[people/arne_f/kernel.git] / security / device_cgroup.c

/*
 * device_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>
#include <linux/slab.h>
#include <linux/rcupdate.h>
#include <linux/mutex.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 */

static DEFINE_MUTEX(devcgroup_mutex);

/*
 * whitelist locking rules:
 * hold devcgroup_mutex for update/read.
 * hold rcu_read_lock() for read.
 */

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;
	bool deny_all;
};

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 *new_cgrp,
				struct cgroup_taskset *set)
{
	struct task_struct *task = cgroup_taskset_first(set);

	if (current != task && !capable(CAP_SYS_ADMIN))
		return -EPERM;
	return 0;
}

/*
 * called under devcgroup_mutex
 */
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 devcgroup_mutex
 */
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;

	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);
	return 0;
}

/*
 * called under devcgroup_mutex
 */
static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
			struct dev_whitelist_item *wh)
{
	struct dev_whitelist_item *walk, *tmp;

	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);
			kfree_rcu(walk, rcu);
		}
	}
}

/*
 * called from kernel/cgroup.c with cgroup_lock() held.
 */
static struct cgroup_subsys_state *devcgroup_create(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;
		dev_cgroup->deny_all = false;
		list_add(&wh->list, &dev_cgroup->whitelist);
	} else {
		parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
		mutex_lock(&devcgroup_mutex);
		ret = dev_whitelist_copy(&dev_cgroup->whitelist,
				&parent_dev_cgroup->whitelist);
		dev_cgroup->deny_all = parent_dev_cgroup->deny_all;
		mutex_unlock(&devcgroup_mutex);
		if (ret) {
			kfree(dev_cgroup);
			return ERR_PTR(ret);
		}
	}

	return &dev_cgroup->css;
}

static void devcgroup_destroy(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 devcgroup_mutex 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;

	if (!pcg)
		return 1;
	parent = cgroup_to_devcgroup(pcg);
	return may_access_whitelist(parent, wh);
}

/*
 * 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)
{
	const char *b;
	char *endp;
	int count;
	struct dev_whitelist_item wh;

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

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

static int devcgroup_access_write(struct cgroup *cgrp, struct cftype *cft,
				  const char *buffer)
{
	int retval;

	mutex_lock(&devcgroup_mutex);
	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
					 cft->private, buffer);
	mutex_unlock(&devcgroup_mutex);
	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,
	},
	{ }	/* terminate */
};

struct cgroup_subsys devices_subsys = {
	.name = "devices",
	.can_attach = devcgroup_can_attach,
	.create = devcgroup_create,
	.destroy = devcgroup_destroy,
	.subsys_id = devices_subsys_id,
	.base_cftypes = dev_cgroup_files,

	/*
	 * While devices cgroup has the rudimentary hierarchy support which
	 * checks the parent's restriction, it doesn't properly propagates
	 * config changes in ancestors to their descendents.  A child
	 * should only be allowed to add more restrictions to the parent's
	 * configuration.  Fix it and remove the following.
	 */
	.broken_hierarchy = true,
};

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

	rcu_read_lock();

	dev_cgroup = task_devcgroup(current);

	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
		if (wh->type & DEV_ALL)
			goto found;
		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;

		if ((mask & MAY_WRITE) && !(wh->access & ACC_WRITE))
			continue;
		if ((mask & MAY_READ) && !(wh->access & ACC_READ))
			continue;
found:
		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;

	if (!S_ISBLK(mode) && !S_ISCHR(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 found;
		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;

		if (!(wh->access & ACC_MKNOD))
			continue;
found:
		rcu_read_unlock();
		return 0;
	}

	rcu_read_unlock();

	return -EPERM;
}
Commit	Line	Data
08ce5f16	1	/*
47c59803	2	* device_cgroup.c - device cgroup subsystem
08ce5f16 SH	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>
5a0e3ad6	13	#include <linux/slab.h>
47c59803	14	#include <linux/rcupdate.h>
b4046f00	15	#include <linux/mutex.h>
08ce5f16 SH	16
	17	#define ACC_MKNOD 1
	18	#define ACC_READ 2
	19	#define ACC_WRITE 4
	20	#define ACC_MASK (ACC_MKNOD \| ACC_READ \| ACC_WRITE)
	21
	22	#define DEV_BLOCK 1
	23	#define DEV_CHAR 2
	24	#define DEV_ALL 4 /* this represents all devices */
	25
b4046f00 LZ	26	static DEFINE_MUTEX(devcgroup_mutex);
b4046f00 LZ	27
08ce5f16 SH	28	/*
08ce5f16 SH	29	* whitelist locking rules:
b4046f00	30	* hold devcgroup_mutex for update/read.
47c59803	31	* hold rcu_read_lock() for read.
08ce5f16 SH	32	*/
	33
	34	struct dev_whitelist_item {
	35	u32 major, minor;
	36	short type;
	37	short access;
	38	struct list_head list;
4efd1a1b	39	struct rcu_head rcu;
08ce5f16 SH	40	};
	41
	42	struct dev_cgroup {
	43	struct cgroup_subsys_state css;
	44	struct list_head whitelist;
66b8ef67	45	bool deny_all;
08ce5f16 SH	46	};
08ce5f16 SH	47
b66862f7 PE	48	static inline struct dev_cgroup css_to_devcgroup(struct cgroup_subsys_state s)
	49	{
	50	return container_of(s, struct dev_cgroup, css);
	51	}
	52
08ce5f16 SH	53	static inline struct dev_cgroup cgroup_to_devcgroup(struct cgroup cgroup)
08ce5f16 SH	54	{
b66862f7	55	return css_to_devcgroup(cgroup_subsys_state(cgroup, devices_subsys_id));
08ce5f16 SH	56	}
08ce5f16 SH	57
f92523e3 PM	58	static inline struct dev_cgroup task_devcgroup(struct task_struct task)
	59	{
	60	return css_to_devcgroup(task_subsys_state(task, devices_subsys_id));
	61	}
	62
08ce5f16 SH	63	struct cgroup_subsys devices_subsys;
08ce5f16 SH	64
761b3ef5 LZ	65	static int devcgroup_can_attach(struct cgroup *new_cgrp,
761b3ef5 LZ	66	struct cgroup_taskset *set)
08ce5f16	67	{
2f7ee569	68	struct task_struct *task = cgroup_taskset_first(set);
08ce5f16	69
2f7ee569 TH	70	if (current != task && !capable(CAP_SYS_ADMIN))
2f7ee569 TH	71	return -EPERM;
08ce5f16 SH	72	return 0;
	73	}
	74
	75	/*
b4046f00	76	* called under devcgroup_mutex
08ce5f16 SH	77	*/
	78	static int dev_whitelist_copy(struct list_head dest, struct list_head orig)
	79	{
	80	struct dev_whitelist_item wh, tmp, *new;
	81
	82	list_for_each_entry(wh, orig, list) {
2cdc7241	83	new = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
08ce5f16 SH	84	if (!new)
08ce5f16 SH	85	goto free_and_exit;
08ce5f16 SH	86	list_add_tail(&new->list, dest);
	87	}
	88
	89	return 0;
	90
	91	free_and_exit:
	92	list_for_each_entry_safe(wh, tmp, dest, list) {
	93	list_del(&wh->list);
	94	kfree(wh);
	95	}
	96	return -ENOMEM;
	97	}
	98
	99	/* Stupid prototype - don't bother combining existing entries */
	100	/*
b4046f00	101	* called under devcgroup_mutex
08ce5f16 SH	102	*/
	103	static int dev_whitelist_add(struct dev_cgroup *dev_cgroup,
	104	struct dev_whitelist_item *wh)
	105	{
d1ee2971	106	struct dev_whitelist_item whcopy, walk;
08ce5f16	107
2cdc7241	108	whcopy = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
08ce5f16 SH	109	if (!whcopy)
	110	return -ENOMEM;
	111
d1ee2971 PE	112	list_for_each_entry(walk, &dev_cgroup->whitelist, list) {
	113	if (walk->type != wh->type)
	114	continue;
	115	if (walk->major != wh->major)
	116	continue;
	117	if (walk->minor != wh->minor)
	118	continue;
	119
	120	walk->access \|= wh->access;
	121	kfree(whcopy);
	122	whcopy = NULL;
	123	}
	124
	125	if (whcopy != NULL)
4efd1a1b	126	list_add_tail_rcu(&whcopy->list, &dev_cgroup->whitelist);
08ce5f16 SH	127	return 0;
	128	}
	129
	130	/*
b4046f00	131	* called under devcgroup_mutex
08ce5f16 SH	132	*/
	133	static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
	134	struct dev_whitelist_item *wh)
	135	{
	136	struct dev_whitelist_item walk, tmp;
	137
08ce5f16 SH	138	list_for_each_entry_safe(walk, tmp, &dev_cgroup->whitelist, list) {
	139	if (walk->type == DEV_ALL)
	140	goto remove;
	141	if (walk->type != wh->type)
	142	continue;
	143	if (walk->major != ~0 && walk->major != wh->major)
	144	continue;
	145	if (walk->minor != ~0 && walk->minor != wh->minor)
	146	continue;
	147
	148	remove:
	149	walk->access &= ~wh->access;
	150	if (!walk->access) {
4efd1a1b	151	list_del_rcu(&walk->list);
6034f7e6	152	kfree_rcu(walk, rcu);
08ce5f16 SH	153	}
08ce5f16 SH	154	}
08ce5f16 SH	155	}
	156
	157	/*
	158	* called from kernel/cgroup.c with cgroup_lock() held.
	159	*/
761b3ef5	160	static struct cgroup_subsys_state devcgroup_create(struct cgroup cgroup)
08ce5f16 SH	161	{
	162	struct dev_cgroup dev_cgroup, parent_dev_cgroup;
	163	struct cgroup *parent_cgroup;
	164	int ret;
	165
	166	dev_cgroup = kzalloc(sizeof(*dev_cgroup), GFP_KERNEL);
	167	if (!dev_cgroup)
	168	return ERR_PTR(-ENOMEM);
	169	INIT_LIST_HEAD(&dev_cgroup->whitelist);
	170	parent_cgroup = cgroup->parent;
	171
	172	if (parent_cgroup == NULL) {
	173	struct dev_whitelist_item *wh;
	174	wh = kmalloc(sizeof(*wh), GFP_KERNEL);
	175	if (!wh) {
	176	kfree(dev_cgroup);
	177	return ERR_PTR(-ENOMEM);
	178	}
	179	wh->minor = wh->major = ~0;
	180	wh->type = DEV_ALL;
7759fc9d	181	wh->access = ACC_MASK;
66b8ef67	182	dev_cgroup->deny_all = false;
08ce5f16 SH	183	list_add(&wh->list, &dev_cgroup->whitelist);
	184	} else {
	185	parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
b4046f00	186	mutex_lock(&devcgroup_mutex);
08ce5f16 SH	187	ret = dev_whitelist_copy(&dev_cgroup->whitelist,
08ce5f16 SH	188	&parent_dev_cgroup->whitelist);
66b8ef67	189	dev_cgroup->deny_all = parent_dev_cgroup->deny_all;
b4046f00	190	mutex_unlock(&devcgroup_mutex);
08ce5f16 SH	191	if (ret) {
	192	kfree(dev_cgroup);
	193	return ERR_PTR(ret);
	194	}
	195	}
	196
08ce5f16 SH	197	return &dev_cgroup->css;
	198	}
	199
761b3ef5	200	static void devcgroup_destroy(struct cgroup *cgroup)
08ce5f16 SH	201	{
	202	struct dev_cgroup *dev_cgroup;
	203	struct dev_whitelist_item wh, tmp;
	204
	205	dev_cgroup = cgroup_to_devcgroup(cgroup);
	206	list_for_each_entry_safe(wh, tmp, &dev_cgroup->whitelist, list) {
	207	list_del(&wh->list);
	208	kfree(wh);
	209	}
	210	kfree(dev_cgroup);
	211	}
	212
	213	#define DEVCG_ALLOW 1
	214	#define DEVCG_DENY 2
29486df3 SH	215	#define DEVCG_LIST 3
29486df3 SH	216
17d213f8	217	#define MAJMINLEN 13
29486df3	218	#define ACCLEN 4
08ce5f16 SH	219
	220	static void set_access(char *acc, short access)
	221	{
	222	int idx = 0;
29486df3	223	memset(acc, 0, ACCLEN);
08ce5f16 SH	224	if (access & ACC_READ)
	225	acc[idx++] = 'r';
	226	if (access & ACC_WRITE)
	227	acc[idx++] = 'w';
	228	if (access & ACC_MKNOD)
	229	acc[idx++] = 'm';
	230	}
	231
	232	static char type_to_char(short type)
	233	{
	234	if (type == DEV_ALL)
	235	return 'a';
	236	if (type == DEV_CHAR)
	237	return 'c';
	238	if (type == DEV_BLOCK)
	239	return 'b';
	240	return 'X';
	241	}
	242
29486df3	243	static void set_majmin(char *str, unsigned m)
08ce5f16	244	{
08ce5f16	245	if (m == ~0)
7759fc9d	246	strcpy(str, "*");
08ce5f16	247	else
7759fc9d	248	sprintf(str, "%u", m);
08ce5f16 SH	249	}
08ce5f16 SH	250
29486df3 SH	251	static int devcgroup_seq_read(struct cgroup cgroup, struct cftype cft,
29486df3 SH	252	struct seq_file *m)
08ce5f16	253	{
29486df3	254	struct dev_cgroup *devcgroup = cgroup_to_devcgroup(cgroup);
08ce5f16	255	struct dev_whitelist_item *wh;
29486df3	256	char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];
08ce5f16	257
4efd1a1b PE	258	rcu_read_lock();
4efd1a1b PE	259	list_for_each_entry_rcu(wh, &devcgroup->whitelist, list) {
08ce5f16	260	set_access(acc, wh->access);
29486df3 SH	261	set_majmin(maj, wh->major);
	262	set_majmin(min, wh->minor);
	263	seq_printf(m, "%c %s:%s %s\n", type_to_char(wh->type),
	264	maj, min, acc);
08ce5f16	265	}
4efd1a1b	266	rcu_read_unlock();
08ce5f16	267
29486df3	268	return 0;
08ce5f16 SH	269	}
	270
	271	/*
	272	* may_access_whitelist:
	273	* does the access granted to dev_cgroup c contain the access
	274	* requested in whitelist item refwh.
	275	* return 1 if yes, 0 if no.
b4046f00	276	* call with devcgroup_mutex held
08ce5f16 SH	277	*/
	278	static int may_access_whitelist(struct dev_cgroup *c,
	279	struct dev_whitelist_item *refwh)
	280	{
	281	struct dev_whitelist_item *whitem;
	282
	283	list_for_each_entry(whitem, &c->whitelist, list) {
	284	if (whitem->type & DEV_ALL)
	285	return 1;
	286	if ((refwh->type & DEV_BLOCK) && !(whitem->type & DEV_BLOCK))
	287	continue;
	288	if ((refwh->type & DEV_CHAR) && !(whitem->type & DEV_CHAR))
	289	continue;
	290	if (whitem->major != ~0 && whitem->major != refwh->major)
	291	continue;
	292	if (whitem->minor != ~0 && whitem->minor != refwh->minor)
	293	continue;
ec229e83	294	if (refwh->access & (~whitem->access))
08ce5f16 SH	295	continue;
	296	return 1;
	297	}
	298	return 0;
	299	}
	300
	301	/*
	302	* parent_has_perm:
	303	* when adding a new allow rule to a device whitelist, the rule
	304	* must be allowed in the parent device
	305	*/
f92523e3	306	static int parent_has_perm(struct dev_cgroup *childcg,
08ce5f16 SH	307	struct dev_whitelist_item *wh)
08ce5f16 SH	308	{
f92523e3	309	struct cgroup *pcg = childcg->css.cgroup->parent;
08ce5f16	310	struct dev_cgroup *parent;
08ce5f16 SH	311
	312	if (!pcg)
	313	return 1;
	314	parent = cgroup_to_devcgroup(pcg);
47c59803	315	return may_access_whitelist(parent, wh);
08ce5f16 SH	316	}
	317
	318	/*
	319	* Modify the whitelist using allow/deny rules.
	320	* CAP_SYS_ADMIN is needed for this. It's at least separate from CAP_MKNOD
	321	* so we can give a container CAP_MKNOD to let it create devices but not
	322	* modify the whitelist.
	323	* It seems likely we'll want to add a CAP_CONTAINER capability to allow
	324	* us to also grant CAP_SYS_ADMIN to containers without giving away the
	325	* device whitelist controls, but for now we'll stick with CAP_SYS_ADMIN
	326	*
	327	* Taking rules away is always allowed (given CAP_SYS_ADMIN). Granting
	328	* new access is only allowed if you're in the top-level cgroup, or your
	329	* parent cgroup has the access you're asking for.
	330	*/
f92523e3 PM	331	static int devcgroup_update_access(struct dev_cgroup *devcgroup,
f92523e3 PM	332	int filetype, const char *buffer)
08ce5f16	333	{
f92523e3	334	const char *b;
7759fc9d	335	char *endp;
c012a54a	336	int count;
08ce5f16 SH	337	struct dev_whitelist_item wh;
	338
	339	if (!capable(CAP_SYS_ADMIN))
	340	return -EPERM;
	341
08ce5f16 SH	342	memset(&wh, 0, sizeof(wh));
	343	b = buffer;
	344
	345	switch (*b) {
	346	case 'a':
	347	wh.type = DEV_ALL;
	348	wh.access = ACC_MASK;
d823f6bf LZ	349	wh.major = ~0;
d823f6bf LZ	350	wh.minor = ~0;
08ce5f16 SH	351	goto handle;
	352	case 'b':
	353	wh.type = DEV_BLOCK;
	354	break;
	355	case 'c':
	356	wh.type = DEV_CHAR;
	357	break;
	358	default:
f92523e3	359	return -EINVAL;
08ce5f16 SH	360	}
08ce5f16 SH	361	b++;
f92523e3 PM	362	if (!isspace(*b))
f92523e3 PM	363	return -EINVAL;
08ce5f16 SH	364	b++;
	365	if (b == '') {
	366	wh.major = ~0;
	367	b++;
	368	} else if (isdigit(*b)) {
7759fc9d LZ	369	wh.major = simple_strtoul(b, &endp, 10);
7759fc9d LZ	370	b = endp;
08ce5f16	371	} else {
f92523e3	372	return -EINVAL;
08ce5f16	373	}
f92523e3 PM	374	if (*b != ':')
f92523e3 PM	375	return -EINVAL;
08ce5f16 SH	376	b++;
	377
	378	/* read minor */
	379	if (b == '') {
	380	wh.minor = ~0;
	381	b++;
	382	} else if (isdigit(*b)) {
7759fc9d LZ	383	wh.minor = simple_strtoul(b, &endp, 10);
7759fc9d LZ	384	b = endp;
08ce5f16	385	} else {
f92523e3	386	return -EINVAL;
08ce5f16	387	}
f92523e3 PM	388	if (!isspace(*b))
f92523e3 PM	389	return -EINVAL;
08ce5f16 SH	390	for (b++, count = 0; count < 3; count++, b++) {
	391	switch (*b) {
	392	case 'r':
	393	wh.access \|= ACC_READ;
	394	break;
	395	case 'w':
	396	wh.access \|= ACC_WRITE;
	397	break;
	398	case 'm':
	399	wh.access \|= ACC_MKNOD;
	400	break;
	401	case '\n':
	402	case '\0':
	403	count = 3;
	404	break;
	405	default:
f92523e3	406	return -EINVAL;
08ce5f16 SH	407	}
	408	}
	409
	410	handle:
08ce5f16 SH	411	switch (filetype) {
08ce5f16 SH	412	case DEVCG_ALLOW:
f92523e3 PM	413	if (!parent_has_perm(devcgroup, &wh))
f92523e3 PM	414	return -EPERM;
66b8ef67	415	devcgroup->deny_all = false;
f92523e3	416	return dev_whitelist_add(devcgroup, &wh);
08ce5f16 SH	417	case DEVCG_DENY:
08ce5f16 SH	418	dev_whitelist_rm(devcgroup, &wh);
66b8ef67	419	devcgroup->deny_all = true;
08ce5f16 SH	420	break;
08ce5f16 SH	421	default:
f92523e3	422	return -EINVAL;
08ce5f16	423	}
f92523e3 PM	424	return 0;
f92523e3 PM	425	}
08ce5f16	426
f92523e3 PM	427	static int devcgroup_access_write(struct cgroup cgrp, struct cftype cft,
	428	const char *buffer)
	429	{
	430	int retval;
b4046f00 LZ	431
b4046f00 LZ	432	mutex_lock(&devcgroup_mutex);
f92523e3 PM	433	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
f92523e3 PM	434	cft->private, buffer);
b4046f00	435	mutex_unlock(&devcgroup_mutex);
08ce5f16 SH	436	return retval;
	437	}
	438
	439	static struct cftype dev_cgroup_files[] = {
	440	{
	441	.name = "allow",
f92523e3	442	.write_string = devcgroup_access_write,
08ce5f16 SH	443	.private = DEVCG_ALLOW,
	444	},
	445	{
	446	.name = "deny",
f92523e3	447	.write_string = devcgroup_access_write,
08ce5f16 SH	448	.private = DEVCG_DENY,
08ce5f16 SH	449	},
29486df3 SH	450	{
	451	.name = "list",
	452	.read_seq_string = devcgroup_seq_read,
	453	.private = DEVCG_LIST,
	454	},
4baf6e33	455	{ } /* terminate */
08ce5f16 SH	456	};
08ce5f16 SH	457
08ce5f16 SH	458	struct cgroup_subsys devices_subsys = {
	459	.name = "devices",
	460	.can_attach = devcgroup_can_attach,
	461	.create = devcgroup_create,
c5b60b5e	462	.destroy = devcgroup_destroy,
08ce5f16	463	.subsys_id = devices_subsys_id,
4baf6e33	464	.base_cftypes = dev_cgroup_files,
8c7f6edb TH	465
	466	/*
	467	* While devices cgroup has the rudimentary hierarchy support which
	468	* checks the parent's restriction, it doesn't properly propagates
	469	* config changes in ancestors to their descendents. A child
	470	* should only be allowed to add more restrictions to the parent's
	471	* configuration. Fix it and remove the following.
	472	*/
	473	.broken_hierarchy = true,
08ce5f16 SH	474	};
08ce5f16 SH	475
482e0cd3	476	int __devcgroup_inode_permission(struct inode *inode, int mask)
08ce5f16	477	{
08ce5f16 SH	478	struct dev_cgroup *dev_cgroup;
	479	struct dev_whitelist_item *wh;
	480
4efd1a1b	481	rcu_read_lock();
36fd71d2 LZ	482
	483	dev_cgroup = task_devcgroup(current);
	484
4efd1a1b	485	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
08ce5f16	486	if (wh->type & DEV_ALL)
cd500819	487	goto found;
08ce5f16 SH	488	if ((wh->type & DEV_BLOCK) && !S_ISBLK(inode->i_mode))
	489	continue;
	490	if ((wh->type & DEV_CHAR) && !S_ISCHR(inode->i_mode))
	491	continue;
	492	if (wh->major != ~0 && wh->major != imajor(inode))
	493	continue;
	494	if (wh->minor != ~0 && wh->minor != iminor(inode))
	495	continue;
cd500819	496
08ce5f16 SH	497	if ((mask & MAY_WRITE) && !(wh->access & ACC_WRITE))
	498	continue;
	499	if ((mask & MAY_READ) && !(wh->access & ACC_READ))
	500	continue;
cd500819	501	found:
4efd1a1b	502	rcu_read_unlock();
08ce5f16 SH	503	return 0;
08ce5f16 SH	504	}
36fd71d2	505
4efd1a1b	506	rcu_read_unlock();
08ce5f16 SH	507
	508	return -EPERM;
	509	}
	510
	511	int devcgroup_inode_mknod(int mode, dev_t dev)
	512	{
08ce5f16 SH	513	struct dev_cgroup *dev_cgroup;
	514	struct dev_whitelist_item *wh;
	515
0b82ac37 SH	516	if (!S_ISBLK(mode) && !S_ISCHR(mode))
	517	return 0;
	518
4efd1a1b	519	rcu_read_lock();
36fd71d2 LZ	520
	521	dev_cgroup = task_devcgroup(current);
	522
116e0575	523	list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) {
08ce5f16	524	if (wh->type & DEV_ALL)
cd500819	525	goto found;
08ce5f16 SH	526	if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode))
	527	continue;
	528	if ((wh->type & DEV_CHAR) && !S_ISCHR(mode))
	529	continue;
	530	if (wh->major != ~0 && wh->major != MAJOR(dev))
	531	continue;
	532	if (wh->minor != ~0 && wh->minor != MINOR(dev))
	533	continue;
cd500819	534
08ce5f16 SH	535	if (!(wh->access & ACC_MKNOD))
08ce5f16 SH	536	continue;
cd500819	537	found:
4efd1a1b	538	rcu_read_unlock();
08ce5f16 SH	539	return 0;
08ce5f16 SH	540	}
36fd71d2	541
4efd1a1b	542	rcu_read_unlock();
36fd71d2	543
08ce5f16 SH	544	return -EPERM;
08ce5f16 SH	545	}