diff options
Diffstat (limited to 'net/dccp/ccids')
| -rw-r--r-- | net/dccp/ccids/Kconfig | 36 | ||||
| -rw-r--r-- | net/dccp/ccids/ccid2.c | 593 | ||||
| -rw-r--r-- | net/dccp/ccids/ccid2.h | 73 | ||||
| -rw-r--r-- | net/dccp/ccids/ccid3.c | 296 | ||||
| -rw-r--r-- | net/dccp/ccids/ccid3.h | 51 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/loss_interval.c | 3 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/loss_interval.h | 8 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/packet_history.c | 42 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/packet_history.h | 47 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/tfrc.c | 3 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/tfrc.h | 23 | ||||
| -rw-r--r-- | net/dccp/ccids/lib/tfrc_equation.c | 16 |
12 files changed, 562 insertions, 629 deletions
diff --git a/net/dccp/ccids/Kconfig b/net/dccp/ccids/Kconfig index 8408398cd44..8ba3fc9d6d1 100644 --- a/net/dccp/ccids/Kconfig +++ b/net/dccp/ccids/Kconfig @@ -1,5 +1,4 @@ -menu "DCCP CCIDs Configuration (EXPERIMENTAL)" - depends on EXPERIMENTAL +menu "DCCP CCIDs Configuration" config IP_DCCP_CCID2_DEBUG bool "CCID-2 debugging messages" @@ -12,7 +11,7 @@ config IP_DCCP_CCID2_DEBUG If in doubt, say N. config IP_DCCP_CCID3 - bool "CCID-3 (TCP-Friendly) (EXPERIMENTAL)" + bool "CCID-3 (TCP-Friendly)" def_bool y if (IP_DCCP = y || IP_DCCP = m) ---help--- CCID-3 denotes TCP-Friendly Rate Control (TFRC), an equation-based @@ -47,37 +46,6 @@ config IP_DCCP_CCID3_DEBUG If in doubt, say N. -config IP_DCCP_CCID3_RTO - int "Use higher bound for nofeedback timer" - default 100 - depends on IP_DCCP_CCID3 && EXPERIMENTAL - ---help--- - Use higher lower bound for nofeedback timer expiration. - - The TFRC nofeedback timer normally expires after the maximum of 4 - RTTs and twice the current send interval (RFC 3448, 4.3). On LANs - with a small RTT this can mean a high processing load and reduced - performance, since then the nofeedback timer is triggered very - frequently. - - This option enables to set a higher lower bound for the nofeedback - value. Values in units of milliseconds can be set here. - - A value of 0 disables this feature by enforcing the value specified - in RFC 3448. The following values have been suggested as bounds for - experimental use: - * 16-20ms to match the typical multimedia inter-frame interval - * 100ms as a reasonable compromise [default] - * 1000ms corresponds to the lower TCP RTO bound (RFC 2988, 2.4) - - The default of 100ms is a compromise between a large value for - efficient DCCP implementations, and a small value to avoid disrupting - the network in times of congestion. - - The purpose of the nofeedback timer is to slow DCCP down when there - is serious network congestion: experimenting with larger values should - therefore not be performed on WANs. - config IP_DCCP_TFRC_LIB def_bool y if IP_DCCP_CCID3 diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c index 9b3ae9922be..f053198e730 100644 --- a/net/dccp/ccids/ccid2.c +++ b/net/dccp/ccids/ccid2.c @@ -25,59 +25,14 @@ */ #include <linux/slab.h> #include "../feat.h" -#include "../ccid.h" -#include "../dccp.h" #include "ccid2.h" #ifdef CONFIG_IP_DCCP_CCID2_DEBUG -static int ccid2_debug; +static bool ccid2_debug; #define ccid2_pr_debug(format, a...) DCCP_PR_DEBUG(ccid2_debug, format, ##a) - -static void ccid2_hc_tx_check_sanity(const struct ccid2_hc_tx_sock *hc) -{ - int len = 0; - int pipe = 0; - struct ccid2_seq *seqp = hc->tx_seqh; - - /* there is data in the chain */ - if (seqp != hc->tx_seqt) { - seqp = seqp->ccid2s_prev; - len++; - if (!seqp->ccid2s_acked) - pipe++; - - while (seqp != hc->tx_seqt) { - struct ccid2_seq *prev = seqp->ccid2s_prev; - - len++; - if (!prev->ccid2s_acked) - pipe++; - - /* packets are sent sequentially */ - BUG_ON(dccp_delta_seqno(seqp->ccid2s_seq, - prev->ccid2s_seq ) >= 0); - BUG_ON(time_before(seqp->ccid2s_sent, - prev->ccid2s_sent)); - - seqp = prev; - } - } - - BUG_ON(pipe != hc->tx_pipe); - ccid2_pr_debug("len of chain=%d\n", len); - - do { - seqp = seqp->ccid2s_prev; - len++; - } while (seqp != hc->tx_seqh); - - ccid2_pr_debug("total len=%d\n", len); - BUG_ON(len != hc->tx_seqbufc * CCID2_SEQBUF_LEN); -} #else #define ccid2_pr_debug(format, a...) -#define ccid2_hc_tx_check_sanity(hc) #endif static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hc) @@ -123,17 +78,13 @@ static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hc) static int ccid2_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb) { - struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); - - if (hc->tx_pipe < hc->tx_cwnd) - return 0; - - return 1; /* XXX CCID should dequeue when ready instead of polling */ + if (ccid2_cwnd_network_limited(ccid2_hc_tx_sk(sk))) + return CCID_PACKET_WILL_DEQUEUE_LATER; + return CCID_PACKET_SEND_AT_ONCE; } static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val) { - struct dccp_sock *dp = dccp_sk(sk); u32 max_ratio = DIV_ROUND_UP(ccid2_hc_tx_sk(sk)->tx_cwnd, 2); /* @@ -146,29 +97,40 @@ static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val) DCCP_WARN("Limiting Ack Ratio (%u) to %u\n", val, max_ratio); val = max_ratio; } - if (val > DCCPF_ACK_RATIO_MAX) - val = DCCPF_ACK_RATIO_MAX; + dccp_feat_signal_nn_change(sk, DCCPF_ACK_RATIO, + min_t(u32, val, DCCPF_ACK_RATIO_MAX)); +} - if (val == dp->dccps_l_ack_ratio) - return; +static void ccid2_check_l_ack_ratio(struct sock *sk) +{ + struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); - ccid2_pr_debug("changing local ack ratio to %u\n", val); - dp->dccps_l_ack_ratio = val; + /* + * After a loss, idle period, application limited period, or RTO we + * need to check that the ack ratio is still less than the congestion + * window. Otherwise, we will send an entire congestion window of + * packets and got no response because we haven't sent ack ratio + * packets yet. + * If the ack ratio does need to be reduced, we reduce it to half of + * the congestion window (or 1 if that's zero) instead of to the + * congestion window. This prevents problems if one ack is lost. + */ + if (dccp_feat_nn_get(sk, DCCPF_ACK_RATIO) > hc->tx_cwnd) + ccid2_change_l_ack_ratio(sk, hc->tx_cwnd/2 ? : 1U); } -static void ccid2_change_srtt(struct ccid2_hc_tx_sock *hc, long val) +static void ccid2_change_l_seq_window(struct sock *sk, u64 val) { - ccid2_pr_debug("change SRTT to %ld\n", val); - hc->tx_srtt = val; + dccp_feat_signal_nn_change(sk, DCCPF_SEQUENCE_WINDOW, + clamp_val(val, DCCPF_SEQ_WMIN, + DCCPF_SEQ_WMAX)); } -static void ccid2_start_rto_timer(struct sock *sk); - static void ccid2_hc_tx_rto_expire(unsigned long data) { struct sock *sk = (struct sock *)data; struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); - long s; + const bool sender_was_blocked = ccid2_cwnd_network_limited(hc); bh_lock_sock(sk); if (sock_owned_by_user(sk)) { @@ -178,23 +140,17 @@ static void ccid2_hc_tx_rto_expire(unsigned long data) ccid2_pr_debug("RTO_EXPIRE\n"); - ccid2_hc_tx_check_sanity(hc); - /* back-off timer */ hc->tx_rto <<= 1; - - s = hc->tx_rto / HZ; - if (s > 60) - hc->tx_rto = 60 * HZ; - - ccid2_start_rto_timer(sk); + if (hc->tx_rto > DCCP_RTO_MAX) + hc->tx_rto = DCCP_RTO_MAX; /* adjust pipe, cwnd etc */ hc->tx_ssthresh = hc->tx_cwnd / 2; if (hc->tx_ssthresh < 2) hc->tx_ssthresh = 2; - hc->tx_cwnd = 1; - hc->tx_pipe = 0; + hc->tx_cwnd = 1; + hc->tx_pipe = 0; /* clear state about stuff we sent */ hc->tx_seqt = hc->tx_seqh; @@ -204,33 +160,108 @@ static void ccid2_hc_tx_rto_expire(unsigned long data) hc->tx_rpseq = 0; hc->tx_rpdupack = -1; ccid2_change_l_ack_ratio(sk, 1); - ccid2_hc_tx_check_sanity(hc); + + /* if we were blocked before, we may now send cwnd=1 packet */ + if (sender_was_blocked) + tasklet_schedule(&dccp_sk(sk)->dccps_xmitlet); + /* restart backed-off timer */ + sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto); out: bh_unlock_sock(sk); sock_put(sk); } -static void ccid2_start_rto_timer(struct sock *sk) +/* + * Congestion window validation (RFC 2861). + */ +static bool ccid2_do_cwv = true; +module_param(ccid2_do_cwv, bool, 0644); +MODULE_PARM_DESC(ccid2_do_cwv, "Perform RFC2861 Congestion Window Validation"); + +/** + * ccid2_update_used_window - Track how much of cwnd is actually used + * This is done in addition to CWV. The sender needs to have an idea of how many + * packets may be in flight, to set the local Sequence Window value accordingly + * (RFC 4340, 7.5.2). The CWV mechanism is exploited to keep track of the + * maximum-used window. We use an EWMA low-pass filter to filter out noise. + */ +static void ccid2_update_used_window(struct ccid2_hc_tx_sock *hc, u32 new_wnd) +{ + hc->tx_expected_wnd = (3 * hc->tx_expected_wnd + new_wnd) / 4; +} + +/* This borrows the code of tcp_cwnd_application_limited() */ +static void ccid2_cwnd_application_limited(struct sock *sk, const u32 now) { struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + /* don't reduce cwnd below the initial window (IW) */ + u32 init_win = rfc3390_bytes_to_packets(dccp_sk(sk)->dccps_mss_cache), + win_used = max(hc->tx_cwnd_used, init_win); + + if (win_used < hc->tx_cwnd) { + hc->tx_ssthresh = max(hc->tx_ssthresh, + (hc->tx_cwnd >> 1) + (hc->tx_cwnd >> 2)); + hc->tx_cwnd = (hc->tx_cwnd + win_used) >> 1; + } + hc->tx_cwnd_used = 0; + hc->tx_cwnd_stamp = now; - ccid2_pr_debug("setting RTO timeout=%ld\n", hc->tx_rto); + ccid2_check_l_ack_ratio(sk); +} - BUG_ON(timer_pending(&hc->tx_rtotimer)); - sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto); +/* This borrows the code of tcp_cwnd_restart() */ +static void ccid2_cwnd_restart(struct sock *sk, const u32 now) +{ + struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + u32 cwnd = hc->tx_cwnd, restart_cwnd, + iwnd = rfc3390_bytes_to_packets(dccp_sk(sk)->dccps_mss_cache); + + hc->tx_ssthresh = max(hc->tx_ssthresh, (cwnd >> 1) + (cwnd >> 2)); + + /* don't reduce cwnd below the initial window (IW) */ + restart_cwnd = min(cwnd, iwnd); + cwnd >>= (now - hc->tx_lsndtime) / hc->tx_rto; + hc->tx_cwnd = max(cwnd, restart_cwnd); + + hc->tx_cwnd_stamp = now; + hc->tx_cwnd_used = 0; + + ccid2_check_l_ack_ratio(sk); } -static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len) +static void ccid2_hc_tx_packet_sent(struct sock *sk, unsigned int len) { struct dccp_sock *dp = dccp_sk(sk); struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + const u32 now = ccid2_time_stamp; struct ccid2_seq *next; - hc->tx_pipe++; + /* slow-start after idle periods (RFC 2581, RFC 2861) */ + if (ccid2_do_cwv && !hc->tx_pipe && + (s32)(now - hc->tx_lsndtime) >= hc->tx_rto) + ccid2_cwnd_restart(sk, now); + + hc->tx_lsndtime = now; + hc->tx_pipe += 1; + + /* see whether cwnd was fully used (RFC 2861), update expected window */ + if (ccid2_cwnd_network_limited(hc)) { + ccid2_update_used_window(hc, hc->tx_cwnd); + hc->tx_cwnd_used = 0; + hc->tx_cwnd_stamp = now; + } else { + if (hc->tx_pipe > hc->tx_cwnd_used) + hc->tx_cwnd_used = hc->tx_pipe; + + ccid2_update_used_window(hc, hc->tx_cwnd_used); + + if (ccid2_do_cwv && (s32)(now - hc->tx_cwnd_stamp) >= hc->tx_rto) + ccid2_cwnd_application_limited(sk, now); + } hc->tx_seqh->ccid2s_seq = dp->dccps_gss; hc->tx_seqh->ccid2s_acked = 0; - hc->tx_seqh->ccid2s_sent = jiffies; + hc->tx_seqh->ccid2s_sent = now; next = hc->tx_seqh->ccid2s_next; /* check if we need to alloc more space */ @@ -296,221 +327,190 @@ static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len) } #endif - /* setup RTO timer */ - if (!timer_pending(&hc->tx_rtotimer)) - ccid2_start_rto_timer(sk); + sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto); #ifdef CONFIG_IP_DCCP_CCID2_DEBUG do { struct ccid2_seq *seqp = hc->tx_seqt; while (seqp != hc->tx_seqh) { - ccid2_pr_debug("out seq=%llu acked=%d time=%lu\n", + ccid2_pr_debug("out seq=%llu acked=%d time=%u\n", (unsigned long long)seqp->ccid2s_seq, seqp->ccid2s_acked, seqp->ccid2s_sent); seqp = seqp->ccid2s_next; } } while (0); ccid2_pr_debug("=========\n"); - ccid2_hc_tx_check_sanity(hc); #endif } -/* XXX Lame code duplication! - * returns -1 if none was found. - * else returns the next offset to use in the function call. +/** + * ccid2_rtt_estimator - Sample RTT and compute RTO using RFC2988 algorithm + * This code is almost identical with TCP's tcp_rtt_estimator(), since + * - it has a higher sampling frequency (recommended by RFC 1323), + * - the RTO does not collapse into RTT due to RTTVAR going towards zero, + * - it is simple (cf. more complex proposals such as Eifel timer or research + * which suggests that the gain should be set according to window size), + * - in tests it was found to work well with CCID2 [gerrit]. */ -static int ccid2_ackvector(struct sock *sk, struct sk_buff *skb, int offset, - unsigned char **vec, unsigned char *veclen) +static void ccid2_rtt_estimator(struct sock *sk, const long mrtt) { - const struct dccp_hdr *dh = dccp_hdr(skb); - unsigned char *options = (unsigned char *)dh + dccp_hdr_len(skb); - unsigned char *opt_ptr; - const unsigned char *opt_end = (unsigned char *)dh + - (dh->dccph_doff * 4); - unsigned char opt, len; - unsigned char *value; - - BUG_ON(offset < 0); - options += offset; - opt_ptr = options; - if (opt_ptr >= opt_end) - return -1; - - while (opt_ptr != opt_end) { - opt = *opt_ptr++; - len = 0; - value = NULL; - - /* Check if this isn't a single byte option */ - if (opt > DCCPO_MAX_RESERVED) { - if (opt_ptr == opt_end) - goto out_invalid_option; - - len = *opt_ptr++; - if (len < 3) - goto out_invalid_option; + struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + long m = mrtt ? : 1; + + if (hc->tx_srtt == 0) { + /* First measurement m */ + hc->tx_srtt = m << 3; + hc->tx_mdev = m << 1; + + hc->tx_mdev_max = max(hc->tx_mdev, tcp_rto_min(sk)); + hc->tx_rttvar = hc->tx_mdev_max; + + hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss; + } else { + /* Update scaled SRTT as SRTT += 1/8 * (m - SRTT) */ + m -= (hc->tx_srtt >> 3); + hc->tx_srtt += m; + + /* Similarly, update scaled mdev with regard to |m| */ + if (m < 0) { + m = -m; + m -= (hc->tx_mdev >> 2); /* - * Remove the type and len fields, leaving - * just the value size + * This neutralises RTO increase when RTT < SRTT - mdev + * (see P. Sarolahti, A. Kuznetsov,"Congestion Control + * in Linux TCP", USENIX 2002, pp. 49-62). */ - len -= 2; - value = opt_ptr; - opt_ptr += len; + if (m > 0) + m >>= 3; + } else { + m -= (hc->tx_mdev >> 2); + } + hc->tx_mdev += m; - if (opt_ptr > opt_end) - goto out_invalid_option; + if (hc->tx_mdev > hc->tx_mdev_max) { + hc->tx_mdev_max = hc->tx_mdev; + if (hc->tx_mdev_max > hc->tx_rttvar) + hc->tx_rttvar = hc->tx_mdev_max; } - switch (opt) { - case DCCPO_ACK_VECTOR_0: - case DCCPO_ACK_VECTOR_1: - *vec = value; - *veclen = len; - return offset + (opt_ptr - options); + /* + * Decay RTTVAR at most once per flight, exploiting that + * 1) pipe <= cwnd <= Sequence_Window = W (RFC 4340, 7.5.2) + * 2) AWL = GSS-W+1 <= GAR <= GSS (RFC 4340, 7.5.1) + * GAR is a useful bound for FlightSize = pipe. + * AWL is probably too low here, as it over-estimates pipe. + */ + if (after48(dccp_sk(sk)->dccps_gar, hc->tx_rtt_seq)) { + if (hc->tx_mdev_max < hc->tx_rttvar) + hc->tx_rttvar -= (hc->tx_rttvar - + hc->tx_mdev_max) >> 2; + hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss; + hc->tx_mdev_max = tcp_rto_min(sk); } } - return -1; - -out_invalid_option: - DCCP_BUG("Invalid option - this should not happen (previous parsing)!"); - return -1; -} - -static void ccid2_hc_tx_kill_rto_timer(struct sock *sk) -{ - struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + /* + * Set RTO from SRTT and RTTVAR + * As in TCP, 4 * RTTVAR >= TCP_RTO_MIN, giving a minimum RTO of 200 ms. + * This agrees with RFC 4341, 5: + * "Because DCCP does not retransmit data, DCCP does not require + * TCP's recommended minimum timeout of one second". + */ + hc->tx_rto = (hc->tx_srtt >> 3) + hc->tx_rttvar; - sk_stop_timer(sk, &hc->tx_rtotimer); - ccid2_pr_debug("deleted RTO timer\n"); + if (hc->tx_rto > DCCP_RTO_MAX) + hc->tx_rto = DCCP_RTO_MAX; } -static inline void ccid2_new_ack(struct sock *sk, - struct ccid2_seq *seqp, - unsigned int *maxincr) +static void ccid2_new_ack(struct sock *sk, struct ccid2_seq *seqp, + unsigned int *maxincr) { struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); - - if (hc->tx_cwnd < hc->tx_ssthresh) { - if (*maxincr > 0 && ++hc->tx_packets_acked == 2) { - hc->tx_cwnd += 1; - *maxincr -= 1; - hc->tx_packets_acked = 0; - } - } else if (++hc->tx_packets_acked >= hc->tx_cwnd) { + struct dccp_sock *dp = dccp_sk(sk); + int r_seq_used = hc->tx_cwnd / dp->dccps_l_ack_ratio; + + if (hc->tx_cwnd < dp->dccps_l_seq_win && + r_seq_used < dp->dccps_r_seq_win) { + if (hc->tx_cwnd < hc->tx_ssthresh) { + if (*maxincr > 0 && ++hc->tx_packets_acked >= 2) { + hc->tx_cwnd += 1; + *maxincr -= 1; + hc->tx_packets_acked = 0; + } + } else if (++hc->tx_packets_acked >= hc->tx_cwnd) { hc->tx_cwnd += 1; hc->tx_packets_acked = 0; - } - - /* update RTO */ - if (hc->tx_srtt == -1 || - time_after(jiffies, hc->tx_lastrtt + hc->tx_srtt)) { - unsigned long r = (long)jiffies - (long)seqp->ccid2s_sent; - int s; - - /* first measurement */ - if (hc->tx_srtt == -1) { - ccid2_pr_debug("R: %lu Time=%lu seq=%llu\n", - r, jiffies, - (unsigned long long)seqp->ccid2s_seq); - ccid2_change_srtt(hc, r); - hc->tx_rttvar = r >> 1; - } else { - /* RTTVAR */ - long tmp = hc->tx_srtt - r; - long srtt; - - if (tmp < 0) - tmp *= -1; - - tmp >>= 2; - hc->tx_rttvar *= 3; - hc->tx_rttvar >>= 2; - hc->tx_rttvar += tmp; - - /* SRTT */ - srtt = hc->tx_srtt; - srtt *= 7; - srtt >>= 3; - tmp = r >> 3; - srtt += tmp; - ccid2_change_srtt(hc, srtt); } - s = hc->tx_rttvar << 2; - /* clock granularity is 1 when based on jiffies */ - if (!s) - s = 1; - hc->tx_rto = hc->tx_srtt + s; - - /* must be at least a second */ - s = hc->tx_rto / HZ; - /* DCCP doesn't require this [but I like it cuz my code sux] */ -#if 1 - if (s < 1) - hc->tx_rto = HZ; -#endif - /* max 60 seconds */ - if (s > 60) - hc->tx_rto = HZ * 60; - - hc->tx_lastrtt = jiffies; - - ccid2_pr_debug("srtt: %ld rttvar: %ld rto: %ld (HZ=%d) R=%lu\n", - hc->tx_srtt, hc->tx_rttvar, - hc->tx_rto, HZ, r); } - /* we got a new ack, so re-start RTO timer */ - ccid2_hc_tx_kill_rto_timer(sk); - ccid2_start_rto_timer(sk); -} - -static void ccid2_hc_tx_dec_pipe(struct sock *sk) -{ - struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + /* + * Adjust the local sequence window and the ack ratio to allow about + * 5 times the number of packets in the network (RFC 4340 7.5.2) + */ + if (r_seq_used * CCID2_WIN_CHANGE_FACTOR >= dp->dccps_r_seq_win) + ccid2_change_l_ack_ratio(sk, dp->dccps_l_ack_ratio * 2); + else if (r_seq_used * CCID2_WIN_CHANGE_FACTOR < dp->dccps_r_seq_win/2) + ccid2_change_l_ack_ratio(sk, dp->dccps_l_ack_ratio / 2 ? : 1U); - if (hc->tx_pipe == 0) - DCCP_BUG("pipe == 0"); - else - hc->tx_pipe--; + if (hc->tx_cwnd * CCID2_WIN_CHANGE_FACTOR >= dp->dccps_l_seq_win) + ccid2_change_l_seq_window(sk, dp->dccps_l_seq_win * 2); + else if (hc->tx_cwnd * CCID2_WIN_CHANGE_FACTOR < dp->dccps_l_seq_win/2) + ccid2_change_l_seq_window(sk, dp->dccps_l_seq_win / 2); - if (hc->tx_pipe == 0) - ccid2_hc_tx_kill_rto_timer(sk); + /* + * FIXME: RTT is sampled several times per acknowledgment (for each + * entry in the Ack Vector), instead of once per Ack (as in TCP SACK). + * This causes the RTT to be over-estimated, since the older entries + * in the Ack Vector have earlier sending times. + * The cleanest solution is to not use the ccid2s_sent field at all + * and instead use DCCP timestamps: requires changes in other places. + */ + ccid2_rtt_estimator(sk, ccid2_time_stamp - seqp->ccid2s_sent); } static void ccid2_congestion_event(struct sock *sk, struct ccid2_seq *seqp) { struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); - if (time_before(seqp->ccid2s_sent, hc->tx_last_cong)) { + if ((s32)(seqp->ccid2s_sent - hc->tx_last_cong) < 0) { ccid2_pr_debug("Multiple losses in an RTT---treating as one\n"); return; } - hc->tx_last_cong = jiffies; + hc->tx_last_cong = ccid2_time_stamp; hc->tx_cwnd = hc->tx_cwnd / 2 ? : 1U; hc->tx_ssthresh = max(hc->tx_cwnd, 2U); - /* Avoid spurious timeouts resulting from Ack Ratio > cwnd */ - if (dccp_sk(sk)->dccps_l_ack_ratio > hc->tx_cwnd) - ccid2_change_l_ack_ratio(sk, hc->tx_cwnd); + ccid2_check_l_ack_ratio(sk); +} + +static int ccid2_hc_tx_parse_options(struct sock *sk, u8 packet_type, + u8 option, u8 *optval, u8 optlen) +{ + struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + + switch (option) { + case DCCPO_ACK_VECTOR_0: + case DCCPO_ACK_VECTOR_1: + return dccp_ackvec_parsed_add(&hc->tx_av_chunks, optval, optlen, + option - DCCPO_ACK_VECTOR_0); + } + return 0; } static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) { struct dccp_sock *dp = dccp_sk(sk); struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); + const bool sender_was_blocked = ccid2_cwnd_network_limited(hc); + struct dccp_ackvec_parsed *avp; u64 ackno, seqno; struct ccid2_seq *seqp; - unsigned char *vector; - unsigned char veclen; - int offset = 0; int done = 0; unsigned int maxincr = 0; - ccid2_hc_tx_check_sanity(hc); /* check reverse path congestion */ seqno = DCCP_SKB_CB(skb)->dccpd_seq; @@ -534,24 +534,27 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) if (hc->tx_rpdupack >= NUMDUPACK) { hc->tx_rpdupack = -1; /* XXX lame */ hc->tx_rpseq = 0; - +#ifdef __CCID2_COPES_GRACEFULLY_WITH_ACK_CONGESTION_CONTROL__ + /* + * FIXME: Ack Congestion Control is broken; in + * the current state instabilities occurred with + * Ack Ratios greater than 1; causing hang-ups + * and long RTO timeouts. This needs to be fixed + * before opening up dynamic changes. -- gerrit + */ ccid2_change_l_ack_ratio(sk, 2 * dp->dccps_l_ack_ratio); +#endif } } } /* check forward path congestion */ - /* still didn't send out new data packets */ - if (hc->tx_seqh == hc->tx_seqt) + if (dccp_packet_without_ack(skb)) return; - switch (DCCP_SKB_CB(skb)->dccpd_type) { - case DCCP_PKT_ACK: - case DCCP_PKT_DATAACK: - break; - default: - return; - } + /* still didn't send out new data packets */ + if (hc->tx_seqh == hc->tx_seqt) + goto done; ackno = DCCP_SKB_CB(skb)->dccpd_ack_seq; if (after48(ackno, hc->tx_high_ack)) @@ -575,16 +578,16 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) maxincr = DIV_ROUND_UP(dp->dccps_l_ack_ratio, 2); /* go through all ack vectors */ - while ((offset = ccid2_ackvector(sk, skb, offset, - &vector, &veclen)) != -1) { + list_for_each_entry(avp, &hc->tx_av_chunks, node) { /* go through this ack vector */ - while (veclen--) { - const u8 rl = *vector & DCCP_ACKVEC_LEN_MASK; - u64 ackno_end_rl = SUB48(ackno, rl); + for (; avp->len--; avp->vec++) { + u64 ackno_end_rl = SUB48(ackno, + dccp_ackvec_runlen(avp->vec)); - ccid2_pr_debug("ackvec start:%llu end:%llu\n", + ccid2_pr_debug("ackvec %llu |%u,%u|\n", (unsigned long long)ackno, - (unsigned long long)ackno_end_rl); + dccp_ackvec_state(avp->vec) >> 6, + dccp_ackvec_runlen(avp->vec)); /* if the seqno we are analyzing is larger than the * current ackno, then move towards the tail of our * seqnos. @@ -603,24 +606,22 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) * run length */ while (between48(seqp->ccid2s_seq,ackno_end_rl,ackno)) { - const u8 state = *vector & - DCCP_ACKVEC_STATE_MASK; + const u8 state = dccp_ackvec_state(avp->vec); /* new packet received or marked */ - if (state != DCCP_ACKVEC_STATE_NOT_RECEIVED && + if (state != DCCPAV_NOT_RECEIVED && !seqp->ccid2s_acked) { - if (state == - DCCP_ACKVEC_STATE_ECN_MARKED) { + if (state == DCCPAV_ECN_MARKED) ccid2_congestion_event(sk, seqp); - } else + else ccid2_new_ack(sk, seqp, &maxincr); seqp->ccid2s_acked = 1; ccid2_pr_debug("Got ack for %llu\n", (unsigned long long)seqp->ccid2s_seq); - ccid2_hc_tx_dec_pipe(sk); + hc->tx_pipe--; } if (seqp == hc->tx_seqt) { done = 1; @@ -632,7 +633,6 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) break; ackno = SUB48(ackno_end_rl, 1); - vector++; } if (done) break; @@ -677,7 +677,7 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) * one ack vector. */ ccid2_congestion_event(sk, seqp); - ccid2_hc_tx_dec_pipe(sk); + hc->tx_pipe--; } if (seqp == hc->tx_seqt) break; @@ -695,7 +695,16 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) hc->tx_seqt = hc->tx_seqt->ccid2s_next; } - ccid2_hc_tx_check_sanity(hc); + /* restart RTO timer if not all outstanding data has been acked */ + if (hc->tx_pipe == 0) + sk_stop_timer(sk, &hc->tx_rtotimer); + else + sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto); +done: + /* check if incoming Acks allow pending packets to be sent */ + if (sender_was_blocked && !ccid2_cwnd_network_limited(hc)) + tasklet_schedule(&dccp_sk(sk)->dccps_xmitlet); + dccp_ackvec_parsed_cleanup(&hc->tx_av_chunks); } static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk) @@ -707,12 +716,9 @@ static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk) /* RFC 4341, 5: initialise ssthresh to arbitrarily high (max) value */ hc->tx_ssthresh = ~0U; - /* - * RFC 4341, 5: "The cwnd parameter is initialized to at most four - * packets for new connections, following the rules from [RFC3390]". - * We need to convert the bytes of RFC3390 into the packets of RFC 4341. - */ - hc->tx_cwnd = clamp(4380U / dp->dccps_mss_cache, 2U, 4U); + /* Use larger initial windows (RFC 4341, section 5). */ + hc->tx_cwnd = rfc3390_bytes_to_packets(dp->dccps_mss_cache); + hc->tx_expected_wnd = hc->tx_cwnd; /* Make sure that Ack Ratio is enabled and within bounds. */ max_ratio = DIV_ROUND_UP(hc->tx_cwnd, 2); @@ -723,15 +729,13 @@ static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk) if (ccid2_hc_tx_alloc_seq(hc)) return -ENOMEM; - hc->tx_rto = 3 * HZ; - ccid2_change_srtt(hc, -1); - hc->tx_rttvar = -1; + hc->tx_rto = DCCP_TIMEOUT_INIT; hc->tx_rpdupack = -1; - hc->tx_last_cong = jiffies; + hc->tx_last_cong = hc->tx_lsndtime = hc->tx_cwnd_stamp = ccid2_time_stamp; + hc->tx_cwnd_used = 0; setup_timer(&hc->tx_rtotimer, ccid2_hc_tx_rto_expire, (unsigned long)sk); - - ccid2_hc_tx_check_sanity(hc); + INIT_LIST_HEAD(&hc->tx_av_chunks); return 0; } @@ -740,7 +744,7 @@ static void ccid2_hc_tx_exit(struct sock *sk) struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk); int i; - ccid2_hc_tx_kill_rto_timer(sk); + sk_stop_timer(sk, &hc->tx_rtotimer); for (i = 0; i < hc->tx_seqbufc; i++) kfree(hc->tx_seqbuf[i]); @@ -749,32 +753,29 @@ static void ccid2_hc_tx_exit(struct sock *sk) static void ccid2_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) { - const struct dccp_sock *dp = dccp_sk(sk); struct ccid2_hc_rx_sock *hc = ccid2_hc_rx_sk(sk); - switch (DCCP_SKB_CB(skb)->dccpd_type) { - case DCCP_PKT_DATA: - case DCCP_PKT_DATAACK: - hc->rx_data++; - if (hc->rx_data >= dp->dccps_r_ack_ratio) { - dccp_send_ack(sk); - hc->rx_data = 0; - } - break; + if (!dccp_data_packet(skb)) + return; + + if (++hc->rx_num_data_pkts >= dccp_sk(sk)->dccps_r_ack_ratio) { + dccp_send_ack(sk); + hc->rx_num_data_pkts = 0; } } struct ccid_operations ccid2_ops = { - .ccid_id = DCCPC_CCID2, - .ccid_name = "TCP-like", - .ccid_hc_tx_obj_size = sizeof(struct ccid2_hc_tx_sock), - .ccid_hc_tx_init = ccid2_hc_tx_init, - .ccid_hc_tx_exit = ccid2_hc_tx_exit, - .ccid_hc_tx_send_packet = ccid2_hc_tx_send_packet, - .ccid_hc_tx_packet_sent = ccid2_hc_tx_packet_sent, - .ccid_hc_tx_packet_recv = ccid2_hc_tx_packet_recv, - .ccid_hc_rx_obj_size = sizeof(struct ccid2_hc_rx_sock), - .ccid_hc_rx_packet_recv = ccid2_hc_rx_packet_recv, + .ccid_id = DCCPC_CCID2, + .ccid_name = "TCP-like", + .ccid_hc_tx_obj_size = sizeof(struct ccid2_hc_tx_sock), + .ccid_hc_tx_init = ccid2_hc_tx_init, + .ccid_hc_tx_exit = ccid2_hc_tx_exit, + .ccid_hc_tx_send_packet = ccid2_hc_tx_send_packet, + .ccid_hc_tx_packet_sent = ccid2_hc_tx_packet_sent, + .ccid_hc_tx_parse_options = ccid2_hc_tx_parse_options, + .ccid_hc_tx_packet_recv = ccid2_hc_tx_packet_recv, + .ccid_hc_rx_obj_size = sizeof(struct ccid2_hc_rx_sock), + .ccid_hc_rx_packet_recv = ccid2_hc_rx_packet_recv, }; #ifdef CONFIG_IP_DCCP_CCID2_DEBUG diff --git a/net/dccp/ccids/ccid2.h b/net/dccp/ccids/ccid2.h index 1ec6a30103b..18c97543e52 100644 --- a/net/dccp/ccids/ccid2.h +++ b/net/dccp/ccids/ccid2.h @@ -18,18 +18,23 @@ #ifndef _DCCP_CCID2_H_ #define _DCCP_CCID2_H_ -#include <linux/dccp.h> #include <linux/timer.h> #include <linux/types.h> #include "../ccid.h" +#include "../dccp.h" + +/* + * CCID-2 timestamping faces the same issues as TCP timestamping. + * Hence we reuse/share as much of the code as possible. + */ +#define ccid2_time_stamp tcp_time_stamp + /* NUMDUPACK parameter from RFC 4341, p. 6 */ #define NUMDUPACK 3 -struct sock; - struct ccid2_seq { u64 ccid2s_seq; - unsigned long ccid2s_sent; + u32 ccid2s_sent; int ccid2s_acked; struct ccid2_seq *ccid2s_prev; struct ccid2_seq *ccid2s_next; @@ -38,13 +43,29 @@ struct ccid2_seq { #define CCID2_SEQBUF_LEN 1024 #define CCID2_SEQBUF_MAX 128 +/* + * Multiple of congestion window to keep the sequence window at + * (RFC 4340 7.5.2) + */ +#define CCID2_WIN_CHANGE_FACTOR 5 + /** * struct ccid2_hc_tx_sock - CCID2 TX half connection * @tx_{cwnd,ssthresh,pipe}: as per RFC 4341, section 5 * @tx_packets_acked: Ack counter for deriving cwnd growth (RFC 3465) - * @tx_lastrtt: time RTT was last measured + * @tx_srtt: smoothed RTT estimate, scaled by 2^3 + * @tx_mdev: smoothed RTT variation, scaled by 2^2 + * @tx_mdev_max: maximum of @mdev during one flight + * @tx_rttvar: moving average/maximum of @mdev_max + * @tx_rto: RTO value deriving from SRTT and RTTVAR (RFC 2988) + * @tx_rtt_seq: to decay RTTVAR at most once per flight + * @tx_cwnd_used: actually used cwnd, W_used of RFC 2861 + * @tx_expected_wnd: moving average of @tx_cwnd_used + * @tx_cwnd_stamp: to track idle periods in CWV + * @tx_lsndtime: last time (in jiffies) a data packet was sent * @tx_rpseq: last consecutive seqno * @tx_rpdupack: dupacks since rpseq + * @tx_av_chunks: list of Ack Vectors received on current skb */ struct ccid2_hc_tx_sock { u32 tx_cwnd; @@ -55,19 +76,49 @@ struct ccid2_hc_tx_sock { int tx_seqbufc; struct ccid2_seq *tx_seqh; struct ccid2_seq *tx_seqt; - long tx_rto; - long tx_srtt; - long tx_rttvar; - unsigned long tx_lastrtt; + + /* RTT measurement: variables/principles are the same as in TCP */ + u32 tx_srtt, + tx_mdev, + tx_mdev_max, + tx_rttvar, + tx_rto; + u64 tx_rtt_seq:48; struct timer_list tx_rtotimer; + + /* Congestion Window validation (optional, RFC 2861) */ + u32 tx_cwnd_used, + tx_expected_wnd, + tx_cwnd_stamp, + tx_lsndtime; + u64 tx_rpseq; int tx_rpdupack; - unsigned long tx_last_cong; + u32 tx_last_cong; u64 tx_high_ack; + struct list_head tx_av_chunks; }; +static inline bool ccid2_cwnd_network_limited(struct ccid2_hc_tx_sock *hc) +{ + return hc->tx_pipe >= hc->tx_cwnd; +} + +/* + * Convert RFC 3390 larger initial window into an equivalent number of packets. + * This is based on the numbers specified in RFC 5681, 3.1. + */ +static inline u32 rfc3390_bytes_to_packets(const u32 smss) +{ + return smss <= 1095 ? 4 : (smss > 2190 ? 2 : 3); +} + +/** + * struct ccid2_hc_rx_sock - Receiving end of CCID-2 half-connection + * @rx_num_data_pkts: number of data packets received since last feedback + */ struct ccid2_hc_rx_sock { - int rx_data; + u32 rx_num_data_pkts; }; static inline struct ccid2_hc_tx_sock *ccid2_hc_tx_sk(const struct sock *sk) diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index d3235899c7e..119c04317d4 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -38,7 +38,7 @@ #include <asm/unaligned.h> #ifdef CONFIG_IP_DCCP_CCID3_DEBUG -static int ccid3_debug; +static bool ccid3_debug; #define ccid3_pr_debug(format, a...) DCCP_PR_DEBUG(ccid3_debug, format, ##a) #else #define ccid3_pr_debug(format, a...) @@ -54,7 +54,6 @@ static const char *ccid3_tx_state_name(enum ccid3_hc_tx_states state) [TFRC_SSTATE_NO_SENT] = "NO_SENT", [TFRC_SSTATE_NO_FBACK] = "NO_FBACK", [TFRC_SSTATE_FBACK] = "FBACK", - [TFRC_SSTATE_TERM] = "TERM", }; return ccid3_state_names[state]; @@ -91,19 +90,17 @@ static inline u64 rfc3390_initial_rate(struct sock *sk) return scaled_div(w_init << 6, hc->tx_rtt); } -/* - * Recalculate t_ipi and delta (should be called whenever X changes) +/** + * ccid3_update_send_interval - Calculate new t_ipi = s / X_inst + * This respects the granularity of X_inst (64 * bytes/second). */ static void ccid3_update_send_interval(struct ccid3_hc_tx_sock *hc) { - /* Calculate new t_ipi = s / X_inst (X_inst is in 64 * bytes/second) */ hc->tx_t_ipi = scaled_div32(((u64)hc->tx_s) << 6, hc->tx_x); - /* Calculate new delta by delta = min(t_ipi / 2, t_gran / 2) */ - hc->tx_delta = min_t(u32, hc->tx_t_ipi / 2, TFRC_OPSYS_HALF_TIME_GRAN); - - ccid3_pr_debug("t_ipi=%u, delta=%u, s=%u, X=%u\n", hc->tx_t_ipi, - hc->tx_delta, hc->tx_s, (unsigned)(hc->tx_x >> 6)); + DCCP_BUG_ON(hc->tx_t_ipi == 0); + ccid3_pr_debug("t_ipi=%u, s=%u, X=%u\n", hc->tx_t_ipi, + hc->tx_s, (unsigned int)(hc->tx_x >> 6)); } static u32 ccid3_hc_tx_idle_rtt(struct ccid3_hc_tx_sock *hc, ktime_t now) @@ -116,6 +113,7 @@ static u32 ccid3_hc_tx_idle_rtt(struct ccid3_hc_tx_sock *hc, ktime_t now) /** * ccid3_hc_tx_update_x - Update allowed sending rate X * @stamp: most recent time if available - can be left NULL. + * * This function tracks draft rfc3448bis, check there for latest details. * * Note: X and X_recv are both stored in units of 64 * bytes/second, to support @@ -156,17 +154,19 @@ static void ccid3_hc_tx_update_x(struct sock *sk, ktime_t *stamp) if (hc->tx_x != old_x) { ccid3_pr_debug("X_prev=%u, X_now=%u, X_calc=%u, " - "X_recv=%u\n", (unsigned)(old_x >> 6), - (unsigned)(hc->tx_x >> 6), hc->tx_x_calc, - (unsigned)(hc->tx_x_recv >> 6)); + "X_recv=%u\n", (unsigned int)(old_x >> 6), + (unsigned int)(hc->tx_x >> 6), hc->tx_x_calc, + (unsigned int)(hc->tx_x_recv >> 6)); ccid3_update_send_interval(hc); } } -/* - * Track the mean packet size `s' (cf. RFC 4342, 5.3 and RFC 3448, 4.1) +/** + * ccid3_hc_tx_update_s - Track the mean packet size `s' * @len: DCCP packet payload size in bytes + * + * cf. RFC 4342, 5.3 and RFC 3448, 4.1 */ static inline void ccid3_hc_tx_update_s(struct ccid3_hc_tx_sock *hc, int len) { @@ -211,16 +211,19 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data) ccid3_pr_debug("%s(%p, state=%s) - entry\n", dccp_role(sk), sk, ccid3_tx_state_name(hc->tx_state)); + /* Ignore and do not restart after leaving the established state */ + if ((1 << sk->sk_state) & ~(DCCPF_OPEN | DCCPF_PARTOPEN)) + goto out; + + /* Reset feedback state to "no feedback received" */ if (hc->tx_state == TFRC_SSTATE_FBACK) ccid3_hc_tx_set_state(sk, TFRC_SSTATE_NO_FBACK); - else if (hc->tx_state != TFRC_SSTATE_NO_FBACK) - goto out; /* * Determine new allowed sending rate X as per draft rfc3448bis-00, 4.4 + * RTO is 0 if and only if no feedback has been received yet. */ - if (hc->tx_t_rto == 0 || /* no feedback received yet */ - hc->tx_p == 0) { + if (hc->tx_t_rto == 0 || hc->tx_p == 0) { /* halve send rate directly */ hc->tx_x = max(hc->tx_x / 2, @@ -237,8 +240,6 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data) * * Note that X_recv is scaled by 2^6 while X_calc is not */ - BUG_ON(hc->tx_p && !hc->tx_x_calc); - if (hc->tx_x_calc > (hc->tx_x_recv >> 5)) hc->tx_x_recv = max(hc->tx_x_recv / 2, @@ -256,7 +257,7 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data) * Set new timeout for the nofeedback timer. * See comments in packet_recv() regarding the value of t_RTO. */ - if (unlikely(hc->tx_t_rto == 0)) /* no feedback yet */ + if (unlikely(hc->tx_t_rto == 0)) /* no feedback received yet */ t_nfb = TFRC_INITIAL_TIMEOUT; else t_nfb = max(hc->tx_t_rto, 2 * hc->tx_t_ipi); @@ -269,11 +270,12 @@ out: sock_put(sk); } -/* - * returns - * > 0: delay (in msecs) that should pass before actually sending - * = 0: can send immediately - * < 0: error condition; do not send packet +/** + * ccid3_hc_tx_send_packet - Delay-based dequeueing of TX packets + * @skb: next packet candidate to send on @sk + * + * This function uses the convention of ccid_packet_dequeue_eval() and + * returns a millisecond-delay value between 0 and t_mbi = 64000 msec. */ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb) { @@ -290,8 +292,7 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb) if (unlikely(skb->len == 0)) return -EBADMSG; - switch (hc->tx_state) { - case TFRC_SSTATE_NO_SENT: + if (hc->tx_state == TFRC_SSTATE_NO_SENT) { sk_reset_timer(sk, &hc->tx_no_feedback_timer, (jiffies + usecs_to_jiffies(TFRC_INITIAL_TIMEOUT))); hc->tx_last_win_count = 0; @@ -326,27 +327,22 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb) ccid3_update_send_interval(hc); ccid3_hc_tx_set_state(sk, TFRC_SSTATE_NO_FBACK); - break; - case TFRC_SSTATE_NO_FBACK: - case TFRC_SSTATE_FBACK: + + } else { delay = ktime_us_delta(hc->tx_t_nom, now); ccid3_pr_debug("delay=%ld\n", (long)delay); /* - * Scheduling of packet transmissions [RFC 3448, 4.6] + * Scheduling of packet transmissions (RFC 5348, 8.3) * * if (t_now > t_nom - delta) * // send the packet now * else * // send the packet in (t_nom - t_now) milliseconds. */ - if (delay - (s64)hc->tx_delta >= 1000) - return (u32)delay / 1000L; + if (delay >= TFRC_T_DELTA) + return (u32)delay / USEC_PER_MSEC; ccid3_hc_tx_update_win_count(hc, now); - break; - case TFRC_SSTATE_TERM: - DCCP_BUG("%s(%p) - Illegal state TERM", dccp_role(sk), sk); - return -EINVAL; } /* prepare to send now (add options etc.) */ @@ -355,11 +351,10 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb) /* set the nominal send time for the next following packet */ hc->tx_t_nom = ktime_add_us(hc->tx_t_nom, hc->tx_t_ipi); - return 0; + return CCID_PACKET_SEND_AT_ONCE; } -static void ccid3_hc_tx_packet_sent(struct sock *sk, int more, - unsigned int len) +static void ccid3_hc_tx_packet_sent(struct sock *sk, unsigned int len) { struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk); @@ -372,48 +367,34 @@ static void ccid3_hc_tx_packet_sent(struct sock *sk, int more, static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) { struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk); - struct ccid3_options_received *opt_recv; + struct tfrc_tx_hist_entry *acked; ktime_t now; unsigned long t_nfb; - u32 pinv, r_sample; + u32 r_sample; /* we are only interested in ACKs */ if (!(DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_ACK || DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_DATAACK)) return; - /* ... and only in the established state */ - if (hc->tx_state != TFRC_SSTATE_FBACK && - hc->tx_state != TFRC_SSTATE_NO_FBACK) - return; - - opt_recv = &hc->tx_options_received; - now = ktime_get_real(); - - /* Estimate RTT from history if ACK number is valid */ - r_sample = tfrc_tx_hist_rtt(hc->tx_hist, - DCCP_SKB_CB(skb)->dccpd_ack_seq, now); - if (r_sample == 0) { - DCCP_WARN("%s(%p): %s with bogus ACK-%llu\n", dccp_role(sk), sk, - dccp_packet_name(DCCP_SKB_CB(skb)->dccpd_type), - (unsigned long long)DCCP_SKB_CB(skb)->dccpd_ack_seq); - return; - } - - /* Update receive rate in units of 64 * bytes/second */ - hc->tx_x_recv = opt_recv->ccid3or_receive_rate; - hc->tx_x_recv <<= 6; - - /* Update loss event rate (which is scaled by 1e6) */ - pinv = opt_recv->ccid3or_loss_event_rate; - if (pinv == ~0U || pinv == 0) /* see RFC 4342, 8.5 */ - hc->tx_p = 0; - else /* can not exceed 100% */ - hc->tx_p = scaled_div(1, pinv); /* - * Validate new RTT sample and update moving average + * Locate the acknowledged packet in the TX history. + * + * Returning "entry not found" here can for instance happen when + * - the host has not sent out anything (e.g. a passive server), + * - the Ack is outdated (packet with higher Ack number was received), + * - it is a bogus Ack (for a packet not sent on this connection). */ - r_sample = dccp_sample_rtt(sk, r_sample); + acked = tfrc_tx_hist_find_entry(hc->tx_hist, dccp_hdr_ack_seq(skb)); + if (acked == NULL) + return; + /* For the sake of RTT sampling, ignore/remove all older entries */ + tfrc_tx_hist_purge(&acked->next); + + /* Update the moving average for the RTT estimate (RFC 3448, 4.3) */ + now = ktime_get_real(); + r_sample = dccp_sample_rtt(sk, ktime_us_delta(now, acked->stamp)); hc->tx_rtt = tfrc_ewma(hc->tx_rtt, r_sample, 9); + /* * Update allowed sending rate X as per draft rfc3448bis-00, 4.2/3 */ @@ -448,8 +429,8 @@ done_computing_x: "p=%u, X_calc=%u, X_recv=%u, X=%u\n", dccp_role(sk), sk, hc->tx_rtt, r_sample, hc->tx_s, hc->tx_p, hc->tx_x_calc, - (unsigned)(hc->tx_x_recv >> 6), - (unsigned)(hc->tx_x >> 6)); + (unsigned int)(hc->tx_x_recv >> 6), + (unsigned int)(hc->tx_x >> 6)); /* unschedule no feedback timer */ sk_stop_timer(sk, &hc->tx_no_feedback_timer); @@ -461,13 +442,12 @@ done_computing_x: sk->sk_write_space(sk); /* - * Update timeout interval for the nofeedback timer. - * We use a configuration option to increase the lower bound. - * This can help avoid triggering the nofeedback timer too - * often ('spinning') on LANs with small RTTs. + * Update timeout interval for the nofeedback timer. In order to control + * rate halving on networks with very low RTTs (<= 1 ms), use per-route + * tunable RTAX_RTO_MIN value as the lower bound. */ - hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt, (CONFIG_IP_DCCP_CCID3_RTO * - (USEC_PER_SEC / 1000))); + hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt, + USEC_PER_SEC/HZ * tcp_rto_min(sk)); /* * Schedule no feedback timer to expire in * max(t_RTO, 2 * s/X) = max(t_RTO, 2 * t_ipi) @@ -482,66 +462,41 @@ done_computing_x: jiffies + usecs_to_jiffies(t_nfb)); } -static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option, - unsigned char len, u16 idx, - unsigned char *value) +static int ccid3_hc_tx_parse_options(struct sock *sk, u8 packet_type, + u8 option, u8 *optval, u8 optlen) { - int rc = 0; - const struct dccp_sock *dp = dccp_sk(sk); struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk); - struct ccid3_options_received *opt_recv; __be32 opt_val; - opt_recv = &hc->tx_options_received; - - if (opt_recv->ccid3or_seqno != dp->dccps_gsr) { - opt_recv->ccid3or_seqno = dp->dccps_gsr; - opt_recv->ccid3or_loss_event_rate = ~0; - opt_recv->ccid3or_loss_intervals_idx = 0; - opt_recv->ccid3or_loss_intervals_len = 0; - opt_recv->ccid3or_receive_rate = 0; - } - switch (option) { + case TFRC_OPT_RECEIVE_RATE: case TFRC_OPT_LOSS_EVENT_RATE: - if (unlikely(len != 4)) { - DCCP_WARN("%s(%p), invalid len %d " - "for TFRC_OPT_LOSS_EVENT_RATE\n", - dccp_role(sk), sk, len); - rc = -EINVAL; - } else { - opt_val = get_unaligned((__be32 *)value); - opt_recv->ccid3or_loss_event_rate = ntohl(opt_val); - ccid3_pr_debug("%s(%p), LOSS_EVENT_RATE=%u\n", - dccp_role(sk), sk, - opt_recv->ccid3or_loss_event_rate); + /* Must be ignored on Data packets, cf. RFC 4342 8.3 and 8.5 */ + if (packet_type == DCCP_PKT_DATA) + break; + if (unlikely(optlen != 4)) { + DCCP_WARN("%s(%p), invalid len %d for %u\n", + dccp_role(sk), sk, optlen, option); + return -EINVAL; } - break; - case TFRC_OPT_LOSS_INTERVALS: - opt_recv->ccid3or_loss_intervals_idx = idx; - opt_recv->ccid3or_loss_intervals_len = len; - ccid3_pr_debug("%s(%p), LOSS_INTERVALS=(%u, %u)\n", - dccp_role(sk), sk, - opt_recv->ccid3or_loss_intervals_idx, - opt_recv->ccid3or_loss_intervals_len); - break; - case TFRC_OPT_RECEIVE_RATE: - if (unlikely(len != 4)) { - DCCP_WARN("%s(%p), invalid len %d " - "for TFRC_OPT_RECEIVE_RATE\n", - dccp_role(sk), sk, len); - rc = -EINVAL; - } else { - opt_val = get_unaligned((__be32 *)value); - opt_recv->ccid3or_receive_rate = ntohl(opt_val); + opt_val = ntohl(get_unaligned((__be32 *)optval)); + + if (option == TFRC_OPT_RECEIVE_RATE) { + /* Receive Rate is kept in units of 64 bytes/second */ + hc->tx_x_recv = opt_val; + hc->tx_x_recv <<= 6; + ccid3_pr_debug("%s(%p), RECEIVE_RATE=%u\n", - dccp_role(sk), sk, - opt_recv->ccid3or_receive_rate); + dccp_role(sk), sk, opt_val); + } else { + /* Update the fixpoint Loss Event Rate fraction */ + hc->tx_p = tfrc_invert_loss_event_rate(opt_val); + + ccid3_pr_debug("%s(%p), LOSS_EVENT_RATE=%u\n", + dccp_role(sk), sk, opt_val); } - break; } - - return rc; + return 0; } static int ccid3_hc_tx_init(struct ccid *ccid, struct sock *sk) @@ -559,42 +514,37 @@ static void ccid3_hc_tx_exit(struct sock *sk) { struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk); - ccid3_hc_tx_set_state(sk, TFRC_SSTATE_TERM); sk_stop_timer(sk, &hc->tx_no_feedback_timer); - tfrc_tx_hist_purge(&hc->tx_hist); } static void ccid3_hc_tx_get_info(struct sock *sk, struct tcp_info *info) { - struct ccid3_hc_tx_sock *hc; - - /* Listen socks doesn't have a private CCID block */ - if (sk->sk_state == DCCP_LISTEN) - return; - - hc = ccid3_hc_tx_sk(sk); - info->tcpi_rto = hc->tx_t_rto; - info->tcpi_rtt = hc->tx_rtt; + info->tcpi_rto = ccid3_hc_tx_sk(sk)->tx_t_rto; + info->tcpi_rtt = ccid3_hc_tx_sk(sk)->tx_rtt; } static int ccid3_hc_tx_getsockopt(struct sock *sk, const int optname, int len, u32 __user *optval, int __user *optlen) { - const struct ccid3_hc_tx_sock *hc; + const struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk); + struct tfrc_tx_info tfrc; const void *val; - /* Listen socks doesn't have a private CCID block */ - if (sk->sk_state == DCCP_LISTEN) - return -EINVAL; - - hc = ccid3_hc_tx_sk(sk); switch (optname) { case DCCP_SOCKOPT_CCID_TX_INFO: - if (len < sizeof(hc->tx_tfrc)) + if (len < sizeof(tfrc)) return -EINVAL; - len = sizeof(hc->tx_tfrc); - val = &hc->tx_tfrc; + memset(&tfrc, 0, sizeof(tfrc)); + tfrc.tfrctx_x = hc->tx_x; + tfrc.tfrctx_x_recv = hc->tx_x_recv; + tfrc.tfrctx_x_calc = hc->tx_x_calc; + tfrc.tfrctx_rtt = hc->tx_rtt; + tfrc.tfrctx_p = hc->tx_p; + tfrc.tfrctx_rto = hc->tx_t_rto; + tfrc.tfrctx_ipi = hc->tx_t_ipi; + len = sizeof(tfrc); + val = &tfrc; break; default: return -ENOPROTOOPT; @@ -624,7 +574,6 @@ static const char *ccid3_rx_state_name(enum ccid3_hc_rx_states state) static const char *const ccid3_rx_state_names[] = { [TFRC_RSTATE_NO_DATA] = "NO_DATA", [TFRC_RSTATE_DATA] = "DATA", - [TFRC_RSTATE_TERM] = "TERM", }; return ccid3_rx_state_names[state]; @@ -650,14 +599,9 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk, { struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk); struct dccp_sock *dp = dccp_sk(sk); - ktime_t now; + ktime_t now = ktime_get_real(); s64 delta = 0; - if (unlikely(hc->rx_state == TFRC_RSTATE_TERM)) - return; - - now = ktime_get_real(); - switch (fbtype) { case CCID3_FBACK_INITIAL: hc->rx_x_recv = 0; @@ -701,23 +645,21 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk, static int ccid3_hc_rx_insert_options(struct sock *sk, struct sk_buff *skb) { - const struct ccid3_hc_rx_sock *hc; + const struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk); __be32 x_recv, pinv; if (!(sk->sk_state == DCCP_OPEN || sk->sk_state == DCCP_PARTOPEN)) return 0; - hc = ccid3_hc_rx_sk(sk); - if (dccp_packet_without_ack(skb)) return 0; x_recv = htonl(hc->rx_x_recv); pinv = htonl(hc->rx_pinv); - if (dccp_insert_option(sk, skb, TFRC_OPT_LOSS_EVENT_RATE, + if (dccp_insert_option(skb, TFRC_OPT_LOSS_EVENT_RATE, &pinv, sizeof(pinv)) || - dccp_insert_option(sk, skb, TFRC_OPT_RECEIVE_RATE, + dccp_insert_option(skb, TFRC_OPT_RECEIVE_RATE, &x_recv, sizeof(x_recv))) return -1; @@ -749,10 +691,11 @@ static u32 ccid3_first_li(struct sock *sk) x_recv = scaled_div32(hc->rx_bytes_recv, delta); if (x_recv == 0) { /* would also trigger divide-by-zero */ DCCP_WARN("X_recv==0\n"); - if ((x_recv = hc->rx_x_recv) == 0) { + if (hc->rx_x_recv == 0) { DCCP_BUG("stored value of X_recv is zero"); return ~0U; } + x_recv = hc->rx_x_recv; } fval = scaled_div(hc->rx_s, hc->rx_rtt); @@ -862,46 +805,31 @@ static void ccid3_hc_rx_exit(struct sock *sk) { struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk); - ccid3_hc_rx_set_state(sk, TFRC_RSTATE_TERM); - tfrc_rx_hist_purge(&hc->rx_hist); tfrc_lh_cleanup(&hc->rx_li_hist); } static void ccid3_hc_rx_get_info(struct sock *sk, struct tcp_info *info) { - const struct ccid3_hc_rx_sock *hc; - - /* Listen socks doesn't have a private CCID block */ - if (sk->sk_state == DCCP_LISTEN) - return; - - hc = ccid3_hc_rx_sk(sk); - info->tcpi_ca_state = hc->rx_state; + info->tcpi_ca_state = ccid3_hc_rx_sk(sk)->rx_state; info->tcpi_options |= TCPI_OPT_TIMESTAMPS; - info->tcpi_rcv_rtt = hc->rx_rtt; + info->tcpi_rcv_rtt = ccid3_hc_rx_sk(sk)->rx_rtt; } static int ccid3_hc_rx_getsockopt(struct sock *sk, const int optname, int len, u32 __user *optval, int __user *optlen) { - const struct ccid3_hc_rx_sock *hc; + const struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk); struct tfrc_rx_info rx_info; const void *val; - /* Listen socks doesn't have a private CCID block */ - if (sk->sk_state == DCCP_LISTEN) - return -EINVAL; - - hc = ccid3_hc_rx_sk(sk); switch (optname) { case DCCP_SOCKOPT_CCID_RX_INFO: if (len < sizeof(rx_info)) return -EINVAL; rx_info.tfrcrx_x_recv = hc->rx_x_recv; rx_info.tfrcrx_rtt = hc->rx_rtt; - rx_info.tfrcrx_p = hc->rx_pinv == 0 ? ~0U : - scaled_div(1, hc->rx_pinv); + rx_info.tfrcrx_p = tfrc_invert_loss_event_rate(hc->rx_pinv); len = sizeof(rx_info); val = &rx_info; break; diff --git a/net/dccp/ccids/ccid3.h b/net/dccp/ccids/ccid3.h index 03263577665..1a9933c2967 100644 --- a/net/dccp/ccids/ccid3.h +++ b/net/dccp/ccids/ccid3.h @@ -42,35 +42,36 @@ #include "lib/tfrc.h" #include "../ccid.h" -/* Two seconds as per RFC 3448 4.2 */ +/* Two seconds as per RFC 5348, 4.2 */ #define TFRC_INITIAL_TIMEOUT (2 * USEC_PER_SEC) -/* In usecs - half the scheduling granularity as per RFC3448 4.6 */ -#define TFRC_OPSYS_HALF_TIME_GRAN (USEC_PER_SEC / (2 * HZ)) - /* Parameter t_mbi from [RFC 3448, 4.3]: backoff interval in seconds */ #define TFRC_T_MBI 64 +/* + * The t_delta parameter (RFC 5348, 8.3): delays of less than %USEC_PER_MSEC are + * rounded down to 0, since sk_reset_timer() here uses millisecond granularity. + * Hence we can use a constant t_delta = %USEC_PER_MSEC when HZ >= 500. A coarse + * resolution of HZ < 500 means that the error is below one timer tick (t_gran) + * when using the constant t_delta = t_gran / 2 = %USEC_PER_SEC / (2 * HZ). + */ +#if (HZ >= 500) +# define TFRC_T_DELTA USEC_PER_MSEC +#else +# define TFRC_T_DELTA (USEC_PER_SEC / (2 * HZ)) +#endif + enum ccid3_options { TFRC_OPT_LOSS_EVENT_RATE = 192, TFRC_OPT_LOSS_INTERVALS = 193, TFRC_OPT_RECEIVE_RATE = 194, }; -struct ccid3_options_received { - u64 ccid3or_seqno:48, - ccid3or_loss_intervals_idx:16; - u16 ccid3or_loss_intervals_len; - u32 ccid3or_loss_event_rate; - u32 ccid3or_receive_rate; -}; - /* TFRC sender states */ enum ccid3_hc_tx_states { TFRC_SSTATE_NO_SENT = 1, TFRC_SSTATE_NO_FBACK, TFRC_SSTATE_FBACK, - TFRC_SSTATE_TERM, }; /** @@ -90,19 +91,16 @@ enum ccid3_hc_tx_states { * @tx_no_feedback_timer: Handle to no feedback timer * @tx_t_ld: Time last doubled during slow start * @tx_t_nom: Nominal send time of next packet - * @tx_delta: Send timer delta (RFC 3448, 4.6) in usecs * @tx_hist: Packet history - * @tx_options_received: Parsed set of retrieved options */ struct ccid3_hc_tx_sock { - struct tfrc_tx_info tx_tfrc; -#define tx_x tx_tfrc.tfrctx_x -#define tx_x_recv tx_tfrc.tfrctx_x_recv -#define tx_x_calc tx_tfrc.tfrctx_x_calc -#define tx_rtt tx_tfrc.tfrctx_rtt -#define tx_p tx_tfrc.tfrctx_p -#define tx_t_rto tx_tfrc.tfrctx_rto -#define tx_t_ipi tx_tfrc.tfrctx_ipi + u64 tx_x; + u64 tx_x_recv; + u32 tx_x_calc; + u32 tx_rtt; + u32 tx_p; + u32 tx_t_rto; + u32 tx_t_ipi; u16 tx_s; enum ccid3_hc_tx_states tx_state:8; u8 tx_last_win_count; @@ -110,9 +108,7 @@ struct ccid3_hc_tx_sock { struct timer_list tx_no_feedback_timer; ktime_t tx_t_ld; ktime_t tx_t_nom; - u32 tx_delta; struct tfrc_tx_hist_entry *tx_hist; - struct ccid3_options_received tx_options_received; }; static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk) @@ -126,21 +122,16 @@ static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk) enum ccid3_hc_rx_states { TFRC_RSTATE_NO_DATA = 1, TFRC_RSTATE_DATA, - TFRC_RSTATE_TERM = 127, }; /** * struct ccid3_hc_rx_sock - CCID3 receiver half-connection socket - * @rx_x_recv: Receiver estimate of send rate (RFC 3448 4.3) - * @rx_rtt: Receiver estimate of rtt (non-standard) - * @rx_p: Current loss event rate (RFC 3448 5.4) * @rx_last_counter: Tracks window counter (RFC 4342, 8.1) * @rx_state: Receiver state, one of %ccid3_hc_rx_states * @rx_bytes_recv: Total sum of DCCP payload bytes * @rx_x_recv: Receiver estimate of send rate (RFC 3448, sec. 4.3) * @rx_rtt: Receiver estimate of RTT * @rx_tstamp_last_feedback: Time at which last feedback was sent - * @rx_tstamp_last_ack: Time at which last feedback was sent * @rx_hist: Packet history (loss detection + RTT sampling) * @rx_li_hist: Loss Interval database * @rx_s: Received packet size in bytes diff --git a/net/dccp/ccids/lib/loss_interval.c b/net/dccp/ccids/lib/loss_interval.c index 8fc3cbf7907..57f9fd78c4d 100644 --- a/net/dccp/ccids/lib/loss_interval.c +++ b/net/dccp/ccids/lib/loss_interval.c @@ -116,7 +116,7 @@ u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *skb) cur->li_length = len; tfrc_lh_calc_i_mean(lh); - return (lh->i_mean < old_i_mean); + return lh->i_mean < old_i_mean; } /* Determine if `new_loss' does begin a new loss interval [RFC 4342, 10.2] */ @@ -133,6 +133,7 @@ static inline u8 tfrc_lh_is_new_loss(struct tfrc_loss_interval *cur, * @rh: Receive history containing a fresh loss event * @calc_first_li: Caller-dependent routine to compute length of first interval * @sk: Used by @calc_first_li in caller-specific way (subtyping) + * * Updates I_mean and returns 1 if a new interval has in fact been added to @lh. */ int tfrc_lh_interval_add(struct tfrc_loss_hist *lh, struct tfrc_rx_hist *rh, diff --git a/net/dccp/ccids/lib/loss_interval.h b/net/dccp/ccids/lib/loss_interval.h index d1d2f5383b7..57f631a86cc 100644 --- a/net/dccp/ccids/lib/loss_interval.h +++ b/net/dccp/ccids/lib/loss_interval.h @@ -65,9 +65,9 @@ static inline u8 tfrc_lh_length(struct tfrc_loss_hist *lh) struct tfrc_rx_hist; -extern int tfrc_lh_interval_add(struct tfrc_loss_hist *, struct tfrc_rx_hist *, - u32 (*first_li)(struct sock *), struct sock *); -extern u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *); -extern void tfrc_lh_cleanup(struct tfrc_loss_hist *lh); +int tfrc_lh_interval_add(struct tfrc_loss_hist *, struct tfrc_rx_hist *, + u32 (*first_li)(struct sock *), struct sock *); +u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *); +void tfrc_lh_cleanup(struct tfrc_loss_hist *lh); #endif /* _DCCP_LI_HIST_ */ diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c index 3a4f414e94a..08df7a3acb3 100644 --- a/net/dccp/ccids/lib/packet_history.c +++ b/net/dccp/ccids/lib/packet_history.c @@ -38,18 +38,6 @@ #include "packet_history.h" #include "../../dccp.h" -/** - * tfrc_tx_hist_entry - Simple singly-linked TX history list - * @next: next oldest entry (LIFO order) - * @seqno: sequence number of this entry - * @stamp: send time of packet with sequence number @seqno - */ -struct tfrc_tx_hist_entry { - struct tfrc_tx_hist_entry *next; - u64 seqno; - ktime_t stamp; -}; - /* * Transmitter History Routines */ @@ -71,15 +59,6 @@ void tfrc_tx_packet_history_exit(void) } } -static struct tfrc_tx_hist_entry * - tfrc_tx_hist_find_entry(struct tfrc_tx_hist_entry *head, u64 seqno) -{ - while (head != NULL && head->seqno != seqno) - head = head->next; - - return head; -} - int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno) { struct tfrc_tx_hist_entry *entry = kmem_cache_alloc(tfrc_tx_hist_slab, gfp_any()); @@ -107,24 +86,6 @@ void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp) *headp = NULL; } -u32 tfrc_tx_hist_rtt(struct tfrc_tx_hist_entry *head, const u64 seqno, - const ktime_t now) -{ - u32 rtt = 0; - struct tfrc_tx_hist_entry *packet = tfrc_tx_hist_find_entry(head, seqno); - - if (packet != NULL) { - rtt = ktime_us_delta(now, packet->stamp); - /* - * Garbage-collect older (irrelevant) entries: - */ - tfrc_tx_hist_purge(&packet->next); - } - - return rtt; -} - - /* * Receiver History Routines */ @@ -354,6 +315,7 @@ static void __three_after_loss(struct tfrc_rx_hist *h) * @ndp: The NDP count belonging to @skb * @calc_first_li: Caller-dependent computation of first loss interval in @lh * @sk: Used by @calc_first_li (see tfrc_lh_interval_add) + * * Chooses action according to pending loss, updates LI database when a new * loss was detected, and does required post-processing. Returns 1 when caller * should send feedback, 0 otherwise. @@ -426,7 +388,7 @@ static inline struct tfrc_rx_hist_entry * } /** - * tfrc_rx_hist_rtt_prev_s: previously suitable (wrt rtt_last_s) RTT-sampling entry + * tfrc_rx_hist_rtt_prev_s - previously suitable (wrt rtt_last_s) RTT-sampling entry */ static inline struct tfrc_rx_hist_entry * tfrc_rx_hist_rtt_prev_s(const struct tfrc_rx_hist *h) diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h index 7df6c529999..ee362b0b630 100644 --- a/net/dccp/ccids/lib/packet_history.h +++ b/net/dccp/ccids/lib/packet_history.h @@ -40,12 +40,28 @@ #include <linux/slab.h> #include "tfrc.h" -struct tfrc_tx_hist_entry; +/** + * tfrc_tx_hist_entry - Simple singly-linked TX history list + * @next: next oldest entry (LIFO order) + * @seqno: sequence number of this entry + * @stamp: send time of packet with sequence number @seqno + */ +struct tfrc_tx_hist_entry { + struct tfrc_tx_hist_entry *next; + u64 seqno; + ktime_t stamp; +}; + +static inline struct tfrc_tx_hist_entry * + tfrc_tx_hist_find_entry(struct tfrc_tx_hist_entry *head, u64 seqno) +{ + while (head != NULL && head->seqno != seqno) + head = head->next; + return head; +} -extern int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno); -extern void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp); -extern u32 tfrc_tx_hist_rtt(struct tfrc_tx_hist_entry *head, - const u64 seqno, const ktime_t now); +int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno); +void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp); /* Subtraction a-b modulo-16, respects circular wrap-around */ #define SUB16(a, b) (((a) + 16 - (b)) & 0xF) @@ -123,20 +139,17 @@ static inline bool tfrc_rx_hist_loss_pending(const struct tfrc_rx_hist *h) return h->loss_count > 0; } -extern void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, - const struct sk_buff *skb, const u64 ndp); +void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, const struct sk_buff *skb, + const u64 ndp); -extern int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb); +int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb); struct tfrc_loss_hist; -extern int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, - struct tfrc_loss_hist *lh, - struct sk_buff *skb, const u64 ndp, - u32 (*first_li)(struct sock *sk), - struct sock *sk); -extern u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, - const struct sk_buff *skb); -extern int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h); -extern void tfrc_rx_hist_purge(struct tfrc_rx_hist *h); +int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, struct tfrc_loss_hist *lh, + struct sk_buff *skb, const u64 ndp, + u32 (*first_li)(struct sock *sk), struct sock *sk); +u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, const struct sk_buff *skb); +int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h); +void tfrc_rx_hist_purge(struct tfrc_rx_hist *h); #endif /* _DCCP_PKT_HIST_ */ diff --git a/net/dccp/ccids/lib/tfrc.c b/net/dccp/ccids/lib/tfrc.c index 4902029854d..62b5828acde 100644 --- a/net/dccp/ccids/lib/tfrc.c +++ b/net/dccp/ccids/lib/tfrc.c @@ -4,10 +4,11 @@ * Copyright (c) 2007 The University of Aberdeen, Scotland, UK * Copyright (c) 2007 Arnaldo Carvalho de Melo <acme@redhat.com> */ +#include <linux/moduleparam.h> #include "tfrc.h" #ifdef CONFIG_IP_DCCP_TFRC_DEBUG -int tfrc_debug; +bool tfrc_debug; module_param(tfrc_debug, bool, 0644); MODULE_PARM_DESC(tfrc_debug, "Enable TFRC debug messages"); #endif diff --git a/net/dccp/ccids/lib/tfrc.h b/net/dccp/ccids/lib/tfrc.h index 01bb48e96c2..40ee7d62b65 100644 --- a/net/dccp/ccids/lib/tfrc.h +++ b/net/dccp/ccids/lib/tfrc.h @@ -21,7 +21,7 @@ #include "packet_history.h" #ifdef CONFIG_IP_DCCP_TFRC_DEBUG -extern int tfrc_debug; +extern bool tfrc_debug; #define tfrc_pr_debug(format, a...) DCCP_PR_DEBUG(tfrc_debug, format, ##a) #else #define tfrc_pr_debug(format, a...) @@ -55,20 +55,21 @@ static inline u32 tfrc_ewma(const u32 avg, const u32 newval, const u8 weight) return avg ? (weight * avg + (10 - weight) * newval) / 10 : newval; } -extern u32 tfrc_calc_x(u16 s, u32 R, u32 p); -extern u32 tfrc_calc_x_reverse_lookup(u32 fvalue); +u32 tfrc_calc_x(u16 s, u32 R, u32 p); +u32 tfrc_calc_x_reverse_lookup(u32 fvalue); +u32 tfrc_invert_loss_event_rate(u32 loss_event_rate); -extern int tfrc_tx_packet_history_init(void); -extern void tfrc_tx_packet_history_exit(void); -extern int tfrc_rx_packet_history_init(void); -extern void tfrc_rx_packet_history_exit(void); +int tfrc_tx_packet_history_init(void); +void tfrc_tx_packet_history_exit(void); +int tfrc_rx_packet_history_init(void); +void tfrc_rx_packet_history_exit(void); -extern int tfrc_li_init(void); -extern void tfrc_li_exit(void); +int tfrc_li_init(void); +void tfrc_li_exit(void); #ifdef CONFIG_IP_DCCP_TFRC_LIB -extern int tfrc_lib_init(void); -extern void tfrc_lib_exit(void); +int tfrc_lib_init(void); +void tfrc_lib_exit(void); #else #define tfrc_lib_init() (0) #define tfrc_lib_exit() diff --git a/net/dccp/ccids/lib/tfrc_equation.c b/net/dccp/ccids/lib/tfrc_equation.c index 22ca1cf0eb5..88ef98285be 100644 --- a/net/dccp/ccids/lib/tfrc_equation.c +++ b/net/dccp/ccids/lib/tfrc_equation.c @@ -611,6 +611,7 @@ static inline u32 tfrc_binsearch(u32 fval, u8 small) * @s: packet size in bytes * @R: RTT scaled by 1000000 (i.e., microseconds) * @p: loss ratio estimate scaled by 1000000 + * * Returns X_calc in bytes per second (not scaled). */ u32 tfrc_calc_x(u16 s, u32 R, u32 p) @@ -659,6 +660,7 @@ u32 tfrc_calc_x(u16 s, u32 R, u32 p) /** * tfrc_calc_x_reverse_lookup - try to find p given f(p) * @fvalue: function value to match, scaled by 1000000 + * * Returns closest match for p, also scaled by 1000000 */ u32 tfrc_calc_x_reverse_lookup(u32 fvalue) @@ -687,3 +689,17 @@ u32 tfrc_calc_x_reverse_lookup(u32 fvalue) index = tfrc_binsearch(fvalue, 0); return (index + 1) * 1000000 / TFRC_CALC_X_ARRSIZE; } + +/** + * tfrc_invert_loss_event_rate - Compute p so that 10^6 corresponds to 100% + * When @loss_event_rate is large, there is a chance that p is truncated to 0. + * To avoid re-entering slow-start in that case, we set p = TFRC_SMALLEST_P > 0. + */ +u32 tfrc_invert_loss_event_rate(u32 loss_event_rate) +{ + if (loss_event_rate == UINT_MAX) /* see RFC 4342, 8.5 */ + return 0; + if (unlikely(loss_event_rate == 0)) /* map 1/0 into 100% */ + return 1000000; + return max_t(u32, scaled_div(1, loss_event_rate), TFRC_SMALLEST_P); +} |