X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=kernel%2Fcgroup.c;h=5a54ff42874ee916997139c859b8cfda725b9a33;hb=f6490438fce5902f840d1f0f905295077c635e7a;hp=1a3c23936d43d99ec3d429123c95182a94de0c02;hpb=cfe36bde59bc1ae868e775ad82386c3acaabb738;p=safe%2Fjmp%2Flinux-2.6 diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 1a3c239..5a54ff4 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -44,6 +44,8 @@ #include #include #include +#include +#include #include @@ -82,17 +84,13 @@ struct cgroupfs_root { /* Tracks how many cgroups are currently defined in hierarchy.*/ int number_of_cgroups; - /* A list running through the mounted hierarchies */ + /* A list running through the active hierarchies */ struct list_head root_list; /* Hierarchy-specific flags */ unsigned long flags; - /* The path to use for release notifications. No locking - * between setting and use - so if userspace updates this - * while child cgroups exist, you could miss a - * notification. We ensure that it's always a valid - * NUL-terminated string */ + /* The path to use for release notifications. */ char release_agent_path[PATH_MAX]; }; @@ -113,22 +111,11 @@ static int root_count; #define dummytop (&rootnode.top_cgroup) /* This flag indicates whether tasks in the fork and exit paths should - * take callback_mutex and 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 be called. + * 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 + * be called. */ -static int need_forkexit_callback; - -/* bits in struct cgroup flags field */ -enum { - /* Control Group is dead */ - CGRP_REMOVED, - /* Control Group has previously had a child cgroup or a task, - * but no longer (only if CGRP_NOTIFY_ON_RELEASE is set) */ - CGRP_RELEASABLE, - /* Control Group requires release notifications to userspace */ - CGRP_NOTIFY_ON_RELEASE, -}; +static int need_forkexit_callback __read_mostly; /* convenient tests for these bits */ inline int cgroup_is_removed(const struct cgroup *cgrp) @@ -141,7 +128,7 @@ enum { ROOT_NOPREFIX, /* mounted subsystems have no named prefix */ }; -inline int cgroup_is_releasable(const struct cgroup *cgrp) +static int cgroup_is_releasable(const struct cgroup *cgrp) { const int bits = (1 << CGRP_RELEASABLE) | @@ -149,7 +136,7 @@ inline int cgroup_is_releasable(const struct cgroup *cgrp) return (cgrp->flags & bits) == bits; } -inline int notify_on_release(const struct cgroup *cgrp) +static int notify_on_release(const struct cgroup *cgrp) { return test_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); } @@ -161,8 +148,8 @@ inline int notify_on_release(const struct cgroup *cgrp) #define for_each_subsys(_root, _ss) \ list_for_each_entry(_ss, &_root->subsys_list, sibling) -/* for_each_root() allows you to iterate across the active hierarchies */ -#define for_each_root(_root) \ +/* 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) /* the list of cgroups eligible for automatic release. Protected by @@ -204,11 +191,32 @@ static struct cg_cgroup_link init_css_set_link; static DEFINE_RWLOCK(css_set_lock); static int css_set_count; +/* hash table for cgroup groups. This improves the performance to + * find an existing css_set */ +#define CSS_SET_HASH_BITS 7 +#define CSS_SET_TABLE_SIZE (1 << CSS_SET_HASH_BITS) +static struct hlist_head css_set_table[CSS_SET_TABLE_SIZE]; + +static struct hlist_head *css_set_hash(struct cgroup_subsys_state *css[]) +{ + int i; + int index; + unsigned long tmp = 0UL; + + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) + tmp += (unsigned long)css[i]; + tmp = (tmp >> 16) ^ tmp; + + index = hash_long(tmp, CSS_SET_HASH_BITS); + + return &css_set_table[index]; +} + /* We don't maintain the lists running through each css_set to its * task until after the first call to cgroup_iter_start(). This * reduces the fork()/exit() overhead for people who have cgroups * compiled into their kernel but not actually in use */ -static int use_task_css_set_links; +static int use_task_css_set_links __read_mostly; /* When we create or destroy a css_set, the operation simply * takes/releases a reference count on all the cgroups referenced @@ -229,30 +237,41 @@ static int use_task_css_set_links; */ static void unlink_css_set(struct css_set *cg) { - write_lock(&css_set_lock); - list_del(&cg->list); + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; + + hlist_del(&cg->hlist); css_set_count--; - while (!list_empty(&cg->cg_links)) { - struct cg_cgroup_link *link; - link = list_entry(cg->cg_links.next, - struct cg_cgroup_link, cg_link_list); + + list_for_each_entry_safe(link, saved_link, &cg->cg_links, + cg_link_list) { list_del(&link->cg_link_list); list_del(&link->cgrp_link_list); kfree(link); } - write_unlock(&css_set_lock); } -static void __release_css_set(struct kref *k, int taskexit) +static void __put_css_set(struct css_set *cg, int taskexit) { int i; - struct css_set *cg = container_of(k, struct css_set, ref); - + /* + * 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)) + return; + write_lock(&css_set_lock); + if (!atomic_dec_and_test(&cg->refcount)) { + write_unlock(&css_set_lock); + return; + } unlink_css_set(cg); + write_unlock(&css_set_lock); rcu_read_lock(); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - struct cgroup *cgrp = cg->subsys[i]->cgroup; + struct cgroup *cgrp = rcu_dereference(cg->subsys[i]->cgroup); if (atomic_dec_and_test(&cgrp->count) && notify_on_release(cgrp)) { if (taskexit) @@ -264,40 +283,28 @@ static void __release_css_set(struct kref *k, int taskexit) kfree(cg); } -static void release_css_set(struct kref *k) -{ - __release_css_set(k, 0); -} - -static void release_css_set_taskexit(struct kref *k) -{ - __release_css_set(k, 1); -} - /* * refcounted get/put for css_set objects */ static inline void get_css_set(struct css_set *cg) { - kref_get(&cg->ref); + atomic_inc(&cg->refcount); } static inline void put_css_set(struct css_set *cg) { - kref_put(&cg->ref, release_css_set); + __put_css_set(cg, 0); } static inline void put_css_set_taskexit(struct css_set *cg) { - kref_put(&cg->ref, release_css_set_taskexit); + __put_css_set(cg, 1); } /* * find_existing_css_set() is a helper for * find_css_set(), and checks to see whether an existing - * css_set is suitable. This currently walks a linked-list for - * simplicity; a later patch will use a hash table for better - * performance + * css_set is suitable. * * oldcg: the cgroup group that we're using before the cgroup * transition @@ -307,7 +314,6 @@ static inline void put_css_set_taskexit(struct css_set *cg) * template: location in which to build the desired set of subsystem * state objects for the new cgroup group */ - static struct css_set *find_existing_css_set( struct css_set *oldcg, struct cgroup *cgrp, @@ -315,12 +321,14 @@ static struct css_set *find_existing_css_set( { int i; struct cgroupfs_root *root = cgrp->root; - struct list_head *l = &init_css_set.list; + struct hlist_head *hhead; + struct hlist_node *node; + struct css_set *cg; /* Built the set of subsystem state objects that we want to * see in the new css_set */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - if (root->subsys_bits & (1ull << i)) { + if (root->subsys_bits & (1UL << i)) { /* Subsystem is in this hierarchy. So we want * the subsystem state from the new * cgroup */ @@ -332,29 +340,34 @@ static struct css_set *find_existing_css_set( } } - /* Look through existing cgroup groups to find one to reuse */ - do { - struct css_set *cg = - list_entry(l, struct css_set, list); - + hhead = css_set_hash(template); + hlist_for_each_entry(cg, node, hhead, hlist) { if (!memcmp(template, cg->subsys, sizeof(cg->subsys))) { /* All subsystems matched */ return cg; } - /* Try the next cgroup group */ - l = l->next; - } while (l != &init_css_set.list); + } /* No existing cgroup group matched */ return NULL; } +static void free_cg_links(struct list_head *tmp) +{ + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; + + list_for_each_entry_safe(link, saved_link, tmp, cgrp_link_list) { + list_del(&link->cgrp_link_list); + 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 */ - static int allocate_cg_links(int count, struct list_head *tmp) { struct cg_cgroup_link *link; @@ -363,13 +376,7 @@ static int allocate_cg_links(int count, struct list_head *tmp) for (i = 0; i < count; i++) { link = kmalloc(sizeof(*link), GFP_KERNEL); if (!link) { - while (!list_empty(tmp)) { - link = list_entry(tmp->next, - struct cg_cgroup_link, - cgrp_link_list); - list_del(&link->cgrp_link_list); - kfree(link); - } + free_cg_links(tmp); return -ENOMEM; } list_add(&link->cgrp_link_list, tmp); @@ -377,16 +384,23 @@ static int allocate_cg_links(int count, struct list_head *tmp) return 0; } -static void free_cg_links(struct list_head *tmp) +/** + * 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 + * @cgrp: the destination cgroup + */ +static void link_css_set(struct list_head *tmp_cg_links, + struct css_set *cg, struct cgroup *cgrp) { - while (!list_empty(tmp)) { - struct cg_cgroup_link *link; - link = list_entry(tmp->next, - struct cg_cgroup_link, - cgrp_link_list); - list_del(&link->cgrp_link_list); - kfree(link); - } + struct cg_cgroup_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; + list_move(&link->cgrp_link_list, &cgrp->css_sets); + list_add(&link->cg_link_list, &cg->cg_links); } /* @@ -396,7 +410,6 @@ static void free_cg_links(struct list_head *tmp) * substituted into the appropriate hierarchy. Must be called with * cgroup_mutex held */ - static struct css_set *find_css_set( struct css_set *oldcg, struct cgroup *cgrp) { @@ -405,15 +418,16 @@ static struct css_set *find_css_set( int i; struct list_head tmp_cg_links; - struct cg_cgroup_link *link; + + struct hlist_head *hhead; /* First see if we already have a cgroup group that matches * the desired set */ - write_lock(&css_set_lock); + read_lock(&css_set_lock); res = find_existing_css_set(oldcg, cgrp, template); if (res) get_css_set(res); - write_unlock(&css_set_lock); + read_unlock(&css_set_lock); if (res) return res; @@ -428,9 +442,10 @@ static struct css_set *find_css_set( return NULL; } - kref_init(&res->ref); + atomic_set(&res->refcount, 1); INIT_LIST_HEAD(&res->cg_links); INIT_LIST_HEAD(&res->tasks); + INIT_HLIST_NODE(&res->hlist); /* Copy the set of subsystem state objects generated in * find_existing_css_set() */ @@ -447,33 +462,20 @@ static struct css_set *find_css_set( * only do it for the first subsystem in each * hierarchy */ - if (ss->root->subsys_list.next == &ss->sibling) { - BUG_ON(list_empty(&tmp_cg_links)); - link = list_entry(tmp_cg_links.next, - struct cg_cgroup_link, - cgrp_link_list); - list_del(&link->cgrp_link_list); - list_add(&link->cgrp_link_list, &cgrp->css_sets); - link->cg = res; - list_add(&link->cg_link_list, &res->cg_links); - } - } - if (list_empty(&rootnode.subsys_list)) { - link = list_entry(tmp_cg_links.next, - struct cg_cgroup_link, - cgrp_link_list); - list_del(&link->cgrp_link_list); - list_add(&link->cgrp_link_list, &dummytop->css_sets); - link->cg = res; - list_add(&link->cg_link_list, &res->cg_links); + if (ss->root->subsys_list.next == &ss->sibling) + link_css_set(&tmp_cg_links, res, cgrp); } + if (list_empty(&rootnode.subsys_list)) + link_css_set(&tmp_cg_links, res, dummytop); BUG_ON(!list_empty(&tmp_cg_links)); - /* Link this cgroup group into the list */ - list_add(&res->list, &init_css_set.list); css_set_count++; - INIT_LIST_HEAD(&res->tasks); + + /* Add this cgroup group to the hash table */ + hhead = css_set_hash(res->subsys); + hlist_add_head(&res->hlist, hhead); + write_unlock(&css_set_lock); return res; @@ -489,7 +491,7 @@ static struct css_set *find_css_set( * Any task can increment and decrement the count field without lock. * So in general, code holding cgroup_mutex can't rely on the count * field not changing. However, if the count goes to zero, then only - * attach_task() can increment it again. Because a count of zero + * cgroup_attach_task() can increment it again. Because a count of zero * means that no tasks are currently attached, therefore there is no * way a task attached to that cgroup can fork (the other way to * increment the count). So code holding cgroup_mutex can safely @@ -498,17 +500,13 @@ static struct css_set *find_css_set( * knows that the cgroup won't be removed, as cgroup_rmdir() * needs that mutex. * - * The cgroup_common_file_write handler for operations that modify - * the cgroup hierarchy holds cgroup_mutex across the entire operation, - * single threading all such cgroup modifications across the system. - * * The fork and exit callbacks cgroup_fork() and cgroup_exit(), don't * (usually) take cgroup_mutex. These are the two most performance * critical pieces of code here. The exception occurs on cgroup_exit(), * when a task in a notify_on_release cgroup exits. Then cgroup_mutex * is taken, and if the cgroup count is zero, a usermode call made - * to /sbin/cgroup_release_agent with the name of the cgroup (path - * relative to the root of cgroup file system) as the argument. + * to the release agent with the name of the cgroup (path relative to + * the root of cgroup file system) as the argument. * * A cgroup can only be deleted if both its 'count' of using tasks * is zero, and its list of 'children' cgroups is empty. Since all @@ -520,24 +518,23 @@ static struct css_set *find_css_set( * The task_lock() exception * * The need for this exception arises from the action of - * attach_task(), which overwrites one tasks cgroup pointer with - * another. It does so using cgroup_mutexe, however there are + * cgroup_attach_task(), which overwrites one tasks cgroup pointer with + * another. It does so using cgroup_mutex, however there are * several performance critical places that need to reference * task->cgroup without the expense of grabbing a system global * mutex. Therefore except as noted below, when dereferencing or, as - * in attach_task(), modifying a task'ss cgroup pointer we use + * in cgroup_attach_task(), modifying a task'ss cgroup pointer we use * task_lock(), which acts on a spinlock (task->alloc_lock) already in * the task_struct routinely used for such matters. * * P.S. One more locking exception. RCU is used to guard the - * update of a tasks cgroup pointer by attach_task() + * update of a tasks cgroup pointer by cgroup_attach_task() */ /** * cgroup_lock - lock out any changes to cgroup structures * */ - void cgroup_lock(void) { mutex_lock(&cgroup_mutex); @@ -548,7 +545,6 @@ void cgroup_lock(void) * * Undo the lock taken in a previous cgroup_lock() call. */ - void cgroup_unlock(void) { mutex_unlock(&cgroup_mutex); @@ -568,7 +564,7 @@ static struct inode_operations cgroup_dir_inode_operations; static struct file_operations proc_cgroupstats_operations; static struct backing_dev_info cgroup_backing_dev_info = { - .capabilities = BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK, + .capabilities = BDI_CAP_NO_ACCT_AND_WRITEBACK, }; static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) @@ -577,20 +573,40 @@ static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) if (inode) { inode->i_mode = mode; - inode->i_uid = current->fsuid; - inode->i_gid = current->fsgid; - inode->i_blocks = 0; + inode->i_uid = current_fsuid(); + inode->i_gid = current_fsgid(); inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; inode->i_mapping->backing_dev_info = &cgroup_backing_dev_info; } return inode; } +/* + * Call subsys's pre_destroy handler. + * This is called before css refcnt check. + */ +static void cgroup_call_pre_destroy(struct cgroup *cgrp) +{ + struct cgroup_subsys *ss; + for_each_subsys(cgrp->root, ss) + if (ss->pre_destroy) + ss->pre_destroy(ss, cgrp); + return; +} + +static void free_cgroup_rcu(struct rcu_head *obj) +{ + struct cgroup *cgrp = container_of(obj, struct cgroup, rcu_head); + + kfree(cgrp); +} + static void cgroup_diput(struct dentry *dentry, struct inode *inode) { /* is dentry a directory ? if so, kfree() associated cgroup */ if (S_ISDIR(inode->i_mode)) { struct cgroup *cgrp = dentry->d_fsdata; + struct cgroup_subsys *ss; BUG_ON(!(cgroup_is_removed(cgrp))); /* It's possible for external users to be holding css * reference counts on a cgroup; css_put() needs to @@ -599,7 +615,24 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode) * queue the cgroup to be handled by the release * agent */ synchronize_rcu(); - kfree(cgrp); + + mutex_lock(&cgroup_mutex); + /* + * Release the subsystem state objects. + */ + for_each_subsys(cgrp->root, ss) + ss->destroy(ss, cgrp); + + cgrp->root->number_of_cgroups--; + mutex_unlock(&cgroup_mutex); + + /* + * Drop the active superblock reference that we took when we + * created the cgroup + */ + deactivate_super(cgrp->root->sb); + + call_rcu(&cgrp->rcu_head, free_cgroup_rcu); } iput(inode); } @@ -663,7 +696,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, added_bits = final_bits & ~root->actual_subsys_bits; /* Check that any added subsystems are currently free */ for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { - unsigned long long bit = 1ull << i; + unsigned long bit = 1UL << i; struct cgroup_subsys *ss = subsys[i]; if (!(bit & added_bits)) continue; @@ -677,7 +710,7 @@ static int rebind_subsystems(struct cgroupfs_root *root, * any child cgroups exist. This is theoretically supportable * but involves complex error handling, so it's being left until * later */ - if (!list_empty(&cgrp->children)) + if (root->number_of_cgroups > 1) return -EBUSY; /* Process each subsystem */ @@ -689,23 +722,26 @@ static int rebind_subsystems(struct cgroupfs_root *root, BUG_ON(cgrp->subsys[i]); BUG_ON(!dummytop->subsys[i]); BUG_ON(dummytop->subsys[i]->cgroup != dummytop); + mutex_lock(&ss->hierarchy_mutex); cgrp->subsys[i] = dummytop->subsys[i]; cgrp->subsys[i]->cgroup = cgrp; - list_add(&ss->sibling, &root->subsys_list); - rcu_assign_pointer(ss->root, root); + list_move(&ss->sibling, &root->subsys_list); + ss->root = root; if (ss->bind) ss->bind(ss, cgrp); - + mutex_unlock(&ss->hierarchy_mutex); } else if (bit & removed_bits) { /* We're removing this subsystem */ BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); BUG_ON(cgrp->subsys[i]->cgroup != cgrp); + mutex_lock(&ss->hierarchy_mutex); if (ss->bind) ss->bind(ss, dummytop); dummytop->subsys[i]->cgroup = dummytop; cgrp->subsys[i] = NULL; - rcu_assign_pointer(subsys[i]->root, &rootnode); - list_del(&ss->sibling); + subsys[i]->root = &rootnode; + list_move(&ss->sibling, &rootnode.subsys_list); + mutex_unlock(&ss->hierarchy_mutex); } else if (bit & final_bits) { /* Subsystem state should already exist */ BUG_ON(!cgrp->subsys[i]); @@ -757,7 +793,14 @@ static int parse_cgroupfs_options(char *data, if (!*token) return -EINVAL; if (!strcmp(token, "all")) { - opts->subsys_bits = (1 << CGROUP_SUBSYS_COUNT) - 1; + /* Add all non-disabled subsystems */ + int i; + opts->subsys_bits = 0; + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + struct cgroup_subsys *ss = subsys[i]; + if (!ss->disabled) + opts->subsys_bits |= 1ul << i; + } } else if (!strcmp(token, "noprefix")) { set_bit(ROOT_NOPREFIX, &opts->flags); } else if (!strncmp(token, "release_agent=", 14)) { @@ -775,7 +818,8 @@ static int parse_cgroupfs_options(char *data, for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { ss = subsys[i]; if (!strcmp(token, ss->name)) { - set_bit(i, &opts->subsys_bits); + if (!ss->disabled) + set_bit(i, &opts->subsys_bits); break; } } @@ -835,6 +879,14 @@ static struct super_operations cgroup_ops = { .remount_fs = cgroup_remount, }; +static void init_cgroup_housekeeping(struct cgroup *cgrp) +{ + INIT_LIST_HEAD(&cgrp->sibling); + INIT_LIST_HEAD(&cgrp->children); + INIT_LIST_HEAD(&cgrp->css_sets); + INIT_LIST_HEAD(&cgrp->release_list); + init_rwsem(&cgrp->pids_mutex); +} static void init_cgroup_root(struct cgroupfs_root *root) { struct cgroup *cgrp = &root->top_cgroup; @@ -843,10 +895,7 @@ static void init_cgroup_root(struct cgroupfs_root *root) root->number_of_cgroups = 1; cgrp->root = root; cgrp->top_cgroup = cgrp; - INIT_LIST_HEAD(&cgrp->sibling); - INIT_LIST_HEAD(&cgrp->children); - INIT_LIST_HEAD(&cgrp->css_sets); - INIT_LIST_HEAD(&cgrp->release_list); + init_cgroup_housekeeping(cgrp); } static int cgroup_test_super(struct super_block *sb, void *data) @@ -894,7 +943,6 @@ static int cgroup_get_rootdir(struct super_block *sb) if (!inode) return -ENOMEM; - inode->i_op = &simple_dir_inode_operations; inode->i_fop = &simple_dir_operations; inode->i_op = &cgroup_dir_inode_operations; /* directories start off with i_nlink == 2 (for "." entry) */ @@ -916,8 +964,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, int ret = 0; struct super_block *sb; struct cgroupfs_root *root; - struct list_head tmp_cg_links, *l; - INIT_LIST_HEAD(&tmp_cg_links); + struct list_head tmp_cg_links; /* First find the desired set of subsystems */ ret = parse_cgroupfs_options(data, &opts); @@ -928,8 +975,11 @@ static int cgroup_get_sb(struct file_system_type *fs_type, } root = kzalloc(sizeof(*root), GFP_KERNEL); - if (!root) + if (!root) { + if (opts.release_agent) + kfree(opts.release_agent); return -ENOMEM; + } init_cgroup_root(root); root->subsys_bits = opts.subsys_bits; @@ -953,8 +1003,9 @@ static int cgroup_get_sb(struct file_system_type *fs_type, root = NULL; } else { /* New superblock */ - struct cgroup *cgrp = &root->top_cgroup; + struct cgroup *root_cgrp = &root->top_cgroup; struct inode *inode; + int i; BUG_ON(sb->s_root != NULL); @@ -984,7 +1035,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, if (ret == -EBUSY) { mutex_unlock(&cgroup_mutex); mutex_unlock(&inode->i_mutex); - goto drop_new_super; + goto free_cg_links; } /* EBUSY should be the only error here */ @@ -993,47 +1044,40 @@ static int cgroup_get_sb(struct file_system_type *fs_type, list_add(&root->root_list, &roots); root_count++; - sb->s_root->d_fsdata = &root->top_cgroup; + sb->s_root->d_fsdata = root_cgrp; root->top_cgroup.dentry = sb->s_root; /* Link the top cgroup in this hierarchy into all * the css_set objects */ write_lock(&css_set_lock); - l = &init_css_set.list; - do { + for (i = 0; i < CSS_SET_TABLE_SIZE; i++) { + struct hlist_head *hhead = &css_set_table[i]; + struct hlist_node *node; struct css_set *cg; - struct cg_cgroup_link *link; - cg = list_entry(l, struct css_set, list); - BUG_ON(list_empty(&tmp_cg_links)); - link = list_entry(tmp_cg_links.next, - struct cg_cgroup_link, - cgrp_link_list); - list_del(&link->cgrp_link_list); - link->cg = cg; - list_add(&link->cgrp_link_list, - &root->top_cgroup.css_sets); - list_add(&link->cg_link_list, &cg->cg_links); - l = l->next; - } while (l != &init_css_set.list); + + hlist_for_each_entry(cg, node, hhead, hlist) + link_css_set(&tmp_cg_links, cg, root_cgrp); + } write_unlock(&css_set_lock); free_cg_links(&tmp_cg_links); - BUG_ON(!list_empty(&cgrp->sibling)); - BUG_ON(!list_empty(&cgrp->children)); + BUG_ON(!list_empty(&root_cgrp->sibling)); + BUG_ON(!list_empty(&root_cgrp->children)); BUG_ON(root->number_of_cgroups != 1); - cgroup_populate_dir(cgrp); + cgroup_populate_dir(root_cgrp); mutex_unlock(&inode->i_mutex); mutex_unlock(&cgroup_mutex); } return simple_set_mnt(mnt, sb); + free_cg_links: + free_cg_links(&tmp_cg_links); drop_new_super: up_write(&sb->s_umount); deactivate_super(sb); - free_cg_links(&tmp_cg_links); return ret; } @@ -1041,6 +1085,8 @@ static void cgroup_kill_sb(struct super_block *sb) { struct cgroupfs_root *root = sb->s_fs_info; struct cgroup *cgrp = &root->top_cgroup; int ret; + struct cg_cgroup_link *link; + struct cg_cgroup_link *saved_link; BUG_ON(!root); @@ -1060,10 +1106,9 @@ static void cgroup_kill_sb(struct super_block *sb) { * root cgroup */ write_lock(&css_set_lock); - while (!list_empty(&cgrp->css_sets)) { - struct cg_cgroup_link *link; - link = list_entry(cgrp->css_sets.next, - struct cg_cgroup_link, cgrp_link_list); + + 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); kfree(link); @@ -1074,6 +1119,7 @@ static void cgroup_kill_sb(struct super_block *sb) { list_del(&root->root_list); root_count--; } + mutex_unlock(&cgroup_mutex); kfree(root); @@ -1096,15 +1142,22 @@ static inline struct cftype *__d_cft(struct dentry *dentry) return dentry->d_fsdata; } -/* - * Called with cgroup_mutex held. Writes path of cgroup into buf. - * Returns 0 on success, -errno on error. +/** + * cgroup_path - generate the path of a cgroup + * @cgrp: the cgroup in question + * @buf: the buffer to write the path into + * @buflen: the length of the buffer + * + * Called with cgroup_mutex held or else with an RCU-protected cgroup + * reference. Writes path of cgroup into buf. Returns 0 on success, + * -errno on error. */ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) { char *start; + struct dentry *dentry = rcu_dereference(cgrp->dentry); - if (cgrp == dummytop) { + if (!dentry || cgrp == dummytop) { /* * Inactive subsystems have no dentry for their root * cgroup @@ -1117,13 +1170,14 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) *--start = '\0'; for (;;) { - int len = cgrp->dentry->d_name.len; + int len = dentry->d_name.len; if ((start -= len) < buf) return -ENAMETOOLONG; memcpy(start, cgrp->dentry->d_name.name, len); cgrp = cgrp->parent; if (!cgrp) break; + dentry = rcu_dereference(cgrp->dentry); if (!cgrp->parent) continue; if (--start < buf) @@ -1155,18 +1209,20 @@ static void get_first_subsys(const struct cgroup *cgrp, *subsys_id = test_ss->subsys_id; } -/* - * Attach task 'tsk' to cgroup 'cgrp' +/** + * cgroup_attach_task - attach task 'tsk' to cgroup 'cgrp' + * @cgrp: the cgroup the task is attaching to + * @tsk: the task to be attached * - * Call holding cgroup_mutex. May take task_lock of - * the task 'pid' during call. + * Call holding cgroup_mutex. May take task_lock of + * the task 'tsk' during call. */ -static int attach_task(struct cgroup *cgrp, struct task_struct *tsk) +int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) { int retval = 0; struct cgroup_subsys *ss; struct cgroup *oldcgrp; - struct css_set *cg = tsk->cgroups; + struct css_set *cg; struct css_set *newcg; struct cgroupfs_root *root = cgrp->root; int subsys_id; @@ -1181,20 +1237,23 @@ static int attach_task(struct cgroup *cgrp, struct task_struct *tsk) for_each_subsys(root, ss) { if (ss->can_attach) { retval = ss->can_attach(ss, cgrp, tsk); - if (retval) { + if (retval) return retval; - } } } + task_lock(tsk); + cg = tsk->cgroups; + get_css_set(cg); + task_unlock(tsk); /* * Locate or allocate a new css_set for this task, * based on its final set of cgroups */ newcg = find_css_set(cg, cgrp); - if (!newcg) { + put_css_set(cg); + if (!newcg) return -ENOMEM; - } task_lock(tsk); if (tsk->flags & PF_EXITING) { @@ -1214,9 +1273,8 @@ static int attach_task(struct cgroup *cgrp, struct task_struct *tsk) write_unlock(&css_set_lock); for_each_subsys(root, ss) { - if (ss->attach) { + if (ss->attach) ss->attach(ss, cgrp, oldcgrp, tsk); - } } set_bit(CGRP_RELEASABLE, &oldcgrp->flags); synchronize_rcu(); @@ -1225,62 +1283,110 @@ static int attach_task(struct cgroup *cgrp, struct task_struct *tsk) } /* - * Attach task with pid 'pid' to cgroup 'cgrp'. Call with - * cgroup_mutex, may take task_lock of task + * Attach task with pid 'pid' to cgroup 'cgrp'. Call with cgroup_mutex + * held. May take task_lock of task */ -static int attach_task_by_pid(struct cgroup *cgrp, char *pidbuf) +static int attach_task_by_pid(struct cgroup *cgrp, u64 pid) { - pid_t pid; struct task_struct *tsk; + const struct cred *cred = current_cred(), *tcred; int ret; - if (sscanf(pidbuf, "%d", &pid) != 1) - return -EIO; - if (pid) { rcu_read_lock(); - tsk = find_task_by_pid(pid); + tsk = find_task_by_vpid(pid); if (!tsk || tsk->flags & PF_EXITING) { rcu_read_unlock(); return -ESRCH; } - get_task_struct(tsk); - rcu_read_unlock(); - if ((current->euid) && (current->euid != tsk->uid) - && (current->euid != tsk->suid)) { - put_task_struct(tsk); + tcred = __task_cred(tsk); + if (cred->euid && + cred->euid != tcred->uid && + cred->euid != tcred->suid) { + rcu_read_unlock(); return -EACCES; } + get_task_struct(tsk); + rcu_read_unlock(); } else { tsk = current; get_task_struct(tsk); } - ret = attach_task(cgrp, tsk); + ret = cgroup_attach_task(cgrp, tsk); put_task_struct(tsk); return ret; } -/* The various types of files and directories in a cgroup file system */ +static int cgroup_tasks_write(struct cgroup *cgrp, struct cftype *cft, u64 pid) +{ + int ret; + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + ret = attach_task_by_pid(cgrp, pid); + cgroup_unlock(); + return ret; +} +/* The various types of files and directories in a cgroup file system */ enum cgroup_filetype { FILE_ROOT, FILE_DIR, FILE_TASKLIST, FILE_NOTIFY_ON_RELEASE, - FILE_RELEASABLE, FILE_RELEASE_AGENT, }; -static ssize_t cgroup_write_uint(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) +/** + * cgroup_lock_live_group - take cgroup_mutex and check that cgrp is alive. + * @cgrp: the cgroup to be checked for liveness + * + * On success, returns true; the lock should be later released with + * cgroup_unlock(). On failure returns false with no lock held. + */ +bool cgroup_lock_live_group(struct cgroup *cgrp) { - char buffer[64]; + mutex_lock(&cgroup_mutex); + if (cgroup_is_removed(cgrp)) { + mutex_unlock(&cgroup_mutex); + return false; + } + return true; +} + +static int cgroup_release_agent_write(struct cgroup *cgrp, struct cftype *cft, + const char *buffer) +{ + BUILD_BUG_ON(sizeof(cgrp->root->release_agent_path) < PATH_MAX); + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + strcpy(cgrp->root->release_agent_path, buffer); + cgroup_unlock(); + return 0; +} + +static int cgroup_release_agent_show(struct cgroup *cgrp, struct cftype *cft, + struct seq_file *seq) +{ + if (!cgroup_lock_live_group(cgrp)) + return -ENODEV; + seq_puts(seq, cgrp->root->release_agent_path); + seq_putc(seq, '\n'); + cgroup_unlock(); + return 0; +} + +/* A buffer size big enough for numbers or short strings */ +#define CGROUP_LOCAL_BUFFER_SIZE 64 + +static ssize_t cgroup_write_X64(struct cgroup *cgrp, struct cftype *cft, + struct file *file, + const char __user *userbuf, + size_t nbytes, loff_t *unused_ppos) +{ + char buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; - u64 val; char *end; if (!nbytes) @@ -1291,93 +1397,56 @@ static ssize_t cgroup_write_uint(struct cgroup *cgrp, struct cftype *cft, return -EFAULT; buffer[nbytes] = 0; /* nul-terminate */ - - /* strip newline if necessary */ - if (nbytes && (buffer[nbytes-1] == '\n')) - buffer[nbytes-1] = 0; - val = simple_strtoull(buffer, &end, 0); - if (*end) - return -EINVAL; - - /* Pass to subsystem */ - retval = cft->write_uint(cgrp, cft, val); + strstrip(buffer); + if (cft->write_u64) { + u64 val = simple_strtoull(buffer, &end, 0); + if (*end) + return -EINVAL; + retval = cft->write_u64(cgrp, cft, val); + } else { + s64 val = simple_strtoll(buffer, &end, 0); + if (*end) + return -EINVAL; + retval = cft->write_s64(cgrp, cft, val); + } if (!retval) retval = nbytes; return retval; } -static ssize_t cgroup_common_file_write(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, - const char __user *userbuf, - size_t nbytes, loff_t *unused_ppos) +static ssize_t cgroup_write_string(struct cgroup *cgrp, struct cftype *cft, + struct file *file, + const char __user *userbuf, + size_t nbytes, loff_t *unused_ppos) { - enum cgroup_filetype type = cft->private; - char *buffer; + char local_buffer[CGROUP_LOCAL_BUFFER_SIZE]; int retval = 0; + size_t max_bytes = cft->max_write_len; + char *buffer = local_buffer; - if (nbytes >= PATH_MAX) + if (!max_bytes) + max_bytes = sizeof(local_buffer) - 1; + if (nbytes >= max_bytes) return -E2BIG; - - /* +1 for nul-terminator */ - buffer = kmalloc(nbytes + 1, GFP_KERNEL); - if (buffer == NULL) - return -ENOMEM; - - if (copy_from_user(buffer, userbuf, nbytes)) { - retval = -EFAULT; - goto out1; - } - buffer[nbytes] = 0; /* nul-terminate */ - - mutex_lock(&cgroup_mutex); - - if (cgroup_is_removed(cgrp)) { - retval = -ENODEV; - goto out2; - } - - switch (type) { - case FILE_TASKLIST: - retval = attach_task_by_pid(cgrp, buffer); - break; - case FILE_NOTIFY_ON_RELEASE: - clear_bit(CGRP_RELEASABLE, &cgrp->flags); - if (simple_strtoul(buffer, NULL, 10) != 0) - set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); - else - clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); - break; - case FILE_RELEASE_AGENT: - { - struct cgroupfs_root *root = cgrp->root; - /* Strip trailing newline */ - if (nbytes && (buffer[nbytes-1] == '\n')) { - buffer[nbytes-1] = 0; - } - if (nbytes < sizeof(root->release_agent_path)) { - /* We never write anything other than '\0' - * into the last char of release_agent_path, - * so it always remains a NUL-terminated - * string */ - strncpy(root->release_agent_path, buffer, nbytes); - root->release_agent_path[nbytes] = 0; - } else { - retval = -ENOSPC; - } - break; + /* Allocate a dynamic buffer if we need one */ + if (nbytes >= sizeof(local_buffer)) { + buffer = kmalloc(nbytes + 1, GFP_KERNEL); + if (buffer == NULL) + return -ENOMEM; } - default: - retval = -EINVAL; - goto out2; + if (nbytes && copy_from_user(buffer, userbuf, nbytes)) { + retval = -EFAULT; + goto out; } - if (retval == 0) + buffer[nbytes] = 0; /* nul-terminate */ + strstrip(buffer); + retval = cft->write_string(cgrp, cft, buffer); + if (!retval) retval = nbytes; -out2: - mutex_unlock(&cgroup_mutex); -out1: - kfree(buffer); +out: + if (buffer != local_buffer) + kfree(buffer); return retval; } @@ -1387,68 +1456,43 @@ 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 (!cft) + if (cgroup_is_removed(cgrp)) return -ENODEV; if (cft->write) return cft->write(cgrp, cft, file, buf, nbytes, ppos); - if (cft->write_uint) - return cgroup_write_uint(cgrp, cft, file, buf, nbytes, ppos); + if (cft->write_u64 || cft->write_s64) + return cgroup_write_X64(cgrp, cft, file, buf, nbytes, ppos); + if (cft->write_string) + return cgroup_write_string(cgrp, cft, file, buf, nbytes, ppos); + if (cft->trigger) { + int ret = cft->trigger(cgrp, (unsigned int)cft->private); + return ret ? ret : nbytes; + } return -EINVAL; } -static ssize_t cgroup_read_uint(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - char __user *buf, size_t nbytes, - loff_t *ppos) +static ssize_t cgroup_read_u64(struct cgroup *cgrp, struct cftype *cft, + struct file *file, + char __user *buf, size_t nbytes, + loff_t *ppos) { - char tmp[64]; - u64 val = cft->read_uint(cgrp, cft); + char tmp[CGROUP_LOCAL_BUFFER_SIZE]; + u64 val = cft->read_u64(cgrp, cft); int len = sprintf(tmp, "%llu\n", (unsigned long long) val); return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); } -static ssize_t cgroup_common_file_read(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, - char __user *buf, - size_t nbytes, loff_t *ppos) +static ssize_t cgroup_read_s64(struct cgroup *cgrp, struct cftype *cft, + struct file *file, + char __user *buf, size_t nbytes, + loff_t *ppos) { - enum cgroup_filetype type = cft->private; - char *page; - ssize_t retval = 0; - char *s; - - if (!(page = (char *)__get_free_page(GFP_KERNEL))) - return -ENOMEM; + char tmp[CGROUP_LOCAL_BUFFER_SIZE]; + s64 val = cft->read_s64(cgrp, cft); + int len = sprintf(tmp, "%lld\n", (long long) val); - s = page; - - switch (type) { - case FILE_RELEASE_AGENT: - { - struct cgroupfs_root *root; - size_t n; - mutex_lock(&cgroup_mutex); - root = cgrp->root; - n = strnlen(root->release_agent_path, - sizeof(root->release_agent_path)); - n = min(n, (size_t) PAGE_SIZE); - strncpy(s, root->release_agent_path, n); - mutex_unlock(&cgroup_mutex); - s += n; - break; - } - default: - retval = -EINVAL; - goto out; - } - *s++ = '\n'; - - retval = simple_read_from_buffer(buf, nbytes, ppos, page, s - page); -out: - free_page((unsigned long)page); - return retval; + return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); } static ssize_t cgroup_file_read(struct file *file, char __user *buf, @@ -1457,16 +1501,62 @@ 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 (!cft) + if (cgroup_is_removed(cgrp)) return -ENODEV; if (cft->read) return cft->read(cgrp, cft, file, buf, nbytes, ppos); - if (cft->read_uint) - return cgroup_read_uint(cgrp, cft, file, buf, nbytes, ppos); + if (cft->read_u64) + return cgroup_read_u64(cgrp, cft, file, buf, nbytes, ppos); + if (cft->read_s64) + return cgroup_read_s64(cgrp, cft, file, buf, nbytes, ppos); return -EINVAL; } +/* + * seqfile ops/methods for returning structured data. Currently just + * supports string->u64 maps, but can be extended in future. + */ + +struct cgroup_seqfile_state { + struct cftype *cft; + struct cgroup *cgroup; +}; + +static int cgroup_map_add(struct cgroup_map_cb *cb, const char *key, u64 value) +{ + struct seq_file *sf = cb->state; + return seq_printf(sf, "%s %llu\n", key, (unsigned long long)value); +} + +static int cgroup_seqfile_show(struct seq_file *m, void *arg) +{ + struct cgroup_seqfile_state *state = m->private; + struct cftype *cft = state->cft; + if (cft->read_map) { + struct cgroup_map_cb cb = { + .fill = cgroup_map_add, + .state = m, + }; + return cft->read_map(state->cgroup, cft, &cb); + } + return cft->read_seq_string(state->cgroup, cft, m); +} + +static int cgroup_seqfile_release(struct inode *inode, struct file *file) +{ + struct seq_file *seq = file->private_data; + kfree(seq->private); + return single_release(inode, file); +} + +static struct file_operations cgroup_seqfile_operations = { + .read = seq_read, + .write = cgroup_file_write, + .llseek = seq_lseek, + .release = cgroup_seqfile_release, +}; + static int cgroup_file_open(struct inode *inode, struct file *file) { int err; @@ -1475,11 +1565,20 @@ static int cgroup_file_open(struct inode *inode, struct file *file) err = generic_file_open(inode, file); if (err) return err; - cft = __d_cft(file->f_dentry); - if (!cft) - return -ENODEV; - if (cft->open) + + if (cft->read_map || cft->read_seq_string) { + struct cgroup_seqfile_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; + err = single_open(file, cgroup_seqfile_show, state); + if (err < 0) + kfree(state); + } else if (cft->open) err = cft->open(inode, file); else err = 0; @@ -1564,12 +1663,11 @@ static int cgroup_create_file(struct dentry *dentry, int mode, } /* - * cgroup_create_dir - create a directory for an object. - * cgrp: the cgroup we create the directory for. - * It must have a valid ->parent field - * And we are going to fill its ->dentry field. - * dentry: dentry of the new cgroup - * mode: mode to set on new directory. + * cgroup_create_dir - create a directory for an object. + * @cgrp: the cgroup we create the directory for. It must have a valid + * ->parent field. And we are going to fill its ->dentry field. + * @dentry: dentry of the new cgroup + * @mode: mode to set on new directory. */ static int cgroup_create_dir(struct cgroup *cgrp, struct dentry *dentry, int mode) @@ -1582,7 +1680,7 @@ static int cgroup_create_dir(struct cgroup *cgrp, struct dentry *dentry, if (!error) { dentry->d_fsdata = cgrp; inc_nlink(parent->d_inode); - cgrp->dentry = dentry; + rcu_assign_pointer(cgrp->dentry, dentry); dget(dentry); } dput(dentry); @@ -1631,20 +1729,20 @@ int cgroup_add_files(struct cgroup *cgrp, return 0; } -/* Count the number of tasks in a cgroup. */ - +/** + * cgroup_task_count - count the number of tasks in a cgroup. + * @cgrp: the cgroup in question + * + * Return the number of tasks in the cgroup. + */ int cgroup_task_count(const struct cgroup *cgrp) { int count = 0; - struct list_head *l; + struct cg_cgroup_link *link; read_lock(&css_set_lock); - l = cgrp->css_sets.next; - while (l != &cgrp->css_sets) { - struct cg_cgroup_link *link = - list_entry(l, struct cg_cgroup_link, cgrp_link_list); - count += atomic_read(&link->cg->ref.refcount); - l = l->next; + list_for_each_entry(link, &cgrp->css_sets, cgrp_link_list) { + count += atomic_read(&link->cg->refcount); } read_unlock(&css_set_lock); return count; @@ -1675,6 +1773,34 @@ static void cgroup_advance_iter(struct cgroup *cgrp, it->task = cg->tasks.next; } +/* + * To reduce the fork() overhead for systems that are not actually + * using their cgroups capability, we don't maintain the lists running + * through each css_set to its tasks until we see the list actually + * used - in other words after the first call to cgroup_iter_start(). + * + * The tasklist_lock is not held here, as do_each_thread() and + * while_each_thread() are protected by RCU. + */ +static void cgroup_enable_task_cg_lists(void) +{ + struct task_struct *p, *g; + write_lock(&css_set_lock); + use_task_css_set_links = 1; + do_each_thread(g, p) { + task_lock(p); + /* + * We should check if the process is exiting, otherwise + * it will race with cgroup_exit() in that the list + * 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); + task_unlock(p); + } while_each_thread(g, p); + write_unlock(&css_set_lock); +} + void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) { /* @@ -1682,18 +1808,9 @@ void cgroup_iter_start(struct cgroup *cgrp, struct cgroup_iter *it) * we need to enable the list linking each css_set to its * tasks, and fix up all existing tasks. */ - if (!use_task_css_set_links) { - struct task_struct *p, *g; - write_lock(&css_set_lock); - use_task_css_set_links = 1; - do_each_thread(g, p) { - task_lock(p); - if (list_empty(&p->cg_list)) - list_add(&p->cg_list, &p->cgroups->tasks); - task_unlock(p); - } while_each_thread(g, p); - write_unlock(&css_set_lock); - } + if (!use_task_css_set_links) + cgroup_enable_task_cg_lists(); + read_lock(&css_set_lock); it->cg_link = &cgrp->css_sets; cgroup_advance_iter(cgrp, it); @@ -1704,6 +1821,7 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, { struct task_struct *res; struct list_head *l = it->task; + struct cg_cgroup_link *link; /* If the iterator cg is NULL, we have no tasks */ if (!it->cg_link) @@ -1711,7 +1829,8 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, res = list_entry(l, struct task_struct, cg_list); /* Advance iterator to find next entry */ l = l->next; - if (l == &res->cgroups->tasks) { + link = list_entry(it->cg_link, struct cg_cgroup_link, cgrp_link_list); + if (l == &link->cg->tasks) { /* We reached the end of this task list - move on to * the next cg_cgroup_link */ cgroup_advance_iter(cgrp, it); @@ -1726,6 +1845,166 @@ void cgroup_iter_end(struct cgroup *cgrp, struct cgroup_iter *it) read_unlock(&css_set_lock); } +static inline int started_after_time(struct task_struct *t1, + struct timespec *time, + struct task_struct *t2) +{ + int start_diff = timespec_compare(&t1->start_time, time); + if (start_diff > 0) { + return 1; + } else if (start_diff < 0) { + return 0; + } else { + /* + * Arbitrarily, if two processes started at the same + * time, we'll say that the lower pointer value + * started first. Note that t2 may have exited by now + * so this may not be a valid pointer any longer, but + * that's fine - it still serves to distinguish + * between two tasks started (effectively) simultaneously. + */ + return t1 > t2; + } +} + +/* + * This function is a callback from heap_insert() and is used to order + * the heap. + * In this case we order the heap in descending task start time. + */ +static inline int started_after(void *p1, void *p2) +{ + struct task_struct *t1 = p1; + struct task_struct *t2 = p2; + return started_after_time(t1, &t2->start_time, t2); +} + +/** + * cgroup_scan_tasks - iterate though all the tasks in a cgroup + * @scan: struct cgroup_scanner containing arguments for the scan + * + * Arguments include pointers to callback functions test_task() and + * process_task(). + * Iterate through all the tasks in a cgroup, calling test_task() for each, + * and if it returns true, call process_task() for it also. + * The test_task pointer may be NULL, meaning always true (select all tasks). + * Effectively duplicates cgroup_iter_{start,next,end}() + * but does not lock css_set_lock for the call to process_task(). + * The struct cgroup_scanner may be embedded in any structure of the caller's + * creation. + * It is guaranteed that process_task() will act on every task that + * is a member of the cgroup for the duration of this call. This + * function may or may not call process_task() for tasks that exit + * or move to a different cgroup during the call, or are forked or + * move into the cgroup during the call. + * + * Note that test_task() may be called with locks held, and may in some + * situations be called multiple times for the same task, so it should + * be cheap. + * If the heap pointer in the struct cgroup_scanner is non-NULL, a heap has been + * pre-allocated and will be used for heap operations (and its "gt" member will + * be overwritten), else a temporary heap will be used (allocation of which + * may cause this function to fail). + */ +int cgroup_scan_tasks(struct cgroup_scanner *scan) +{ + int retval, i; + struct cgroup_iter it; + struct task_struct *p, *dropped; + /* Never dereference latest_task, since it's not refcounted */ + struct task_struct *latest_task = NULL; + struct ptr_heap tmp_heap; + struct ptr_heap *heap; + struct timespec latest_time = { 0, 0 }; + + if (scan->heap) { + /* The caller supplied our heap and pre-allocated its memory */ + heap = scan->heap; + heap->gt = &started_after; + } else { + /* We need to allocate our own heap memory */ + heap = &tmp_heap; + retval = heap_init(heap, PAGE_SIZE, GFP_KERNEL, &started_after); + if (retval) + /* cannot allocate the heap */ + return retval; + } + + again: + /* + * Scan tasks in the cgroup, using the scanner's "test_task" callback + * to determine which are of interest, and using the scanner's + * "process_task" callback to process any of them that need an update. + * Since we don't want to hold any locks during the task updates, + * gather tasks to be processed in a heap structure. + * The heap is sorted by descending task start time. + * If the statically-sized heap fills up, we overflow tasks that + * started later, and in future iterations only consider tasks that + * started after the latest task in the previous pass. This + * guarantees forward progress and that we don't miss any tasks. + */ + heap->size = 0; + cgroup_iter_start(scan->cg, &it); + while ((p = cgroup_iter_next(scan->cg, &it))) { + /* + * Only affect tasks that qualify per the caller's callback, + * if he provided one + */ + if (scan->test_task && !scan->test_task(p, scan)) + continue; + /* + * Only process tasks that started after the last task + * we processed + */ + if (!started_after_time(p, &latest_time, latest_task)) + continue; + dropped = heap_insert(heap, p); + if (dropped == NULL) { + /* + * The new task was inserted; the heap wasn't + * previously full + */ + get_task_struct(p); + } else if (dropped != p) { + /* + * The new task was inserted, and pushed out a + * different task + */ + get_task_struct(p); + put_task_struct(dropped); + } + /* + * Else the new task was newer than anything already in + * the heap and wasn't inserted + */ + } + cgroup_iter_end(scan->cg, &it); + + if (heap->size) { + for (i = 0; i < heap->size; i++) { + struct task_struct *q = heap->ptrs[i]; + if (i == 0) { + latest_time = q->start_time; + latest_task = q; + } + /* Process the task per the caller's callback */ + scan->process_task(q, scan); + put_task_struct(q); + } + /* + * If we had to process any tasks at all, scan again + * in case some of them were in the middle of forking + * children that didn't get processed. + * Not the most efficient way to do it, but it avoids + * having to take callback_mutex in the fork path + */ + goto again; + } + if (heap == &tmp_heap) + heap_free(&tmp_heap); + return 0; +} + /* * Stuff for reading the 'tasks' file. * @@ -1734,16 +2013,7 @@ void cgroup_iter_end(struct cgroup *cgrp, struct cgroup_iter *it) * but we cannot guarantee that the information we produce is correct * unless we produce it entirely atomically. * - * Upon tasks file open(), a struct ctr_struct is allocated, that - * will have a pointer to an array (also allocated here). The struct - * ctr_struct * is stored in file->private_data. Its resources will - * be freed by release() when the file is closed. The array is used - * to sprintf the PIDs and then used by read(). */ -struct ctr_struct { - char *buf; - int bufsz; -}; /* * Load into 'pidarray' up to 'npids' of the tasks using cgroup @@ -1754,26 +2024,29 @@ struct ctr_struct { */ static int pid_array_load(pid_t *pidarray, int npids, struct cgroup *cgrp) { - int n = 0; + int n = 0, pid; struct cgroup_iter it; struct task_struct *tsk; cgroup_iter_start(cgrp, &it); while ((tsk = cgroup_iter_next(cgrp, &it))) { if (unlikely(n == npids)) break; - pidarray[n++] = task_pid_nr(tsk); + pid = task_pid_vnr(tsk); + if (pid > 0) + pidarray[n++] = pid; } cgroup_iter_end(cgrp, &it); return n; } /** - * Build and fill cgroupstats so that taskstats can export it to user - * space. - * + * cgroupstats_build - build and fill cgroupstats * @stats: cgroupstats to fill information into * @dentry: A dentry entry belonging to the cgroup for which stats have * been requested. + * + * Build and fill cgroupstats so that taskstats can export it to user + * space. */ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) { @@ -1781,15 +2054,17 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) struct cgroup *cgrp; struct cgroup_iter it; struct task_struct *tsk; + /* - * Validate dentry by checking the superblock operations + * Validate dentry by checking the superblock operations, + * and make sure it's a directory. */ - if (dentry->d_sb->s_op != &cgroup_ops) + if (dentry->d_sb->s_op != &cgroup_ops || + !S_ISDIR(dentry->d_inode->i_mode)) goto err; ret = 0; cgrp = dentry->d_fsdata; - rcu_read_lock(); cgroup_iter_start(cgrp, &it); while ((tsk = cgroup_iter_next(cgrp, &it))) { @@ -1814,7 +2089,6 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) } cgroup_iter_end(cgrp, &it); - rcu_read_unlock(); err: return ret; } @@ -1824,42 +2098,132 @@ static int cmppid(const void *a, const void *b) return *(pid_t *)a - *(pid_t *)b; } + /* - * Convert array 'a' of 'npids' pid_t's to a string of newline separated - * decimal pids in 'buf'. Don't write more than 'sz' chars, but return - * count 'cnt' of how many chars would be written if buf were large enough. + * seq_file methods for the "tasks" file. The seq_file position is the + * next pid to display; the seq_file iterator is a pointer to the pid + * in the cgroup->tasks_pids array. */ -static int pid_array_to_buf(char *buf, int sz, pid_t *a, int npids) + +static void *cgroup_tasks_start(struct seq_file *s, loff_t *pos) { - int cnt = 0; - int i; + /* + * Initially we receive a position value that corresponds to + * one more than the last pid shown (or 0 on the first call or + * after a seek to the start). Use a binary-search to find the + * next pid to display, if any + */ + struct cgroup *cgrp = s->private; + int index = 0, pid = *pos; + int *iter; + + down_read(&cgrp->pids_mutex); + if (pid) { + int end = cgrp->pids_length; + + while (index < end) { + int mid = (index + end) / 2; + if (cgrp->tasks_pids[mid] == pid) { + index = mid; + break; + } else if (cgrp->tasks_pids[mid] <= pid) + index = mid + 1; + else + end = mid; + } + } + /* If we're off the end of the array, we're done */ + if (index >= cgrp->pids_length) + return NULL; + /* Update the abstract position to be the actual pid that we found */ + iter = cgrp->tasks_pids + index; + *pos = *iter; + return iter; +} + +static void cgroup_tasks_stop(struct seq_file *s, void *v) +{ + struct cgroup *cgrp = s->private; + up_read(&cgrp->pids_mutex); +} + +static void *cgroup_tasks_next(struct seq_file *s, void *v, loff_t *pos) +{ + struct cgroup *cgrp = s->private; + int *p = v; + int *end = cgrp->tasks_pids + cgrp->pids_length; + + /* + * Advance to the next pid in the array. If this goes off the + * end, we're done + */ + p++; + if (p >= end) { + return NULL; + } else { + *pos = *p; + return p; + } +} + +static int cgroup_tasks_show(struct seq_file *s, void *v) +{ + return seq_printf(s, "%d\n", *(int *)v); +} - for (i = 0; i < npids; i++) - cnt += snprintf(buf + cnt, max(sz - cnt, 0), "%d\n", a[i]); - return cnt; +static struct seq_operations cgroup_tasks_seq_operations = { + .start = cgroup_tasks_start, + .stop = cgroup_tasks_stop, + .next = cgroup_tasks_next, + .show = cgroup_tasks_show, +}; + +static void release_cgroup_pid_array(struct cgroup *cgrp) +{ + down_write(&cgrp->pids_mutex); + BUG_ON(!cgrp->pids_use_count); + if (!--cgrp->pids_use_count) { + kfree(cgrp->tasks_pids); + cgrp->tasks_pids = NULL; + cgrp->pids_length = 0; + } + up_write(&cgrp->pids_mutex); +} + +static int cgroup_tasks_release(struct inode *inode, struct file *file) +{ + struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); + + if (!(file->f_mode & FMODE_READ)) + return 0; + + release_cgroup_pid_array(cgrp); + return seq_release(inode, file); } +static struct file_operations cgroup_tasks_operations = { + .read = seq_read, + .llseek = seq_lseek, + .write = cgroup_file_write, + .release = cgroup_tasks_release, +}; + /* - * Handle an open on 'tasks' file. Prepare a buffer listing the + * Handle an open on 'tasks' file. Prepare an array containing the * process id's of tasks currently attached to the cgroup being opened. - * - * Does not require any specific cgroup mutexes, and does not take any. */ + static int cgroup_tasks_open(struct inode *unused, struct file *file) { struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); - struct ctr_struct *ctr; pid_t *pidarray; int npids; - char c; + int retval; + /* Nothing to do for write-only files */ if (!(file->f_mode & FMODE_READ)) return 0; - ctr = kmalloc(sizeof(*ctr), GFP_KERNEL); - if (!ctr) - goto err0; - /* * If cgroup gets more users after we read count, we won't have * enough space - tough. This race is indistinguishable to the @@ -1867,57 +2231,31 @@ static int cgroup_tasks_open(struct inode *unused, struct file *file) * show up until sometime later on. */ npids = cgroup_task_count(cgrp); - if (npids) { - pidarray = kmalloc(npids * sizeof(pid_t), GFP_KERNEL); - if (!pidarray) - goto err1; - - npids = pid_array_load(pidarray, npids, cgrp); - sort(pidarray, npids, sizeof(pid_t), cmppid, NULL); - - /* Call pid_array_to_buf() twice, first just to get bufsz */ - ctr->bufsz = pid_array_to_buf(&c, sizeof(c), pidarray, npids) + 1; - ctr->buf = kmalloc(ctr->bufsz, GFP_KERNEL); - if (!ctr->buf) - goto err2; - ctr->bufsz = pid_array_to_buf(ctr->buf, ctr->bufsz, pidarray, npids); - - kfree(pidarray); - } else { - ctr->buf = 0; - ctr->bufsz = 0; - } - file->private_data = ctr; - return 0; - -err2: - kfree(pidarray); -err1: - kfree(ctr); -err0: - return -ENOMEM; -} - -static ssize_t cgroup_tasks_read(struct cgroup *cgrp, - struct cftype *cft, - struct file *file, char __user *buf, - size_t nbytes, loff_t *ppos) -{ - struct ctr_struct *ctr = file->private_data; + pidarray = kmalloc(npids * sizeof(pid_t), GFP_KERNEL); + if (!pidarray) + return -ENOMEM; + npids = pid_array_load(pidarray, npids, cgrp); + sort(pidarray, npids, sizeof(pid_t), cmppid, NULL); - return simple_read_from_buffer(buf, nbytes, ppos, ctr->buf, ctr->bufsz); -} + /* + * Store the array in the cgroup, freeing the old + * array if necessary + */ + down_write(&cgrp->pids_mutex); + kfree(cgrp->tasks_pids); + cgrp->tasks_pids = pidarray; + cgrp->pids_length = npids; + cgrp->pids_use_count++; + up_write(&cgrp->pids_mutex); -static int cgroup_tasks_release(struct inode *unused_inode, - struct file *file) -{ - struct ctr_struct *ctr; + file->f_op = &cgroup_tasks_operations; - if (file->f_mode & FMODE_READ) { - ctr = file->private_data; - kfree(ctr->buf); - kfree(ctr); + retval = seq_open(file, &cgroup_tasks_seq_operations); + if (retval) { + release_cgroup_pid_array(cgrp); + return retval; } + ((struct seq_file *)file->private_data)->private = cgrp; return 0; } @@ -1927,9 +2265,16 @@ static u64 cgroup_read_notify_on_release(struct cgroup *cgrp, return notify_on_release(cgrp); } -static u64 cgroup_read_releasable(struct cgroup *cgrp, struct cftype *cft) +static int cgroup_write_notify_on_release(struct cgroup *cgrp, + struct cftype *cft, + u64 val) { - return test_bit(CGRP_RELEASABLE, &cgrp->flags); + clear_bit(CGRP_RELEASABLE, &cgrp->flags); + if (val) + set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + else + clear_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + return 0; } /* @@ -1939,30 +2284,24 @@ static struct cftype files[] = { { .name = "tasks", .open = cgroup_tasks_open, - .read = cgroup_tasks_read, - .write = cgroup_common_file_write, + .write_u64 = cgroup_tasks_write, .release = cgroup_tasks_release, .private = FILE_TASKLIST, }, { .name = "notify_on_release", - .read_uint = cgroup_read_notify_on_release, - .write = cgroup_common_file_write, + .read_u64 = cgroup_read_notify_on_release, + .write_u64 = cgroup_write_notify_on_release, .private = FILE_NOTIFY_ON_RELEASE, }, - - { - .name = "releasable", - .read_uint = cgroup_read_releasable, - .private = FILE_RELEASABLE, - } }; static struct cftype cft_release_agent = { .name = "release_agent", - .read = cgroup_common_file_read, - .write = cgroup_common_file_write, + .read_seq_string = cgroup_release_agent_show, + .write_string = cgroup_release_agent_write, + .max_write_len = PATH_MAX, .private = FILE_RELEASE_AGENT, }; @@ -1996,7 +2335,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, struct cgroup *cgrp) { css->cgroup = cgrp; - atomic_set(&css->refcnt, 0); + atomic_set(&css->refcnt, 1); css->flags = 0; if (cgrp == dummytop) set_bit(CSS_ROOT, &css->flags); @@ -2004,15 +2343,37 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, cgrp->subsys[ss->subsys_id] = css; } +static void cgroup_lock_hierarchy(struct cgroupfs_root *root) +{ + /* We need to take each hierarchy_mutex in a consistent order */ + int i; + + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + struct cgroup_subsys *ss = subsys[i]; + if (ss->root == root) + mutex_lock_nested(&ss->hierarchy_mutex, i); + } +} + +static void cgroup_unlock_hierarchy(struct cgroupfs_root *root) +{ + int i; + + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + struct cgroup_subsys *ss = subsys[i]; + if (ss->root == root) + mutex_unlock(&ss->hierarchy_mutex); + } +} + /* - * cgroup_create - create a cgroup - * parent: cgroup that will be parent of the new cgroup. - * name: name of the new cgroup. Will be strcpy'ed. - * mode: mode to set on new inode + * cgroup_create - create a cgroup + * @parent: cgroup that will be parent of the new cgroup + * @dentry: dentry of the new cgroup + * @mode: mode to set on new inode * - * Must be called with the mutex on the parent inode held + * Must be called with the mutex on the parent inode held */ - static long cgroup_create(struct cgroup *parent, struct dentry *dentry, int mode) { @@ -2035,16 +2396,15 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, mutex_lock(&cgroup_mutex); - cgrp->flags = 0; - INIT_LIST_HEAD(&cgrp->sibling); - INIT_LIST_HEAD(&cgrp->children); - INIT_LIST_HEAD(&cgrp->css_sets); - INIT_LIST_HEAD(&cgrp->release_list); + init_cgroup_housekeeping(cgrp); cgrp->parent = parent; cgrp->root = parent->root; cgrp->top_cgroup = parent->top_cgroup; + if (notify_on_release(parent)) + set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); + for_each_subsys(root, ss) { struct cgroup_subsys_state *css = ss->create(ss, cgrp); if (IS_ERR(css)) { @@ -2054,7 +2414,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, init_cgroup_css(css, ss, cgrp); } + cgroup_lock_hierarchy(root); list_add(&cgrp->sibling, &cgrp->parent->children); + cgroup_unlock_hierarchy(root); root->number_of_cgroups++; err = cgroup_create_dir(cgrp, dentry, mode); @@ -2074,7 +2436,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err_remove: + cgroup_lock_hierarchy(root); list_del(&cgrp->sibling); + cgroup_unlock_hierarchy(root); root->number_of_cgroups--; err_destroy: @@ -2101,11 +2465,11 @@ static int cgroup_mkdir(struct inode *dir, struct dentry *dentry, int mode) return cgroup_create(c_parent, dentry, mode | S_IFDIR); } -static inline int cgroup_has_css_refs(struct cgroup *cgrp) +static int cgroup_has_css_refs(struct cgroup *cgrp) { /* Check the reference count on each subsystem. Since we * already established that there are no tasks in the - * cgroup, if the css refcount is also 0, then there should + * cgroup, if the css refcount is also 1, then there should * be no outstanding references, so the subsystem is safe to * destroy. We scan across all subsystems rather than using * the per-hierarchy linked list of mounted subsystems since @@ -2126,21 +2490,70 @@ static inline int cgroup_has_css_refs(struct cgroup *cgrp) * matter, since it can only happen if the cgroup * has been deleted and hence no longer needs the * release agent to be called anyway. */ - if (css && atomic_read(&css->refcnt)) { + if (css && (atomic_read(&css->refcnt) > 1)) return 1; - } } return 0; } +/* + * Atomically mark all (or else none) of the cgroup's CSS objects as + * CSS_REMOVED. Return true on success, or false if the cgroup has + * busy subsystems. Call with cgroup_mutex held + */ + +static int cgroup_clear_css_refs(struct cgroup *cgrp) +{ + struct cgroup_subsys *ss; + unsigned long flags; + bool failed = false; + local_irq_save(flags); + for_each_subsys(cgrp->root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + int refcnt; + while (1) { + /* We can only remove a CSS with a refcnt==1 */ + refcnt = atomic_read(&css->refcnt); + if (refcnt > 1) { + failed = true; + goto done; + } + BUG_ON(!refcnt); + /* + * Drop the refcnt to 0 while we check other + * subsystems. This will cause any racing + * css_tryget() to spin until we set the + * CSS_REMOVED bits or abort + */ + if (atomic_cmpxchg(&css->refcnt, refcnt, 0) == refcnt) + break; + cpu_relax(); + } + } + done: + for_each_subsys(cgrp->root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + if (failed) { + /* + * Restore old refcnt if we previously managed + * to clear it from 1 to 0 + */ + if (!atomic_read(&css->refcnt)) + atomic_set(&css->refcnt, 1); + } else { + /* Commit the fact that the CSS is removed */ + set_bit(CSS_REMOVED, &css->flags); + } + } + local_irq_restore(flags); + return !failed; +} + static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) { struct cgroup *cgrp = dentry->d_fsdata; struct dentry *d; struct cgroup *parent; - struct cgroup_subsys *ss; - struct super_block *sb; - struct cgroupfs_root *root; /* the vfs holds both inode->i_mutex already */ @@ -2153,108 +2566,95 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) mutex_unlock(&cgroup_mutex); return -EBUSY; } + mutex_unlock(&cgroup_mutex); + + /* + * Call pre_destroy handlers of subsys. Notify subsystems + * that rmdir() request comes. + */ + cgroup_call_pre_destroy(cgrp); + mutex_lock(&cgroup_mutex); parent = cgrp->parent; - root = cgrp->root; - sb = root->sb; - if (cgroup_has_css_refs(cgrp)) { + if (atomic_read(&cgrp->count) + || !list_empty(&cgrp->children) + || !cgroup_clear_css_refs(cgrp)) { mutex_unlock(&cgroup_mutex); return -EBUSY; } - for_each_subsys(root, ss) { - if (cgrp->subsys[ss->subsys_id]) - ss->destroy(ss, cgrp); - } - spin_lock(&release_list_lock); set_bit(CGRP_REMOVED, &cgrp->flags); if (!list_empty(&cgrp->release_list)) list_del(&cgrp->release_list); spin_unlock(&release_list_lock); - /* delete my sibling from parent->children */ + + cgroup_lock_hierarchy(cgrp->root); + /* delete this cgroup from parent->children */ list_del(&cgrp->sibling); + cgroup_unlock_hierarchy(cgrp->root); + spin_lock(&cgrp->dentry->d_lock); d = dget(cgrp->dentry); - cgrp->dentry = NULL; spin_unlock(&d->d_lock); cgroup_d_remove_dir(d); dput(d); - root->number_of_cgroups--; set_bit(CGRP_RELEASABLE, &parent->flags); check_for_release(parent); mutex_unlock(&cgroup_mutex); - /* Drop the active superblock reference that we took when we - * created the cgroup */ - deactivate_super(sb); return 0; } -static void cgroup_init_subsys(struct cgroup_subsys *ss) +static void __init cgroup_init_subsys(struct cgroup_subsys *ss) { struct cgroup_subsys_state *css; - struct list_head *l; printk(KERN_INFO "Initializing cgroup subsys %s\n", ss->name); /* Create the top cgroup state for this subsystem */ + list_add(&ss->sibling, &rootnode.subsys_list); ss->root = &rootnode; css = ss->create(ss, dummytop); /* We don't handle early failures gracefully */ BUG_ON(IS_ERR(css)); init_cgroup_css(css, ss, dummytop); - /* Update all cgroup groups to contain a subsys + /* Update the init_css_set to contain a subsys * pointer to this state - since the subsystem is - * newly registered, all tasks and hence all cgroup - * groups are in the subsystem's top cgroup. */ - write_lock(&css_set_lock); - l = &init_css_set.list; - do { - struct css_set *cg = - list_entry(l, struct css_set, list); - cg->subsys[ss->subsys_id] = dummytop->subsys[ss->subsys_id]; - l = l->next; - } while (l != &init_css_set.list); - write_unlock(&css_set_lock); - - /* If this subsystem requested that it be notified with fork - * events, we should send it one now for every process in the - * system */ - if (ss->fork) { - struct task_struct *g, *p; - - read_lock(&tasklist_lock); - do_each_thread(g, p) { - ss->fork(ss, p); - } while_each_thread(g, p); - read_unlock(&tasklist_lock); - } + * newly registered, all tasks and hence the + * init_css_set is in the subsystem's top cgroup. */ + init_css_set.subsys[ss->subsys_id] = dummytop->subsys[ss->subsys_id]; need_forkexit_callback |= ss->fork || ss->exit; + /* At system boot, before all subsystems have been + * registered, no tasks have been forked, so we don't + * need to invoke fork callbacks here. */ + BUG_ON(!list_empty(&init_task.tasks)); + + mutex_init(&ss->hierarchy_mutex); ss->active = 1; } /** - * cgroup_init_early - initialize cgroups at system boot, and - * initialize any subsystems that request early init. + * cgroup_init_early - cgroup initialization at system boot + * + * Initialize cgroups at system boot, and initialize any + * subsystems that request early init. */ int __init cgroup_init_early(void) { int i; - kref_init(&init_css_set.ref); - kref_get(&init_css_set.ref); - INIT_LIST_HEAD(&init_css_set.list); + atomic_set(&init_css_set.refcount, 1); INIT_LIST_HEAD(&init_css_set.cg_links); INIT_LIST_HEAD(&init_css_set.tasks); + INIT_HLIST_NODE(&init_css_set.hlist); css_set_count = 1; init_cgroup_root(&rootnode); - list_add(&rootnode.root_list, &roots); root_count = 1; init_task.cgroups = &init_css_set; @@ -2264,6 +2664,9 @@ int __init cgroup_init_early(void) list_add(&init_css_set_link.cg_link_list, &init_css_set.cg_links); + for (i = 0; i < CSS_SET_TABLE_SIZE; i++) + INIT_HLIST_HEAD(&css_set_table[i]); + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; @@ -2284,14 +2687,16 @@ int __init cgroup_init_early(void) } /** - * cgroup_init - register cgroup filesystem and /proc file, and - * initialize any subsystems that didn't request early init. + * cgroup_init - cgroup initialization + * + * Register cgroup filesystem and /proc file, and initialize + * any subsystems that didn't request early init. */ int __init cgroup_init(void) { int err; int i; - struct proc_dir_entry *entry; + struct hlist_head *hhead; err = bdi_init(&cgroup_backing_dev_info); if (err) @@ -2303,13 +2708,15 @@ int __init cgroup_init(void) cgroup_init_subsys(ss); } + /* Add init_css_set to the hash table */ + hhead = css_set_hash(init_css_set.subsys); + hlist_add_head(&init_css_set.hlist, hhead); + err = register_filesystem(&cgroup_fs_type); if (err < 0) goto out; - entry = create_proc_entry("cgroups", 0, NULL); - if (entry) - entry->proc_fops = &proc_cgroupstats_operations; + proc_create("cgroups", 0, NULL, &proc_cgroupstats_operations); out: if (err) @@ -2324,7 +2731,7 @@ out: * - Used for /proc//cgroup. * - No need to task_lock(tsk) on this tsk->cgroup reference, as it * doesn't really matter if tsk->cgroup changes after we read it, - * and we take cgroup_mutex, keeping attach_task() from changing it + * and we take cgroup_mutex, keeping cgroup_attach_task() from changing it * anyway. No need to check that tsk->cgroup != NULL, thanks to * the_top_cgroup_hack in cgroup_exit(), which sets an exiting tasks * cgroup to top_cgroup. @@ -2354,15 +2761,13 @@ static int proc_cgroup_show(struct seq_file *m, void *v) mutex_lock(&cgroup_mutex); - for_each_root(root) { + for_each_active_root(root) { struct cgroup_subsys *ss; struct cgroup *cgrp; int subsys_id; int count = 0; - /* Skip this hierarchy if it has no active subsystems */ - if (!root->actual_subsys_bits) - continue; + seq_printf(m, "%lu:", root->subsys_bits); for_each_subsys(root, ss) seq_printf(m, "%s%s", count++ ? "," : "", ss->name); seq_putc(m, ':'); @@ -2402,13 +2807,13 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) { int i; - seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\n"); + seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\tenabled\n"); mutex_lock(&cgroup_mutex); for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; - seq_printf(m, "%s\t%lu\t%d\n", + seq_printf(m, "%s\t%lu\t%d\t%d\n", ss->name, ss->root->subsys_bits, - ss->root->number_of_cgroups); + ss->root->number_of_cgroups, !ss->disabled); } mutex_unlock(&cgroup_mutex); return 0; @@ -2416,7 +2821,7 @@ static int proc_cgroupstats_show(struct seq_file *m, void *v) static int cgroupstats_open(struct inode *inode, struct file *file) { - return single_open(file, proc_cgroupstats_show, 0); + return single_open(file, proc_cgroupstats_show, NULL); } static struct file_operations proc_cgroupstats_operations = { @@ -2428,14 +2833,14 @@ static struct file_operations proc_cgroupstats_operations = { /** * cgroup_fork - attach newly forked task to its parents cgroup. - * @tsk: pointer to task_struct of forking parent process. + * @child: pointer to task_struct of forking parent process. * * Description: A task inherits its parent's cgroup at fork(). * * A pointer to the shared css_set was automatically copied in * fork.c by dup_task_struct(). However, we ignore that copy, since * it was not made under the protection of RCU or cgroup_mutex, so - * might no longer be a valid cgroup pointer. attach_task() might + * might no longer be a valid cgroup pointer. cgroup_attach_task() might * have already changed current->cgroups, allowing the previously * referenced cgroup group to be removed and freed. * @@ -2452,9 +2857,12 @@ void cgroup_fork(struct task_struct *child) } /** - * cgroup_fork_callbacks - called on a new task very soon before - * adding it to the tasklist. No need to take any locks since no-one - * can be operating on this task + * cgroup_fork_callbacks - run fork callbacks + * @child: the new task + * + * Called on a new task very soon before adding it to the + * tasklist. No need to take any locks since no-one can + * be operating on this task. */ void cgroup_fork_callbacks(struct task_struct *child) { @@ -2469,23 +2877,29 @@ void cgroup_fork_callbacks(struct task_struct *child) } /** - * cgroup_post_fork - called on a new task after adding it to the - * task list. Adds the task to the list running through its css_set - * if necessary. Has to be after the task is visible on the task list - * in case we race with the first call to cgroup_iter_start() - to - * guarantee that the new task ends up on its list. */ + * cgroup_post_fork - called on a new task after adding it to the task list + * @child: the task in question + * + * Adds the task to the list running through its css_set if necessary. + * Has to be after the task is visible on the task list in case we race + * with the first call to cgroup_iter_start() - to guarantee that the + * new task ends up on its list. + */ void cgroup_post_fork(struct task_struct *child) { if (use_task_css_set_links) { write_lock(&css_set_lock); + task_lock(child); if (list_empty(&child->cg_list)) list_add(&child->cg_list, &child->cgroups->tasks); + task_unlock(child); write_unlock(&css_set_lock); } } /** * cgroup_exit - detach cgroup from exiting task * @tsk: pointer to task_struct of exiting process + * @run_callback: run exit callbacks? * * Description: Detach cgroup from @tsk and release it. * @@ -2514,9 +2928,8 @@ void cgroup_post_fork(struct task_struct *child) * attach us to a different cgroup, decrementing the count on * the first cgroup that we never incremented. But in this case, * top_cgroup isn't going away, and either task has PF_EXITING set, - * which wards off any attach_task() attempts, or task is a failed - * fork, never visible to attach_task. - * + * which wards off any cgroup_attach_task() attempts, or task is a failed + * fork, never visible to cgroup_attach_task. */ void cgroup_exit(struct task_struct *tsk, int run_callbacks) { @@ -2553,15 +2966,20 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks) } /** - * cgroup_clone - duplicate the current cgroup in the hierarchy - * that the given subsystem is attached to, and move this task into - * the new child + * cgroup_clone - clone the cgroup the given subsystem is attached to + * @tsk: the task to be moved + * @subsys: the given subsystem + * @nodename: the name for the new cgroup + * + * Duplicate the current cgroup in the hierarchy that the given + * subsystem is attached to, and move this task into the new + * child. */ -int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) +int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, + char *nodename) { struct dentry *dentry; int ret = 0; - char nodename[MAX_CGROUP_TYPE_NAMELEN]; struct cgroup *parent, *child; struct inode *inode; struct css_set *cg; @@ -2577,22 +2995,24 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) again: root = subsys->root; if (root == &rootnode) { - printk(KERN_INFO - "Not cloning cgroup for unused subsystem %s\n", - subsys->name); mutex_unlock(&cgroup_mutex); return 0; } - cg = tsk->cgroups; - parent = task_cgroup(tsk, subsys->subsys_id); - - snprintf(nodename, MAX_CGROUP_TYPE_NAMELEN, "node_%d", tsk->pid); /* Pin the hierarchy */ - atomic_inc(&parent->root->sb->s_active); + if (!atomic_inc_not_zero(&root->sb->s_active)) { + /* We race with the final deactivate_super() */ + mutex_unlock(&cgroup_mutex); + return 0; + } /* Keep the cgroup alive */ + task_lock(tsk); + parent = task_cgroup(tsk, subsys->subsys_id); + cg = tsk->cgroups; get_css_set(cg); + task_unlock(tsk); + mutex_unlock(&cgroup_mutex); /* Now do the VFS work to create a cgroup */ @@ -2611,7 +3031,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) } /* Create the cgroup directory, which also creates the cgroup */ - ret = vfs_mkdir(inode, dentry, S_IFDIR | 0755); + ret = vfs_mkdir(inode, dentry, 0755); child = __d_cgrp(dentry); dput(dentry); if (ret) { @@ -2621,13 +3041,6 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) goto out_release; } - if (!child) { - printk(KERN_INFO - "Couldn't find new cgroup %s\n", nodename); - ret = -ENOMEM; - goto out_release; - } - /* The cgroup now exists. Retake cgroup_mutex and check * that we're still in the same state that we thought we * were. */ @@ -2638,7 +3051,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) mutex_unlock(&inode->i_mutex); put_css_set(cg); - deactivate_super(parent->root->sb); + deactivate_super(root->sb); /* The cgroup is still accessible in the VFS, but * we're not going to try to rmdir() it at this * point. */ @@ -2655,7 +3068,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) } /* All seems fine. Finish by moving the task into the new cgroup */ - ret = attach_task(child, tsk); + ret = cgroup_attach_task(child, tsk); mutex_unlock(&cgroup_mutex); out_release: @@ -2664,13 +3077,16 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys) mutex_lock(&cgroup_mutex); put_css_set(cg); mutex_unlock(&cgroup_mutex); - deactivate_super(parent->root->sb); + deactivate_super(root->sb); return ret; } -/* - * See if "cgrp" is a descendant of the current task's cgroup in - * the appropriate hierarchy +/** + * cgroup_is_descendant - see if @cgrp is a descendant of current task's cgrp + * @cgrp: the cgroup in question + * + * See if @cgrp is a descendant of the current task's cgroup in + * the appropriate hierarchy. * * If we are sending in dummytop, then presumably we are creating * the top cgroup in the subsystem. @@ -2720,7 +3136,8 @@ void __css_put(struct cgroup_subsys_state *css) { struct cgroup *cgrp = css->cgroup; rcu_read_lock(); - if (atomic_dec_and_test(&css->refcnt) && notify_on_release(cgrp)) { + if ((atomic_dec_return(&css->refcnt) == 1) && + notify_on_release(cgrp)) { set_bit(CGRP_RELEASABLE, &cgrp->flags); check_for_release(cgrp); } @@ -2749,9 +3166,7 @@ void __css_put(struct cgroup_subsys_state *css) * release agent task. We don't bother to wait because the caller of * this routine has no use for the exit status of the release agent * task, so no sense holding our caller up for that. - * */ - static void cgroup_release_agent(struct work_struct *work) { BUG_ON(work != &release_agent_work); @@ -2760,27 +3175,24 @@ static void cgroup_release_agent(struct work_struct *work) while (!list_empty(&release_list)) { char *argv[3], *envp[3]; int i; - char *pathbuf; + char *pathbuf = NULL, *agentbuf = NULL; struct cgroup *cgrp = list_entry(release_list.next, struct cgroup, release_list); list_del_init(&cgrp->release_list); spin_unlock(&release_list_lock); pathbuf = kmalloc(PAGE_SIZE, GFP_KERNEL); - if (!pathbuf) { - spin_lock(&release_list_lock); - continue; - } - - if (cgroup_path(cgrp, pathbuf, PAGE_SIZE) < 0) { - kfree(pathbuf); - spin_lock(&release_list_lock); - continue; - } + if (!pathbuf) + goto continue_free; + if (cgroup_path(cgrp, pathbuf, PAGE_SIZE) < 0) + goto continue_free; + agentbuf = kstrdup(cgrp->root->release_agent_path, GFP_KERNEL); + if (!agentbuf) + goto continue_free; i = 0; - argv[i++] = cgrp->root->release_agent_path; - argv[i++] = (char *)pathbuf; + argv[i++] = agentbuf; + argv[i++] = pathbuf; argv[i] = NULL; i = 0; @@ -2794,10 +3206,36 @@ static void cgroup_release_agent(struct work_struct *work) * be a slow process */ mutex_unlock(&cgroup_mutex); call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC); - kfree(pathbuf); mutex_lock(&cgroup_mutex); + continue_free: + kfree(pathbuf); + kfree(agentbuf); spin_lock(&release_list_lock); } spin_unlock(&release_list_lock); mutex_unlock(&cgroup_mutex); } + +static int __init cgroup_disable(char *str) +{ + int i; + char *token; + + while ((token = strsep(&str, ",")) != NULL) { + if (!*token) + continue; + + for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { + struct cgroup_subsys *ss = subsys[i]; + + if (!strcmp(token, ss->name)) { + ss->disabled = 1; + printk(KERN_INFO "Disabling %s control group" + " subsystem\n", ss->name); + break; + } + } + } + return 1; +} +__setup("cgroup_disable=", cgroup_disable);