[linux-2.6-block.git] / include / linux / filter.h

/*
 * Linux Socket Filter Data Structures
 */
#ifndef __LINUX_FILTER_H__
#define __LINUX_FILTER_H__

#include <stdarg.h>

#include <linux/atomic.h>
#include <linux/compat.h>
#include <linux/skbuff.h>
#include <linux/linkage.h>
#include <linux/printk.h>
#include <linux/workqueue.h>
#include <linux/sched.h>
#include <net/sch_generic.h>

#include <asm/cacheflush.h>

#include <uapi/linux/filter.h>
#include <uapi/linux/bpf.h>

struct sk_buff;
struct sock;
struct seccomp_data;
struct bpf_prog_aux;

/* ArgX, context and stack frame pointer register positions. Note,
 * Arg1, Arg2, Arg3, etc are used as argument mappings of function
 * calls in BPF_CALL instruction.
 */
#define BPF_REG_ARG1	BPF_REG_1
#define BPF_REG_ARG2	BPF_REG_2
#define BPF_REG_ARG3	BPF_REG_3
#define BPF_REG_ARG4	BPF_REG_4
#define BPF_REG_ARG5	BPF_REG_5
#define BPF_REG_CTX	BPF_REG_6
#define BPF_REG_FP	BPF_REG_10

/* Additional register mappings for converted user programs. */
#define BPF_REG_A	BPF_REG_0
#define BPF_REG_X	BPF_REG_7
#define BPF_REG_TMP	BPF_REG_8

/* BPF program can access up to 512 bytes of stack space. */
#define MAX_BPF_STACK	512

/* Helper macros for filter block array initializers. */

/* ALU ops on registers, bpf_add|sub|...: dst_reg += src_reg */

#define BPF_ALU64_REG(OP, DST, SRC)				\
	((struct bpf_insn) {					\
		.code  = BPF_ALU64 | BPF_OP(OP) | BPF_X,	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = 0 })

#define BPF_ALU32_REG(OP, DST, SRC)				\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_OP(OP) | BPF_X,		\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = 0 })

/* ALU ops on immediates, bpf_add|sub|...: dst_reg += imm32 */

#define BPF_ALU64_IMM(OP, DST, IMM)				\
	((struct bpf_insn) {					\
		.code  = BPF_ALU64 | BPF_OP(OP) | BPF_K,	\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = IMM })

#define BPF_ALU32_IMM(OP, DST, IMM)				\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_OP(OP) | BPF_K,		\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = IMM })

/* Endianess conversion, cpu_to_{l,b}e(), {l,b}e_to_cpu() */

#define BPF_ENDIAN(TYPE, DST, LEN)				\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_END | BPF_SRC(TYPE),	\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = LEN })

/* Short form of mov, dst_reg = src_reg */

#define BPF_MOV64_REG(DST, SRC)					\
	((struct bpf_insn) {					\
		.code  = BPF_ALU64 | BPF_MOV | BPF_X,		\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = 0 })

#define BPF_MOV32_REG(DST, SRC)					\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_MOV | BPF_X,		\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = 0 })

/* Short form of mov, dst_reg = imm32 */

#define BPF_MOV64_IMM(DST, IMM)					\
	((struct bpf_insn) {					\
		.code  = BPF_ALU64 | BPF_MOV | BPF_K,		\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = IMM })

#define BPF_MOV32_IMM(DST, IMM)					\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_MOV | BPF_K,		\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = IMM })

/* BPF_LD_IMM64 macro encodes single 'load 64-bit immediate' insn */
#define BPF_LD_IMM64(DST, IMM)					\
	BPF_LD_IMM64_RAW(DST, 0, IMM)

#define BPF_LD_IMM64_RAW(DST, SRC, IMM)				\
	((struct bpf_insn) {					\
		.code  = BPF_LD | BPF_DW | BPF_IMM,		\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = (__u32) (IMM) }),			\
	((struct bpf_insn) {					\
		.code  = 0, /* zero is reserved opcode */	\
		.dst_reg = 0,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = ((__u64) (IMM)) >> 32 })

/* pseudo BPF_LD_IMM64 insn used to refer to process-local map_fd */
#define BPF_LD_MAP_FD(DST, MAP_FD)				\
	BPF_LD_IMM64_RAW(DST, BPF_PSEUDO_MAP_FD, MAP_FD)

/* Short form of mov based on type, BPF_X: dst_reg = src_reg, BPF_K: dst_reg = imm32 */

#define BPF_MOV64_RAW(TYPE, DST, SRC, IMM)			\
	((struct bpf_insn) {					\
		.code  = BPF_ALU64 | BPF_MOV | BPF_SRC(TYPE),	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = IMM })

#define BPF_MOV32_RAW(TYPE, DST, SRC, IMM)			\
	((struct bpf_insn) {					\
		.code  = BPF_ALU | BPF_MOV | BPF_SRC(TYPE),	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = IMM })

/* Direct packet access, R0 = *(uint *) (skb->data + imm32) */

#define BPF_LD_ABS(SIZE, IMM)					\
	((struct bpf_insn) {					\
		.code  = BPF_LD | BPF_SIZE(SIZE) | BPF_ABS,	\
		.dst_reg = 0,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = IMM })

/* Indirect packet access, R0 = *(uint *) (skb->data + src_reg + imm32) */

#define BPF_LD_IND(SIZE, SRC, IMM)				\
	((struct bpf_insn) {					\
		.code  = BPF_LD | BPF_SIZE(SIZE) | BPF_IND,	\
		.dst_reg = 0,					\
		.src_reg = SRC,					\
		.off   = 0,					\
		.imm   = IMM })

/* Memory load, dst_reg = *(uint *) (src_reg + off16) */

#define BPF_LDX_MEM(SIZE, DST, SRC, OFF)			\
	((struct bpf_insn) {					\
		.code  = BPF_LDX | BPF_SIZE(SIZE) | BPF_MEM,	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = OFF,					\
		.imm   = 0 })

/* Memory store, *(uint *) (dst_reg + off16) = src_reg */

#define BPF_STX_MEM(SIZE, DST, SRC, OFF)			\
	((struct bpf_insn) {					\
		.code  = BPF_STX | BPF_SIZE(SIZE) | BPF_MEM,	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = OFF,					\
		.imm   = 0 })

/* Atomic memory add, *(uint *)(dst_reg + off16) += src_reg */

#define BPF_STX_XADD(SIZE, DST, SRC, OFF)			\
	((struct bpf_insn) {					\
		.code  = BPF_STX | BPF_SIZE(SIZE) | BPF_XADD,	\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = OFF,					\
		.imm   = 0 })

/* Memory store, *(uint *) (dst_reg + off16) = imm32 */

#define BPF_ST_MEM(SIZE, DST, OFF, IMM)				\
	((struct bpf_insn) {					\
		.code  = BPF_ST | BPF_SIZE(SIZE) | BPF_MEM,	\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = OFF,					\
		.imm   = IMM })

/* Conditional jumps against registers, if (dst_reg 'op' src_reg) goto pc + off16 */

#define BPF_JMP_REG(OP, DST, SRC, OFF)				\
	((struct bpf_insn) {					\
		.code  = BPF_JMP | BPF_OP(OP) | BPF_X,		\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = OFF,					\
		.imm   = 0 })

/* Conditional jumps against immediates, if (dst_reg 'op' imm32) goto pc + off16 */

#define BPF_JMP_IMM(OP, DST, IMM, OFF)				\
	((struct bpf_insn) {					\
		.code  = BPF_JMP | BPF_OP(OP) | BPF_K,		\
		.dst_reg = DST,					\
		.src_reg = 0,					\
		.off   = OFF,					\
		.imm   = IMM })

/* Function call */

#define BPF_EMIT_CALL(FUNC)					\
	((struct bpf_insn) {					\
		.code  = BPF_JMP | BPF_CALL,			\
		.dst_reg = 0,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = ((FUNC) - __bpf_call_base) })

/* Raw code statement block */

#define BPF_RAW_INSN(CODE, DST, SRC, OFF, IMM)			\
	((struct bpf_insn) {					\
		.code  = CODE,					\
		.dst_reg = DST,					\
		.src_reg = SRC,					\
		.off   = OFF,					\
		.imm   = IMM })

/* Program exit */

#define BPF_EXIT_INSN()						\
	((struct bpf_insn) {					\
		.code  = BPF_JMP | BPF_EXIT,			\
		.dst_reg = 0,					\
		.src_reg = 0,					\
		.off   = 0,					\
		.imm   = 0 })

/* Internal classic blocks for direct assignment */

#define __BPF_STMT(CODE, K)					\
	((struct sock_filter) BPF_STMT(CODE, K))

#define __BPF_JUMP(CODE, K, JT, JF)				\
	((struct sock_filter) BPF_JUMP(CODE, K, JT, JF))

#define bytes_to_bpf_size(bytes)				\
({								\
	int bpf_size = -EINVAL;					\
								\
	if (bytes == sizeof(u8))				\
		bpf_size = BPF_B;				\
	else if (bytes == sizeof(u16))				\
		bpf_size = BPF_H;				\
	else if (bytes == sizeof(u32))				\
		bpf_size = BPF_W;				\
	else if (bytes == sizeof(u64))				\
		bpf_size = BPF_DW;				\
								\
	bpf_size;						\
})

#ifdef CONFIG_COMPAT
/* A struct sock_filter is architecture independent. */
struct compat_sock_fprog {
	u16		len;
	compat_uptr_t	filter;	/* struct sock_filter * */
};
#endif

struct sock_fprog_kern {
	u16			len;
	struct sock_filter	*filter;
};

struct bpf_binary_header {
	unsigned int pages;
	u8 image[];
};

struct bpf_prog {
	u16			pages;		/* Number of allocated pages */
	kmemcheck_bitfield_begin(meta);
	u16			jited:1,	/* Is our filter JIT'ed? */
				gpl_compatible:1, /* Is filter GPL compatible? */
				cb_access:1,	/* Is control block accessed? */
				dst_needed:1;	/* Do we need dst entry? */
	kmemcheck_bitfield_end(meta);
	u32			len;		/* Number of filter blocks */
	enum bpf_prog_type	type;		/* Type of BPF program */
	struct bpf_prog_aux	*aux;		/* Auxiliary fields */
	struct sock_fprog_kern	*orig_prog;	/* Original BPF program */
	unsigned int		(*bpf_func)(const struct sk_buff *skb,
					    const struct bpf_insn *filter);
	/* Instructions for interpreter */
	union {
		struct sock_filter	insns[0];
		struct bpf_insn		insnsi[0];
	};
};

struct sk_filter {
	atomic_t	refcnt;
	struct rcu_head	rcu;
	struct bpf_prog	*prog;
};

#define BPF_PROG_RUN(filter, ctx)  (*filter->bpf_func)(ctx, filter->insnsi)

#define BPF_SKB_CB_LEN QDISC_CB_PRIV_LEN

static inline u8 *bpf_skb_cb(struct sk_buff *skb)
{
	/* eBPF programs may read/write skb->cb[] area to transfer meta
	 * data between tail calls. Since this also needs to work with
	 * tc, that scratch memory is mapped to qdisc_skb_cb's data area.
	 *
	 * In some socket filter cases, the cb unfortunately needs to be
	 * saved/restored so that protocol specific skb->cb[] data won't
	 * be lost. In any case, due to unpriviledged eBPF programs
	 * attached to sockets, we need to clear the bpf_skb_cb() area
	 * to not leak previous contents to user space.
	 */
	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) != BPF_SKB_CB_LEN);
	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) !=
		     FIELD_SIZEOF(struct qdisc_skb_cb, data));

	return qdisc_skb_cb(skb)->data;
}

static inline u32 bpf_prog_run_save_cb(const struct bpf_prog *prog,
				       struct sk_buff *skb)
{
	u8 *cb_data = bpf_skb_cb(skb);
	u8 cb_saved[BPF_SKB_CB_LEN];
	u32 res;

	if (unlikely(prog->cb_access)) {
		memcpy(cb_saved, cb_data, sizeof(cb_saved));
		memset(cb_data, 0, sizeof(cb_saved));
	}

	res = BPF_PROG_RUN(prog, skb);

	if (unlikely(prog->cb_access))
		memcpy(cb_data, cb_saved, sizeof(cb_saved));

	return res;
}

static inline u32 bpf_prog_run_clear_cb(const struct bpf_prog *prog,
					struct sk_buff *skb)
{
	u8 *cb_data = bpf_skb_cb(skb);

	if (unlikely(prog->cb_access))
		memset(cb_data, 0, BPF_SKB_CB_LEN);

	return BPF_PROG_RUN(prog, skb);
}

static inline unsigned int bpf_prog_size(unsigned int proglen)
{
	return max(sizeof(struct bpf_prog),
		   offsetof(struct bpf_prog, insns[proglen]));
}

static inline bool bpf_prog_was_classic(const struct bpf_prog *prog)
{
	/* When classic BPF programs have been loaded and the arch
	 * does not have a classic BPF JIT (anymore), they have been
	 * converted via bpf_migrate_filter() to eBPF and thus always
	 * have an unspec program type.
	 */
	return prog->type == BPF_PROG_TYPE_UNSPEC;
}

#define bpf_classic_proglen(fprog) (fprog->len * sizeof(fprog->filter[0]))

#ifdef CONFIG_DEBUG_SET_MODULE_RONX
static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
{
	set_memory_ro((unsigned long)fp, fp->pages);
}

static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
{
	set_memory_rw((unsigned long)fp, fp->pages);
}
#else
static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
{
}

static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
{
}
#endif /* CONFIG_DEBUG_SET_MODULE_RONX */

int sk_filter(struct sock *sk, struct sk_buff *skb);

int bpf_prog_select_runtime(struct bpf_prog *fp);
void bpf_prog_free(struct bpf_prog *fp);

struct bpf_prog *bpf_prog_alloc(unsigned int size, gfp_t gfp_extra_flags);
struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
				  gfp_t gfp_extra_flags);
void __bpf_prog_free(struct bpf_prog *fp);

static inline void bpf_prog_unlock_free(struct bpf_prog *fp)
{
	bpf_prog_unlock_ro(fp);
	__bpf_prog_free(fp);
}

typedef int (*bpf_aux_classic_check_t)(struct sock_filter *filter,
				       unsigned int flen);

int bpf_prog_create(struct bpf_prog **pfp, struct sock_fprog_kern *fprog);
int bpf_prog_create_from_user(struct bpf_prog **pfp, struct sock_fprog *fprog,
			      bpf_aux_classic_check_t trans, bool save_orig);
void bpf_prog_destroy(struct bpf_prog *fp);

int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk);
int __sk_attach_filter(struct sock_fprog *fprog, struct sock *sk,
		       bool locked);
int sk_attach_bpf(u32 ufd, struct sock *sk);
int sk_reuseport_attach_filter(struct sock_fprog *fprog, struct sock *sk);
int sk_reuseport_attach_bpf(u32 ufd, struct sock *sk);
int sk_detach_filter(struct sock *sk);
int __sk_detach_filter(struct sock *sk, bool locked);

int sk_get_filter(struct sock *sk, struct sock_filter __user *filter,
		  unsigned int len);

bool sk_filter_charge(struct sock *sk, struct sk_filter *fp);
void sk_filter_uncharge(struct sock *sk, struct sk_filter *fp);

u64 __bpf_call_base(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
void bpf_int_jit_compile(struct bpf_prog *fp);
bool bpf_helper_changes_skb_data(void *func);

#ifdef CONFIG_BPF_JIT
typedef void (*bpf_jit_fill_hole_t)(void *area, unsigned int size);

struct bpf_binary_header *
bpf_jit_binary_alloc(unsigned int proglen, u8 **image_ptr,
		     unsigned int alignment,
		     bpf_jit_fill_hole_t bpf_fill_ill_insns);
void bpf_jit_binary_free(struct bpf_binary_header *hdr);

void bpf_jit_compile(struct bpf_prog *fp);
void bpf_jit_free(struct bpf_prog *fp);

static inline void bpf_jit_dump(unsigned int flen, unsigned int proglen,
				u32 pass, void *image)
{
	pr_err("flen=%u proglen=%u pass=%u image=%pK from=%s pid=%d\n", flen,
	       proglen, pass, image, current->comm, task_pid_nr(current));

	if (image)
		print_hex_dump(KERN_ERR, "JIT code: ", DUMP_PREFIX_OFFSET,
			       16, 1, image, proglen, false);
}
#else
static inline void bpf_jit_compile(struct bpf_prog *fp)
{
}

static inline void bpf_jit_free(struct bpf_prog *fp)
{
	bpf_prog_unlock_free(fp);
}
#endif /* CONFIG_BPF_JIT */

#define BPF_ANC		BIT(15)

static inline bool bpf_needs_clear_a(const struct sock_filter *first)
{
	switch (first->code) {
	case BPF_RET | BPF_K:
	case BPF_LD | BPF_W | BPF_LEN:
		return false;

	case BPF_LD | BPF_W | BPF_ABS:
	case BPF_LD | BPF_H | BPF_ABS:
	case BPF_LD | BPF_B | BPF_ABS:
		if (first->k == SKF_AD_OFF + SKF_AD_ALU_XOR_X)
			return true;
		return false;

	default:
		return true;
	}
}

static inline u16 bpf_anc_helper(const struct sock_filter *ftest)
{
	BUG_ON(ftest->code & BPF_ANC);

	switch (ftest->code) {
	case BPF_LD | BPF_W | BPF_ABS:
	case BPF_LD | BPF_H | BPF_ABS:
	case BPF_LD | BPF_B | BPF_ABS:
#define BPF_ANCILLARY(CODE)	case SKF_AD_OFF + SKF_AD_##CODE:	\
				return BPF_ANC | SKF_AD_##CODE
		switch (ftest->k) {
		BPF_ANCILLARY(PROTOCOL);
		BPF_ANCILLARY(PKTTYPE);
		BPF_ANCILLARY(IFINDEX);
		BPF_ANCILLARY(NLATTR);
		BPF_ANCILLARY(NLATTR_NEST);
		BPF_ANCILLARY(MARK);
		BPF_ANCILLARY(QUEUE);
		BPF_ANCILLARY(HATYPE);
		BPF_ANCILLARY(RXHASH);
		BPF_ANCILLARY(CPU);
		BPF_ANCILLARY(ALU_XOR_X);
		BPF_ANCILLARY(VLAN_TAG);
		BPF_ANCILLARY(VLAN_TAG_PRESENT);
		BPF_ANCILLARY(PAY_OFFSET);
		BPF_ANCILLARY(RANDOM);
		BPF_ANCILLARY(VLAN_TPID);
		}
		/* Fallthrough. */
	default:
		return ftest->code;
	}
}

void *bpf_internal_load_pointer_neg_helper(const struct sk_buff *skb,
					   int k, unsigned int size);

static inline void *bpf_load_pointer(const struct sk_buff *skb, int k,
				     unsigned int size, void *buffer)
{
	if (k >= 0)
		return skb_header_pointer(skb, k, size, buffer);

	return bpf_internal_load_pointer_neg_helper(skb, k, size);
}

static inline int bpf_tell_extensions(void)
{
	return SKF_AD_MAX;
}

#endif /* __LINUX_FILTER_H__ */
Commit	Line	Data
	1	/*
	2	* Linux Socket Filter Data Structures
	3	*/
	4	#ifndef __LINUX_FILTER_H__
	5	#define __LINUX_FILTER_H__
	6
	7	#include <stdarg.h>
	8
	9	#include <linux/atomic.h>
	10	#include <linux/compat.h>
	11	#include <linux/skbuff.h>
	12	#include <linux/linkage.h>
	13	#include <linux/printk.h>
	14	#include <linux/workqueue.h>
	15	#include <linux/sched.h>
	16	#include <net/sch_generic.h>
	17
	18	#include <asm/cacheflush.h>
	19
	20	#include <uapi/linux/filter.h>
	21	#include <uapi/linux/bpf.h>
	22
	23	struct sk_buff;
	24	struct sock;
	25	struct seccomp_data;
	26	struct bpf_prog_aux;
	27
	28	/* ArgX, context and stack frame pointer register positions. Note,
	29	* Arg1, Arg2, Arg3, etc are used as argument mappings of function
	30	* calls in BPF_CALL instruction.
	31	*/
	32	#define BPF_REG_ARG1 BPF_REG_1
	33	#define BPF_REG_ARG2 BPF_REG_2
	34	#define BPF_REG_ARG3 BPF_REG_3
	35	#define BPF_REG_ARG4 BPF_REG_4
	36	#define BPF_REG_ARG5 BPF_REG_5
	37	#define BPF_REG_CTX BPF_REG_6
	38	#define BPF_REG_FP BPF_REG_10
	39
	40	/* Additional register mappings for converted user programs. */
	41	#define BPF_REG_A BPF_REG_0
	42	#define BPF_REG_X BPF_REG_7
	43	#define BPF_REG_TMP BPF_REG_8
	44
	45	/* BPF program can access up to 512 bytes of stack space. */
	46	#define MAX_BPF_STACK 512
	47
	48	/* Helper macros for filter block array initializers. */
	49
	50	/* ALU ops on registers, bpf_add\|sub\|...: dst_reg += src_reg */
	51
	52	#define BPF_ALU64_REG(OP, DST, SRC) \
	53	((struct bpf_insn) { \
	54	.code = BPF_ALU64 \| BPF_OP(OP) \| BPF_X, \
	55	.dst_reg = DST, \
	56	.src_reg = SRC, \
	57	.off = 0, \
	58	.imm = 0 })
	59
	60	#define BPF_ALU32_REG(OP, DST, SRC) \
	61	((struct bpf_insn) { \
	62	.code = BPF_ALU \| BPF_OP(OP) \| BPF_X, \
	63	.dst_reg = DST, \
	64	.src_reg = SRC, \
	65	.off = 0, \
	66	.imm = 0 })
	67
	68	/* ALU ops on immediates, bpf_add\|sub\|...: dst_reg += imm32 */
	69
	70	#define BPF_ALU64_IMM(OP, DST, IMM) \
	71	((struct bpf_insn) { \
	72	.code = BPF_ALU64 \| BPF_OP(OP) \| BPF_K, \
	73	.dst_reg = DST, \
	74	.src_reg = 0, \
	75	.off = 0, \
	76	.imm = IMM })
	77
	78	#define BPF_ALU32_IMM(OP, DST, IMM) \
	79	((struct bpf_insn) { \
	80	.code = BPF_ALU \| BPF_OP(OP) \| BPF_K, \
	81	.dst_reg = DST, \
	82	.src_reg = 0, \
	83	.off = 0, \
	84	.imm = IMM })
	85
	86	/* Endianess conversion, cpu_to_{l,b}e(), {l,b}e_to_cpu() */
	87
	88	#define BPF_ENDIAN(TYPE, DST, LEN) \
	89	((struct bpf_insn) { \
	90	.code = BPF_ALU \| BPF_END \| BPF_SRC(TYPE), \
	91	.dst_reg = DST, \
	92	.src_reg = 0, \
	93	.off = 0, \
	94	.imm = LEN })
	95
	96	/* Short form of mov, dst_reg = src_reg */
	97
	98	#define BPF_MOV64_REG(DST, SRC) \
	99	((struct bpf_insn) { \
	100	.code = BPF_ALU64 \| BPF_MOV \| BPF_X, \
	101	.dst_reg = DST, \
	102	.src_reg = SRC, \
	103	.off = 0, \
	104	.imm = 0 })
	105
	106	#define BPF_MOV32_REG(DST, SRC) \
	107	((struct bpf_insn) { \
	108	.code = BPF_ALU \| BPF_MOV \| BPF_X, \
	109	.dst_reg = DST, \
	110	.src_reg = SRC, \
	111	.off = 0, \
	112	.imm = 0 })
	113
	114	/* Short form of mov, dst_reg = imm32 */
	115
	116	#define BPF_MOV64_IMM(DST, IMM) \
	117	((struct bpf_insn) { \
	118	.code = BPF_ALU64 \| BPF_MOV \| BPF_K, \
	119	.dst_reg = DST, \
	120	.src_reg = 0, \
	121	.off = 0, \
	122	.imm = IMM })
	123
	124	#define BPF_MOV32_IMM(DST, IMM) \
	125	((struct bpf_insn) { \
	126	.code = BPF_ALU \| BPF_MOV \| BPF_K, \
	127	.dst_reg = DST, \
	128	.src_reg = 0, \
	129	.off = 0, \
	130	.imm = IMM })
	131
	132	/* BPF_LD_IMM64 macro encodes single 'load 64-bit immediate' insn */
	133	#define BPF_LD_IMM64(DST, IMM) \
	134	BPF_LD_IMM64_RAW(DST, 0, IMM)
	135
	136	#define BPF_LD_IMM64_RAW(DST, SRC, IMM) \
	137	((struct bpf_insn) { \
	138	.code = BPF_LD \| BPF_DW \| BPF_IMM, \
	139	.dst_reg = DST, \
	140	.src_reg = SRC, \
	141	.off = 0, \
	142	.imm = (__u32) (IMM) }), \
	143	((struct bpf_insn) { \
	144	.code = 0, /* zero is reserved opcode */ \
	145	.dst_reg = 0, \
	146	.src_reg = 0, \
	147	.off = 0, \
	148	.imm = ((__u64) (IMM)) >> 32 })
	149
	150	/* pseudo BPF_LD_IMM64 insn used to refer to process-local map_fd */
	151	#define BPF_LD_MAP_FD(DST, MAP_FD) \
	152	BPF_LD_IMM64_RAW(DST, BPF_PSEUDO_MAP_FD, MAP_FD)
	153
	154	/* Short form of mov based on type, BPF_X: dst_reg = src_reg, BPF_K: dst_reg = imm32 */
	155
	156	#define BPF_MOV64_RAW(TYPE, DST, SRC, IMM) \
	157	((struct bpf_insn) { \
	158	.code = BPF_ALU64 \| BPF_MOV \| BPF_SRC(TYPE), \
	159	.dst_reg = DST, \
	160	.src_reg = SRC, \
	161	.off = 0, \
	162	.imm = IMM })
	163
	164	#define BPF_MOV32_RAW(TYPE, DST, SRC, IMM) \
	165	((struct bpf_insn) { \
	166	.code = BPF_ALU \| BPF_MOV \| BPF_SRC(TYPE), \
	167	.dst_reg = DST, \
	168	.src_reg = SRC, \
	169	.off = 0, \
	170	.imm = IMM })
	171
	172	/* Direct packet access, R0 = (uint ) (skb->data + imm32) */
	173
	174	#define BPF_LD_ABS(SIZE, IMM) \
	175	((struct bpf_insn) { \
	176	.code = BPF_LD \| BPF_SIZE(SIZE) \| BPF_ABS, \
	177	.dst_reg = 0, \
	178	.src_reg = 0, \
	179	.off = 0, \
	180	.imm = IMM })
	181
	182	/* Indirect packet access, R0 = (uint ) (skb->data + src_reg + imm32) */
	183
	184	#define BPF_LD_IND(SIZE, SRC, IMM) \
	185	((struct bpf_insn) { \
	186	.code = BPF_LD \| BPF_SIZE(SIZE) \| BPF_IND, \
	187	.dst_reg = 0, \
	188	.src_reg = SRC, \
	189	.off = 0, \
	190	.imm = IMM })
	191
	192	/* Memory load, dst_reg = (uint ) (src_reg + off16) */
	193
	194	#define BPF_LDX_MEM(SIZE, DST, SRC, OFF) \
	195	((struct bpf_insn) { \
	196	.code = BPF_LDX \| BPF_SIZE(SIZE) \| BPF_MEM, \
	197	.dst_reg = DST, \
	198	.src_reg = SRC, \
	199	.off = OFF, \
	200	.imm = 0 })
	201
	202	/* Memory store, (uint ) (dst_reg + off16) = src_reg */
	203
	204	#define BPF_STX_MEM(SIZE, DST, SRC, OFF) \
	205	((struct bpf_insn) { \
	206	.code = BPF_STX \| BPF_SIZE(SIZE) \| BPF_MEM, \
	207	.dst_reg = DST, \
	208	.src_reg = SRC, \
	209	.off = OFF, \
	210	.imm = 0 })
	211
	212	/* Atomic memory add, (uint )(dst_reg + off16) += src_reg */
	213
	214	#define BPF_STX_XADD(SIZE, DST, SRC, OFF) \
	215	((struct bpf_insn) { \
	216	.code = BPF_STX \| BPF_SIZE(SIZE) \| BPF_XADD, \
	217	.dst_reg = DST, \
	218	.src_reg = SRC, \
	219	.off = OFF, \
	220	.imm = 0 })
	221
	222	/* Memory store, (uint ) (dst_reg + off16) = imm32 */
	223
	224	#define BPF_ST_MEM(SIZE, DST, OFF, IMM) \
	225	((struct bpf_insn) { \
	226	.code = BPF_ST \| BPF_SIZE(SIZE) \| BPF_MEM, \
	227	.dst_reg = DST, \
	228	.src_reg = 0, \
	229	.off = OFF, \
	230	.imm = IMM })
	231
	232	/* Conditional jumps against registers, if (dst_reg 'op' src_reg) goto pc + off16 */
	233
	234	#define BPF_JMP_REG(OP, DST, SRC, OFF) \
	235	((struct bpf_insn) { \
	236	.code = BPF_JMP \| BPF_OP(OP) \| BPF_X, \
	237	.dst_reg = DST, \
	238	.src_reg = SRC, \
	239	.off = OFF, \
	240	.imm = 0 })
	241
	242	/* Conditional jumps against immediates, if (dst_reg 'op' imm32) goto pc + off16 */
	243
	244	#define BPF_JMP_IMM(OP, DST, IMM, OFF) \
	245	((struct bpf_insn) { \
	246	.code = BPF_JMP \| BPF_OP(OP) \| BPF_K, \
	247	.dst_reg = DST, \
	248	.src_reg = 0, \
	249	.off = OFF, \
	250	.imm = IMM })
	251
	252	/* Function call */
	253
	254	#define BPF_EMIT_CALL(FUNC) \
	255	((struct bpf_insn) { \
	256	.code = BPF_JMP \| BPF_CALL, \
	257	.dst_reg = 0, \
	258	.src_reg = 0, \
	259	.off = 0, \
	260	.imm = ((FUNC) - __bpf_call_base) })
	261
	262	/* Raw code statement block */
	263
	264	#define BPF_RAW_INSN(CODE, DST, SRC, OFF, IMM) \
	265	((struct bpf_insn) { \
	266	.code = CODE, \
	267	.dst_reg = DST, \
	268	.src_reg = SRC, \
	269	.off = OFF, \
	270	.imm = IMM })
	271
	272	/* Program exit */
	273
	274	#define BPF_EXIT_INSN() \
	275	((struct bpf_insn) { \
	276	.code = BPF_JMP \| BPF_EXIT, \
	277	.dst_reg = 0, \
	278	.src_reg = 0, \
	279	.off = 0, \
	280	.imm = 0 })
	281
	282	/* Internal classic blocks for direct assignment */
	283
	284	#define __BPF_STMT(CODE, K) \
	285	((struct sock_filter) BPF_STMT(CODE, K))
	286
	287	#define __BPF_JUMP(CODE, K, JT, JF) \
	288	((struct sock_filter) BPF_JUMP(CODE, K, JT, JF))
	289
	290	#define bytes_to_bpf_size(bytes) \
	291	({ \
	292	int bpf_size = -EINVAL; \
	293	\
	294	if (bytes == sizeof(u8)) \
	295	bpf_size = BPF_B; \
	296	else if (bytes == sizeof(u16)) \
	297	bpf_size = BPF_H; \
	298	else if (bytes == sizeof(u32)) \
	299	bpf_size = BPF_W; \
	300	else if (bytes == sizeof(u64)) \
	301	bpf_size = BPF_DW; \
	302	\
	303	bpf_size; \
	304	})
	305
	306	#ifdef CONFIG_COMPAT
	307	/* A struct sock_filter is architecture independent. */
	308	struct compat_sock_fprog {
	309	u16 len;
	310	compat_uptr_t filter; /* struct sock_filter * */
	311	};
	312	#endif
	313
	314	struct sock_fprog_kern {
	315	u16 len;
	316	struct sock_filter *filter;
	317	};
	318
	319	struct bpf_binary_header {
	320	unsigned int pages;
	321	u8 image[];
	322	};
	323
	324	struct bpf_prog {
	325	u16 pages; /* Number of allocated pages */
	326	kmemcheck_bitfield_begin(meta);
	327	u16 jited:1, /* Is our filter JIT'ed? */
	328	gpl_compatible:1, /* Is filter GPL compatible? */
	329	cb_access:1, /* Is control block accessed? */
	330	dst_needed:1; /* Do we need dst entry? */
	331	kmemcheck_bitfield_end(meta);
	332	u32 len; /* Number of filter blocks */
	333	enum bpf_prog_type type; /* Type of BPF program */
	334	struct bpf_prog_aux aux; / Auxiliary fields */
	335	struct sock_fprog_kern orig_prog; / Original BPF program */
	336	unsigned int (bpf_func)(const struct sk_buff skb,
	337	const struct bpf_insn *filter);
	338	/* Instructions for interpreter */
	339	union {
	340	struct sock_filter insns[0];
	341	struct bpf_insn insnsi[0];
	342	};
	343	};
	344
	345	struct sk_filter {
	346	atomic_t refcnt;
	347	struct rcu_head rcu;
	348	struct bpf_prog *prog;
	349	};
	350
	351	#define BPF_PROG_RUN(filter, ctx) (*filter->bpf_func)(ctx, filter->insnsi)
	352
	353	#define BPF_SKB_CB_LEN QDISC_CB_PRIV_LEN
	354
	355	static inline u8 bpf_skb_cb(struct sk_buff skb)
	356	{
	357	/* eBPF programs may read/write skb->cb[] area to transfer meta
	358	* data between tail calls. Since this also needs to work with
	359	* tc, that scratch memory is mapped to qdisc_skb_cb's data area.
	360	*
	361	* In some socket filter cases, the cb unfortunately needs to be
	362	* saved/restored so that protocol specific skb->cb[] data won't
	363	* be lost. In any case, due to unpriviledged eBPF programs
	364	* attached to sockets, we need to clear the bpf_skb_cb() area
	365	* to not leak previous contents to user space.
	366	*/
	367	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) != BPF_SKB_CB_LEN);
	368	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) !=
	369	FIELD_SIZEOF(struct qdisc_skb_cb, data));
	370
	371	return qdisc_skb_cb(skb)->data;
	372	}
	373
	374	static inline u32 bpf_prog_run_save_cb(const struct bpf_prog *prog,
	375	struct sk_buff *skb)
	376	{
	377	u8 *cb_data = bpf_skb_cb(skb);
	378	u8 cb_saved[BPF_SKB_CB_LEN];
	379	u32 res;
	380
	381	if (unlikely(prog->cb_access)) {
	382	memcpy(cb_saved, cb_data, sizeof(cb_saved));
	383	memset(cb_data, 0, sizeof(cb_saved));
	384	}
	385
	386	res = BPF_PROG_RUN(prog, skb);
	387
	388	if (unlikely(prog->cb_access))
	389	memcpy(cb_data, cb_saved, sizeof(cb_saved));
	390
	391	return res;
	392	}
	393
	394	static inline u32 bpf_prog_run_clear_cb(const struct bpf_prog *prog,
	395	struct sk_buff *skb)
	396	{
	397	u8 *cb_data = bpf_skb_cb(skb);
	398
	399	if (unlikely(prog->cb_access))
	400	memset(cb_data, 0, BPF_SKB_CB_LEN);
	401
	402	return BPF_PROG_RUN(prog, skb);
	403	}
	404
	405	static inline unsigned int bpf_prog_size(unsigned int proglen)
	406	{
	407	return max(sizeof(struct bpf_prog),
	408	offsetof(struct bpf_prog, insns[proglen]));
	409	}
	410
	411	static inline bool bpf_prog_was_classic(const struct bpf_prog *prog)
	412	{
	413	/* When classic BPF programs have been loaded and the arch
	414	* does not have a classic BPF JIT (anymore), they have been
	415	* converted via bpf_migrate_filter() to eBPF and thus always
	416	* have an unspec program type.
	417	*/
	418	return prog->type == BPF_PROG_TYPE_UNSPEC;
	419	}
	420
	421	#define bpf_classic_proglen(fprog) (fprog->len * sizeof(fprog->filter[0]))
	422
	423	#ifdef CONFIG_DEBUG_SET_MODULE_RONX
	424	static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
	425	{
	426	set_memory_ro((unsigned long)fp, fp->pages);
	427	}
	428
	429	static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
	430	{
	431	set_memory_rw((unsigned long)fp, fp->pages);
	432	}
	433	#else
	434	static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
	435	{
	436	}
	437
	438	static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
	439	{
	440	}
	441	#endif /* CONFIG_DEBUG_SET_MODULE_RONX */
	442
	443	int sk_filter(struct sock sk, struct sk_buff skb);
	444
	445	int bpf_prog_select_runtime(struct bpf_prog *fp);
	446	void bpf_prog_free(struct bpf_prog *fp);
	447
	448	struct bpf_prog *bpf_prog_alloc(unsigned int size, gfp_t gfp_extra_flags);
	449	struct bpf_prog bpf_prog_realloc(struct bpf_prog fp_old, unsigned int size,
	450	gfp_t gfp_extra_flags);
	451	void __bpf_prog_free(struct bpf_prog *fp);
	452
	453	static inline void bpf_prog_unlock_free(struct bpf_prog *fp)
	454	{
	455	bpf_prog_unlock_ro(fp);
	456	__bpf_prog_free(fp);
	457	}
	458
	459	typedef int (bpf_aux_classic_check_t)(struct sock_filter filter,
	460	unsigned int flen);
	461
	462	int bpf_prog_create(struct bpf_prog *pfp, struct sock_fprog_kern fprog);
	463	int bpf_prog_create_from_user(struct bpf_prog *pfp, struct sock_fprog fprog,
	464	bpf_aux_classic_check_t trans, bool save_orig);
	465	void bpf_prog_destroy(struct bpf_prog *fp);
	466
	467	int sk_attach_filter(struct sock_fprog fprog, struct sock sk);
	468	int __sk_attach_filter(struct sock_fprog fprog, struct sock sk,
	469	bool locked);
	470	int sk_attach_bpf(u32 ufd, struct sock *sk);
	471	int sk_reuseport_attach_filter(struct sock_fprog fprog, struct sock sk);
	472	int sk_reuseport_attach_bpf(u32 ufd, struct sock *sk);
	473	int sk_detach_filter(struct sock *sk);
	474	int __sk_detach_filter(struct sock *sk, bool locked);
	475
	476	int sk_get_filter(struct sock sk, struct sock_filter __user filter,
	477	unsigned int len);
	478
	479	bool sk_filter_charge(struct sock sk, struct sk_filter fp);
	480	void sk_filter_uncharge(struct sock sk, struct sk_filter fp);
	481
	482	u64 __bpf_call_base(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
	483	void bpf_int_jit_compile(struct bpf_prog *fp);
	484	bool bpf_helper_changes_skb_data(void *func);
	485
	486	#ifdef CONFIG_BPF_JIT
	487	typedef void (bpf_jit_fill_hole_t)(void area, unsigned int size);
	488
	489	struct bpf_binary_header *
	490	bpf_jit_binary_alloc(unsigned int proglen, u8 **image_ptr,
	491	unsigned int alignment,
	492	bpf_jit_fill_hole_t bpf_fill_ill_insns);
	493	void bpf_jit_binary_free(struct bpf_binary_header *hdr);
	494
	495	void bpf_jit_compile(struct bpf_prog *fp);
	496	void bpf_jit_free(struct bpf_prog *fp);
	497
	498	static inline void bpf_jit_dump(unsigned int flen, unsigned int proglen,
	499	u32 pass, void *image)
	500	{
	501	pr_err("flen=%u proglen=%u pass=%u image=%pK from=%s pid=%d\n", flen,
	502	proglen, pass, image, current->comm, task_pid_nr(current));
	503
	504	if (image)
	505	print_hex_dump(KERN_ERR, "JIT code: ", DUMP_PREFIX_OFFSET,
	506	16, 1, image, proglen, false);
	507	}
	508	#else
	509	static inline void bpf_jit_compile(struct bpf_prog *fp)
	510	{
	511	}
	512
	513	static inline void bpf_jit_free(struct bpf_prog *fp)
	514	{
	515	bpf_prog_unlock_free(fp);
	516	}
	517	#endif /* CONFIG_BPF_JIT */
	518
	519	#define BPF_ANC BIT(15)
	520
	521	static inline bool bpf_needs_clear_a(const struct sock_filter *first)
	522	{
	523	switch (first->code) {
	524	case BPF_RET \| BPF_K:
	525	case BPF_LD \| BPF_W \| BPF_LEN:
	526	return false;
	527
	528	case BPF_LD \| BPF_W \| BPF_ABS:
	529	case BPF_LD \| BPF_H \| BPF_ABS:
	530	case BPF_LD \| BPF_B \| BPF_ABS:
	531	if (first->k == SKF_AD_OFF + SKF_AD_ALU_XOR_X)
	532	return true;
	533	return false;
	534
	535	default:
	536	return true;
	537	}
	538	}
	539
	540	static inline u16 bpf_anc_helper(const struct sock_filter *ftest)
	541	{
	542	BUG_ON(ftest->code & BPF_ANC);
	543
	544	switch (ftest->code) {
	545	case BPF_LD \| BPF_W \| BPF_ABS:
	546	case BPF_LD \| BPF_H \| BPF_ABS:
	547	case BPF_LD \| BPF_B \| BPF_ABS:
	548	#define BPF_ANCILLARY(CODE) case SKF_AD_OFF + SKF_AD_##CODE: \
	549	return BPF_ANC \| SKF_AD_##CODE
	550	switch (ftest->k) {
	551	BPF_ANCILLARY(PROTOCOL);
	552	BPF_ANCILLARY(PKTTYPE);
	553	BPF_ANCILLARY(IFINDEX);
	554	BPF_ANCILLARY(NLATTR);
	555	BPF_ANCILLARY(NLATTR_NEST);
	556	BPF_ANCILLARY(MARK);
	557	BPF_ANCILLARY(QUEUE);
	558	BPF_ANCILLARY(HATYPE);
	559	BPF_ANCILLARY(RXHASH);
	560	BPF_ANCILLARY(CPU);
	561	BPF_ANCILLARY(ALU_XOR_X);
	562	BPF_ANCILLARY(VLAN_TAG);
	563	BPF_ANCILLARY(VLAN_TAG_PRESENT);
	564	BPF_ANCILLARY(PAY_OFFSET);
	565	BPF_ANCILLARY(RANDOM);
	566	BPF_ANCILLARY(VLAN_TPID);
	567	}
	568	/* Fallthrough. */
	569	default:
	570	return ftest->code;
	571	}
	572	}
	573
	574	void bpf_internal_load_pointer_neg_helper(const struct sk_buff skb,
	575	int k, unsigned int size);
	576
	577	static inline void bpf_load_pointer(const struct sk_buff skb, int k,
	578	unsigned int size, void *buffer)
	579	{
	580	if (k >= 0)
	581	return skb_header_pointer(skb, k, size, buffer);
	582
	583	return bpf_internal_load_pointer_neg_helper(skb, k, size);
	584	}
	585
	586	static inline int bpf_tell_extensions(void)
	587	{
	588	return SKF_AD_MAX;
	589	}
	590
	591	#endif /* __LINUX_FILTER_H__ */