X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=net%2Fipv4%2Ftcp_cubic.c;h=80bd084a9f9198d5b4c31f380ffcd5448a8390b0;hb=c6fda282294da882f8d8cc4c513940277dd380f5;hp=78b7a6b9e4defed2a15d99ee6cc7356e52f69042;hpb=72dc5b9225c53310c010b68a70ea97c8c8e24bdf;p=safe%2Fjmp%2Flinux-2.6 diff --git a/net/ipv4/tcp_cubic.c b/net/ipv4/tcp_cubic.c index 78b7a6b..80bd084 100644 --- a/net/ipv4/tcp_cubic.c +++ b/net/ipv4/tcp_cubic.c @@ -1,5 +1,5 @@ /* - * TCP CUBIC: Binary Increase Congestion control for TCP v2.0 + * TCP CUBIC: Binary Increase Congestion control for TCP v2.1 * * This is from the implementation of CUBIC TCP in * Injong Rhee, Lisong Xu. @@ -12,7 +12,6 @@ * this behaves the same as the original Reno. */ -#include #include #include #include @@ -27,16 +26,16 @@ */ #define BICTCP_HZ 10 /* BIC HZ 2^10 = 1024 */ -static int fast_convergence = 1; -static int max_increment = 16; -static int beta = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */ -static int initial_ssthresh = 100; -static int bic_scale = 41; -static int tcp_friendliness = 1; +static int fast_convergence __read_mostly = 1; +static int max_increment __read_mostly = 16; +static int beta __read_mostly = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */ +static int initial_ssthresh __read_mostly; +static int bic_scale __read_mostly = 41; +static int tcp_friendliness __read_mostly = 1; -static u32 cube_rtt_scale; -static u32 beta_scale; -static u64 cube_factor; +static u32 cube_rtt_scale __read_mostly; +static u32 beta_scale __read_mostly; +static u64 cube_factor __read_mostly; /* Note parameters that are used for precomputing scale factors are read-only */ module_param(fast_convergence, int, 0644); @@ -52,8 +51,6 @@ MODULE_PARM_DESC(bic_scale, "scale (scaled by 1024) value for bic function (bic_ module_param(tcp_friendliness, int, 0644); MODULE_PARM_DESC(tcp_friendliness, "turn on/off tcp friendliness"); -#include - /* BIC TCP Parameters */ struct bictcp { u32 cnt; /* increase cwnd by 1 after ACKs */ @@ -94,50 +91,51 @@ static void bictcp_init(struct sock *sk) tcp_sk(sk)->snd_ssthresh = initial_ssthresh; } -/* 64bit divisor, dividend and result. dynamic precision */ -static inline u_int64_t div64_64(u_int64_t dividend, u_int64_t divisor) -{ - u_int32_t d = divisor; - - if (divisor > 0xffffffffULL) { - unsigned int shift = fls(divisor >> 32); - - d = divisor >> shift; - dividend >>= shift; - } - - /* avoid 64 bit division if possible */ - if (dividend >> 32) - do_div(dividend, d); - else - dividend = (uint32_t) dividend / d; - - return dividend; -} - -/* - * calculate the cubic root of x using Newton-Raphson +/* calculate the cubic root of x using a table lookup followed by one + * Newton-Raphson iteration. + * Avg err ~= 0.195% */ static u32 cubic_root(u64 a) { - u32 x, x1; - - /* Initial estimate is based on: - * cbrt(x) = exp(log(x) / 3) + u32 x, b, shift; + /* + * cbrt(x) MSB values for x MSB values in [0..63]. + * Precomputed then refined by hand - Willy Tarreau + * + * For x in [0..63], + * v = cbrt(x << 18) - 1 + * cbrt(x) = (v[x] + 10) >> 6 */ - x = 1u << (fls64(a)/3); + static const u8 v[] = { + /* 0x00 */ 0, 54, 54, 54, 118, 118, 118, 118, + /* 0x08 */ 123, 129, 134, 138, 143, 147, 151, 156, + /* 0x10 */ 157, 161, 164, 168, 170, 173, 176, 179, + /* 0x18 */ 181, 185, 187, 190, 192, 194, 197, 199, + /* 0x20 */ 200, 202, 204, 206, 209, 211, 213, 215, + /* 0x28 */ 217, 219, 221, 222, 224, 225, 227, 229, + /* 0x30 */ 231, 232, 234, 236, 237, 239, 240, 242, + /* 0x38 */ 244, 245, 246, 248, 250, 251, 252, 254, + }; + + b = fls64(a); + if (b < 7) { + /* a in [0..63] */ + return ((u32)v[(u32)a] + 35) >> 6; + } + + b = ((b * 84) >> 8) - 1; + shift = (a >> (b * 3)); + + x = ((u32)(((u32)v[shift] + 10) << b)) >> 6; /* - * Iteration based on: + * Newton-Raphson iteration * 2 * x = ( 2 * x + a / x ) / 3 * k+1 k k */ - do { - x1 = x; - x = (2 * x + (uint32_t) div64_64(a, x*x)) / 3; - } while (abs(x1 - x) > 1); - + x = (2 * x + (u32)div64_64(a, (u64)x * (u64)(x - 1))); + x = ((x * 341) >> 10); return x; } @@ -176,51 +174,53 @@ static inline void bictcp_update(struct bictcp *ca, u32 cwnd) } } - /* cubic function - calc*/ - /* calculate c * time^3 / rtt, - * while considering overflow in calculation of time^3 + /* cubic function - calc*/ + /* calculate c * time^3 / rtt, + * while considering overflow in calculation of time^3 * (so time^3 is done by using 64 bit) * and without the support of division of 64bit numbers * (so all divisions are done by using 32 bit) - * also NOTE the unit of those veriables - * time = (t - K) / 2^bictcp_HZ - * c = bic_scale >> 10 + * also NOTE the unit of those veriables + * time = (t - K) / 2^bictcp_HZ + * c = bic_scale >> 10 * rtt = (srtt >> 3) / HZ * !!! The following code does not have overflow problems, * if the cwnd < 1 million packets !!! - */ + */ /* change the unit from HZ to bictcp_HZ */ - t = ((tcp_time_stamp + ca->delay_min - ca->epoch_start) + t = ((tcp_time_stamp + (ca->delay_min>>3) - ca->epoch_start) << BICTCP_HZ) / HZ; - if (t < ca->bic_K) /* t - K */ + if (t < ca->bic_K) /* t - K */ offs = ca->bic_K - t; - else - offs = t - ca->bic_K; + else + offs = t - ca->bic_K; /* c/rtt * (t-K)^3 */ delta = (cube_rtt_scale * offs * offs * offs) >> (10+3*BICTCP_HZ); - if (t < ca->bic_K) /* below origin*/ - bic_target = ca->bic_origin_point - delta; - else /* above origin*/ - bic_target = ca->bic_origin_point + delta; + if (t < ca->bic_K) /* below origin*/ + bic_target = ca->bic_origin_point - delta; + else /* above origin*/ + bic_target = ca->bic_origin_point + delta; - /* cubic function - calc bictcp_cnt*/ - if (bic_target > cwnd) { + /* cubic function - calc bictcp_cnt*/ + if (bic_target > cwnd) { ca->cnt = cwnd / (bic_target - cwnd); - } else { - ca->cnt = 100 * cwnd; /* very small increment*/ - } + } else { + ca->cnt = 100 * cwnd; /* very small increment*/ + } if (ca->delay_min > 0) { /* max increment = Smax * rtt / 0.1 */ min_cnt = (cwnd * HZ * 8)/(10 * max_increment * ca->delay_min); - if (ca->cnt < min_cnt) + + /* use concave growth when the target is above the origin */ + if (ca->cnt < min_cnt && t >= ca->bic_K) ca->cnt = min_cnt; } - /* slow start and low utilization */ + /* slow start and low utilization */ if (ca->loss_cwnd == 0) /* could be aggressive in slow start */ ca->cnt = 50; @@ -228,9 +228,9 @@ static inline void bictcp_update(struct bictcp *ca, u32 cwnd) if (tcp_friendliness) { u32 scale = beta_scale; delta = (cwnd * scale) >> 3; - while (ca->ack_cnt > delta) { /* update tcp cwnd */ - ca->ack_cnt -= delta; - ca->tcp_cwnd++; + while (ca->ack_cnt > delta) { /* update tcp cwnd */ + ca->ack_cnt -= delta; + ca->tcp_cwnd++; } if (ca->tcp_cwnd > cwnd){ /* if bic is slower than tcp */ @@ -239,45 +239,19 @@ static inline void bictcp_update(struct bictcp *ca, u32 cwnd) if (ca->cnt > max_cnt) ca->cnt = max_cnt; } - } + } ca->cnt = (ca->cnt << ACK_RATIO_SHIFT) / ca->delayed_ack; if (ca->cnt == 0) /* cannot be zero */ ca->cnt = 1; } - -/* Keep track of minimum rtt */ -static inline void measure_delay(struct sock *sk) -{ - const struct tcp_sock *tp = tcp_sk(sk); - struct bictcp *ca = inet_csk_ca(sk); - u32 delay; - - /* No time stamp */ - if (!(tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr) || - /* Discard delay samples right after fast recovery */ - (s32)(tcp_time_stamp - ca->epoch_start) < HZ) - return; - - delay = tcp_time_stamp - tp->rx_opt.rcv_tsecr; - if (delay == 0) - delay = 1; - - /* first time call or link delay decreases */ - if (ca->delay_min == 0 || ca->delay_min > delay) - ca->delay_min = delay; -} - static void bictcp_cong_avoid(struct sock *sk, u32 ack, - u32 seq_rtt, u32 in_flight, int data_acked) + u32 in_flight, int data_acked) { struct tcp_sock *tp = tcp_sk(sk); struct bictcp *ca = inet_csk_ca(sk); - if (data_acked) - measure_delay(sk); - if (!tcp_is_cwnd_limited(sk, in_flight)) return; @@ -334,17 +308,33 @@ static void bictcp_state(struct sock *sk, u8 new_state) /* Track delayed acknowledgment ratio using sliding window * ratio = (15*ratio + sample) / 16 */ -static void bictcp_acked(struct sock *sk, u32 cnt) +static void bictcp_acked(struct sock *sk, u32 cnt, s32 rtt_us) { const struct inet_connection_sock *icsk = inet_csk(sk); + struct bictcp *ca = inet_csk_ca(sk); + u32 delay; - if (cnt > 0 && icsk->icsk_ca_state == TCP_CA_Open) { - struct bictcp *ca = inet_csk_ca(sk); + if (icsk->icsk_ca_state == TCP_CA_Open) { cnt -= ca->delayed_ack >> ACK_RATIO_SHIFT; ca->delayed_ack += cnt; } -} + /* Some calls are for duplicates without timetamps */ + if (rtt_us < 0) + return; + + /* Discard delay samples right after fast recovery */ + if ((s32)(tcp_time_stamp - ca->epoch_start) < HZ) + return; + + delay = usecs_to_jiffies(rtt_us) << 3; + if (delay == 0) + delay = 1; + + /* first time call or link delay decreases */ + if (ca->delay_min == 0 || ca->delay_min > delay) + ca->delay_min = delay; +} static struct tcp_congestion_ops cubictcp = { .init = bictcp_init, @@ -359,7 +349,7 @@ static struct tcp_congestion_ops cubictcp = { static int __init cubictcp_register(void) { - BUG_ON(sizeof(struct bictcp) > ICSK_CA_PRIV_SIZE); + BUILD_BUG_ON(sizeof(struct bictcp) > ICSK_CA_PRIV_SIZE); /* Precompute a bunch of the scaling factors that are used per-packet * based on SRTT of 100ms @@ -367,7 +357,7 @@ static int __init cubictcp_register(void) beta_scale = 8*(BICTCP_BETA_SCALE+beta)/ 3 / (BICTCP_BETA_SCALE - beta); - cube_rtt_scale = (bic_scale << 3) / 10; /* 1024*c/rtt */ + cube_rtt_scale = (bic_scale * 10); /* 1024*c/rtt */ /* calculate the "K" for (wmax-cwnd) = c/rtt * K^3 * so K = cubic_root( (wmax-cwnd)*rtt/c ) @@ -402,4 +392,4 @@ module_exit(cubictcp_unregister); MODULE_AUTHOR("Sangtae Ha, Stephen Hemminger"); MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("CUBIC TCP"); -MODULE_VERSION("2.0"); +MODULE_VERSION("2.1");