summaryrefslogtreecommitdiffstats
path: root/net/dccp/ccids
diff options
context:
space:
mode:
Diffstat (limited to 'net/dccp/ccids')
-rw-r--r--net/dccp/ccids/Kconfig75
-rw-r--r--net/dccp/ccids/ccid2.c73
-rw-r--r--net/dccp/ccids/ccid2.h3
-rw-r--r--net/dccp/ccids/ccid3.c930
-rw-r--r--net/dccp/ccids/ccid3.h129
-rw-r--r--net/dccp/ccids/lib/loss_interval.c8
-rw-r--r--net/dccp/ccids/lib/loss_interval.h2
-rw-r--r--net/dccp/ccids/lib/packet_history.c219
-rw-r--r--net/dccp/ccids/lib/packet_history.h130
-rw-r--r--net/dccp/ccids/lib/tfrc.h23
-rw-r--r--net/dccp/ccids/lib/tfrc_equation.c241
11 files changed, 996 insertions, 837 deletions
diff --git a/net/dccp/ccids/Kconfig b/net/dccp/ccids/Kconfig
index 32752f75044..80f46988769 100644
--- a/net/dccp/ccids/Kconfig
+++ b/net/dccp/ccids/Kconfig
@@ -22,19 +22,26 @@ config IP_DCCP_CCID2
for lost packets, would prefer CCID 2 to CCID 3. On-line games may
also prefer CCID 2.
- CCID 2 is further described in:
- http://www.icir.org/kohler/dccp/draft-ietf-dccp-ccid2-10.txt
+ CCID 2 is further described in RFC 4341,
+ http://www.ietf.org/rfc/rfc4341.txt
- This text was extracted from:
- http://www.icir.org/kohler/dccp/draft-ietf-dccp-spec-13.txt
+ This text was extracted from RFC 4340 (sec. 10.1),
+ http://www.ietf.org/rfc/rfc4340.txt
+
+ To compile this CCID as a module, choose M here: the module will be
+ called dccp_ccid2.
If in doubt, say M.
config IP_DCCP_CCID2_DEBUG
- bool "CCID2 debug"
+ bool "CCID2 debugging messages"
depends on IP_DCCP_CCID2
---help---
- Enable CCID2 debug messages.
+ Enable CCID2-specific debugging messages.
+
+ When compiling CCID2 as a module, this debugging output can
+ additionally be toggled by setting the ccid2_debug module
+ parameter to 0 or 1.
If in doubt, say N.
@@ -53,20 +60,66 @@ config IP_DCCP_CCID3
suitable than CCID 2 for applications such streaming media where a
relatively smooth sending rate is of importance.
- CCID 3 is further described in:
-
- http://www.icir.org/kohler/dccp/draft-ietf-dccp-ccid3-11.txt.
+ CCID 3 is further described in RFC 4342,
+ http://www.ietf.org/rfc/rfc4342.txt
The TFRC congestion control algorithms were initially described in
RFC 3448.
- This text was extracted from:
- http://www.icir.org/kohler/dccp/draft-ietf-dccp-spec-13.txt
+ This text was extracted from RFC 4340 (sec. 10.2),
+ http://www.ietf.org/rfc/rfc4340.txt
+ To compile this CCID as a module, choose M here: the module will be
+ called dccp_ccid3.
+
If in doubt, say M.
config IP_DCCP_TFRC_LIB
depends on IP_DCCP_CCID3
def_tristate IP_DCCP_CCID3
+config IP_DCCP_CCID3_DEBUG
+ bool "CCID3 debugging messages"
+ depends on IP_DCCP_CCID3
+ ---help---
+ Enable CCID3-specific debugging messages.
+
+ When compiling CCID3 as a module, this debugging output can
+ additionally be toggled by setting the ccid3_debug module
+ parameter to 0 or 1.
+
+ 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.
+
+
endmenu
diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c
index 2efb505aeb3..fd38b05d6f7 100644
--- a/net/dccp/ccids/ccid2.c
+++ b/net/dccp/ccids/ccid2.c
@@ -23,7 +23,7 @@
*/
/*
- * This implementation should follow: draft-ietf-dccp-ccid2-10.txt
+ * This implementation should follow RFC 4341
*
* BUGS:
* - sequence number wrapping
@@ -33,18 +33,11 @@
#include "../dccp.h"
#include "ccid2.h"
-static int ccid2_debug;
#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
-#define ccid2_pr_debug(format, a...) \
- do { if (ccid2_debug) \
- printk(KERN_DEBUG "%s: " format, __FUNCTION__, ##a); \
- } while (0)
-#else
-#define ccid2_pr_debug(format, a...)
-#endif
+static int ccid2_debug;
+#define ccid2_pr_debug(format, a...) DCCP_PR_DEBUG(ccid2_debug, format, ##a)
-#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
static void ccid2_hc_tx_check_sanity(const struct ccid2_hc_tx_sock *hctx)
{
int len = 0;
@@ -86,7 +79,8 @@ static void ccid2_hc_tx_check_sanity(const struct ccid2_hc_tx_sock *hctx)
BUG_ON(len != hctx->ccid2hctx_seqbufc * CCID2_SEQBUF_LEN);
}
#else
-#define ccid2_hc_tx_check_sanity(hctx) do {} while (0)
+#define ccid2_pr_debug(format, a...)
+#define ccid2_hc_tx_check_sanity(hctx)
#endif
static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hctx, int num,
@@ -131,8 +125,7 @@ static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hctx, int num,
return 0;
}
-static int ccid2_hc_tx_send_packet(struct sock *sk,
- struct sk_buff *skb, int len)
+static int ccid2_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
{
struct ccid2_hc_tx_sock *hctx;
@@ -274,7 +267,7 @@ static void ccid2_start_rto_timer(struct sock *sk)
jiffies + hctx->ccid2hctx_rto);
}
-static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, int len)
+static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len)
{
struct dccp_sock *dp = dccp_sk(sk);
struct ccid2_hc_tx_sock *hctx = ccid2_hc_tx_sk(sk);
@@ -352,14 +345,14 @@ static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, int len)
#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
ccid2_pr_debug("pipe=%d\n", hctx->ccid2hctx_pipe);
- ccid2_pr_debug("Sent: seq=%llu\n", seq);
+ ccid2_pr_debug("Sent: seq=%llu\n", (unsigned long long)seq);
do {
struct ccid2_seq *seqp = hctx->ccid2hctx_seqt;
while (seqp != hctx->ccid2hctx_seqh) {
ccid2_pr_debug("out seq=%llu acked=%d time=%lu\n",
- seqp->ccid2s_seq, seqp->ccid2s_acked,
- seqp->ccid2s_sent);
+ (unsigned long long)seqp->ccid2s_seq,
+ seqp->ccid2s_acked, seqp->ccid2s_sent);
seqp = seqp->ccid2s_next;
}
} while (0);
@@ -426,7 +419,7 @@ static int ccid2_ackvector(struct sock *sk, struct sk_buff *skb, int offset,
return -1;
out_invalid_option:
- BUG_ON(1); /* should never happen... options were previously parsed ! */
+ DCCP_BUG("Invalid option - this should not happen (previous parsing)!");
return -1;
}
@@ -480,7 +473,8 @@ static inline void ccid2_new_ack(struct sock *sk,
/* first measurement */
if (hctx->ccid2hctx_srtt == -1) {
ccid2_pr_debug("R: %lu Time=%lu seq=%llu\n",
- r, jiffies, seqp->ccid2s_seq);
+ r, jiffies,
+ (unsigned long long)seqp->ccid2s_seq);
ccid2_change_srtt(hctx, r);
hctx->ccid2hctx_rttvar = r >> 1;
} else {
@@ -524,8 +518,8 @@ static inline void ccid2_new_ack(struct sock *sk,
hctx->ccid2hctx_lastrtt = jiffies;
ccid2_pr_debug("srtt: %ld rttvar: %ld rto: %ld (HZ=%d) R=%lu\n",
- hctx->ccid2hctx_srtt, hctx->ccid2hctx_rttvar,
- hctx->ccid2hctx_rto, HZ, r);
+ hctx->ccid2hctx_srtt, hctx->ccid2hctx_rttvar,
+ hctx->ccid2hctx_rto, HZ, r);
hctx->ccid2hctx_sent = 0;
}
@@ -618,7 +612,17 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
}
ackno = DCCP_SKB_CB(skb)->dccpd_ack_seq;
- seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
+ if (after48(ackno, hctx->ccid2hctx_high_ack))
+ hctx->ccid2hctx_high_ack = ackno;
+
+ seqp = hctx->ccid2hctx_seqt;
+ while (before48(seqp->ccid2s_seq, ackno)) {
+ seqp = seqp->ccid2s_next;
+ if (seqp == hctx->ccid2hctx_seqh) {
+ seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
+ break;
+ }
+ }
/* If in slow-start, cwnd can increase at most Ack Ratio / 2 packets for
* this single ack. I round up.
@@ -636,8 +640,9 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
u64 ackno_end_rl;
dccp_set_seqno(&ackno_end_rl, ackno - rl);
- ccid2_pr_debug("ackvec start:%llu end:%llu\n", ackno,
- ackno_end_rl);
+ ccid2_pr_debug("ackvec start:%llu end:%llu\n",
+ (unsigned long long)ackno,
+ (unsigned long long)ackno_end_rl);
/* if the seqno we are analyzing is larger than the
* current ackno, then move towards the tail of our
* seqnos.
@@ -662,9 +667,9 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
/* new packet received or marked */
if (state != DCCP_ACKVEC_STATE_NOT_RECEIVED &&
!seqp->ccid2s_acked) {
- if (state ==
+ if (state ==
DCCP_ACKVEC_STATE_ECN_MARKED) {
- ccid2_congestion_event(hctx,
+ ccid2_congestion_event(hctx,
seqp);
} else
ccid2_new_ack(sk, seqp,
@@ -672,7 +677,7 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
seqp->ccid2s_acked = 1;
ccid2_pr_debug("Got ack for %llu\n",
- seqp->ccid2s_seq);
+ (unsigned long long)seqp->ccid2s_seq);
ccid2_hc_tx_dec_pipe(sk);
}
if (seqp == hctx->ccid2hctx_seqt) {
@@ -695,7 +700,14 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
/* The state about what is acked should be correct now
* Check for NUMDUPACK
*/
- seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
+ seqp = hctx->ccid2hctx_seqt;
+ while (before48(seqp->ccid2s_seq, hctx->ccid2hctx_high_ack)) {
+ seqp = seqp->ccid2s_next;
+ if (seqp == hctx->ccid2hctx_seqh) {
+ seqp = hctx->ccid2hctx_seqh->ccid2s_prev;
+ break;
+ }
+ }
done = 0;
while (1) {
if (seqp->ccid2s_acked) {
@@ -718,7 +730,7 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
while (1) {
if (!seqp->ccid2s_acked) {
ccid2_pr_debug("Packet lost: %llu\n",
- seqp->ccid2s_seq);
+ (unsigned long long)seqp->ccid2s_seq);
/* XXX need to traverse from tail -> head in
* order to detect multiple congestion events in
* one ack vector.
@@ -769,6 +781,7 @@ static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk)
hctx->ccid2hctx_lastrtt = 0;
hctx->ccid2hctx_rpdupack = -1;
hctx->ccid2hctx_last_cong = jiffies;
+ hctx->ccid2hctx_high_ack = 0;
hctx->ccid2hctx_rtotimer.function = &ccid2_hc_tx_rto_expire;
hctx->ccid2hctx_rtotimer.data = (unsigned long)sk;
@@ -821,8 +834,10 @@ static struct ccid_operations ccid2 = {
.ccid_hc_rx_packet_recv = ccid2_hc_rx_packet_recv,
};
+#ifdef CONFIG_IP_DCCP_CCID2_DEBUG
module_param(ccid2_debug, int, 0444);
MODULE_PARM_DESC(ccid2_debug, "Enable debug messages");
+#endif
static __init int ccid2_module_init(void)
{
diff --git a/net/dccp/ccids/ccid2.h b/net/dccp/ccids/ccid2.h
index 5b2ef4acb30..ebd79499c85 100644
--- a/net/dccp/ccids/ccid2.h
+++ b/net/dccp/ccids/ccid2.h
@@ -35,7 +35,7 @@ struct ccid2_seq {
struct ccid2_seq *ccid2s_next;
};
-#define CCID2_SEQBUF_LEN 256
+#define CCID2_SEQBUF_LEN 1024
#define CCID2_SEQBUF_MAX 128
/** struct ccid2_hc_tx_sock - CCID2 TX half connection
@@ -72,6 +72,7 @@ struct ccid2_hc_tx_sock {
int ccid2hctx_rpdupack;
int ccid2hctx_sendwait;
unsigned long ccid2hctx_last_cong;
+ u64 ccid2hctx_high_ack;
};
struct ccid2_hc_rx_sock {
diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c
index 67d2dc0e7c6..40402c59506 100644
--- a/net/dccp/ccids/ccid3.c
+++ b/net/dccp/ccids/ccid3.c
@@ -41,32 +41,9 @@
#include "lib/tfrc.h"
#include "ccid3.h"
-/*
- * Reason for maths here is to avoid 32 bit overflow when a is big.
- * With this we get close to the limit.
- */
-static u32 usecs_div(const u32 a, const u32 b)
-{
- const u32 div = a < (UINT_MAX / (USEC_PER_SEC / 10)) ? 10 :
- a < (UINT_MAX / (USEC_PER_SEC / 50)) ? 50 :
- a < (UINT_MAX / (USEC_PER_SEC / 100)) ? 100 :
- a < (UINT_MAX / (USEC_PER_SEC / 500)) ? 500 :
- a < (UINT_MAX / (USEC_PER_SEC / 1000)) ? 1000 :
- a < (UINT_MAX / (USEC_PER_SEC / 5000)) ? 5000 :
- a < (UINT_MAX / (USEC_PER_SEC / 10000)) ? 10000 :
- a < (UINT_MAX / (USEC_PER_SEC / 50000)) ? 50000 :
- 100000;
- const u32 tmp = a * (USEC_PER_SEC / div);
- return (b >= 2 * div) ? tmp / (b / div) : tmp;
-}
-
+#ifdef CONFIG_IP_DCCP_CCID3_DEBUG
static int ccid3_debug;
-
-#ifdef CCID3_DEBUG
-#define ccid3_pr_debug(format, a...) \
- do { if (ccid3_debug) \
- printk(KERN_DEBUG "%s: " format, __FUNCTION__, ##a); \
- } while (0)
+#define ccid3_pr_debug(format, a...) DCCP_PR_DEBUG(ccid3_debug, format, ##a)
#else
#define ccid3_pr_debug(format, a...)
#endif
@@ -75,15 +52,7 @@ static struct dccp_tx_hist *ccid3_tx_hist;
static struct dccp_rx_hist *ccid3_rx_hist;
static struct dccp_li_hist *ccid3_li_hist;
-/* TFRC sender states */
-enum ccid3_hc_tx_states {
- TFRC_SSTATE_NO_SENT = 1,
- TFRC_SSTATE_NO_FBACK,
- TFRC_SSTATE_FBACK,
- TFRC_SSTATE_TERM,
-};
-
-#ifdef CCID3_DEBUG
+#ifdef CONFIG_IP_DCCP_CCID3_DEBUG
static const char *ccid3_tx_state_name(enum ccid3_hc_tx_states state)
{
static char *ccid3_state_names[] = {
@@ -110,325 +79,319 @@ static void ccid3_hc_tx_set_state(struct sock *sk,
hctx->ccid3hctx_state = state;
}
-/* Calculate new t_ipi (inter packet interval) by t_ipi = s / X_inst */
-static inline void ccid3_calc_new_t_ipi(struct ccid3_hc_tx_sock *hctx)
+/*
+ * Recalculate scheduled nominal send time t_nom, inter-packet interval
+ * t_ipi, and delta value. Should be called after each change to X.
+ */
+static inline void ccid3_update_send_time(struct ccid3_hc_tx_sock *hctx)
{
- /*
- * If no feedback spec says t_ipi is 1 second (set elsewhere and then
- * doubles after every no feedback timer (separate function)
- */
- if (hctx->ccid3hctx_state != TFRC_SSTATE_NO_FBACK)
- hctx->ccid3hctx_t_ipi = usecs_div(hctx->ccid3hctx_s,
- hctx->ccid3hctx_x);
-}
+ timeval_sub_usecs(&hctx->ccid3hctx_t_nom, hctx->ccid3hctx_t_ipi);
-/* Calculate new delta by delta = min(t_ipi / 2, t_gran / 2) */
-static inline void ccid3_calc_new_delta(struct ccid3_hc_tx_sock *hctx)
-{
+ /* Calculate new t_ipi = s / X_inst (X_inst is in 64 * bytes/second) */
+ hctx->ccid3hctx_t_ipi = scaled_div(hctx->ccid3hctx_s,
+ hctx->ccid3hctx_x >> 6);
+
+ /* Update nominal send time with regard to the new t_ipi */
+ timeval_add_usecs(&hctx->ccid3hctx_t_nom, hctx->ccid3hctx_t_ipi);
+
+ /* Calculate new delta by delta = min(t_ipi / 2, t_gran / 2) */
hctx->ccid3hctx_delta = min_t(u32, hctx->ccid3hctx_t_ipi / 2,
TFRC_OPSYS_HALF_TIME_GRAN);
}
-
/*
* Update X by
* If (p > 0)
- * x_calc = calcX(s, R, p);
+ * X_calc = calcX(s, R, p);
* X = max(min(X_calc, 2 * X_recv), s / t_mbi);
* Else
* If (now - tld >= R)
* X = max(min(2 * X, 2 * X_recv), s / R);
* tld = now;
- */
-static void ccid3_hc_tx_update_x(struct sock *sk)
+ *
+ * Note: X and X_recv are both stored in units of 64 * bytes/second, to support
+ * fine-grained resolution of sending rates. This requires scaling by 2^6
+ * throughout the code. Only X_calc is unscaled (in bytes/second).
+ *
+ * If X has changed, we also update the scheduled send time t_now,
+ * the inter-packet interval t_ipi, and the delta value.
+ */
+static void ccid3_hc_tx_update_x(struct sock *sk, struct timeval *now)
+
{
struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk);
+ const __u64 old_x = hctx->ccid3hctx_x;
- /* To avoid large error in calcX */
- if (hctx->ccid3hctx_p >= TFRC_SMALLEST_P) {
- hctx->ccid3hctx_x_calc = tfrc_calc_x(hctx->ccid3hctx_s,
- hctx->ccid3hctx_rtt,
- hctx->ccid3hctx_p);
- hctx->ccid3hctx_x = max_t(u32, min_t(u32, hctx->ccid3hctx_x_calc,
- 2 * hctx->ccid3hctx_x_recv),
- (hctx->ccid3hctx_s /
- TFRC_MAX_BACK_OFF_TIME));
- } else {
- struct timeval now;
+ if (hctx->ccid3hctx_p > 0) {
- dccp_timestamp(sk, &now);
- if (timeval_delta(&now, &hctx->ccid3hctx_t_ld) >=
- hctx->ccid3hctx_rtt) {
- hctx->ccid3hctx_x = max_t(u32, min_t(u32, hctx->ccid3hctx_x_recv,
- hctx->ccid3hctx_x) * 2,
- usecs_div(hctx->ccid3hctx_s,
- hctx->ccid3hctx_rtt));
- hctx->ccid3hctx_t_ld = now;
- }
+ hctx->ccid3hctx_x = min(((__u64)hctx->ccid3hctx_x_calc) << 6,
+ hctx->ccid3hctx_x_recv * 2);
+ hctx->ccid3hctx_x = max(hctx->ccid3hctx_x,
+ (((__u64)hctx->ccid3hctx_s) << 6) /
+ TFRC_T_MBI);
+
+ } else if (timeval_delta(now, &hctx->ccid3hctx_t_ld) -
+ (suseconds_t)hctx->ccid3hctx_rtt >= 0) {
+
+ hctx->ccid3hctx_x =
+ max(2 * min(hctx->ccid3hctx_x, hctx->ccid3hctx_x_recv),
+ scaled_div(((__u64)hctx->ccid3hctx_s) << 6,
+ hctx->ccid3hctx_rtt));
+ hctx->ccid3hctx_t_ld = *now;
+ }
+
+ if (hctx->ccid3hctx_x != old_x)
+ ccid3_update_send_time(hctx);
+}
+
+/*
+ * Track the mean packet size `s' (cf. RFC 4342, 5.3 and RFC 3448, 4.1)
+ * @len: DCCP packet payload size in bytes
+ */
+static inline void ccid3_hc_tx_update_s(struct ccid3_hc_tx_sock *hctx, int len)
+{
+ if (unlikely(len == 0))
+ ccid3_pr_debug("Packet payload length is 0 - not updating\n");
+ else
+ hctx->ccid3hctx_s = hctx->ccid3hctx_s == 0 ? len :
+ (9 * hctx->ccid3hctx_s + len) / 10;
+ /*
+ * Note: We could do a potential optimisation here - when `s' changes,
+ * recalculate sending rate and consequently t_ipi, t_delta, and
+ * t_now. This is however non-standard, and the benefits are not
+ * clear, so it is currently left out.
+ */
+}
+
+/*
+ * Update Window Counter using the algorithm from [RFC 4342, 8.1].
+ * The algorithm is not applicable if RTT < 4 microseconds.
+ */
+static inline void ccid3_hc_tx_update_win_count(struct ccid3_hc_tx_sock *hctx,
+ struct timeval *now)
+{
+ suseconds_t delta;
+ u32 quarter_rtts;
+
+ if (unlikely(hctx->ccid3hctx_rtt < 4)) /* avoid divide-by-zero */
+ return;
+
+ delta = timeval_delta(now, &hctx->ccid3hctx_t_last_win_count);
+ DCCP_BUG_ON(delta < 0);
+
+ quarter_rtts = (u32)delta / (hctx->ccid3hctx_rtt / 4);
+
+ if (quarter_rtts > 0) {
+ hctx->ccid3hctx_t_last_win_count = *now;
+ hctx->ccid3hctx_last_win_count += min_t(u32, quarter_rtts, 5);
+ hctx->ccid3hctx_last_win_count &= 0xF; /* mod 16 */
+
+ ccid3_pr_debug("now at %#X\n", hctx->ccid3hctx_last_win_count);
}
}
static void ccid3_hc_tx_no_feedback_timer(unsigned long data)
{
struct sock *sk = (struct sock *)data;
- unsigned long next_tmout = 0;
struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk);
+ unsigned long t_nfb = USEC_PER_SEC / 5;
bh_lock_sock(sk);
if (sock_owned_by_user(sk)) {
/* Try again later. */
/* XXX: set some sensible MIB */
- sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
- jiffies + HZ / 5);
- goto out;
+ goto restart_timer;
}
- ccid3_pr_debug("%s, sk=%p, state=%s\n", dccp_role(sk), sk,
+ ccid3_pr_debug("%s(%p, state=%s) - entry \n", dccp_role(sk), sk,
ccid3_tx_state_name(hctx->ccid3hctx_state));
-
+
switch (hctx->ccid3hctx_state) {
- case TFRC_SSTATE_TERM:
- goto out;
case TFRC_SSTATE_NO_FBACK:
- /* Halve send rate */
- hctx->ccid3hctx_x /= 2;
- if (hctx->ccid3hctx_x < (hctx->ccid3hctx_s /
- TFRC_MAX_BACK_OFF_TIME))
- hctx->ccid3hctx_x = (hctx->ccid3hctx_s /
- TFRC_MAX_BACK_OFF_TIME);
-
- ccid3_pr_debug("%s, sk=%p, state=%s, updated tx rate to %d "
- "bytes/s\n",
- dccp_role(sk), sk,
+ /* RFC 3448, 4.4: Halve send rate directly */
+ hctx->ccid3hctx_x = max(hctx->ccid3hctx_x / 2,
+ (((__u64)hctx->ccid3hctx_s) << 6) /
+ TFRC_T_MBI);
+
+ ccid3_pr_debug("%s(%p, state=%s), updated tx rate to %u "
+ "bytes/s\n", dccp_role(sk), sk,
ccid3_tx_state_name(hctx->ccid3hctx_state),
- hctx->ccid3hctx_x);
- next_tmout = max_t(u32, 2 * usecs_div(hctx->ccid3hctx_s,
- hctx->ccid3hctx_x),
- TFRC_INITIAL_TIMEOUT);
- /*
- * FIXME - not sure above calculation is correct. See section
- * 5 of CCID3 11 should adjust tx_t_ipi and double that to
- * achieve it really
- */
+ (unsigned)(hctx->ccid3hctx_x >> 6));
+ /* The value of R is still undefined and so we can not recompute
+ * the timout value. Keep initial value as per [RFC 4342, 5]. */
+ t_nfb = TFRC_INITIAL_TIMEOUT;
+ ccid3_update_send_time(hctx);
break;
case TFRC_SSTATE_FBACK:
/*
* Check if IDLE since last timeout and recv rate is less than
- * 4 packets per RTT
+ * 4 packets (in units of 64*bytes/sec) per RTT
*/
if (!hctx->ccid3hctx_idle ||
- (hctx->ccid3hctx_x_recv >=
- 4 * usecs_div(hctx->ccid3hctx_s, hctx->ccid3hctx_rtt))) {
- ccid3_pr_debug("%s, sk=%p, state=%s, not idle\n",
+ (hctx->ccid3hctx_x_recv >= 4 *
+ scaled_div(((__u64)hctx->ccid3hctx_s) << 6,
+ hctx->ccid3hctx_rtt))) {
+ struct timeval now;
+
+ ccid3_pr_debug("%s(%p, state=%s), not idle\n",
dccp_role(sk), sk,
- ccid3_tx_state_name(hctx->ccid3hctx_state));
- /* Halve sending rate */
+ ccid3_tx_state_name(hctx->ccid3hctx_state));
- /* If (X_calc > 2 * X_recv)
+ /*
+ * Modify the cached value of X_recv [RFC 3448, 4.4]
+ *
+ * If (p == 0 || X_calc > 2 * X_recv)
* X_recv = max(X_recv / 2, s / (2 * t_mbi));
* Else
* X_recv = X_calc / 4;
+ *
+ * Note that X_recv is scaled by 2^6 while X_calc is not
*/
- BUG_ON(hctx->ccid3hctx_p >= TFRC_SMALLEST_P &&
- hctx->ccid3hctx_x_calc == 0);
-
- /* check also if p is zero -> x_calc is infinity? */
- if (hctx->ccid3hctx_p < TFRC_SMALLEST_P ||
- hctx->ccid3hctx_x_calc > 2 * hctx->ccid3hctx_x_recv)
- hctx->ccid3hctx_x_recv = max_t(u32, hctx->ccid3hctx_x_recv / 2,
- hctx->ccid3hctx_s / (2 * TFRC_MAX_BACK_OFF_TIME));
- else
- hctx->ccid3hctx_x_recv = hctx->ccid3hctx_x_calc / 4;
-
- /* Update sending rate */
- ccid3_hc_tx_update_x(sk);
+ BUG_ON(hctx->ccid3hctx_p && !hctx->ccid3hctx_x_calc);
+
+ if (hctx->ccid3hctx_p == 0 ||
+ (hctx->ccid3hctx_x_calc >
+ (hctx->ccid3hctx_x_recv >> 5))) {
+
+ hctx->ccid3hctx_x_recv =
+ max(hctx->ccid3hctx_x_recv / 2,
+ (((__u64)hctx->ccid3hctx_s) << 6) /
+ (2 * TFRC_T_MBI));
+
+ if (hctx->ccid3hctx_p == 0)
+ dccp_timestamp(sk, &now);
+ } else {
+ hctx->ccid3hctx_x_recv = hctx->ccid3hctx_x_calc;
+ hctx->ccid3hctx_x_recv <<= 4;
+ }
+ /* Now recalculate X [RFC 3448, 4.3, step (4)] */
+ ccid3_hc_tx_update_x(sk, &now);
}
/*
* Schedule no feedback timer to expire in
- * max(4 * R, 2 * s / X)
+ * max(t_RTO, 2 * s/X) = max(t_RTO, 2 * t_ipi)
+ * See comments in packet_recv() regarding the value of t_RTO.
*/
- next_tmout = max_t(u32, hctx->ccid3hctx_t_rto,
- 2 * usecs_div(hctx->ccid3hctx_s,
- hctx->ccid3hctx_x));
+ t_nfb = max(hctx->ccid3hctx_t_rto, 2 * hctx->ccid3hctx_t_ipi);
break;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hctx->ccid3hctx_state);
- dump_stack();
+ case TFRC_SSTATE_NO_SENT:
+ DCCP_BUG("%s(%p) - Illegal state NO_SENT", dccp_role(sk), sk);
+ /* fall through */
+ case TFRC_SSTATE_TERM:
goto out;
}
- sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
- jiffies + max_t(u32, 1, usecs_to_jiffies(next_tmout)));
hctx->ccid3hctx_idle = 1;
+
+restart_timer:
+ sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
+ jiffies + usecs_to_jiffies(t_nfb));
out:
bh_unlock_sock(sk);
sock_put(sk);
}
-static int ccid3_hc_tx_send_packet(struct sock *sk,
- struct sk_buff *skb, int len)
+/*
+ * returns
+ * > 0: delay (in msecs) that should pass before actually sending
+ * = 0: can send immediately
+ * < 0: error condition; do not send packet
+ */
+static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
{
struct dccp_sock *dp = dccp_sk(sk);
struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk);
- struct dccp_tx_hist_entry *new_packet;
struct timeval now;
- long delay;
- int rc = -ENOTCONN;
+ suseconds_t delay;
- BUG_ON(hctx == NULL || hctx->ccid3hctx_state == TFRC_SSTATE_TERM);
+ BUG_ON(hctx == NULL);
- /* Check if pure ACK or Terminating*/
/*
- * XXX: We only call this function for DATA and DATAACK, on, these
- * packets can have zero length, but why the comment about "pure ACK"?
+ * This function is called only for Data and DataAck packets. Sending
+ * zero-sized Data(Ack)s is theoretically possible, but for congestion
+ * control this case is pathological - ignore it.
*/
- if (unlikely(len == 0))
- goto out;
-
- /* See if last packet allocated was not sent */
- new_packet = dccp_tx_hist_head(&hctx->ccid3hctx_hist);
- if (new_packet == NULL || new_packet->dccphtx_sent) {
- new_packet = dccp_tx_hist_entry_new(ccid3_tx_hist,
- SLAB_ATOMIC);
-
- rc = -ENOBUFS;
- if (unlikely(new_packet == NULL)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, not enough "
- "mem to add to history, send refused\n",
- __FUNCTION__, dccp_role(sk), sk);
- goto out;
- }
-
- dccp_tx_hist_add_entry(&hctx->ccid3hctx_hist, new_packet);
- }
+ if (unlikely(skb->len == 0))
+ return -EBADMSG;
dccp_timestamp(sk, &now);
switch (hctx->ccid3hctx_state) {
case TFRC_SSTATE_NO_SENT:
sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
- jiffies + usecs_to_jiffies(TFRC_INITIAL_TIMEOUT));
+ (jiffies +
+ usecs_to_jiffies(TFRC_INITIAL_TIMEOUT)));
hctx->ccid3hctx_last_win_count = 0;
hctx->ccid3hctx_t_last_win_count = now;
ccid3_hc_tx_set_state(sk, TFRC_SSTATE_NO_FBACK);
- hctx->ccid3hctx_t_ipi = TFRC_INITIAL_IPI;
- /* Set nominal send time for initial packet */
+ /* Set initial sending rate X/s to 1pps (X is scaled by 2^6) */
+ ccid3_hc_tx_update_s(hctx, skb->len);
+ hctx->ccid3hctx_x = hctx->ccid3hctx_s;
+ hctx->ccid3hctx_x <<= 6;
+
+ /* First timeout, according to [RFC 3448, 4.2], is 1 second */
+ hctx->ccid3hctx_t_ipi = USEC_PER_SEC;
+ /* Initial delta: minimum of 0.5 sec and t_gran/2 */
+ hctx->ccid3hctx_delta = TFRC_OPSYS_HALF_TIME_GRAN;
+
+ /* Set t_0 for initial packet */
hctx->ccid3hctx_t_nom = now;
- timeval_add_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- ccid3_calc_new_delta(hctx);
- rc = 0;
break;
case TFRC_SSTATE_NO_FBACK:
case TFRC_SSTATE_FBACK:
- delay = (timeval_delta(&now, &hctx->ccid3hctx_t_nom) -
- hctx->ccid3hctx_delta);
- delay /= -1000;
- /* divide by -1000 is to convert to ms and get sign right */
- rc = delay > 0 ? delay : 0;
- break;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hctx->ccid3hctx_state);
- dump_stack();
- rc = -EINVAL;
+ delay = timeval_delta(&hctx->ccid3hctx_t_nom, &now);
+ /*
+ * Scheduling of packet transmissions [RFC 3448, 4.6]
+ *
+ * if (t_now > t_nom - delta)
+ * // send the packet now
+ * else
+ * // send the packet in (t_nom - t_now) milliseconds.
+ */
+ if (delay - (suseconds_t)hctx->ccid3hctx_delta >= 0)
+ return delay / 1000L;
+
+ ccid3_hc_tx_update_win_count(hctx, &now);
break;
+ case TFRC_SSTATE_TERM:
+ DCCP_BUG("%s(%p) - Illegal state TERM", dccp_role(sk), sk);
+ return -EINVAL;
}
- /* Can we send? if so add options and add to packet history */
- if (rc == 0) {
- dp->dccps_hc_tx_insert_options = 1;
- new_packet->dccphtx_ccval =
- DCCP_SKB_CB(skb)->dccpd_ccval =
- hctx->ccid3hctx_last_win_count;
- timeval_add_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- }
-out:
- return rc;
+ /* prepare to send now (add options etc.) */
+ dp->dccps_hc_tx_insert_options = 1;
+ DCCP_SKB_CB(skb)->dccpd_ccval = hctx->ccid3hctx_last_win_count;
+
+ /* set the nominal send time for the next following packet */
+ timeval_add_usecs(&hctx->ccid3hctx_t_nom, hctx->ccid3hctx_t_ipi);
+
+ return 0;
}
-static void ccid3_hc_tx_packet_sent(struct sock *sk, int more, int len)
+static void ccid3_hc_tx_packet_sent(struct sock *sk, int more,
+ unsigned int len)
{
- const struct dccp_sock *dp = dccp_sk(sk);
struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk);
struct timeval now;
+ struct dccp_tx_hist_entry *packet;
- BUG_ON(hctx == NULL || hctx->ccid3hctx_state == TFRC_SSTATE_TERM);
-
- dccp_timestamp(sk, &now);
-
- /* check if we have sent a data packet */
- if (len > 0) {
- unsigned long quarter_rtt;
- struct dccp_tx_hist_entry *packet;
-
- packet = dccp_tx_hist_head(&hctx->ccid3hctx_hist);
- if (unlikely(packet == NULL)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: packet doesn't "
- "exists in history!\n", __FUNCTION__);
- return;
- }
- if (unlikely(packet->dccphtx_sent)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: no unsent packet in "
- "history!\n", __FUNCTION__);
- return;
- }
- packet->dccphtx_tstamp = now;
- packet->dccphtx_seqno = dp->dccps_gss;
- /*
- * Check if win_count have changed
- * Algorithm in "8.1. Window Counter Valuer" in
- * draft-ietf-dccp-ccid3-11.txt
- */
- quarter_rtt = timeval_delta(&now, &hctx->ccid3hctx_t_last_win_count);
- if (likely(hctx->ccid3hctx_rtt > 8))
- quarter_rtt /= hctx->ccid3hctx_rtt / 4;
-
- if (quarter_rtt > 0) {
- hctx->ccid3hctx_t_last_win_count = now;
- hctx->ccid3hctx_last_win_count = (hctx->ccid3hctx_last_win_count +
- min_t(unsigned long, quarter_rtt, 5)) % 16;
- ccid3_pr_debug("%s, sk=%p, window changed from "
- "%u to %u!\n",
- dccp_role(sk), sk,
- packet->dccphtx_ccval,
- hctx->ccid3hctx_last_win_count);
- }
+ BUG_ON(hctx == NULL);
- hctx->ccid3hctx_idle = 0;
- packet->dccphtx_rtt = hctx->ccid3hctx_rtt;
- packet->dccphtx_sent = 1;
- } else
- ccid3_pr_debug("%s, sk=%p, seqno=%llu NOT inserted!\n",
- dccp_role(sk), sk, dp->dccps_gss);
+ ccid3_hc_tx_update_s(hctx, len);
- switch (hctx->ccid3hctx_state) {
- case TFRC_SSTATE_NO_SENT:
- /* if first wasn't pure ack */
- if (len != 0)
- printk(KERN_CRIT "%s: %s, First packet sent is noted "
- "as a data packet\n",
- __FUNCTION__, dccp_role(sk));
+ packet = dccp_tx_hist_entry_new(ccid3_tx_hist, GFP_ATOMIC);
+ if (unlikely(packet == NULL)) {
+ DCCP_CRIT("packet history - out of memory!");
return;
- case TFRC_SSTATE_NO_FBACK:
- case TFRC_SSTATE_FBACK:
- if (len > 0) {
- timeval_sub_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- ccid3_calc_new_t_ipi(hctx);
- ccid3_calc_new_delta(hctx);
- timeval_add_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- }
- break;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hctx->ccid3hctx_state);
- dump_stack();
- break;
}
+ dccp_tx_hist_add_entry(&hctx->ccid3hctx_hist, packet);
+
+ dccp_timestamp(sk, &now);
+ packet->dccphtx_tstamp = now;
+ packet->dccphtx_seqno = dccp_sk(sk)->dccps_gss;
+ packet->dccphtx_rtt = hctx->ccid3hctx_rtt;
+ packet->dccphtx_sent = 1;
+ hctx->ccid3hctx_idle = 0;
}
static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
@@ -438,13 +401,11 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
struct ccid3_options_received *opt_recv;
struct dccp_tx_hist_entry *packet;
struct timeval now;
- unsigned long next_tmout;
- u32 t_elapsed;
+ unsigned long t_nfb;
u32 pinv;
- u32 x_recv;
- u32 r_sample;
+ suseconds_t r_sample, t_elapsed;
- BUG_ON(hctx == NULL || hctx->ccid3hctx_state == TFRC_SSTATE_TERM);
+ BUG_ON(hctx == NULL);
/* we are only interested in ACKs */
if (!(DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_ACK ||
@@ -453,41 +414,49 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
opt_recv = &hctx->ccid3hctx_options_received;
- t_elapsed = dp->dccps_options_received.dccpor_elapsed_time * 10;
- x_recv = opt_recv->ccid3or_receive_rate;
- pinv = opt_recv->ccid3or_loss_event_rate;
-
switch (hctx->ccid3hctx_state) {
- case TFRC_SSTATE_NO_SENT:
- /* FIXME: what to do here? */
- return;
case TFRC_SSTATE_NO_FBACK:
case TFRC_SSTATE_FBACK:
- /* Calculate new round trip sample by
- * R_sample = (now - t_recvdata) - t_delay */
- /* get t_recvdata from history */
+ /* get packet from history to look up t_recvdata */
packet = dccp_tx_hist_find_entry(&hctx->ccid3hctx_hist,
- DCCP_SKB_CB(skb)->dccpd_ack_seq);
+ DCCP_SKB_CB(skb)->dccpd_ack_seq);
if (unlikely(packet == NULL)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, seqno "
- "%llu(%s) does't exist in history!\n",
- __FUNCTION__, dccp_role(sk), sk,
+ DCCP_WARN("%s(%p), seqno %llu(%s) doesn't exist "
+ "in history!\n", dccp_role(sk), sk,
(unsigned long long)DCCP_SKB_CB(skb)->dccpd_ack_seq,
dccp_packet_name(DCCP_SKB_CB(skb)->dccpd_type));
return;
}
- /* Update RTT */
+ /* Update receive rate in units of 64 * bytes/second */
+ hctx->ccid3hctx_x_recv = opt_recv->ccid3or_receive_rate;
+ hctx->ccid3hctx_x_recv <<= 6;
+
+ /* Update loss event rate */
+ pinv = opt_recv->ccid3or_loss_event_rate;
+ if (pinv == ~0U || pinv == 0) /* see RFC 4342, 8.5 */
+ hctx->ccid3hctx_p = 0;
+ else /* can not exceed 100% */
+ hctx->ccid3hctx_p = 1000000 / pinv;
+
dccp_timestamp(sk, &now);
- r_sample = timeval_delta(&now, &packet->dccphtx_tstamp);
+
+ /*
+ * Calculate new round trip sample as per [RFC 3448, 4.3] by
+ * R_sample = (now - t_recvdata) - t_elapsed
+ */
+ r_sample = timeval_delta(&now, &packet->dccphtx_tstamp);
+ t_elapsed = dp->dccps_options_received.dccpor_elapsed_time * 10;
+
+ DCCP_BUG_ON(r_sample < 0);
if (unlikely(r_sample <= t_elapsed))
- LIMIT_NETDEBUG(KERN_WARNING "%s: r_sample=%uus, "
- "t_elapsed=%uus\n",
- __FUNCTION__, r_sample, t_elapsed);
+ DCCP_WARN("WARNING: r_sample=%dus <= t_elapsed=%dus\n",
+ (int)r_sample, (int)t_elapsed);
else
r_sample -= t_elapsed;
+ CCID3_RTT_SANITY_CHECK(r_sample);
- /* Update RTT estimate by
+ /* Update RTT estimate by
* If (No feedback recv)
* R = R_sample;
* Else
@@ -496,97 +465,96 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
* q is a constant, RFC 3448 recomments 0.9
*/
if (hctx->ccid3hctx_state == TFRC_SSTATE_NO_FBACK) {
- ccid3_hc_tx_set_state(sk, TFRC_SSTATE_FBACK);
- hctx->ccid3hctx_rtt = r_sample;
- } else
- hctx->ccid3hctx_rtt = (hctx->ccid3hctx_rtt * 9) / 10 +
- r_sample / 10;
-
- ccid3_pr_debug("%s, sk=%p, New RTT estimate=%uus, "
- "r_sample=%us\n", dccp_role(sk), sk,
- hctx->ccid3hctx_rtt, r_sample);
-
- /* Update timeout interval */
- hctx->ccid3hctx_t_rto = max_t(u32, 4 * hctx->ccid3hctx_rtt,
- USEC_PER_SEC);
+ /*
+ * Larger Initial Windows [RFC 4342, sec. 5]
+ * We deviate in that we use `s' instead of `MSS'.
+ */
+ __u64 w_init = min(4 * hctx->ccid3hctx_s,
+ max(2 * hctx->ccid3hctx_s, 4380));
+ hctx->ccid3hctx_rtt = r_sample;
+ hctx->ccid3hctx_x = scaled_div(w_init << 6, r_sample);
+ hctx->ccid3hctx_t_ld = now;
- /* Update receive rate */
- hctx->ccid3hctx_x_recv = x_recv;/* X_recv in bytes per sec */
+ ccid3_update_send_time(hctx);
- /* Update loss event rate */
- if (pinv == ~0 || pinv == 0)
- hctx->ccid3hctx_p = 0;
- else {
- hctx->ccid3hctx_p = 1000000 / pinv;
+ ccid3_pr_debug("%s(%p), s=%u, w_init=%llu, "
+ "R_sample=%dus, X=%u\n", dccp_role(sk),
+ sk, hctx->ccid3hctx_s, w_init,
+ (int)r_sample,
+ (unsigned)(hctx->ccid3hctx_x >> 6));
- if (hctx->ccid3hctx_p < TFRC_SMALLEST_P) {
- hctx->ccid3hctx_p = TFRC_SMALLEST_P;
- ccid3_pr_debug("%s, sk=%p, Smallest p used!\n",
- dccp_role(sk), sk);
- }
+ ccid3_hc_tx_set_state(sk, TFRC_SSTATE_FBACK);
+ } else {
+ hctx->ccid3hctx_rtt = (9 * hctx->ccid3hctx_rtt +
+ (u32)r_sample) / 10;
+
+ /* Update sending rate (step 4 of [RFC 3448, 4.3]) */
+ if (hctx->ccid3hctx_p > 0)
+ hctx->ccid3hctx_x_calc =
+ tfrc_calc_x(hctx->ccid3hctx_s,
+ hctx->ccid3hctx_rtt,
+ hctx->ccid3hctx_p);
+ ccid3_hc_tx_update_x(sk, &now);
+
+ ccid3_pr_debug("%s(%p), RTT=%uus (sample=%dus), s=%u, "
+ "p=%u, X_calc=%u, X_recv=%u, X=%u\n",
+ dccp_role(sk),
+ sk, hctx->ccid3hctx_rtt, (int)r_sample,
+ hctx->ccid3hctx_s, hctx->ccid3hctx_p,
+ hctx->ccid3hctx_x_calc,
+ (unsigned)(hctx->ccid3hctx_x_recv >> 6),
+ (unsigned)(hctx->ccid3hctx_x >> 6));
}
/* unschedule no feedback timer */
sk_stop_timer(sk, &hctx->ccid3hctx_no_feedback_timer);
- /* Update sending rate */
- ccid3_hc_tx_update_x(sk);
-
- /* Update next send time */
- timeval_sub_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- ccid3_calc_new_t_ipi(hctx);
- timeval_add_usecs(&hctx->ccid3hctx_t_nom,
- hctx->ccid3hctx_t_ipi);
- ccid3_calc_new_delta(hctx);
-
/* remove all packets older than the one acked from history */
dccp_tx_hist_purge_older(ccid3_tx_hist,
&hctx->ccid3hctx_hist, packet);
/*
- * As we have calculated new ipi, delta, t_nom it is possible that
- * we now can send a packet, so wake up dccp_wait_for_ccids.
+ * As we have calculated new ipi, delta, t_nom it is possible
+ * that we now can send a packet, so wake up dccp_wait_for_ccid
*/
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.
+ */
+ hctx->ccid3hctx_t_rto = max_t(u32, 4 * hctx->ccid3hctx_rtt,
+ CONFIG_IP_DCCP_CCID3_RTO *
+ (USEC_PER_SEC/1000));
+ /*
* Schedule no feedback timer to expire in
- * max(4 * R, 2 * s / X)
+ * max(t_RTO, 2 * s/X) = max(t_RTO, 2 * t_ipi)
*/
- next_tmout = max(hctx->ccid3hctx_t_rto,
- 2 * usecs_div(hctx->ccid3hctx_s,
- hctx->ccid3hctx_x));
-
- ccid3_pr_debug("%s, sk=%p, Scheduled no feedback timer to "
+ t_nfb = max(hctx->ccid3hctx_t_rto, 2 * hctx->ccid3hctx_t_ipi);
+
+ ccid3_pr_debug("%s(%p), Scheduled no feedback timer to "
"expire in %lu jiffies (%luus)\n",
- dccp_role(sk), sk,
- usecs_to_jiffies(next_tmout), next_tmout);
+ dccp_role(sk),
+ sk, usecs_to_jiffies(t_nfb), t_nfb);
- sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
- jiffies + max_t(u32, 1, usecs_to_jiffies(next_tmout)));
+ sk_reset_timer(sk, &hctx->ccid3hctx_no_feedback_timer,
+ jiffies + usecs_to_jiffies(t_nfb));
/* set idle flag */
- hctx->ccid3hctx_idle = 1;
+ hctx->ccid3hctx_idle = 1;
break;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hctx->ccid3hctx_state);
- dump_stack();
+ case TFRC_SSTATE_NO_SENT:
+ /*
+ * XXX when implementing bidirectional rx/tx check this again
+ */
+ DCCP_WARN("Illegal ACK received - no packet sent\n");
+ /* fall through */
+ case TFRC_SSTATE_TERM: /* ignore feedback when closing */
break;
}
}
-static int ccid3_hc_tx_insert_options(struct sock *sk, struct sk_buff *skb)
-{
- const struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk);
-
- BUG_ON(hctx == NULL);
-
- if (sk->sk_state == DCCP_OPEN || sk->sk_state == DCCP_PARTOPEN)
- DCCP_SKB_CB(skb)->dccpd_ccval = hctx->ccid3hctx_last_win_count;
- return 0;
-}
-
static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option,
unsigned char len, u16 idx,
unsigned char *value)
@@ -611,13 +579,14 @@ static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option,
switch (option) {
case TFRC_OPT_LOSS_EVENT_RATE:
if (unlikely(len != 4)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, invalid "
- "len for TFRC_OPT_LOSS_EVENT_RATE\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), invalid len %d "
+ "for TFRC_OPT_LOSS_EVENT_RATE\n",
+ dccp_role(sk), sk, len);
rc = -EINVAL;
} else {
- opt_recv->ccid3or_loss_event_rate = ntohl(*(__be32 *)value);
- ccid3_pr_debug("%s, sk=%p, LOSS_EVENT_RATE=%u\n",
+ opt_recv->ccid3or_loss_event_rate =
+ ntohl(*(__be32 *)value);
+ ccid3_pr_debug("%s(%p), LOSS_EVENT_RATE=%u\n",
dccp_role(sk), sk,
opt_recv->ccid3or_loss_event_rate);
}
@@ -625,20 +594,21 @@ static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option,
case TFRC_OPT_LOSS_INTERVALS:
opt_recv->ccid3or_loss_intervals_idx = idx;
opt_recv->ccid3or_loss_intervals_len = len;
- ccid3_pr_debug("%s, sk=%p, LOSS_INTERVALS=(%u, %u)\n",
+ 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)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, invalid "
- "len for TFRC_OPT_RECEIVE_RATE\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), invalid len %d "
+ "for TFRC_OPT_RECEIVE_RATE\n",
+ dccp_role(sk), sk, len);
rc = -EINVAL;
} else {
- opt_recv->ccid3or_receive_rate = ntohl(*(__be32 *)value);
- ccid3_pr_debug("%s, sk=%p, RECEIVE_RATE=%u\n",
+ opt_recv->ccid3or_receive_rate =
+ ntohl(*(__be32 *)value);
+ ccid3_pr_debug("%s(%p), RECEIVE_RATE=%u\n",
dccp_role(sk), sk,
opt_recv->ccid3or_receive_rate);
}
@@ -650,22 +620,15 @@ static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option,
static int ccid3_hc_tx_init(struct ccid *ccid, struct sock *sk)
{
- struct dccp_sock *dp = dccp_sk(sk);
struct ccid3_hc_tx_sock *hctx = ccid_priv(ccid);
- if (dp->dccps_packet_size >= TFRC_MIN_PACKET_SIZE &&
- dp->dccps_packet_size <= TFRC_MAX_PACKET_SIZE)
- hctx->ccid3hctx_s = dp->dccps_packet_size;
- else
- hctx->ccid3hctx_s = TFRC_STD_PACKET_SIZE;
-
- /* Set transmission rate to 1 packet per second */
- hctx->ccid3hctx_x = hctx->ccid3hctx_s;
- hctx->ccid3hctx_t_rto = USEC_PER_SEC;
+ hctx->ccid3hctx_s = 0;
+ hctx->ccid3hctx_rtt = 0;
hctx->ccid3hctx_state = TFRC_SSTATE_NO_SENT;
INIT_LIST_HEAD(&hctx->ccid3hctx_hist);
- hctx->ccid3hctx_no_feedback_timer.function = ccid3_hc_tx_no_feedback_timer;
+ hctx->ccid3hctx_no_feedback_timer.function =
+ ccid3_hc_tx_no_feedback_timer;
hctx->ccid3hctx_no_feedback_timer.data = (unsigned long)sk;
init_timer(&hctx->ccid3hctx_no_feedback_timer);
@@ -689,14 +652,7 @@ static void ccid3_hc_tx_exit(struct sock *sk)
* RX Half Connection methods
*/
-/* TFRC receiver states */
-enum ccid3_hc_rx_states {
- TFRC_RSTATE_NO_DATA = 1,
- TFRC_RSTATE_DATA,
- TFRC_RSTATE_TERM = 127,
-};
-
-#ifdef CCID3_DEBUG
+#ifdef CONFIG_IP_DCCP_CCID3_DEBUG
static const char *ccid3_rx_state_name(enum ccid3_hc_rx_states state)
{
static char *ccid3_rx_state_names[] = {
@@ -722,14 +678,24 @@ static void ccid3_hc_rx_set_state(struct sock *sk,
hcrx->ccid3hcrx_state = state;
}
+static inline void ccid3_hc_rx_update_s(struct ccid3_hc_rx_sock *hcrx, int len)
+{
+ if (unlikely(len == 0)) /* don't update on empty packets (e.g. ACKs) */
+ ccid3_pr_debug("Packet payload length is 0 - not updating\n");
+ else
+ hcrx->ccid3hcrx_s = hcrx->ccid3hcrx_s == 0 ? len :
+ (9 * hcrx->ccid3hcrx_s + len) / 10;
+}
+
static void ccid3_hc_rx_send_feedback(struct sock *sk)
{
struct ccid3_hc_rx_sock *hcrx = ccid3_hc_rx_sk(sk);
struct dccp_sock *dp = dccp_sk(sk);
struct dccp_rx_hist_entry *packet;
struct timeval now;
+ suseconds_t delta;
- ccid3_pr_debug("%s, sk=%p\n", dccp_role(sk), sk);
+ ccid3_pr_debug("%s(%p) - entry \n", dccp_role(sk), sk);
dccp_timestamp(sk, &now);
@@ -737,25 +703,22 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk)
case TFRC_RSTATE_NO_DATA:
hcrx->ccid3hcrx_x_recv = 0;
break;
- case TFRC_RSTATE_DATA: {
- const u32 delta = timeval_delta(&now,
- &hcrx->ccid3hcrx_tstamp_last_feedback);
- hcrx->ccid3hcrx_x_recv = usecs_div(hcrx->ccid3hcrx_bytes_recv,
- delta);
- }
+ case TFRC_RSTATE_DATA:
+ delta = timeval_delta(&now,
+ &hcrx->ccid3hcrx_tstamp_last_feedback);
+ DCCP_BUG_ON(delta < 0);
+ hcrx->ccid3hcrx_x_recv =
+ scaled_div32(hcrx->ccid3hcrx_bytes_recv, delta);
break;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hcrx->ccid3hcrx_state);
- dump_stack();
+ case TFRC_RSTATE_TERM:
+ DCCP_BUG("%s(%p) - Illegal state TERM", dccp_role(sk), sk);
return;
}
packet = dccp_rx_hist_find_data_packet(&hcrx->ccid3hcrx_hist);
if (unlikely(packet == NULL)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, no data packet "
- "in history!\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), no data packet in history!\n",
+ dccp_role(sk), sk);
return;
}
@@ -763,13 +726,19 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk)
hcrx->ccid3hcrx_ccval_last_counter = packet->dccphrx_ccval;
hcrx->ccid3hcrx_bytes_recv = 0;
- /* Convert to multiples of 10us */
- hcrx->ccid3hcrx_elapsed_time =
- timeval_delta(&now, &packet->dccphrx_tstamp) / 10;
+ /* Elapsed time information [RFC 4340, 13.2] in units of 10 * usecs */
+ delta = timeval_delta(&now, &packet->dccphrx_tstamp);
+ DCCP_BUG_ON(delta < 0);
+ hcrx->ccid3hcrx_elapsed_time = delta / 10;
+
if (hcrx->ccid3hcrx_p == 0)
- hcrx->ccid3hcrx_pinv = ~0;
- else
+ hcrx->ccid3hcrx_pinv = ~0U; /* see RFC 4342, 8.5 */
+ else if (hcrx->ccid3hcrx_p > 1000000) {
+ DCCP_WARN("p (%u) > 100%%\n", hcrx->ccid3hcrx_p);
+ hcrx->ccid3hcrx_pinv = 1; /* use 100% in this case */
+ } else
hcrx->ccid3hcrx_pinv = 1000000 / hcrx->ccid3hcrx_p;
+
dp->dccps_hc_rx_insert_options = 1;
dccp_send_ack(sk);
}
@@ -797,9 +766,9 @@ static int ccid3_hc_rx_insert_options(struct sock *sk, struct sk_buff *skb)
hcrx->ccid3hcrx_elapsed_time)) ||
dccp_insert_option_timestamp(sk, skb) ||
dccp_insert_option(sk, skb, TFRC_OPT_LOSS_EVENT_RATE,
- &pinv, sizeof(pinv)) ||
+ &pinv, sizeof(pinv)) ||
dccp_insert_option(sk, skb, TFRC_OPT_RECEIVE_RATE,
- &x_recv, sizeof(x_recv)))
+ &x_recv, sizeof(x_recv)))
return -1;
return 0;
@@ -813,12 +782,13 @@ static u32 ccid3_hc_rx_calc_first_li(struct sock *sk)
{
struct ccid3_hc_rx_sock *hcrx = ccid3_hc_rx_sk(sk);
struct dccp_rx_hist_entry *entry, *next, *tail = NULL;
- u32 rtt, delta, x_recv, fval, p, tmp2;
+ u32 x_recv, p;
+ suseconds_t rtt, delta;
struct timeval tstamp = { 0, };
int interval = 0;
int win_count = 0;
int step = 0;
- u64 tmp1;
+ u64 fval;
list_for_each_entry_safe(entry, next, &hcrx->ccid3hcrx_hist,
dccphrx_node) {
@@ -843,58 +813,66 @@ static u32 ccid3_hc_rx_calc_first_li(struct sock *sk)
}
if (unlikely(step == 0)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, packet history "
- "contains no data packets!\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), packet history has no data packets!\n",
+ dccp_role(sk), sk);
return ~0;
}
if (unlikely(interval == 0)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, Could not find a "
- "win_count interval > 0. Defaulting to 1\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), Could not find a win_count interval > 0."
+ "Defaulting to 1\n", dccp_role(sk), sk);
interval = 1;
}
found:
if (!tail) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: tail is null\n",
- __FUNCTION__);
+ DCCP_CRIT("tail is null\n");
return ~0;
}
- rtt = timeval_delta(&tstamp, &tail->dccphrx_tstamp) * 4 / interval;
- ccid3_pr_debug("%s, sk=%p, approximated RTT to %uus\n",
- dccp_role(sk), sk, rtt);
- if (rtt == 0)
- rtt = 1;
-
- dccp_timestamp(sk, &tstamp);
- delta = timeval_delta(&tstamp, &hcrx->ccid3hcrx_tstamp_last_feedback);
- x_recv = usecs_div(hcrx->ccid3hcrx_bytes_recv, delta);
- if (x_recv == 0)
- x_recv = hcrx->ccid3hcrx_x_recv;
+ delta = timeval_delta(&tstamp, &tail->dccphrx_tstamp);
+ DCCP_BUG_ON(delta < 0);
- tmp1 = (u64)x_recv * (u64)rtt;
- do_div(tmp1,10000000);
- tmp2 = (u32)tmp1;
+ rtt = delta * 4 / interval;
+ ccid3_pr_debug("%s(%p), approximated RTT to %dus\n",
+ dccp_role(sk), sk, (int)rtt);
- if (!tmp2) {
- LIMIT_NETDEBUG(KERN_WARNING "tmp2 = 0 "
- "%s: x_recv = %u, rtt =%u\n",
- __FUNCTION__, x_recv, rtt);
+ /*
+ * Determine the length of the first loss interval via inverse lookup.
+ * Assume that X_recv can be computed by the throughput equation
+ * s
+ * X_recv = --------
+ * R * fval
+ * Find some p such that f(p) = fval; return 1/p [RFC 3448, 6.3.1].
+ */
+ if (rtt == 0) { /* would result in divide-by-zero */
+ DCCP_WARN("RTT==0\n");
return ~0;
}
- fval = (hcrx->ccid3hcrx_s * 100000) / tmp2;
- /* do not alter order above or you will get overflow on 32 bit */
+ dccp_timestamp(sk, &tstamp);
+ delta = timeval_delta(&tstamp, &hcrx->ccid3hcrx_tstamp_last_feedback);
+ DCCP_BUG_ON(delta <= 0);
+
+ x_recv = scaled_div32(hcrx->ccid3hcrx_bytes_recv, delta);
+ if (x_recv == 0) { /* would also trigger divide-by-zero */
+ DCCP_WARN("X_recv==0\n");
+ if ((x_recv = hcrx->ccid3hcrx_x_recv) == 0) {
+ DCCP_BUG("stored value of X_recv is zero");
+ return ~0;
+ }
+ }
+
+ fval = scaled_div(hcrx->ccid3hcrx_s, rtt);
+ fval = scaled_div32(fval, x_recv);
p = tfrc_calc_x_reverse_lookup(fval);
- ccid3_pr_debug("%s, sk=%p, receive rate=%u bytes/s, implied "
+
+ ccid3_pr_debug("%s(%p), receive rate=%u bytes/s, implied "
"loss rate=%u\n", dccp_role(sk), sk, x_recv, p);
if (p == 0)
return ~0;
else
- return 1000000 / p;
+ return 1000000 / p;
}
static void ccid3_hc_rx_update_li(struct sock *sk, u64 seq_loss, u8 win_loss)
@@ -924,11 +902,10 @@ static void ccid3_hc_rx_update_li(struct sock *sk, u64 seq_loss, u8 win_loss)
/* new loss event detected */
/* calculate last interval length */
seq_temp = dccp_delta_seqno(head->dccplih_seqno, seq_loss);
- entry = dccp_li_hist_entry_new(ccid3_li_hist, SLAB_ATOMIC);
+ entry = dccp_li_hist_entry_new(ccid3_li_hist, GFP_ATOMIC);
if (entry == NULL) {
- printk(KERN_CRIT "%s: out of memory\n",__FUNCTION__);
- dump_stack();
+ DCCP_BUG("out of memory - can not allocate entry");
return;
}
@@ -949,7 +926,8 @@ static int ccid3_hc_rx_detect_loss(struct sock *sk,
struct dccp_rx_hist_entry *packet)
{
struct ccid3_hc_rx_sock *hcrx = ccid3_hc_rx_sk(sk);
- struct dccp_rx_hist_entry *rx_hist = dccp_rx_hist_head(&hcrx->ccid3hcrx_hist);
+ struct dccp_rx_hist_entry *rx_hist =
+ dccp_rx_hist_head(&hcrx->ccid3hcrx_hist);
u64 seqno = packet->dccphrx_seqno;
u64 tmp_seqno;
int loss = 0;
@@ -977,7 +955,7 @@ static int ccid3_hc_rx_detect_loss(struct sock *sk,
dccp_inc_seqno(&tmp_seqno);
while (dccp_rx_hist_find_entry(&hcrx->ccid3hcrx_hist,
tmp_seqno, &ccval)) {
- hcrx->ccid3hcrx_seqno_nonloss = tmp_seqno;
+ hcrx->ccid3hcrx_seqno_nonloss = tmp_seqno;
hcrx->ccid3hcrx_ccval_nonloss = ccval;
dccp_inc_seqno(&tmp_seqno);
}
@@ -1003,13 +981,11 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
const struct dccp_options_received *opt_recv;
struct dccp_rx_hist_entry *packet;
struct timeval now;
- u8 win_count;
- u32 p_prev, rtt_prev, r_sample, t_elapsed;
- int loss;
+ u32 p_prev, rtt_prev;
+ suseconds_t r_sample, t_elapsed;
+ int loss, payload_size;
- BUG_ON(hcrx == NULL ||
- !(hcrx->ccid3hcrx_state == TFRC_RSTATE_NO_DATA ||
- hcrx->ccid3hcrx_state == TFRC_RSTATE_DATA));
+ BUG_ON(hcrx == NULL);
opt_recv = &dccp_sk(sk)->dccps_options_received;
@@ -1026,12 +1002,13 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
r_sample = timeval_usecs(&now);
t_elapsed = opt_recv->dccpor_elapsed_time * 10;
+ DCCP_BUG_ON(r_sample < 0);
if (unlikely(r_sample <= t_elapsed))
- LIMIT_NETDEBUG(KERN_WARNING "%s: r_sample=%uus, "
- "t_elapsed=%uus\n",
- __FUNCTION__, r_sample, t_elapsed);
+ DCCP_WARN("r_sample=%ldus, t_elapsed=%ldus\n",
+ r_sample, t_elapsed);
else
r_sample -= t_elapsed;
+ CCID3_RTT_SANITY_CHECK(r_sample);
if (hcrx->ccid3hcrx_state == TFRC_RSTATE_NO_DATA)
hcrx->ccid3hcrx_rtt = r_sample;
@@ -1040,8 +1017,8 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
r_sample / 10;
if (rtt_prev != hcrx->ccid3hcrx_rtt)
- ccid3_pr_debug("%s, New RTT=%uus, elapsed time=%u\n",
- dccp_role(sk), hcrx->ccid3hcrx_rtt,
+ ccid3_pr_debug("%s(%p), New RTT=%uus, elapsed time=%u\n",
+ dccp_role(sk), sk, hcrx->ccid3hcrx_rtt,
opt_recv->dccpor_elapsed_time);
break;
case DCCP_PKT_DATA:
@@ -1051,52 +1028,48 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
}
packet = dccp_rx_hist_entry_new(ccid3_rx_hist, sk, opt_recv->dccpor_ndp,
- skb, SLAB_ATOMIC);
+ skb, GFP_ATOMIC);
if (unlikely(packet == NULL)) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: %s, sk=%p, Not enough mem to "
- "add rx packet to history, consider it lost!\n",
- __FUNCTION__, dccp_role(sk), sk);
+ DCCP_WARN("%s(%p), Not enough mem to add rx packet "
+ "to history, consider it lost!\n", dccp_role(sk), sk);
return;
}
- win_count = packet->dccphrx_ccval;
-
loss = ccid3_hc_rx_detect_loss(sk, packet);
if (DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_ACK)
return;
+ payload_size = skb->len - dccp_hdr(skb)->dccph_doff * 4;
+ ccid3_hc_rx_update_s(hcrx, payload_size);
+
switch (hcrx->ccid3hcrx_state) {
case TFRC_RSTATE_NO_DATA:
- ccid3_pr_debug("%s, sk=%p(%s), skb=%p, sending initial "
- "feedback\n",
- dccp_role(sk), sk,
+ ccid3_pr_debug("%s(%p, state=%s), skb=%p, sending initial "
+ "feedback\n", dccp_role(sk), sk,
dccp_state_name(sk->sk_state), skb);
ccid3_hc_rx_send_feedback(sk);
ccid3_hc_rx_set_state(sk, TFRC_RSTATE_DATA);
return;
case TFRC_RSTATE_DATA:
- hcrx->ccid3hcrx_bytes_recv += skb->len -
- dccp_hdr(skb)->dccph_doff * 4;
+ hcrx->ccid3hcrx_bytes_recv += payload_size;
if (loss)
break;
dccp_timestamp(sk, &now);
- if (timeval_delta(&now, &hcrx->ccid3hcrx_tstamp_last_ack) >=
- hcrx->ccid3hcrx_rtt) {
+ if ((timeval_delta(&now, &hcrx->ccid3hcrx_tstamp_last_ack) -
+ (suseconds_t)hcrx->ccid3hcrx_rtt) >= 0) {
hcrx->ccid3hcrx_tstamp_last_ack = now;
ccid3_hc_rx_send_feedback(sk);
}
return;
- default:
- printk(KERN_CRIT "%s: %s, sk=%p, Illegal state (%d)!\n",
- __FUNCTION__, dccp_role(sk), sk, hcrx->ccid3hcrx_state);
- dump_stack();
+ case TFRC_RSTATE_TERM:
+ DCCP_BUG("%s(%p) - Illegal state TERM", dccp_role(sk), sk);
return;
}
/* Dealing with packet loss */
- ccid3_pr_debug("%s, sk=%p(%s), data loss! Reacting...\n",
+ ccid3_pr_debug("%s(%p, state=%s), data loss! Reacting...\n",
dccp_role(sk), sk, dccp_state_name(sk->sk_state));
p_prev = hcrx->ccid3hcrx_p;
@@ -1108,10 +1081,8 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
/* Scaling up by 1000000 as fixed decimal */
if (i_mean != 0)
hcrx->ccid3hcrx_p = 1000000 / i_mean;
- } else {
- printk(KERN_CRIT "%s: empty loss hist\n",__FUNCTION__);
- dump_stack();
- }
+ } else
+ DCCP_BUG("empty loss history");
if (hcrx->ccid3hcrx_p > p_prev) {
ccid3_hc_rx_send_feedback(sk);
@@ -1121,23 +1092,17 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb)
static int ccid3_hc_rx_init(struct ccid *ccid, struct sock *sk)
{
- struct dccp_sock *dp = dccp_sk(sk);
struct ccid3_hc_rx_sock *hcrx = ccid_priv(ccid);
- ccid3_pr_debug("%s, sk=%p\n", dccp_role(sk), sk);
-
- if (dp->dccps_packet_size >= TFRC_MIN_PACKET_SIZE &&
- dp->dccps_packet_size <= TFRC_MAX_PACKET_SIZE)
- hcrx->ccid3hcrx_s = dp->dccps_packet_size;
- else
- hcrx->ccid3hcrx_s = TFRC_STD_PACKET_SIZE;
+ ccid3_pr_debug("entry\n");
hcrx->ccid3hcrx_state = TFRC_RSTATE_NO_DATA;
INIT_LIST_HEAD(&hcrx->ccid3hcrx_hist);
INIT_LIST_HEAD(&hcrx->ccid3hcrx_li_hist);
dccp_timestamp(sk, &hcrx->ccid3hcrx_tstamp_last_ack);
hcrx->ccid3hcrx_tstamp_last_feedback = hcrx->ccid3hcrx_tstamp_last_ack;
- hcrx->ccid3hcrx_rtt = 5000; /* XXX 5ms for now... */
+ hcrx->ccid3hcrx_s = 0;
+ hcrx->ccid3hcrx_rtt = 0;
return 0;
}
@@ -1166,9 +1131,9 @@ static void ccid3_hc_rx_get_info(struct sock *sk, struct tcp_info *info)
BUG_ON(hcrx == NULL);
- info->tcpi_ca_state = hcrx->ccid3hcrx_state;
- info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
- info->tcpi_rcv_rtt = hcrx->ccid3hcrx_rtt;
+ info->tcpi_ca_state = hcrx->ccid3hcrx_state;
+ info->tcpi_options |= TCPI_OPT_TIMESTAMPS;
+ info->tcpi_rcv_rtt = hcrx->ccid3hcrx_rtt;
}
static void ccid3_hc_tx_get_info(struct sock *sk, struct tcp_info *info)
@@ -1249,7 +1214,6 @@ static struct ccid_operations ccid3 = {
.ccid_hc_tx_send_packet = ccid3_hc_tx_send_packet,
.ccid_hc_tx_packet_sent = ccid3_hc_tx_packet_sent,
.ccid_hc_tx_packet_recv = ccid3_hc_tx_packet_recv,
- .ccid_hc_tx_insert_options = ccid3_hc_tx_insert_options,
.ccid_hc_tx_parse_options = ccid3_hc_tx_parse_options,
.ccid_hc_rx_obj_size = sizeof(struct ccid3_hc_rx_sock),
.ccid_hc_rx_init = ccid3_hc_rx_init,
@@ -1261,9 +1225,11 @@ static struct ccid_operations ccid3 = {
.ccid_hc_rx_getsockopt = ccid3_hc_rx_getsockopt,
.ccid_hc_tx_getsockopt = ccid3_hc_tx_getsockopt,
};
-
+
+#ifdef CONFIG_IP_DCCP_CCID3_DEBUG
module_param(ccid3_debug, int, 0444);
MODULE_PARM_DESC(ccid3_debug, "Enable debug messages");
+#endif
static __init int ccid3_module_init(void)
{
@@ -1282,7 +1248,7 @@ static __init int ccid3_module_init(void)
goto out_free_tx;
rc = ccid_register(&ccid3);
- if (rc != 0)
+ if (rc != 0)
goto out_free_loss_interval_history;
out:
return rc;
diff --git a/net/dccp/ccids/ccid3.h b/net/dccp/ccids/ccid3.h
index 0a2cb7536d2..15776a88c09 100644
--- a/net/dccp/ccids/ccid3.h
+++ b/net/dccp/ccids/ccid3.h
@@ -42,22 +42,24 @@
#include <linux/tfrc.h>
#include "../ccid.h"
-#define TFRC_MIN_PACKET_SIZE 16
-#define TFRC_STD_PACKET_SIZE 256
-#define TFRC_MAX_PACKET_SIZE 65535
-
-/* Two seconds as per CCID3 spec */
+/* Two seconds as per RFC 3448 4.2 */
#define TFRC_INITIAL_TIMEOUT (2 * USEC_PER_SEC)
-#define TFRC_INITIAL_IPI (USEC_PER_SEC / 4)
-
/* In usecs - half the scheduling granularity as per RFC3448 4.6 */
#define TFRC_OPSYS_HALF_TIME_GRAN (USEC_PER_SEC / (2 * HZ))
-/* In seconds */
-#define TFRC_MAX_BACK_OFF_TIME 64
+/* Parameter t_mbi from [RFC 3448, 4.3]: backoff interval in seconds */
+#define TFRC_T_MBI 64
+
+/* What we think is a reasonable upper limit on RTT values */
+#define CCID3_SANE_RTT_MAX ((suseconds_t)(4 * USEC_PER_SEC))
-#define TFRC_SMALLEST_P 40
+#define CCID3_RTT_SANITY_CHECK(rtt) do { \
+ if (rtt > CCID3_SANE_RTT_MAX) { \
+ DCCP_CRIT("RTT (%d) too large, substituting %d", \
+ (int)rtt, (int)CCID3_SANE_RTT_MAX); \
+ rtt = CCID3_SANE_RTT_MAX; \
+ } } while (0)
enum ccid3_options {
TFRC_OPT_LOSS_EVENT_RATE = 192,
@@ -73,26 +75,36 @@ struct ccid3_options_received {
u32 ccid3or_receive_rate;
};
-/** struct ccid3_hc_tx_sock - CCID3 sender half connection sock
+/* TFRC sender states */
+enum ccid3_hc_tx_states {
+ TFRC_SSTATE_NO_SENT = 1,
+ TFRC_SSTATE_NO_FBACK,
+ TFRC_SSTATE_FBACK,
+ TFRC_SSTATE_TERM,
+};
+
+/** struct ccid3_hc_tx_sock - CCID3 sender half-connection socket
*
- * @ccid3hctx_state - Sender state
- * @ccid3hctx_x - Current sending rate
- * @ccid3hctx_x_recv - Receive rate
- * @ccid3hctx_x_calc - Calculated send (?) rate
- * @ccid3hctx_s - Packet size
- * @ccid3hctx_rtt - Estimate of current round trip time in usecs
- * @@ccid3hctx_p - Current loss event rate (0-1) scaled by 1000000
- * @ccid3hctx_last_win_count - Last window counter sent
- * @ccid3hctx_t_last_win_count - Timestamp of earliest packet
- * with last_win_count value sent
- * @ccid3hctx_no_feedback_timer - Handle to no feedback timer
- * @ccid3hctx_idle - FIXME
- * @ccid3hctx_t_ld - Time last doubled during slow start
- * @ccid3hctx_t_nom - Nominal send time of next packet
- * @ccid3hctx_t_ipi - Interpacket (send) interval
- * @ccid3hctx_delta - Send timer delta
- * @ccid3hctx_hist - Packet history
- */
+ * @ccid3hctx_x - Current sending rate in 64 * bytes per second
+ * @ccid3hctx_x_recv - Receive rate in 64 * bytes per second
+ * @ccid3hctx_x_calc - Calculated rate in bytes per second
+ * @ccid3hctx_rtt - Estimate of current round trip time in usecs
+ * @ccid3hctx_p - Current loss event rate (0-1) scaled by 1000000
+ * @ccid3hctx_s - Packet size in bytes
+ * @ccid3hctx_t_rto - Nofeedback Timer setting in usecs
+ * @ccid3hctx_t_ipi - Interpacket (send) interval (RFC 3448, 4.6) in usecs
+ * @ccid3hctx_state - Sender state, one of %ccid3_hc_tx_states
+ * @ccid3hctx_last_win_count - Last window counter sent
+ * @ccid3hctx_t_last_win_count - Timestamp of earliest packet
+ * with last_win_count value sent
+ * @ccid3hctx_no_feedback_timer - Handle to no feedback timer
+ * @ccid3hctx_idle - Flag indicating that sender is idling
+ * @ccid3hctx_t_ld - Time last doubled during slow start
+ * @ccid3hctx_t_nom - Nominal send time of next packet
+ * @ccid3hctx_delta - Send timer delta (RFC 3448, 4.6) in usecs
+ * @ccid3hctx_hist - Packet history
+ * @ccid3hctx_options_received - Parsed set of retrieved options
+ */
struct ccid3_hc_tx_sock {
struct tfrc_tx_info ccid3hctx_tfrc;
#define ccid3hctx_x ccid3hctx_tfrc.tfrctx_x
@@ -103,7 +115,7 @@ struct ccid3_hc_tx_sock {
#define ccid3hctx_t_rto ccid3hctx_tfrc.tfrctx_rto
#define ccid3hctx_t_ipi ccid3hctx_tfrc.tfrctx_ipi
u16 ccid3hctx_s;
- u8 ccid3hctx_state;
+ enum ccid3_hc_tx_states ccid3hctx_state:8;
u8 ccid3hctx_last_win_count;
u8 ccid3hctx_idle;
struct timeval ccid3hctx_t_last_win_count;
@@ -115,23 +127,48 @@ struct ccid3_hc_tx_sock {
struct ccid3_options_received ccid3hctx_options_received;
};
+/* TFRC receiver states */
+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
+ *
+ * @ccid3hcrx_x_recv - Receiver estimate of send rate (RFC 3448 4.3)
+ * @ccid3hcrx_rtt - Receiver estimate of rtt (non-standard)
+ * @ccid3hcrx_p - current loss event rate (RFC 3448 5.4)
+ * @ccid3hcrx_seqno_nonloss - Last received non-loss sequence number
+ * @ccid3hcrx_ccval_nonloss - Last received non-loss Window CCVal
+ * @ccid3hcrx_ccval_last_counter - Tracks window counter (RFC 4342, 8.1)
+ * @ccid3hcrx_state - receiver state, one of %ccid3_hc_rx_states
+ * @ccid3hcrx_bytes_recv - Total sum of DCCP payload bytes
+ * @ccid3hcrx_tstamp_last_feedback - Time at which last feedback was sent
+ * @ccid3hcrx_tstamp_last_ack - Time at which last feedback was sent
+ * @ccid3hcrx_hist - Packet history
+ * @ccid3hcrx_li_hist - Loss Interval History
+ * @ccid3hcrx_s - Received packet size in bytes
+ * @ccid3hcrx_pinv - Inverse of Loss Event Rate (RFC 4342, sec. 8.5)
+ * @ccid3hcrx_elapsed_time - Time since packet reception
+ */
struct ccid3_hc_rx_sock {
- struct tfrc_rx_info ccid3hcrx_tfrc;
-#define ccid3hcrx_x_recv ccid3hcrx_tfrc.tfrcrx_x_recv
-#define ccid3hcrx_rtt ccid3hcrx_tfrc.tfrcrx_rtt
-#define ccid3hcrx_p ccid3hcrx_tfrc.tfrcrx_p
- u64 ccid3hcrx_seqno_nonloss:48,
- ccid3hcrx_ccval_nonloss:4,
- ccid3hcrx_state:8,
- ccid3hcrx_ccval_last_counter:4;
- u32 ccid3hcrx_bytes_recv;
- struct timeval ccid3hcrx_tstamp_last_feedback;
- struct timeval ccid3hcrx_tstamp_last_ack;
- struct list_head ccid3hcrx_hist;
- struct list_head ccid3hcrx_li_hist;
- u16 ccid3hcrx_s;
- u32 ccid3hcrx_pinv;
- u32 ccid3hcrx_elapsed_time;
+ struct tfrc_rx_info ccid3hcrx_tfrc;
+#define ccid3hcrx_x_recv ccid3hcrx_tfrc.tfrcrx_x_recv
+#define ccid3hcrx_rtt ccid3hcrx_tfrc.tfrcrx_rtt
+#define ccid3hcrx_p ccid3hcrx_tfrc.tfrcrx_p
+ u64 ccid3hcrx_seqno_nonloss:48,
+ ccid3hcrx_ccval_nonloss:4,
+ ccid3hcrx_ccval_last_counter:4;
+ enum ccid3_hc_rx_states ccid3hcrx_state:8;
+ u32 ccid3hcrx_bytes_recv;
+ struct timeval ccid3hcrx_tstamp_last_feedback;
+ struct timeval ccid3hcrx_tstamp_last_ack;
+ struct list_head ccid3hcrx_hist;
+ struct list_head ccid3hcrx_li_hist;
+ u16 ccid3hcrx_s;
+ u32 ccid3hcrx_pinv;
+ u32 ccid3hcrx_elapsed_time;
};
static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk)
diff --git a/net/dccp/ccids/lib/loss_interval.c b/net/dccp/ccids/lib/loss_interval.c
index 906c81ab9d4..0a0baef16b3 100644
--- a/net/dccp/ccids/lib/loss_interval.c
+++ b/net/dccp/ccids/lib/loss_interval.c
@@ -13,7 +13,7 @@
#include <linux/module.h>
#include <net/sock.h>
-
+#include "../../dccp.h"
#include "loss_interval.h"
struct dccp_li_hist *dccp_li_hist_new(const char *name)
@@ -109,7 +109,7 @@ u32 dccp_li_hist_calc_i_mean(struct list_head *list)
i_tot = max(i_tot0, i_tot1);
if (!w_tot) {
- LIMIT_NETDEBUG(KERN_WARNING "%s: w_tot = 0\n", __FUNCTION__);
+ DCCP_WARN("w_tot = 0\n");
return 1;
}
@@ -125,10 +125,10 @@ int dccp_li_hist_interval_new(struct dccp_li_hist *hist,
int i;
for (i = 0; i < DCCP_LI_HIST_IVAL_F_LENGTH; i++) {
- entry = dccp_li_hist_entry_new(hist, SLAB_ATOMIC);
+ entry = dccp_li_hist_entry_new(hist, GFP_ATOMIC);
if (entry == NULL) {
dccp_li_hist_purge(hist, list);
- dump_stack();
+ DCCP_BUG("loss interval list entry is NULL");
return 0;
}
entry->dccplih_interval = ~0;
diff --git a/net/dccp/ccids/lib/loss_interval.h b/net/dccp/ccids/lib/loss_interval.h
index 0ae85f0340b..eb257014dd7 100644
--- a/net/dccp/ccids/lib/loss_interval.h
+++ b/net/dccp/ccids/lib/loss_interval.h
@@ -20,7 +20,7 @@
#define DCCP_LI_HIST_IVAL_F_LENGTH 8
struct dccp_li_hist {
- kmem_cache_t *dccplih_slab;
+ struct kmem_cache *dccplih_slab;
};
extern struct dccp_li_hist *dccp_li_hist_new(const char *name);
diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c
index b876c9c81c6..2e8ef42721e 100644
--- a/net/dccp/ccids/lib/packet_history.c
+++ b/net/dccp/ccids/lib/packet_history.c
@@ -36,9 +36,100 @@
#include <linux/module.h>
#include <linux/string.h>
-
#include "packet_history.h"
+/*
+ * Transmitter History Routines
+ */
+struct dccp_tx_hist *dccp_tx_hist_new(const char *name)
+{
+ struct dccp_tx_hist *hist = kmalloc(sizeof(*hist), GFP_ATOMIC);
+ static const char dccp_tx_hist_mask[] = "tx_hist_%s";
+ char *slab_name;
+
+ if (hist == NULL)
+ goto out;
+
+ slab_name = kmalloc(strlen(name) + sizeof(dccp_tx_hist_mask) - 1,
+ GFP_ATOMIC);
+ if (slab_name == NULL)
+ goto out_free_hist;
+
+ sprintf(slab_name, dccp_tx_hist_mask, name);
+ hist->dccptxh_slab = kmem_cache_create(slab_name,
+ sizeof(struct dccp_tx_hist_entry),
+ 0, SLAB_HWCACHE_ALIGN,
+ NULL, NULL);
+ if (hist->dccptxh_slab == NULL)
+ goto out_free_slab_name;
+out:
+ return hist;
+out_free_slab_name:
+ kfree(slab_name);
+out_free_hist:
+ kfree(hist);
+ hist = NULL;
+ goto out;
+}
+
+EXPORT_SYMBOL_GPL(dccp_tx_hist_new);
+
+void dccp_tx_hist_delete(struct dccp_tx_hist *hist)
+{
+ const char* name = kmem_cache_name(hist->dccptxh_slab);
+
+ kmem_cache_destroy(hist->dccptxh_slab);
+ kfree(name);
+ kfree(hist);
+}
+
+EXPORT_SYMBOL_GPL(dccp_tx_hist_delete);
+
+struct dccp_tx_hist_entry *
+ dccp_tx_hist_find_entry(const struct list_head *list, const u64 seq)
+{
+ struct dccp_tx_hist_entry *packet = NULL, *entry;
+
+ list_for_each_entry(entry, list, dccphtx_node)
+ if (entry->dccphtx_seqno == seq) {
+ packet = entry;
+ break;
+ }
+
+ return packet;
+}
+
+EXPORT_SYMBOL_GPL(dccp_tx_hist_find_entry);
+
+void dccp_tx_hist_purge(struct dccp_tx_hist *hist, struct list_head *list)
+{
+ struct dccp_tx_hist_entry *entry, *next;
+
+ list_for_each_entry_safe(entry, next, list, dccphtx_node) {
+ list_del_init(&entry->dccphtx_node);
+ dccp_tx_hist_entry_delete(hist, entry);
+ }
+}
+
+EXPORT_SYMBOL_GPL(dccp_tx_hist_purge);
+
+void dccp_tx_hist_purge_older(struct dccp_tx_hist *hist,
+ struct list_head *list,
+ struct dccp_tx_hist_entry *packet)
+{
+ struct dccp_tx_hist_entry *next;
+
+ list_for_each_entry_safe_continue(packet, next, list, dccphtx_node) {
+ list_del_init(&packet->dccphtx_node);
+ dccp_tx_hist_entry_delete(hist, packet);
+ }
+}
+
+EXPORT_SYMBOL_GPL(dccp_tx_hist_purge_older);
+
+/*
+ * Receiver History Routines
+ */
struct dccp_rx_hist *dccp_rx_hist_new(const char *name)
{
struct dccp_rx_hist *hist = kmalloc(sizeof(*hist), GFP_ATOMIC);
@@ -83,18 +174,24 @@ void dccp_rx_hist_delete(struct dccp_rx_hist *hist)
EXPORT_SYMBOL_GPL(dccp_rx_hist_delete);
-void dccp_rx_hist_purge(struct dccp_rx_hist *hist, struct list_head *list)
+int dccp_rx_hist_find_entry(const struct list_head *list, const u64 seq,
+ u8 *ccval)
{
- struct dccp_rx_hist_entry *entry, *next;
+ struct dccp_rx_hist_entry *packet = NULL, *entry;
- list_for_each_entry_safe(entry, next, list, dccphrx_node) {
- list_del_init(&entry->dccphrx_node);
- kmem_cache_free(hist->dccprxh_slab, entry);
- }
-}
+ list_for_each_entry(entry, list, dccphrx_node)
+ if (entry->dccphrx_seqno == seq) {
+ packet = entry;
+ break;
+ }
-EXPORT_SYMBOL_GPL(dccp_rx_hist_purge);
+ if (packet)
+ *ccval = packet->dccphrx_ccval;
+ return packet != NULL;
+}
+
+EXPORT_SYMBOL_GPL(dccp_rx_hist_find_entry);
struct dccp_rx_hist_entry *
dccp_rx_hist_find_data_packet(const struct list_head *list)
{
@@ -184,110 +281,18 @@ void dccp_rx_hist_add_packet(struct dccp_rx_hist *hist,
EXPORT_SYMBOL_GPL(dccp_rx_hist_add_packet);
-struct dccp_tx_hist *dccp_tx_hist_new(const char *name)
-{
- struct dccp_tx_hist *hist = kmalloc(sizeof(*hist), GFP_ATOMIC);
- static const char dccp_tx_hist_mask[] = "tx_hist_%s";
- char *slab_name;
-
- if (hist == NULL)
- goto out;
-
- slab_name = kmalloc(strlen(name) + sizeof(dccp_tx_hist_mask) - 1,
- GFP_ATOMIC);
- if (slab_name == NULL)
- goto out_free_hist;
-
- sprintf(slab_name, dccp_tx_hist_mask, name);
- hist->dccptxh_slab = kmem_cache_create(slab_name,
- sizeof(struct dccp_tx_hist_entry),
- 0, SLAB_HWCACHE_ALIGN,
- NULL, NULL);
- if (hist->dccptxh_slab == NULL)
- goto out_free_slab_name;
-out:
- return hist;
-out_free_slab_name:
- kfree(slab_name);
-out_free_hist:
- kfree(hist);
- hist = NULL;
- goto out;
-}
-
-EXPORT_SYMBOL_GPL(dccp_tx_hist_new);
-
-void dccp_tx_hist_delete(struct dccp_tx_hist *hist)
-{
- const char* name = kmem_cache_name(hist->dccptxh_slab);
-
- kmem_cache_destroy(hist->dccptxh_slab);
- kfree(name);
- kfree(hist);
-}
-
-EXPORT_SYMBOL_GPL(dccp_tx_hist_delete);
-
-struct dccp_tx_hist_entry *
- dccp_tx_hist_find_entry(const struct list_head *list, const u64 seq)
-{
- struct dccp_tx_hist_entry *packet = NULL, *entry;
-
- list_for_each_entry(entry, list, dccphtx_node)
- if (entry->dccphtx_seqno == seq) {
- packet = entry;
- break;
- }
-
- return packet;
-}
-
-EXPORT_SYMBOL_GPL(dccp_tx_hist_find_entry);
-
-int dccp_rx_hist_find_entry(const struct list_head *list, const u64 seq,
- u8 *ccval)
-{
- struct dccp_rx_hist_entry *packet = NULL, *entry;
-
- list_for_each_entry(entry, list, dccphrx_node)
- if (entry->dccphrx_seqno == seq) {
- packet = entry;
- break;
- }
-
- if (packet)
- *ccval = packet->dccphrx_ccval;
-
- return packet != NULL;
-}
-
-EXPORT_SYMBOL_GPL(dccp_rx_hist_find_entry);
-
-void dccp_tx_hist_purge_older(struct dccp_tx_hist *hist,
- struct list_head *list,
- struct dccp_tx_hist_entry *packet)
+void dccp_rx_hist_purge(struct dccp_rx_hist *hist, struct list_head *list)
{
- struct dccp_tx_hist_entry *next;
+ struct dccp_rx_hist_entry *entry, *next;
- list_for_each_entry_safe_continue(packet, next, list, dccphtx_node) {
- list_del_init(&packet->dccphtx_node);
- dccp_tx_hist_entry_delete(hist, packet);
+ list_for_each_entry_safe(entry, next, list, dccphrx_node) {
+ list_del_init(&entry->dccphrx_node);
+ kmem_cache_free(hist->dccprxh_slab, entry);
}
}
-EXPORT_SYMBOL_GPL(dccp_tx_hist_purge_older);
-
-void dccp_tx_hist_purge(struct dccp_tx_hist *hist, struct list_head *list)
-{
- struct dccp_tx_hist_entry *entry, *next;
-
- list_for_each_entry_safe(entry, next, list, dccphtx_node) {
- list_del_init(&entry->dccphtx_node);
- dccp_tx_hist_entry_delete(hist, entry);
- }
-}
+EXPORT_SYMBOL_GPL(dccp_rx_hist_purge);
-EXPORT_SYMBOL_GPL(dccp_tx_hist_purge);
MODULE_AUTHOR("Ian McDonald <ian.mcdonald@jandi.co.nz>, "
"Arnaldo Carvalho de Melo <acme@ghostprotocols.net>");
diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h
index 067cf1c85a3..1f960c19ea1 100644
--- a/net/dccp/ccids/lib/packet_history.h
+++ b/net/dccp/ccids/lib/packet_history.h
@@ -49,43 +49,27 @@
#define TFRC_WIN_COUNT_PER_RTT 4
#define TFRC_WIN_COUNT_LIMIT 16
+/*
+ * Transmitter History data structures and declarations
+ */
struct dccp_tx_hist_entry {
struct list_head dccphtx_node;
u64 dccphtx_seqno:48,
- dccphtx_ccval:4,
dccphtx_sent:1;
u32 dccphtx_rtt;
struct timeval dccphtx_tstamp;
};
-struct dccp_rx_hist_entry {
- struct list_head dccphrx_node;
- u64 dccphrx_seqno:48,
- dccphrx_ccval:4,
- dccphrx_type:4;
- u32 dccphrx_ndp; /* In fact it is from 8 to 24 bits */
- struct timeval dccphrx_tstamp;
-};
-
struct dccp_tx_hist {
- kmem_cache_t *dccptxh_slab;
+ struct kmem_cache *dccptxh_slab;
};
extern struct dccp_tx_hist *dccp_tx_hist_new(const char *name);
-extern void dccp_tx_hist_delete(struct dccp_tx_hist *hist);
-
-struct dccp_rx_hist {
- kmem_cache_t *dccprxh_slab;
-};
-
-extern struct dccp_rx_hist *dccp_rx_hist_new(const char *name);
-extern void dccp_rx_hist_delete(struct dccp_rx_hist *hist);
-extern struct dccp_rx_hist_entry *
- dccp_rx_hist_find_data_packet(const struct list_head *list);
+extern void dccp_tx_hist_delete(struct dccp_tx_hist *hist);
static inline struct dccp_tx_hist_entry *
- dccp_tx_hist_entry_new(struct dccp_tx_hist *hist,
- const gfp_t prio)
+ dccp_tx_hist_entry_new(struct dccp_tx_hist *hist,
+ const gfp_t prio)
{
struct dccp_tx_hist_entry *entry = kmem_cache_alloc(hist->dccptxh_slab,
prio);
@@ -96,18 +80,20 @@ static inline struct dccp_tx_hist_entry *
return entry;
}
-static inline void dccp_tx_hist_entry_delete(struct dccp_tx_hist *hist,
- struct dccp_tx_hist_entry *entry)
+static inline struct dccp_tx_hist_entry *
+ dccp_tx_hist_head(struct list_head *list)
{
- if (entry != NULL)
- kmem_cache_free(hist->dccptxh_slab, entry);
+ struct dccp_tx_hist_entry *head = NULL;
+
+ if (!list_empty(list))
+ head = list_entry(list->next, struct dccp_tx_hist_entry,
+ dccphtx_node);
+ return head;
}
extern struct dccp_tx_hist_entry *
dccp_tx_hist_find_entry(const struct list_head *list,
const u64 seq);
-extern int dccp_rx_hist_find_entry(const struct list_head *list, const u64 seq,
- u8 *ccval);
static inline void dccp_tx_hist_add_entry(struct list_head *list,
struct dccp_tx_hist_entry *entry)
@@ -115,30 +101,45 @@ static inline void dccp_tx_hist_add_entry(struct list_head *list,
list_add(&entry->dccphtx_node, list);
}
+static inline void dccp_tx_hist_entry_delete(struct dccp_tx_hist *hist,
+ struct dccp_tx_hist_entry *entry)
+{
+ if (entry != NULL)
+ kmem_cache_free(hist->dccptxh_slab, entry);
+}
+
+extern void dccp_tx_hist_purge(struct dccp_tx_hist *hist,
+ struct list_head *list);
+
extern void dccp_tx_hist_purge_older(struct dccp_tx_hist *hist,
struct list_head *list,
struct dccp_tx_hist_entry *next);
-extern void dccp_tx_hist_purge(struct dccp_tx_hist *hist,
- struct list_head *list);
+/*
+ * Receiver History data structures and declarations
+ */
+struct dccp_rx_hist_entry {
+ struct list_head dccphrx_node;
+ u64 dccphrx_seqno:48,
+ dccphrx_ccval:4,
+ dccphrx_type:4;
+ u32 dccphrx_ndp; /* In fact it is from 8 to 24 bits */
+ struct timeval dccphrx_tstamp;
+};
-static inline struct dccp_tx_hist_entry *
- dccp_tx_hist_head(struct list_head *list)
-{
- struct dccp_tx_hist_entry *head = NULL;
+struct dccp_rx_hist {
+ struct kmem_cache *dccprxh_slab;
+};
- if (!list_empty(list))
- head = list_entry(list->next, struct dccp_tx_hist_entry,
- dccphtx_node);
- return head;
-}
+extern struct dccp_rx_hist *dccp_rx_hist_new(const char *name);
+extern void dccp_rx_hist_delete(struct dccp_rx_hist *hist);
static inline struct dccp_rx_hist_entry *
- dccp_rx_hist_entry_new(struct dccp_rx_hist *hist,
- const struct sock *sk,
- const u32 ndp,
- const struct sk_buff *skb,
- const gfp_t prio)
+ dccp_rx_hist_entry_new(struct dccp_rx_hist *hist,
+ const struct sock *sk,
+ const u32 ndp,
+ const struct sk_buff *skb,
+ const gfp_t prio)
{
struct dccp_rx_hist_entry *entry = kmem_cache_alloc(hist->dccprxh_slab,
prio);
@@ -156,18 +157,8 @@ static inline struct dccp_rx_hist_entry *
return entry;
}
-static inline void dccp_rx_hist_entry_delete(struct dccp_rx_hist *hist,
- struct dccp_rx_hist_entry *entry)
-{
- if (entry != NULL)
- kmem_cache_free(hist->dccprxh_slab, entry);
-}
-
-extern void dccp_rx_hist_purge(struct dccp_rx_hist *hist,
- struct list_head *list);
-
static inline struct dccp_rx_hist_entry *
- dccp_rx_hist_head(struct list_head *list)
+ dccp_rx_hist_head(struct list_head *list)
{
struct dccp_rx_hist_entry *head = NULL;
@@ -177,6 +168,27 @@ static inline struct dccp_rx_hist_entry *
return head;
}
+extern int dccp_rx_hist_find_entry(const struct list_head *list, const u64 seq,
+ u8 *ccval);
+extern struct dccp_rx_hist_entry *
+ dccp_rx_hist_find_data_packet(const struct list_head *list);
+
+extern void dccp_rx_hist_add_packet(struct dccp_rx_hist *hist,
+ struct list_head *rx_list,
+ struct list_head *li_list,
+ struct dccp_rx_hist_entry *packet,
+ u64 nonloss_seqno);
+
+static inline void dccp_rx_hist_entry_delete(struct dccp_rx_hist *hist,
+ struct dccp_rx_hist_entry *entry)
+{
+ if (entry != NULL)
+ kmem_cache_free(hist->dccprxh_slab, entry);
+}
+
+extern void dccp_rx_hist_purge(struct dccp_rx_hist *hist,
+ struct list_head *list);
+
static inline int
dccp_rx_hist_entry_data_packet(const struct dccp_rx_hist_entry *entry)
{
@@ -184,12 +196,6 @@ static inline int
entry->dccphrx_type == DCCP_PKT_DATAACK;
}
-extern void dccp_rx_hist_add_packet(struct dccp_rx_hist *hist,
- struct list_head *rx_list,
- struct list_head *li_list,
- struct dccp_rx_hist_entry *packet,
- u64 nonloss_seqno);
-
extern u64 dccp_rx_hist_detect_loss(struct list_head *rx_list,
struct list_head *li_list, u8 *win_loss);
diff --git a/net/dccp/ccids/lib/tfrc.h b/net/dccp/ccids/lib/tfrc.h
index 45f30f59ea2..faf5f7e219e 100644
--- a/net/dccp/ccids/lib/tfrc.h
+++ b/net/dccp/ccids/lib/tfrc.h
@@ -13,8 +13,29 @@
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*/
-
#include <linux/types.h>
+#include <asm/div64.h>
+
+/* integer-arithmetic divisions of type (a * 1000000)/b */
+static inline u64 scaled_div(u64 a, u32 b)
+{
+ BUG_ON(b==0);
+ a *= 1000000;
+ do_div(a, b);
+ return a;
+}
+
+static inline u32 scaled_div32(u64 a, u32 b)
+{
+ u64 result = scaled_div(a, b);
+
+ if (result > UINT_MAX) {
+ DCCP_CRIT("Overflow: a(%llu)/b(%u) > ~0U",
+ (unsigned long long)a, b);
+ return UINT_MAX;
+ }
+ return result;
+}
extern u32 tfrc_calc_x(u16 s, u32 R, u32 p);
extern u32 tfrc_calc_x_reverse_lookup(u32 fvalue);
diff --git a/net/dccp/ccids/lib/tfrc_equation.c b/net/dccp/ccids/lib/tfrc_equation.c
index 44076e0c659..90009fd77e1 100644
--- a/net/dccp/ccids/lib/tfrc_equation.c
+++ b/net/dccp/ccids/lib/tfrc_equation.c
@@ -13,16 +13,83 @@
*/
#include <linux/module.h>
-
-#include <asm/div64.h>
-
+#include "../../dccp.h"
#include "tfrc.h"
#define TFRC_CALC_X_ARRSIZE 500
+#define TFRC_CALC_X_SPLIT 50000 /* 0.05 * 1000000, details below */
+#define TFRC_SMALLEST_P (TFRC_CALC_X_SPLIT/TFRC_CALC_X_ARRSIZE)
-#define TFRC_CALC_X_SPLIT 50000
-/* equivalent to 0.05 */
-
+/*
+ TFRC TCP Reno Throughput Equation Lookup Table for f(p)
+
+ The following two-column lookup table implements a part of the TCP throughput
+ equation from [RFC 3448, sec. 3.1]:
+
+ s
+ X_calc = --------------------------------------------------------------
+ R * sqrt(2*b*p/3) + (3 * t_RTO * sqrt(3*b*p/8) * (p + 32*p^3))
+
+ Where:
+ X is the transmit rate in bytes/second
+ s is the packet size in bytes
+ R is the round trip time in seconds
+ p is the loss event rate, between 0 and 1.0, of the number of loss
+ events as a fraction of the number of packets transmitted
+ t_RTO is the TCP retransmission timeout value in seconds
+ b is the number of packets acknowledged by a single TCP ACK
+
+ We can assume that b = 1 and t_RTO is 4 * R. The equation now becomes:
+
+ s
+ X_calc = -------------------------------------------------------
+ R * sqrt(p*2/3) + (12 * R * sqrt(p*3/8) * (p + 32*p^3))
+
+ which we can break down into:
+
+ s
+ X_calc = ---------
+ R * f(p)
+
+ where f(p) is given for 0 < p <= 1 by:
+
+ f(p) = sqrt(2*p/3) + 12 * sqrt(3*p/8) * (p + 32*p^3)
+
+ Since this is kernel code, floating-point arithmetic is avoided in favour of
+ integer arithmetic. This means that nearly all fractional parameters are
+ scaled by 1000000:
+ * the parameters p and R
+ * the return result f(p)
+ The lookup table therefore actually tabulates the following function g(q):
+
+ g(q) = 1000000 * f(q/1000000)
+
+ Hence, when p <= 1, q must be less than or equal to 1000000. To achieve finer
+ granularity for the practically more relevant case of small values of p (up to
+ 5%), the second column is used; the first one ranges up to 100%. This split
+ corresponds to the value of q = TFRC_CALC_X_SPLIT. At the same time this also
+ determines the smallest resolution possible with this lookup table:
+
+ TFRC_SMALLEST_P = TFRC_CALC_X_SPLIT / TFRC_CALC_X_ARRSIZE
+
+ The entire table is generated by:
+ for(i=0; i < TFRC_CALC_X_ARRSIZE; i++) {
+ lookup[i][0] = g((i+1) * 1000000/TFRC_CALC_X_ARRSIZE);
+ lookup[i][1] = g((i+1) * TFRC_CALC_X_SPLIT/TFRC_CALC_X_ARRSIZE);
+ }
+
+ With the given configuration, we have, with M = TFRC_CALC_X_ARRSIZE-1,
+ lookup[0][0] = g(1000000/(M+1)) = 1000000 * f(0.2%)
+ lookup[M][0] = g(1000000) = 1000000 * f(100%)
+ lookup[0][1] = g(TFRC_SMALLEST_P) = 1000000 * f(0.01%)
+ lookup[M][1] = g(TFRC_CALC_X_SPLIT) = 1000000 * f(5%)
+
+ In summary, the two columns represent f(p) for the following ranges:
+ * The first column is for 0.002 <= p <= 1.0
+ * The second column is for 0.0001 <= p <= 0.05
+ Where the columns overlap, the second (finer-grained) is given preference,
+ i.e. the first column is used only for p >= 0.05.
+ */
static const u32 tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE][2] = {
{ 37172, 8172 },
{ 53499, 11567 },
@@ -526,117 +593,105 @@ static const u32 tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE][2] = {
{ 243315981, 271305 }
};
-/* Calculate the send rate as per section 3.1 of RFC3448
-
-Returns send rate in bytes per second
-
-Integer maths and lookups are used as not allowed floating point in kernel
-
-The function for Xcalc as per section 3.1 of RFC3448 is:
-
-X = s
- -------------------------------------------------------------
- R*sqrt(2*b*p/3) + (t_RTO * (3*sqrt(3*b*p/8) * p * (1+32*p^2)))
-
-where
-X is the trasmit rate in bytes/second
-s is the packet size in bytes
-R is the round trip time in seconds
-p is the loss event rate, between 0 and 1.0, of the number of loss events
- as a fraction of the number of packets transmitted
-t_RTO is the TCP retransmission timeout value in seconds
-b is the number of packets acknowledged by a single TCP acknowledgement
-
-we can assume that b = 1 and t_RTO is 4 * R. With this the equation becomes:
-
-X = s
- -----------------------------------------------------------------------
- R * sqrt(2 * p / 3) + (12 * R * (sqrt(3 * p / 8) * p * (1 + 32 * p^2)))
-
-
-which we can break down into:
-
-X = s
- --------
- R * f(p)
-
-where f(p) = sqrt(2 * p / 3) + (12 * sqrt(3 * p / 8) * p * (1 + 32 * p * p))
-
-Function parameters:
-s - bytes
-R - RTT in usecs
-p - loss rate (decimal fraction multiplied by 1,000,000)
-
-Returns Xcalc in bytes per second
-
-DON'T alter this code unless you run test cases against it as the code
-has been manipulated to stop underflow/overlow.
+/* return largest index i such that fval <= lookup[i][small] */
+static inline u32 tfrc_binsearch(u32 fval, u8 small)
+{
+ u32 try, low = 0, high = TFRC_CALC_X_ARRSIZE - 1;
+
+ while (low < high) {
+ try = (low + high) / 2;
+ if (fval <= tfrc_calc_x_lookup[try][small])
+ high = try;
+ else
+ low = try + 1;
+ }
+ return high;
+}
-*/
+/**
+ * tfrc_calc_x - Calculate the send rate as per section 3.1 of RFC3448
+ *
+ * @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)
{
- int index;
+ u16 index;
u32 f;
- u64 tmp1, tmp2;
+ u64 result;
- if (p < TFRC_CALC_X_SPLIT)
- index = (p / (TFRC_CALC_X_SPLIT / TFRC_CALC_X_ARRSIZE)) - 1;
- else
- index = (p / (1000000 / TFRC_CALC_X_ARRSIZE)) - 1;
+ /* check against invalid parameters and divide-by-zero */
+ BUG_ON(p > 1000000); /* p must not exceed 100% */
+ BUG_ON(p == 0); /* f(0) = 0, divide by zero */
+ if (R == 0) { /* possible divide by zero */
+ DCCP_CRIT("WARNING: RTT is 0, returning maximum X_calc.");
+ return ~0U;
+ }
- if (index < 0)
- /* p should be 0 unless there is a bug in my code */
- index = 0;
+ if (p <= TFRC_CALC_X_SPLIT) { /* 0.0000 < p <= 0.05 */
+ if (p < TFRC_SMALLEST_P) { /* 0.0000 < p < 0.0001 */
+ DCCP_WARN("Value of p (%d) below resolution. "
+ "Substituting %d\n", p, TFRC_SMALLEST_P);
+ index = 0;
+ } else /* 0.0001 <= p <= 0.05 */
+ index = p/TFRC_SMALLEST_P - 1;
- if (R == 0)
- R = 1; /* RTT can't be zero or else divide by zero */
+ f = tfrc_calc_x_lookup[index][1];
- BUG_ON(index >= TFRC_CALC_X_ARRSIZE);
+ } else { /* 0.05 < p <= 1.00 */
+ index = p/(1000000/TFRC_CALC_X_ARRSIZE) - 1;
- if (p >= TFRC_CALC_X_SPLIT)
f = tfrc_calc_x_lookup[index][0];
- else
- f = tfrc_calc_x_lookup[index][1];
-
- tmp1 = ((u64)s * 100000000);
- tmp2 = ((u64)R * (u64)f);
- do_div(tmp2, 10000);
- do_div(tmp1, tmp2);
- /* Don't alter above math unless you test due to overflow on 32 bit */
-
- return (u32)tmp1;
+ }
+
+ /*
+ * Compute X = s/(R*f(p)) in bytes per second.
+ * Since f(p) and R are both scaled by 1000000, we need to multiply by
+ * 1000000^2. To avoid overflow, the result is computed in two stages.
+ * This works under almost all reasonable operational conditions, for a
+ * wide range of parameters. Yet, should some strange combination of
+ * parameters result in overflow, the use of scaled_div32 will catch
+ * this and return UINT_MAX - which is a logically adequate consequence.
+ */
+ result = scaled_div(s, R);
+ return scaled_div32(result, f);
}
EXPORT_SYMBOL_GPL(tfrc_calc_x);
/*
- * args: fvalue - function value to match
- * returns: p closest to that value
+ * tfrc_calc_x_reverse_lookup - try to find p given f(p)
*
- * both fvalue and p are multiplied by 1,000,000 to use ints
+ * @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)
{
- int ctr = 0;
- int small;
+ int index;
- if (fvalue < tfrc_calc_x_lookup[0][1])
+ if (fvalue == 0) /* f(p) = 0 whenever p = 0 */
return 0;
- if (fvalue <= tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE - 1][1])
- small = 1;
- else if (fvalue > tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE - 1][0])
+ /* Error cases. */
+ if (fvalue < tfrc_calc_x_lookup[0][1]) {
+ DCCP_WARN("fvalue %d smaller than resolution\n", fvalue);
+ return tfrc_calc_x_lookup[0][1];
+ }
+ if (fvalue > tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE - 1][0]) {
+ DCCP_WARN("fvalue %d exceeds bounds!\n", fvalue);
return 1000000;
- else
- small = 0;
-
- while (fvalue > tfrc_calc_x_lookup[ctr][small])
- ctr++;
+ }
- if (small)
- return TFRC_CALC_X_SPLIT * ctr / TFRC_CALC_X_ARRSIZE;
- else
- return 1000000 * ctr / TFRC_CALC_X_ARRSIZE;
+ if (fvalue <= tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE - 1][1]) {
+ index = tfrc_binsearch(fvalue, 1);
+ return (index + 1) * TFRC_CALC_X_SPLIT / TFRC_CALC_X_ARRSIZE;
+ }
+
+ /* else ... it must be in the coarse-grained column */
+ index = tfrc_binsearch(fvalue, 0);
+ return (index + 1) * 1000000 / TFRC_CALC_X_ARRSIZE;
}
EXPORT_SYMBOL_GPL(tfrc_calc_x_reverse_lookup);
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-11-25 07:42:28 +0000