X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=kernel%2Fuser.c;h=46d0165ca70c6b10240256fb1ccc2b3420c6c91e;hb=5da9a0fb673a0ea0a093862f95f6b89b3390c31e;hp=7e8215d87b40e70a13166dcc0efa064ee19a94d2;hpb=b1a8c172c318534b96d0f0f1aecdad3898118b98;p=safe%2Fjmp%2Flinux-2.6 diff --git a/kernel/user.c b/kernel/user.c index 7e8215d..46d0165 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -16,6 +16,15 @@ #include #include #include +#include "cred-internals.h" + +struct user_namespace init_user_ns = { + .kref = { + .refcount = ATOMIC_INIT(2), + }, + .creator = &root_user, +}; +EXPORT_SYMBOL_GPL(init_user_ns); /* * UID task count cache, to get fast user lookup in "alloc_uid" @@ -39,18 +48,15 @@ static struct kmem_cache *uid_cachep; */ static DEFINE_SPINLOCK(uidhash_lock); +/* root_user.__count is 2, 1 for init task cred, 1 for init_user_ns->creator */ struct user_struct root_user = { - .__count = ATOMIC_INIT(1), + .__count = ATOMIC_INIT(2), .processes = ATOMIC_INIT(1), .files = ATOMIC_INIT(0), .sigpending = ATOMIC_INIT(0), - .mq_bytes = 0, .locked_shm = 0, -#ifdef CONFIG_KEYS - .uid_keyring = &root_user_keyring, - .session_keyring = &root_session_keyring, -#endif -#ifdef CONFIG_FAIR_USER_SCHED + .user_ns = &init_user_ns, +#ifdef CONFIG_USER_SCHED .tg = &init_task_group, #endif }; @@ -58,34 +64,18 @@ struct user_struct root_user = { /* * These routines must be called with the uidhash spinlock held! */ -static inline void uid_hash_insert(struct user_struct *up, - struct hlist_head *hashent) +static void uid_hash_insert(struct user_struct *up, struct hlist_head *hashent) { hlist_add_head(&up->uidhash_node, hashent); } -static inline void uid_hash_remove(struct user_struct *up) +static void uid_hash_remove(struct user_struct *up) { hlist_del_init(&up->uidhash_node); + put_user_ns(up->user_ns); } -static inline struct user_struct *uid_hash_find(uid_t uid, - struct hlist_head *hashent) -{ - struct user_struct *user; - struct hlist_node *h; - - hlist_for_each_entry(user, h, hashent, uidhash_node) { - if (user->uid == uid) { - atomic_inc(&user->__count); - return user; - } - } - - return NULL; -} - -#ifdef CONFIG_FAIR_USER_SCHED +#ifdef CONFIG_USER_SCHED static void sched_destroy_user(struct user_struct *up) { @@ -96,29 +86,42 @@ static int sched_create_user(struct user_struct *up) { int rc = 0; - up->tg = sched_create_group(); + up->tg = sched_create_group(&root_task_group); if (IS_ERR(up->tg)) rc = -ENOMEM; - return rc; -} + set_tg_uid(up); -static void sched_switch_user(struct task_struct *p) -{ - sched_move_task(p); + return rc; } -#else /* CONFIG_FAIR_USER_SCHED */ +#else /* CONFIG_USER_SCHED */ static void sched_destroy_user(struct user_struct *up) { } static int sched_create_user(struct user_struct *up) { return 0; } -static void sched_switch_user(struct task_struct *p) { } -#endif /* CONFIG_FAIR_USER_SCHED */ +#endif /* CONFIG_USER_SCHED */ + +#if defined(CONFIG_USER_SCHED) && defined(CONFIG_SYSFS) + +static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) +{ + struct user_struct *user; + struct hlist_node *h; + + hlist_for_each_entry(user, h, hashent, uidhash_node) { + if (user->uid == uid) { + /* possibly resurrect an "almost deleted" object */ + if (atomic_inc_return(&user->__count) == 1) + cancel_delayed_work(&user->work); + return user; + } + } -#if defined(CONFIG_FAIR_USER_SCHED) && defined(CONFIG_SYSFS) + return NULL; +} -static struct kobject uids_kobject; /* represents /sys/kernel/uids directory */ +static struct kset *uids_kset; /* represents the /sys/kernel/uids/ directory */ static DEFINE_MUTEX(uids_mutex); static inline void uids_mutex_lock(void) @@ -131,94 +134,163 @@ static inline void uids_mutex_unlock(void) mutex_unlock(&uids_mutex); } -/* return cpu shares held by the user */ -ssize_t cpu_shares_show(struct kset *kset, char *buffer) +/* uid directory attributes */ +#ifdef CONFIG_FAIR_GROUP_SCHED +static ssize_t cpu_shares_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) { - struct user_struct *up = container_of(kset, struct user_struct, kset); + struct user_struct *up = container_of(kobj, struct user_struct, kobj); - return sprintf(buffer, "%lu\n", sched_group_shares(up->tg)); + return sprintf(buf, "%lu\n", sched_group_shares(up->tg)); } -/* modify cpu shares held by the user */ -ssize_t cpu_shares_store(struct kset *kset, const char *buffer, size_t size) +static ssize_t cpu_shares_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t size) { - struct user_struct *up = container_of(kset, struct user_struct, kset); + struct user_struct *up = container_of(kobj, struct user_struct, kobj); unsigned long shares; int rc; - sscanf(buffer, "%lu", &shares); + sscanf(buf, "%lu", &shares); rc = sched_group_set_shares(up->tg, shares); return (rc ? rc : size); } -static void user_attr_init(struct subsys_attribute *sa, char *name, int mode) +static struct kobj_attribute cpu_share_attr = + __ATTR(cpu_share, 0644, cpu_shares_show, cpu_shares_store); +#endif + +#ifdef CONFIG_RT_GROUP_SCHED +static ssize_t cpu_rt_runtime_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + struct user_struct *up = container_of(kobj, struct user_struct, kobj); + + return sprintf(buf, "%ld\n", sched_group_rt_runtime(up->tg)); +} + +static ssize_t cpu_rt_runtime_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t size) +{ + struct user_struct *up = container_of(kobj, struct user_struct, kobj); + unsigned long rt_runtime; + int rc; + + sscanf(buf, "%ld", &rt_runtime); + + rc = sched_group_set_rt_runtime(up->tg, rt_runtime); + + return (rc ? rc : size); +} + +static struct kobj_attribute cpu_rt_runtime_attr = + __ATTR(cpu_rt_runtime, 0644, cpu_rt_runtime_show, cpu_rt_runtime_store); + +static ssize_t cpu_rt_period_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + struct user_struct *up = container_of(kobj, struct user_struct, kobj); + + return sprintf(buf, "%lu\n", sched_group_rt_period(up->tg)); +} + +static ssize_t cpu_rt_period_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t size) +{ + struct user_struct *up = container_of(kobj, struct user_struct, kobj); + unsigned long rt_period; + int rc; + + sscanf(buf, "%lu", &rt_period); + + rc = sched_group_set_rt_period(up->tg, rt_period); + + return (rc ? rc : size); +} + +static struct kobj_attribute cpu_rt_period_attr = + __ATTR(cpu_rt_period, 0644, cpu_rt_period_show, cpu_rt_period_store); +#endif + +/* default attributes per uid directory */ +static struct attribute *uids_attributes[] = { +#ifdef CONFIG_FAIR_GROUP_SCHED + &cpu_share_attr.attr, +#endif +#ifdef CONFIG_RT_GROUP_SCHED + &cpu_rt_runtime_attr.attr, + &cpu_rt_period_attr.attr, +#endif + NULL +}; + +/* the lifetime of user_struct is not managed by the core (now) */ +static void uids_release(struct kobject *kobj) { - sa->attr.name = name; - sa->attr.mode = mode; - sa->show = cpu_shares_show; - sa->store = cpu_shares_store; + return; } -/* Create "/sys/kernel/uids/" directory and - * "/sys/kernel/uids//cpu_share" file for this user. +static struct kobj_type uids_ktype = { + .sysfs_ops = &kobj_sysfs_ops, + .default_attrs = uids_attributes, + .release = uids_release, +}; + +/* + * Create /sys/kernel/uids//cpu_share file for this user + * We do not create this file for users in a user namespace (until + * sysfs tagging is implemented). + * + * See Documentation/scheduler/sched-design-CFS.txt for ramifications. */ -static int user_kobject_create(struct user_struct *up) +static int uids_user_create(struct user_struct *up) { - struct kset *kset = &up->kset; - struct kobject *kobj = &kset->kobj; + struct kobject *kobj = &up->kobj; int error; - memset(kset, 0, sizeof(struct kset)); - kobj->parent = &uids_kobject; /* create under /sys/kernel/uids dir */ - kobject_set_name(kobj, "%d", up->uid); - kset_init(kset); - user_attr_init(&up->user_attr, "cpu_share", 0644); - - error = kobject_add(kobj); - if (error) + memset(kobj, 0, sizeof(struct kobject)); + if (up->user_ns != &init_user_ns) + return 0; + kobj->kset = uids_kset; + error = kobject_init_and_add(kobj, &uids_ktype, NULL, "%d", up->uid); + if (error) { + kobject_put(kobj); goto done; - - error = sysfs_create_file(kobj, &up->user_attr.attr); - if (error) - kobject_del(kobj); + } kobject_uevent(kobj, KOBJ_ADD); - done: return error; } -/* create these in sysfs filesystem: +/* create these entries in sysfs: * "/sys/kernel/uids" directory * "/sys/kernel/uids/0" directory (for root user) * "/sys/kernel/uids/0/cpu_share" file (for root user) */ -int __init uids_kobject_init(void) +int __init uids_sysfs_init(void) { - int error; - - /* create under /sys/kernel dir */ - uids_kobject.parent = &kernel_subsys.kobj; - uids_kobject.kset = &kernel_subsys; - kobject_set_name(&uids_kobject, "uids"); - kobject_init(&uids_kobject); + uids_kset = kset_create_and_add("uids", NULL, kernel_kobj); + if (!uids_kset) + return -ENOMEM; - error = kobject_add(&uids_kobject); - if (!error) - error = user_kobject_create(&root_user); - - return error; + return uids_user_create(&root_user); } -/* work function to remove sysfs directory for a user and free up +/* delayed work function to remove sysfs directory for a user and free up * corresponding structures. */ -static void remove_user_sysfs_dir(struct work_struct *w) +static void cleanup_user_struct(struct work_struct *w) { - struct user_struct *up = container_of(w, struct user_struct, work); - struct kobject *kobj = &up->kset.kobj; + struct user_struct *up = container_of(w, struct user_struct, work.work); unsigned long flags; int remove_user = 0; @@ -227,22 +299,21 @@ static void remove_user_sysfs_dir(struct work_struct *w) */ uids_mutex_lock(); - local_irq_save(flags); - - if (atomic_dec_and_lock(&up->__count, &uidhash_lock)) { + spin_lock_irqsave(&uidhash_lock, flags); + if (atomic_read(&up->__count) == 0) { uid_hash_remove(up); remove_user = 1; - spin_unlock_irqrestore(&uidhash_lock, flags); - } else { - local_irq_restore(flags); } + spin_unlock_irqrestore(&uidhash_lock, flags); if (!remove_user) goto done; - sysfs_remove_file(kobj, &up->user_attr.attr); - kobject_uevent(kobj, KOBJ_REMOVE); - kobject_del(kobj); + if (up->user_ns == &init_user_ns) { + kobject_uevent(&up->kobj, KOBJ_REMOVE); + kobject_del(&up->kobj); + kobject_put(&up->kobj); + } sched_destroy_user(up); key_put(up->uid_keyring); @@ -257,19 +328,32 @@ done: * IRQ state (as stored in flags) is restored and uidhash_lock released * upon function exit. */ -static inline void free_user(struct user_struct *up, unsigned long flags) +static void free_user(struct user_struct *up, unsigned long flags) { - /* restore back the count */ - atomic_inc(&up->__count); + INIT_DELAYED_WORK(&up->work, cleanup_user_struct); + schedule_delayed_work(&up->work, msecs_to_jiffies(1000)); spin_unlock_irqrestore(&uidhash_lock, flags); - - INIT_WORK(&up->work, remove_user_sysfs_dir); - schedule_work(&up->work); } -#else /* CONFIG_FAIR_USER_SCHED && CONFIG_SYSFS */ +#else /* CONFIG_USER_SCHED && CONFIG_SYSFS */ + +static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) +{ + struct user_struct *user; + struct hlist_node *h; + + hlist_for_each_entry(user, h, hashent, uidhash_node) { + if (user->uid == uid) { + atomic_inc(&user->__count); + return user; + } + } + + return NULL; +} -static inline int user_kobject_create(struct user_struct *up) { return 0; } +int uids_sysfs_init(void) { return 0; } +static inline int uids_user_create(struct user_struct *up) { return 0; } static inline void uids_mutex_lock(void) { } static inline void uids_mutex_unlock(void) { } @@ -277,7 +361,7 @@ static inline void uids_mutex_unlock(void) { } * IRQ state (as stored in flags) is restored and uidhash_lock released * upon function exit. */ -static inline void free_user(struct user_struct *up, unsigned long flags) +static void free_user(struct user_struct *up, unsigned long flags) { uid_hash_remove(up); spin_unlock_irqrestore(&uidhash_lock, flags); @@ -289,6 +373,24 @@ static inline void free_user(struct user_struct *up, unsigned long flags) #endif +#if defined(CONFIG_RT_GROUP_SCHED) && defined(CONFIG_USER_SCHED) +/* + * We need to check if a setuid can take place. This function should be called + * before successfully completing the setuid. + */ +int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) +{ + + return sched_rt_can_attach(up->tg, tsk); + +} +#else +int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) +{ + return 1; +} +#endif + /* * Locate the user_struct for the passed UID. If found, take a ref on it. The * caller must undo that ref with free_uid(). @@ -299,7 +401,7 @@ struct user_struct *find_user(uid_t uid) { struct user_struct *ret; unsigned long flags; - struct user_namespace *ns = current->nsproxy->user_ns; + struct user_namespace *ns = current_user_ns(); spin_lock_irqsave(&uidhash_lock, flags); ret = uid_hash_find(uid, uidhashentry(ns, uid)); @@ -321,12 +423,12 @@ void free_uid(struct user_struct *up) local_irq_restore(flags); } -struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid) +struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) { struct hlist_head *hashent = uidhashentry(ns, uid); - struct user_struct *up; + struct user_struct *up, *new; - /* Make uid_hash_find() + user_kobject_create() + uid_hash_insert() + /* Make uid_hash_find() + uids_user_create() + uid_hash_insert() * atomic. */ uids_mutex_lock(); @@ -336,44 +438,20 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid) spin_unlock_irq(&uidhash_lock); if (!up) { - struct user_struct *new; - - new = kmem_cache_alloc(uid_cachep, GFP_KERNEL); + new = kmem_cache_zalloc(uid_cachep, GFP_KERNEL); if (!new) - return NULL; + goto out_unlock; + new->uid = uid; atomic_set(&new->__count, 1); - atomic_set(&new->processes, 0); - atomic_set(&new->files, 0); - atomic_set(&new->sigpending, 0); -#ifdef CONFIG_INOTIFY_USER - atomic_set(&new->inotify_watches, 0); - atomic_set(&new->inotify_devs, 0); -#endif - new->mq_bytes = 0; - new->locked_shm = 0; + if (sched_create_user(new) < 0) + goto out_free_user; - if (alloc_uid_keyring(new, current) < 0) { - kmem_cache_free(uid_cachep, new); - return NULL; - } - - if (sched_create_user(new) < 0) { - key_put(new->uid_keyring); - key_put(new->session_keyring); - kmem_cache_free(uid_cachep, new); - return NULL; - } + new->user_ns = get_user_ns(ns); - if (user_kobject_create(new)) { - sched_destroy_user(new); - key_put(new->uid_keyring); - key_put(new->session_keyring); - kmem_cache_free(uid_cachep, new); - uids_mutex_unlock(); - return NULL; - } + if (uids_user_create(new)) + goto out_destoy_sched; /* * Before adding this, check whether we raced @@ -382,7 +460,7 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid) spin_lock_irq(&uidhash_lock); up = uid_hash_find(uid, hashent); if (up) { - /* This case is not possible when CONFIG_FAIR_USER_SCHED + /* This case is not possible when CONFIG_USER_SCHED * is defined, since we serialize alloc_uid() using * uids_mutex. Hence no need to call * sched_destroy_user() or remove_user_sysfs_dir(). @@ -395,67 +473,20 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid) up = new; } spin_unlock_irq(&uidhash_lock); - } uids_mutex_unlock(); return up; -} -void switch_uid(struct user_struct *new_user) -{ - struct user_struct *old_user; - - /* What if a process setreuid()'s and this brings the - * new uid over his NPROC rlimit? We can check this now - * cheaply with the new uid cache, so if it matters - * we should be checking for it. -DaveM - */ - old_user = current->user; - atomic_inc(&new_user->processes); - atomic_dec(&old_user->processes); - switch_uid_keyring(new_user); - current->user = new_user; - sched_switch_user(current); - - /* - * We need to synchronize with __sigqueue_alloc() - * doing a get_uid(p->user).. If that saw the old - * user value, we need to wait until it has exited - * its critical region before we can free the old - * structure. - */ - smp_mb(); - spin_unlock_wait(¤t->sighand->siglock); - - free_uid(old_user); - suid_keys(current); -} - -void release_uids(struct user_namespace *ns) -{ - int i; - unsigned long flags; - struct hlist_head *head; - struct hlist_node *nd; - - spin_lock_irqsave(&uidhash_lock, flags); - /* - * collapse the chains so that the user_struct-s will - * be still alive, but not in hashes. subsequent free_uid() - * will free them. - */ - for (i = 0; i < UIDHASH_SZ; i++) { - head = ns->uidhash_table + i; - while (!hlist_empty(head)) { - nd = head->first; - hlist_del_init(nd); - } - } - spin_unlock_irqrestore(&uidhash_lock, flags); - - free_uid(ns->root_user); +out_destoy_sched: + sched_destroy_user(new); + put_user_ns(new->user_ns); +out_free_user: + kmem_cache_free(uid_cachep, new); +out_unlock: + uids_mutex_unlock(); + return NULL; } static int __init uid_cache_init(void)