[linux-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/ptrace.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;
	struct sk_filter *prog;
};

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

/*
 * Endianness is explicitly ignored and left for BPF program authors to manage
 * as per the specific architecture.
 */
static void populate_seccomp_data(struct seccomp_data *sd)
{
	struct task_struct *task = current;
	struct pt_regs *regs = task_pt_regs(task);
	unsigned long args[6];

	sd->nr = syscall_get_nr(task, regs);
	sd->arch = syscall_get_arch();
	syscall_get_arguments(task, regs, 0, 6, args);
	sd->args[0] = args[0];
	sd->args[1] = args[1];
	sd->args[2] = args[2];
	sd->args[3] = args[3];
	sd->args[4] = args[4];
	sd->args[5] = args[5];
	sd->instruction_pointer = KSTK_EIP(task);
}

/**
 *	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_LDX | BPF_W | BPF_ABS;
			/* 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_LD | BPF_IMM;
			ftest->k = sizeof(struct seccomp_data);
			continue;
		case BPF_S_LDX_W_LEN:
			ftest->code = BPF_LDX | BPF_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_XOR_K:
		case BPF_S_ALU_XOR_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:
			sk_decode_filter(ftest, ftest);
			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;
	struct seccomp_data sd;
	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;

	populate_seccomp_data(&sd);

	/*
	 * 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(f->prog, (void *)&sd);

		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;
	struct sock_filter *fp;
	int new_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->prog->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;

	fp = kzalloc(fp_size, GFP_KERNEL|__GFP_NOWARN);
	if (!fp)
		return -ENOMEM;

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

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

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

	/* Convert 'sock_filter' insns to 'sock_filter_int' insns */
	ret = sk_convert_filter(fp, fprog->len, NULL, &new_len);
	if (ret)
		goto free_prog;

	/* Allocate a new seccomp_filter */
	ret = -ENOMEM;
	filter = kzalloc(sizeof(struct seccomp_filter),
			 GFP_KERNEL|__GFP_NOWARN);
	if (!filter)
		goto free_prog;

	filter->prog = kzalloc(sk_filter_size(new_len),
			       GFP_KERNEL|__GFP_NOWARN);
	if (!filter->prog)
		goto free_filter;

	ret = sk_convert_filter(fp, fprog->len, filter->prog->insnsi, &new_len);
	if (ret)
		goto free_filter_prog;
	kfree(fp);

	atomic_set(&filter->usage, 1);
	filter->prog->len = new_len;

	sk_filter_select_runtime(filter->prog);

	/*
	 * 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;

free_filter_prog:
	kfree(filter->prog);
free_filter:
	kfree(filter);
free_prog:
	kfree(fp);
	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.
 */
static 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;
		sk_filter_free(freeme->prog);
		kfree(freeme);
	}
}

/**
 * seccomp_send_sigsys - signals the task to allow in-process syscall emulation
 * @syscall: syscall number to send to userland
 * @reason: filter-supplied reason code to send to userland (via si_errno)
 *
 * Forces a SIGSYS with a code of SYS_SECCOMP and related sigsys info.
 */
static void seccomp_send_sigsys(int syscall, int reason)
{
	struct siginfo info;
	memset(&info, 0, sizeof(info));
	info.si_signo = SIGSYS;
	info.si_code = SYS_SECCOMP;
	info.si_call_addr = (void __user *)KSTK_EIP(current);
	info.si_errno = reason;
	info.si_arch = syscall_get_arch();
	info.si_syscall = syscall;
	force_sig_info(SIGSYS, &info, current);
}
#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;

	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;
		ret = SECCOMP_RET_KILL;
		break;
#ifdef CONFIG_SECCOMP_FILTER
	case SECCOMP_MODE_FILTER: {
		int data;
		struct pt_regs *regs = task_pt_regs(current);
		ret = seccomp_run_filters(this_syscall);
		data = ret & SECCOMP_RET_DATA;
		ret &= SECCOMP_RET_ACTION;
		switch (ret) {
		case SECCOMP_RET_ERRNO:
			/* Set the low-order 16-bits as a errno. */
			syscall_set_return_value(current, regs,
						 -data, 0);
			goto skip;
		case SECCOMP_RET_TRAP:
			/* Show the handler the original registers. */
			syscall_rollback(current, regs);
			/* Let the filter pass back 16 bits of data. */
			seccomp_send_sigsys(this_syscall, data);
			goto skip;
		case SECCOMP_RET_TRACE:
			/* Skip these calls if there is no tracer. */
			if (!ptrace_event_enabled(current, PTRACE_EVENT_SECCOMP)) {
				syscall_set_return_value(current, regs,
							 -ENOSYS, 0);
				goto skip;
			}
			/* Allow the BPF to provide the event message */
			ptrace_event(PTRACE_EVENT_SECCOMP, data);
			/*
			 * The delivery of a fatal signal during event
			 * notification may silently skip tracer notification.
			 * Terminating the task now avoids executing a system
			 * call that may not be intended.
			 */
			if (fatal_signal_pending(current))
				break;
			if (syscall_get_nr(current, regs) < 0)
				goto skip;  /* Explicit request to skip. */

			return 0;
		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);
#ifdef CONFIG_SECCOMP_FILTER
skip:
	audit_seccomp(this_syscall, exit_sig, ret);
#endif
	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>
fb0fadf9	27	#include <linux/ptrace.h>
e2cfabdf WD	28	#include <linux/security.h>
	29	#include <linux/slab.h>
	30	#include <linux/tracehook.h>
	31	#include <linux/uaccess.h>
	32
	33	/**
	34	* struct seccomp_filter - container for seccomp BPF programs
	35	*
	36	* @usage: reference count to manage the object lifetime.
	37	* get/put helpers should be used when accessing an instance
	38	* outside of a lifetime-guarded section. In general, this
	39	* is only needed for handling filters shared across tasks.
	40	* @prev: points to a previously installed, or inherited, filter
	41	* @len: the number of instructions in the program
	42	* @insns: the BPF program instructions to evaluate
	43	*
	44	* seccomp_filter objects are organized in a tree linked via the @prev
	45	* pointer. For any task, it appears to be a singly-linked list starting
	46	* with current->seccomp.filter, the most recently attached or inherited filter.
	47	* However, multiple filters may share a @prev node, by way of fork(), which
	48	* results in a unidirectional tree existing in memory. This is similar to
	49	* how namespaces work.
	50	*
	51	* seccomp_filter objects should never be modified after being attached
	52	* to a task_struct (other than @usage).
	53	*/
	54	struct seccomp_filter {
	55	atomic_t usage;
	56	struct seccomp_filter *prev;
8f577cad	57	struct sk_filter *prog;
e2cfabdf WD	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
bd4cf0ed	63	/*
e2cfabdf WD	64	* Endianness is explicitly ignored and left for BPF program authors to manage
	65	* as per the specific architecture.
	66	*/
bd4cf0ed	67	static void populate_seccomp_data(struct seccomp_data *sd)
e2cfabdf	68	{
bd4cf0ed AS	69	struct task_struct *task = current;
bd4cf0ed AS	70	struct pt_regs *regs = task_pt_regs(task);
2eac7648	71	unsigned long args[6];
e2cfabdf	72
bd4cf0ed	73	sd->nr = syscall_get_nr(task, regs);
0b747172	74	sd->arch = syscall_get_arch();
2eac7648 DB	75	syscall_get_arguments(task, regs, 0, 6, args);
	76	sd->args[0] = args[0];
	77	sd->args[1] = args[1];
	78	sd->args[2] = args[2];
	79	sd->args[3] = args[3];
	80	sd->args[4] = args[4];
	81	sd->args[5] = args[5];
bd4cf0ed	82	sd->instruction_pointer = KSTK_EIP(task);
e2cfabdf WD	83	}
	84
	85	/**
	86	* seccomp_check_filter - verify seccomp filter code
	87	* @filter: filter to verify
	88	* @flen: length of filter
	89	*
	90	* Takes a previously checked filter (by sk_chk_filter) and
	91	* redirects all filter code that loads struct sk_buff data
	92	* and related data through seccomp_bpf_load. It also
	93	* enforces length and alignment checking of those loads.
	94	*
	95	* Returns 0 if the rule set is legal or -EINVAL if not.
	96	*/
	97	static int seccomp_check_filter(struct sock_filter *filter, unsigned int flen)
	98	{
	99	int pc;
	100	for (pc = 0; pc < flen; pc++) {
	101	struct sock_filter *ftest = &filter[pc];
	102	u16 code = ftest->code;
	103	u32 k = ftest->k;
	104
	105	switch (code) {
	106	case BPF_S_LD_W_ABS:
bd4cf0ed	107	ftest->code = BPF_LDX \| BPF_W \| BPF_ABS;
e2cfabdf WD	108	/* 32-bit aligned and not out of bounds. */
	109	if (k >= sizeof(struct seccomp_data) \|\| k & 3)
	110	return -EINVAL;
	111	continue;
	112	case BPF_S_LD_W_LEN:
bd4cf0ed	113	ftest->code = BPF_LD \| BPF_IMM;
e2cfabdf WD	114	ftest->k = sizeof(struct seccomp_data);
	115	continue;
	116	case BPF_S_LDX_W_LEN:
bd4cf0ed	117	ftest->code = BPF_LDX \| BPF_IMM;
e2cfabdf WD	118	ftest->k = sizeof(struct seccomp_data);
	119	continue;
	120	/* Explicitly include allowed calls. */
	121	case BPF_S_RET_K:
	122	case BPF_S_RET_A:
	123	case BPF_S_ALU_ADD_K:
	124	case BPF_S_ALU_ADD_X:
	125	case BPF_S_ALU_SUB_K:
	126	case BPF_S_ALU_SUB_X:
	127	case BPF_S_ALU_MUL_K:
	128	case BPF_S_ALU_MUL_X:
	129	case BPF_S_ALU_DIV_X:
	130	case BPF_S_ALU_AND_K:
	131	case BPF_S_ALU_AND_X:
	132	case BPF_S_ALU_OR_K:
	133	case BPF_S_ALU_OR_X:
d1327479 NS	134	case BPF_S_ALU_XOR_K:
d1327479 NS	135	case BPF_S_ALU_XOR_X:
e2cfabdf WD	136	case BPF_S_ALU_LSH_K:
	137	case BPF_S_ALU_LSH_X:
	138	case BPF_S_ALU_RSH_K:
	139	case BPF_S_ALU_RSH_X:
	140	case BPF_S_ALU_NEG:
	141	case BPF_S_LD_IMM:
	142	case BPF_S_LDX_IMM:
	143	case BPF_S_MISC_TAX:
	144	case BPF_S_MISC_TXA:
	145	case BPF_S_ALU_DIV_K:
	146	case BPF_S_LD_MEM:
	147	case BPF_S_LDX_MEM:
	148	case BPF_S_ST:
	149	case BPF_S_STX:
	150	case BPF_S_JMP_JA:
	151	case BPF_S_JMP_JEQ_K:
	152	case BPF_S_JMP_JEQ_X:
	153	case BPF_S_JMP_JGE_K:
	154	case BPF_S_JMP_JGE_X:
	155	case BPF_S_JMP_JGT_K:
	156	case BPF_S_JMP_JGT_X:
	157	case BPF_S_JMP_JSET_K:
	158	case BPF_S_JMP_JSET_X:
bd4cf0ed	159	sk_decode_filter(ftest, ftest);
e2cfabdf WD	160	continue;
	161	default:
	162	return -EINVAL;
	163	}
	164	}
	165	return 0;
	166	}
	167
	168	/**
	169	* seccomp_run_filters - evaluates all seccomp filters against @syscall
	170	* @syscall: number of the current system call
	171	*
	172	* Returns valid seccomp BPF response codes.
	173	*/
	174	static u32 seccomp_run_filters(int syscall)
	175	{
	176	struct seccomp_filter *f;
bd4cf0ed	177	struct seccomp_data sd;
acf3b2c7 WD	178	u32 ret = SECCOMP_RET_ALLOW;
	179
	180	/* Ensure unexpected behavior doesn't result in failing open. */
	181	if (WARN_ON(current->seccomp.filter == NULL))
	182	return SECCOMP_RET_KILL;
	183
bd4cf0ed AS	184	populate_seccomp_data(&sd);
bd4cf0ed AS	185
e2cfabdf WD	186	/*
e2cfabdf WD	187	* All filters in the list are evaluated and the lowest BPF return
acf3b2c7	188	* value always takes priority (ignoring the DATA).
e2cfabdf WD	189	*/
e2cfabdf WD	190	for (f = current->seccomp.filter; f; f = f->prev) {
8f577cad AS	191	u32 cur_ret = SK_RUN_FILTER(f->prog, (void *)&sd);
8f577cad AS	192
acf3b2c7 WD	193	if ((cur_ret & SECCOMP_RET_ACTION) < (ret & SECCOMP_RET_ACTION))
acf3b2c7 WD	194	ret = cur_ret;
e2cfabdf WD	195	}
	196	return ret;
	197	}
	198
	199	/**
	200	* seccomp_attach_filter: Attaches a seccomp filter to current.
	201	* @fprog: BPF program to install
	202	*
	203	* Returns 0 on success or an errno on failure.
	204	*/
	205	static long seccomp_attach_filter(struct sock_fprog *fprog)
	206	{
	207	struct seccomp_filter *filter;
	208	unsigned long fp_size = fprog->len * sizeof(struct sock_filter);
	209	unsigned long total_insns = fprog->len;
bd4cf0ed AS	210	struct sock_filter *fp;
bd4cf0ed AS	211	int new_len;
e2cfabdf WD	212	long ret;
	213
	214	if (fprog->len == 0 \|\| fprog->len > BPF_MAXINSNS)
	215	return -EINVAL;
	216
	217	for (filter = current->seccomp.filter; filter; filter = filter->prev)
8f577cad	218	total_insns += filter->prog->len + 4; /* include a 4 instr penalty */
e2cfabdf WD	219	if (total_insns > MAX_INSNS_PER_PATH)
	220	return -ENOMEM;
	221
	222	/*
	223	* Installing a seccomp filter requires that the task have
	224	* CAP_SYS_ADMIN in its namespace or be running with no_new_privs.
	225	* This avoids scenarios where unprivileged tasks can affect the
	226	* behavior of privileged children.
	227	*/
	228	if (!current->no_new_privs &&
	229	security_capable_noaudit(current_cred(), current_user_ns(),
	230	CAP_SYS_ADMIN) != 0)
	231	return -EACCES;
	232
bd4cf0ed AS	233	fp = kzalloc(fp_size, GFP_KERNEL\|__GFP_NOWARN);
bd4cf0ed AS	234	if (!fp)
e2cfabdf	235	return -ENOMEM;
e2cfabdf WD	236
	237	/* Copy the instructions from fprog. */
	238	ret = -EFAULT;
bd4cf0ed AS	239	if (copy_from_user(fp, fprog->filter, fp_size))
bd4cf0ed AS	240	goto free_prog;
e2cfabdf WD	241
e2cfabdf WD	242	/* Check and rewrite the fprog via the skb checker */
bd4cf0ed	243	ret = sk_chk_filter(fp, fprog->len);
e2cfabdf	244	if (ret)
bd4cf0ed	245	goto free_prog;
e2cfabdf WD	246
e2cfabdf WD	247	/* Check and rewrite the fprog for seccomp use */
bd4cf0ed AS	248	ret = seccomp_check_filter(fp, fprog->len);
	249	if (ret)
	250	goto free_prog;
	251
	252	/* Convert 'sock_filter' insns to 'sock_filter_int' insns */
	253	ret = sk_convert_filter(fp, fprog->len, NULL, &new_len);
	254	if (ret)
	255	goto free_prog;
	256
	257	/* Allocate a new seccomp_filter */
0acf07d2	258	ret = -ENOMEM;
8f577cad	259	filter = kzalloc(sizeof(struct seccomp_filter),
bd4cf0ed AS	260	GFP_KERNEL\|__GFP_NOWARN);
	261	if (!filter)
	262	goto free_prog;
	263
8f577cad AS	264	filter->prog = kzalloc(sk_filter_size(new_len),
	265	GFP_KERNEL\|__GFP_NOWARN);
	266	if (!filter->prog)
bd4cf0ed	267	goto free_filter;
8f577cad AS	268
	269	ret = sk_convert_filter(fp, fprog->len, filter->prog->insnsi, &new_len);
	270	if (ret)
	271	goto free_filter_prog;
0acf07d2	272	kfree(fp);
bd4cf0ed AS	273
bd4cf0ed AS	274	atomic_set(&filter->usage, 1);
8f577cad	275	filter->prog->len = new_len;
8f577cad	276
5fe821a9	277	sk_filter_select_runtime(filter->prog);
e2cfabdf WD	278
	279	/*
	280	* If there is an existing filter, make it the prev and don't drop its
	281	* task reference.
	282	*/
	283	filter->prev = current->seccomp.filter;
	284	current->seccomp.filter = filter;
	285	return 0;
bd4cf0ed	286
8f577cad AS	287	free_filter_prog:
8f577cad AS	288	kfree(filter->prog);
bd4cf0ed	289	free_filter:
e2cfabdf	290	kfree(filter);
bd4cf0ed AS	291	free_prog:
bd4cf0ed AS	292	kfree(fp);
e2cfabdf WD	293	return ret;
	294	}
	295
	296	/**
	297	* seccomp_attach_user_filter - attaches a user-supplied sock_fprog
	298	* @user_filter: pointer to the user data containing a sock_fprog.
	299	*
	300	* Returns 0 on success and non-zero otherwise.
	301	*/
864f32a5	302	static long seccomp_attach_user_filter(char __user *user_filter)
e2cfabdf WD	303	{
	304	struct sock_fprog fprog;
	305	long ret = -EFAULT;
	306
	307	#ifdef CONFIG_COMPAT
	308	if (is_compat_task()) {
	309	struct compat_sock_fprog fprog32;
	310	if (copy_from_user(&fprog32, user_filter, sizeof(fprog32)))
	311	goto out;
	312	fprog.len = fprog32.len;
	313	fprog.filter = compat_ptr(fprog32.filter);
	314	} else /* falls through to the if below. */
	315	#endif
	316	if (copy_from_user(&fprog, user_filter, sizeof(fprog)))
	317	goto out;
	318	ret = seccomp_attach_filter(&fprog);
	319	out:
	320	return ret;
	321	}
	322
	323	/* get_seccomp_filter - increments the reference count of the filter on @tsk */
	324	void get_seccomp_filter(struct task_struct *tsk)
	325	{
	326	struct seccomp_filter *orig = tsk->seccomp.filter;
	327	if (!orig)
	328	return;
	329	/* Reference count is bounded by the number of total processes. */
	330	atomic_inc(&orig->usage);
	331	}
	332
	333	/* put_seccomp_filter - decrements the ref count of tsk->seccomp.filter */
	334	void put_seccomp_filter(struct task_struct *tsk)
	335	{
	336	struct seccomp_filter *orig = tsk->seccomp.filter;
	337	/* Clean up single-reference branches iteratively. */
	338	while (orig && atomic_dec_and_test(&orig->usage)) {
	339	struct seccomp_filter *freeme = orig;
	340	orig = orig->prev;
5fe821a9	341	sk_filter_free(freeme->prog);
e2cfabdf WD	342	kfree(freeme);
	343	}
	344	}
bb6ea430 WD	345
	346	/**
	347	* seccomp_send_sigsys - signals the task to allow in-process syscall emulation
	348	* @syscall: syscall number to send to userland
	349	* @reason: filter-supplied reason code to send to userland (via si_errno)
	350	*
	351	* Forces a SIGSYS with a code of SYS_SECCOMP and related sigsys info.
	352	*/
	353	static void seccomp_send_sigsys(int syscall, int reason)
	354	{
	355	struct siginfo info;
	356	memset(&info, 0, sizeof(info));
	357	info.si_signo = SIGSYS;
	358	info.si_code = SYS_SECCOMP;
	359	info.si_call_addr = (void __user *)KSTK_EIP(current);
	360	info.si_errno = reason;
5e937a9a	361	info.si_arch = syscall_get_arch();
bb6ea430 WD	362	info.si_syscall = syscall;
	363	force_sig_info(SIGSYS, &info, current);
	364	}
e2cfabdf	365	#endif /* CONFIG_SECCOMP_FILTER */
1da177e4 LT	366
	367	/*
	368	* Secure computing mode 1 allows only read/write/exit/sigreturn.
	369	* To be fully secure this must be combined with rlimit
	370	* to limit the stack allocations too.
	371	*/
	372	static int mode1_syscalls[] = {
	373	__NR_seccomp_read, __NR_seccomp_write, __NR_seccomp_exit, __NR_seccomp_sigreturn,
	374	0, /* null terminated */
	375	};
	376
5b101740	377	#ifdef CONFIG_COMPAT
1da177e4 LT	378	static int mode1_syscalls_32[] = {
	379	__NR_seccomp_read_32, __NR_seccomp_write_32, __NR_seccomp_exit_32, __NR_seccomp_sigreturn_32,
	380	0, /* null terminated */
	381	};
	382	#endif
	383
acf3b2c7	384	int __secure_computing(int this_syscall)
1da177e4 LT	385	{
1da177e4 LT	386	int mode = current->seccomp.mode;
e2cfabdf WD	387	int exit_sig = 0;
e2cfabdf WD	388	int *syscall;
8156b451	389	u32 ret;
1da177e4 LT	390
1da177e4 LT	391	switch (mode) {
e2cfabdf	392	case SECCOMP_MODE_STRICT:
1da177e4	393	syscall = mode1_syscalls;
5b101740 RM	394	#ifdef CONFIG_COMPAT
5b101740 RM	395	if (is_compat_task())
1da177e4 LT	396	syscall = mode1_syscalls_32;
	397	#endif
	398	do {
	399	if (*syscall == this_syscall)
acf3b2c7	400	return 0;
1da177e4	401	} while (*++syscall);
e2cfabdf	402	exit_sig = SIGKILL;
8156b451	403	ret = SECCOMP_RET_KILL;
1da177e4	404	break;
e2cfabdf	405	#ifdef CONFIG_SECCOMP_FILTER
8156b451 WD	406	case SECCOMP_MODE_FILTER: {
8156b451 WD	407	int data;
87b526d3	408	struct pt_regs *regs = task_pt_regs(current);
acf3b2c7 WD	409	ret = seccomp_run_filters(this_syscall);
acf3b2c7 WD	410	data = ret & SECCOMP_RET_DATA;
8156b451 WD	411	ret &= SECCOMP_RET_ACTION;
8156b451 WD	412	switch (ret) {
acf3b2c7 WD	413	case SECCOMP_RET_ERRNO:
acf3b2c7 WD	414	/* Set the low-order 16-bits as a errno. */
87b526d3	415	syscall_set_return_value(current, regs,
acf3b2c7 WD	416	-data, 0);
acf3b2c7 WD	417	goto skip;
bb6ea430 WD	418	case SECCOMP_RET_TRAP:
bb6ea430 WD	419	/* Show the handler the original registers. */
87b526d3	420	syscall_rollback(current, regs);
bb6ea430 WD	421	/* Let the filter pass back 16 bits of data. */
	422	seccomp_send_sigsys(this_syscall, data);
	423	goto skip;
fb0fadf9 WD	424	case SECCOMP_RET_TRACE:
fb0fadf9 WD	425	/* Skip these calls if there is no tracer. */
87b526d3 AL	426	if (!ptrace_event_enabled(current, PTRACE_EVENT_SECCOMP)) {
	427	syscall_set_return_value(current, regs,
	428	-ENOSYS, 0);
fb0fadf9	429	goto skip;
87b526d3	430	}
fb0fadf9 WD	431	/* Allow the BPF to provide the event message */
	432	ptrace_event(PTRACE_EVENT_SECCOMP, data);
	433	/*
	434	* The delivery of a fatal signal during event
	435	* notification may silently skip tracer notification.
	436	* Terminating the task now avoids executing a system
	437	* call that may not be intended.
	438	*/
	439	if (fatal_signal_pending(current))
	440	break;
87b526d3 AL	441	if (syscall_get_nr(current, regs) < 0)
	442	goto skip; /* Explicit request to skip. */
	443
fb0fadf9	444	return 0;
acf3b2c7 WD	445	case SECCOMP_RET_ALLOW:
	446	return 0;
	447	case SECCOMP_RET_KILL:
	448	default:
	449	break;
	450	}
e2cfabdf WD	451	exit_sig = SIGSYS;
e2cfabdf WD	452	break;
8156b451	453	}
e2cfabdf	454	#endif
1da177e4 LT	455	default:
	456	BUG();
	457	}
	458
	459	#ifdef SECCOMP_DEBUG
	460	dump_stack();
	461	#endif
acf3b2c7	462	audit_seccomp(this_syscall, exit_sig, ret);
e2cfabdf	463	do_exit(exit_sig);
8156b451	464	#ifdef CONFIG_SECCOMP_FILTER
acf3b2c7 WD	465	skip:
acf3b2c7 WD	466	audit_seccomp(this_syscall, exit_sig, ret);
8156b451	467	#endif
acf3b2c7	468	return -1;
1da177e4	469	}
1d9d02fe AA	470
	471	long prctl_get_seccomp(void)
	472	{
	473	return current->seccomp.mode;
	474	}
	475
e2cfabdf WD	476	/**
	477	* prctl_set_seccomp: configures current->seccomp.mode
	478	* @seccomp_mode: requested mode to use
	479	* @filter: optional struct sock_fprog for use with SECCOMP_MODE_FILTER
	480	*
	481	* This function may be called repeatedly with a @seccomp_mode of
	482	* SECCOMP_MODE_FILTER to install additional filters. Every filter
	483	* successfully installed will be evaluated (in reverse order) for each system
	484	* call the task makes.
	485	*
	486	* Once current->seccomp.mode is non-zero, it may not be changed.
	487	*
	488	* Returns 0 on success or -EINVAL on failure.
	489	*/
	490	long prctl_set_seccomp(unsigned long seccomp_mode, char __user *filter)
1d9d02fe	491	{
e2cfabdf	492	long ret = -EINVAL;
1d9d02fe	493
e2cfabdf WD	494	if (current->seccomp.mode &&
e2cfabdf WD	495	current->seccomp.mode != seccomp_mode)
1d9d02fe AA	496	goto out;
1d9d02fe AA	497
e2cfabdf WD	498	switch (seccomp_mode) {
	499	case SECCOMP_MODE_STRICT:
	500	ret = 0;
cf99abac AA	501	#ifdef TIF_NOTSC
	502	disable_TSC();
	503	#endif
e2cfabdf WD	504	break;
	505	#ifdef CONFIG_SECCOMP_FILTER
	506	case SECCOMP_MODE_FILTER:
	507	ret = seccomp_attach_user_filter(filter);
	508	if (ret)
	509	goto out;
	510	break;
	511	#endif
	512	default:
	513	goto out;
1d9d02fe AA	514	}
1d9d02fe AA	515
e2cfabdf WD	516	current->seccomp.mode = seccomp_mode;
	517	set_thread_flag(TIF_SECCOMP);
	518	out:
1d9d02fe AA	519	return ret;
1d9d02fe AA	520	}