[linux-2.6-block.git] / kernel / seccomp.c

/*
 * linux/kernel/seccomp.c
 *
 * Copyright 2004-2005  Andrea Arcangeli <andrea@cpushare.com>
 *
 * Copyright (C) 2012 Google, Inc.
 * Will Drewry <wad@chromium.org>
 *
 * This defines a simple but solid secure-computing facility.
 *
 * Mode 1 uses a fixed list of allowed system calls.
 * Mode 2 allows user-defined system call filters in the form
 *        of Berkeley Packet Filters/Linux Socket Filters.
 */

#include <linux/atomic.h>
#include <linux/audit.h>
#include <linux/compat.h>
#include <linux/sched.h>
#include <linux/seccomp.h>

/* #define SECCOMP_DEBUG 1 */

#ifdef CONFIG_SECCOMP_FILTER
#include <asm/syscall.h>
#include <linux/filter.h>
#include <linux/security.h>
#include <linux/slab.h>
#include <linux/tracehook.h>
#include <linux/uaccess.h>

/**
 * struct seccomp_filter - container for seccomp BPF programs
 *
 * @usage: reference count to manage the object lifetime.
 *         get/put helpers should be used when accessing an instance
 *         outside of a lifetime-guarded section.  In general, this
 *         is only needed for handling filters shared across tasks.
 * @prev: points to a previously installed, or inherited, filter
 * @len: the number of instructions in the program
 * @insns: the BPF program instructions to evaluate
 *
 * seccomp_filter objects are organized in a tree linked via the @prev
 * pointer.  For any task, it appears to be a singly-linked list starting
 * with current->seccomp.filter, the most recently attached or inherited filter.
 * However, multiple filters may share a @prev node, by way of fork(), which
 * results in a unidirectional tree existing in memory.  This is similar to
 * how namespaces work.
 *
 * seccomp_filter objects should never be modified after being attached
 * to a task_struct (other than @usage).
 */
struct seccomp_filter {
	atomic_t usage;
	struct seccomp_filter *prev;
	unsigned short len;  /* Instruction count */
	struct sock_filter insns[];
};

/* Limit any path through the tree to 256KB worth of instructions. */
#define MAX_INSNS_PER_PATH ((1 << 18) / sizeof(struct sock_filter))

/**
 * get_u32 - returns a u32 offset into data
 * @data: a unsigned 64 bit value
 * @index: 0 or 1 to return the first or second 32-bits
 *
 * This inline exists to hide the length of unsigned long.  If a 32-bit
 * unsigned long is passed in, it will be extended and the top 32-bits will be
 * 0. If it is a 64-bit unsigned long, then whatever data is resident will be
 * properly returned.
 *
 * Endianness is explicitly ignored and left for BPF program authors to manage
 * as per the specific architecture.
 */
static inline u32 get_u32(u64 data, int index)
{
	return ((u32 *)&data)[index];
}

/* Helper for bpf_load below. */
#define BPF_DATA(_name) offsetof(struct seccomp_data, _name)
/**
 * bpf_load: checks and returns a pointer to the requested offset
 * @off: offset into struct seccomp_data to load from
 *
 * Returns the requested 32-bits of data.
 * seccomp_check_filter() should assure that @off is 32-bit aligned
 * and not out of bounds.  Failure to do so is a BUG.
 */
u32 seccomp_bpf_load(int off)
{
	struct pt_regs *regs = task_pt_regs(current);
	if (off == BPF_DATA(nr))
		return syscall_get_nr(current, regs);
	if (off == BPF_DATA(arch))
		return syscall_get_arch(current, regs);
	if (off >= BPF_DATA(args[0]) && off < BPF_DATA(args[6])) {
		unsigned long value;
		int arg = (off - BPF_DATA(args[0])) / sizeof(u64);
		int index = !!(off % sizeof(u64));
		syscall_get_arguments(current, regs, arg, 1, &value);
		return get_u32(value, index);
	}
	if (off == BPF_DATA(instruction_pointer))
		return get_u32(KSTK_EIP(current), 0);
	if (off == BPF_DATA(instruction_pointer) + sizeof(u32))
		return get_u32(KSTK_EIP(current), 1);
	/* seccomp_check_filter should make this impossible. */
	BUG();
}

/**
 *	seccomp_check_filter - verify seccomp filter code
 *	@filter: filter to verify
 *	@flen: length of filter
 *
 * Takes a previously checked filter (by sk_chk_filter) and
 * redirects all filter code that loads struct sk_buff data
 * and related data through seccomp_bpf_load.  It also
 * enforces length and alignment checking of those loads.
 *
 * Returns 0 if the rule set is legal or -EINVAL if not.
 */
static int seccomp_check_filter(struct sock_filter *filter, unsigned int flen)
{
	int pc;
	for (pc = 0; pc < flen; pc++) {
		struct sock_filter *ftest = &filter[pc];
		u16 code = ftest->code;
		u32 k = ftest->k;

		switch (code) {
		case BPF_S_LD_W_ABS:
			ftest->code = BPF_S_ANC_SECCOMP_LD_W;
			/* 32-bit aligned and not out of bounds. */
			if (k >= sizeof(struct seccomp_data) || k & 3)
				return -EINVAL;
			continue;
		case BPF_S_LD_W_LEN:
			ftest->code = BPF_S_LD_IMM;
			ftest->k = sizeof(struct seccomp_data);
			continue;
		case BPF_S_LDX_W_LEN:
			ftest->code = BPF_S_LDX_IMM;
			ftest->k = sizeof(struct seccomp_data);
			continue;
		/* Explicitly include allowed calls. */
		case BPF_S_RET_K:
		case BPF_S_RET_A:
		case BPF_S_ALU_ADD_K:
		case BPF_S_ALU_ADD_X:
		case BPF_S_ALU_SUB_K:
		case BPF_S_ALU_SUB_X:
		case BPF_S_ALU_MUL_K:
		case BPF_S_ALU_MUL_X:
		case BPF_S_ALU_DIV_X:
		case BPF_S_ALU_AND_K:
		case BPF_S_ALU_AND_X:
		case BPF_S_ALU_OR_K:
		case BPF_S_ALU_OR_X:
		case BPF_S_ALU_LSH_K:
		case BPF_S_ALU_LSH_X:
		case BPF_S_ALU_RSH_K:
		case BPF_S_ALU_RSH_X:
		case BPF_S_ALU_NEG:
		case BPF_S_LD_IMM:
		case BPF_S_LDX_IMM:
		case BPF_S_MISC_TAX:
		case BPF_S_MISC_TXA:
		case BPF_S_ALU_DIV_K:
		case BPF_S_LD_MEM:
		case BPF_S_LDX_MEM:
		case BPF_S_ST:
		case BPF_S_STX:
		case BPF_S_JMP_JA:
		case BPF_S_JMP_JEQ_K:
		case BPF_S_JMP_JEQ_X:
		case BPF_S_JMP_JGE_K:
		case BPF_S_JMP_JGE_X:
		case BPF_S_JMP_JGT_K:
		case BPF_S_JMP_JGT_X:
		case BPF_S_JMP_JSET_K:
		case BPF_S_JMP_JSET_X:
			continue;
		default:
			return -EINVAL;
		}
	}
	return 0;
}

/**
 * seccomp_run_filters - evaluates all seccomp filters against @syscall
 * @syscall: number of the current system call
 *
 * Returns valid seccomp BPF response codes.
 */
static u32 seccomp_run_filters(int syscall)
{
	struct seccomp_filter *f;
	u32 ret = SECCOMP_RET_ALLOW;

	/* Ensure unexpected behavior doesn't result in failing open. */
	if (WARN_ON(current->seccomp.filter == NULL))
		return SECCOMP_RET_KILL;

	/*
	 * All filters in the list are evaluated and the lowest BPF return
	 * value always takes priority (ignoring the DATA).
	 */
	for (f = current->seccomp.filter; f; f = f->prev) {
		u32 cur_ret = sk_run_filter(NULL, f->insns);
		if ((cur_ret & SECCOMP_RET_ACTION) < (ret & SECCOMP_RET_ACTION))
			ret = cur_ret;
	}
	return ret;
}

/**
 * seccomp_attach_filter: Attaches a seccomp filter to current.
 * @fprog: BPF program to install
 *
 * Returns 0 on success or an errno on failure.
 */
static long seccomp_attach_filter(struct sock_fprog *fprog)
{
	struct seccomp_filter *filter;
	unsigned long fp_size = fprog->len * sizeof(struct sock_filter);
	unsigned long total_insns = fprog->len;
	long ret;

	if (fprog->len == 0 || fprog->len > BPF_MAXINSNS)
		return -EINVAL;

	for (filter = current->seccomp.filter; filter; filter = filter->prev)
		total_insns += filter->len + 4;  /* include a 4 instr penalty */
	if (total_insns > MAX_INSNS_PER_PATH)
		return -ENOMEM;

	/*
	 * Installing a seccomp filter requires that the task have
	 * CAP_SYS_ADMIN in its namespace or be running with no_new_privs.
	 * This avoids scenarios where unprivileged tasks can affect the
	 * behavior of privileged children.
	 */
	if (!current->no_new_privs &&
	    security_capable_noaudit(current_cred(), current_user_ns(),
				     CAP_SYS_ADMIN) != 0)
		return -EACCES;

	/* Allocate a new seccomp_filter */
	filter = kzalloc(sizeof(struct seccomp_filter) + fp_size,
			 GFP_KERNEL|__GFP_NOWARN);
	if (!filter)
		return -ENOMEM;
	atomic_set(&filter->usage, 1);
	filter->len = fprog->len;

	/* Copy the instructions from fprog. */
	ret = -EFAULT;
	if (copy_from_user(filter->insns, fprog->filter, fp_size))
		goto fail;

	/* Check and rewrite the fprog via the skb checker */
	ret = sk_chk_filter(filter->insns, filter->len);
	if (ret)
		goto fail;

	/* Check and rewrite the fprog for seccomp use */
	ret = seccomp_check_filter(filter->insns, filter->len);
	if (ret)
		goto fail;

	/*
	 * If there is an existing filter, make it the prev and don't drop its
	 * task reference.
	 */
	filter->prev = current->seccomp.filter;
	current->seccomp.filter = filter;
	return 0;
fail:
	kfree(filter);
	return ret;
}

/**
 * seccomp_attach_user_filter - attaches a user-supplied sock_fprog
 * @user_filter: pointer to the user data containing a sock_fprog.
 *
 * Returns 0 on success and non-zero otherwise.
 */
long seccomp_attach_user_filter(char __user *user_filter)
{
	struct sock_fprog fprog;
	long ret = -EFAULT;

#ifdef CONFIG_COMPAT
	if (is_compat_task()) {
		struct compat_sock_fprog fprog32;
		if (copy_from_user(&fprog32, user_filter, sizeof(fprog32)))
			goto out;
		fprog.len = fprog32.len;
		fprog.filter = compat_ptr(fprog32.filter);
	} else /* falls through to the if below. */
#endif
	if (copy_from_user(&fprog, user_filter, sizeof(fprog)))
		goto out;
	ret = seccomp_attach_filter(&fprog);
out:
	return ret;
}

/* get_seccomp_filter - increments the reference count of the filter on @tsk */
void get_seccomp_filter(struct task_struct *tsk)
{
	struct seccomp_filter *orig = tsk->seccomp.filter;
	if (!orig)
		return;
	/* Reference count is bounded by the number of total processes. */
	atomic_inc(&orig->usage);
}

/* put_seccomp_filter - decrements the ref count of tsk->seccomp.filter */
void put_seccomp_filter(struct task_struct *tsk)
{
	struct seccomp_filter *orig = tsk->seccomp.filter;
	/* Clean up single-reference branches iteratively. */
	while (orig && atomic_dec_and_test(&orig->usage)) {
		struct seccomp_filter *freeme = orig;
		orig = orig->prev;
		kfree(freeme);
	}
}
#endif	/* CONFIG_SECCOMP_FILTER */

/*
 * Secure computing mode 1 allows only read/write/exit/sigreturn.
 * To be fully secure this must be combined with rlimit
 * to limit the stack allocations too.
 */
static int mode1_syscalls[] = {
	__NR_seccomp_read, __NR_seccomp_write, __NR_seccomp_exit, __NR_seccomp_sigreturn,
	0, /* null terminated */
};

#ifdef CONFIG_COMPAT
static int mode1_syscalls_32[] = {
	__NR_seccomp_read_32, __NR_seccomp_write_32, __NR_seccomp_exit_32, __NR_seccomp_sigreturn_32,
	0, /* null terminated */
};
#endif

int __secure_computing(int this_syscall)
{
	int mode = current->seccomp.mode;
	int exit_sig = 0;
	int *syscall;
	u32 ret = SECCOMP_RET_KILL;
	int data;

	switch (mode) {
	case SECCOMP_MODE_STRICT:
		syscall = mode1_syscalls;
#ifdef CONFIG_COMPAT
		if (is_compat_task())
			syscall = mode1_syscalls_32;
#endif
		do {
			if (*syscall == this_syscall)
				return 0;
		} while (*++syscall);
		exit_sig = SIGKILL;
		break;
#ifdef CONFIG_SECCOMP_FILTER
	case SECCOMP_MODE_FILTER:
		ret = seccomp_run_filters(this_syscall);
		data = ret & SECCOMP_RET_DATA;
		switch (ret & SECCOMP_RET_ACTION) {
		case SECCOMP_RET_ERRNO:
			/* Set the low-order 16-bits as a errno. */
			syscall_set_return_value(current, task_pt_regs(current),
						 -data, 0);
			goto skip;
		case SECCOMP_RET_ALLOW:
			return 0;
		case SECCOMP_RET_KILL:
		default:
			break;
		}
		exit_sig = SIGSYS;
		break;
#endif
	default:
		BUG();
	}

#ifdef SECCOMP_DEBUG
	dump_stack();
#endif
	audit_seccomp(this_syscall, exit_sig, ret);
	do_exit(exit_sig);
skip:
	audit_seccomp(this_syscall, exit_sig, ret);
	return -1;
}

long prctl_get_seccomp(void)
{
	return current->seccomp.mode;
}

/**
 * prctl_set_seccomp: configures current->seccomp.mode
 * @seccomp_mode: requested mode to use
 * @filter: optional struct sock_fprog for use with SECCOMP_MODE_FILTER
 *
 * This function may be called repeatedly with a @seccomp_mode of
 * SECCOMP_MODE_FILTER to install additional filters.  Every filter
 * successfully installed will be evaluated (in reverse order) for each system
 * call the task makes.
 *
 * Once current->seccomp.mode is non-zero, it may not be changed.
 *
 * Returns 0 on success or -EINVAL on failure.
 */
long prctl_set_seccomp(unsigned long seccomp_mode, char __user *filter)
{
	long ret = -EINVAL;

	if (current->seccomp.mode &&
	    current->seccomp.mode != seccomp_mode)
		goto out;

	switch (seccomp_mode) {
	case SECCOMP_MODE_STRICT:
		ret = 0;
#ifdef TIF_NOTSC
		disable_TSC();
#endif
		break;
#ifdef CONFIG_SECCOMP_FILTER
	case SECCOMP_MODE_FILTER:
		ret = seccomp_attach_user_filter(filter);
		if (ret)
			goto out;
		break;
#endif
	default:
		goto out;
	}

	current->seccomp.mode = seccomp_mode;
	set_thread_flag(TIF_SECCOMP);
out:
	return ret;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/kernel/seccomp.c
	3	*
	4	* Copyright 2004-2005 Andrea Arcangeli <andrea@cpushare.com>
	5	*
e2cfabdf WD	6	* Copyright (C) 2012 Google, Inc.
	7	* Will Drewry <wad@chromium.org>
	8	*
	9	* This defines a simple but solid secure-computing facility.
	10	*
	11	* Mode 1 uses a fixed list of allowed system calls.
	12	* Mode 2 allows user-defined system call filters in the form
	13	* of Berkeley Packet Filters/Linux Socket Filters.
1da177e4 LT	14	*/
1da177e4 LT	15
e2cfabdf	16	#include <linux/atomic.h>
85e7bac3	17	#include <linux/audit.h>
5b101740	18	#include <linux/compat.h>
e2cfabdf WD	19	#include <linux/sched.h>
e2cfabdf WD	20	#include <linux/seccomp.h>
1da177e4 LT	21
1da177e4 LT	22	/* #define SECCOMP_DEBUG 1 */
e2cfabdf WD	23
	24	#ifdef CONFIG_SECCOMP_FILTER
	25	#include <asm/syscall.h>
	26	#include <linux/filter.h>
	27	#include <linux/security.h>
	28	#include <linux/slab.h>
	29	#include <linux/tracehook.h>
	30	#include <linux/uaccess.h>
	31
	32	/**
	33	* struct seccomp_filter - container for seccomp BPF programs
	34	*
	35	* @usage: reference count to manage the object lifetime.
	36	* get/put helpers should be used when accessing an instance
	37	* outside of a lifetime-guarded section. In general, this
	38	* is only needed for handling filters shared across tasks.
	39	* @prev: points to a previously installed, or inherited, filter
	40	* @len: the number of instructions in the program
	41	* @insns: the BPF program instructions to evaluate
	42	*
	43	* seccomp_filter objects are organized in a tree linked via the @prev
	44	* pointer. For any task, it appears to be a singly-linked list starting
	45	* with current->seccomp.filter, the most recently attached or inherited filter.
	46	* However, multiple filters may share a @prev node, by way of fork(), which
	47	* results in a unidirectional tree existing in memory. This is similar to
	48	* how namespaces work.
	49	*
	50	* seccomp_filter objects should never be modified after being attached
	51	* to a task_struct (other than @usage).
	52	*/
	53	struct seccomp_filter {
	54	atomic_t usage;
	55	struct seccomp_filter *prev;
	56	unsigned short len; /* Instruction count */
	57	struct sock_filter insns[];
	58	};
	59
	60	/* Limit any path through the tree to 256KB worth of instructions. */
	61	#define MAX_INSNS_PER_PATH ((1 << 18) / sizeof(struct sock_filter))
	62
e2cfabdf WD	63	/**
	64	* get_u32 - returns a u32 offset into data
	65	* @data: a unsigned 64 bit value
	66	* @index: 0 or 1 to return the first or second 32-bits
	67	*
	68	* This inline exists to hide the length of unsigned long. If a 32-bit
	69	* unsigned long is passed in, it will be extended and the top 32-bits will be
	70	* 0. If it is a 64-bit unsigned long, then whatever data is resident will be
	71	* properly returned.
	72	*
	73	* Endianness is explicitly ignored and left for BPF program authors to manage
	74	* as per the specific architecture.
	75	*/
	76	static inline u32 get_u32(u64 data, int index)
	77	{
	78	return ((u32 *)&data)[index];
	79	}
	80
	81	/* Helper for bpf_load below. */
	82	#define BPF_DATA(_name) offsetof(struct seccomp_data, _name)
	83	/**
	84	* bpf_load: checks and returns a pointer to the requested offset
	85	* @off: offset into struct seccomp_data to load from
	86	*
	87	* Returns the requested 32-bits of data.
	88	* seccomp_check_filter() should assure that @off is 32-bit aligned
	89	* and not out of bounds. Failure to do so is a BUG.
	90	*/
	91	u32 seccomp_bpf_load(int off)
	92	{
	93	struct pt_regs *regs = task_pt_regs(current);
	94	if (off == BPF_DATA(nr))
	95	return syscall_get_nr(current, regs);
	96	if (off == BPF_DATA(arch))
	97	return syscall_get_arch(current, regs);
	98	if (off >= BPF_DATA(args[0]) && off < BPF_DATA(args[6])) {
	99	unsigned long value;
	100	int arg = (off - BPF_DATA(args[0])) / sizeof(u64);
	101	int index = !!(off % sizeof(u64));
	102	syscall_get_arguments(current, regs, arg, 1, &value);
	103	return get_u32(value, index);
	104	}
	105	if (off == BPF_DATA(instruction_pointer))
	106	return get_u32(KSTK_EIP(current), 0);
	107	if (off == BPF_DATA(instruction_pointer) + sizeof(u32))
	108	return get_u32(KSTK_EIP(current), 1);
	109	/* seccomp_check_filter should make this impossible. */
	110	BUG();
	111	}
	112
	113	/**
	114	* seccomp_check_filter - verify seccomp filter code
	115	* @filter: filter to verify
	116	* @flen: length of filter
	117	*
	118	* Takes a previously checked filter (by sk_chk_filter) and
	119	* redirects all filter code that loads struct sk_buff data
	120	* and related data through seccomp_bpf_load. It also
	121	* enforces length and alignment checking of those loads.
	122	*
	123	* Returns 0 if the rule set is legal or -EINVAL if not.
	124	*/
	125	static int seccomp_check_filter(struct sock_filter *filter, unsigned int flen)
	126	{
127	int pc;
128	for (pc = 0; pc < flen; pc++) {
129	struct sock_filter *ftest = &filter[pc];
130	u16 code = ftest->code;
131	u32 k = ftest->k;
132
133	switch (code) {
134	case BPF_S_LD_W_ABS:
135	ftest->code = BPF_S_ANC_SECCOMP_LD_W;
136	/* 32-bit aligned and not out of bounds. */
137	if (k >= sizeof(struct seccomp_data) \|\| k & 3)
138	return -EINVAL;
139	continue;
140	case BPF_S_LD_W_LEN:
141	ftest->code = BPF_S_LD_IMM;
142	ftest->k = sizeof(struct seccomp_data);
143	continue;
144	case BPF_S_LDX_W_LEN:
145	ftest->code = BPF_S_LDX_IMM;
146	ftest->k = sizeof(struct seccomp_data);
147	continue;
148	/* Explicitly include allowed calls. */
149	case BPF_S_RET_K:
150	case BPF_S_RET_A:
151	case BPF_S_ALU_ADD_K:
152	case BPF_S_ALU_ADD_X:
153	case BPF_S_ALU_SUB_K:
154	case BPF_S_ALU_SUB_X:
155	case BPF_S_ALU_MUL_K:
156	case BPF_S_ALU_MUL_X:
157	case BPF_S_ALU_DIV_X:
158	case BPF_S_ALU_AND_K:
159	case BPF_S_ALU_AND_X:
160	case BPF_S_ALU_OR_K:
161	case BPF_S_ALU_OR_X:
162	case BPF_S_ALU_LSH_K:
163	case BPF_S_ALU_LSH_X:
164	case BPF_S_ALU_RSH_K:
165	case BPF_S_ALU_RSH_X:
166	case BPF_S_ALU_NEG:
167	case BPF_S_LD_IMM:
168	case BPF_S_LDX_IMM:
169	case BPF_S_MISC_TAX:
170	case BPF_S_MISC_TXA:
171	case BPF_S_ALU_DIV_K:
172	case BPF_S_LD_MEM:
173	case BPF_S_LDX_MEM:
174	case BPF_S_ST:
175	case BPF_S_STX:
176	case BPF_S_JMP_JA:
177	case BPF_S_JMP_JEQ_K:
178	case BPF_S_JMP_JEQ_X:
179	case BPF_S_JMP_JGE_K:
180	case BPF_S_JMP_JGE_X:
181	case BPF_S_JMP_JGT_K:
182	case BPF_S_JMP_JGT_X:
183	case BPF_S_JMP_JSET_K:
184	case BPF_S_JMP_JSET_X:
185	continue;
186	default:
187	return -EINVAL;
188	}
189	}
190	return 0;
191	}
192
193	/**
194	* seccomp_run_filters - evaluates all seccomp filters against @syscall
195	* @syscall: number of the current system call
196	*
197	* Returns valid seccomp BPF response codes.
198	*/
199	static u32 seccomp_run_filters(int syscall)
200	{
201	struct seccomp_filter *f;
acf3b2c7 WD	202	u32 ret = SECCOMP_RET_ALLOW;
	203
	204	/* Ensure unexpected behavior doesn't result in failing open. */
	205	if (WARN_ON(current->seccomp.filter == NULL))
	206	return SECCOMP_RET_KILL;
	207
e2cfabdf WD	208	/*
e2cfabdf WD	209	* All filters in the list are evaluated and the lowest BPF return
acf3b2c7	210	* value always takes priority (ignoring the DATA).
e2cfabdf WD	211	*/
e2cfabdf WD	212	for (f = current->seccomp.filter; f; f = f->prev) {
acf3b2c7 WD	213	u32 cur_ret = sk_run_filter(NULL, f->insns);
	214	if ((cur_ret & SECCOMP_RET_ACTION) < (ret & SECCOMP_RET_ACTION))
	215	ret = cur_ret;
e2cfabdf WD	216	}
	217	return ret;
	218	}
	219
	220	/**
	221	* seccomp_attach_filter: Attaches a seccomp filter to current.
	222	* @fprog: BPF program to install
	223	*
	224	* Returns 0 on success or an errno on failure.
	225	*/
	226	static long seccomp_attach_filter(struct sock_fprog *fprog)
	227	{
	228	struct seccomp_filter *filter;
	229	unsigned long fp_size = fprog->len * sizeof(struct sock_filter);
	230	unsigned long total_insns = fprog->len;
	231	long ret;
	232
	233	if (fprog->len == 0 \|\| fprog->len > BPF_MAXINSNS)
	234	return -EINVAL;
	235
	236	for (filter = current->seccomp.filter; filter; filter = filter->prev)
	237	total_insns += filter->len + 4; /* include a 4 instr penalty */
	238	if (total_insns > MAX_INSNS_PER_PATH)
	239	return -ENOMEM;
	240
	241	/*
	242	* Installing a seccomp filter requires that the task have
	243	* CAP_SYS_ADMIN in its namespace or be running with no_new_privs.
	244	* This avoids scenarios where unprivileged tasks can affect the
	245	* behavior of privileged children.
	246	*/
	247	if (!current->no_new_privs &&
	248	security_capable_noaudit(current_cred(), current_user_ns(),
	249	CAP_SYS_ADMIN) != 0)
	250	return -EACCES;
	251
	252	/* Allocate a new seccomp_filter */
	253	filter = kzalloc(sizeof(struct seccomp_filter) + fp_size,
	254	GFP_KERNEL\|__GFP_NOWARN);
	255	if (!filter)
	256	return -ENOMEM;
	257	atomic_set(&filter->usage, 1);
	258	filter->len = fprog->len;
	259
	260	/* Copy the instructions from fprog. */
	261	ret = -EFAULT;
	262	if (copy_from_user(filter->insns, fprog->filter, fp_size))
	263	goto fail;
	264
	265	/* Check and rewrite the fprog via the skb checker */
	266	ret = sk_chk_filter(filter->insns, filter->len);
	267	if (ret)
	268	goto fail;
	269
	270	/* Check and rewrite the fprog for seccomp use */
	271	ret = seccomp_check_filter(filter->insns, filter->len);
	272	if (ret)
	273	goto fail;
	274
	275	/*
	276	* If there is an existing filter, make it the prev and don't drop its
	277	* task reference.
	278	*/
	279	filter->prev = current->seccomp.filter;
280	current->seccomp.filter = filter;
281	return 0;
282	fail:
283	kfree(filter);
284	return ret;
285	}
286
287	/**
288	* seccomp_attach_user_filter - attaches a user-supplied sock_fprog
289	* @user_filter: pointer to the user data containing a sock_fprog.
290	*
291	* Returns 0 on success and non-zero otherwise.
292	*/
293	long seccomp_attach_user_filter(char __user *user_filter)
294	{
295	struct sock_fprog fprog;
296	long ret = -EFAULT;
297
298	#ifdef CONFIG_COMPAT
299	if (is_compat_task()) {
300	struct compat_sock_fprog fprog32;
301	if (copy_from_user(&fprog32, user_filter, sizeof(fprog32)))
302	goto out;
303	fprog.len = fprog32.len;
304	fprog.filter = compat_ptr(fprog32.filter);
305	} else /* falls through to the if below. */
306	#endif
307	if (copy_from_user(&fprog, user_filter, sizeof(fprog)))
308	goto out;
309	ret = seccomp_attach_filter(&fprog);
310	out:
311	return ret;
312	}
313
314	/* get_seccomp_filter - increments the reference count of the filter on @tsk */
315	void get_seccomp_filter(struct task_struct *tsk)
316	{
317	struct seccomp_filter *orig = tsk->seccomp.filter;
318	if (!orig)
319	return;
320	/* Reference count is bounded by the number of total processes. */
321	atomic_inc(&orig->usage);
322	}
323
324	/* put_seccomp_filter - decrements the ref count of tsk->seccomp.filter */
325	void put_seccomp_filter(struct task_struct *tsk)
326	{
327	struct seccomp_filter *orig = tsk->seccomp.filter;
328	/* Clean up single-reference branches iteratively. */
329	while (orig && atomic_dec_and_test(&orig->usage)) {
330	struct seccomp_filter *freeme = orig;
331	orig = orig->prev;
332	kfree(freeme);
333	}
334	}
335	#endif /* CONFIG_SECCOMP_FILTER */
1da177e4 LT	336
	337	/*
	338	* Secure computing mode 1 allows only read/write/exit/sigreturn.
	339	* To be fully secure this must be combined with rlimit
	340	* to limit the stack allocations too.
	341	*/
	342	static int mode1_syscalls[] = {
	343	__NR_seccomp_read, __NR_seccomp_write, __NR_seccomp_exit, __NR_seccomp_sigreturn,
	344	0, /* null terminated */
	345	};
	346
5b101740	347	#ifdef CONFIG_COMPAT
1da177e4 LT	348	static int mode1_syscalls_32[] = {
	349	__NR_seccomp_read_32, __NR_seccomp_write_32, __NR_seccomp_exit_32, __NR_seccomp_sigreturn_32,
	350	0, /* null terminated */
	351	};
	352	#endif
	353
acf3b2c7	354	int __secure_computing(int this_syscall)
1da177e4 LT	355	{
1da177e4 LT	356	int mode = current->seccomp.mode;
e2cfabdf WD	357	int exit_sig = 0;
e2cfabdf WD	358	int *syscall;
acf3b2c7 WD	359	u32 ret = SECCOMP_RET_KILL;
acf3b2c7 WD	360	int data;
1da177e4 LT	361
1da177e4 LT	362	switch (mode) {
e2cfabdf	363	case SECCOMP_MODE_STRICT:
1da177e4	364	syscall = mode1_syscalls;
5b101740 RM	365	#ifdef CONFIG_COMPAT
5b101740 RM	366	if (is_compat_task())
1da177e4 LT	367	syscall = mode1_syscalls_32;
	368	#endif
	369	do {
	370	if (*syscall == this_syscall)
acf3b2c7	371	return 0;
1da177e4	372	} while (*++syscall);
e2cfabdf	373	exit_sig = SIGKILL;
1da177e4	374	break;
e2cfabdf WD	375	#ifdef CONFIG_SECCOMP_FILTER
e2cfabdf WD	376	case SECCOMP_MODE_FILTER:
acf3b2c7 WD	377	ret = seccomp_run_filters(this_syscall);
	378	data = ret & SECCOMP_RET_DATA;
	379	switch (ret & SECCOMP_RET_ACTION) {
	380	case SECCOMP_RET_ERRNO:
	381	/* Set the low-order 16-bits as a errno. */
	382	syscall_set_return_value(current, task_pt_regs(current),
	383	-data, 0);
	384	goto skip;
	385	case SECCOMP_RET_ALLOW:
	386	return 0;
	387	case SECCOMP_RET_KILL:
	388	default:
	389	break;
	390	}
e2cfabdf WD	391	exit_sig = SIGSYS;
	392	break;
	393	#endif
1da177e4 LT	394	default:
	395	BUG();
	396	}
	397
	398	#ifdef SECCOMP_DEBUG
	399	dump_stack();
	400	#endif
acf3b2c7	401	audit_seccomp(this_syscall, exit_sig, ret);
e2cfabdf	402	do_exit(exit_sig);
acf3b2c7 WD	403	skip:
	404	audit_seccomp(this_syscall, exit_sig, ret);
	405	return -1;
1da177e4	406	}
1d9d02fe AA	407
	408	long prctl_get_seccomp(void)
	409	{
	410	return current->seccomp.mode;
	411	}
	412
e2cfabdf WD	413	/**
	414	* prctl_set_seccomp: configures current->seccomp.mode
	415	* @seccomp_mode: requested mode to use
	416	* @filter: optional struct sock_fprog for use with SECCOMP_MODE_FILTER
	417	*
	418	* This function may be called repeatedly with a @seccomp_mode of
	419	* SECCOMP_MODE_FILTER to install additional filters. Every filter
	420	* successfully installed will be evaluated (in reverse order) for each system
	421	* call the task makes.
	422	*
	423	* Once current->seccomp.mode is non-zero, it may not be changed.
	424	*
	425	* Returns 0 on success or -EINVAL on failure.
	426	*/
	427	long prctl_set_seccomp(unsigned long seccomp_mode, char __user *filter)
1d9d02fe	428	{
e2cfabdf	429	long ret = -EINVAL;
1d9d02fe	430
e2cfabdf WD	431	if (current->seccomp.mode &&
e2cfabdf WD	432	current->seccomp.mode != seccomp_mode)
1d9d02fe AA	433	goto out;
1d9d02fe AA	434
e2cfabdf WD	435	switch (seccomp_mode) {
	436	case SECCOMP_MODE_STRICT:
	437	ret = 0;
cf99abac AA	438	#ifdef TIF_NOTSC
	439	disable_TSC();
	440	#endif
e2cfabdf WD	441	break;
	442	#ifdef CONFIG_SECCOMP_FILTER
	443	case SECCOMP_MODE_FILTER:
	444	ret = seccomp_attach_user_filter(filter);
	445	if (ret)
	446	goto out;
	447	break;
	448	#endif
	449	default:
	450	goto out;
1d9d02fe AA	451	}
1d9d02fe AA	452
e2cfabdf WD	453	current->seccomp.mode = seccomp_mode;
	454	set_thread_flag(TIF_SECCOMP);
	455	out:
1d9d02fe AA	456	return ret;
1d9d02fe AA	457	}