[linux-2.6-block.git] / security / device_cgroup.c

/*
 * dev_cgroup.c - device cgroup subsystem
 *
 * Copyright 2007 IBM Corp
 */

#include <linux/device_cgroup.h>
#include <linux/cgroup.h>
#include <linux/ctype.h>
#include <linux/list.h>
#include <linux/uaccess.h>
#include <linux/seq_file.h>

#define ACC_MKNOD 1
#define ACC_READ  2
#define ACC_WRITE 4
#define ACC_MASK (ACC_MKNOD | ACC_READ | ACC_WRITE)

#define DEV_BLOCK 1
#define DEV_CHAR  2
#define DEV_ALL   4  /* this represents all devices */

/*
 * whitelist locking rules:
 * cgroup_lock() cannot be taken under dev_cgroup->lock.
 * dev_cgroup->lock can be taken with or without cgroup_lock().
 *
 * modifications always require cgroup_lock
 * modifications to a list which is visible require the
 *   dev_cgroup->lock *and* cgroup_lock()
 * walking the list requires dev_cgroup->lock or cgroup_lock().
 *
 * reasoning: dev_whitelist_copy() needs to kmalloc, so needs
 *   a mutex, which the cgroup_lock() is.  Since modifying
 *   a visible list requires both locks, either lock can be
 *   taken for walking the list.
 */

struct dev_whitelist_item {
	u32 major, minor;
	short type;
	short access;
	struct list_head list;
	struct rcu_head rcu;
};

struct dev_cgroup {
	struct cgroup_subsys_state css;
	struct list_head whitelist;
	spinlock_t lock;
};

static inline struct dev_cgroup *css_to_devcgroup(struct cgroup_subsys_state *s)
{
	return container_of(s, struct dev_cgroup, css);
}

static inline struct dev_cgroup *cgroup_to_devcgroup(struct cgroup *cgroup)
{
	return css_to_devcgroup(cgroup_subsys_state(cgroup, devices_subsys_id));
}

static inline struct dev_cgroup *task_devcgroup(struct task_struct *task)
{
	return css_to_devcgroup(task_subsys_state(task, devices_subsys_id));
}

struct cgroup_subsys devices_subsys;

static int devcgroup_can_attach(struct cgroup_subsys *ss,
		struct cgroup *new_cgroup, struct task_struct *task)
{
	if (current != task && !capable(CAP_SYS_ADMIN))
			return -EPERM;

	return 0;
}

/*
 * called under cgroup_lock()
 */
static int dev_whitelist_copy(struct list_head *dest, struct list_head *orig)
{
	struct dev_whitelist_item *wh, *tmp, *new;

	list_for_each_entry(wh, orig, list) {
		new = kmalloc(sizeof(*wh), GFP_KERNEL);
		if (!new)
			goto free_and_exit;
		new->major = wh->major;
		new->minor = wh->minor;
		new->type = wh->type;
		new->access = wh->access;
		list_add_tail(&new->list, dest);
	}

	return 0;

free_and_exit:
	list_for_each_entry_safe(wh, tmp, dest, list) {
		list_del(&wh->list);
		kfree(wh);
	}
	return -ENOMEM;
}

/* Stupid prototype - don't bother combining existing entries */
/*
 * called under cgroup_lock()
 * since the list is visible to other tasks, we need the spinlock also
 */
static int dev_whitelist_add(struct dev_cgroup *dev_cgroup,
			struct dev_whitelist_item *wh)
{
	struct dev_whitelist_item *whcopy, *walk;

	whcopy = kmalloc(sizeof(*whcopy), GFP_KERNEL);
	if (!whcopy)
		return -ENOMEM;

	memcpy(whcopy, wh, sizeof(*whcopy));
	spin_lock(&dev_cgroup->lock);
	list_for_each_entry(walk, &dev_cgroup->whitelist, list) {
		if (walk->type != wh->type)
			continue;
		if (walk->major != wh->major)
			continue;
		if (walk->minor != wh->minor)
			continue;

		walk->access |= wh->access;
		kfree(whcopy);
		whcopy = NULL;
	}

	if (whcopy != NULL)
		list_add_tail_rcu(&whcopy->list, &dev_cgroup->whitelist);
	spin_unlock(&dev_cgroup->lock);
	return 0;
}

static void whitelist_item_free(struct rcu_head *rcu)
{
	struct dev_whitelist_item *item;

	item = container_of(rcu, struct dev_whitelist_item, rcu);
	kfree(item);
}

/*
 * called under cgroup_lock()
 * since the list is visible to other tasks, we need the spinlock also
 */
static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
			struct dev_whitelist_item *wh)
{
	struct dev_whitelist_item *walk, *tmp;

	spin_lock(&dev_cgroup->lock);
	list_for_each_entry_safe(walk, tmp, &dev_cgroup->whitelist, list) {
		if (walk->type == DEV_ALL)
			goto remove;
		if (walk->type != wh->type)
			continue;
		if (walk->major != ~0 && walk->major != wh->major)
			continue;
		if (walk->minor != ~0 && walk->minor != wh->minor)
			continue;

remove:
		walk->access &= ~wh->access;
		if (!walk->access) {
			list_del_rcu(&walk->list);
			call_rcu(&walk->rcu, whitelist_item_free);
		}
	}
	spin_unlock(&dev_cgroup->lock);
}

/*
 * called from kernel/cgroup.c with cgroup_lock() held.
 */
static struct cgroup_subsys_state *devcgroup_create(struct cgroup_subsys *ss,
						struct cgroup *cgroup)
{
	struct dev_cgroup *dev_cgroup, *parent_dev_cgroup;
	struct cgroup *parent_cgroup;
	int ret;

	dev_cgroup = kzalloc(sizeof(*dev_cgroup), GFP_KERNEL);
	if (!dev_cgroup)
		return ERR_PTR(-ENOMEM);
	INIT_LIST_HEAD(&dev_cgroup->whitelist);
	parent_cgroup = cgroup->parent;

	if (parent_cgroup == NULL) {
		struct dev_whitelist_item *wh;
		wh = kmalloc(sizeof(*wh), GFP_KERNEL);
		if (!wh) {
			kfree(dev_cgroup);
			return ERR_PTR(-ENOMEM);
		}
		wh->minor = wh->major = ~0;
		wh->type = DEV_ALL;
		wh->access = ACC_MASK;
		list_add(&wh->list, &dev_cgroup->whitelist);
	} else {
		parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
		ret = dev_whitelist_copy(&dev_cgroup->whitelist,
				&parent_dev_cgroup->whitelist);
		if (ret) {
			kfree(dev_cgroup);
			return ERR_PTR(ret);
		}
	}

	spin_lock_init(&dev_cgroup->lock);
	return &dev_cgroup->css;
}

static void devcgroup_destroy(struct cgroup_subsys *ss,
			struct cgroup *cgroup)
{
	struct dev_cgroup *dev_cgroup;
	struct dev_whitelist_item *wh, *tmp;

	dev_cgroup = cgroup_to_devcgroup(cgroup);
	list_for_each_entry_safe(wh, tmp, &dev_cgroup->whitelist, list) {
		list_del(&wh->list);
		kfree(wh);
	}
	kfree(dev_cgroup);
}

#define DEVCG_ALLOW 1
#define DEVCG_DENY 2
#define DEVCG_LIST 3

#define MAJMINLEN 13
#define ACCLEN 4

static void set_access(char *acc, short access)
{
	int idx = 0;
	memset(acc, 0, ACCLEN);
	if (access & ACC_READ)
		acc[idx++] = 'r';
	if (access & ACC_WRITE)
		acc[idx++] = 'w';
	if (access & ACC_MKNOD)
		acc[idx++] = 'm';
}

static char type_to_char(short type)
{
	if (type == DEV_ALL)
		return 'a';
	if (type == DEV_CHAR)
		return 'c';
	if (type == DEV_BLOCK)
		return 'b';
	return 'X';
}

static void set_majmin(char *str, unsigned m)
{
	if (m == ~0)
		strcpy(str, "*");
	else
		sprintf(str, "%u", m);
}

static int devcgroup_seq_read(struct cgroup *cgroup, struct cftype *cft,
				struct seq_file *m)
{
	struct dev_cgroup *devcgroup = cgroup_to_devcgroup(cgroup);
	struct dev_whitelist_item *wh;
	char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];

	rcu_read_lock();
	list_for_each_entry_rcu(wh, &devcgroup->whitelist, list) {
		set_access(acc, wh->access);
		set_majmin(maj, wh->major);
		set_majmin(min, wh->minor);
		seq_printf(m, "%c %s:%s %s\n", type_to_char(wh->type),
			   maj, min, acc);
	}
	rcu_read_unlock();

	return 0;
}

/*
 * may_access_whitelist:
 * does the access granted to dev_cgroup c contain the access
 * requested in whitelist item refwh.
 * return 1 if yes, 0 if no.
 * call with c->lock held
 */
static int may_access_whitelist(struct dev_cgroup *c,
				       struct dev_whitelist_item *refwh)
{
	struct dev_whitelist_item *whitem;

	list_for_each_entry(whitem, &c->whitelist, list) {
		if (whitem->type & DEV_ALL)
			return 1;
		if ((refwh->type & DEV_BLOCK) && !(whitem->type & DEV_BLOCK))
			continue;
		if ((refwh->type & DEV_CHAR) && !(whitem->type & DEV_CHAR))
			continue;
		if (whitem->major != ~0 && whitem->major != refwh->major)
			continue;
		if (whitem->minor != ~0 && whitem->minor != refwh->minor)
			continue;
		if (refwh->access & (~whitem->access))
			continue;
		return 1;
	}
	return 0;
}

/*
 * parent_has_perm:
 * when adding a new allow rule to a device whitelist, the rule
 * must be allowed in the parent device
 */
static int parent_has_perm(struct dev_cgroup *childcg,
				  struct dev_whitelist_item *wh)
{
	struct cgroup *pcg = childcg->css.cgroup->parent;
	struct dev_cgroup *parent;
	int ret;

	if (!pcg)
		return 1;
	parent = cgroup_to_devcgroup(pcg);
	spin_lock(&parent->lock);
	ret = may_access_whitelist(parent, wh);
	spin_unlock(&parent->lock);
	return ret;
}

/*
 * Modify the whitelist using allow/deny rules.
 * CAP_SYS_ADMIN is needed for this.  It's at least separate from CAP_MKNOD
 * so we can give a container CAP_MKNOD to let it create devices but not
 * modify the whitelist.
 * It seems likely we'll want to add a CAP_CONTAINER capability to allow
 * us to also grant CAP_SYS_ADMIN to containers without giving away the
 * device whitelist controls, but for now we'll stick with CAP_SYS_ADMIN
 *
 * Taking rules away is always allowed (given CAP_SYS_ADMIN).  Granting
 * new access is only allowed if you're in the top-level cgroup, or your
 * parent cgroup has the access you're asking for.
 */
static int devcgroup_update_access(struct dev_cgroup *devcgroup,
				   int filetype, const char *buffer)
{
	struct dev_cgroup *cur_devcgroup;
	const char *b;
	char *endp;
	int retval = 0, count;
	struct dev_whitelist_item wh;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	cur_devcgroup = task_devcgroup(current);

	memset(&wh, 0, sizeof(wh));
	b = buffer;

	switch (*b) {
	case 'a':
		wh.type = DEV_ALL;
		wh.access = ACC_MASK;
		wh.major = ~0;
		wh.minor = ~0;
		goto handle;
	case 'b':
		wh.type = DEV_BLOCK;
		break;
	case 'c':
		wh.type = DEV_CHAR;
		break;
	default:
		return -EINVAL;
	}
	b++;
	if (!isspace(*b))
		return -EINVAL;
	b++;
	if (*b == '*') {
		wh.major = ~0;
		b++;
	} else if (isdigit(*b)) {
		wh.major = simple_strtoul(b, &endp, 10);
		b = endp;
	} else {
		return -EINVAL;
	}
	if (*b != ':')
		return -EINVAL;
	b++;

	/* read minor */
	if (*b == '*') {
		wh.minor = ~0;
		b++;
	} else if (isdigit(*b)) {
		wh.minor = simple_strtoul(b, &endp, 10);
		b = endp;
	} else {
		return -EINVAL;
	}
	if (!isspace(*b))
		return -EINVAL;
	for (b++, count = 0; count < 3; count++, b++) {
		switch (*b) {
		case 'r':
			wh.access |= ACC_READ;
			break;
		case 'w':
			wh.access |= ACC_WRITE;
			break;
		case 'm':
			wh.access |= ACC_MKNOD;
			break;
		case '\n':
		case '\0':
			count = 3;
			break;
		default:
			return -EINVAL;
		}
	}

handle:
	retval = 0;
	switch (filetype) {
	case DEVCG_ALLOW:
		if (!parent_has_perm(devcgroup, &wh))
			return -EPERM;
		return dev_whitelist_add(devcgroup, &wh);
	case DEVCG_DENY:
		dev_whitelist_rm(devcgroup, &wh);
		break;
	default:
		return -EINVAL;
	}
	return 0;
}

static int devcgroup_access_write(struct cgroup *cgrp, struct cftype *cft,
				  const char *buffer)
{
	int retval;
	if (!cgroup_lock_live_group(cgrp))
		return -ENODEV;
	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
					 cft->private, buffer);
	cgroup_unlock();
	return retval;
}

static struct cftype dev_cgroup_files[] = {
	{
		.name = "allow",
		.write_string  = devcgroup_access_write,
		.private = DEVCG_ALLOW,
	},
	{
		.name = "deny",
		.write_string = devcgroup_access_write,
		.private = DEVCG_DENY,
	},
	{
		.name = "list",
		.read_seq_string = devcgroup_seq_read,
		.private = DEVCG_LIST,
	},
};

static int devcgroup_populate(struct cgroup_subsys *ss,
				struct cgroup *cgroup)
{
	return cgroup_add_files(cgroup, ss, dev_cgroup_files,
					ARRAY_SIZE(dev_cgroup_files));
}

struct cgroup_subsys devices_subsys = {
	.name = "devices",
	.can_attach = devcgroup_can_attach,
	.create = devcgroup_create,
	.destroy  = devcgroup_destroy,
	.populate = devcgroup_populate,
	.subsys_id = devices_subsys_id,
};

int devcgroup_inode_permission(struct inode *inode, int mask)
{
	struct dev_cgroup *dev_cgroup;
	struct dev_whitelist_item *wh;

	dev_t device = inode->i_rdev;
	if (!device)
		return 0;
	if (!S_ISBLK(inode->i_mode) && !S_ISCHR(inode->i_mode))
		return 0;
	dev_cgroup = css_to_devcgroup(task_subsys_state(current,
				devices_subsys_id));
	if (!dev_cgroup)
		return 0;

	rcu_read_lock();
	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
		if (wh->type & DEV_ALL)
			goto acc_check;
		if ((wh->type & DEV_BLOCK) && !S_ISBLK(inode->i_mode))
			continue;
		if ((wh->type & DEV_CHAR) && !S_ISCHR(inode->i_mode))
			continue;
		if (wh->major != ~0 && wh->major != imajor(inode))
			continue;
		if (wh->minor != ~0 && wh->minor != iminor(inode))
			continue;
acc_check:
		if ((mask & MAY_WRITE) && !(wh->access & ACC_WRITE))
			continue;
		if ((mask & MAY_READ) && !(wh->access & ACC_READ))
			continue;
		rcu_read_unlock();
		return 0;
	}
	rcu_read_unlock();

	return -EPERM;
}

int devcgroup_inode_mknod(int mode, dev_t dev)
{
	struct dev_cgroup *dev_cgroup;
	struct dev_whitelist_item *wh;

	dev_cgroup = css_to_devcgroup(task_subsys_state(current,
				devices_subsys_id));
	if (!dev_cgroup)
		return 0;

	rcu_read_lock();
	list_for_each_entry(wh, &dev_cgroup->whitelist, list) {
		if (wh->type & DEV_ALL)
			goto acc_check;
		if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode))
			continue;
		if ((wh->type & DEV_CHAR) && !S_ISCHR(mode))
			continue;
		if (wh->major != ~0 && wh->major != MAJOR(dev))
			continue;
		if (wh->minor != ~0 && wh->minor != MINOR(dev))
			continue;
acc_check:
		if (!(wh->access & ACC_MKNOD))
			continue;
		rcu_read_unlock();
		return 0;
	}
	rcu_read_unlock();
	return -EPERM;
}
Commit	Line	Data
08ce5f16 SH	1	/*
	2	* dev_cgroup.c - device cgroup subsystem
	3	*
	4	* Copyright 2007 IBM Corp
	5	*/
	6
	7	#include <linux/device_cgroup.h>
	8	#include <linux/cgroup.h>
	9	#include <linux/ctype.h>
	10	#include <linux/list.h>
	11	#include <linux/uaccess.h>
29486df3	12	#include <linux/seq_file.h>
08ce5f16 SH	13
	14	#define ACC_MKNOD 1
	15	#define ACC_READ 2
	16	#define ACC_WRITE 4
	17	#define ACC_MASK (ACC_MKNOD \| ACC_READ \| ACC_WRITE)
	18
	19	#define DEV_BLOCK 1
	20	#define DEV_CHAR 2
	21	#define DEV_ALL 4 /* this represents all devices */
	22
	23	/*
	24	* whitelist locking rules:
	25	* cgroup_lock() cannot be taken under dev_cgroup->lock.
	26	* dev_cgroup->lock can be taken with or without cgroup_lock().
	27	*
	28	* modifications always require cgroup_lock
	29	* modifications to a list which is visible require the
	30	* dev_cgroup->lock and cgroup_lock()
	31	* walking the list requires dev_cgroup->lock or cgroup_lock().
	32	*
	33	* reasoning: dev_whitelist_copy() needs to kmalloc, so needs
	34	* a mutex, which the cgroup_lock() is. Since modifying
	35	* a visible list requires both locks, either lock can be
	36	* taken for walking the list.
	37	*/
	38
	39	struct dev_whitelist_item {
	40	u32 major, minor;
	41	short type;
	42	short access;
	43	struct list_head list;
4efd1a1b	44	struct rcu_head rcu;
08ce5f16 SH	45	};
	46
	47	struct dev_cgroup {
	48	struct cgroup_subsys_state css;
	49	struct list_head whitelist;
	50	spinlock_t lock;
	51	};
	52
b66862f7 PE	53	static inline struct dev_cgroup css_to_devcgroup(struct cgroup_subsys_state s)
	54	{
	55	return container_of(s, struct dev_cgroup, css);
	56	}
	57
08ce5f16 SH	58	static inline struct dev_cgroup cgroup_to_devcgroup(struct cgroup cgroup)
08ce5f16 SH	59	{
b66862f7	60	return css_to_devcgroup(cgroup_subsys_state(cgroup, devices_subsys_id));
08ce5f16 SH	61	}
08ce5f16 SH	62
f92523e3 PM	63	static inline struct dev_cgroup task_devcgroup(struct task_struct task)
	64	{
	65	return css_to_devcgroup(task_subsys_state(task, devices_subsys_id));
	66	}
	67
08ce5f16 SH	68	struct cgroup_subsys devices_subsys;
	69
	70	static int devcgroup_can_attach(struct cgroup_subsys *ss,
	71	struct cgroup new_cgroup, struct task_struct task)
	72	{
	73	if (current != task && !capable(CAP_SYS_ADMIN))
	74	return -EPERM;
	75
	76	return 0;
	77	}
	78
	79	/*
	80	* called under cgroup_lock()
	81	*/
	82	static int dev_whitelist_copy(struct list_head dest, struct list_head orig)
	83	{
	84	struct dev_whitelist_item wh, tmp, *new;
	85
	86	list_for_each_entry(wh, orig, list) {
	87	new = kmalloc(sizeof(*wh), GFP_KERNEL);
	88	if (!new)
	89	goto free_and_exit;
	90	new->major = wh->major;
	91	new->minor = wh->minor;
	92	new->type = wh->type;
	93	new->access = wh->access;
	94	list_add_tail(&new->list, dest);
	95	}
	96
	97	return 0;
	98
	99	free_and_exit:
	100	list_for_each_entry_safe(wh, tmp, dest, list) {
	101	list_del(&wh->list);
	102	kfree(wh);
	103	}
	104	return -ENOMEM;
	105	}
	106
	107	/* Stupid prototype - don't bother combining existing entries */
	108	/*
	109	* called under cgroup_lock()
	110	* since the list is visible to other tasks, we need the spinlock also
	111	*/
	112	static int dev_whitelist_add(struct dev_cgroup *dev_cgroup,
	113	struct dev_whitelist_item *wh)
	114	{
d1ee2971	115	struct dev_whitelist_item whcopy, walk;
08ce5f16 SH	116
	117	whcopy = kmalloc(sizeof(*whcopy), GFP_KERNEL);
	118	if (!whcopy)
	119	return -ENOMEM;
	120
	121	memcpy(whcopy, wh, sizeof(*whcopy));
	122	spin_lock(&dev_cgroup->lock);
d1ee2971 PE	123	list_for_each_entry(walk, &dev_cgroup->whitelist, list) {
	124	if (walk->type != wh->type)
	125	continue;
	126	if (walk->major != wh->major)
	127	continue;
	128	if (walk->minor != wh->minor)
	129	continue;
	130
	131	walk->access \|= wh->access;
	132	kfree(whcopy);
	133	whcopy = NULL;
	134	}
	135
	136	if (whcopy != NULL)
4efd1a1b	137	list_add_tail_rcu(&whcopy->list, &dev_cgroup->whitelist);
08ce5f16 SH	138	spin_unlock(&dev_cgroup->lock);
	139	return 0;
	140	}
	141
4efd1a1b PE	142	static void whitelist_item_free(struct rcu_head *rcu)
	143	{
	144	struct dev_whitelist_item *item;
	145
	146	item = container_of(rcu, struct dev_whitelist_item, rcu);
	147	kfree(item);
	148	}
	149
08ce5f16 SH	150	/*
	151	* called under cgroup_lock()
	152	* since the list is visible to other tasks, we need the spinlock also
	153	*/
	154	static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
	155	struct dev_whitelist_item *wh)
	156	{
	157	struct dev_whitelist_item walk, tmp;
	158
	159	spin_lock(&dev_cgroup->lock);
	160	list_for_each_entry_safe(walk, tmp, &dev_cgroup->whitelist, list) {
	161	if (walk->type == DEV_ALL)
	162	goto remove;
	163	if (walk->type != wh->type)
	164	continue;
	165	if (walk->major != ~0 && walk->major != wh->major)
	166	continue;
	167	if (walk->minor != ~0 && walk->minor != wh->minor)
	168	continue;
	169
	170	remove:
	171	walk->access &= ~wh->access;
	172	if (!walk->access) {
4efd1a1b PE	173	list_del_rcu(&walk->list);
4efd1a1b PE	174	call_rcu(&walk->rcu, whitelist_item_free);
08ce5f16 SH	175	}
	176	}
	177	spin_unlock(&dev_cgroup->lock);
	178	}
	179
	180	/*
	181	* called from kernel/cgroup.c with cgroup_lock() held.
	182	*/
	183	static struct cgroup_subsys_state devcgroup_create(struct cgroup_subsys ss,
	184	struct cgroup *cgroup)
	185	{
	186	struct dev_cgroup dev_cgroup, parent_dev_cgroup;
	187	struct cgroup *parent_cgroup;
	188	int ret;
	189
	190	dev_cgroup = kzalloc(sizeof(*dev_cgroup), GFP_KERNEL);
	191	if (!dev_cgroup)
	192	return ERR_PTR(-ENOMEM);
	193	INIT_LIST_HEAD(&dev_cgroup->whitelist);
	194	parent_cgroup = cgroup->parent;
	195
	196	if (parent_cgroup == NULL) {
	197	struct dev_whitelist_item *wh;
	198	wh = kmalloc(sizeof(*wh), GFP_KERNEL);
	199	if (!wh) {
	200	kfree(dev_cgroup);
	201	return ERR_PTR(-ENOMEM);
	202	}
	203	wh->minor = wh->major = ~0;
	204	wh->type = DEV_ALL;
7759fc9d	205	wh->access = ACC_MASK;
08ce5f16 SH	206	list_add(&wh->list, &dev_cgroup->whitelist);
	207	} else {
	208	parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
	209	ret = dev_whitelist_copy(&dev_cgroup->whitelist,
	210	&parent_dev_cgroup->whitelist);
	211	if (ret) {
	212	kfree(dev_cgroup);
	213	return ERR_PTR(ret);
	214	}
	215	}
	216
	217	spin_lock_init(&dev_cgroup->lock);
	218	return &dev_cgroup->css;
	219	}
	220
	221	static void devcgroup_destroy(struct cgroup_subsys *ss,
	222	struct cgroup *cgroup)
	223	{
	224	struct dev_cgroup *dev_cgroup;
	225	struct dev_whitelist_item wh, tmp;
	226
	227	dev_cgroup = cgroup_to_devcgroup(cgroup);
	228	list_for_each_entry_safe(wh, tmp, &dev_cgroup->whitelist, list) {
	229	list_del(&wh->list);
	230	kfree(wh);
	231	}
	232	kfree(dev_cgroup);
	233	}
	234
	235	#define DEVCG_ALLOW 1
	236	#define DEVCG_DENY 2
29486df3 SH	237	#define DEVCG_LIST 3
29486df3 SH	238
17d213f8	239	#define MAJMINLEN 13
29486df3	240	#define ACCLEN 4
08ce5f16 SH	241
	242	static void set_access(char *acc, short access)
	243	{
	244	int idx = 0;
29486df3	245	memset(acc, 0, ACCLEN);
08ce5f16 SH	246	if (access & ACC_READ)
	247	acc[idx++] = 'r';
	248	if (access & ACC_WRITE)
	249	acc[idx++] = 'w';
	250	if (access & ACC_MKNOD)
	251	acc[idx++] = 'm';
	252	}
	253
	254	static char type_to_char(short type)
	255	{
	256	if (type == DEV_ALL)
	257	return 'a';
	258	if (type == DEV_CHAR)
	259	return 'c';
	260	if (type == DEV_BLOCK)
	261	return 'b';
	262	return 'X';
	263	}
	264
29486df3	265	static void set_majmin(char *str, unsigned m)
08ce5f16	266	{
08ce5f16	267	if (m == ~0)
7759fc9d	268	strcpy(str, "*");
08ce5f16	269	else
7759fc9d	270	sprintf(str, "%u", m);
08ce5f16 SH	271	}
08ce5f16 SH	272
29486df3 SH	273	static int devcgroup_seq_read(struct cgroup cgroup, struct cftype cft,
29486df3 SH	274	struct seq_file *m)
08ce5f16	275	{
29486df3	276	struct dev_cgroup *devcgroup = cgroup_to_devcgroup(cgroup);
08ce5f16	277	struct dev_whitelist_item *wh;
29486df3	278	char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];
08ce5f16	279
4efd1a1b PE	280	rcu_read_lock();
4efd1a1b PE	281	list_for_each_entry_rcu(wh, &devcgroup->whitelist, list) {
08ce5f16	282	set_access(acc, wh->access);
29486df3 SH	283	set_majmin(maj, wh->major);
	284	set_majmin(min, wh->minor);
	285	seq_printf(m, "%c %s:%s %s\n", type_to_char(wh->type),
	286	maj, min, acc);
08ce5f16	287	}
4efd1a1b	288	rcu_read_unlock();
08ce5f16	289
29486df3	290	return 0;
08ce5f16 SH	291	}
	292
	293	/*
	294	* may_access_whitelist:
	295	* does the access granted to dev_cgroup c contain the access
	296	* requested in whitelist item refwh.
	297	* return 1 if yes, 0 if no.
	298	* call with c->lock held
	299	*/
	300	static int may_access_whitelist(struct dev_cgroup *c,
	301	struct dev_whitelist_item *refwh)
	302	{
	303	struct dev_whitelist_item *whitem;
	304
	305	list_for_each_entry(whitem, &c->whitelist, list) {
	306	if (whitem->type & DEV_ALL)
	307	return 1;
	308	if ((refwh->type & DEV_BLOCK) && !(whitem->type & DEV_BLOCK))
	309	continue;
	310	if ((refwh->type & DEV_CHAR) && !(whitem->type & DEV_CHAR))
	311	continue;
	312	if (whitem->major != ~0 && whitem->major != refwh->major)
	313	continue;
	314	if (whitem->minor != ~0 && whitem->minor != refwh->minor)
	315	continue;
ec229e83	316	if (refwh->access & (~whitem->access))
08ce5f16 SH	317	continue;
	318	return 1;
	319	}
	320	return 0;
	321	}
	322
	323	/*
	324	* parent_has_perm:
	325	* when adding a new allow rule to a device whitelist, the rule
	326	* must be allowed in the parent device
	327	*/
f92523e3	328	static int parent_has_perm(struct dev_cgroup *childcg,
08ce5f16 SH	329	struct dev_whitelist_item *wh)
08ce5f16 SH	330	{
f92523e3	331	struct cgroup *pcg = childcg->css.cgroup->parent;
08ce5f16 SH	332	struct dev_cgroup *parent;
	333	int ret;
	334
	335	if (!pcg)
	336	return 1;
	337	parent = cgroup_to_devcgroup(pcg);
	338	spin_lock(&parent->lock);
	339	ret = may_access_whitelist(parent, wh);
	340	spin_unlock(&parent->lock);
	341	return ret;
	342	}
	343
	344	/*
	345	* Modify the whitelist using allow/deny rules.
	346	* CAP_SYS_ADMIN is needed for this. It's at least separate from CAP_MKNOD
	347	* so we can give a container CAP_MKNOD to let it create devices but not
	348	* modify the whitelist.
	349	* It seems likely we'll want to add a CAP_CONTAINER capability to allow
	350	* us to also grant CAP_SYS_ADMIN to containers without giving away the
	351	* device whitelist controls, but for now we'll stick with CAP_SYS_ADMIN
	352	*
	353	* Taking rules away is always allowed (given CAP_SYS_ADMIN). Granting
	354	* new access is only allowed if you're in the top-level cgroup, or your
	355	* parent cgroup has the access you're asking for.
	356	*/
f92523e3 PM	357	static int devcgroup_update_access(struct dev_cgroup *devcgroup,
f92523e3 PM	358	int filetype, const char *buffer)
08ce5f16	359	{
f92523e3 PM	360	struct dev_cgroup *cur_devcgroup;
f92523e3 PM	361	const char *b;
7759fc9d	362	char *endp;
08ce5f16 SH	363	int retval = 0, count;
	364	struct dev_whitelist_item wh;
	365
	366	if (!capable(CAP_SYS_ADMIN))
	367	return -EPERM;
	368
f92523e3	369	cur_devcgroup = task_devcgroup(current);
08ce5f16 SH	370
	371	memset(&wh, 0, sizeof(wh));
	372	b = buffer;
	373
	374	switch (*b) {
	375	case 'a':
	376	wh.type = DEV_ALL;
	377	wh.access = ACC_MASK;
d823f6bf LZ	378	wh.major = ~0;
d823f6bf LZ	379	wh.minor = ~0;
08ce5f16 SH	380	goto handle;
	381	case 'b':
	382	wh.type = DEV_BLOCK;
	383	break;
	384	case 'c':
	385	wh.type = DEV_CHAR;
	386	break;
	387	default:
f92523e3	388	return -EINVAL;
08ce5f16 SH	389	}
08ce5f16 SH	390	b++;
f92523e3 PM	391	if (!isspace(*b))
f92523e3 PM	392	return -EINVAL;
08ce5f16 SH	393	b++;
	394	if (b == '') {
	395	wh.major = ~0;
	396	b++;
	397	} else if (isdigit(*b)) {
7759fc9d LZ	398	wh.major = simple_strtoul(b, &endp, 10);
7759fc9d LZ	399	b = endp;
08ce5f16	400	} else {
f92523e3	401	return -EINVAL;
08ce5f16	402	}
f92523e3 PM	403	if (*b != ':')
f92523e3 PM	404	return -EINVAL;
08ce5f16 SH	405	b++;
	406
	407	/* read minor */
	408	if (b == '') {
	409	wh.minor = ~0;
	410	b++;
	411	} else if (isdigit(*b)) {
7759fc9d LZ	412	wh.minor = simple_strtoul(b, &endp, 10);
7759fc9d LZ	413	b = endp;
08ce5f16	414	} else {
f92523e3	415	return -EINVAL;
08ce5f16	416	}
f92523e3 PM	417	if (!isspace(*b))
f92523e3 PM	418	return -EINVAL;
08ce5f16 SH	419	for (b++, count = 0; count < 3; count++, b++) {
	420	switch (*b) {
	421	case 'r':
	422	wh.access \|= ACC_READ;
	423	break;
	424	case 'w':
	425	wh.access \|= ACC_WRITE;
	426	break;
	427	case 'm':
	428	wh.access \|= ACC_MKNOD;
	429	break;
	430	case '\n':
	431	case '\0':
	432	count = 3;
	433	break;
	434	default:
f92523e3	435	return -EINVAL;
08ce5f16 SH	436	}
	437	}
	438
	439	handle:
	440	retval = 0;
	441	switch (filetype) {
	442	case DEVCG_ALLOW:
f92523e3 PM	443	if (!parent_has_perm(devcgroup, &wh))
	444	return -EPERM;
	445	return dev_whitelist_add(devcgroup, &wh);
08ce5f16 SH	446	case DEVCG_DENY:
	447	dev_whitelist_rm(devcgroup, &wh);
	448	break;
	449	default:
f92523e3	450	return -EINVAL;
08ce5f16	451	}
f92523e3 PM	452	return 0;
f92523e3 PM	453	}
08ce5f16	454
f92523e3 PM	455	static int devcgroup_access_write(struct cgroup cgrp, struct cftype cft,
	456	const char *buffer)
	457	{
	458	int retval;
	459	if (!cgroup_lock_live_group(cgrp))
	460	return -ENODEV;
	461	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
	462	cft->private, buffer);
08ce5f16	463	cgroup_unlock();
08ce5f16 SH	464	return retval;
	465	}
	466
	467	static struct cftype dev_cgroup_files[] = {
	468	{
	469	.name = "allow",
f92523e3	470	.write_string = devcgroup_access_write,
08ce5f16 SH	471	.private = DEVCG_ALLOW,
	472	},
	473	{
	474	.name = "deny",
f92523e3	475	.write_string = devcgroup_access_write,
08ce5f16 SH	476	.private = DEVCG_DENY,
08ce5f16 SH	477	},
29486df3 SH	478	{
	479	.name = "list",
	480	.read_seq_string = devcgroup_seq_read,
	481	.private = DEVCG_LIST,
	482	},
08ce5f16 SH	483	};
	484
	485	static int devcgroup_populate(struct cgroup_subsys *ss,
	486	struct cgroup *cgroup)
	487	{
	488	return cgroup_add_files(cgroup, ss, dev_cgroup_files,
	489	ARRAY_SIZE(dev_cgroup_files));
	490	}
	491
	492	struct cgroup_subsys devices_subsys = {
	493	.name = "devices",
	494	.can_attach = devcgroup_can_attach,
	495	.create = devcgroup_create,
	496	.destroy = devcgroup_destroy,
	497	.populate = devcgroup_populate,
	498	.subsys_id = devices_subsys_id,
	499	};
	500
	501	int devcgroup_inode_permission(struct inode *inode, int mask)
	502	{
08ce5f16 SH	503	struct dev_cgroup *dev_cgroup;
	504	struct dev_whitelist_item *wh;
	505
	506	dev_t device = inode->i_rdev;
	507	if (!device)
	508	return 0;
	509	if (!S_ISBLK(inode->i_mode) && !S_ISCHR(inode->i_mode))
	510	return 0;
cc9cb219 PE	511	dev_cgroup = css_to_devcgroup(task_subsys_state(current,
cc9cb219 PE	512	devices_subsys_id));
08ce5f16 SH	513	if (!dev_cgroup)
	514	return 0;
	515
4efd1a1b PE	516	rcu_read_lock();
4efd1a1b PE	517	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
08ce5f16 SH	518	if (wh->type & DEV_ALL)
	519	goto acc_check;
	520	if ((wh->type & DEV_BLOCK) && !S_ISBLK(inode->i_mode))
	521	continue;
	522	if ((wh->type & DEV_CHAR) && !S_ISCHR(inode->i_mode))
	523	continue;
	524	if (wh->major != ~0 && wh->major != imajor(inode))
	525	continue;
	526	if (wh->minor != ~0 && wh->minor != iminor(inode))
	527	continue;
	528	acc_check:
	529	if ((mask & MAY_WRITE) && !(wh->access & ACC_WRITE))
	530	continue;
	531	if ((mask & MAY_READ) && !(wh->access & ACC_READ))
	532	continue;
4efd1a1b	533	rcu_read_unlock();
08ce5f16 SH	534	return 0;
08ce5f16 SH	535	}
4efd1a1b	536	rcu_read_unlock();
08ce5f16 SH	537
	538	return -EPERM;
	539	}
	540
	541	int devcgroup_inode_mknod(int mode, dev_t dev)
	542	{
08ce5f16 SH	543	struct dev_cgroup *dev_cgroup;
	544	struct dev_whitelist_item *wh;
	545
cc9cb219 PE	546	dev_cgroup = css_to_devcgroup(task_subsys_state(current,
cc9cb219 PE	547	devices_subsys_id));
08ce5f16 SH	548	if (!dev_cgroup)
	549	return 0;
	550
4efd1a1b	551	rcu_read_lock();
08ce5f16 SH	552	list_for_each_entry(wh, &dev_cgroup->whitelist, list) {
	553	if (wh->type & DEV_ALL)
	554	goto acc_check;
	555	if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode))
	556	continue;
	557	if ((wh->type & DEV_CHAR) && !S_ISCHR(mode))
	558	continue;
	559	if (wh->major != ~0 && wh->major != MAJOR(dev))
	560	continue;
	561	if (wh->minor != ~0 && wh->minor != MINOR(dev))
	562	continue;
	563	acc_check:
	564	if (!(wh->access & ACC_MKNOD))
	565	continue;
4efd1a1b	566	rcu_read_unlock();
08ce5f16 SH	567	return 0;
08ce5f16 SH	568	}
4efd1a1b	569	rcu_read_unlock();
08ce5f16 SH	570	return -EPERM;
08ce5f16 SH	571	}