[net-next-2.6.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 = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
		if (!new)
			goto free_and_exit;
		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 = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
	if (!whcopy)
		return -ENOMEM;

	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;

	rcu_read_lock();

	dev_cgroup = task_devcgroup(current);

	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;

	rcu_read_lock();

	dev_cgroup = task_devcgroup(current);

	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) {
2cdc7241	87	new = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
08ce5f16 SH	88	if (!new)
08ce5f16 SH	89	goto free_and_exit;
08ce5f16 SH	90	list_add_tail(&new->list, dest);
	91	}
	92
	93	return 0;
	94
	95	free_and_exit:
	96	list_for_each_entry_safe(wh, tmp, dest, list) {
	97	list_del(&wh->list);
	98	kfree(wh);
	99	}
	100	return -ENOMEM;
	101	}
	102
	103	/* Stupid prototype - don't bother combining existing entries */
	104	/*
	105	* called under cgroup_lock()
	106	* since the list is visible to other tasks, we need the spinlock also
	107	*/
	108	static int dev_whitelist_add(struct dev_cgroup *dev_cgroup,
	109	struct dev_whitelist_item *wh)
	110	{
d1ee2971	111	struct dev_whitelist_item whcopy, walk;
08ce5f16	112
2cdc7241	113	whcopy = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
08ce5f16 SH	114	if (!whcopy)
	115	return -ENOMEM;
	116
08ce5f16	117	spin_lock(&dev_cgroup->lock);
d1ee2971 PE	118	list_for_each_entry(walk, &dev_cgroup->whitelist, list) {
	119	if (walk->type != wh->type)
	120	continue;
	121	if (walk->major != wh->major)
	122	continue;
	123	if (walk->minor != wh->minor)
	124	continue;
	125
	126	walk->access \|= wh->access;
	127	kfree(whcopy);
	128	whcopy = NULL;
	129	}
	130
	131	if (whcopy != NULL)
4efd1a1b	132	list_add_tail_rcu(&whcopy->list, &dev_cgroup->whitelist);
08ce5f16 SH	133	spin_unlock(&dev_cgroup->lock);
	134	return 0;
	135	}
	136
4efd1a1b PE	137	static void whitelist_item_free(struct rcu_head *rcu)
	138	{
	139	struct dev_whitelist_item *item;
	140
	141	item = container_of(rcu, struct dev_whitelist_item, rcu);
	142	kfree(item);
	143	}
	144
08ce5f16 SH	145	/*
	146	* called under cgroup_lock()
	147	* since the list is visible to other tasks, we need the spinlock also
	148	*/
	149	static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
	150	struct dev_whitelist_item *wh)
	151	{
	152	struct dev_whitelist_item walk, tmp;
	153
	154	spin_lock(&dev_cgroup->lock);
	155	list_for_each_entry_safe(walk, tmp, &dev_cgroup->whitelist, list) {
	156	if (walk->type == DEV_ALL)
	157	goto remove;
	158	if (walk->type != wh->type)
	159	continue;
	160	if (walk->major != ~0 && walk->major != wh->major)
	161	continue;
	162	if (walk->minor != ~0 && walk->minor != wh->minor)
	163	continue;
	164
	165	remove:
	166	walk->access &= ~wh->access;
	167	if (!walk->access) {
4efd1a1b PE	168	list_del_rcu(&walk->list);
4efd1a1b PE	169	call_rcu(&walk->rcu, whitelist_item_free);
08ce5f16 SH	170	}
	171	}
	172	spin_unlock(&dev_cgroup->lock);
	173	}
	174
	175	/*
	176	* called from kernel/cgroup.c with cgroup_lock() held.
	177	*/
	178	static struct cgroup_subsys_state devcgroup_create(struct cgroup_subsys ss,
	179	struct cgroup *cgroup)
	180	{
	181	struct dev_cgroup dev_cgroup, parent_dev_cgroup;
	182	struct cgroup *parent_cgroup;
	183	int ret;
	184
	185	dev_cgroup = kzalloc(sizeof(*dev_cgroup), GFP_KERNEL);
	186	if (!dev_cgroup)
	187	return ERR_PTR(-ENOMEM);
	188	INIT_LIST_HEAD(&dev_cgroup->whitelist);
	189	parent_cgroup = cgroup->parent;
	190
	191	if (parent_cgroup == NULL) {
	192	struct dev_whitelist_item *wh;
	193	wh = kmalloc(sizeof(*wh), GFP_KERNEL);
	194	if (!wh) {
	195	kfree(dev_cgroup);
	196	return ERR_PTR(-ENOMEM);
	197	}
	198	wh->minor = wh->major = ~0;
	199	wh->type = DEV_ALL;
7759fc9d	200	wh->access = ACC_MASK;
08ce5f16 SH	201	list_add(&wh->list, &dev_cgroup->whitelist);
	202	} else {
	203	parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
	204	ret = dev_whitelist_copy(&dev_cgroup->whitelist,
	205	&parent_dev_cgroup->whitelist);
	206	if (ret) {
	207	kfree(dev_cgroup);
	208	return ERR_PTR(ret);
	209	}
	210	}
	211
	212	spin_lock_init(&dev_cgroup->lock);
	213	return &dev_cgroup->css;
	214	}
	215
	216	static void devcgroup_destroy(struct cgroup_subsys *ss,
	217	struct cgroup *cgroup)
	218	{
	219	struct dev_cgroup *dev_cgroup;
	220	struct dev_whitelist_item wh, tmp;
	221
	222	dev_cgroup = cgroup_to_devcgroup(cgroup);
	223	list_for_each_entry_safe(wh, tmp, &dev_cgroup->whitelist, list) {
	224	list_del(&wh->list);
	225	kfree(wh);
	226	}
	227	kfree(dev_cgroup);
	228	}
	229
	230	#define DEVCG_ALLOW 1
	231	#define DEVCG_DENY 2
29486df3 SH	232	#define DEVCG_LIST 3
29486df3 SH	233
17d213f8	234	#define MAJMINLEN 13
29486df3	235	#define ACCLEN 4
08ce5f16 SH	236
	237	static void set_access(char *acc, short access)
	238	{
	239	int idx = 0;
29486df3	240	memset(acc, 0, ACCLEN);
08ce5f16 SH	241	if (access & ACC_READ)
	242	acc[idx++] = 'r';
	243	if (access & ACC_WRITE)
	244	acc[idx++] = 'w';
	245	if (access & ACC_MKNOD)
	246	acc[idx++] = 'm';
	247	}
	248
	249	static char type_to_char(short type)
	250	{
	251	if (type == DEV_ALL)
	252	return 'a';
	253	if (type == DEV_CHAR)
	254	return 'c';
	255	if (type == DEV_BLOCK)
	256	return 'b';
	257	return 'X';
	258	}
	259
29486df3	260	static void set_majmin(char *str, unsigned m)
08ce5f16	261	{
08ce5f16	262	if (m == ~0)
7759fc9d	263	strcpy(str, "*");
08ce5f16	264	else
7759fc9d	265	sprintf(str, "%u", m);
08ce5f16 SH	266	}
08ce5f16 SH	267
29486df3 SH	268	static int devcgroup_seq_read(struct cgroup cgroup, struct cftype cft,
29486df3 SH	269	struct seq_file *m)
08ce5f16	270	{
29486df3	271	struct dev_cgroup *devcgroup = cgroup_to_devcgroup(cgroup);
08ce5f16	272	struct dev_whitelist_item *wh;
29486df3	273	char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];
08ce5f16	274
4efd1a1b PE	275	rcu_read_lock();
4efd1a1b PE	276	list_for_each_entry_rcu(wh, &devcgroup->whitelist, list) {
08ce5f16	277	set_access(acc, wh->access);
29486df3 SH	278	set_majmin(maj, wh->major);
	279	set_majmin(min, wh->minor);
	280	seq_printf(m, "%c %s:%s %s\n", type_to_char(wh->type),
	281	maj, min, acc);
08ce5f16	282	}
4efd1a1b	283	rcu_read_unlock();
08ce5f16	284
29486df3	285	return 0;
08ce5f16 SH	286	}
	287
	288	/*
	289	* may_access_whitelist:
	290	* does the access granted to dev_cgroup c contain the access
	291	* requested in whitelist item refwh.
	292	* return 1 if yes, 0 if no.
	293	* call with c->lock held
	294	*/
	295	static int may_access_whitelist(struct dev_cgroup *c,
	296	struct dev_whitelist_item *refwh)
	297	{
	298	struct dev_whitelist_item *whitem;
	299
	300	list_for_each_entry(whitem, &c->whitelist, list) {
	301	if (whitem->type & DEV_ALL)
	302	return 1;
	303	if ((refwh->type & DEV_BLOCK) && !(whitem->type & DEV_BLOCK))
	304	continue;
	305	if ((refwh->type & DEV_CHAR) && !(whitem->type & DEV_CHAR))
	306	continue;
	307	if (whitem->major != ~0 && whitem->major != refwh->major)
	308	continue;
	309	if (whitem->minor != ~0 && whitem->minor != refwh->minor)
	310	continue;
ec229e83	311	if (refwh->access & (~whitem->access))
08ce5f16 SH	312	continue;
	313	return 1;
	314	}
	315	return 0;
	316	}
	317
	318	/*
	319	* parent_has_perm:
	320	* when adding a new allow rule to a device whitelist, the rule
	321	* must be allowed in the parent device
	322	*/
f92523e3	323	static int parent_has_perm(struct dev_cgroup *childcg,
08ce5f16 SH	324	struct dev_whitelist_item *wh)
08ce5f16 SH	325	{
f92523e3	326	struct cgroup *pcg = childcg->css.cgroup->parent;
08ce5f16 SH	327	struct dev_cgroup *parent;
	328	int ret;
	329
	330	if (!pcg)
	331	return 1;
	332	parent = cgroup_to_devcgroup(pcg);
	333	spin_lock(&parent->lock);
	334	ret = may_access_whitelist(parent, wh);
	335	spin_unlock(&parent->lock);
	336	return ret;
	337	}
	338
	339	/*
	340	* Modify the whitelist using allow/deny rules.
	341	* CAP_SYS_ADMIN is needed for this. It's at least separate from CAP_MKNOD
	342	* so we can give a container CAP_MKNOD to let it create devices but not
	343	* modify the whitelist.
	344	* It seems likely we'll want to add a CAP_CONTAINER capability to allow
	345	* us to also grant CAP_SYS_ADMIN to containers without giving away the
	346	* device whitelist controls, but for now we'll stick with CAP_SYS_ADMIN
	347	*
	348	* Taking rules away is always allowed (given CAP_SYS_ADMIN). Granting
	349	* new access is only allowed if you're in the top-level cgroup, or your
	350	* parent cgroup has the access you're asking for.
	351	*/
f92523e3 PM	352	static int devcgroup_update_access(struct dev_cgroup *devcgroup,
f92523e3 PM	353	int filetype, const char *buffer)
08ce5f16	354	{
f92523e3 PM	355	struct dev_cgroup *cur_devcgroup;
f92523e3 PM	356	const char *b;
7759fc9d	357	char *endp;
08ce5f16 SH	358	int retval = 0, count;
	359	struct dev_whitelist_item wh;
	360
	361	if (!capable(CAP_SYS_ADMIN))
	362	return -EPERM;
	363
f92523e3	364	cur_devcgroup = task_devcgroup(current);
08ce5f16 SH	365
	366	memset(&wh, 0, sizeof(wh));
	367	b = buffer;
	368
	369	switch (*b) {
	370	case 'a':
	371	wh.type = DEV_ALL;
	372	wh.access = ACC_MASK;
d823f6bf LZ	373	wh.major = ~0;
d823f6bf LZ	374	wh.minor = ~0;
08ce5f16 SH	375	goto handle;
	376	case 'b':
	377	wh.type = DEV_BLOCK;
	378	break;
	379	case 'c':
	380	wh.type = DEV_CHAR;
	381	break;
	382	default:
f92523e3	383	return -EINVAL;
08ce5f16 SH	384	}
08ce5f16 SH	385	b++;
f92523e3 PM	386	if (!isspace(*b))
f92523e3 PM	387	return -EINVAL;
08ce5f16 SH	388	b++;
	389	if (b == '') {
	390	wh.major = ~0;
	391	b++;
	392	} else if (isdigit(*b)) {
7759fc9d LZ	393	wh.major = simple_strtoul(b, &endp, 10);
7759fc9d LZ	394	b = endp;
08ce5f16	395	} else {
f92523e3	396	return -EINVAL;
08ce5f16	397	}
f92523e3 PM	398	if (*b != ':')
f92523e3 PM	399	return -EINVAL;
08ce5f16 SH	400	b++;
	401
	402	/* read minor */
	403	if (b == '') {
	404	wh.minor = ~0;
	405	b++;
	406	} else if (isdigit(*b)) {
7759fc9d LZ	407	wh.minor = simple_strtoul(b, &endp, 10);
7759fc9d LZ	408	b = endp;
08ce5f16	409	} else {
f92523e3	410	return -EINVAL;
08ce5f16	411	}
f92523e3 PM	412	if (!isspace(*b))
f92523e3 PM	413	return -EINVAL;
08ce5f16 SH	414	for (b++, count = 0; count < 3; count++, b++) {
	415	switch (*b) {
	416	case 'r':
	417	wh.access \|= ACC_READ;
	418	break;
	419	case 'w':
	420	wh.access \|= ACC_WRITE;
	421	break;
	422	case 'm':
	423	wh.access \|= ACC_MKNOD;
	424	break;
	425	case '\n':
	426	case '\0':
	427	count = 3;
	428	break;
	429	default:
f92523e3	430	return -EINVAL;
08ce5f16 SH	431	}
	432	}
	433
	434	handle:
	435	retval = 0;
	436	switch (filetype) {
	437	case DEVCG_ALLOW:
f92523e3 PM	438	if (!parent_has_perm(devcgroup, &wh))
	439	return -EPERM;
	440	return dev_whitelist_add(devcgroup, &wh);
08ce5f16 SH	441	case DEVCG_DENY:
	442	dev_whitelist_rm(devcgroup, &wh);
	443	break;
	444	default:
f92523e3	445	return -EINVAL;
08ce5f16	446	}
f92523e3 PM	447	return 0;
f92523e3 PM	448	}
08ce5f16	449
f92523e3 PM	450	static int devcgroup_access_write(struct cgroup cgrp, struct cftype cft,
	451	const char *buffer)
	452	{
	453	int retval;
	454	if (!cgroup_lock_live_group(cgrp))
	455	return -ENODEV;
	456	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
	457	cft->private, buffer);
08ce5f16	458	cgroup_unlock();
08ce5f16 SH	459	return retval;
	460	}
	461
	462	static struct cftype dev_cgroup_files[] = {
	463	{
	464	.name = "allow",
f92523e3	465	.write_string = devcgroup_access_write,
08ce5f16 SH	466	.private = DEVCG_ALLOW,
	467	},
	468	{
	469	.name = "deny",
f92523e3	470	.write_string = devcgroup_access_write,
08ce5f16 SH	471	.private = DEVCG_DENY,
08ce5f16 SH	472	},
29486df3 SH	473	{
	474	.name = "list",
	475	.read_seq_string = devcgroup_seq_read,
	476	.private = DEVCG_LIST,
	477	},
08ce5f16 SH	478	};
	479
	480	static int devcgroup_populate(struct cgroup_subsys *ss,
	481	struct cgroup *cgroup)
	482	{
	483	return cgroup_add_files(cgroup, ss, dev_cgroup_files,
	484	ARRAY_SIZE(dev_cgroup_files));
	485	}
	486
	487	struct cgroup_subsys devices_subsys = {
	488	.name = "devices",
	489	.can_attach = devcgroup_can_attach,
	490	.create = devcgroup_create,
	491	.destroy = devcgroup_destroy,
	492	.populate = devcgroup_populate,
	493	.subsys_id = devices_subsys_id,
	494	};
	495
	496	int devcgroup_inode_permission(struct inode *inode, int mask)
	497	{
08ce5f16 SH	498	struct dev_cgroup *dev_cgroup;
	499	struct dev_whitelist_item *wh;
	500
	501	dev_t device = inode->i_rdev;
	502	if (!device)
	503	return 0;
	504	if (!S_ISBLK(inode->i_mode) && !S_ISCHR(inode->i_mode))
	505	return 0;
08ce5f16	506
4efd1a1b	507	rcu_read_lock();
36fd71d2 LZ	508
	509	dev_cgroup = task_devcgroup(current);
	510
4efd1a1b	511	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
08ce5f16 SH	512	if (wh->type & DEV_ALL)
	513	goto acc_check;
	514	if ((wh->type & DEV_BLOCK) && !S_ISBLK(inode->i_mode))
	515	continue;
	516	if ((wh->type & DEV_CHAR) && !S_ISCHR(inode->i_mode))
	517	continue;
	518	if (wh->major != ~0 && wh->major != imajor(inode))
	519	continue;
	520	if (wh->minor != ~0 && wh->minor != iminor(inode))
	521	continue;
	522	acc_check:
	523	if ((mask & MAY_WRITE) && !(wh->access & ACC_WRITE))
	524	continue;
	525	if ((mask & MAY_READ) && !(wh->access & ACC_READ))
	526	continue;
4efd1a1b	527	rcu_read_unlock();
08ce5f16 SH	528	return 0;
08ce5f16 SH	529	}
36fd71d2	530
4efd1a1b	531	rcu_read_unlock();
08ce5f16 SH	532
	533	return -EPERM;
	534	}
	535
	536	int devcgroup_inode_mknod(int mode, dev_t dev)
	537	{
08ce5f16 SH	538	struct dev_cgroup *dev_cgroup;
	539	struct dev_whitelist_item *wh;
	540
4efd1a1b	541	rcu_read_lock();
36fd71d2 LZ	542
	543	dev_cgroup = task_devcgroup(current);
	544
08ce5f16 SH	545	list_for_each_entry(wh, &dev_cgroup->whitelist, list) {
	546	if (wh->type & DEV_ALL)
	547	goto acc_check;
	548	if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode))
	549	continue;
	550	if ((wh->type & DEV_CHAR) && !S_ISCHR(mode))
	551	continue;
	552	if (wh->major != ~0 && wh->major != MAJOR(dev))
	553	continue;
	554	if (wh->minor != ~0 && wh->minor != MINOR(dev))
	555	continue;
	556	acc_check:
	557	if (!(wh->access & ACC_MKNOD))
	558	continue;
4efd1a1b	559	rcu_read_unlock();
08ce5f16 SH	560	return 0;
08ce5f16 SH	561	}
36fd71d2	562
4efd1a1b	563	rcu_read_unlock();
36fd71d2	564
08ce5f16 SH	565	return -EPERM;
08ce5f16 SH	566	}