[linux-block.git] / net / ipv4 / bpf_tcp_ca.c

// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2019 Facebook  */

#include <linux/types.h>
#include <linux/bpf_verifier.h>
#include <linux/bpf.h>
#include <linux/btf.h>
#include <linux/filter.h>
#include <net/tcp.h>
#include <net/bpf_sk_storage.h>

static u32 optional_ops[] = {
	offsetof(struct tcp_congestion_ops, init),
	offsetof(struct tcp_congestion_ops, release),
	offsetof(struct tcp_congestion_ops, set_state),
	offsetof(struct tcp_congestion_ops, cwnd_event),
	offsetof(struct tcp_congestion_ops, in_ack_event),
	offsetof(struct tcp_congestion_ops, pkts_acked),
	offsetof(struct tcp_congestion_ops, min_tso_segs),
	offsetof(struct tcp_congestion_ops, sndbuf_expand),
	offsetof(struct tcp_congestion_ops, cong_control),
};

static u32 unsupported_ops[] = {
	offsetof(struct tcp_congestion_ops, get_info),
};

static const struct btf_type *tcp_sock_type;
static u32 tcp_sock_id, sock_id;

static int bpf_tcp_ca_init(struct btf *btf)
{
	s32 type_id;

	type_id = btf_find_by_name_kind(btf, "sock", BTF_KIND_STRUCT);
	if (type_id < 0)
		return -EINVAL;
	sock_id = type_id;

	type_id = btf_find_by_name_kind(btf, "tcp_sock", BTF_KIND_STRUCT);
	if (type_id < 0)
		return -EINVAL;
	tcp_sock_id = type_id;
	tcp_sock_type = btf_type_by_id(btf, tcp_sock_id);

	return 0;
}

static bool is_optional(u32 member_offset)
{
	unsigned int i;

	for (i = 0; i < ARRAY_SIZE(optional_ops); i++) {
		if (member_offset == optional_ops[i])
			return true;
	}

	return false;
}

static bool is_unsupported(u32 member_offset)
{
	unsigned int i;

	for (i = 0; i < ARRAY_SIZE(unsupported_ops); i++) {
		if (member_offset == unsupported_ops[i])
			return true;
	}

	return false;
}

extern struct btf *btf_vmlinux;

static bool bpf_tcp_ca_is_valid_access(int off, int size,
				       enum bpf_access_type type,
				       const struct bpf_prog *prog,
				       struct bpf_insn_access_aux *info)
{
	if (off < 0 || off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
		return false;
	if (type != BPF_READ)
		return false;
	if (off % size != 0)
		return false;

	if (!btf_ctx_access(off, size, type, prog, info))
		return false;

	if (info->reg_type == PTR_TO_BTF_ID && info->btf_id == sock_id)
		/* promote it to tcp_sock */
		info->btf_id = tcp_sock_id;

	return true;
}

static int bpf_tcp_ca_btf_struct_access(struct bpf_verifier_log *log,
					const struct btf *btf,
					const struct btf_type *t, int off,
					int size, enum bpf_access_type atype,
					u32 *next_btf_id)
{
	size_t end;

	if (atype == BPF_READ)
		return btf_struct_access(log, btf, t, off, size, atype, next_btf_id);

	if (t != tcp_sock_type) {
		bpf_log(log, "only read is supported\n");
		return -EACCES;
	}

	switch (off) {
	case bpf_ctx_range(struct inet_connection_sock, icsk_ca_priv):
		end = offsetofend(struct inet_connection_sock, icsk_ca_priv);
		break;
	case offsetof(struct inet_connection_sock, icsk_ack.pending):
		end = offsetofend(struct inet_connection_sock,
				  icsk_ack.pending);
		break;
	case offsetof(struct tcp_sock, snd_cwnd):
		end = offsetofend(struct tcp_sock, snd_cwnd);
		break;
	case offsetof(struct tcp_sock, snd_cwnd_cnt):
		end = offsetofend(struct tcp_sock, snd_cwnd_cnt);
		break;
	case offsetof(struct tcp_sock, snd_ssthresh):
		end = offsetofend(struct tcp_sock, snd_ssthresh);
		break;
	case offsetof(struct tcp_sock, ecn_flags):
		end = offsetofend(struct tcp_sock, ecn_flags);
		break;
	default:
		bpf_log(log, "no write support to tcp_sock at off %d\n", off);
		return -EACCES;
	}

	if (off + size > end) {
		bpf_log(log,
			"write access at off %d with size %d beyond the member of tcp_sock ended at %zu\n",
			off, size, end);
		return -EACCES;
	}

	return NOT_INIT;
}

BPF_CALL_2(bpf_tcp_send_ack, struct tcp_sock *, tp, u32, rcv_nxt)
{
	/* bpf_tcp_ca prog cannot have NULL tp */
	__tcp_send_ack((struct sock *)tp, rcv_nxt);
	return 0;
}

static const struct bpf_func_proto bpf_tcp_send_ack_proto = {
	.func		= bpf_tcp_send_ack,
	.gpl_only	= false,
	/* In case we want to report error later */
	.ret_type	= RET_INTEGER,
	.arg1_type	= ARG_PTR_TO_BTF_ID,
	.arg1_btf_id	= &tcp_sock_id,
	.arg2_type	= ARG_ANYTHING,
};

static const struct bpf_func_proto *
bpf_tcp_ca_get_func_proto(enum bpf_func_id func_id,
			  const struct bpf_prog *prog)
{
	switch (func_id) {
	case BPF_FUNC_tcp_send_ack:
		return &bpf_tcp_send_ack_proto;
	case BPF_FUNC_sk_storage_get:
		return &bpf_sk_storage_get_proto;
	case BPF_FUNC_sk_storage_delete:
		return &bpf_sk_storage_delete_proto;
	default:
		return bpf_base_func_proto(func_id);
	}
}

static const struct bpf_verifier_ops bpf_tcp_ca_verifier_ops = {
	.get_func_proto		= bpf_tcp_ca_get_func_proto,
	.is_valid_access	= bpf_tcp_ca_is_valid_access,
	.btf_struct_access	= bpf_tcp_ca_btf_struct_access,
};

static int bpf_tcp_ca_init_member(const struct btf_type *t,
				  const struct btf_member *member,
				  void *kdata, const void *udata)
{
	const struct tcp_congestion_ops *utcp_ca;
	struct tcp_congestion_ops *tcp_ca;
	int prog_fd;
	u32 moff;

	utcp_ca = (const struct tcp_congestion_ops *)udata;
	tcp_ca = (struct tcp_congestion_ops *)kdata;

	moff = btf_member_bit_offset(t, member) / 8;
	switch (moff) {
	case offsetof(struct tcp_congestion_ops, flags):
		if (utcp_ca->flags & ~TCP_CONG_MASK)
			return -EINVAL;
		tcp_ca->flags = utcp_ca->flags;
		return 1;
	case offsetof(struct tcp_congestion_ops, name):
		if (bpf_obj_name_cpy(tcp_ca->name, utcp_ca->name,
				     sizeof(tcp_ca->name)) <= 0)
			return -EINVAL;
		if (tcp_ca_find(utcp_ca->name))
			return -EEXIST;
		return 1;
	}

	if (!btf_type_resolve_func_ptr(btf_vmlinux, member->type, NULL))
		return 0;

	/* Ensure bpf_prog is provided for compulsory func ptr */
	prog_fd = (int)(*(unsigned long *)(udata + moff));
	if (!prog_fd && !is_optional(moff) && !is_unsupported(moff))
		return -EINVAL;

	return 0;
}

static int bpf_tcp_ca_check_member(const struct btf_type *t,
				   const struct btf_member *member)
{
	if (is_unsupported(btf_member_bit_offset(t, member) / 8))
		return -ENOTSUPP;
	return 0;
}

static int bpf_tcp_ca_reg(void *kdata)
{
	return tcp_register_congestion_control(kdata);
}

static void bpf_tcp_ca_unreg(void *kdata)
{
	tcp_unregister_congestion_control(kdata);
}

/* Avoid sparse warning.  It is only used in bpf_struct_ops.c. */
extern struct bpf_struct_ops bpf_tcp_congestion_ops;

struct bpf_struct_ops bpf_tcp_congestion_ops = {
	.verifier_ops = &bpf_tcp_ca_verifier_ops,
	.reg = bpf_tcp_ca_reg,
	.unreg = bpf_tcp_ca_unreg,
	.check_member = bpf_tcp_ca_check_member,
	.init_member = bpf_tcp_ca_init_member,
	.init = bpf_tcp_ca_init,
	.name = "tcp_congestion_ops",
};
Commit	Line	Data
0baf26b0 MKL	1	// SPDX-License-Identifier: GPL-2.0
	2	/* Copyright (c) 2019 Facebook */
	3
	4	#include <linux/types.h>
	5	#include <linux/bpf_verifier.h>
	6	#include <linux/bpf.h>
	7	#include <linux/btf.h>
	8	#include <linux/filter.h>
	9	#include <net/tcp.h>
ab14fd4e	10	#include <net/bpf_sk_storage.h>
0baf26b0 MKL	11
	12	static u32 optional_ops[] = {
	13	offsetof(struct tcp_congestion_ops, init),
	14	offsetof(struct tcp_congestion_ops, release),
	15	offsetof(struct tcp_congestion_ops, set_state),
	16	offsetof(struct tcp_congestion_ops, cwnd_event),
	17	offsetof(struct tcp_congestion_ops, in_ack_event),
	18	offsetof(struct tcp_congestion_ops, pkts_acked),
	19	offsetof(struct tcp_congestion_ops, min_tso_segs),
	20	offsetof(struct tcp_congestion_ops, sndbuf_expand),
	21	offsetof(struct tcp_congestion_ops, cong_control),
	22	};
	23
	24	static u32 unsupported_ops[] = {
	25	offsetof(struct tcp_congestion_ops, get_info),
	26	};
	27
	28	static const struct btf_type *tcp_sock_type;
	29	static u32 tcp_sock_id, sock_id;
	30
	31	static int bpf_tcp_ca_init(struct btf *btf)
	32	{
	33	s32 type_id;
	34
	35	type_id = btf_find_by_name_kind(btf, "sock", BTF_KIND_STRUCT);
	36	if (type_id < 0)
	37	return -EINVAL;
	38	sock_id = type_id;
	39
	40	type_id = btf_find_by_name_kind(btf, "tcp_sock", BTF_KIND_STRUCT);
	41	if (type_id < 0)
	42	return -EINVAL;
	43	tcp_sock_id = type_id;
	44	tcp_sock_type = btf_type_by_id(btf, tcp_sock_id);
	45
	46	return 0;
	47	}
	48
	49	static bool is_optional(u32 member_offset)
	50	{
	51	unsigned int i;
	52
	53	for (i = 0; i < ARRAY_SIZE(optional_ops); i++) {
	54	if (member_offset == optional_ops[i])
	55	return true;
	56	}
	57
	58	return false;
	59	}
	60
	61	static bool is_unsupported(u32 member_offset)
	62	{
	63	unsigned int i;
	64
	65	for (i = 0; i < ARRAY_SIZE(unsupported_ops); i++) {
	66	if (member_offset == unsupported_ops[i])
	67	return true;
	68	}
	69
	70	return false;
	71	}
	72
	73	extern struct btf *btf_vmlinux;
	74
75	static bool bpf_tcp_ca_is_valid_access(int off, int size,
76	enum bpf_access_type type,
77	const struct bpf_prog *prog,
78	struct bpf_insn_access_aux *info)
79	{
80	if (off < 0 \|\| off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
81	return false;
82	if (type != BPF_READ)
83	return false;
84	if (off % size != 0)
85	return false;
86
87	if (!btf_ctx_access(off, size, type, prog, info))
88	return false;
89
90	if (info->reg_type == PTR_TO_BTF_ID && info->btf_id == sock_id)
91	/* promote it to tcp_sock */
92	info->btf_id = tcp_sock_id;
93
94	return true;
95	}
96
97	static int bpf_tcp_ca_btf_struct_access(struct bpf_verifier_log *log,
22dc4a0f	98	const struct btf *btf,
0baf26b0 MKL	99	const struct btf_type *t, int off,
	100	int size, enum bpf_access_type atype,
	101	u32 *next_btf_id)
	102	{
	103	size_t end;
	104
	105	if (atype == BPF_READ)
22dc4a0f	106	return btf_struct_access(log, btf, t, off, size, atype, next_btf_id);
0baf26b0 MKL	107
	108	if (t != tcp_sock_type) {
	109	bpf_log(log, "only read is supported\n");
	110	return -EACCES;
	111	}
	112
	113	switch (off) {
	114	case bpf_ctx_range(struct inet_connection_sock, icsk_ca_priv):
	115	end = offsetofend(struct inet_connection_sock, icsk_ca_priv);
	116	break;
	117	case offsetof(struct inet_connection_sock, icsk_ack.pending):
	118	end = offsetofend(struct inet_connection_sock,
	119	icsk_ack.pending);
	120	break;
	121	case offsetof(struct tcp_sock, snd_cwnd):
	122	end = offsetofend(struct tcp_sock, snd_cwnd);
	123	break;
	124	case offsetof(struct tcp_sock, snd_cwnd_cnt):
	125	end = offsetofend(struct tcp_sock, snd_cwnd_cnt);
	126	break;
	127	case offsetof(struct tcp_sock, snd_ssthresh):
	128	end = offsetofend(struct tcp_sock, snd_ssthresh);
	129	break;
	130	case offsetof(struct tcp_sock, ecn_flags):
	131	end = offsetofend(struct tcp_sock, ecn_flags);
	132	break;
	133	default:
	134	bpf_log(log, "no write support to tcp_sock at off %d\n", off);
	135	return -EACCES;
	136	}
	137
	138	if (off + size > end) {
	139	bpf_log(log,
	140	"write access at off %d with size %d beyond the member of tcp_sock ended at %zu\n",
	141	off, size, end);
	142	return -EACCES;
	143	}
	144
	145	return NOT_INIT;
	146	}
	147
206057fe MKL	148	BPF_CALL_2(bpf_tcp_send_ack, struct tcp_sock *, tp, u32, rcv_nxt)
	149	{
	150	/* bpf_tcp_ca prog cannot have NULL tp */
	151	__tcp_send_ack((struct sock *)tp, rcv_nxt);
	152	return 0;
	153	}
	154
	155	static const struct bpf_func_proto bpf_tcp_send_ack_proto = {
	156	.func = bpf_tcp_send_ack,
	157	.gpl_only = false,
	158	/* In case we want to report error later */
	159	.ret_type = RET_INTEGER,
	160	.arg1_type = ARG_PTR_TO_BTF_ID,
9436ef6e	161	.arg1_btf_id = &tcp_sock_id,
206057fe	162	.arg2_type = ARG_ANYTHING,
206057fe MKL	163	};
206057fe MKL	164
0baf26b0 MKL	165	static const struct bpf_func_proto *
	166	bpf_tcp_ca_get_func_proto(enum bpf_func_id func_id,
	167	const struct bpf_prog *prog)
	168	{
206057fe MKL	169	switch (func_id) {
	170	case BPF_FUNC_tcp_send_ack:
	171	return &bpf_tcp_send_ack_proto;
ab14fd4e	172	case BPF_FUNC_sk_storage_get:
592a3498	173	return &bpf_sk_storage_get_proto;
ab14fd4e	174	case BPF_FUNC_sk_storage_delete:
592a3498	175	return &bpf_sk_storage_delete_proto;
206057fe MKL	176	default:
	177	return bpf_base_func_proto(func_id);
	178	}
0baf26b0 MKL	179	}
	180
	181	static const struct bpf_verifier_ops bpf_tcp_ca_verifier_ops = {
	182	.get_func_proto = bpf_tcp_ca_get_func_proto,
	183	.is_valid_access = bpf_tcp_ca_is_valid_access,
	184	.btf_struct_access = bpf_tcp_ca_btf_struct_access,
	185	};
	186
	187	static int bpf_tcp_ca_init_member(const struct btf_type *t,
	188	const struct btf_member *member,
	189	void kdata, const void udata)
	190	{
	191	const struct tcp_congestion_ops *utcp_ca;
	192	struct tcp_congestion_ops *tcp_ca;
0baf26b0 MKL	193	int prog_fd;
	194	u32 moff;
	195
	196	utcp_ca = (const struct tcp_congestion_ops *)udata;
	197	tcp_ca = (struct tcp_congestion_ops *)kdata;
	198
	199	moff = btf_member_bit_offset(t, member) / 8;
	200	switch (moff) {
	201	case offsetof(struct tcp_congestion_ops, flags):
	202	if (utcp_ca->flags & ~TCP_CONG_MASK)
	203	return -EINVAL;
	204	tcp_ca->flags = utcp_ca->flags;
	205	return 1;
	206	case offsetof(struct tcp_congestion_ops, name):
8e7ae251 MKL	207	if (bpf_obj_name_cpy(tcp_ca->name, utcp_ca->name,
8e7ae251 MKL	208	sizeof(tcp_ca->name)) <= 0)
0baf26b0 MKL	209	return -EINVAL;
	210	if (tcp_ca_find(utcp_ca->name))
	211	return -EEXIST;
0baf26b0 MKL	212	return 1;
	213	}
	214
	215	if (!btf_type_resolve_func_ptr(btf_vmlinux, member->type, NULL))
	216	return 0;
	217
	218	/* Ensure bpf_prog is provided for compulsory func ptr */
	219	prog_fd = (int)((unsigned long )(udata + moff));
	220	if (!prog_fd && !is_optional(moff) && !is_unsupported(moff))
	221	return -EINVAL;
	222
	223	return 0;
	224	}
	225
	226	static int bpf_tcp_ca_check_member(const struct btf_type *t,
	227	const struct btf_member *member)
	228	{
	229	if (is_unsupported(btf_member_bit_offset(t, member) / 8))
	230	return -ENOTSUPP;
	231	return 0;
	232	}
	233
	234	static int bpf_tcp_ca_reg(void *kdata)
	235	{
	236	return tcp_register_congestion_control(kdata);
	237	}
	238
	239	static void bpf_tcp_ca_unreg(void *kdata)
	240	{
	241	tcp_unregister_congestion_control(kdata);
	242	}
	243
	244	/* Avoid sparse warning. It is only used in bpf_struct_ops.c. */
	245	extern struct bpf_struct_ops bpf_tcp_congestion_ops;
	246
	247	struct bpf_struct_ops bpf_tcp_congestion_ops = {
	248	.verifier_ops = &bpf_tcp_ca_verifier_ops,
	249	.reg = bpf_tcp_ca_reg,
	250	.unreg = bpf_tcp_ca_unreg,
	251	.check_member = bpf_tcp_ca_check_member,
	252	.init_member = bpf_tcp_ca_init_member,
	253	.init = bpf_tcp_ca_init,
	254	.name = "tcp_congestion_ops",
	255	};