate, just leave. */ if (rcu_fanout_leaf == CONFIG_RCU_FANOUT_LEAF && nr_cpu_ids == NR_CPUS) diff --git a/kernel/rcutree.h b/kernel/rcutree.h index da77a8f57ff9..4a39d364493c 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -27,6 +27,7 @@ #include <linux/threads.h> #include <linux/cpumask.h> #include <linux/seqlock.h> +#include <linux/irq_work.h> /* * Define shape of hierarchy based on NR_CPUS, CONFIG_RCU_FANOUT, and @@ -342,12 +343,17 @@ struct rcu_data { #define RCU_FORCE_QS 3 /* Need to force quiescent state. */ #define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK -#define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ +#define RCU_JIFFIES_TILL_FORCE_QS (1 + (HZ > 250) + (HZ > 500)) + /* For jiffies_till_first_fqs and */ + /* and jiffies_till_next_fqs. */ -#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ - /* to take at least one */ - /* scheduling clock irq */ - /* before ratting on them. */ +#define RCU_JIFFIES_FQS_DIV 256 /* Very large systems need more */ + /* delay between bouts of */ + /* quiescent-state forcing. */ + +#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time to take */ + /* at least one scheduling clock */ + /* irq before ratting on them. */ #define rcu_wait(cond) \ do { \ @@ -442,6 +448,7 @@ struct rcu_state { char *name; /* Name of structure. */ char abbr; /* Abbreviated name. */ struct list_head flavors; /* List of RCU flavors. */ + struct irq_work wakeup_work; /* Postponed wakeups */ }; /* Values for rcu_state structure's gp_flags field. */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 3db5a375d8dd..63098a59216e 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -53,38 +53,37 @@ static char __initdata nocb_buf[NR_CPUS * 5]; static void __init rcu_bootup_announce_oddness(void) { #ifdef CONFIG_RCU_TRACE - printk(KERN_INFO "\tRCU debugfs-based tracing is enabled.\n"); + pr_info("\tRCU debugfs-based tracing is enabled.\n"); #endif #if (defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 64) || (!defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 32) - printk(KERN_INFO "\tCONFIG_RCU_FANOUT set to non-default value of %d\n", + pr_info("\tCONFIG_RCU_FANOUT set to non-default value of %d\n", CONFIG_RCU_FANOUT); #endif #ifdef CONFIG_RCU_FANOUT_EXACT - printk(KERN_INFO "\tHierarchical RCU autobalancing is disabled.\n"); + pr_info("\tHierarchical RCU autobalancing is disabled.\n"); #endif #ifdef CONFIG_RCU_FAST_NO_HZ - printk(KERN_INFO - "\tRCU dyntick-idle grace-period acceleration is enabled.\n"); + pr_info("\tRCU dyntick-idle grace-period acceleration is enabled.\n"); #endif #ifdef CONFIG_PROVE_RCU - printk(KERN_INFO "\tRCU lockdep checking is enabled.\n"); + pr_info("\tRCU lockdep checking is enabled.\n"); #endif #ifdef CONFIG_RCU_TORTURE_TEST_RUNNABLE - printk(KERN_INFO "\tRCU torture testing starts during boot.\n"); + pr_info("\tRCU torture testing starts during boot.\n"); #endif #if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE) - printk(KERN_INFO "\tDump stacks of tasks blocking RCU-preempt GP.\n"); + pr_info("\tDump stacks of tasks blocking RCU-preempt GP.\n"); #endif #if defined(CONFIG_RCU_CPU_STALL_INFO) - printk(KERN_INFO "\tAdditional per-CPU info printed with stalls.\n"); + pr_info("\tAdditional per-CPU info printed with stalls.\n"); #endif #if NUM_RCU_LVL_4 != 0 - printk(KERN_INFO "\tFour-level hierarchy is enabled.\n"); + pr_info("\tFour-level hierarchy is enabled.\n"); #endif if (rcu_fanout_leaf != CONFIG_RCU_FANOUT_LEAF) - printk(KERN_INFO "\tExperimental boot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); + pr_info("\tBoot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf); if (nr_cpu_ids != NR_CPUS) - printk(KERN_INFO "\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids); + pr_info("\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids); #ifdef CONFIG_RCU_NOCB_CPU #ifndef CONFIG_RCU_NOCB_CPU_NONE if (!have_rcu_nocb_mask) { @@ -92,19 +91,19 @@ static void __init rcu_bootup_announce_oddness(void) have_rcu_nocb_mask = true; } #ifdef CONFIG_RCU_NOCB_CPU_ZERO - pr_info("\tExperimental no-CBs CPU 0\n"); + pr_info("\tOffload RCU callbacks from CPU 0\n"); cpumask_set_cpu(0, rcu_nocb_mask); #endif /* #ifdef CONFIG_RCU_NOCB_CPU_ZERO */ #ifdef CONFIG_RCU_NOCB_CPU_ALL - pr_info("\tExperimental no-CBs for all CPUs\n"); + pr_info("\tOffload RCU callbacks from all CPUs\n"); cpumask_setall(rcu_nocb_mask); #endif /* #ifdef CONFIG_RCU_NOCB_CPU_ALL */ #endif /* #ifndef CONFIG_RCU_NOCB_CPU_NONE */ if (have_rcu_nocb_mask) { cpulist_scnprintf(nocb_buf, sizeof(nocb_buf), rcu_nocb_mask); - pr_info("\tExperimental no-CBs CPUs: %s.\n", nocb_buf); + pr_info("\tOffload RCU callbacks from CPUs: %s.\n", nocb_buf); if (rcu_nocb_poll) - pr_info("\tExperimental polled no-CBs CPUs.\n"); + pr_info("\tPoll for callbacks from no-CBs CPUs.\n"); } #endif /* #ifdef CONFIG_RCU_NOCB_CPU */ } @@ -123,7 +122,7 @@ static int rcu_preempted_readers_exp(struct rcu_node *rnp); */ static void __init rcu_bootup_announce(void) { - printk(KERN_INFO "Preemptible hierarchical RCU implementation.\n"); + pr_info("Preemptible hierarchical RCU implementation.\n"); rcu_bootup_announce_oddness(); } @@ -490,13 +489,13 @@ static void rcu_print_detail_task_stall(struct rcu_state *rsp) static void rcu_print_task_stall_begin(struct rcu_node *rnp) { - printk(KERN_ERR "\tTasks blocked on level-%d rcu_node (CPUs %d-%d):", + pr_err("\tTasks blocked on level-%d rcu_node (CPUs %d-%d):", rnp->level, rnp->grplo, rnp->grphi); } static void rcu_print_task_stall_end(void) { - printk(KERN_CONT "\n"); + pr_cont("\n"); } #else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */ @@ -526,7 +525,7 @@ static int rcu_print_task_stall(struct rcu_node *rnp) t = list_entry(rnp->gp_tasks, struct task_struct, rcu_node_entry); list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry) { - printk(KERN_CONT " P%d", t->pid); + pr_cont(" P%d", t->pid); ndetected++; } rcu_print_task_stall_end(); @@ -933,6 +932,24 @@ static void __init __rcu_init_preempt(void) rcu_init_one(&rcu_preempt_state, &rcu_preempt_data); } +/* + * Check for a task exiting while in a preemptible-RCU read-side + * critical section, clean up if so. No need to issue warnings, + * as debug_check_no_locks_held() already does this if lockdep + * is enabled. + */ +void exit_rcu(void) +{ + struct task_struct *t = current; + + if (likely(list_empty(¤t->rcu_node_entry))) + return; + t->rcu_read_lock_nesting = 1; + barrier(); + t->rcu_read_unlock_special = RCU_READ_UNLOCK_BLOCKED; + __rcu_read_unlock(); +} + #else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ static struct rcu_state *rcu_state = &rcu_sched_state; @@ -942,7 +959,7 @@ static struct rcu_state *rcu_state = &rcu_sched_state; */ static void __init rcu_bootup_announce(void) { - printk(KERN_INFO "Hierarchical RCU implementation.\n"); + pr_info("Hierarchical RCU implementation.\n"); rcu_bootup_announce_oddness(); } @@ -1101,6 +1118,14 @@ static void __init __rcu_init_preempt(void) { } +/* + * Because preemptible RCU does not exist, tasks cannot possibly exit + * while in preemptible RCU read-side critical sections. + */ +void exit_rcu(void) +{ +} + #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ #ifdef CONFIG_RCU_BOOST @@ -1629,7 +1654,7 @@ static bool rcu_try_advance_all_cbs(void) */ if (rdp->completed != rnp->completed && rdp->nxttail[RCU_DONE_TAIL] != rdp->nxttail[RCU_NEXT_TAIL]) - rcu_process_gp_end(rsp, rdp); + note_gp_changes(rsp, rdp); if (cpu_has_callbacks_ready_to_invoke(rdp)) cbs_ready = true; @@ -1883,7 +1908,7 @@ static void print_cpu_stall_fast_no_hz(char *cp, int cpu) /* Initiate the stall-info list. */ static void print_cpu_stall_info_begin(void) { - printk(KERN_CONT "\n"); + pr_cont("\n"); } /* @@ -1914,7 +1939,7 @@ static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) ticks_value = rsp->gpnum - rdp->gpnum; } print_cpu_stall_fast_no_hz(fast_no_hz, cpu); - printk(KERN_ERR "\t%d: (%lu %s) idle=%03x/%llx/%d softirq=%u/%u %s\n", + pr_err("\t%d: (%lu %s) idle=%03x/%llx/%d softirq=%u/%u %s\n", cpu, ticks_value, ticks_title, atomic_read(&rdtp->dynticks) & 0xfff, rdtp->dynticks_nesting, rdtp->dynticks_nmi_nesting, @@ -1925,7 +1950,7 @@ static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) /* Terminate the stall-info list. */ static void print_cpu_stall_info_end(void) { - printk(KERN_ERR "\t"); + pr_err("\t"); } /* Zero ->ticks_this_gp for all flavors of RCU. */ @@ -1948,17 +1973,17 @@ static void increment_cpu_stall_ticks(void) static void print_cpu_stall_info_begin(void) { - printk(KERN_CONT " {"); + pr_cont(" {"); } static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) { - printk(KERN_CONT " %d", cpu); + pr_cont(" %d", cpu); } static void print_cpu_stall_info_end(void) { - printk(KERN_CONT "} "); + pr_cont("} "); } static void zero_cpu_stall_ticks(struct rcu_data *rdp) diff --git a/kernel/resource.c b/kernel/resource.c index d7386986e10e..3f285dce9347 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -409,6 +409,7 @@ int __weak page_is_ram(unsigned long pfn) { return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; } +EXPORT_SYMBOL_GPL(page_is_ram); void __weak arch_remove_reservations(struct resource *avail) { @@ -448,7 +449,6 @@ static int __find_resource(struct resource *root, struct resource *old, struct resource *this = root->child; struct resource tmp = *new, avail, alloc; - tmp.flags = new->flags; tmp.start = root->start; /* * Skip past an allocated resource that starts at 0, since the assignment diff --git a/kernel/rtmutex.c b/kernel/rtmutex.c index 1e09308bf2a1..0dd6aec1cb6a 100644 --- a/kernel/rtmutex.c +++ b/kernel/rtmutex.c @@ -145,6 +145,19 @@ int max_lock_depth = 1024; /* * Adjust the priority chain. Also used for deadlock detection. * Decreases task's usage by one - may thus free the task. + * + * @task: the task owning the mutex (owner) for which a chain walk is probably + * needed + * @deadlock_detect: do we have to carry out deadlock detection? + * @orig_lock: the mutex (can be NULL if we are walking the chain to recheck + * things for a task that has just got its priority adjusted, and + * is waiting on a mutex) + * @orig_waiter: rt_mutex_waiter struct for the task that has just donated + * its priority to the mutex owner (can be NULL in the case + * depicted above or if the top waiter is gone away and we are + * actually deboosting the owner) + * @top_task: the current top waiter + * * Returns 0 or -EDEADLK. */ static int rt_mutex_adjust_prio_chain(struct task_struct *task, diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile index deaf90e4a1de..54adcf35f495 100644 --- a/kernel/sched/Makefile +++ b/kernel/sched/Makefile @@ -11,7 +11,7 @@ ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) CFLAGS_core.o := $(PROFILING) -fno-omit-frame-pointer endif -obj-y += core.o clock.o cputime.o idle_task.o fair.o rt.o stop_task.o +obj-y += core.o proc.o clock.o cputime.o idle_task.o fair.o rt.o stop_task.o obj-$(CONFIG_SMP) += cpupri.o obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o obj-$(CONFIG_SCHEDSTATS) += stats.o diff --git a/kernel/sched/auto_group.c b/kernel/sched/auto_group.c index 64de5f8b0c9e..4a073539c58e 100644 --- a/kernel/sched/auto_group.c +++ b/kernel/sched/auto_group.c @@ -77,8 +77,6 @@ static inline struct autogroup *autogroup_create(void) if (IS_ERR(tg)) goto out_free; - sched_online_group(tg, &root_task_group); - kref_init(&ag->kref); init_rwsem(&ag->lock); ag->id = atomic_inc_return(&autogroup_seq_nr); @@ -98,6 +96,7 @@ static inline struct autogroup *autogroup_create(void) #endif tg->autogroup = ag; + sched_online_group(tg, &root_task_group); return ag; out_free: diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 58453b8272fd..9b1f2e533b95 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -633,7 +633,19 @@ void wake_up_nohz_cpu(int cpu) static inline bool got_nohz_idle_kick(void) { int cpu = smp_processor_id(); - return idle_cpu(cpu) && test_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu)); + + if (!test_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu))) + return false; + + if (idle_cpu(cpu) && !need_resched()) + return true; + + /* + * We can't run Idle Load Balance on this CPU for this time so we + * cancel it and clear NOHZ_BALANCE_KICK + */ + clear_bit(NOHZ_BALANCE_KICK, nohz_flags(cpu)); + return false; } #else /* CONFIG_NO_HZ_COMMON */ @@ -667,7 +679,7 @@ void sched_avg_update(struct rq *rq) { s64 period = sched_avg_period(); - while ((s64)(rq->clock - rq->age_stamp) > period) { + while ((s64)(rq_clock(rq) - rq->age_stamp) > period) { /* * Inline assembly required to prevent the compiler * optimising this loop into a divmod call. @@ -1328,7 +1340,7 @@ ttwu_do_wakeup(struct rq *rq, struct task_struct *p, int wake_flags) p->sched_class->task_woken(rq, p); if (rq->idle_stamp) { - u64 delta = rq->clock - rq->idle_stamp; + u64 delta = rq_clock(rq) - rq->idle_stamp; u64 max = 2*sysctl_sched_migration_cost; if (delta > max) @@ -1365,6 +1377,8 @@ static int ttwu_remote(struct task_struct *p, int wake_flags) rq = __task_rq_lock(p); if (p->on_rq) { + /* check_preempt_curr() may use rq clock */ + update_rq_clock(rq); ttwu_do_wakeup(rq, p, wake_flags); ret = 1; } @@ -1393,8 +1407,9 @@ static void sched_ttwu_pending(void) void scheduler_ipi(void) { - if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick() - && !tick_nohz_full_cpu(smp_processor_id())) + if (llist_empty(&this_rq()->wake_list) + && !tick_nohz_full_cpu(smp_processor_id()) + && !got_nohz_idle_kick()) return; /* @@ -1417,7 +1432,7 @@ void scheduler_ipi(void) /* * Check if someone kicked us for doing the nohz idle load balance. */ - if (unlikely(got_nohz_idle_kick() && !need_resched())) { + if (unlikely(got_nohz_idle_kick())) { this_rq()->idle_balance = 1; raise_softirq_irqoff(SCHED_SOFTIRQ); } @@ -1596,15 +1611,6 @@ static void __sched_fork(struct task_struct *p) p->se.vruntime = 0; INIT_LIST_HEAD(&p->se.group_node); -/* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) - p->se.avg.runnable_avg_period = 0; - p->se.avg.runnable_avg_sum = 0; -#endif #ifdef CONFIG_SCHEDSTATS memset(&p->se.statistics, 0, sizeof(p->se.statistics)); #endif @@ -1748,6 +1754,8 @@ void wake_up_new_task(struct task_struct *p) set_task_cpu(p, select_task_rq(p, SD_BALANCE_FORK, 0)); #endif + /* Initialize new task's runnable average */ + init_task_runnable_average(p); rq = __task_rq_lock(p); activate_task(rq, p, 0); p->on_rq = 1; @@ -2056,575 +2064,6 @@ unsigned long nr_iowait_cpu(int cpu) return atomic_read(&this->nr_iowait); } -unsigned long this_cpu_load(void) -{ - struct rq *this = this_rq(); - return this->cpu_load[0]; -} - - -/* - * Global load-average calculations - * - * We take a distributed and async approach to calculating the global load-avg - * in order to minimize overhead. - * - * The global load average is an exponentially decaying average of nr_running + - * nr_uninterruptible. - * - * Once every LOAD_FREQ: - * - * nr_active = 0; - * for_each_possible_cpu(cpu) - * nr_active += cpu_of(cpu)->nr_running + cpu_of(cpu)->nr_uninterruptible; - * - * avenrun[n] = avenrun[0] * exp_n + nr_active * (1 - exp_n) - * - * Due to a number of reasons the above turns in the mess below: - * - * - for_each_possible_cpu() is prohibitively expensive on machines with - * serious number of cpus, therefore we need to take a distributed approach - * to calculating nr_active. - * - * \Sum_i x_i(t) = \Sum_i x_i(t) - x_i(t_0) | x_i(t_0) := 0 - * = \Sum_i { \Sum_j=1 x_i(t_j) - x_i(t_j-1) } - * - * So assuming nr_active := 0 when we start out -- true per definition, we - * can simply take per-cpu deltas and fold those into a global accumulate - * to obtain the same result. See calc_load_fold_active(). - * - * Furthermore, in order to avoid synchronizing all per-cpu delta folding - * across the machine, we assume 10 ticks is sufficient time for every - * cpu to have completed this task. - * - * This places an upper-bound on the IRQ-off latency of the machine. Then - * again, being late doesn't loose the delta, just wrecks the sample. - * - * - cpu_rq()->nr_uninterruptible isn't accurately tracked per-cpu because - * this would add another cross-cpu cacheline miss and atomic operation - * to the wakeup path. Instead we increment on whatever cpu the task ran - * when it went into uninterruptible state and decrement on whatever cpu - * did the wakeup. This means that only the sum of nr_uninterruptible over - * all cpus yields the correct result. - * - * This covers the NO_HZ=n code, for extra head-aches, see the comment below. - */ - -/* 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); /* should be removed */ - -/** - * 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 long calc_load_fold_active(struct rq *this_rq) -{ - long nr_active, delta = 0; - - 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; - } - - return delta; -} - -/* - * a1 = a0 * e + a * (1 - e) - */ -static unsigned long -calc_load(unsigned long load, unsigned long exp, unsigned long active) -{ - load *= exp; - load += active * (FIXED_1 - exp); - load += 1UL << (FSHIFT - 1); - return load >> FSHIFT; -} - -#ifdef CONFIG_NO_HZ_COMMON -/* - * Handle NO_HZ for the global load-average. - * - * Since the above described distributed algorithm to compute the global - * load-average relies on per-cpu sampling from the tick, it is affected by - * NO_HZ. - * - * The basic idea is to fold the nr_active delta into a global idle-delta upon - * entering NO_HZ state such that we can include this as an 'extra' cpu delta - * when we read the global state. - * - * Obviously reality has to ruin such a delightfully simple scheme: - * - * - When we go NO_HZ idle during the window, we can negate our sample - * contribution, causing under-accounting. - * - * We avoid this by keeping two idle-delta counters and flipping them - * when the window starts, thus separating old and new NO_HZ load. - * - * The only trick is the slight shift in index flip for read vs write. - * - * 0s 5s 10s 15s - * +10 +10 +10 +10 - * |-|-----------|-|-----------|-|-----------|-| - * r:0 0 1 1 0 0 1 1 0 - * w:0 1 1 0 0 1 1 0 0 - * - * This ensures we'll fold the old idle contribution in this window while - * accumlating the new one. - * - * - When we wake up from NO_HZ idle during the window, we push up our - * contribution, since we effectively move our sample point to a known - * busy state. - * - * This is solved by pushing the window forward, and thus skipping the - * sample, for this cpu (effectively using the idle-delta for this cpu which - * was in effect at the time the window opened). This also solves the issue - * of having to deal with a cpu having been in NOHZ idle for multiple - * LOAD_FREQ intervals. - * - * When making the ILB scale, we should try to pull this in as well. - */ -static atomic_long_t calc_load_idle[2]; -static int calc_load_idx; - -static inline int calc_load_write_idx(void) -{ - int idx = calc_load_idx; - - /* - * See calc_global_nohz(), if we observe the new index, we also - * need to observe the new update time. - */ - smp_rmb(); - - /* - * If the folding window started, make sure we start writing in the - * next idle-delta. - */ - if (!time_before(jiffies, calc_load_update)) - idx++; - - return idx & 1; -} - -static inline int calc_load_read_idx(void) -{ - return calc_load_idx & 1; -} - -void calc_load_enter_idle(void) -{ - struct rq *this_rq = this_rq(); - long delta; - - /* - * We're going into NOHZ mode, if there's any pending delta, fold it - * into the pending idle delta. - */ - delta = calc_load_fold_active(this_rq); - if (delta) { - int idx = calc_load_write_idx(); - atomic_long_add(delta, &calc_load_idle[idx]); - } -} - -void calc_load_exit_idle(void) -{ - struct rq *this_rq = this_rq(); - - /* - * If we're still before the sample window, we're done. - */ - if (time_before(jiffies, this_rq->calc_load_update)) - return; - - /* - * We woke inside or after the sample window, this means we're already - * accounted through the nohz accounting, so skip the entire deal and - * sync up for the next window. - */ - this_rq->calc_load_update = calc_load_update; - if (time_before(jiffies, this_rq->calc_load_update + 10)) - this_rq->calc_load_update += LOAD_FREQ; -} - -static long calc_load_fold_idle(void) -{ - int idx = calc_load_read_idx(); - long delta = 0; - - if (atomic_long_read(&calc_load_idle[idx])) - delta = atomic_long_xchg(&calc_load_idle[idx], 0); - - return delta; -} - -/** - * fixed_power_int - compute: x^n, in O(log n) time - * - * @x: base of the power - * @frac_bits: fractional bits of @x - * @n: power to raise @x to. - * - * By exploiting the relation between the definition of the natural power - * function: x^n := x*x*...*x (x multiplied by itself for n times), and - * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, - * (where: n_i \elem {0, 1}, the binary vector representing n), - * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is - * of course trivially computable in O(log_2 n), the length of our binary - * vector. - */ -static unsigned long -fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) -{ - unsigned long result = 1UL << frac_bits; - - if (n) for (;;) { - if (n & 1) { - result *= x; - result += 1UL << (frac_bits - 1); - result >>= frac_bits; - } - n >>= 1; - if (!n) - break; - x *= x; - x += 1UL << (frac_bits - 1); - x >>= frac_bits; - } - - return result; -} - -/* - * a1 = a0 * e + a * (1 - e) - * - * a2 = a1 * e + a * (1 - e) - * = (a0 * e + a * (1 - e)) * e + a * (1 - e) - * = a0 * e^2 + a * (1 - e) * (1 + e) - * - * a3 = a2 * e + a * (1 - e) - * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) - * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) - * - * ... - * - * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] - * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) - * = a0 * e^n + a * (1 - e^n) - * - * [1] application of the geometric series: - * - * n 1 - x^(n+1) - * S_n := \Sum x^i = ------------- - * i=0 1 - x - */ -static unsigned long -calc_load_n(unsigned long load, unsigned long exp, - unsigned long active, unsigned int n) -{ - - return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); -} - -/* - * NO_HZ can leave us missing all per-cpu ticks calling - * calc_load_account_active(), but since an idle CPU folds its delta into - * calc_load_tasks_idle per calc_load_account_idle(), all we need to do is fold - * in the pending idle delta if our idle period crossed a load cycle boundary. - * - * Once we've updated the global active value, we need to apply the exponential - * weights adjusted to the number of cycles missed. - */ -static void calc_global_nohz(void) -{ - long delta, active, n; - - if (!time_before(jiffies, calc_load_update + 10)) { - /* - * Catch-up, fold however many we are behind still - */ - delta = jiffies - calc_load_update - 10; - n = 1 + (delta / LOAD_FREQ); - - active = atomic_long_read(&calc_load_tasks); - active = active > 0 ? active * FIXED_1 : 0; - - avenrun[0] = calc_load_n(avenrun[0], EXP_1, active, n); - avenrun[1] = calc_load_n(avenrun[1], EXP_5, active, n); - avenrun[2] = calc_load_n(avenrun[2], EXP_15, active, n); - - calc_load_update += n * LOAD_FREQ; - } - - /* - * Flip the idle index... - * - * Make sure we first write the new time then flip the index, so that - * calc_load_write_idx() will see the new time when it reads the new - * index, this avoids a double flip messing things up. - */ - smp_wmb(); - calc_load_idx++; -} -#else /* !CONFIG_NO_HZ_COMMON */ - -static inline long calc_load_fold_idle(void) { return 0; } -static inline void calc_global_nohz(void) { } - -#endif /* CONFIG_NO_HZ_COMMON */ - -/* - * calc_load - update the avenrun load estimates 10 ticks after the - * CPUs have updated calc_load_tasks. - */ -void calc_global_load(unsigned long ticks) -{ - long active, delta; - - if (time_before(jiffies, calc_load_update + 10)) - return; - - /* - * Fold the 'old' idle-delta to include all NO_HZ cpus. - */ - delta = calc_load_fold_idle(); - if (delta) - atomic_long_add(delta, &calc_load_tasks); - - active = atomic_long_read(&calc_load_tasks); - active = active > 0 ? active * FIXED_1 : 0; - - 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; - - /* - * In case we idled for multiple LOAD_FREQ intervals, catch up in bulk. - */ - calc_global_nohz(); -} - -/* - * Called from update_cpu_load() to periodically update this CPU's - * active count. - */ -static void calc_load_account_active(struct rq *this_rq) -{ - long delta; - - if (time_before(jiffies, this_rq->calc_load_update)) - return; - - delta = calc_load_fold_active(this_rq); - if (delta) - atomic_long_add(delta, &calc_load_tasks); - - this_rq->calc_load_update += LOAD_FREQ; -} - -/* - * End of global load-average stuff - */ - -/* - * The exact cpuload at various idx values, calculated at every tick would be - * load = (2^idx - 1) / 2^idx * load + 1 / 2^idx * cur_load - * - * If a cpu misses updates for n-1 ticks (as it was idle) and update gets called - * on nth tick when cpu may be busy, then we have: - * load = ((2^idx - 1) / 2^idx)^(n-1) * load - * load = (2^idx - 1) / 2^idx) * load + 1 / 2^idx * cur_load - * - * decay_load_missed() below does efficient calculation of - * load = ((2^idx - 1) / 2^idx)^(n-1) * load - * avoiding 0..n-1 loop doing load = ((2^idx - 1) / 2^idx) * load - * - * The calculation is approximated on a 128 point scale. - * degrade_zero_ticks is the number of ticks after which load at any - * particular idx is approximated to be zero. - * degrade_factor is a precomputed table, a row for each load idx. - * Each column corresponds to degradation factor for a power of two ticks, - * based on 128 point scale. - * Example: - * row 2, col 3 (=12) says that the degradation at load idx 2 after - * 8 ticks is 12/128 (which is an approximation of exact factor 3^8/4^8). - * - * With this power of 2 load factors, we can degrade the load n times - * by looking at 1 bits in n and doing as many mult/shift instead of - * n mult/shifts needed by the exact degradation. - */ -#define DEGRADE_SHIFT 7 -static const unsigned char - degrade_zero_ticks[CPU_LOAD_IDX_MAX] = {0, 8, 32, 64, 128}; -static const unsigned char - degrade_factor[CPU_LOAD_IDX_MAX][DEGRADE_SHIFT + 1] = { - {0, 0, 0, 0, 0, 0, 0, 0}, - {64, 32, 8, 0, 0, 0, 0, 0}, - {96, 72, 40, 12, 1, 0, 0}, - {112, 98, 75, 43, 15, 1, 0}, - {120, 112, 98, 76, 45, 16, 2} }; - -/* - * Update cpu_load for any missed ticks, due to tickless idle. The backlog - * would be when CPU is idle and so we just decay the old load without - * adding any new load. - */ -static unsigned long -decay_load_missed(unsigned long load, unsigned long missed_updates, int idx) -{ - int j = 0; - - if (!missed_updates) - return load; - - if (missed_updates >= degrade_zero_ticks[idx]) - return 0; - - if (idx == 1) - return load >> missed_updates; - - while (missed_updates) { - if (missed_updates % 2) - load = (load * degrade_factor[idx][j]) >> DEGRADE_SHIFT; - - missed_updates >>= 1; - j++; - } - return load; -} - -/* - * Update rq->cpu_load[] statistics. This function is usually called every - * scheduler tick (TICK_NSEC). With tickless idle this will not be called - * every tick. We fix it up based on jiffies. - */ -static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, - unsigned long pending_updates) -{ - int i, scale; - - this_rq->nr_load_updates++; - - /* Update our load: */ - this_rq->cpu_load[0] = this_load; /* Fasttrack for idx 0 */ - for (i = 1, scale = 2; i < CPU_LOAD_IDX_MAX; i++, scale += scale) { - unsigned long old_load, new_load; - - /* scale is effectively 1 << i now, and >> i divides by scale */ - - old_load = this_rq->cpu_load[i]; - old_load = decay_load_missed(old_load, pending_updates - 1, i); - new_load = this_load; - /* - * Round up the averaging division if load is increasing. This - * prevents us from getting stuck on 9 if the load is 10, for - * example. - */ - if (new_load > old_load) - new_load += scale - 1; - - this_rq->cpu_load[i] = (old_load * (scale - 1) + new_load) >> i; - } - - sched_avg_update(this_rq); -} - -#ifdef CONFIG_NO_HZ_COMMON -/* - * There is no sane way to deal with nohz on smp when using jiffies because the - * cpu doing the jiffies update might drift wrt the cpu doing the jiffy reading - * causing off-by-one errors in observed deltas; {0,2} instead of {1,1}. - * - * Therefore we cannot use the delta approach from the regular tick since that - * would seriously skew the load calculation. However we'll make do for those - * updates happening while idle (nohz_idle_balance) or coming out of idle - * (tick_nohz_idle_exit). - * - * This means we might still be one tick off for nohz periods. - */ - -/* - * Called from nohz_idle_balance() to update the load ratings before doing the - * idle balance. - */ -void update_idle_cpu_load(struct rq *this_rq) -{ - unsigned long curr_jiffies = ACCESS_ONCE(jiffies); - unsigned long load = this_rq->load.weight; - unsigned long pending_updates; - - /* - * bail if there's load or we're actually up-to-date. - */ - if (load || curr_jiffies == this_rq->last_load_update_tick) - return; - - pending_updates = curr_jiffies - this_rq->last_load_update_tick; - this_rq->last_load_update_tick = curr_jiffies; - - __update_cpu_load(this_rq, load, pending_updates); -} - -/* - * Called from tick_nohz_idle_exit() -- try and fix up the ticks we missed. - */ -void update_cpu_load_nohz(void) -{ - struct rq *this_rq = this_rq(); - unsigned long curr_jiffies = ACCESS_ONCE(jiffies); - unsigned long pending_updates; - - if (curr_jiffies == this_rq->last_load_update_tick) - return; - - raw_spin_lock(&this_rq->lock); - pending_updates = curr_jiffies - this_rq->last_load_update_tick; - if (pending_updates) { - this_rq->last_load_update_tick = curr_jiffies; - /* - * We were idle, this means load 0, the current load might be - * !0 due to remote wakeups and the sort. - */ - __update_cpu_load(this_rq, 0, pending_updates); - } - raw_spin_unlock(&this_rq->lock); -} -#endif /* CONFIG_NO_HZ_COMMON */ - -/* - * Called from scheduler_tick() - */ -static void update_cpu_load_active(struct rq *this_rq) -{ - /* - * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). - */ - this_rq->last_load_update_tick = jiffies; - __update_cpu_load(this_rq, this_rq->load.weight, 1); - - calc_load_account_active(this_rq); -} - #ifdef CONFIG_SMP /* @@ -2673,7 +2112,7 @@ static u64 do_task_delta_exec(struct task_struct *p, struct rq *rq) if (task_current(rq, p)) { update_rq_clock(rq); - ns = rq->clock_task - p->se.exec_start; + ns = rq_clock_task(rq) - p->se.exec_start; if ((s64)ns < 0) ns = 0; } @@ -2726,8 +2165,8 @@ void scheduler_tick(void) raw_spin_lock(&rq->lock); update_rq_clock(rq); - update_cpu_load_active(rq); curr->sched_class->task_tick(rq, curr, 0); + update_cpu_load_active(rq); raw_spin_unlock(&rq->lock); perf_event_task_tick(); @@ -4745,7 +4184,7 @@ void __cpuinit init_idle(struct task_struct *idle, int cpu) */ idle->sched_class = &idle_sched_class; ftrace_graph_init_idle_task(idle, cpu); - vtime_init_idle(idle); + vtime_init_idle(idle, cpu); #if defined(CONFIG_SMP) sprintf(idle->comm, "%s/%d", INIT_TASK_COMM, cpu); #endif @@ -4947,6 +4386,13 @@ static void migrate_tasks(unsigned int dead_cpu) */ rq->stop = NULL; + /* + * put_prev_task() and pick_next_task() sched + * class method both need to have an up-to-date + * value of rq->clock[_task] + */ + update_rq_clock(rq); + for ( ; ; ) { /* * There's this thread running, bail when that's the only @@ -5080,7 +4526,7 @@ sd_alloc_ctl_domain_table(struct sched_domain *sd) return table; } -static ctl_table *sd_alloc_ctl_cpu_table(int cpu) +static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu) { struct ctl_table *entry, *table; struct sched_domain *sd; @@ -5894,7 +5340,7 @@ build_sched_groups(struct sched_domain *sd, int cpu) get_group(cpu, sdd, &sd->groups); atomic_inc(&sd->groups->ref); - if (cpu != cpumask_first(sched_domain_span(sd))) + if (cpu != cpumask_first(span)) return 0; lockdep_assert_held(&sched_domains_mutex); @@ -5904,12 +5350,12 @@ build_sched_groups(struct sched_domain *sd, int cpu) for_each_cpu(i, span) { struct sched_group *sg; - int group = get_group(i, sdd, &sg); - int j; + int group, j; if (cpumask_test_cpu(i, covered)) continue; + group = get_group(i, sdd, &sg); cpumask_clear(sched_group_cpus(sg)); sg->sgp->power = 0; cpumask_setall(sched_group_mask(sg)); @@ -5947,7 +5393,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) { struct sched_group *sg = sd->groups; - WARN_ON(!sd || !sg); + WARN_ON(!sg); do { sg->group_weight = cpumask_weight(sched_group_cpus(sg)); @@ -6112,6 +5558,9 @@ static struct sched_domain_topology_level default_topology[] = { static struct sched_domain_topology_level *sched_domain_topology = default_topology; +#define for_each_sd_topology(tl) \ + for (tl = sched_domain_topology; tl->init; tl++) + #ifdef CONFIG_NUMA static int sched_domains_numa_levels; @@ -6409,7 +5858,7 @@ static int __sdt_alloc(const struct cpumask *cpu_map) struct sched_domain_topology_level *tl; int j; - for (tl = sched_domain_topology; tl->init; tl++) { + for_each_sd_topology(tl) { struct sd_data *sdd = &tl->data; sdd->sd = alloc_percpu(struct sched_domain *); @@ -6462,7 +5911,7 @@ static void __sdt_free(const struct cpumask *cpu_map) struct sched_domain_topology_level *tl; int j; - for (tl = sched_domain_topology; tl->init; tl++) { + for_each_sd_topology(tl) { struct sd_data *sdd = &tl->data; for_each_cpu(j, cpu_map) { @@ -6490,9 +5939,8 @@ static void __sdt_free(const struct cpumask *cpu_map) } struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, - struct s_data *d, const struct cpumask *cpu_map, - struct sched_domain_attr *attr, struct sched_domain *child, - int cpu) + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *child, int cpu) { struct sched_domain *sd = tl->init(tl, cpu); if (!sd) @@ -6503,8 +5951,8 @@ struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, sd->level = child->level + 1; sched_domain_level_max = max(sched_domain_level_max, sd->level); child->parent = sd; + sd->child = child; } - sd->child = child; set_domain_attribute(sd, attr); return sd; @@ -6517,7 +5965,7 @@ struct sched_domain *build_sched_domain(struct sched_domain_topology_level *tl, static int build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_attr *attr) { - enum s_alloc alloc_state = sa_none; + enum s_alloc alloc_state; struct sched_domain *sd; struct s_data d; int i, ret = -ENOMEM; @@ -6531,18 +5979,15 @@ static int build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_topology_level *tl; sd = NULL; - for (tl = sched_domain_topology; tl->init; tl++) { - sd = build_sched_domain(tl, &d, cpu_map, attr, sd, i); + for_each_sd_topology(tl) { + sd = build_sched_domain(tl, cpu_map, attr, sd, i); + if (tl == sched_domain_topology) + *per_cpu_ptr(d.sd, i) = sd; if (tl->flags & SDTL_OVERLAP || sched_feat(FORCE_SD_OVERLAP)) sd->flags |= SD_OVERLAP; if (cpumask_equal(cpu_map, sched_domain_span(sd))) break; } - - while (sd->child) - sd = sd->child; - - *per_cpu_ptr(d.sd, i) = sd; } /* Build the groups for the domains */ @@ -6854,9 +6299,6 @@ void __init sched_init_smp(void) hotcpu_notifier(cpuset_cpu_active, CPU_PRI_CPUSET_ACTIVE); hotcpu_notifier(cpuset_cpu_inactive, CPU_PRI_CPUSET_INACTIVE); - /* RT runtime code needs to handle some hotplug events */ - hotcpu_notifier(update_runtime, 0); - init_hrtick(); /* Move init over to a non-isolated CPU */ diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index cc2dc3eea8a3..a7959e05a9d5 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -515,9 +515,8 @@ static cputime_t scale_stime(u64 stime, u64 rtime, u64 total) for (;;) { /* Make sure "rtime" is the bigger of stime/rtime */ - if (stime > rtime) { - u64 tmp = rtime; rtime = stime; stime = tmp; - } + if (stime > rtime) + swap(rtime, stime); /* Make sure 'total' fits in 32 bits */ if (total >> 32) @@ -747,17 +746,17 @@ void arch_vtime_task_switch(struct task_struct *prev) write_seqlock(¤t->vtime_seqlock); current->vtime_snap_whence = VTIME_SYS; - current->vtime_snap = sched_clock(); + current->vtime_snap = sched_clock_cpu(smp_processor_id()); write_sequnlock(¤t->vtime_seqlock); } -void vtime_init_idle(struct task_struct *t) +void vtime_init_idle(struct task_struct *t, int cpu) { unsigned long flags; write_seqlock_irqsave(&t->vtime_seqlock, flags); t->vtime_snap_whence = VTIME_SYS; - t->vtime_snap = sched_clock(); + t->vtime_snap = sched_clock_cpu(cpu); write_sequnlock_irqrestore(&t->vtime_seqlock, flags); } diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 75024a673520..e076bddd4c66 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -209,22 +209,24 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) cfs_rq->nr_spread_over); SEQ_printf(m, " .%-30s: %d\n", "nr_running", cfs_rq->nr_running); SEQ_printf(m, " .%-30s: %ld\n", "load", cfs_rq->load.weight); -#ifdef CONFIG_FAIR_GROUP_SCHED #ifdef CONFIG_SMP - SEQ_printf(m, " .%-30s: %lld\n", "runnable_load_avg", + SEQ_printf(m, " .%-30s: %ld\n", "runnable_load_avg", cfs_rq->runnable_load_avg); - SEQ_printf(m, " .%-30s: %lld\n", "blocked_load_avg", + SEQ_printf(m, " .%-30s: %ld\n", "blocked_load_avg", cfs_rq->blocked_load_avg); - SEQ_printf(m, " .%-30s: %lld\n", "tg_load_avg", - (unsigned long long)atomic64_read(&cfs_rq->tg->load_avg)); - SEQ_printf(m, " .%-30s: %lld\n", "tg_load_contrib", +#ifdef CONFIG_FAIR_GROUP_SCHED + SEQ_printf(m, " .%-30s: %ld\n", "tg_load_contrib", cfs_rq->tg_load_contrib); SEQ_printf(m, " .%-30s: %d\n", "tg_runnable_contrib", cfs_rq->tg_runnable_contrib); + SEQ_printf(m, " .%-30s: %ld\n", "tg_load_avg", + atomic_long_read(&cfs_rq->tg->load_avg)); SEQ_printf(m, " .%-30s: %d\n", "tg->runnable_avg", atomic_read(&cfs_rq->tg->runnable_avg)); #endif +#endif +#ifdef CONFIG_FAIR_GROUP_SCHED print_cfs_group_stats(m, cpu, cfs_rq->tg); #endif } @@ -493,15 +495,16 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) SEQ_printf(m, "%s (%d, #threads: %d)\n", p->comm, p->pid, get_nr_threads(p)); SEQ_printf(m, - "---------------------------------------------------------\n"); + "---------------------------------------------------------" + "----------\n"); #define __P(F) \ - SEQ_printf(m, "%-35s:%21Ld\n", #F, (long long)F) + SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)F) #define P(F) \ - SEQ_printf(m, "%-35s:%21Ld\n", #F, (long long)p->F) + SEQ_printf(m, "%-45s:%21Ld\n", #F, (long long)p->F) #define __PN(F) \ - SEQ_printf(m, "%-35s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F)) + SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)F)) #define PN(F) \ - SEQ_printf(m, "%-35s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F)) + SEQ_printf(m, "%-45s:%14Ld.%06ld\n", #F, SPLIT_NS((long long)p->F)) PN(se.exec_start); PN(se.vruntime); @@ -560,12 +563,18 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) } #endif __P(nr_switches); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "nr_voluntary_switches", (long long)p->nvcsw); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "nr_involuntary_switches", (long long)p->nivcsw); P(se.load.weight); +#ifdef CONFIG_SMP + P(se.avg.runnable_avg_sum); + P(se.avg.runnable_avg_period); + P(se.avg.load_avg_contrib); + P(se.avg.decay_count); +#endif P(policy); P(prio); #undef PN @@ -579,7 +588,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) t0 = cpu_clock(this_cpu); t1 = cpu_clock(this_cpu); - SEQ_printf(m, "%-35s:%21Ld\n", + SEQ_printf(m, "%-45s:%21Ld\n", "clock-delta", (long long)(t1-t0)); } } diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index c61a614465c8..f77f9c527449 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -113,6 +113,24 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL; unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL; #endif +static inline void update_load_add(struct load_weight *lw, unsigned long inc) +{ + lw->weight += inc; + lw->inv_weight = 0; +} + +static inline void update_load_sub(struct load_weight *lw, unsigned long dec) +{ + lw->weight -= dec; + lw->inv_weight = 0; +} + +static inline void update_load_set(struct load_weight *lw, unsigned long w) +{ + lw->weight = w; + lw->inv_weight = 0; +} + /* * Increase the granularity value when there are more CPUs, * because with more CPUs the 'effective latency' as visible @@ -662,6 +680,26 @@ static u64 sched_vslice(struct cfs_rq *cfs_rq, struct sched_entity *se) return calc_delta_fair(sched_slice(cfs_rq, se), se); } +#ifdef CONFIG_SMP +static inline void __update_task_entity_contrib(struct sched_entity *se); + +/* Give new task start runnable values to heavy its load in infant time */ +void init_task_runnable_average(struct task_struct *p) +{ + u32 slice; + + p->se.avg.decay_count = 0; + slice = sched_slice(task_cfs_rq(p), &p->se) >> 10; + p->se.avg.runnable_avg_sum = slice; + p->se.avg.runnable_avg_period = slice; + __update_task_entity_contrib(&p->se); +} +#else +void init_task_runnable_average(struct task_struct *p) +{ +} +#endif + /* * Update the current task's runtime statistics. Skip current tasks that * are not in our scheduling class. @@ -686,7 +724,7 @@ __update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, static void update_curr(struct cfs_rq *cfs_rq) { struct sched_entity *curr = cfs_rq->curr; - u64 now = rq_of(cfs_rq)->clock_task; + u64 now = rq_clock_task(rq_of(cfs_rq)); unsigned long delta_exec; if (unlikely(!curr)) @@ -718,7 +756,7 @@ static void update_curr(struct cfs_rq *cfs_rq) static inline void update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se) { - schedstat_set(se->statistics.wait_start, rq_of(cfs_rq)->clock); + schedstat_set(se->statistics.wait_start, rq_clock(rq_of(cfs_rq))); } /* @@ -738,14 +776,14 @@ static void update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se) { schedstat_set(se->statistics.wait_max, max(se->statistics.wait_max, - rq_of(cfs_rq)->clock - se->statistics.wait_start)); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start)); schedstat_set(se->statistics.wait_count, se->statistics.wait_count + 1); schedstat_set(se->statistics.wait_sum, se->statistics.wait_sum + - rq_of(cfs_rq)->clock - se->statistics.wait_start); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start); #ifdef CONFIG_SCHEDSTATS if (entity_is_task(se)) { trace_sched_stat_wait(task_of(se), - rq_of(cfs_rq)->clock - se->statistics.wait_start); + rq_clock(rq_of(cfs_rq)) - se->statistics.wait_start); } #endif schedstat_set(se->statistics.wait_start, 0); @@ -771,7 +809,7 @@ update_stats_curr_start(struct cfs_rq *cfs_rq, struct sched_entity *se) /* * We are starting a new run period: */ - se->exec_start = rq_of(cfs_rq)->clock_task; + se->exec_start = rq_clock_task(rq_of(cfs_rq)); } /************************************************** @@ -1037,7 +1075,7 @@ static inline long calc_tg_weight(struct task_group *tg, struct cfs_rq *cfs_rq) * to gain a more accurate current total weight. See * update_cfs_rq_load_contribution(). */ - tg_weight = atomic64_read(&tg->load_avg); + tg_weight = atomic_long_read(&tg->load_avg); tg_weight -= cfs_rq->tg_load_contrib; tg_weight += cfs_rq->load.weight; @@ -1110,8 +1148,7 @@ static inline void update_cfs_shares(struct cfs_rq *cfs_rq) } #endif /* CONFIG_FAIR_GROUP_SCHED */ -/* Only depends on SMP, FAIR_GROUP_SCHED may be removed when useful in lb */ -#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED) +#ifdef CONFIG_SMP /* * We choose a half-life close to 1 scheduling period. * Note: The tables below are dependent on this value. @@ -1319,13 +1356,13 @@ static inline void __update_cfs_rq_tg_load_contrib(struct cfs_rq *cfs_rq, int force_update) { struct task_group *tg = cfs_rq->tg; - s64 tg_contrib; + long tg_contrib; tg_contrib = cfs_rq->runnable_load_avg + cfs_rq->blocked_load_avg; tg_contrib -= cfs_rq->tg_load_contrib; - if (force_update || abs64(tg_contrib) > cfs_rq->tg_load_contrib / 8) { - atomic64_add(tg_contrib, &tg->load_avg); + if (force_update || abs(tg_contrib) > cfs_rq->tg_load_contrib / 8) { + atomic_long_add(tg_contrib, &tg->load_avg); cfs_rq->tg_load_contrib += tg_contrib; } } @@ -1360,8 +1397,8 @@ static inline void __update_group_entity_contrib(struct sched_entity *se) u64 contrib; contrib = cfs_rq->tg_load_contrib * tg->shares; - se->avg.load_avg_contrib = div64_u64(contrib, - atomic64_read(&tg->load_avg) + 1); + se->avg.load_avg_contrib = div_u64(contrib, + atomic_long_read(&tg->load_avg) + 1); /* * For group entities we need to compute a correction term in the case @@ -1480,8 +1517,9 @@ static void update_cfs_rq_blocked_load(struct cfs_rq *cfs_rq, int force_update) if (!decays && !force_update) return; - if (atomic64_read(&cfs_rq->removed_load)) { - u64 removed_load = atomic64_xchg(&cfs_rq->removed_load, 0); + if (atomic_long_read(&cfs_rq->removed_load)) { + unsigned long removed_load; + removed_load = atomic_long_xchg(&cfs_rq->removed_load, 0); subtract_blocked_load_contrib(cfs_rq, removed_load); } @@ -1497,7 +1535,7 @@ static void update_cfs_rq_blocked_load(struct cfs_rq *cfs_rq, int force_update) static inline void update_rq_runnable_avg(struct rq *rq, int runnable) { - __update_entity_runnable_avg(rq->clock_task, &rq->avg, runnable); + __update_entity_runnable_avg(rq_clock_task(rq), &rq->avg, runnable); __update_tg_runnable_avg(&rq->avg, &rq->cfs); } @@ -1510,9 +1548,13 @@ static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq, * We track migrations using entity decay_count <= 0, on a wake-up * migration we use a negative decay count to track the remote decays * accumulated while sleeping. + * + * Newly forked tasks are enqueued with se->avg.decay_count == 0, they + * are seen by enqueue_entity_load_avg() as a migration with an already + * constructed load_avg_contrib. */ if (unlikely(se->avg.decay_count <= 0)) { - se->avg.last_runnable_update = rq_of(cfs_rq)->clock_task; + se->avg.last_runnable_update = rq_clock_task(rq_of(cfs_rq)); if (se->avg.decay_count) { /* * In a wake-up migration we have to approximate the @@ -1530,7 +1572,13 @@ static inline void enqueue_entity_load_avg(struct cfs_rq *cfs_rq, } wakeup = 0; } else { - __synchronize_entity_decay(se); + /* + * Task re-woke on same cpu (or else migrate_task_rq_fair() + * would have made count negative); we must be careful to avoid + * double-accounting blocked time after synchronizing decays. + */ + se->avg.last_runnable_update += __synchronize_entity_decay(se) + << 20; } /* migrated tasks did not contribute to our blocked load */ @@ -1607,7 +1655,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) tsk = task_of(se); if (se->statistics.sleep_start) { - u64 delta = rq_of(cfs_rq)->clock - se->statistics.sleep_start; + u64 delta = rq_clock(rq_of(cfs_rq)) - se->statistics.sleep_start; if ((s64)delta < 0) delta = 0; @@ -1624,7 +1672,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) } } if (se->statistics.block_start) { - u64 delta = rq_of(cfs_rq)->clock - se->statistics.block_start; + u64 delta = rq_clock(rq_of(cfs_rq)) - se->statistics.block_start; if ((s64)delta < 0) delta = 0; @@ -1712,7 +1760,7 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) { /* * Update the normalized vruntime before updating min_vruntime - * through callig update_curr(). + * through calling update_curr(). */ if (!(flags & ENQUEUE_WAKEUP) || (flags & ENQUEUE_WAKING)) se->vruntime += cfs_rq->min_vruntime; @@ -1805,9 +1853,9 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) struct task_struct *tsk = task_of(se); if (tsk->state & TASK_INTERRUPTIBLE) - se->statistics.sleep_start = rq_of(cfs_rq)->clock; + se->statistics.sleep_start = rq_clock(rq_of(cfs_rq)); if (tsk->state & TASK_UNINTERRUPTIBLE) - se->statistics.block_start = rq_of(cfs_rq)->clock; + se->statistics.block_start = rq_clock(rq_of(cfs_rq)); } #endif } @@ -2082,7 +2130,7 @@ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) if (unlikely(cfs_rq->throttle_count)) return cfs_rq->throttled_clock_task; - return rq_of(cfs_rq)->clock_task - cfs_rq->throttled_clock_task_time; + return rq_clock_task(rq_of(cfs_rq)) - cfs_rq->throttled_clock_task_time; } /* returns 0 on failure to allocate runtime */ @@ -2138,10 +2186,9 @@ static int assign_cfs_rq_runtime(struct cfs_rq *cfs_rq) static void expire_cfs_rq_runtime(struct cfs_rq *cfs_rq) { struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg); - struct rq *rq = rq_of(cfs_rq); /* if the deadline is ahead of our clock, nothing to do */ - if (likely((s64)(rq->clock - cfs_rq->runtime_expires) < 0)) + if (likely((s64)(rq_clock(rq_of(cfs_rq)) - cfs_rq->runtime_expires) < 0)) return; if (cfs_rq->runtime_remaining < 0) @@ -2230,7 +2277,7 @@ static int tg_unthrottle_up(struct task_group *tg, void *data) #ifdef CONFIG_SMP if (!cfs_rq->throttle_count) { /* adjust cfs_rq_clock_task() */ - cfs_rq->throttled_clock_task_time += rq->clock_task - + cfs_rq->throttled_clock_task_time += rq_clock_task(rq) - cfs_rq->throttled_clock_task; } #endif @@ -2245,7 +2292,7 @@ static int tg_throttle_down(struct task_group *tg, void *data) /* group is entering throttled state, stop time */ if (!cfs_rq->throttle_count) - cfs_rq->throttled_clock_task = rq->clock_task; + cfs_rq->throttled_clock_task = rq_clock_task(rq); cfs_rq->throttle_count++; return 0; @@ -2284,7 +2331,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq) rq->nr_running -= task_delta; cfs_rq->throttled = 1; - cfs_rq->throttled_clock = rq->clock; + cfs_rq->throttled_clock = rq_clock(rq); raw_spin_lock(&cfs_b->lock); list_add_tail_rcu(&cfs_rq->throttled_list, &cfs_b->throttled_cfs_rq); raw_spin_unlock(&cfs_b->lock); @@ -2298,15 +2345,17 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq) int enqueue = 1; long task_delta; - se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))]; + se = cfs_rq->tg->se[cpu_of(rq)]; cfs_rq->throttled = 0; + + update_rq_clock(rq); + raw_spin_lock(&cfs_b->lock); - cfs_b->throttled_time += rq->clock - cfs_rq->throttled_clock; + cfs_b->throttled_time += rq_clock(rq) - cfs_rq->throttled_clock; list_del_rcu(&cfs_rq->throttled_list); raw_spin_unlock(&cfs_b->lock); - update_rq_clock(rq); /* update hierarchical throttle state */ walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq); @@ -2599,10 +2648,6 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq) throttle_cfs_rq(cfs_rq); } -static inline u64 default_cfs_period(void); -static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun); -static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b); - static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer) { struct cfs_bandwidth *cfs_b = @@ -2706,7 +2751,7 @@ static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq) #else /* CONFIG_CFS_BANDWIDTH */ static inline u64 cfs_rq_clock_task(struct cfs_rq *cfs_rq) { - return rq_of(cfs_rq)->clock_task; + return rq_clock_task(rq_of(cfs_rq)); } static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq, @@ -2919,7 +2964,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) /* Used instead of source_load when we know the type == 0 */ static unsigned long weighted_cpuload(const int cpu) { - return cpu_rq(cpu)->load.weight; + return cpu_rq(cpu)->cfs.runnable_load_avg; } /* @@ -2964,9 +3009,10 @@ static unsigned long cpu_avg_load_per_task(int cpu) { struct rq *rq = cpu_rq(cpu); unsigned long nr_running = ACCESS_ONCE(rq->nr_running); + unsigned long load_avg = rq->cfs.runnable_load_avg; if (nr_running) - return rq->load.weight / nr_running; + return load_avg / nr_running; return 0; } @@ -3416,12 +3462,6 @@ unlock: } /* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#ifdef CONFIG_FAIR_GROUP_SCHED -/* * Called immediately before a task is migrated to a new cpu; task_cpu(p) and * cfs_rq_of(p) references at time of call are still valid and identify the * previous cpu. However, the caller only guarantees p->pi_lock is held; no @@ -3441,10 +3481,10 @@ migrate_task_rq_fair(struct task_struct *p, int next_cpu) */ if (se->avg.decay_count) { se->avg.decay_count = -__synchronize_entity_decay(se); - atomic64_add(se->avg.load_avg_contrib, &cfs_rq->removed_load); + atomic_long_add(se->avg.load_avg_contrib, + &cfs_rq->removed_load); } } -#endif #endif /* CONFIG_SMP */ static unsigned long @@ -3946,7 +3986,7 @@ int can_migrate_task(struct task_struct *p, struct lb_env *env) * 2) too many balance attempts have failed. */ - tsk_cache_hot = task_hot(p, env->src_rq->clock_task, env->sd); + tsk_cache_hot = task_hot(p, rq_clock_task(env->src_rq), env->sd); if (!tsk_cache_hot || env->sd->nr_balance_failed > env->sd->cache_nice_tries) { @@ -4141,11 +4181,11 @@ static int tg_load_down(struct task_group *tg, void *data) long cpu = (long)data; if (!tg->parent) { - load = cpu_rq(cpu)->load.weight; + load = cpu_rq(cpu)->avg.load_avg_contrib; } else { load = tg->parent->cfs_rq[cpu]->h_load; - load *= tg->se[cpu]->load.weight; - load /= tg->parent->cfs_rq[cpu]->load.weight + 1; + load = div64_ul(load * tg->se[cpu]->avg.load_avg_contrib, + tg->parent->cfs_rq[cpu]->runnable_load_avg + 1); } tg->cfs_rq[cpu]->h_load = load; @@ -4171,12 +4211,9 @@ static void update_h_load(long cpu) static unsigned long task_h_load(struct task_struct *p) { struct cfs_rq *cfs_rq = task_cfs_rq(p); - unsigned long load; - - load = p->se.load.weight; - load = div_u64(load * cfs_rq->h_load, cfs_rq->load.weight + 1); - return load; + return div64_ul(p->se.avg.load_avg_contrib * cfs_rq->h_load, + cfs_rq->runnable_load_avg + 1); } #else static inline void update_blocked_averages(int cpu) @@ -4189,7 +4226,7 @@ static inline void update_h_load(long cpu) static unsigned long task_h_load(struct task_struct *p) { - return p->se.load.weight; + return p->se.avg.load_avg_contrib; } #endif @@ -4302,7 +4339,7 @@ static unsigned long scale_rt_power(int cpu) age_stamp = ACCESS_ONCE(rq->age_stamp); avg = ACCESS_ONCE(rq->rt_avg); - total = sched_avg_period() + (rq->clock - age_stamp); + total = sched_avg_period() + (rq_clock(rq) - age_stamp); if (unlikely(total < avg)) { /* Ensures that power won't end up being negative */ @@ -5241,7 +5278,7 @@ void idle_balance(int this_cpu, struct rq *this_rq) int pulled_task = 0; unsigned long next_balance = jiffies + HZ; - this_rq->idle_stamp = this_rq->clock; + this_rq->idle_stamp = rq_clock(this_rq); if (this_rq->avg_idle < sysctl_sched_migration_cost) return; @@ -5418,10 +5455,9 @@ static inline void nohz_balance_exit_idle(int cpu) static inline void set_cpu_sd_state_busy(void) { struct sched_domain *sd; - int cpu = smp_processor_id(); rcu_read_lock(); - sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); + sd = rcu_dereference_check_sched_domain(this_rq()->sd); if (!sd || !sd->nohz_idle) goto unlock; @@ -5436,10 +5472,9 @@ unlock: void set_cpu_sd_state_idle(void) { struct sched_domain *sd; - int cpu = smp_processor_id(); rcu_read_lock(); - sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); + sd = rcu_dereference_check_sched_domain(this_rq()->sd); if (!sd || sd->nohz_idle) goto unlock; @@ -5848,7 +5883,7 @@ static void switched_from_fair(struct rq *rq, struct task_struct *p) se->vruntime -= cfs_rq->min_vruntime; } -#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP) +#ifdef CONFIG_SMP /* * Remove our load from contribution when we leave sched_fair * and ensure we don't carry in an old decay_count if we @@ -5907,9 +5942,9 @@ void init_cfs_rq(struct cfs_rq *cfs_rq) #ifndef CONFIG_64BIT cfs_rq->min_vruntime_copy = cfs_rq->min_vruntime; #endif -#if defined(CONFIG_FAIR_GROUP_SCHED) && defined(CONFIG_SMP) +#ifdef CONFIG_SMP atomic64_set(&cfs_rq->decay_counter, 1); - atomic64_set(&cfs_rq->removed_load, 0); + atomic_long_set(&cfs_rq->removed_load, 0); #endif } @@ -6091,6 +6126,9 @@ int sched_group_set_shares(struct task_group *tg, unsigned long shares) se = tg->se[i]; /* Propagate contribution to hierarchy */ raw_spin_lock_irqsave(&rq->lock, flags); + + /* Possible calls to update_curr() need rq clock */ + update_rq_clock(rq); for_each_sched_entity(se) update_cfs_shares(group_cfs_rq(se)); raw_spin_unlock_irqrestore(&rq->lock, flags); @@ -6146,9 +6184,8 @@ const struct sched_class fair_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_fair, -#ifdef CONFIG_FAIR_GROUP_SCHED .migrate_task_rq = migrate_task_rq_fair, -#endif + .rq_online = rq_online_fair, .rq_offline = rq_offline_fair, diff --git a/kernel/sched/proc.c b/kernel/sched/proc.c new file mode 100644 index 000000000000..16f5a30f9c88 --- /dev/null +++ b/kernel/sched/proc.c @@ -0,0 +1,591 @@ +/* + * kernel/sched/proc.c + * + * Kernel load calculations, forked from sched/core.c + */ + +#include <linux/export.h> + +#include "sched.h" + +unsigned long this_cpu_load(void) +{ + struct rq *this = this_rq(); + return this->cpu_load[0]; +} + + +/* + * Global load-average calculations + * + * We take a distributed and async approach to calculating the global load-avg + * in order to minimize overhead. + * + * The global load average is an exponentially decaying average of nr_running + + * nr_uninterruptible. + * + * Once every LOAD_FREQ: + * + * nr_active = 0; + * for_each_possible_cpu(cpu) + * nr_active += cpu_of(cpu)->nr_running + cpu_of(cpu)->nr_uninterruptible; + * + * avenrun[n] = avenrun[0] * exp_n + nr_active * (1 - exp_n) + * + * Due to a number of reasons the above turns in the mess below: + * + * - for_each_possible_cpu() is prohibitively expensive on machines with + * serious number of cpus, therefore we need to take a distributed approach + * to calculating nr_active. + * + * \Sum_i x_i(t) = \Sum_i x_i(t) - x_i(t_0) | x_i(t_0) := 0 + * = \Sum_i { \Sum_j=1 x_i(t_j) - x_i(t_j-1) } + * + * So assuming nr_active := 0 when we start out -- true per definition, we + * can simply take per-cpu deltas and fold those into a global accumulate + * to obtain the same result. See calc_load_fold_active(). + * + * Furthermore, in order to avoid synchronizing all per-cpu delta folding + * across the machine, we assume 10 ticks is sufficient time for every + * cpu to have completed this task. + * + * This places an upper-bound on the IRQ-off latency of the machine. Then + * again, being late doesn't loose the delta, just wrecks the sample. + * + * - cpu_rq()->nr_uninterruptible isn't accurately tracked per-cpu because + * this would add another cross-cpu cacheline miss and atomic operation + * to the wakeup path. Instead we increment on whatever cpu the task ran + * when it went into uninterruptible state and decrement on whatever cpu + * did the wakeup. This means that only the sum of nr_uninterruptible over + * all cpus yields the correct result. + * + * This covers the NO_HZ=n code, for extra head-aches, see the comment below. + */ + +/* Variables and functions for calc_load */ +atomic_long_t calc_load_tasks; +unsigned long calc_load_update; +unsigned long avenrun[3]; +EXPORT_SYMBOL(avenrun); /* should be removed */ + +/** + * 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; +} + +long calc_load_fold_active(struct rq *this_rq) +{ + long nr_active, delta = 0; + + 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; + } + + return delta; +} + +/* + * a1 = a0 * e + a * (1 - e) + */ +static unsigned long +calc_load(unsigned long load, unsigned long exp, unsigned long active) +{ + load *= exp; + load += active * (FIXED_1 - exp); + load += 1UL << (FSHIFT - 1); + return load >> FSHIFT; +} + +#ifdef CONFIG_NO_HZ_COMMON +/* + * Handle NO_HZ for the global load-average. + * + * Since the above described distributed algorithm to compute the global + * load-average relies on per-cpu sampling from the tick, it is affected by + * NO_HZ. + * + * The basic idea is to fold the nr_active delta into a global idle-delta upon + * entering NO_HZ state such that we can include this as an 'extra' cpu delta + * when we read the global state. + * + * Obviously reality has to ruin such a delightfully simple scheme: + * + * - When we go NO_HZ idle during the window, we can negate our sample + * contribution, causing under-accounting. + * + * We avoid this by keeping two idle-delta counters and flipping them + * when the window starts, thus separating old and new NO_HZ load. + * + * The only trick is the slight shift in index flip for read vs write. + * + * 0s 5s 10s 15s + * +10 +10 +10 +10 + * |-|-----------|-|-----------|-|-----------|-| + * r:0 0 1 1 0 0 1 1 0 + * w:0 1 1 0 0 1 1 0 0 + * + * This ensures we'll fold the old idle contribution in this window while + * accumlating the new one. + * + * - When we wake up from NO_HZ idle during the window, we push up our + * contribution, since we effectively move our sample point to a known + * busy state. + * + * This is solved by pushing the window forward, and thus skipping the + * sample, for this cpu (effectively using the idle-delta for this cpu which + * was in effect at the time the window opened). This also solves the issue + * of having to deal with a cpu having been in NOHZ idle for multiple + * LOAD_FREQ intervals. + * + * When making the ILB scale, we should try to pull this in as well. + */ +static atomic_long_t calc_load_idle[2]; +static int calc_load_idx; + +static inline int calc_load_write_idx(void) +{ + int idx = calc_load_idx; + + /* + * See calc_global_nohz(), if we observe the new index, we also + * need to observe the new update time. + */ + smp_rmb(); + + /* + * If the folding window started, make sure we start writing in the + * next idle-delta. + */ + if (!time_before(jiffies, calc_load_update)) + idx++; + + return idx & 1; +} + +static inline int calc_load_read_idx(void) +{ + return calc_load_idx & 1; +} + +void calc_load_enter_idle(void) +{ + struct rq *this_rq = this_rq(); + long delta; + + /* + * We're going into NOHZ mode, if there's any pending delta, fold it + * into the pending idle delta. + */ + delta = calc_load_fold_active(this_rq); + if (delta) { + int idx = calc_load_write_idx(); + atomic_long_add(delta, &calc_load_idle[idx]); + } +} + +void calc_load_exit_idle(void) +{ + struct rq *this_rq = this_rq(); + + /* + * If we're still before the sample window, we're done. + */ + if (time_before(jiffies, this_rq->calc_load_update)) + return; + + /* + * We woke inside or after the sample window, this means we're already + * accounted through the nohz accounting, so skip the entire deal and + * sync up for the next window. + */ + this_rq->calc_load_update = calc_load_update; + if (time_before(jiffies, this_rq->calc_load_update + 10)) + this_rq->calc_load_update += LOAD_FREQ; +} + +static long calc_load_fold_idle(void) +{ + int idx = calc_load_read_idx(); + long delta = 0; + + if (atomic_long_read(&calc_load_idle[idx])) + delta = atomic_long_xchg(&calc_load_idle[idx], 0); + + return delta; +} + +/** + * fixed_power_int - compute: x^n, in O(log n) time + * + * @x: base of the power + * @frac_bits: fractional bits of @x + * @n: power to raise @x to. + * + * By exploiting the relation between the definition of the natural power + * function: x^n := x*x*...*x (x multiplied by itself for n times), and + * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, + * (where: n_i \elem {0, 1}, the binary vector representing n), + * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is + * of course trivially computable in O(log_2 n), the length of our binary + * vector. + */ +static unsigned long +fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) +{ + unsigned long result = 1UL << frac_bits; + + if (n) for (;;) { + if (n & 1) { + result *= x; + result += 1UL << (frac_bits - 1); + result >>= frac_bits; + } + n >>= 1; + if (!n) + break; + x *= x; + x += 1UL << (frac_bits - 1); + x >>= frac_bits; + } + + return result; +} + +/* + * a1 = a0 * e + a * (1 - e) + * + * a2 = a1 * e + a * (1 - e) + * = (a0 * e + a * (1 - e)) * e + a * (1 - e) + * = a0 * e^2 + a * (1 - e) * (1 + e) + * + * a3 = a2 * e + a * (1 - e) + * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) + * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) + * + * ... + * + * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] + * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) + * = a0 * e^n + a * (1 - e^n) + * + * [1] application of the geometric series: + * + * n 1 - x^(n+1) + * S_n := \Sum x^i = ------------- + * i=0 1 - x + */ +static unsigned long +calc_load_n(unsigned long load, unsigned long exp, + unsigned long active, unsigned int n) +{ + + return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); +} + +/* + * NO_HZ can leave us missing all per-cpu ticks calling + * calc_load_account_active(), but since an idle CPU folds its delta into + * calc_load_tasks_idle per calc_load_account_idle(), all we need to do is fold + * in the pending idle delta if our idle period crossed a load cycle boundary. + * + * Once we've updated the global active value, we need to apply the exponential + * weights adjusted to the number of cycles missed. + */ +static void calc_global_nohz(void) +{ + long delta, active, n; + + if (!time_before(jiffies, calc_load_update + 10)) { + /* + * Catch-up, fold however many we are behind still + */ + delta = jiffies - calc_load_update - 10; + n = 1 + (delta / LOAD_FREQ); + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + avenrun[0] = calc_load_n(avenrun[0], EXP_1, active, n); + avenrun[1] = calc_load_n(avenrun[1], EXP_5, active, n); + avenrun[2] = calc_load_n(avenrun[2], EXP_15, active, n); + + calc_load_update += n * LOAD_FREQ; + } + + /* + * Flip the idle index... + * + * Make sure we first write the new time then flip the index, so that + * calc_load_write_idx() will see the new time when it reads the new + * index, this avoids a double flip messing things up. + */ + smp_wmb(); + calc_load_idx++; +} +#else /* !CONFIG_NO_HZ_COMMON */ + +static inline long calc_load_fold_idle(void) { return 0; } +static inline void calc_global_nohz(void) { } + +#endif /* CONFIG_NO_HZ_COMMON */ + +/* + * calc_load - update the avenrun load estimates 10 ticks after the + * CPUs have updated calc_load_tasks. + */ +void calc_global_load(unsigned long ticks) +{ + long active, delta; + + if (time_before(jiffies, calc_load_update + 10)) + return; + + /* + * Fold the 'old' idle-delta to include all NO_HZ cpus. + */ + delta = calc_load_fold_idle(); + if (delta) + atomic_long_add(delta, &calc_load_tasks); + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + 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; + + /* + * In case we idled for multiple LOAD_FREQ intervals, catch up in bulk. + */ + calc_global_nohz(); +} + +/* + * Called from update_cpu_load() to periodically update this CPU's + * active count. + */ +static void calc_load_account_active(struct rq *this_rq) +{ + long delta; + + if (time_before(jiffies, this_rq->calc_load_update)) + return; + + delta = calc_load_fold_active(this_rq); + if (delta) + atomic_long_add(delta, &calc_load_tasks); + + this_rq->calc_load_update += LOAD_FREQ; +} + +/* + * End of global load-average stuff + */ + +/* + * The exact cpuload at various idx values, calculated at every tick would be + * load = (2^idx - 1) / 2^idx * load + 1 / 2^idx * cur_load + * + * If a cpu misses updates for n-1 ticks (as it was idle) and update gets called + * on nth tick when cpu may be busy, then we have: + * load = ((2^idx - 1) / 2^idx)^(n-1) * load + * load = (2^idx - 1) / 2^idx) * load + 1 / 2^idx * cur_load + * + * decay_load_missed() below does efficient calculation of + * load = ((2^idx - 1) / 2^idx)^(n-1) * load + * avoiding 0..n-1 loop doing load = ((2^idx - 1) / 2^idx) * load + * + * The calculation is approximated on a 128 point scale. + * degrade_zero_ticks is the number of ticks after which load at any + * particular idx is approximated to be zero. + * degrade_factor is a precomputed table, a row for each load idx. + * Each column corresponds to degradation factor for a power of two ticks, + * based on 128 point scale. + * Example: + * row 2, col 3 (=12) says that the degradation at load idx 2 after + * 8 ticks is 12/128 (which is an approximation of exact factor 3^8/4^8). + * + * With this power of 2 load factors, we can degrade the load n times + * by looking at 1 bits in n and doing as many mult/shift instead of + * n mult/shifts needed by the exact degradation. + */ +#define DEGRADE_SHIFT 7 +static const unsigned char + degrade_zero_ticks[CPU_LOAD_IDX_MAX] = {0, 8, 32, 64, 128}; +static const unsigned char + degrade_factor[CPU_LOAD_IDX_MAX][DEGRADE_SHIFT + 1] = { + {0, 0, 0, 0, 0, 0, 0, 0}, + {64, 32, 8, 0, 0, 0, 0, 0}, + {96, 72, 40, 12, 1, 0, 0}, + {112, 98, 75, 43, 15, 1, 0}, + {120, 112, 98, 76, 45, 16, 2} }; + +/* + * Update cpu_load for any missed ticks, due to tickless idle. The backlog + * would be when CPU is idle and so we just decay the old load without + * adding any new load. + */ +static unsigned long +decay_load_missed(unsigned long load, unsigned long missed_updates, int idx) +{ + int j = 0; + + if (!missed_updates) + return load; + + if (missed_updates >= degrade_zero_ticks[idx]) + return 0; + + if (idx == 1) + return load >> missed_updates; + + while (missed_updates) { + if (missed_updates % 2) + load = (load * degrade_factor[idx][j]) >> DEGRADE_SHIFT; + + missed_updates >>= 1; + j++; + } + return load; +} + +/* + * Update rq->cpu_load[] statistics. This function is usually called every + * scheduler tick (TICK_NSEC). With tickless idle this will not be called + * every tick. We fix it up based on jiffies. + */ +static void __update_cpu_load(struct rq *this_rq, unsigned long this_load, + unsigned long pending_updates) +{ + int i, scale; + + this_rq->nr_load_updates++; + + /* Update our load: */ + this_rq->cpu_load[0] = this_load; /* Fasttrack for idx 0 */ + for (i = 1, scale = 2; i < CPU_LOAD_IDX_MAX; i++, scale += scale) { + unsigned long old_load, new_load; + + /* scale is effectively 1 << i now, and >> i divides by scale */ + + old_load = this_rq->cpu_load[i]; + old_load = decay_load_missed(old_load, pending_updates - 1, i); + new_load = this_load; + /* + * Round up the averaging division if load is increasing. This + * prevents us from getting stuck on 9 if the load is 10, for + * example. + */ + if (new_load > old_load) + new_load += scale - 1; + + this_rq->cpu_load[i] = (old_load * (scale - 1) + new_load) >> i; + } + + sched_avg_update(this_rq); +} + +#ifdef CONFIG_SMP +static inline unsigned long get_rq_runnable_load(struct rq *rq) +{ + return rq->cfs.runnable_load_avg; +} +#else +static inline unsigned long get_rq_runnable_load(struct rq *rq) +{ + return rq->load.weight; +} +#endif + +#ifdef CONFIG_NO_HZ_COMMON +/* + * There is no sane way to deal with nohz on smp when using jiffies because the + * cpu doing the jiffies update might drift wrt the cpu doing the jiffy reading + * causing off-by-one errors in observed deltas; {0,2} instead of {1,1}. + * + * Therefore we cannot use the delta approach from the regular tick since that + * would seriously skew the load calculation. However we'll make do for those + * updates happening while idle (nohz_idle_balance) or coming out of idle + * (tick_nohz_idle_exit). + * + * This means we might still be one tick off for nohz periods. + */ + +/* + * Called from nohz_idle_balance() to update the load ratings before doing the + * idle balance. + */ +void update_idle_cpu_load(struct rq *this_rq) +{ + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); + unsigned long load = get_rq_runnable_load(this_rq); + unsigned long pending_updates; + + /* + * bail if there's load or we're actually up-to-date. + */ + if (load || curr_jiffies == this_rq->last_load_update_tick) + return; + + pending_updates = curr_jiffies - this_rq->last_load_update_tick; + this_rq->last_load_update_tick = curr_jiffies; + + __update_cpu_load(this_rq, load, pending_updates); +} + +/* + * Called from tick_nohz_idle_exit() -- try and fix up the ticks we missed. + */ +void update_cpu_load_nohz(void) +{ + struct rq *this_rq = this_rq(); + unsigned long curr_jiffies = ACCESS_ONCE(jiffies); + unsigned long pending_updates; + + if (curr_jiffies == this_rq->last_load_update_tick) + return; + + raw_spin_lock(&this_rq->lock); + pending_updates = curr_jiffies - this_rq->last_load_update_tick; + if (pending_updates) { + this_rq->last_load_update_tick = curr_jiffies; + /* + * We were idle, this means load 0, the current load might be + * !0 due to remote wakeups and the sort. + */ + __update_cpu_load(this_rq, 0, pending_updates); + } + raw_spin_unlock(&this_rq->lock); +} +#endif /* CONFIG_NO_HZ */ + +/* + * Called from scheduler_tick() + */ +void update_cpu_load_active(struct rq *this_rq) +{ + unsigned long load = get_rq_runnable_load(this_rq); + /* + * See the mess around update_idle_cpu_load() / update_cpu_load_nohz(). + */ + this_rq->last_load_update_tick = jiffies; + __update_cpu_load(this_rq, load, 1); + + calc_load_account_active(this_rq); +} diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 127a2c4cf4ab..01970c8e64df 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -399,20 +399,6 @@ static inline struct task_group *next_task_group(struct task_group *tg) (iter = next_task_group(iter)) && \ (rt_rq = iter->rt_rq[cpu_of(rq)]);) -static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq) -{ - list_add_rcu(&rt_rq->leaf_rt_rq_list, - &rq_of_rt_rq(rt_rq)->leaf_rt_rq_list); -} - -static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq) -{ - list_del_rcu(&rt_rq->leaf_rt_rq_list); -} - -#define for_each_leaf_rt_rq(rt_rq, rq) \ - list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list) - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = rt_se->parent) @@ -472,7 +458,7 @@ static int rt_se_boosted(struct sched_rt_entity *rt_se) #ifdef CONFIG_SMP static inline const struct cpumask *sched_rt_period_mask(void) { - return cpu_rq(smp_processor_id())->rd->span; + return this_rq()->rd->span; } #else static inline const struct cpumask *sched_rt_period_mask(void) @@ -509,17 +495,6 @@ typedef struct rt_rq *rt_rq_iter_t; #define for_each_rt_rq(rt_rq, iter, rq) \ for ((void) iter, rt_rq = &rq->rt; rt_rq; rt_rq = NULL) -static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq) -{ -} - -static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq) -{ -} - -#define for_each_leaf_rt_rq(rt_rq, rq) \ - for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL) - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = NULL) @@ -699,15 +674,6 @@ balanced: } } -static void disable_runtime(struct rq *rq) -{ - unsigned long flags; - - raw_spin_lock_irqsave(&rq->lock, flags); - __disable_runtime(rq); - raw_spin_unlock_irqrestore(&rq->lock, flags); -} - static void __enable_runtime(struct rq *rq) { rt_rq_iter_t iter; @@ -732,37 +698,6 @@ static void __enable_runtime(struct rq *rq) } } -static void enable_runtime(struct rq *rq) -{ - unsigned long flags; - - raw_spin_lock_irqsave(&rq->lock, flags); - __enable_runtime(rq); - raw_spin_unlock_irqrestore(&rq->lock, flags); -} - -int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu) -{ - int cpu = (int)(long)hcpu; - - switch (action) { - case CPU_DOWN_PREPARE: - case CPU_DOWN_PREPARE_FROZEN: - disable_runtime(cpu_rq(cpu)); - return NOTIFY_OK; - - case CPU_DOWN_FAILED: - case CPU_DOWN_FAILED_FROZEN: - case CPU_ONLINE: - case CPU_ONLINE_FROZEN: - enable_runtime(cpu_rq(cpu)); - return NOTIFY_OK; - - default: - return NOTIFY_DONE; - } -} - static int balance_runtime(struct rt_rq *rt_rq) { int more = 0; @@ -926,7 +861,7 @@ static void update_curr_rt(struct rq *rq) if (curr->sched_class != &rt_sched_class) return; - delta_exec = rq->clock_task - curr->se.exec_start; + delta_exec = rq_clock_task(rq) - curr->se.exec_start; if (unlikely((s64)delta_exec <= 0)) return; @@ -936,7 +871,7 @@ static void update_curr_rt(struct rq *rq) curr->se.sum_exec_runtime += delta_exec; account_group_exec_runtime(curr, delta_exec); - curr->se.exec_start = rq->clock_task; + curr->se.exec_start = rq_clock_task(rq); cpuacct_charge(curr, delta_exec); sched_rt_avg_update(rq, delta_exec); @@ -1106,9 +1041,6 @@ static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running)) return; - if (!rt_rq->rt_nr_running) - list_add_leaf_rt_rq(rt_rq); - if (head) list_add(&rt_se->run_list, queue); else @@ -1128,8 +1060,6 @@ static void __dequeue_rt_entity(struct sched_rt_entity *rt_se) __clear_bit(rt_se_prio(rt_se), array->bitmap); dec_rt_tasks(rt_se, rt_rq); - if (!rt_rq->rt_nr_running) - list_del_leaf_rt_rq(rt_rq); } /* @@ -1385,7 +1315,7 @@ static struct task_struct *_pick_next_task_rt(struct rq *rq) } while (rt_rq); p = rt_task_of(rt_se); - p->se.exec_start = rq->clock_task; + p->se.exec_start = rq_clock_task(rq); return p; } @@ -1434,42 +1364,24 @@ static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu) return 0; } -/* Return the second highest RT task, NULL otherwise */ -static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) +/* + * Return the highest pushable rq's task, which is suitable to be executed + * on the cpu, NULL otherwise + */ +static struct task_struct *pick_highest_pushable_task(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; - - 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; + struct plist_head *head = &rq->rt.pushable_tasks; + struct task_struct *p; - if (!rt_entity_is_task(rt_se)) - continue; + if (!has_pushable_tasks(rq)) + return NULL; - 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; - } + plist_for_each_entry(p, head, pushable_tasks) { + if (pick_rt_task(rq, p, cpu)) + return p; } - return next; + return NULL; } static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask); @@ -1743,12 +1655,10 @@ static int pull_rt_task(struct rq *this_rq) double_lock_balance(this_rq, src_rq); /* - * Are there still pullable RT tasks? + * We can pull only a task, which is pushable + * on its rq, and no others. */ - if (src_rq->rt.rt_nr_running <= 1) - goto skip; - - p = pick_next_highest_task_rt(src_rq, this_cpu); + p = pick_highest_pushable_task(src_rq, this_cpu); /* * Do we have an RT task that preempts @@ -2037,7 +1947,7 @@ static void set_curr_task_rt(struct rq *rq) { struct task_struct *p = rq->curr; - p->se.exec_start = rq->clock_task; + p->se.exec_start = rq_clock_task(rq); /* The running task is never eligible for pushing */ dequeue_pushable_task(rq, p); diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ce39224d6155..ef0a7b2439dd 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -10,8 +10,16 @@ #include "cpupri.h" #include "cpuacct.h" +struct rq; + extern __read_mostly int scheduler_running; +extern unsigned long calc_load_update; +extern atomic_long_t calc_load_tasks; + +extern long calc_load_fold_active(struct rq *this_rq); +extern void update_cpu_load_active(struct rq *this_rq); + /* * Convert user-nice values [ -20 ... 0 ... 19 ] * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ], @@ -140,10 +148,11 @@ struct task_group { struct cfs_rq **cfs_rq; unsigned long shares; - atomic_t load_weight; - atomic64_t load_avg; +#ifdef CONFIG_SMP + atomic_long_t load_avg; atomic_t runnable_avg; #endif +#endif #ifdef CONFIG_RT_GROUP_SCHED struct sched_rt_entity **rt_se; @@ -261,26 +270,21 @@ struct cfs_rq { #endif #ifdef CONFIG_SMP -/* - * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be - * removed when useful for applications beyond shares distribution (e.g. - * load-balance). - */ -#ifdef CONFIG_FAIR_GROUP_SCHED /* * CFS Load tracking * Under CFS, load is tracked on a per-entity basis and aggregated up. * This allows for the description of both thread and group usage (in * the FAIR_GROUP_SCHED case). */ - u64 runnable_load_avg, blocked_load_avg; - atomic64_t decay_counter, removed_load; + unsigned long runnable_load_avg, blocked_load_avg; + atomic64_t decay_counter; u64 last_decay; -#endif /* CONFIG_FAIR_GROUP_SCHED */ -/* These always depend on CONFIG_FAIR_GROUP_SCHED */ + atomic_long_t removed_load; + #ifdef CONFIG_FAIR_GROUP_SCHED + /* Required to track per-cpu representation of a task_group */ u32 tg_runnable_contrib; - u64 tg_load_contrib; + unsigned long tg_load_contrib; #endif /* CONFIG_FAIR_GROUP_SCHED */ /* @@ -353,7 +357,6 @@ struct rt_rq { unsigned long rt_nr_boosted; struct rq *rq; - struct list_head leaf_rt_rq_list; struct task_group *tg; #endif }; @@ -540,6 +543,16 @@ DECLARE_PER_CPU(struct rq, runqueues); #define cpu_curr(cpu) (cpu_rq(cpu)->curr) #define raw_rq() (&__raw_get_cpu_var(runqueues)) +static inline u64 rq_clock(struct rq *rq) +{ + return rq->clock; +} + +static inline u64 rq_clock_task(struct rq *rq) +{ + return rq->clock_task; +} + #ifdef CONFIG_SMP #define rcu_dereference_check_sched_domain(p) \ @@ -884,24 +897,6 @@ static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) #define WF_FORK 0x02 /* child wakeup after fork */ #define WF_MIGRATED 0x4 /* internal use, task got migrated */ -static inline void update_load_add(struct load_weight *lw, unsigned long inc) -{ - lw->weight += inc; - lw->inv_weight = 0; -} - -static inline void update_load_sub(struct load_weight *lw, unsigned long dec) -{ - lw->weight -= dec; - lw->inv_weight = 0; -} - -static inline void update_load_set(struct load_weight *lw, unsigned long w) -{ - lw->weight = w; - lw->inv_weight = 0; -} - /* * To aid in avoiding the subversion of "niceness" due to uneven distribution * of tasks with abnormal "nice" values across CPUs the contribution that @@ -1028,17 +1023,8 @@ extern void update_group_power(struct sched_domain *sd, int cpu); extern void trigger_load_balance(struct rq *rq, int cpu); extern void idle_balance(int this_cpu, struct rq *this_rq); -/* - * Only depends on SMP, FAIR_GROUP_SCHED may be removed when runnable_avg - * becomes useful in lb - */ -#if defined(CONFIG_FAIR_GROUP_SCHED) extern void idle_enter_fair(struct rq *this_rq); extern void idle_exit_fair(struct rq *this_rq); -#else -static inline void idle_enter_fair(struct rq *this_rq) {} -static inline void idle_exit_fair(struct rq *this_rq) {} -#endif #else /* CONFIG_SMP */ @@ -1051,7 +1037,6 @@ static inline void idle_balance(int cpu, struct rq *rq) extern void sysrq_sched_debug_show(void); extern void sched_init_granularity(void); extern void update_max_interval(void); -extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu); extern void init_sched_rt_class(void); extern void init_sched_fair_class(void); @@ -1063,6 +1048,8 @@ extern void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime extern void update_idle_cpu_load(struct rq *this_rq); +extern void init_task_runnable_average(struct task_struct *p); + #ifdef CONFIG_PARAVIRT static inline u64 steal_ticks(u64 steal) { diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h index 2ef90a51ec5e..17d7065c3872 100644 --- a/kernel/sched/stats.h +++ b/kernel/sched/stats.h @@ -61,7 +61,7 @@ static inline void sched_info_reset_dequeued(struct task_struct *t) */ static inline void sched_info_dequeued(struct task_struct *t) { - unsigned long long now = task_rq(t)->clock, delta = 0; + unsigned long long now = rq_clock(task_rq(t)), delta = 0; if (unlikely(sched_info_on())) if (t->sched_info.last_queued) @@ -79,7 +79,7 @@ static inline void sched_info_dequeued(struct task_struct *t) */ static void sched_info_arrive(struct task_struct *t) { - unsigned long long now = task_rq(t)->clock, delta = 0; + unsigned long long now = rq_clock(task_rq(t)), delta = 0; if (t->sched_info.last_queued) delta = now - t->sched_info.last_queued; @@ -100,7 +100,7 @@ static inline void sched_info_queued(struct task_struct *t) { if (unlikely(sched_info_on())) if (!t->sched_info.last_queued) - t->sched_info.last_queued = task_rq(t)->clock; + t->sched_info.last_queued = rq_clock(task_rq(t)); } /* @@ -112,7 +112,7 @@ static inline void sched_info_queued(struct task_struct *t) */ static inline void sched_info_depart(struct task_struct *t) { - unsigned long long delta = task_rq(t)->clock - + unsigned long long delta = rq_clock(task_rq(t)) - t->sched_info.last_arrival; rq_sched_info_depart(task_rq(t), delta); diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c index da5eb5bed84a..e08fbeeb54b9 100644 --- a/kernel/sched/stop_task.c +++ b/kernel/sched/stop_task.c @@ -28,7 +28,7 @@ static struct task_struct *pick_next_task_stop(struct rq *rq) struct task_struct *stop = rq->stop; if (stop && stop->on_rq) { - stop->se.exec_start = rq->clock_task; + stop->se.exec_start = rq_clock_task(rq); return stop; } @@ -57,7 +57,7 @@ static void put_prev_task_stop(struct rq *rq, struct task_struct *prev) struct task_struct *curr = rq->curr; u64 delta_exec; - delta_exec = rq->clock_task - curr->se.exec_start; + delta_exec = rq_clock_task(rq) - curr->se.exec_start; if (unlikely((s64)delta_exec < 0)) delta_exec = 0; @@ -67,7 +67,7 @@ static void put_prev_task_stop(struct rq *rq, struct task_struct *prev) curr->se.sum_exec_runtime += delta_exec; account_group_exec_runtime(curr, delta_exec); - curr->se.exec_start = rq->clock_task; + curr->se.exec_start = rq_clock_task(rq); cpuacct_charge(curr, delta_exec); } @@ -79,7 +79,7 @@ static void set_curr_task_stop(struct rq *rq) { struct task_struct *stop = rq->stop; - stop->se.exec_start = rq->clock_task; + stop->se.exec_start = rq_clock_task(rq); } static void switched_to_stop(struct rq *rq, struct task_struct *p) diff --git a/kernel/signal.c b/kernel/signal.c index 113411bfe8b1..50e41075ac77 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -2848,7 +2848,7 @@ int do_sigtimedwait(const sigset_t *which, siginfo_t *info, recalc_sigpending(); spin_unlock_irq(&tsk->sighand->siglock); - timeout = schedule_timeout_interruptible(timeout); + timeout = freezable_schedule_timeout_interruptible(timeout); spin_lock_irq(&tsk->sighand->siglock); __set_task_blocked(tsk, &tsk->real_blocked); diff --git a/kernel/softirq.c b/kernel/softirq.c index b5197dcb0dad..ca25e6e704a2 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -127,8 +127,7 @@ static inline void __local_bh_disable(unsigned long ip, unsigned int cnt) void local_bh_disable(void) { - __local_bh_disable((unsigned long)__builtin_return_address(0), - SOFTIRQ_DISABLE_OFFSET); + __local_bh_disable(_RET_IP_, SOFTIRQ_DISABLE_OFFSET); } EXPORT_SYMBOL(local_bh_disable); @@ -139,7 +138,7 @@ static void __local_bh_enable(unsigned int cnt) WARN_ON_ONCE(!irqs_disabled()); if (softirq_count() == cnt) - trace_softirqs_on((unsigned long)__builtin_return_address(0)); + trace_softirqs_on(_RET_IP_); sub_preempt_count(cnt); } @@ -184,7 +183,7 @@ static inline void _local_bh_enable_ip(unsigned long ip) void local_bh_enable(void) { - _local_bh_enable_ip((unsigned long)__builtin_return_address(0)); + _local_bh_enable_ip(_RET_IP_); } EXPORT_SYMBOL(local_bh_enable); @@ -195,8 +194,12 @@ void local_bh_enable_ip(unsigned long ip) EXPORT_SYMBOL(local_bh_enable_ip); /* - * We restart softirq processing for at most 2 ms, - * and if need_resched() is not set. + * We restart softirq processing for at most MAX_SOFTIRQ_RESTART times, + * but break the loop if need_resched() is set or after 2 ms. + * The MAX_SOFTIRQ_TIME provides a nice upper bound in most cases, but in + * certain cases, such as stop_machine(), jiffies may cease to + * increment and so we need the MAX_SOFTIRQ_RESTART limit as + * well to make sure we eventually return from this method. * * These limits have been established via experimentation. * The two things to balance is latency against fairness - @@ -204,6 +207,7 @@ EXPORT_SYMBOL(local_bh_enable_ip); * should not be able to lock up the box. */ #define MAX_SOFTIRQ_TIME msecs_to_jiffies(2) +#define MAX_SOFTIRQ_RESTART 10 asmlinkage void __do_softirq(void) { @@ -212,6 +216,7 @@ asmlinkage void __do_softirq(void) unsigned long end = jiffies + MAX_SOFTIRQ_TIME; int cpu; unsigned long old_flags = current->flags; + int max_restart = MAX_SOFTIRQ_RESTART; /* * Mask out PF_MEMALLOC s current task context is borrowed for the @@ -223,8 +228,7 @@ asmlinkage void __do_softirq(void) pending = local_softirq_pending(); account_irq_enter_time(current); - __local_bh_disable((unsigned long)__builtin_return_address(0), - SOFTIRQ_OFFSET); + __local_bh_disable(_RET_IP_, SOFTIRQ_OFFSET); lockdep_softirq_enter(); cpu = smp_processor_id(); @@ -265,7 +269,8 @@ restart: pending = local_softirq_pending(); if (pending) { - if (time_before(jiffies, end) && !need_resched()) + if (time_before(jiffies, end) && !need_resched() && + --max_restart) goto restart; wakeup_softirqd(); diff --git a/kernel/sys.c b/kernel/sys.c index b95d3c72ba21..071de900c824 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -362,6 +362,29 @@ int unregister_reboot_notifier(struct notifier_block *nb) } EXPORT_SYMBOL(unregister_reboot_notifier); +/* Add backwards compatibility for stable trees. */ +#ifndef PF_NO_SETAFFINITY +#define PF_NO_SETAFFINITY PF_THREAD_BOUND +#endif + +static void migrate_to_reboot_cpu(void) +{ + /* The boot cpu is always logical cpu 0 */ + int cpu = 0; + + cpu_hotplug_disable(); + + /* Make certain the cpu I'm about to reboot on is online */ + if (!cpu_online(cpu)) + cpu = cpumask_first(cpu_online_mask); + + /* Prevent races with other tasks migrating this task */ + current->flags |= PF_NO_SETAFFINITY; + + /* Make certain I only run on the appropriate processor */ + set_cpus_allowed_ptr(current, cpumask_of(cpu)); +} + /** * kernel_restart - reboot the system * @cmd: pointer to buffer containing command to execute for restart @@ -373,7 +396,7 @@ EXPORT_SYMBOL(unregister_reboot_notifier); void kernel_restart(char *cmd) { kernel_restart_prepare(cmd); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); if (!cmd) printk(KERN_EMERG "Restarting system.\n"); @@ -400,7 +423,7 @@ static void kernel_shutdown_prepare(enum system_states state) void kernel_halt(void) { kernel_shutdown_prepare(SYSTEM_HALT); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); printk(KERN_EMERG "System halted.\n"); kmsg_dump(KMSG_DUMP_HALT); @@ -419,7 +442,7 @@ void kernel_power_off(void) kernel_shutdown_prepare(SYSTEM_POWER_OFF); if (pm_power_off_prepare) pm_power_off_prepare(); - disable_nonboot_cpus(); + migrate_to_reboot_cpu(); syscore_shutdown(); printk(KERN_EMERG "Power down.\n"); kmsg_dump(KMSG_DUMP_POWEROFF); @@ -488,7 +511,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, case LINUX_REBOOT_CMD_HALT: kernel_halt(); do_exit(0); - panic("cannot halt"); + panic("cannot halt.\n"); case LINUX_REBOOT_CMD_POWER_OFF: kernel_power_off(); @@ -1286,6 +1309,17 @@ out: return retval; } +static void set_special_pids(struct pid *pid) +{ + struct task_struct *curr = current->group_leader; + + if (task_session(curr) != pid) + change_pid(curr, PIDTYPE_SID, pid); + + if (task_pgrp(curr) != pid) + change_pid(curr, PIDTYPE_PGID, pid); +} + SYSCALL_DEFINE0(setsid) { struct task_struct *group_leader = current->group_leader; @@ -1305,7 +1339,7 @@ SYSCALL_DEFINE0(setsid) goto out; group_leader->signal->leader = 1; - __set_special_pids(sid); + set_special_pids(sid); proc_clear_tty(group_leader); @@ -2332,8 +2366,7 @@ static int do_sysinfo(struct sysinfo *info) memset(info, 0, sizeof(struct sysinfo)); - ktime_get_ts(&tp); - monotonic_to_bootbased(&tp); + get_monotonic_boottime(&tp); info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); get_avenrun(info->loads, 0, SI_LOAD_SHIFT - FSHIFT); diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 9edcf456e0fc..4ce13c3cedb9 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -120,7 +120,6 @@ extern int blk_iopoll_enabled; /* Constants used for minimum and maximum */ #ifdef CONFIG_LOCKUP_DETECTOR static int sixty = 60; -static int neg_one = -1; #endif static int zero; @@ -814,7 +813,7 @@ static struct ctl_table kern_table[] = { .maxlen = sizeof(int), .mode = 0644, .proc_handler = proc_dowatchdog, - .extra1 = &neg_one, + .extra1 = &zero, .extra2 = &sixty, }, { @@ -1044,6 +1043,15 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = perf_proc_update_handler, }, + { + .procname = "perf_cpu_time_max_percent", + .data = &sysctl_perf_cpu_time_max_percent, + .maxlen = sizeof(sysctl_perf_cpu_time_max_percent), + .mode = 0644, + .proc_handler = perf_cpu_time_max_percent_handler, + .extra1 = &zero, + .extra2 = &one_hundred, + }, #endif #ifdef CONFIG_KMEMCHECK { diff --git a/kernel/time.c b/kernel/time.c index d3617dbd3dca..7c7964c33ae7 100644 --- a/kernel/time.c +++ b/kernel/time.c @@ -11,7 +11,7 @@ * Modification history kernel/time.c * * 1993-09-02 Philip Gladstone - * Created file with time related functions from sched.c and adjtimex() + * Created file with time related functions from sched/core.c and adjtimex() * 1993-10-08 Torsten Duwe * adjtime interface update and CMOS clock write code * 1995-08-13 Torsten Duwe diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index 12ff13a838c6..8f5b3b98577b 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -874,7 +874,6 @@ static void hardpps_update_phase(long error) void __hardpps(const struct timespec *phase_ts, const struct timespec *raw_ts) { struct pps_normtime pts_norm, freq_norm; - unsigned long flags; pts_norm = pps_normalize_ts(*phase_ts); diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 24938d577669..20d6fba70652 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -511,6 +511,12 @@ again: } } + /* + * Remove the current cpu from the pending mask. The event is + * delivered immediately in tick_do_broadcast() ! + */ + cpumask_clear_cpu(smp_processor_id(), tick_broadcast_pending_mask); + /* Take care of enforced broadcast requests */ cpumask_or(tmpmask, tmpmask, tick_broadcast_force_mask); cpumask_clear(tick_broadcast_force_mask); @@ -575,8 +581,8 @@ void tick_broadcast_oneshot_control(unsigned long reason) raw_spin_lock_irqsave(&tick_broadcast_lock, flags); if (reason == CLOCK_EVT_NOTIFY_BROADCAST_ENTER) { - WARN_ON_ONCE(cpumask_test_cpu(cpu, tick_broadcast_pending_mask)); if (!cpumask_test_and_set_cpu(cpu, tick_broadcast_oneshot_mask)) { + WARN_ON_ONCE(cpumask_test_cpu(cpu, tick_broadcast_pending_mask)); clockevents_set_mode(dev, CLOCK_EVT_MODE_SHUTDOWN); /* * We only reprogram the broadcast timer if we @@ -593,8 +599,6 @@ void tick_broadcast_oneshot_control(unsigned long reason) } else { if (cpumask_test_and_clear_cpu(cpu, tick_broadcast_oneshot_mask)) { clockevents_set_mode(dev, CLOCK_EVT_MODE_ONESHOT); - if (dev->next_event.tv64 == KTIME_MAX) - goto out; /* * The cpu which was handling the broadcast * timer marked this cpu in the broadcast @@ -609,6 +613,11 @@ void tick_broadcast_oneshot_control(unsigned long reason) goto out; /* + * Bail out if there is no next event. + */ + if (dev->next_event.tv64 == KTIME_MAX) + goto out; + /* * If the pending bit is not set, then we are * either the CPU handling the broadcast * interrupt or we got woken by something else. @@ -692,10 +701,6 @@ void tick_broadcast_setup_oneshot(struct clock_event_device *bc) bc->event_handler = tick_handle_oneshot_broadcast; - /* Take the do_timer update */ - if (!tick_nohz_full_cpu(cpu)) - tick_do_timer_cpu = cpu; - /* * We must be careful here. There might be other CPUs * waiting for periodic broadcast. We need to set the diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index f4208138fbf4..0cf1c1453181 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -306,7 +306,7 @@ static int __cpuinit tick_nohz_cpu_down_callback(struct notifier_block *nfb, * we can't safely shutdown that CPU. */ if (have_nohz_full_mask && tick_do_timer_cpu == cpu) - return -EINVAL; + return NOTIFY_BAD; break; } return NOTIFY_OK; diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 98cd470bbe49..baeeb5c87cf1 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -975,6 +975,14 @@ static int timekeeping_suspend(void) read_persistent_clock(&timekeeping_suspend_time); + /* + * On some systems the persistent_clock can not be detected at + * timekeeping_init by its return value, so if we see a valid + * value returned, update the persistent_clock_exists flag. + */ + if (timekeeping_suspend_time.tv_sec || timekeeping_suspend_time.tv_nsec) + persistent_clock_exist = true; + raw_spin_lock_irqsave(&timekeeper_lock, flags); write_seqcount_begin(&timekeeper_seq); timekeeping_forward_now(tk); diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index b549b0f5b977..6c508ff33c62 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -120,22 +120,22 @@ static void ftrace_ops_no_ops(unsigned long ip, unsigned long parent_ip); /* * Traverse the ftrace_global_list, invoking all entries. The reason that we - * can use rcu_dereference_raw() is that elements removed from this list + * can use rcu_dereference_raw_notrace() is that elements removed from this list * are simply leaked, so there is no need to interact with a grace-period - * mechanism. The rcu_dereference_raw() calls are needed to handle + * mechanism. The rcu_dereference_raw_notrace() calls are needed to handle * concurrent insertions into the ftrace_global_list. * * Silly Alpha and silly pointer-speculation compiler optimizations! */ #define do_for_each_ftrace_op(op, list) \ - op = rcu_dereference_raw(list); \ + op = rcu_dereference_raw_notrace(list); \ do /* * Optimized for just a single item in the list (as that is the normal case). */ #define while_for_each_ftrace_op(op) \ - while (likely(op = rcu_dereference_raw((op)->next)) && \ + while (likely(op = rcu_dereference_raw_notrace((op)->next)) && \ unlikely((op) != &ftrace_list_end)) static inline void ftrace_ops_init(struct ftrace_ops *ops) @@ -779,7 +779,7 @@ ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip) if (hlist_empty(hhd)) return NULL; - hlist_for_each_entry_rcu(rec, hhd, node) { + hlist_for_each_entry_rcu_notrace(rec, hhd, node) { if (rec->ip == ip) return rec; } @@ -1165,7 +1165,7 @@ ftrace_lookup_ip(struct ftrace_hash *hash, unsigned long ip) hhd = &hash->buckets[key]; - hlist_for_each_entry_rcu(entry, hhd, hlist) { + hlist_for_each_entry_rcu_notrace(entry, hhd, hlist) { if (entry->ip == ip) return entry; } @@ -1422,8 +1422,8 @@ ftrace_ops_test(struct ftrace_ops *ops, unsigned long ip) struct ftrace_hash *notrace_hash; int ret; - filter_hash = rcu_dereference_raw(ops->filter_hash); - notrace_hash = rcu_dereference_raw(ops->notrace_hash); + filter_hash = rcu_dereference_raw_notrace(ops->filter_hash); + notrace_hash = rcu_dereference_raw_notrace(ops->notrace_hash); if ((ftrace_hash_empty(filter_hash) || ftrace_lookup_ip(filter_hash, ip)) && @@ -2920,7 +2920,7 @@ static void function_trace_probe_call(unsigned long ip, unsigned long parent_ip, * on the hash. rcu_read_lock is too dangerous here. */ preempt_disable_notrace(); - hlist_for_each_entry_rcu(entry, hhd, node) { + hlist_for_each_entry_rcu_notrace(entry, hhd, node) { if (entry->ip == ip) entry->ops->func(ip, parent_ip, &entry->data); } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 4d79485b3237..e71a8be4a6ee 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -652,8 +652,6 @@ static struct { ARCH_TRACE_CLOCKS }; -int trace_clock_id; - /* * trace_parser_get_init - gets the buffer for trace parser */ @@ -843,7 +841,15 @@ __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) memcpy(max_data->comm, tsk->comm, TASK_COMM_LEN); max_data->pid = tsk->pid; - max_data->uid = task_uid(tsk); + /* + * If tsk == current, then use current_uid(), as that does not use + * RCU. The irq tracer can be called out of RCU scope. + */ + if (tsk == current) + max_data->uid = current_uid(); + else + max_data->uid = task_uid(tsk); + max_data->nice = tsk->static_prio - 20 - MAX_RT_PRIO; max_data->policy = tsk->policy; max_data->rt_priority = tsk->rt_priority; @@ -2818,7 +2824,7 @@ __tracing_open(struct inode *inode, struct file *file, bool snapshot) iter->iter_flags |= TRACE_FILE_ANNOTATE; /* Output in nanoseconds only if we are using a clock in nanoseconds. */ - if (trace_clocks[trace_clock_id].in_ns) + if (trace_clocks[tr->clock_id].in_ns) iter->iter_flags |= TRACE_FILE_TIME_IN_NS; /* stop the trace while dumping if we are not opening "snapshot" */ @@ -3817,7 +3823,7 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) iter->iter_flags |= TRACE_FILE_LAT_FMT; /* Output in nanoseconds only if we are using a clock in nanoseconds. */ - if (trace_clocks[trace_clock_id].in_ns) + if (trace_clocks[tr->clock_id].in_ns) iter->iter_flags |= TRACE_FILE_TIME_IN_NS; iter->cpu_file = tc->cpu; @@ -5087,7 +5093,7 @@ tracing_stats_read(struct file *filp, char __user *ubuf, cnt = ring_buffer_bytes_cpu(trace_buf->buffer, cpu); trace_seq_printf(s, "bytes: %ld\n", cnt); - if (trace_clocks[trace_clock_id].in_ns) { + if (trace_clocks[tr->clock_id].in_ns) { /* local or global for trace_clock */ t = ns2usecs(ring_buffer_oldest_event_ts(trace_buf->buffer, cpu)); usec_rem = do_div(t, USEC_PER_SEC); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 711ca7d3e7f1..20572ed88c5c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -700,8 +700,6 @@ enum print_line_t print_trace_line(struct trace_iterator *iter); extern unsigned long trace_flags; -extern int trace_clock_id; - /* Standard output formatting function used for function return traces */ #ifdef CONFIG_FUNCTION_GRAPH_TRACER diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 55e2cf66967b..2901e3b88590 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -1159,7 +1159,7 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) /* stop the tracing. */ tracing_stop(); /* check the trace buffer */ - ret = trace_test_buffer(tr, &count); + ret = trace_test_buffer(&tr->trace_buffer, &count); trace->reset(tr); tracing_start(); diff --git a/kernel/wait.c b/kernel/wait.c index 6698e0c04ead..ce0daa320a26 100644 --- a/kernel/wait.c +++ b/kernel/wait.c @@ -287,3 +287,91 @@ wait_queue_head_t *bit_waitqueue(void *word, int bit) return &zone->wait_table[hash_long(val, zone->wait_table_bits)]; } EXPORT_SYMBOL(bit_waitqueue); + +/* + * Manipulate the atomic_t address to produce a better bit waitqueue table hash + * index (we're keying off bit -1, but that would produce a horrible hash + * value). + */ +static inline wait_queue_head_t *atomic_t_waitqueue(atomic_t *p) +{ + if (BITS_PER_LONG == 64) { + unsigned long q = (unsigned long)p; + return bit_waitqueue((void *)(q & ~1), q & 1); + } + return bit_waitqueue(p, 0); +} + +static int wake_atomic_t_function(wait_queue_t *wait, unsigned mode, int sync, + void *arg) +{ + struct wait_bit_key *key = arg; + struct wait_bit_queue *wait_bit + = container_of(wait, struct wait_bit_queue, wait); + atomic_t *val = key->flags; + + if (wait_bit->key.flags != key->flags || + wait_bit->key.bit_nr != key->bit_nr || + atomic_read(val) != 0) + return 0; + return autoremove_wake_function(wait, mode, sync, key); +} + +/* + * To allow interruptible waiting and asynchronous (i.e. nonblocking) waiting, + * the actions of __wait_on_atomic_t() are permitted return codes. Nonzero + * return codes halt waiting and return. + */ +static __sched +int __wait_on_atomic_t(wait_queue_head_t *wq, struct wait_bit_queue *q, + int (*action)(atomic_t *), unsigned mode) +{ + atomic_t *val; + int ret = 0; + + do { + prepare_to_wait(wq, &q->wait, mode); + val = q->key.flags; + if (atomic_read(val) == 0) + ret = (*action)(val); + } while (!ret && atomic_read(val) != 0); + finish_wait(wq, &q->wait); + return ret; +} + +#define DEFINE_WAIT_ATOMIC_T(name, p) \ + struct wait_bit_queue name = { \ + .key = __WAIT_ATOMIC_T_KEY_INITIALIZER(p), \ + .wait = { \ + .private = current, \ + .func = wake_atomic_t_function, \ + .task_list = \ + LIST_HEAD_INIT((name).wait.task_list), \ + }, \ + } + +__sched int out_of_line_wait_on_atomic_t(atomic_t *p, int (*action)(atomic_t *), + unsigned mode) +{ + wait_queue_head_t *wq = atomic_t_waitqueue(p); + DEFINE_WAIT_ATOMIC_T(wait, p); + + return __wait_on_atomic_t(wq, &wait, action, mode); +} +EXPORT_SYMBOL(out_of_line_wait_on_atomic_t); + +/** + * wake_up_atomic_t - Wake up a waiter on a atomic_t + * @word: The word being waited on, a kernel virtual address + * @bit: The bit of the word being waited on + * + * Wake up anyone waiting for the atomic_t to go to zero. + * + * Abuse the bit-waker function and its waitqueue hash table set (the atomic_t + * check is done by the waiter's wake function, not the by the waker itself). + */ +void wake_up_atomic_t(atomic_t *p) +{ + __wake_up_bit(atomic_t_waitqueue(p), p, WAIT_ATOMIC_T_BIT_NR); +} +EXPORT_SYMBOL(wake_up_atomic_t); diff --git a/kernel/workqueue.c b/kernel/workqueue.c index ee8e29a2320c..f02c4a4a0c3c 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -272,6 +272,15 @@ static cpumask_var_t *wq_numa_possible_cpumask; static bool wq_disable_numa; module_param_named(disable_numa, wq_disable_numa, bool, 0444); +/* see the comment above the definition of WQ_POWER_EFFICIENT */ +#ifdef CONFIG_WQ_POWER_EFFICIENT_DEFAULT +static bool wq_power_efficient = true; +#else +static bool wq_power_efficient; +#endif + +module_param_named(power_efficient, wq_power_efficient, bool, 0444); + static bool wq_numa_enabled; /* unbound NUMA affinity enabled */ /* buf for wq_update_unbound_numa_attrs(), protected by CPU hotplug exclusion */ @@ -305,6 +314,10 @@ struct workqueue_struct *system_unbound_wq __read_mostly; EXPORT_SYMBOL_GPL(system_unbound_wq); struct workqueue_struct *system_freezable_wq __read_mostly; EXPORT_SYMBOL_GPL(system_freezable_wq); +struct workqueue_struct *system_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_power_efficient_wq); +struct workqueue_struct *system_freezable_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_freezable_power_efficient_wq); static int worker_thread(void *__worker); static void copy_workqueue_attrs(struct workqueue_attrs *to, @@ -4086,6 +4099,10 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, struct workqueue_struct *wq; struct pool_workqueue *pwq; + /* see the comment above the definition of WQ_POWER_EFFICIENT */ + if ((flags & WQ_POWER_EFFICIENT) && wq_power_efficient) + flags |= WQ_UNBOUND; + /* allocate wq and format name */ if (flags & WQ_UNBOUND) tbl_size = wq_numa_tbl_len * sizeof(wq->numa_pwq_tbl[0]); @@ -4985,8 +5002,15 @@ static int __init init_workqueues(void) WQ_UNBOUND_MAX_ACTIVE); system_freezable_wq = alloc_workqueue("events_freezable", WQ_FREEZABLE, 0); + system_power_efficient_wq = alloc_workqueue("events_power_efficient", + WQ_POWER_EFFICIENT, 0); + system_freezable_power_efficient_wq = alloc_workqueue("events_freezable_power_efficient", + WQ_FREEZABLE | WQ_POWER_EFFICIENT, + 0); BUG_ON(!system_wq || !system_highpri_wq || !system_long_wq || - !system_unbound_wq || !system_freezable_wq); + !system_unbound_wq || !system_freezable_wq || + !system_power_efficient_wq || + !system_freezable_power_efficient_wq); return 0; } early_initcall(init_workqueues); diff --git a/kernel/workqueue_internal.h b/kernel/workqueue_internal.h index ad83c96b2ece..7e2204db0b1a 100644 --- a/kernel/workqueue_internal.h +++ b/kernel/workqueue_internal.h @@ -64,7 +64,7 @@ static inline struct worker *current_wq_worker(void) /* * Scheduler hooks for concurrency managed workqueue. Only to be used from - * sched.c and workqueue.c. + * sched/core.c and workqueue.c. */ void wq_worker_waking_up(struct task_struct *task, int cpu); struct task_struct *wq_worker_sleeping(struct task_struct *task, int cpu); |