^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) #include <linux/rcupdate.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) #include <linux/spinlock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) #include <linux/jiffies.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include <linux/cache.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include <linux/init.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include <linux/tcp.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <linux/hash.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/tcp_metrics.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <linux/vmalloc.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <net/inet_connection_sock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <net/net_namespace.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <net/request_sock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <net/inetpeer.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <net/sock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <net/ipv6.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <net/dst.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <net/tcp.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include <net/genetlink.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) static struct tcp_metrics_block *__tcp_get_metrics(const struct inetpeer_addr *saddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) const struct inetpeer_addr *daddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) struct net *net, unsigned int hash);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) struct tcp_fastopen_metrics {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) u16 mss;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) u16 syn_loss:10, /* Recurring Fast Open SYN losses */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) try_exp:2; /* Request w/ exp. option (once) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) unsigned long last_syn_loss; /* Last Fast Open SYN loss */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) struct tcp_fastopen_cookie cookie;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) /* TCP_METRIC_MAX includes 2 extra fields for userspace compatibility
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * Kernel only stores RTT and RTTVAR in usec resolution
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) #define TCP_METRIC_MAX_KERNEL (TCP_METRIC_MAX - 2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) struct tcp_metrics_block {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) struct tcp_metrics_block __rcu *tcpm_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) possible_net_t tcpm_net;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) struct inetpeer_addr tcpm_saddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) struct inetpeer_addr tcpm_daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) unsigned long tcpm_stamp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) u32 tcpm_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) u32 tcpm_vals[TCP_METRIC_MAX_KERNEL + 1];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) struct tcp_fastopen_metrics tcpm_fastopen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) struct rcu_head rcu_head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) static inline struct net *tm_net(struct tcp_metrics_block *tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) return read_pnet(&tm->tcpm_net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) static bool tcp_metric_locked(struct tcp_metrics_block *tm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) enum tcp_metric_index idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) return tm->tcpm_lock & (1 << idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) static u32 tcp_metric_get(struct tcp_metrics_block *tm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) enum tcp_metric_index idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) return tm->tcpm_vals[idx];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) static void tcp_metric_set(struct tcp_metrics_block *tm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) enum tcp_metric_index idx,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) u32 val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) tm->tcpm_vals[idx] = val;
^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) static bool addr_same(const struct inetpeer_addr *a,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) const struct inetpeer_addr *b)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) return inetpeer_addr_cmp(a, b) == 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) struct tcpm_hash_bucket {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) struct tcp_metrics_block __rcu *chain;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) static struct tcpm_hash_bucket *tcp_metrics_hash __read_mostly;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) static unsigned int tcp_metrics_hash_log __read_mostly;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) static DEFINE_SPINLOCK(tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) static void tcpm_suck_dst(struct tcp_metrics_block *tm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) const struct dst_entry *dst,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) bool fastopen_clear)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) u32 msval;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) u32 val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) tm->tcpm_stamp = jiffies;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) val = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) if (dst_metric_locked(dst, RTAX_RTT))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) val |= 1 << TCP_METRIC_RTT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) if (dst_metric_locked(dst, RTAX_RTTVAR))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) val |= 1 << TCP_METRIC_RTTVAR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) if (dst_metric_locked(dst, RTAX_SSTHRESH))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) val |= 1 << TCP_METRIC_SSTHRESH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) if (dst_metric_locked(dst, RTAX_CWND))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) val |= 1 << TCP_METRIC_CWND;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) if (dst_metric_locked(dst, RTAX_REORDERING))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) val |= 1 << TCP_METRIC_REORDERING;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) tm->tcpm_lock = val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) msval = dst_metric_raw(dst, RTAX_RTT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) tm->tcpm_vals[TCP_METRIC_RTT] = msval * USEC_PER_MSEC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) msval = dst_metric_raw(dst, RTAX_RTTVAR);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) tm->tcpm_vals[TCP_METRIC_RTTVAR] = msval * USEC_PER_MSEC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) tm->tcpm_vals[TCP_METRIC_SSTHRESH] = dst_metric_raw(dst, RTAX_SSTHRESH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) tm->tcpm_vals[TCP_METRIC_CWND] = dst_metric_raw(dst, RTAX_CWND);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) tm->tcpm_vals[TCP_METRIC_REORDERING] = dst_metric_raw(dst, RTAX_REORDERING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) if (fastopen_clear) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) tm->tcpm_fastopen.mss = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) tm->tcpm_fastopen.syn_loss = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) tm->tcpm_fastopen.try_exp = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) tm->tcpm_fastopen.cookie.exp = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) tm->tcpm_fastopen.cookie.len = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) #define TCP_METRICS_TIMEOUT (60 * 60 * HZ)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) static void tcpm_check_stamp(struct tcp_metrics_block *tm, struct dst_entry *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) if (tm && unlikely(time_after(jiffies, tm->tcpm_stamp + TCP_METRICS_TIMEOUT)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) tcpm_suck_dst(tm, dst, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) #define TCP_METRICS_RECLAIM_DEPTH 5
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) #define TCP_METRICS_RECLAIM_PTR (struct tcp_metrics_block *) 0x1UL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) #define deref_locked(p) \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) rcu_dereference_protected(p, lockdep_is_held(&tcp_metrics_lock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) static struct tcp_metrics_block *tcpm_new(struct dst_entry *dst,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) struct inetpeer_addr *saddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) struct inetpeer_addr *daddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) unsigned int hash)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) struct net *net;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) bool reclaim = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) spin_lock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) net = dev_net(dst->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) /* While waiting for the spin-lock the cache might have been populated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * with this entry and so we have to check again.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) tm = __tcp_get_metrics(saddr, daddr, net, hash);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) if (tm == TCP_METRICS_RECLAIM_PTR) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) reclaim = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) tm = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) if (tm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) tcpm_check_stamp(tm, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) if (unlikely(reclaim)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) struct tcp_metrics_block *oldest;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) oldest = deref_locked(tcp_metrics_hash[hash].chain);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) for (tm = deref_locked(oldest->tcpm_next); tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) tm = deref_locked(tm->tcpm_next)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) if (time_before(tm->tcpm_stamp, oldest->tcpm_stamp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) oldest = tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) tm = oldest;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) tm = kmalloc(sizeof(*tm), GFP_ATOMIC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) if (!tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) write_pnet(&tm->tcpm_net, net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) tm->tcpm_saddr = *saddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) tm->tcpm_daddr = *daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) tcpm_suck_dst(tm, dst, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) if (likely(!reclaim)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) tm->tcpm_next = tcp_metrics_hash[hash].chain;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) rcu_assign_pointer(tcp_metrics_hash[hash].chain, tm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) spin_unlock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) return tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) static struct tcp_metrics_block *tcp_get_encode(struct tcp_metrics_block *tm, int depth)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) if (tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) return tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) if (depth > TCP_METRICS_RECLAIM_DEPTH)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) return TCP_METRICS_RECLAIM_PTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) static struct tcp_metrics_block *__tcp_get_metrics(const struct inetpeer_addr *saddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) const struct inetpeer_addr *daddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) struct net *net, unsigned int hash)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) int depth = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) for (tm = rcu_dereference(tcp_metrics_hash[hash].chain); tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) tm = rcu_dereference(tm->tcpm_next)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) if (addr_same(&tm->tcpm_saddr, saddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) addr_same(&tm->tcpm_daddr, daddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) net_eq(tm_net(tm), net))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) depth++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) return tcp_get_encode(tm, depth);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) static struct tcp_metrics_block *__tcp_get_metrics_req(struct request_sock *req,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) struct dst_entry *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) struct inetpeer_addr saddr, daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) unsigned int hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) struct net *net;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) saddr.family = req->rsk_ops->family;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) daddr.family = req->rsk_ops->family;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) switch (daddr.family) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) case AF_INET:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) inetpeer_set_addr_v4(&saddr, inet_rsk(req)->ir_loc_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) inetpeer_set_addr_v4(&daddr, inet_rsk(req)->ir_rmt_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) hash = ipv4_addr_hash(inet_rsk(req)->ir_rmt_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) case AF_INET6:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) inetpeer_set_addr_v6(&saddr, &inet_rsk(req)->ir_v6_loc_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) inetpeer_set_addr_v6(&daddr, &inet_rsk(req)->ir_v6_rmt_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) hash = ipv6_addr_hash(&inet_rsk(req)->ir_v6_rmt_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) net = dev_net(dst->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) hash ^= net_hash_mix(net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) hash = hash_32(hash, tcp_metrics_hash_log);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) for (tm = rcu_dereference(tcp_metrics_hash[hash].chain); tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) tm = rcu_dereference(tm->tcpm_next)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) if (addr_same(&tm->tcpm_saddr, &saddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) addr_same(&tm->tcpm_daddr, &daddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) net_eq(tm_net(tm), net))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) tcpm_check_stamp(tm, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) return tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) static struct tcp_metrics_block *tcp_get_metrics(struct sock *sk,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) struct dst_entry *dst,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) bool create)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) struct inetpeer_addr saddr, daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) unsigned int hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) struct net *net;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) if (sk->sk_family == AF_INET) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) inetpeer_set_addr_v4(&saddr, inet_sk(sk)->inet_saddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) inetpeer_set_addr_v4(&daddr, inet_sk(sk)->inet_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) hash = ipv4_addr_hash(inet_sk(sk)->inet_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) #if IS_ENABLED(CONFIG_IPV6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) else if (sk->sk_family == AF_INET6) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) if (ipv6_addr_v4mapped(&sk->sk_v6_daddr)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) inetpeer_set_addr_v4(&saddr, inet_sk(sk)->inet_saddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) inetpeer_set_addr_v4(&daddr, inet_sk(sk)->inet_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) hash = ipv4_addr_hash(inet_sk(sk)->inet_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) inetpeer_set_addr_v6(&saddr, &sk->sk_v6_rcv_saddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) inetpeer_set_addr_v6(&daddr, &sk->sk_v6_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) hash = ipv6_addr_hash(&sk->sk_v6_daddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) net = dev_net(dst->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) hash ^= net_hash_mix(net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) hash = hash_32(hash, tcp_metrics_hash_log);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) tm = __tcp_get_metrics(&saddr, &daddr, net, hash);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) if (tm == TCP_METRICS_RECLAIM_PTR)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) tm = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) if (!tm && create)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) tm = tcpm_new(dst, &saddr, &daddr, hash);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) tcpm_check_stamp(tm, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) return tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) /* Save metrics learned by this TCP session. This function is called
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) * only, when TCP finishes successfully i.e. when it enters TIME-WAIT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) * or goes from LAST-ACK to CLOSE.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) void tcp_update_metrics(struct sock *sk)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) const struct inet_connection_sock *icsk = inet_csk(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) struct dst_entry *dst = __sk_dst_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) struct tcp_sock *tp = tcp_sk(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) struct net *net = sock_net(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) unsigned long rtt;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) u32 val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) int m;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) sk_dst_confirm(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) if (net->ipv4.sysctl_tcp_nometrics_save || !dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) if (icsk->icsk_backoff || !tp->srtt_us) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) /* This session failed to estimate rtt. Why?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) * Probably, no packets returned in time. Reset our
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) * results.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) tm = tcp_get_metrics(sk, dst, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) if (tm && !tcp_metric_locked(tm, TCP_METRIC_RTT))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) tcp_metric_set(tm, TCP_METRIC_RTT, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) tm = tcp_get_metrics(sk, dst, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) if (!tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) rtt = tcp_metric_get(tm, TCP_METRIC_RTT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) m = rtt - tp->srtt_us;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) /* If newly calculated rtt larger than stored one, store new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) * one. Otherwise, use EWMA. Remember, rtt overestimation is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) * always better than underestimation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) if (!tcp_metric_locked(tm, TCP_METRIC_RTT)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) if (m <= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) rtt = tp->srtt_us;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) rtt -= (m >> 3);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) tcp_metric_set(tm, TCP_METRIC_RTT, rtt);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) if (!tcp_metric_locked(tm, TCP_METRIC_RTTVAR)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) unsigned long var;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) if (m < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) m = -m;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) /* Scale deviation to rttvar fixed point */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) m >>= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) if (m < tp->mdev_us)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) m = tp->mdev_us;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) var = tcp_metric_get(tm, TCP_METRIC_RTTVAR);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) if (m >= var)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) var = m;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) var -= (var - m) >> 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) tcp_metric_set(tm, TCP_METRIC_RTTVAR, var);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) if (tcp_in_initial_slowstart(tp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) /* Slow start still did not finish. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) if (!net->ipv4.sysctl_tcp_no_ssthresh_metrics_save &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) !tcp_metric_locked(tm, TCP_METRIC_SSTHRESH)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) val = tcp_metric_get(tm, TCP_METRIC_SSTHRESH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) if (val && (tp->snd_cwnd >> 1) > val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) tcp_metric_set(tm, TCP_METRIC_SSTHRESH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) tp->snd_cwnd >> 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) if (!tcp_metric_locked(tm, TCP_METRIC_CWND)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) val = tcp_metric_get(tm, TCP_METRIC_CWND);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) if (tp->snd_cwnd > val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) tcp_metric_set(tm, TCP_METRIC_CWND,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) tp->snd_cwnd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) } else if (!tcp_in_slow_start(tp) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) icsk->icsk_ca_state == TCP_CA_Open) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) /* Cong. avoidance phase, cwnd is reliable. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) if (!net->ipv4.sysctl_tcp_no_ssthresh_metrics_save &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) !tcp_metric_locked(tm, TCP_METRIC_SSTHRESH))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) tcp_metric_set(tm, TCP_METRIC_SSTHRESH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) max(tp->snd_cwnd >> 1, tp->snd_ssthresh));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) if (!tcp_metric_locked(tm, TCP_METRIC_CWND)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) val = tcp_metric_get(tm, TCP_METRIC_CWND);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) tcp_metric_set(tm, TCP_METRIC_CWND, (val + tp->snd_cwnd) >> 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) /* Else slow start did not finish, cwnd is non-sense,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) * ssthresh may be also invalid.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) if (!tcp_metric_locked(tm, TCP_METRIC_CWND)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) val = tcp_metric_get(tm, TCP_METRIC_CWND);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) tcp_metric_set(tm, TCP_METRIC_CWND,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) (val + tp->snd_ssthresh) >> 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) if (!net->ipv4.sysctl_tcp_no_ssthresh_metrics_save &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) !tcp_metric_locked(tm, TCP_METRIC_SSTHRESH)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) val = tcp_metric_get(tm, TCP_METRIC_SSTHRESH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) if (val && tp->snd_ssthresh > val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) tcp_metric_set(tm, TCP_METRIC_SSTHRESH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) tp->snd_ssthresh);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) if (!tcp_metric_locked(tm, TCP_METRIC_REORDERING)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) val = tcp_metric_get(tm, TCP_METRIC_REORDERING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) if (val < tp->reordering &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) tp->reordering != net->ipv4.sysctl_tcp_reordering)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) tcp_metric_set(tm, TCP_METRIC_REORDERING,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) tp->reordering);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) tm->tcpm_stamp = jiffies;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) /* Initialize metrics on socket. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) void tcp_init_metrics(struct sock *sk)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) struct dst_entry *dst = __sk_dst_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) struct tcp_sock *tp = tcp_sk(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) struct net *net = sock_net(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) u32 val, crtt = 0; /* cached RTT scaled by 8 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) sk_dst_confirm(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) if (!dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) goto reset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) tm = tcp_get_metrics(sk, dst, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) if (!tm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) goto reset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) if (tcp_metric_locked(tm, TCP_METRIC_CWND))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) tp->snd_cwnd_clamp = tcp_metric_get(tm, TCP_METRIC_CWND);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) val = net->ipv4.sysctl_tcp_no_ssthresh_metrics_save ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) 0 : tcp_metric_get(tm, TCP_METRIC_SSTHRESH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) if (val) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) tp->snd_ssthresh = val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) if (tp->snd_ssthresh > tp->snd_cwnd_clamp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) tp->snd_ssthresh = tp->snd_cwnd_clamp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) /* ssthresh may have been reduced unnecessarily during.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) * 3WHS. Restore it back to its initial default.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) val = tcp_metric_get(tm, TCP_METRIC_REORDERING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) if (val && tp->reordering != val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) tp->reordering = val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) crtt = tcp_metric_get(tm, TCP_METRIC_RTT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) reset:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) /* The initial RTT measurement from the SYN/SYN-ACK is not ideal
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) * to seed the RTO for later data packets because SYN packets are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) * small. Use the per-dst cached values to seed the RTO but keep
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) * the RTT estimator variables intact (e.g., srtt, mdev, rttvar).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) * Later the RTO will be updated immediately upon obtaining the first
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) * data RTT sample (tcp_rtt_estimator()). Hence the cached RTT only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) * influences the first RTO but not later RTT estimation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) * But if RTT is not available from the SYN (due to retransmits or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) * syn cookies) or the cache, force a conservative 3secs timeout.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) * A bit of theory. RTT is time passed after "normal" sized packet
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) * is sent until it is ACKed. In normal circumstances sending small
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) * packets force peer to delay ACKs and calculation is correct too.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) * The algorithm is adaptive and, provided we follow specs, it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) * NEVER underestimate RTT. BUT! If peer tries to make some clever
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) * tricks sort of "quick acks" for time long enough to decrease RTT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) * to low value, and then abruptly stops to do it and starts to delay
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) * ACKs, wait for troubles.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) if (crtt > tp->srtt_us) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) /* Set RTO like tcp_rtt_estimator(), but from cached RTT. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) crtt /= 8 * USEC_PER_SEC / HZ;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) inet_csk(sk)->icsk_rto = crtt + max(2 * crtt, tcp_rto_min(sk));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) } else if (tp->srtt_us == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) /* RFC6298: 5.7 We've failed to get a valid RTT sample from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) * 3WHS. This is most likely due to retransmission,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) * including spurious one. Reset the RTO back to 3secs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) * from the more aggressive 1sec to avoid more spurious
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) * retransmission.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) tp->rttvar_us = jiffies_to_usecs(TCP_TIMEOUT_FALLBACK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) tp->mdev_us = tp->mdev_max_us = tp->rttvar_us;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) inet_csk(sk)->icsk_rto = TCP_TIMEOUT_FALLBACK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) bool tcp_peer_is_proven(struct request_sock *req, struct dst_entry *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) bool ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) if (!dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) tm = __tcp_get_metrics_req(req, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) if (tm && tcp_metric_get(tm, TCP_METRIC_RTT))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) ret = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) ret = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) static DEFINE_SEQLOCK(fastopen_seqlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) void tcp_fastopen_cache_get(struct sock *sk, u16 *mss,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) struct tcp_fastopen_cookie *cookie)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) tm = tcp_get_metrics(sk, __sk_dst_get(sk), false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) if (tm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) struct tcp_fastopen_metrics *tfom = &tm->tcpm_fastopen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) unsigned int seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) seq = read_seqbegin(&fastopen_seqlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) if (tfom->mss)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) *mss = tfom->mss;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) *cookie = tfom->cookie;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) if (cookie->len <= 0 && tfom->try_exp == 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) cookie->exp = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) } while (read_seqretry(&fastopen_seqlock, seq));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) void tcp_fastopen_cache_set(struct sock *sk, u16 mss,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) struct tcp_fastopen_cookie *cookie, bool syn_lost,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) u16 try_exp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) struct dst_entry *dst = __sk_dst_get(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) if (!dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) tm = tcp_get_metrics(sk, dst, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) if (tm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) struct tcp_fastopen_metrics *tfom = &tm->tcpm_fastopen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) write_seqlock_bh(&fastopen_seqlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (mss)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) tfom->mss = mss;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) if (cookie && cookie->len > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) tfom->cookie = *cookie;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) else if (try_exp > tfom->try_exp &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) tfom->cookie.len <= 0 && !tfom->cookie.exp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) tfom->try_exp = try_exp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) if (syn_lost) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) ++tfom->syn_loss;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) tfom->last_syn_loss = jiffies;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) tfom->syn_loss = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) write_sequnlock_bh(&fastopen_seqlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) rcu_read_unlock();
^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) static struct genl_family tcp_metrics_nl_family;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) static const struct nla_policy tcp_metrics_nl_policy[TCP_METRICS_ATTR_MAX + 1] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) [TCP_METRICS_ATTR_ADDR_IPV4] = { .type = NLA_U32, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) [TCP_METRICS_ATTR_ADDR_IPV6] = { .type = NLA_BINARY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) .len = sizeof(struct in6_addr), },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) /* Following attributes are not received for GET/DEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) * we keep them for reference
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) #if 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) [TCP_METRICS_ATTR_AGE] = { .type = NLA_MSECS, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) [TCP_METRICS_ATTR_TW_TSVAL] = { .type = NLA_U32, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) [TCP_METRICS_ATTR_TW_TS_STAMP] = { .type = NLA_S32, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) [TCP_METRICS_ATTR_VALS] = { .type = NLA_NESTED, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) [TCP_METRICS_ATTR_FOPEN_MSS] = { .type = NLA_U16, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) [TCP_METRICS_ATTR_FOPEN_SYN_DROPS] = { .type = NLA_U16, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) [TCP_METRICS_ATTR_FOPEN_SYN_DROP_TS] = { .type = NLA_MSECS, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) [TCP_METRICS_ATTR_FOPEN_COOKIE] = { .type = NLA_BINARY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) .len = TCP_FASTOPEN_COOKIE_MAX, },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) /* Add attributes, caller cancels its header on failure */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) static int tcp_metrics_fill_info(struct sk_buff *msg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) struct tcp_metrics_block *tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) struct nlattr *nest;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) switch (tm->tcpm_daddr.family) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) case AF_INET:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) if (nla_put_in_addr(msg, TCP_METRICS_ATTR_ADDR_IPV4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) inetpeer_get_addr_v4(&tm->tcpm_daddr)) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) if (nla_put_in_addr(msg, TCP_METRICS_ATTR_SADDR_IPV4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) inetpeer_get_addr_v4(&tm->tcpm_saddr)) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) case AF_INET6:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) if (nla_put_in6_addr(msg, TCP_METRICS_ATTR_ADDR_IPV6,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) inetpeer_get_addr_v6(&tm->tcpm_daddr)) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) if (nla_put_in6_addr(msg, TCP_METRICS_ATTR_SADDR_IPV6,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) inetpeer_get_addr_v6(&tm->tcpm_saddr)) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) return -EAFNOSUPPORT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) if (nla_put_msecs(msg, TCP_METRICS_ATTR_AGE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) jiffies - tm->tcpm_stamp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) TCP_METRICS_ATTR_PAD) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) int n = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) nest = nla_nest_start_noflag(msg, TCP_METRICS_ATTR_VALS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) if (!nest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) for (i = 0; i < TCP_METRIC_MAX_KERNEL + 1; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) u32 val = tm->tcpm_vals[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) if (!val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) if (i == TCP_METRIC_RTT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) if (nla_put_u32(msg, TCP_METRIC_RTT_US + 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) val) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) n++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) val = max(val / 1000, 1U);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) if (i == TCP_METRIC_RTTVAR) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) if (nla_put_u32(msg, TCP_METRIC_RTTVAR_US + 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) val) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) n++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) val = max(val / 1000, 1U);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) if (nla_put_u32(msg, i + 1, val) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) n++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) if (n)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) nla_nest_end(msg, nest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) nla_nest_cancel(msg, nest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) struct tcp_fastopen_metrics tfom_copy[1], *tfom;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) unsigned int seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) seq = read_seqbegin(&fastopen_seqlock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) tfom_copy[0] = tm->tcpm_fastopen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) } while (read_seqretry(&fastopen_seqlock, seq));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) tfom = tfom_copy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) if (tfom->mss &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) nla_put_u16(msg, TCP_METRICS_ATTR_FOPEN_MSS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) tfom->mss) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) if (tfom->syn_loss &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) (nla_put_u16(msg, TCP_METRICS_ATTR_FOPEN_SYN_DROPS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) tfom->syn_loss) < 0 ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) nla_put_msecs(msg, TCP_METRICS_ATTR_FOPEN_SYN_DROP_TS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) jiffies - tfom->last_syn_loss,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) TCP_METRICS_ATTR_PAD) < 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) if (tfom->cookie.len > 0 &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) nla_put(msg, TCP_METRICS_ATTR_FOPEN_COOKIE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) tfom->cookie.len, tfom->cookie.val) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) nla_put_failure:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) return -EMSGSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) static int tcp_metrics_dump_info(struct sk_buff *skb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) struct netlink_callback *cb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) struct tcp_metrics_block *tm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) void *hdr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) &tcp_metrics_nl_family, NLM_F_MULTI,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) TCP_METRICS_CMD_GET);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) if (!hdr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) return -EMSGSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) if (tcp_metrics_fill_info(skb, tm) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) genlmsg_end(skb, hdr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) nla_put_failure:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) genlmsg_cancel(skb, hdr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) return -EMSGSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) static int tcp_metrics_nl_dump(struct sk_buff *skb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) struct netlink_callback *cb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) struct net *net = sock_net(skb->sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) unsigned int max_rows = 1U << tcp_metrics_hash_log;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) unsigned int row, s_row = cb->args[0];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) int s_col = cb->args[1], col = s_col;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) for (row = s_row; row < max_rows; row++, s_col = 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) struct tcpm_hash_bucket *hb = tcp_metrics_hash + row;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) for (col = 0, tm = rcu_dereference(hb->chain); tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) tm = rcu_dereference(tm->tcpm_next), col++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) if (!net_eq(tm_net(tm), net))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) if (col < s_col)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) if (tcp_metrics_dump_info(skb, cb, tm) < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) done:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) cb->args[0] = row;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) cb->args[1] = col;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) return skb->len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) static int __parse_nl_addr(struct genl_info *info, struct inetpeer_addr *addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) unsigned int *hash, int optional, int v4, int v6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) struct nlattr *a;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) a = info->attrs[v4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) if (a) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) inetpeer_set_addr_v4(addr, nla_get_in_addr(a));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) if (hash)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) *hash = ipv4_addr_hash(inetpeer_get_addr_v4(addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) a = info->attrs[v6];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) if (a) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) struct in6_addr in6;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) if (nla_len(a) != sizeof(struct in6_addr))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) in6 = nla_get_in6_addr(a);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) inetpeer_set_addr_v6(addr, &in6);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) if (hash)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) *hash = ipv6_addr_hash(inetpeer_get_addr_v6(addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) return optional ? 1 : -EAFNOSUPPORT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) static int parse_nl_addr(struct genl_info *info, struct inetpeer_addr *addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) unsigned int *hash, int optional)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) return __parse_nl_addr(info, addr, hash, optional,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) TCP_METRICS_ATTR_ADDR_IPV4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) TCP_METRICS_ATTR_ADDR_IPV6);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) static int parse_nl_saddr(struct genl_info *info, struct inetpeer_addr *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) return __parse_nl_addr(info, addr, NULL, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) TCP_METRICS_ATTR_SADDR_IPV4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) TCP_METRICS_ATTR_SADDR_IPV6);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) static int tcp_metrics_nl_cmd_get(struct sk_buff *skb, struct genl_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) struct inetpeer_addr saddr, daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) unsigned int hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) struct sk_buff *msg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) struct net *net = genl_info_net(info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) void *reply;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) bool src = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) ret = parse_nl_addr(info, &daddr, &hash, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) ret = parse_nl_saddr(info, &saddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) src = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) if (!msg)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) reply = genlmsg_put_reply(msg, info, &tcp_metrics_nl_family, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) info->genlhdr->cmd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) if (!reply)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) goto nla_put_failure;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) hash ^= net_hash_mix(net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) hash = hash_32(hash, tcp_metrics_hash_log);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) ret = -ESRCH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) for (tm = rcu_dereference(tcp_metrics_hash[hash].chain); tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) tm = rcu_dereference(tm->tcpm_next)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) if (addr_same(&tm->tcpm_daddr, &daddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) (!src || addr_same(&tm->tcpm_saddr, &saddr)) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) net_eq(tm_net(tm), net)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) ret = tcp_metrics_fill_info(msg, tm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) genlmsg_end(msg, reply);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) return genlmsg_reply(msg, info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) nla_put_failure:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) ret = -EMSGSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) out_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) nlmsg_free(msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) static void tcp_metrics_flush_all(struct net *net)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) unsigned int max_rows = 1U << tcp_metrics_hash_log;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) struct tcpm_hash_bucket *hb = tcp_metrics_hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) unsigned int row;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) for (row = 0; row < max_rows; row++, hb++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) struct tcp_metrics_block __rcu **pp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) bool match;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) spin_lock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) pp = &hb->chain;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) for (tm = deref_locked(*pp); tm; tm = deref_locked(*pp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) match = net ? net_eq(tm_net(tm), net) :
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890) !refcount_read(&tm_net(tm)->count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) if (match) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) *pp = tm->tcpm_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) kfree_rcu(tm, rcu_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) pp = &tm->tcpm_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) spin_unlock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902) static int tcp_metrics_nl_cmd_del(struct sk_buff *skb, struct genl_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) struct tcpm_hash_bucket *hb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) struct tcp_metrics_block *tm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) struct tcp_metrics_block __rcu **pp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) struct inetpeer_addr saddr, daddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) unsigned int hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) struct net *net = genl_info_net(info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) bool src = true, found = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) ret = parse_nl_addr(info, &daddr, &hash, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) if (ret > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) tcp_metrics_flush_all(net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) ret = parse_nl_saddr(info, &saddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) src = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) hash ^= net_hash_mix(net);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) hash = hash_32(hash, tcp_metrics_hash_log);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) hb = tcp_metrics_hash + hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) pp = &hb->chain;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) spin_lock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) for (tm = deref_locked(*pp); tm; tm = deref_locked(*pp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) if (addr_same(&tm->tcpm_daddr, &daddr) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) (!src || addr_same(&tm->tcpm_saddr, &saddr)) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) net_eq(tm_net(tm), net)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) *pp = tm->tcpm_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) kfree_rcu(tm, rcu_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) found = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) pp = &tm->tcpm_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) spin_unlock_bh(&tcp_metrics_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) if (!found)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) return -ESRCH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) static const struct genl_small_ops tcp_metrics_nl_ops[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) .cmd = TCP_METRICS_CMD_GET,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) .validate = GENL_DONT_VALIDATE_STRICT | GENL_DONT_VALIDATE_DUMP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) .doit = tcp_metrics_nl_cmd_get,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) .dumpit = tcp_metrics_nl_dump,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) .cmd = TCP_METRICS_CMD_DEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) .validate = GENL_DONT_VALIDATE_STRICT | GENL_DONT_VALIDATE_DUMP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) .doit = tcp_metrics_nl_cmd_del,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) .flags = GENL_ADMIN_PERM,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) static struct genl_family tcp_metrics_nl_family __ro_after_init = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) .hdrsize = 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) .name = TCP_METRICS_GENL_NAME,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) .version = TCP_METRICS_GENL_VERSION,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) .maxattr = TCP_METRICS_ATTR_MAX,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) .policy = tcp_metrics_nl_policy,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) .netnsok = true,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) .module = THIS_MODULE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) .small_ops = tcp_metrics_nl_ops,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) .n_small_ops = ARRAY_SIZE(tcp_metrics_nl_ops),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) static unsigned int tcpmhash_entries;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) static int __init set_tcpmhash_entries(char *str)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) ssize_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) if (!str)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) ret = kstrtouint(str, 0, &tcpmhash_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) __setup("tcpmhash_entries=", set_tcpmhash_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) static int __net_init tcp_net_metrics_init(struct net *net)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) size_t size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) unsigned int slots;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) if (!net_eq(net, &init_net))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) slots = tcpmhash_entries;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) if (!slots) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) if (totalram_pages() >= 128 * 1024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) slots = 16 * 1024;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) slots = 8 * 1024;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) tcp_metrics_hash_log = order_base_2(slots);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) size = sizeof(struct tcpm_hash_bucket) << tcp_metrics_hash_log;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) tcp_metrics_hash = kvzalloc(size, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) if (!tcp_metrics_hash)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) static void __net_exit tcp_net_metrics_exit_batch(struct list_head *net_exit_list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) tcp_metrics_flush_all(NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) static __net_initdata struct pernet_operations tcp_net_metrics_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) .init = tcp_net_metrics_init,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) .exit_batch = tcp_net_metrics_exit_batch,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) void __init tcp_metrics_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) ret = register_pernet_subsys(&tcp_net_metrics_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) panic("Could not allocate the tcp_metrics hash table\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) ret = genl_register_family(&tcp_metrics_nl_family);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) panic("Could not register tcp_metrics generic netlink\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) }
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags