[mirror_ubuntu-artful-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;
}

/*
 * 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 != wh->type)
			continue;
		if (walk->major != wh->major)
			continue;
		if (walk->minor != wh->minor)
			continue;

		walk->access &= ~wh->access;
		if (!walk->access) {
			list_del_rcu(&walk->list);
			kfree_rcu(walk, rcu);
		}
	}
}

/**
 * dev_whitelist_clean - frees all entries of the whitelist
 * @dev_cgroup: dev_cgroup with the whitelist to be cleaned
 *
 * called under devcgroup_mutex
 */
static void dev_whitelist_clean(struct dev_cgroup *dev_cgroup)
{
	struct dev_whitelist_item *wh, *tmp;

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

/*
 * 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)
		dev_cgroup->deny_all = false;
	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;

	dev_cgroup = cgroup_to_devcgroup(cgroup);
	dev_whitelist_clean(dev_cgroup);
	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();
	/*
	 * To preserve the compatibility:
	 * - Only show the "all devices" when the default policy is to allow
	 * - List the exceptions in case the default policy is to deny
	 * This way, the file remains as a "whitelist of devices"
	 */
	if (devcgroup->deny_all == false) {
		set_access(acc, ACC_MASK);
		set_majmin(maj, ~0);
		set_majmin(min, ~0);
		seq_printf(m, "%c %s:%s %s\n", type_to_char(DEV_ALL),
			   maj, min, acc);
	} else {
		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 - verifies if a new rule is part of what is allowed
 *			  by a dev cgroup based on the default policy +
 *			  exceptions. This is used to make sure a child cgroup
 *			  won't have more privileges than its parent or to
 *			  verify if a certain access is allowed.
 * @dev_cgroup: dev cgroup to be tested against
 * @refwh: new rule
 */
static int may_access_whitelist(struct dev_cgroup *dev_cgroup,
			        struct dev_whitelist_item *refwh)
{
	struct dev_whitelist_item *whitem;
	bool match = false;

	list_for_each_entry(whitem, &dev_cgroup->whitelist, list) {
		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;
		match = true;
		break;
	}

	/*
	 * In two cases we'll consider this new rule valid:
	 * - the dev cgroup has its default policy to allow + exception list:
	 *   the new rule should *not* match any of the exceptions
	 *   (!deny_all, !match)
	 * - the dev cgroup has its default policy to deny + exception list:
	 *   the new rule *should* match the exceptions
	 *   (deny_all, match)
	 */
	if (dev_cgroup->deny_all == match)
		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':
		switch (filetype) {
		case DEVCG_ALLOW:
			if (!parent_has_perm(devcgroup, &wh))
				return -EPERM;
			dev_whitelist_clean(devcgroup);
			devcgroup->deny_all = false;
			break;
		case DEVCG_DENY:
			dev_whitelist_clean(devcgroup);
			devcgroup->deny_all = true;
			break;
		default:
			return -EINVAL;
		}
		return 0;
	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;
		}
	}

	switch (filetype) {
	case DEVCG_ALLOW:
		if (!parent_has_perm(devcgroup, &wh))
			return -EPERM;
		/*
		 * If the default policy is to allow by default, try to remove
		 * an matching exception instead. And be silent about it: we
		 * don't want to break compatibility
		 */
		if (devcgroup->deny_all == false) {
			dev_whitelist_rm(devcgroup, &wh);
			return 0;
		}
		return dev_whitelist_add(devcgroup, &wh);
	case DEVCG_DENY:
		/*
		 * If the default policy is to deny by default, try to remove
		 * an matching exception instead. And be silent about it: we
		 * don't want to break compatibility
		 */
		if (devcgroup->deny_all == true) {
			dev_whitelist_rm(devcgroup, &wh);
			return 0;
		}
		return dev_whitelist_add(devcgroup, &wh);
	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,
};

/**
 * __devcgroup_check_permission - checks if an inode operation is permitted
 * @dev_cgroup: the dev cgroup to be tested against
 * @type: device type
 * @major: device major number
 * @minor: device minor number
 * @access: combination of ACC_WRITE, ACC_READ and ACC_MKNOD
 *
 * returns 0 on success, -EPERM case the operation is not permitted
 */
static int __devcgroup_check_permission(struct dev_cgroup *dev_cgroup,
					short type, u32 major, u32 minor,
				        short access)
{
	struct dev_whitelist_item wh;
	int rc;

	memset(&wh, 0, sizeof(wh));
	wh.type = type;
	wh.major = major;
	wh.minor = minor;
	wh.access = access;

	rcu_read_lock();
	rc = may_access_whitelist(dev_cgroup, &wh);
	rcu_read_unlock();

	if (!rc)
		return -EPERM;

	return 0;
}

int __devcgroup_inode_permission(struct inode *inode, int mask)
{
	struct dev_cgroup *dev_cgroup = task_devcgroup(current);
	short type, access = 0;

	if (S_ISBLK(inode->i_mode))
		type = DEV_BLOCK;
	if (S_ISCHR(inode->i_mode))
		type = DEV_CHAR;
	if (mask & MAY_WRITE)
		access |= ACC_WRITE;
	if (mask & MAY_READ)
		access |= ACC_READ;

	return __devcgroup_check_permission(dev_cgroup, type, imajor(inode),
					    iminor(inode), access);
}

int devcgroup_inode_mknod(int mode, dev_t dev)
{
	struct dev_cgroup *dev_cgroup = task_devcgroup(current);
	short type;

	if (!S_ISBLK(mode) && !S_ISCHR(mode))
		return 0;

	if (S_ISBLK(mode))
		type = DEV_BLOCK;
	else
		type = DEV_CHAR;

	return __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev),
					    MINOR(dev), ACC_MKNOD);

}
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
08ce5f16	99	/*
b4046f00	100	* called under devcgroup_mutex
08ce5f16 SH	101	*/
	102	static int dev_whitelist_add(struct dev_cgroup *dev_cgroup,
	103	struct dev_whitelist_item *wh)
	104	{
d1ee2971	105	struct dev_whitelist_item whcopy, walk;
08ce5f16	106
2cdc7241	107	whcopy = kmemdup(wh, sizeof(*wh), GFP_KERNEL);
08ce5f16 SH	108	if (!whcopy)
	109	return -ENOMEM;
	110
d1ee2971 PE	111	list_for_each_entry(walk, &dev_cgroup->whitelist, list) {
	112	if (walk->type != wh->type)
	113	continue;
	114	if (walk->major != wh->major)
	115	continue;
	116	if (walk->minor != wh->minor)
	117	continue;
	118
	119	walk->access \|= wh->access;
	120	kfree(whcopy);
	121	whcopy = NULL;
	122	}
	123
	124	if (whcopy != NULL)
4efd1a1b	125	list_add_tail_rcu(&whcopy->list, &dev_cgroup->whitelist);
08ce5f16 SH	126	return 0;
	127	}
	128
	129	/*
b4046f00	130	* called under devcgroup_mutex
08ce5f16 SH	131	*/
	132	static void dev_whitelist_rm(struct dev_cgroup *dev_cgroup,
	133	struct dev_whitelist_item *wh)
	134	{
	135	struct dev_whitelist_item walk, tmp;
	136
08ce5f16	137	list_for_each_entry_safe(walk, tmp, &dev_cgroup->whitelist, list) {
08ce5f16 SH	138	if (walk->type != wh->type)
08ce5f16 SH	139	continue;
ad676077	140	if (walk->major != wh->major)
08ce5f16	141	continue;
ad676077	142	if (walk->minor != wh->minor)
08ce5f16 SH	143	continue;
08ce5f16 SH	144
08ce5f16 SH	145	walk->access &= ~wh->access;
08ce5f16 SH	146	if (!walk->access) {
4efd1a1b	147	list_del_rcu(&walk->list);
6034f7e6	148	kfree_rcu(walk, rcu);
08ce5f16 SH	149	}
08ce5f16 SH	150	}
08ce5f16 SH	151	}
08ce5f16 SH	152
868539a3 AR	153	/**
	154	* dev_whitelist_clean - frees all entries of the whitelist
	155	* @dev_cgroup: dev_cgroup with the whitelist to be cleaned
	156	*
	157	* called under devcgroup_mutex
	158	*/
	159	static void dev_whitelist_clean(struct dev_cgroup *dev_cgroup)
	160	{
	161	struct dev_whitelist_item wh, tmp;
	162
	163	list_for_each_entry_safe(wh, tmp, &dev_cgroup->whitelist, list) {
	164	list_del(&wh->list);
	165	kfree(wh);
	166	}
	167	}
	168
08ce5f16 SH	169	/*
	170	* called from kernel/cgroup.c with cgroup_lock() held.
	171	*/
761b3ef5	172	static struct cgroup_subsys_state devcgroup_create(struct cgroup cgroup)
08ce5f16 SH	173	{
	174	struct dev_cgroup dev_cgroup, parent_dev_cgroup;
	175	struct cgroup *parent_cgroup;
	176	int ret;
	177
	178	dev_cgroup = kzalloc(sizeof(*dev_cgroup), GFP_KERNEL);
	179	if (!dev_cgroup)
	180	return ERR_PTR(-ENOMEM);
	181	INIT_LIST_HEAD(&dev_cgroup->whitelist);
	182	parent_cgroup = cgroup->parent;
	183
ad676077	184	if (parent_cgroup == NULL)
66b8ef67	185	dev_cgroup->deny_all = false;
ad676077	186	else {
08ce5f16	187	parent_dev_cgroup = cgroup_to_devcgroup(parent_cgroup);
b4046f00	188	mutex_lock(&devcgroup_mutex);
08ce5f16 SH	189	ret = dev_whitelist_copy(&dev_cgroup->whitelist,
08ce5f16 SH	190	&parent_dev_cgroup->whitelist);
66b8ef67	191	dev_cgroup->deny_all = parent_dev_cgroup->deny_all;
b4046f00	192	mutex_unlock(&devcgroup_mutex);
08ce5f16 SH	193	if (ret) {
	194	kfree(dev_cgroup);
	195	return ERR_PTR(ret);
	196	}
	197	}
	198
08ce5f16 SH	199	return &dev_cgroup->css;
	200	}
	201
761b3ef5	202	static void devcgroup_destroy(struct cgroup *cgroup)
08ce5f16 SH	203	{
08ce5f16 SH	204	struct dev_cgroup *dev_cgroup;
08ce5f16 SH	205
08ce5f16 SH	206	dev_cgroup = cgroup_to_devcgroup(cgroup);
868539a3	207	dev_whitelist_clean(dev_cgroup);
08ce5f16 SH	208	kfree(dev_cgroup);
	209	}
	210
	211	#define DEVCG_ALLOW 1
	212	#define DEVCG_DENY 2
29486df3 SH	213	#define DEVCG_LIST 3
29486df3 SH	214
17d213f8	215	#define MAJMINLEN 13
29486df3	216	#define ACCLEN 4
08ce5f16 SH	217
	218	static void set_access(char *acc, short access)
	219	{
	220	int idx = 0;
29486df3	221	memset(acc, 0, ACCLEN);
08ce5f16 SH	222	if (access & ACC_READ)
	223	acc[idx++] = 'r';
	224	if (access & ACC_WRITE)
	225	acc[idx++] = 'w';
	226	if (access & ACC_MKNOD)
	227	acc[idx++] = 'm';
	228	}
	229
	230	static char type_to_char(short type)
	231	{
	232	if (type == DEV_ALL)
	233	return 'a';
	234	if (type == DEV_CHAR)
	235	return 'c';
	236	if (type == DEV_BLOCK)
	237	return 'b';
	238	return 'X';
	239	}
	240
29486df3	241	static void set_majmin(char *str, unsigned m)
08ce5f16	242	{
08ce5f16	243	if (m == ~0)
7759fc9d	244	strcpy(str, "*");
08ce5f16	245	else
7759fc9d	246	sprintf(str, "%u", m);
08ce5f16 SH	247	}
08ce5f16 SH	248
29486df3 SH	249	static int devcgroup_seq_read(struct cgroup cgroup, struct cftype cft,
29486df3 SH	250	struct seq_file *m)
08ce5f16	251	{
29486df3	252	struct dev_cgroup *devcgroup = cgroup_to_devcgroup(cgroup);
08ce5f16	253	struct dev_whitelist_item *wh;
29486df3	254	char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];
08ce5f16	255
4efd1a1b	256	rcu_read_lock();
ad676077 AR	257	/*
	258	* To preserve the compatibility:
	259	* - Only show the "all devices" when the default policy is to allow
	260	* - List the exceptions in case the default policy is to deny
	261	* This way, the file remains as a "whitelist of devices"
	262	*/
	263	if (devcgroup->deny_all == false) {
	264	set_access(acc, ACC_MASK);
	265	set_majmin(maj, ~0);
	266	set_majmin(min, ~0);
	267	seq_printf(m, "%c %s:%s %s\n", type_to_char(DEV_ALL),
29486df3	268	maj, min, acc);
ad676077 AR	269	} else {
	270	list_for_each_entry_rcu(wh, &devcgroup->whitelist, list) {
	271	set_access(acc, wh->access);
	272	set_majmin(maj, wh->major);
	273	set_majmin(min, wh->minor);
	274	seq_printf(m, "%c %s:%s %s\n", type_to_char(wh->type),
	275	maj, min, acc);
	276	}
08ce5f16	277	}
4efd1a1b	278	rcu_read_unlock();
08ce5f16	279
29486df3	280	return 0;
08ce5f16 SH	281	}
08ce5f16 SH	282
ad676077 AR	283	/**
	284	* may_access_whitelist - verifies if a new rule is part of what is allowed
	285	* by a dev cgroup based on the default policy +
	286	* exceptions. This is used to make sure a child cgroup
	287	* won't have more privileges than its parent or to
	288	* verify if a certain access is allowed.
	289	* @dev_cgroup: dev cgroup to be tested against
	290	* @refwh: new rule
08ce5f16	291	*/
ad676077 AR	292	static int may_access_whitelist(struct dev_cgroup *dev_cgroup,
ad676077 AR	293	struct dev_whitelist_item *refwh)
08ce5f16 SH	294	{
08ce5f16 SH	295	struct dev_whitelist_item *whitem;
ad676077	296	bool match = false;
08ce5f16	297
ad676077	298	list_for_each_entry(whitem, &dev_cgroup->whitelist, list) {
08ce5f16 SH	299	if ((refwh->type & DEV_BLOCK) && !(whitem->type & DEV_BLOCK))
	300	continue;
	301	if ((refwh->type & DEV_CHAR) && !(whitem->type & DEV_CHAR))
	302	continue;
	303	if (whitem->major != ~0 && whitem->major != refwh->major)
	304	continue;
	305	if (whitem->minor != ~0 && whitem->minor != refwh->minor)
	306	continue;
ec229e83	307	if (refwh->access & (~whitem->access))
08ce5f16	308	continue;
ad676077 AR	309	match = true;
ad676077 AR	310	break;
08ce5f16	311	}
ad676077 AR	312
	313	/*
	314	* In two cases we'll consider this new rule valid:
	315	* - the dev cgroup has its default policy to allow + exception list:
	316	* the new rule should not match any of the exceptions
	317	* (!deny_all, !match)
	318	* - the dev cgroup has its default policy to deny + exception list:
	319	* the new rule should match the exceptions
	320	* (deny_all, match)
	321	*/
	322	if (dev_cgroup->deny_all == match)
	323	return 1;
08ce5f16 SH	324	return 0;
	325	}
	326
	327	/*
	328	* parent_has_perm:
	329	* when adding a new allow rule to a device whitelist, the rule
	330	* must be allowed in the parent device
	331	*/
f92523e3	332	static int parent_has_perm(struct dev_cgroup *childcg,
08ce5f16 SH	333	struct dev_whitelist_item *wh)
08ce5f16 SH	334	{
f92523e3	335	struct cgroup *pcg = childcg->css.cgroup->parent;
08ce5f16	336	struct dev_cgroup *parent;
08ce5f16 SH	337
	338	if (!pcg)
	339	return 1;
	340	parent = cgroup_to_devcgroup(pcg);
47c59803	341	return may_access_whitelist(parent, wh);
08ce5f16 SH	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	360	const char *b;
7759fc9d	361	char *endp;
c012a54a	362	int count;
08ce5f16 SH	363	struct dev_whitelist_item wh;
	364
	365	if (!capable(CAP_SYS_ADMIN))
	366	return -EPERM;
	367
08ce5f16 SH	368	memset(&wh, 0, sizeof(wh));
	369	b = buffer;
	370
	371	switch (*b) {
	372	case 'a':
ad676077 AR	373	switch (filetype) {
	374	case DEVCG_ALLOW:
	375	if (!parent_has_perm(devcgroup, &wh))
	376	return -EPERM;
	377	dev_whitelist_clean(devcgroup);
	378	devcgroup->deny_all = false;
	379	break;
	380	case DEVCG_DENY:
	381	dev_whitelist_clean(devcgroup);
	382	devcgroup->deny_all = true;
	383	break;
	384	default:
	385	return -EINVAL;
	386	}
	387	return 0;
08ce5f16 SH	388	case 'b':
	389	wh.type = DEV_BLOCK;
	390	break;
	391	case 'c':
	392	wh.type = DEV_CHAR;
	393	break;
	394	default:
f92523e3	395	return -EINVAL;
08ce5f16 SH	396	}
08ce5f16 SH	397	b++;
f92523e3 PM	398	if (!isspace(*b))
f92523e3 PM	399	return -EINVAL;
08ce5f16 SH	400	b++;
	401	if (b == '') {
	402	wh.major = ~0;
	403	b++;
	404	} else if (isdigit(*b)) {
7759fc9d LZ	405	wh.major = simple_strtoul(b, &endp, 10);
7759fc9d LZ	406	b = endp;
08ce5f16	407	} else {
f92523e3	408	return -EINVAL;
08ce5f16	409	}
f92523e3 PM	410	if (*b != ':')
f92523e3 PM	411	return -EINVAL;
08ce5f16 SH	412	b++;
	413
	414	/* read minor */
	415	if (b == '') {
	416	wh.minor = ~0;
	417	b++;
	418	} else if (isdigit(*b)) {
7759fc9d LZ	419	wh.minor = simple_strtoul(b, &endp, 10);
7759fc9d LZ	420	b = endp;
08ce5f16	421	} else {
f92523e3	422	return -EINVAL;
08ce5f16	423	}
f92523e3 PM	424	if (!isspace(*b))
f92523e3 PM	425	return -EINVAL;
08ce5f16 SH	426	for (b++, count = 0; count < 3; count++, b++) {
	427	switch (*b) {
	428	case 'r':
	429	wh.access \|= ACC_READ;
	430	break;
	431	case 'w':
	432	wh.access \|= ACC_WRITE;
	433	break;
	434	case 'm':
	435	wh.access \|= ACC_MKNOD;
	436	break;
	437	case '\n':
	438	case '\0':
	439	count = 3;
	440	break;
	441	default:
f92523e3	442	return -EINVAL;
08ce5f16 SH	443	}
	444	}
	445
08ce5f16 SH	446	switch (filetype) {
08ce5f16 SH	447	case DEVCG_ALLOW:
f92523e3 PM	448	if (!parent_has_perm(devcgroup, &wh))
f92523e3 PM	449	return -EPERM;
ad676077 AR	450	/*
	451	* If the default policy is to allow by default, try to remove
	452	* an matching exception instead. And be silent about it: we
	453	* don't want to break compatibility
	454	*/
	455	if (devcgroup->deny_all == false) {
	456	dev_whitelist_rm(devcgroup, &wh);
	457	return 0;
	458	}
f92523e3	459	return dev_whitelist_add(devcgroup, &wh);
08ce5f16	460	case DEVCG_DENY:
ad676077 AR	461	/*
	462	* If the default policy is to deny by default, try to remove
	463	* an matching exception instead. And be silent about it: we
	464	* don't want to break compatibility
	465	*/
	466	if (devcgroup->deny_all == true) {
	467	dev_whitelist_rm(devcgroup, &wh);
	468	return 0;
	469	}
	470	return dev_whitelist_add(devcgroup, &wh);
08ce5f16	471	default:
f92523e3	472	return -EINVAL;
08ce5f16	473	}
f92523e3 PM	474	return 0;
f92523e3 PM	475	}
08ce5f16	476
f92523e3 PM	477	static int devcgroup_access_write(struct cgroup cgrp, struct cftype cft,
	478	const char *buffer)
	479	{
	480	int retval;
b4046f00 LZ	481
b4046f00 LZ	482	mutex_lock(&devcgroup_mutex);
f92523e3 PM	483	retval = devcgroup_update_access(cgroup_to_devcgroup(cgrp),
f92523e3 PM	484	cft->private, buffer);
b4046f00	485	mutex_unlock(&devcgroup_mutex);
08ce5f16 SH	486	return retval;
	487	}
	488
	489	static struct cftype dev_cgroup_files[] = {
	490	{
	491	.name = "allow",
f92523e3	492	.write_string = devcgroup_access_write,
08ce5f16 SH	493	.private = DEVCG_ALLOW,
	494	},
	495	{
	496	.name = "deny",
f92523e3	497	.write_string = devcgroup_access_write,
08ce5f16 SH	498	.private = DEVCG_DENY,
08ce5f16 SH	499	},
29486df3 SH	500	{
	501	.name = "list",
	502	.read_seq_string = devcgroup_seq_read,
	503	.private = DEVCG_LIST,
	504	},
4baf6e33	505	{ } /* terminate */
08ce5f16 SH	506	};
08ce5f16 SH	507
08ce5f16 SH	508	struct cgroup_subsys devices_subsys = {
	509	.name = "devices",
	510	.can_attach = devcgroup_can_attach,
	511	.create = devcgroup_create,
c5b60b5e	512	.destroy = devcgroup_destroy,
08ce5f16	513	.subsys_id = devices_subsys_id,
4baf6e33	514	.base_cftypes = dev_cgroup_files,
8c7f6edb TH	515
	516	/*
	517	* While devices cgroup has the rudimentary hierarchy support which
	518	* checks the parent's restriction, it doesn't properly propagates
	519	* config changes in ancestors to their descendents. A child
	520	* should only be allowed to add more restrictions to the parent's
	521	* configuration. Fix it and remove the following.
	522	*/
	523	.broken_hierarchy = true,
08ce5f16 SH	524	};
08ce5f16 SH	525
ad676077 AR	526	/**
	527	* __devcgroup_check_permission - checks if an inode operation is permitted
	528	* @dev_cgroup: the dev cgroup to be tested against
	529	* @type: device type
	530	* @major: device major number
	531	* @minor: device minor number
	532	* @access: combination of ACC_WRITE, ACC_READ and ACC_MKNOD
	533	*
	534	* returns 0 on success, -EPERM case the operation is not permitted
	535	*/
	536	static int __devcgroup_check_permission(struct dev_cgroup *dev_cgroup,
	537	short type, u32 major, u32 minor,
	538	short access)
08ce5f16	539	{
ad676077 AR	540	struct dev_whitelist_item wh;
ad676077 AR	541	int rc;
36fd71d2	542
ad676077 AR	543	memset(&wh, 0, sizeof(wh));
	544	wh.type = type;
	545	wh.major = major;
	546	wh.minor = minor;
	547	wh.access = access;
36fd71d2	548
ad676077 AR	549	rcu_read_lock();
	550	rc = may_access_whitelist(dev_cgroup, &wh);
	551	rcu_read_unlock();
cd500819	552
ad676077 AR	553	if (!rc)
ad676077 AR	554	return -EPERM;
36fd71d2	555
ad676077 AR	556	return 0;
ad676077 AR	557	}
08ce5f16	558
ad676077 AR	559	int __devcgroup_inode_permission(struct inode *inode, int mask)
	560	{
	561	struct dev_cgroup *dev_cgroup = task_devcgroup(current);
	562	short type, access = 0;
	563
	564	if (S_ISBLK(inode->i_mode))
	565	type = DEV_BLOCK;
	566	if (S_ISCHR(inode->i_mode))
	567	type = DEV_CHAR;
	568	if (mask & MAY_WRITE)
	569	access \|= ACC_WRITE;
	570	if (mask & MAY_READ)
	571	access \|= ACC_READ;
	572
	573	return __devcgroup_check_permission(dev_cgroup, type, imajor(inode),
	574	iminor(inode), access);
08ce5f16 SH	575	}
	576
	577	int devcgroup_inode_mknod(int mode, dev_t dev)
	578	{
ad676077 AR	579	struct dev_cgroup *dev_cgroup = task_devcgroup(current);
ad676077 AR	580	short type;
08ce5f16	581
0b82ac37 SH	582	if (!S_ISBLK(mode) && !S_ISCHR(mode))
	583	return 0;
	584
ad676077 AR	585	if (S_ISBLK(mode))
	586	type = DEV_BLOCK;
	587	else
	588	type = DEV_CHAR;
36fd71d2	589
ad676077 AR	590	return __devcgroup_check_permission(dev_cgroup, type, MAJOR(dev),
ad676077 AR	591	MINOR(dev), ACC_MKNOD);
36fd71d2	592
08ce5f16	593	}