^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /* Copyright (c) 2017 - 2018 Covalent IO, Inc. http://covalent.io */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) #include <linux/skmsg.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) #include <linux/filter.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include <linux/bpf.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include <linux/init.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include <linux/wait.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <net/inet_common.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <net/tls.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) int __tcp_bpf_recvmsg(struct sock *sk, struct sk_psock *psock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) struct msghdr *msg, int len, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) struct iov_iter *iter = &msg->msg_iter;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) int peek = flags & MSG_PEEK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) struct sk_msg *msg_rx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) int i, copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) msg_rx = list_first_entry_or_null(&psock->ingress_msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) struct sk_msg, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) while (copied != len) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) struct scatterlist *sge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) if (unlikely(!msg_rx))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) i = msg_rx->sg.start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) int copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) sge = sk_msg_elem(msg_rx, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) copy = sge->length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) page = sg_page(sge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) if (copied + copy > len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) copy = len - copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) copy = copy_page_to_iter(page, sge->offset, copy, iter);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) if (!copy)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) return copied ? copied : -EFAULT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) copied += copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) if (likely(!peek)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) sge->offset += copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) sge->length -= copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) if (!msg_rx->skb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) sk_mem_uncharge(sk, copy);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) msg_rx->sg.size -= copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) if (!sge->length) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) sk_msg_iter_var_next(i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) if (!msg_rx->skb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) /* Lets not optimize peek case if copy_page_to_iter
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * didn't copy the entire length lets just break.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) if (copy != sge->length)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) return copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) sk_msg_iter_var_next(i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) if (copied == len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) } while (i != msg_rx->sg.end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) if (unlikely(peek)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) if (msg_rx == list_last_entry(&psock->ingress_msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) struct sk_msg, list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) msg_rx = list_next_entry(msg_rx, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) msg_rx->sg.start = i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) if (!sge->length && msg_rx->sg.start == msg_rx->sg.end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) list_del(&msg_rx->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) if (msg_rx->skb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) consume_skb(msg_rx->skb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) kfree(msg_rx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) msg_rx = list_first_entry_or_null(&psock->ingress_msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) struct sk_msg, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) return copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) EXPORT_SYMBOL_GPL(__tcp_bpf_recvmsg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) static int bpf_tcp_ingress(struct sock *sk, struct sk_psock *psock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) struct sk_msg *msg, u32 apply_bytes, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) bool apply = apply_bytes;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) struct scatterlist *sge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) u32 size, copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) struct sk_msg *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) int i, ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) tmp = kzalloc(sizeof(*tmp), __GFP_NOWARN | GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) if (unlikely(!tmp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) tmp->sg.start = msg->sg.start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) i = msg->sg.start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) sge = sk_msg_elem(msg, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) size = (apply && apply_bytes < sge->length) ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) apply_bytes : sge->length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) if (!sk_wmem_schedule(sk, size)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) if (!copied)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) ret = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) sk_mem_charge(sk, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) sk_msg_xfer(tmp, msg, i, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) copied += size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) if (sge->length)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) get_page(sk_msg_page(tmp, i));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) sk_msg_iter_var_next(i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) tmp->sg.end = i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) if (apply) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) apply_bytes -= size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) if (!apply_bytes)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) } while (i != msg->sg.end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) if (!ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) msg->sg.start = i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) sk_psock_queue_msg(psock, tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) sk_psock_data_ready(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) sk_msg_free(sk, tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) kfree(tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) static int tcp_bpf_push(struct sock *sk, struct sk_msg *msg, u32 apply_bytes,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) int flags, bool uncharge)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) bool apply = apply_bytes;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) struct scatterlist *sge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) int size, ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) u32 off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) while (1) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) bool has_tx_ulp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) sge = sk_msg_elem(msg, msg->sg.start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) size = (apply && apply_bytes < sge->length) ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) apply_bytes : sge->length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) off = sge->offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) page = sg_page(sge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) tcp_rate_check_app_limited(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) has_tx_ulp = tls_sw_has_ctx_tx(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) if (has_tx_ulp) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) flags |= MSG_SENDPAGE_NOPOLICY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) ret = kernel_sendpage_locked(sk,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) page, off, size, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) ret = do_tcp_sendpages(sk, page, off, size, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) if (ret <= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) if (apply)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) apply_bytes -= ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) msg->sg.size -= ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) sge->offset += ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) sge->length -= ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) if (uncharge)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) sk_mem_uncharge(sk, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) if (ret != size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) size -= ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) off += ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) if (!sge->length) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) sk_msg_iter_next(msg, start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) sg_init_table(sge, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) if (msg->sg.start == msg->sg.end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) if (apply && !apply_bytes)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) static int tcp_bpf_push_locked(struct sock *sk, struct sk_msg *msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) u32 apply_bytes, int flags, bool uncharge)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) ret = tcp_bpf_push(sk, msg, apply_bytes, flags, uncharge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) int tcp_bpf_sendmsg_redir(struct sock *sk, struct sk_msg *msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) u32 bytes, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) bool ingress = sk_msg_to_ingress(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) struct sk_psock *psock = sk_psock_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) if (unlikely(!psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) return -EPIPE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) ret = ingress ? bpf_tcp_ingress(sk, psock, msg, bytes, flags) :
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) tcp_bpf_push_locked(sk, msg, bytes, flags, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) EXPORT_SYMBOL_GPL(tcp_bpf_sendmsg_redir);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) #ifdef CONFIG_BPF_STREAM_PARSER
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) static bool tcp_bpf_stream_read(const struct sock *sk)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) struct sk_psock *psock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) bool empty = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) psock = sk_psock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) if (likely(psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) empty = list_empty(&psock->ingress_msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) return !empty;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) static int tcp_bpf_wait_data(struct sock *sk, struct sk_psock *psock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) int flags, long timeo, int *err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) DEFINE_WAIT_FUNC(wait, woken_wake_function);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) if (sk->sk_shutdown & RCV_SHUTDOWN)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) if (!timeo)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) add_wait_queue(sk_sleep(sk), &wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) sk_set_bit(SOCKWQ_ASYNC_WAITDATA, sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) ret = sk_wait_event(sk, &timeo,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) !list_empty(&psock->ingress_msg) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) !skb_queue_empty(&sk->sk_receive_queue), &wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) sk_clear_bit(SOCKWQ_ASYNC_WAITDATA, sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) remove_wait_queue(sk_sleep(sk), &wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) static int tcp_bpf_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) int nonblock, int flags, int *addr_len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) struct sk_psock *psock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) int copied, ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) if (unlikely(flags & MSG_ERRQUEUE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) return inet_recv_error(sk, msg, len, addr_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) psock = sk_psock_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) if (unlikely(!psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) return tcp_recvmsg(sk, msg, len, nonblock, flags, addr_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) if (!skb_queue_empty(&sk->sk_receive_queue) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) sk_psock_queue_empty(psock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) return tcp_recvmsg(sk, msg, len, nonblock, flags, addr_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) msg_bytes_ready:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) copied = __tcp_bpf_recvmsg(sk, psock, msg, len, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) if (!copied) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) int data, err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) long timeo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) timeo = sock_rcvtimeo(sk, nonblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) data = tcp_bpf_wait_data(sk, psock, flags, timeo, &err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) if (data) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) if (!sk_psock_queue_empty(psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) goto msg_bytes_ready;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) return tcp_recvmsg(sk, msg, len, nonblock, flags, addr_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) ret = err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) copied = -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) ret = copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) static int tcp_bpf_send_verdict(struct sock *sk, struct sk_psock *psock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) struct sk_msg *msg, int *copied, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) bool cork = false, enospc = sk_msg_full(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) struct sock *sk_redir;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) u32 tosend, delta = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) u32 eval = __SK_NONE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) more_data:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) if (psock->eval == __SK_NONE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) /* Track delta in msg size to add/subtract it on SK_DROP from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) * returned to user copied size. This ensures user doesn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) * get a positive return code with msg_cut_data and SK_DROP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) * verdict.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) delta = msg->sg.size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) psock->eval = sk_psock_msg_verdict(sk, psock, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) delta -= msg->sg.size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) if (msg->cork_bytes &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) msg->cork_bytes > msg->sg.size && !enospc) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) psock->cork_bytes = msg->cork_bytes - msg->sg.size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) if (!psock->cork) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) psock->cork = kzalloc(sizeof(*psock->cork),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) GFP_ATOMIC | __GFP_NOWARN);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) if (!psock->cork)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) memcpy(psock->cork, msg, sizeof(*msg));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) tosend = msg->sg.size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) if (psock->apply_bytes && psock->apply_bytes < tosend)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) tosend = psock->apply_bytes;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) switch (psock->eval) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) case __SK_PASS:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) ret = tcp_bpf_push(sk, msg, tosend, flags, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) if (unlikely(ret)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) *copied -= sk_msg_free(sk, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) sk_msg_apply_bytes(psock, tosend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) case __SK_REDIRECT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) sk_redir = psock->sk_redir;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) sk_msg_apply_bytes(psock, tosend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) if (!psock->apply_bytes) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) /* Clean up before releasing the sock lock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) eval = psock->eval;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) psock->eval = __SK_NONE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) psock->sk_redir = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) if (psock->cork) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) cork = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) psock->cork = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) sk_msg_return(sk, msg, msg->sg.size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) ret = tcp_bpf_sendmsg_redir(sk_redir, msg, tosend, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) if (eval == __SK_REDIRECT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) sock_put(sk_redir);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) if (unlikely(ret < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) int free = sk_msg_free_nocharge(sk, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) if (!cork)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) *copied -= free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) if (cork) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) sk_msg_free(sk, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) kfree(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) msg = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) case __SK_DROP:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) sk_msg_free_partial(sk, msg, tosend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) sk_msg_apply_bytes(psock, tosend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) *copied -= (tosend + delta);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) return -EACCES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) if (likely(!ret)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) if (!psock->apply_bytes) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) psock->eval = __SK_NONE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) if (psock->sk_redir) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) sock_put(psock->sk_redir);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) psock->sk_redir = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) if (msg &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) msg->sg.data[msg->sg.start].page_link &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) msg->sg.data[msg->sg.start].length) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) if (eval == __SK_REDIRECT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) sk_mem_charge(sk, msg->sg.size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) goto more_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) static int tcp_bpf_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) struct sk_msg tmp, *msg_tx = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) int copied = 0, err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) struct sk_psock *psock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) long timeo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) int flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) /* Don't let internal do_tcp_sendpages() flags through */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) flags = (msg->msg_flags & ~MSG_SENDPAGE_DECRYPTED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) flags |= MSG_NO_SHARED_FRAGS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) psock = sk_psock_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) if (unlikely(!psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) return tcp_sendmsg(sk, msg, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) while (msg_data_left(msg)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) bool enospc = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) u32 copy, osize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) if (sk->sk_err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) err = -sk->sk_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) copy = msg_data_left(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) if (!sk_stream_memory_free(sk))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) goto wait_for_sndbuf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) if (psock->cork) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) msg_tx = psock->cork;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) msg_tx = &tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) sk_msg_init(msg_tx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) osize = msg_tx->sg.size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) err = sk_msg_alloc(sk, msg_tx, msg_tx->sg.size + copy, msg_tx->sg.end - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) if (err != -ENOSPC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) goto wait_for_memory;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) enospc = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) copy = msg_tx->sg.size - osize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) err = sk_msg_memcopy_from_iter(sk, &msg->msg_iter, msg_tx,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) copy);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) if (err < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) sk_msg_trim(sk, msg_tx, osize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) copied += copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) if (psock->cork_bytes) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) if (size > psock->cork_bytes)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) psock->cork_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) psock->cork_bytes -= size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) if (psock->cork_bytes && !enospc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) /* All cork bytes are accounted, rerun the prog. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) psock->eval = __SK_NONE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) psock->cork_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) err = tcp_bpf_send_verdict(sk, psock, msg_tx, &copied, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) if (unlikely(err < 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) wait_for_sndbuf:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) wait_for_memory:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) err = sk_stream_wait_memory(sk, &timeo);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) if (msg_tx && msg_tx != psock->cork)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) sk_msg_free(sk, msg_tx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) out_err:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) if (err < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) err = sk_stream_error(sk, msg->msg_flags, err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) return copied ? copied : err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) static int tcp_bpf_sendpage(struct sock *sk, struct page *page, int offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) size_t size, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) struct sk_msg tmp, *msg = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) int err = 0, copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) struct sk_psock *psock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) bool enospc = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) psock = sk_psock_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) if (unlikely(!psock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) return tcp_sendpage(sk, page, offset, size, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) lock_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) if (psock->cork) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) msg = psock->cork;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) msg = &tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) sk_msg_init(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) /* Catch case where ring is full and sendpage is stalled. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) if (unlikely(sk_msg_full(msg)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) sk_msg_page_add(msg, page, size, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) sk_mem_charge(sk, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) copied = size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) if (sk_msg_full(msg))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) enospc = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) if (psock->cork_bytes) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) if (size > psock->cork_bytes)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) psock->cork_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) psock->cork_bytes -= size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) if (psock->cork_bytes && !enospc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) /* All cork bytes are accounted, rerun the prog. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) psock->eval = __SK_NONE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) psock->cork_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) err = tcp_bpf_send_verdict(sk, psock, msg, &copied, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) out_err:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) release_sock(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) sk_psock_put(sk, psock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) return copied ? copied : err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) enum {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) TCP_BPF_IPV4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) TCP_BPF_IPV6,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) TCP_BPF_NUM_PROTS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) enum {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) TCP_BPF_BASE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) TCP_BPF_TX,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) TCP_BPF_NUM_CFGS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) static struct proto *tcpv6_prot_saved __read_mostly;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) static DEFINE_SPINLOCK(tcpv6_prot_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) static struct proto tcp_bpf_prots[TCP_BPF_NUM_PROTS][TCP_BPF_NUM_CFGS];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) static void tcp_bpf_rebuild_protos(struct proto prot[TCP_BPF_NUM_CFGS],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) struct proto *base)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) prot[TCP_BPF_BASE] = *base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) prot[TCP_BPF_BASE].close = sock_map_close;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) prot[TCP_BPF_BASE].recvmsg = tcp_bpf_recvmsg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) prot[TCP_BPF_BASE].stream_memory_read = tcp_bpf_stream_read;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) prot[TCP_BPF_TX] = prot[TCP_BPF_BASE];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) prot[TCP_BPF_TX].sendmsg = tcp_bpf_sendmsg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) prot[TCP_BPF_TX].sendpage = tcp_bpf_sendpage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) static void tcp_bpf_check_v6_needs_rebuild(struct proto *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) if (unlikely(ops != smp_load_acquire(&tcpv6_prot_saved))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) spin_lock_bh(&tcpv6_prot_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) if (likely(ops != tcpv6_prot_saved)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) tcp_bpf_rebuild_protos(tcp_bpf_prots[TCP_BPF_IPV6], ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) smp_store_release(&tcpv6_prot_saved, ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) spin_unlock_bh(&tcpv6_prot_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) static int __init tcp_bpf_v4_build_proto(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) tcp_bpf_rebuild_protos(tcp_bpf_prots[TCP_BPF_IPV4], &tcp_prot);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) late_initcall(tcp_bpf_v4_build_proto);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) static int tcp_bpf_assert_proto_ops(struct proto *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) /* In order to avoid retpoline, we make assumptions when we call
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) * into ops if e.g. a psock is not present. Make sure they are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) * indeed valid assumptions.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) return ops->recvmsg == tcp_recvmsg &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) ops->sendmsg == tcp_sendmsg &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) ops->sendpage == tcp_sendpage ? 0 : -ENOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) int config = psock->progs.msg_parser ? TCP_BPF_TX : TCP_BPF_BASE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) if (sk->sk_family == AF_INET6) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) if (tcp_bpf_assert_proto_ops(psock->sk_proto))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) return ERR_PTR(-EINVAL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) tcp_bpf_check_v6_needs_rebuild(psock->sk_proto);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) return &tcp_bpf_prots[family][config];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) /* If a child got cloned from a listening socket that had tcp_bpf
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) * protocol callbacks installed, we need to restore the callbacks to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) * the default ones because the child does not inherit the psock state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) * that tcp_bpf callbacks expect.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) void tcp_bpf_clone(const struct sock *sk, struct sock *newsk)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) struct proto *prot = newsk->sk_prot;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) if (prot == &tcp_bpf_prots[family][TCP_BPF_BASE])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) newsk->sk_prot = sk->sk_prot_creator;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) #endif /* CONFIG_BPF_STREAM_PARSER */
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags