8 files changed, 902 insertions, 684 deletions

diff --git a/net/dccp/ccids/lib/Makefile b/net/dccp/ccids/lib/Makefile
deleted file mode 100644
index 5f940a6cbac..00000000000
--- a/net/dccp/ccids/lib/Makefile
+++ /dev/null
@@ -1,3 +0,0 @@
-obj-$(CONFIG_IP_DCCP_TFRC_LIB) += dccp_tfrc_lib.o
-
-dccp_tfrc_lib-y := loss_interval.o packet_history.o tfrc_equation.o
diff --git a/net/dccp/ccids/lib/loss_interval.c b/net/dccp/ccids/lib/loss_interval.c
index 4c01a54143a..57f9fd78c4d 100644
--- a/net/dccp/ccids/lib/loss_interval.c
+++ b/net/dccp/ccids/lib/loss_interval.c
@@ -1,8 +1,7 @@
 /*
- *  net/dccp/ccids/lib/loss_interval.c
- *
- *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
- *  Copyright (c) 2005 Ian McDonald <iam4@cs.waikato.ac.nz>
+ *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
+ *  Copyright (c) 2005-7 The University of Waikato, Hamilton, New Zealand.
+ *  Copyright (c) 2005-7 Ian McDonald <ian.mcdonald@jandi.co.nz>
  *  Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
  *
  *  This program is free software; you can redistribute it and/or modify
@@ -10,135 +9,177 @@
  *  the Free Software Foundation; either version 2 of the License, or
  *  (at your option) any later version.
  */
+#include <net/sock.h>
+#include "tfrc.h"
 
-#include <linux/config.h>
-#include <linux/module.h>
-
-#include "loss_interval.h"
+static struct kmem_cache  *tfrc_lh_slab  __read_mostly;
+/* Loss Interval weights from [RFC 3448, 5.4], scaled by 10 */
+static const int tfrc_lh_weights[NINTERVAL] = { 10, 10, 10, 10, 8, 6, 4, 2 };
 
-struct dccp_li_hist *dccp_li_hist_new(const char *name)
+/* implements LIFO semantics on the array */
+static inline u8 LIH_INDEX(const u8 ctr)
 {
-	struct dccp_li_hist *hist = kmalloc(sizeof(*hist), GFP_ATOMIC);
-	static const char dccp_li_hist_mask[] = "li_hist_%s";
-	char *slab_name;
-
-	if (hist == NULL)
-		goto out;
-
-	slab_name = kmalloc(strlen(name) + sizeof(dccp_li_hist_mask) - 1,
-			    GFP_ATOMIC);
-	if (slab_name == NULL)
-		goto out_free_hist;
-
-	sprintf(slab_name, dccp_li_hist_mask, name);
-	hist->dccplih_slab = kmem_cache_create(slab_name,
-					     sizeof(struct dccp_li_hist_entry),
-					       0, SLAB_HWCACHE_ALIGN,
-					       NULL, NULL);
-	if (hist->dccplih_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;
+	return LIH_SIZE - 1 - (ctr % LIH_SIZE);
 }
 
-EXPORT_SYMBOL_GPL(dccp_li_hist_new);
+/* the `counter' index always points at the next entry to be populated */
+static inline struct tfrc_loss_interval *tfrc_lh_peek(struct tfrc_loss_hist *lh)
+{
+	return lh->counter ? lh->ring[LIH_INDEX(lh->counter - 1)] : NULL;
+}
 
-void dccp_li_hist_delete(struct dccp_li_hist *hist)
+/* given i with 0 <= i <= k, return I_i as per the rfc3448bis notation */
+static inline u32 tfrc_lh_get_interval(struct tfrc_loss_hist *lh, const u8 i)
 {
-	const char* name = kmem_cache_name(hist->dccplih_slab);
+	BUG_ON(i >= lh->counter);
+	return lh->ring[LIH_INDEX(lh->counter - i - 1)]->li_length;
+}
 
-	kmem_cache_destroy(hist->dccplih_slab);
-	kfree(name);
-	kfree(hist);
+/*
+ *	On-demand allocation and de-allocation of entries
+ */
+static struct tfrc_loss_interval *tfrc_lh_demand_next(struct tfrc_loss_hist *lh)
+{
+	if (lh->ring[LIH_INDEX(lh->counter)] == NULL)
+		lh->ring[LIH_INDEX(lh->counter)] = kmem_cache_alloc(tfrc_lh_slab,
+								    GFP_ATOMIC);
+	return lh->ring[LIH_INDEX(lh->counter)];
 }
 
-EXPORT_SYMBOL_GPL(dccp_li_hist_delete);
+void tfrc_lh_cleanup(struct tfrc_loss_hist *lh)
+{
+	if (!tfrc_lh_is_initialised(lh))
+		return;
+
+	for (lh->counter = 0; lh->counter < LIH_SIZE; lh->counter++)
+		if (lh->ring[LIH_INDEX(lh->counter)] != NULL) {
+			kmem_cache_free(tfrc_lh_slab,
+					lh->ring[LIH_INDEX(lh->counter)]);
+			lh->ring[LIH_INDEX(lh->counter)] = NULL;
+		}
+}
 
-void dccp_li_hist_purge(struct dccp_li_hist *hist, struct list_head *list)
+static void tfrc_lh_calc_i_mean(struct tfrc_loss_hist *lh)
 {
-	struct dccp_li_hist_entry *entry, *next;
+	u32 i_i, i_tot0 = 0, i_tot1 = 0, w_tot = 0;
+	int i, k = tfrc_lh_length(lh) - 1; /* k is as in rfc3448bis, 5.4 */
+
+	if (k <= 0)
+		return;
+
+	for (i = 0; i <= k; i++) {
+		i_i = tfrc_lh_get_interval(lh, i);
 
-	list_for_each_entry_safe(entry, next, list, dccplih_node) {
-		list_del_init(&entry->dccplih_node);
-		kmem_cache_free(hist->dccplih_slab, entry);
+		if (i < k) {
+			i_tot0 += i_i * tfrc_lh_weights[i];
+			w_tot  += tfrc_lh_weights[i];
+		}
+		if (i > 0)
+			i_tot1 += i_i * tfrc_lh_weights[i-1];
 	}
-}
 
-EXPORT_SYMBOL_GPL(dccp_li_hist_purge);
+	lh->i_mean = max(i_tot0, i_tot1) / w_tot;
+}
 
-/* Weights used to calculate loss event rate */
-/*
- * These are integers as per section 8 of RFC3448. We can then divide by 4 *
- * when we use it.
+/**
+ * tfrc_lh_update_i_mean  -  Update the `open' loss interval I_0
+ * For recomputing p: returns `true' if p > p_prev  <=>  1/p < 1/p_prev
  */
-static const int dccp_li_hist_w[DCCP_LI_HIST_IVAL_F_LENGTH] = {
-	4, 4, 4, 4, 3, 2, 1, 1,
-};
-
-u32 dccp_li_hist_calc_i_mean(struct list_head *list)
+u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *skb)
 {
-	struct dccp_li_hist_entry *li_entry, *li_next;
-	int i = 0;
-	u32 i_tot;
-	u32 i_tot0 = 0;
-	u32 i_tot1 = 0;
-	u32 w_tot  = 0;
-
-	list_for_each_entry_safe(li_entry, li_next, list, dccplih_node) {
-		if (i < DCCP_LI_HIST_IVAL_F_LENGTH) {
-			i_tot0 += li_entry->dccplih_interval * dccp_li_hist_w[i];
-			w_tot  += dccp_li_hist_w[i];
-		}
+	struct tfrc_loss_interval *cur = tfrc_lh_peek(lh);
+	u32 old_i_mean = lh->i_mean;
+	s64 len;
 
-		if (i != 0)
-			i_tot1 += li_entry->dccplih_interval * dccp_li_hist_w[i - 1];
+	if (cur == NULL)			/* not initialised */
+		return 0;
 
-		if (++i > DCCP_LI_HIST_IVAL_F_LENGTH)
-			break;
-	}
+	len = dccp_delta_seqno(cur->li_seqno, DCCP_SKB_CB(skb)->dccpd_seq) + 1;
 
-	if (i != DCCP_LI_HIST_IVAL_F_LENGTH)
+	if (len - (s64)cur->li_length <= 0)	/* duplicate or reordered */
 		return 0;
 
-	i_tot = max(i_tot0, i_tot1);
+	if (SUB16(dccp_hdr(skb)->dccph_ccval, cur->li_ccval) > 4)
+		/*
+		 * Implements RFC 4342, 10.2:
+		 * If a packet S (skb) exists whose seqno comes `after' the one
+		 * starting the current loss interval (cur) and if the modulo-16
+		 * distance from C(cur) to C(S) is greater than 4, consider all
+		 * subsequent packets as belonging to a new loss interval. This
+		 * test is necessary since CCVal may wrap between intervals.
+		 */
+		cur->li_is_closed = 1;
+
+	if (tfrc_lh_length(lh) == 1)		/* due to RFC 3448, 6.3.1 */
+		return 0;
 
-	/* FIXME: Why do we do this? -Ian McDonald */
-	if (i_tot * 4 < w_tot)
-		i_tot = w_tot * 4;
+	cur->li_length = len;
+	tfrc_lh_calc_i_mean(lh);
 
-	return i_tot * 4 / w_tot;
+	return lh->i_mean < old_i_mean;
 }
 
-EXPORT_SYMBOL_GPL(dccp_li_hist_calc_i_mean);
+/* Determine if `new_loss' does begin a new loss interval [RFC 4342, 10.2] */
+static inline u8 tfrc_lh_is_new_loss(struct tfrc_loss_interval *cur,
+				     struct tfrc_rx_hist_entry *new_loss)
+{
+	return	dccp_delta_seqno(cur->li_seqno, new_loss->tfrchrx_seqno) > 0 &&
+		(cur->li_is_closed || SUB16(new_loss->tfrchrx_ccval, cur->li_ccval) > 4);
+}
 
-struct dccp_li_hist_entry *dccp_li_hist_interval_new(struct dccp_li_hist *hist,
-						     struct list_head *list,
-						     const u64 seq_loss,
-						     const u8 win_loss)
+/**
+ * tfrc_lh_interval_add  -  Insert new record into the Loss Interval database
+ * @lh:		   Loss Interval database
+ * @rh:		   Receive history containing a fresh loss event
+ * @calc_first_li: Caller-dependent routine to compute length of first interval
+ * @sk:		   Used by @calc_first_li in caller-specific way (subtyping)
+ *
+ * Updates I_mean and returns 1 if a new interval has in fact been added to @lh.
+ */
+int tfrc_lh_interval_add(struct tfrc_loss_hist *lh, struct tfrc_rx_hist *rh,
+			 u32 (*calc_first_li)(struct sock *), struct sock *sk)
 {
-	struct dccp_li_hist_entry *tail = NULL, *entry;
-	int i;
-
-	for (i = 0; i <= DCCP_LI_HIST_IVAL_F_LENGTH; ++i) {
-		entry = dccp_li_hist_entry_new(hist, SLAB_ATOMIC);
-		if (entry == NULL) {
-			dccp_li_hist_purge(hist, list);
-			return NULL;
-		}
-		if (tail == NULL)
-			tail = entry;
-		list_add(&entry->dccplih_node, list);
+	struct tfrc_loss_interval *cur = tfrc_lh_peek(lh), *new;
+
+	if (cur != NULL && !tfrc_lh_is_new_loss(cur, tfrc_rx_hist_loss_prev(rh)))
+		return 0;
+
+	new = tfrc_lh_demand_next(lh);
+	if (unlikely(new == NULL)) {
+		DCCP_CRIT("Cannot allocate/add loss record.");
+		return 0;
+	}
+
+	new->li_seqno	  = tfrc_rx_hist_loss_prev(rh)->tfrchrx_seqno;
+	new->li_ccval	  = tfrc_rx_hist_loss_prev(rh)->tfrchrx_ccval;
+	new->li_is_closed = 0;
+
+	if (++lh->counter == 1)
+		lh->i_mean = new->li_length = (*calc_first_li)(sk);
+	else {
+		cur->li_length = dccp_delta_seqno(cur->li_seqno, new->li_seqno);
+		new->li_length = dccp_delta_seqno(new->li_seqno,
+				  tfrc_rx_hist_last_rcv(rh)->tfrchrx_seqno) + 1;
+		if (lh->counter > (2*LIH_SIZE))
+			lh->counter -= LIH_SIZE;
+
+		tfrc_lh_calc_i_mean(lh);
 	}
+	return 1;
+}
 
-	entry->dccplih_seqno     = seq_loss;
-	entry->dccplih_win_count = win_loss;
-	return tail;
+int __init tfrc_li_init(void)
+{
+	tfrc_lh_slab = kmem_cache_create("tfrc_li_hist",
+					 sizeof(struct tfrc_loss_interval), 0,
+					 SLAB_HWCACHE_ALIGN, NULL);
+	return tfrc_lh_slab == NULL ? -ENOBUFS : 0;
 }
 
-EXPORT_SYMBOL_GPL(dccp_li_hist_interval_new);
+void tfrc_li_exit(void)
+{
+	if (tfrc_lh_slab != NULL) {
+		kmem_cache_destroy(tfrc_lh_slab);
+		tfrc_lh_slab = NULL;
+	}
+}
diff --git a/net/dccp/ccids/lib/loss_interval.h b/net/dccp/ccids/lib/loss_interval.h
index 417d9d82df3..57f631a86cc 100644
--- a/net/dccp/ccids/lib/loss_interval.h
+++ b/net/dccp/ccids/lib/loss_interval.h
@@ -1,10 +1,9 @@
 #ifndef _DCCP_LI_HIST_
 #define _DCCP_LI_HIST_
 /*
- *  net/dccp/ccids/lib/loss_interval.h
- *
- *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
- *  Copyright (c) 2005 Ian McDonald <iam4@cs.waikato.ac.nz>
+ *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
+ *  Copyright (c) 2005-7 The University of Waikato, Hamilton, New Zealand.
+ *  Copyright (c) 2005-7 Ian McDonald <ian.mcdonald@jandi.co.nz>
  *  Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
  *
  *  This program is free software; you can redistribute it and/or modify it
@@ -12,50 +11,63 @@
  *  Software Foundation; either version 2 of the License, or (at your option)
  *  any later version.
  */
-
-#include <linux/config.h>
+#include <linux/ktime.h>
 #include <linux/list.h>
 #include <linux/slab.h>
-#include <linux/time.h>
 
-#define DCCP_LI_HIST_IVAL_F_LENGTH  8
+/*
+ * Number of loss intervals (RFC 4342, 8.6.1). The history size is one more than
+ * NINTERVAL, since the `open' interval I_0 is always stored as the first entry.
+ */
+#define NINTERVAL	8
+#define LIH_SIZE	(NINTERVAL + 1)
 
-struct dccp_li_hist {
-	kmem_cache_t *dccplih_slab;
+/**
+ *  tfrc_loss_interval  -  Loss history record for TFRC-based protocols
+ *  @li_seqno:		Highest received seqno before the start of loss
+ *  @li_ccval:		The CCVal belonging to @li_seqno
+ *  @li_is_closed:	Whether @li_seqno is older than 1 RTT
+ *  @li_length:		Loss interval sequence length
+ */
+struct tfrc_loss_interval {
+	u64		 li_seqno:48,
+			 li_ccval:4,
+			 li_is_closed:1;
+	u32		 li_length;
 };
 
-extern struct dccp_li_hist *dccp_li_hist_new(const char *name);
-extern void dccp_li_hist_delete(struct dccp_li_hist *hist);
-
-struct dccp_li_hist_entry {
-	struct list_head dccplih_node;
-	u64		 dccplih_seqno:48,
-			 dccplih_win_count:4;
-	u32		 dccplih_interval;
+/**
+ *  tfrc_loss_hist  -  Loss record database
+ *  @ring:	Circular queue managed in LIFO manner
+ *  @counter:	Current count of entries (can be more than %LIH_SIZE)
+ *  @i_mean:	Current Average Loss Interval [RFC 3448, 5.4]
+ */
+struct tfrc_loss_hist {
+	struct tfrc_loss_interval	*ring[LIH_SIZE];
+	u8				counter;
+	u32				i_mean;
 };
 
-static inline struct dccp_li_hist_entry *
-		dccp_li_hist_entry_new(struct dccp_li_hist *hist,
-				       const gfp_t prio)
+static inline void tfrc_lh_init(struct tfrc_loss_hist *lh)
+{
+	memset(lh, 0, sizeof(struct tfrc_loss_hist));
+}
+
+static inline u8 tfrc_lh_is_initialised(struct tfrc_loss_hist *lh)
 {
-	return kmem_cache_alloc(hist->dccplih_slab, prio);
+	return lh->counter > 0;
 }
 
-static inline void dccp_li_hist_entry_delete(struct dccp_li_hist *hist,
-					     struct dccp_li_hist_entry *entry)
+static inline u8 tfrc_lh_length(struct tfrc_loss_hist *lh)
 {
-	if (entry != NULL)
-		kmem_cache_free(hist->dccplih_slab, entry);
+	return min(lh->counter, (u8)LIH_SIZE);
 }
 
-extern void dccp_li_hist_purge(struct dccp_li_hist *hist,
-			       struct list_head *list);
+struct tfrc_rx_hist;
 
-extern u32 dccp_li_hist_calc_i_mean(struct list_head *list);
+int tfrc_lh_interval_add(struct tfrc_loss_hist *, struct tfrc_rx_hist *,
+			 u32 (*first_li)(struct sock *), struct sock *);
+u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *);
+void tfrc_lh_cleanup(struct tfrc_loss_hist *lh);
 
-extern struct dccp_li_hist_entry *
-			dccp_li_hist_interval_new(struct dccp_li_hist *hist,
-						  struct list_head *list,
-						  const u64 seq_loss,
-						  const u8 win_loss);
 #endif /* _DCCP_LI_HIST_ */
diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c
index d3f9d205383..08df7a3acb3 100644
--- a/net/dccp/ccids/lib/packet_history.c
+++ b/net/dccp/ccids/lib/packet_history.c
@@ -1,13 +1,12 @@
 /*
- *  net/dccp/packet_history.h
- *
- *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
+ *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
+ *  Copyright (c) 2005-7 The University of Waikato, Hamilton, New Zealand.
  *
  *  An implementation of the DCCP protocol
  *
  *  This code has been developed by the University of Waikato WAND
  *  research group. For further information please see http://www.wand.net.nz/
- *  or e-mail Ian McDonald - iam4@cs.waikato.ac.nz
+ *  or e-mail Ian McDonald - ian.mcdonald@jandi.co.nz
  *
  *  This code also uses code from Lulea University, rereleased as GPL by its
  *  authors:
@@ -34,365 +33,417 @@
  *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  */
 
-#include <linux/config.h>
-#include <linux/module.h>
 #include <linux/string.h>
-
+#include <linux/slab.h>
 #include "packet_history.h"
+#include "../../dccp.h"
 
-struct dccp_rx_hist *dccp_rx_hist_new(const char *name)
+/*
+ * Transmitter History Routines
+ */
+static struct kmem_cache *tfrc_tx_hist_slab;
+
+int __init tfrc_tx_packet_history_init(void)
 {
-	struct dccp_rx_hist *hist = kmalloc(sizeof(*hist), GFP_ATOMIC);
-	static const char dccp_rx_hist_mask[] = "rx_hist_%s";
-	char *slab_name;
-
-	if (hist == NULL)
-		goto out;
-
-	slab_name = kmalloc(strlen(name) + sizeof(dccp_rx_hist_mask) - 1,
-			    GFP_ATOMIC);
-	if (slab_name == NULL)
-		goto out_free_hist;
-
-	sprintf(slab_name, dccp_rx_hist_mask, name);
-	hist->dccprxh_slab = kmem_cache_create(slab_name,
-					     sizeof(struct dccp_rx_hist_entry),
-					       0, SLAB_HWCACHE_ALIGN,
-					       NULL, NULL);
-	if (hist->dccprxh_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;
+	tfrc_tx_hist_slab = kmem_cache_create("tfrc_tx_hist",
+					      sizeof(struct tfrc_tx_hist_entry),
+					      0, SLAB_HWCACHE_ALIGN, NULL);
+	return tfrc_tx_hist_slab == NULL ? -ENOBUFS : 0;
 }
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_new);
+void tfrc_tx_packet_history_exit(void)
+{
+	if (tfrc_tx_hist_slab != NULL) {
+		kmem_cache_destroy(tfrc_tx_hist_slab);
+		tfrc_tx_hist_slab = NULL;
+	}
+}
+
+int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno)
+{
+	struct tfrc_tx_hist_entry *entry = kmem_cache_alloc(tfrc_tx_hist_slab, gfp_any());
+
+	if (entry == NULL)
+		return -ENOBUFS;
+	entry->seqno = seqno;
+	entry->stamp = ktime_get_real();
+	entry->next  = *headp;
+	*headp	     = entry;
+	return 0;
+}
 
-void dccp_rx_hist_delete(struct dccp_rx_hist *hist)
+void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp)
 {
-	const char* name = kmem_cache_name(hist->dccprxh_slab);
+	struct tfrc_tx_hist_entry *head = *headp;
+
+	while (head != NULL) {
+		struct tfrc_tx_hist_entry *next = head->next;
 
-	kmem_cache_destroy(hist->dccprxh_slab);
-	kfree(name);
-	kfree(hist);
+		kmem_cache_free(tfrc_tx_hist_slab, head);
+		head = next;
+	}
+
+	*headp = NULL;
 }
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_delete);
+/*
+ *	Receiver History Routines
+ */
+static struct kmem_cache *tfrc_rx_hist_slab;
 
-void dccp_rx_hist_purge(struct dccp_rx_hist *hist, struct list_head *list)
+int __init tfrc_rx_packet_history_init(void)
 {
-	struct dccp_rx_hist_entry *entry, *next;
+	tfrc_rx_hist_slab = kmem_cache_create("tfrc_rxh_cache",
+					      sizeof(struct tfrc_rx_hist_entry),
+					      0, SLAB_HWCACHE_ALIGN, NULL);
+	return tfrc_rx_hist_slab == NULL ? -ENOBUFS : 0;
+}
 
-	list_for_each_entry_safe(entry, next, list, dccphrx_node) {
-		list_del_init(&entry->dccphrx_node);
-		kmem_cache_free(hist->dccprxh_slab, entry);
+void tfrc_rx_packet_history_exit(void)
+{
+	if (tfrc_rx_hist_slab != NULL) {
+		kmem_cache_destroy(tfrc_rx_hist_slab);
+		tfrc_rx_hist_slab = NULL;
 	}
 }
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_purge);
+static inline void tfrc_rx_hist_entry_from_skb(struct tfrc_rx_hist_entry *entry,
+					       const struct sk_buff *skb,
+					       const u64 ndp)
+{
+	const struct dccp_hdr *dh = dccp_hdr(skb);
+
+	entry->tfrchrx_seqno = DCCP_SKB_CB(skb)->dccpd_seq;
+	entry->tfrchrx_ccval = dh->dccph_ccval;
+	entry->tfrchrx_type  = dh->dccph_type;
+	entry->tfrchrx_ndp   = ndp;
+	entry->tfrchrx_tstamp = ktime_get_real();
+}
 
-struct dccp_rx_hist_entry *
-		dccp_rx_hist_find_data_packet(const struct list_head *list)
+void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h,
+			     const struct sk_buff *skb,
+			     const u64 ndp)
 {
-	struct dccp_rx_hist_entry *entry, *packet = NULL;
+	struct tfrc_rx_hist_entry *entry = tfrc_rx_hist_last_rcv(h);
 
-	list_for_each_entry(entry, list, dccphrx_node)
-		if (entry->dccphrx_type == DCCP_PKT_DATA ||
-		    entry->dccphrx_type == DCCP_PKT_DATAACK) {
-			packet = entry;
-			break;
-		}
+	tfrc_rx_hist_entry_from_skb(entry, skb, ndp);
+}
 
-	return packet;
+/* has the packet contained in skb been seen before? */
+int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb)
+{
+	const u64 seq = DCCP_SKB_CB(skb)->dccpd_seq;
+	int i;
+
+	if (dccp_delta_seqno(tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, seq) <= 0)
+		return 1;
+
+	for (i = 1; i <= h->loss_count; i++)
+		if (tfrc_rx_hist_entry(h, i)->tfrchrx_seqno == seq)
+			return 1;
+
+	return 0;
 }
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_find_data_packet);
+static void tfrc_rx_hist_swap(struct tfrc_rx_hist *h, const u8 a, const u8 b)
+{
+	const u8 idx_a = tfrc_rx_hist_index(h, a),
+		 idx_b = tfrc_rx_hist_index(h, b);
+	struct tfrc_rx_hist_entry *tmp = h->ring[idx_a];
+
+	h->ring[idx_a] = h->ring[idx_b];
+	h->ring[idx_b] = tmp;
+}
 
-int 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)
+/*
+ * Private helper functions for loss detection.
+ *
+ * In the descriptions, `Si' refers to the sequence number of entry number i,
+ * whose NDP count is `Ni' (lower case is used for variables).
+ * Note: All __xxx_loss functions expect that a test against duplicates has been
+ *       performed already: the seqno of the skb must not be less than the seqno
+ *       of loss_prev; and it must not equal that of any valid history entry.
+ */
+static void __do_track_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u64 n1)
 {
-	struct dccp_rx_hist_entry *entry, *next, *iter;
-	u8 num_later = 0;
-
-	iter = dccp_rx_hist_head(rx_list);
-	if (iter == NULL)
-		dccp_rx_hist_add_entry(rx_list, packet);
-	else {
-		const u64 seqno = packet->dccphrx_seqno;
-
-		if (after48(seqno, iter->dccphrx_seqno))
-			dccp_rx_hist_add_entry(rx_list, packet);
-		else {
-			if (dccp_rx_hist_entry_data_packet(iter))
-				num_later = 1;
-
-			list_for_each_entry_continue(iter, rx_list,
-						     dccphrx_node) {
-				if (after48(seqno, iter->dccphrx_seqno)) {
-					dccp_rx_hist_add_entry(&iter->dccphrx_node,
-							       packet);
-					goto trim_history;
-				}
-
-				if (dccp_rx_hist_entry_data_packet(iter))
-					num_later++;
-
-				if (num_later == TFRC_RECV_NUM_LATE_LOSS) {
-					dccp_rx_hist_entry_delete(hist, packet);
-					return 1;
-				}
-			}
+	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
+	    s1 = DCCP_SKB_CB(skb)->dccpd_seq;
 
-			if (num_later < TFRC_RECV_NUM_LATE_LOSS)
-				dccp_rx_hist_add_entry(rx_list, packet);
-			/*
-			 * FIXME: else what? should we destroy the packet
-			 * like above?
-			 */
-		}
+	if (!dccp_loss_free(s0, s1, n1)) {	/* gap between S0 and S1 */
+		h->loss_count = 1;
+		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n1);
 	}
+}
 
-trim_history:
-	/*
-	 * Trim history (remove all packets after the NUM_LATE_LOSS + 1
-	 * data packets)
-	 */
-	num_later = TFRC_RECV_NUM_LATE_LOSS + 1;
-
-	if (!list_empty(li_list)) {
-		list_for_each_entry_safe(entry, next, rx_list, dccphrx_node) {
-			if (num_later == 0) {
-				list_del_init(&entry->dccphrx_node);
-				dccp_rx_hist_entry_delete(hist, entry);
-			} else if (dccp_rx_hist_entry_data_packet(entry))
-				--num_later;
-		}
-	} else {
-		int step = 0;
-		u8 win_count = 0; /* Not needed, but lets shut up gcc */
-		int tmp;
+static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2)
+{
+	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
+	    s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
+	    s2 = DCCP_SKB_CB(skb)->dccpd_seq;
+
+	if (likely(dccp_delta_seqno(s1, s2) > 0)) {	/* S1  <  S2 */
+		h->loss_count = 2;
+		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n2);
+		return;
+	}
+
+	/* S0  <  S2  <  S1 */
+
+	if (dccp_loss_free(s0, s2, n2)) {
+		u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp;
+
+		if (dccp_loss_free(s2, s1, n1)) {
+			/* hole is filled: S0, S2, and S1 are consecutive */
+			h->loss_count = 0;
+			h->loss_start = tfrc_rx_hist_index(h, 1);
+		} else
+			/* gap between S2 and S1: just update loss_prev */
+			tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n2);
+
+	} else {	/* gap between S0 and S2 */
 		/*
-		 * We have no loss interval history so we need at least one
-		 * rtt:s of data packets to approximate rtt.
+		 * Reorder history to insert S2 between S0 and S1
 		 */
-		list_for_each_entry_safe(entry, next, rx_list, dccphrx_node) {
-			if (num_later == 0) {
-				switch (step) {
-				case 0:
-					step = 1;
-					/* OK, find next data packet */
-					num_later = 1;
-					break;
-				case 1:
-					step = 2;
-					/* OK, find next data packet */
-					num_later = 1;
-					win_count = entry->dccphrx_ccval;
-					break;
-				case 2:
-					tmp = win_count - entry->dccphrx_ccval;
-					if (tmp < 0)
-						tmp += TFRC_WIN_COUNT_LIMIT;
-					if (tmp > TFRC_WIN_COUNT_PER_RTT + 1) {
-						/*
-						 * We have found a packet older
-						 * than one rtt remove the rest
-						 */
-						step = 3;
-					} else /* OK, find next data packet */
-						num_later = 1;
-					break;
-				case 3:
-					list_del_init(&entry->dccphrx_node);
-					dccp_rx_hist_entry_delete(hist, entry);
-					break;
-				}
-			} else if (dccp_rx_hist_entry_data_packet(entry))
-				--num_later;
-		}
+		tfrc_rx_hist_swap(h, 0, 3);
+		h->loss_start = tfrc_rx_hist_index(h, 3);
+		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n2);
+		h->loss_count = 2;
 	}
-
-	return 0;
 }
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_add_packet);
-
-u64 dccp_rx_hist_detect_loss(struct list_head *rx_list,
-			     struct list_head *li_list, u8 *win_loss)
+/* return 1 if a new loss event has been identified */
+static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3)
 {
-	struct dccp_rx_hist_entry *entry, *next, *packet;
-	struct dccp_rx_hist_entry *a_loss = NULL;
-	struct dccp_rx_hist_entry *b_loss = NULL;
-	u64 seq_loss = DCCP_MAX_SEQNO + 1;
-	u8 num_later = TFRC_RECV_NUM_LATE_LOSS;
-
-	list_for_each_entry_safe(entry, next, rx_list, dccphrx_node) {
-		if (num_later == 0) {
-			b_loss = entry;
-			break;
-		} else if (dccp_rx_hist_entry_data_packet(entry))
-			--num_later;
+	u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno,
+	    s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
+	    s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno,
+	    s3 = DCCP_SKB_CB(skb)->dccpd_seq;
+
+	if (likely(dccp_delta_seqno(s2, s3) > 0)) {	/* S2  <  S3 */
+		h->loss_count = 3;
+		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 3), skb, n3);
+		return 1;
 	}
 
-	if (b_loss == NULL)
-		goto out;
+	/* S3  <  S2 */
 
-	num_later = 1;
-	list_for_each_entry_safe_continue(entry, next, rx_list, dccphrx_node) {
-		if (num_later == 0) {
-			a_loss = entry;
-			break;
-		} else if (dccp_rx_hist_entry_data_packet(entry))
-			--num_later;
+	if (dccp_delta_seqno(s1, s3) > 0) {		/* S1  <  S3  <  S2 */
+		/*
+		 * Reorder history to insert S3 between S1 and S2
+		 */
+		tfrc_rx_hist_swap(h, 2, 3);
+		tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n3);
+		h->loss_count = 3;
+		return 1;
 	}
 
-	if (a_loss == NULL) {
-		if (list_empty(li_list)) {
-			/* no loss event have occured yet */
-			LIMIT_NETDEBUG("%s: TODO: find a lost data packet by "
-				       "comparing to initial seqno\n",
-				       __FUNCTION__);
-			goto out;
-		} else {
-			LIMIT_NETDEBUG("%s: Less than 4 data pkts in history!",
-				       __FUNCTION__);
-			goto out;
-		}
-	}
+	/* S0  <  S3  <  S1 */
 
-	/* Locate a lost data packet */
-	entry = packet = b_loss;
-	list_for_each_entry_safe_continue(entry, next, rx_list, dccphrx_node) {
-		u64 delta = dccp_delta_seqno(entry->dccphrx_seqno,
-					     packet->dccphrx_seqno);
-
-		if (delta != 0) {
-			if (dccp_rx_hist_entry_data_packet(packet))
-				--delta;
-			/*
-			 * FIXME: check this, probably this % usage is because
-			 * in earlier drafts the ndp count was just 8 bits
-			 * long, but now it cam be up to 24 bits long.
-			 */
-#if 0
-			if (delta % DCCP_NDP_LIMIT !=
-			    (packet->dccphrx_ndp -
-			     entry->dccphrx_ndp) % DCCP_NDP_LIMIT)
-#endif
-			if (delta != packet->dccphrx_ndp - entry->dccphrx_ndp) {
-				seq_loss = entry->dccphrx_seqno;
-				dccp_inc_seqno(&seq_loss);
+	if (dccp_loss_free(s0, s3, n3)) {
+		u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp;
+
+		if (dccp_loss_free(s3, s1, n1)) {
+			/* hole between S0 and S1 filled by S3 */
+			u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp;
+
+			if (dccp_loss_free(s1, s2, n2)) {
+				/* entire hole filled by S0, S3, S1, S2 */
+				h->loss_start = tfrc_rx_hist_index(h, 2);
+				h->loss_count = 0;
+			} else {
+				/* gap remains between S1 and S2 */
+				h->loss_start = tfrc_rx_hist_index(h, 1);
+				h->loss_count = 1;
 			}
-		}
-		packet = entry;
-		if (packet == a_loss)
-			break;
+
+		} else /* gap exists between S3 and S1, loss_count stays at 2 */
+			tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n3);
+
+		return 0;
 	}
-out:
-	if (seq_loss != DCCP_MAX_SEQNO + 1)
-		*win_loss = a_loss->dccphrx_ccval;
-	else
-		*win_loss = 0; /* Paranoia */
 
-	return seq_loss;
-}
+	/*
+	 * The remaining case:  S0  <  S3  <  S1  <  S2;  gap between S0 and S3
+	 * Reorder history to insert S3 between S0 and S1.
+	 */
+	tfrc_rx_hist_swap(h, 0, 3);
+	h->loss_start = tfrc_rx_hist_index(h, 3);
+	tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n3);
+	h->loss_count = 3;
 
-EXPORT_SYMBOL_GPL(dccp_rx_hist_detect_loss);
+	return 1;
+}
 
-struct dccp_tx_hist *dccp_tx_hist_new(const char *name)
+/* recycle RX history records to continue loss detection if necessary */
+static void __three_after_loss(struct tfrc_rx_hist *h)
 {
-	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;
+	/*
+	 * At this stage we know already that there is a gap between S0 and S1
+	 * (since S0 was the highest sequence number received before detecting
+	 * the loss). To recycle the loss record, it is	thus only necessary to
+	 * check for other possible gaps between S1/S2 and between S2/S3.
+	 */
+	u64 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno,
+	    s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno,
+	    s3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_seqno;
+	u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp,
+	    n3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_ndp;
+
+	if (dccp_loss_free(s1, s2, n2)) {
+
+		if (dccp_loss_free(s2, s3, n3)) {
+			/* no gap between S2 and S3: entire hole is filled */
+			h->loss_start = tfrc_rx_hist_index(h, 3);
+			h->loss_count = 0;
+		} else {
+			/* gap between S2 and S3 */
+			h->loss_start = tfrc_rx_hist_index(h, 2);
+			h->loss_count = 1;
+		}
+
+	} else {	/* gap between S1 and S2 */
+		h->loss_start = tfrc_rx_hist_index(h, 1);
+		h->loss_count = 2;
+	}
 }
 
-EXPORT_SYMBOL_GPL(dccp_tx_hist_new);
+/**
+ *  tfrc_rx_handle_loss  -  Loss detection and further processing
+ *  @h:		    The non-empty RX history object
+ *  @lh:	    Loss Intervals database to update
+ *  @skb:	    Currently received packet
+ *  @ndp:	    The NDP count belonging to @skb
+ *  @calc_first_li: Caller-dependent computation of first loss interval in @lh
+ *  @sk:	    Used by @calc_first_li (see tfrc_lh_interval_add)
+ *
+ *  Chooses action according to pending loss, updates LI database when a new
+ *  loss was detected, and does required post-processing. Returns 1 when caller
+ *  should send feedback, 0 otherwise.
+ *  Since it also takes care of reordering during loss detection and updates the
+ *  records accordingly, the caller should not perform any more RX history
+ *  operations when loss_count is greater than 0 after calling this function.
+ */
+int tfrc_rx_handle_loss(struct tfrc_rx_hist *h,
+			struct tfrc_loss_hist *lh,
+			struct sk_buff *skb, const u64 ndp,
+			u32 (*calc_first_li)(struct sock *), struct sock *sk)
+{
+	int is_new_loss = 0;
+
+	if (h->loss_count == 0) {
+		__do_track_loss(h, skb, ndp);
+	} else if (h->loss_count == 1) {
+		__one_after_loss(h, skb, ndp);
+	} else if (h->loss_count != 2) {
+		DCCP_BUG("invalid loss_count %d", h->loss_count);
+	} else if (__two_after_loss(h, skb, ndp)) {
+		/*
+		 * Update Loss Interval database and recycle RX records
+		 */
+		is_new_loss = tfrc_lh_interval_add(lh, h, calc_first_li, sk);
+		__three_after_loss(h);
+	}
+	return is_new_loss;
+}
 
-void dccp_tx_hist_delete(struct dccp_tx_hist *hist)
+int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h)
 {
-	const char* name = kmem_cache_name(hist->dccptxh_slab);
+	int i;
 
-	kmem_cache_destroy(hist->dccptxh_slab);
-	kfree(name);
-	kfree(hist);
-}
+	for (i = 0; i <= TFRC_NDUPACK; i++) {
+		h->ring[i] = kmem_cache_alloc(tfrc_rx_hist_slab, GFP_ATOMIC);
+		if (h->ring[i] == NULL)
+			goto out_free;
+	}
 
-EXPORT_SYMBOL_GPL(dccp_tx_hist_delete);
+	h->loss_count = h->loss_start = 0;
+	return 0;
+
+out_free:
+	while (i-- != 0) {
+		kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]);
+		h->ring[i] = NULL;
+	}
+	return -ENOBUFS;
+}
 
-struct dccp_tx_hist_entry *
-	dccp_tx_hist_find_entry(const struct list_head *list, const u64 seq)
+void tfrc_rx_hist_purge(struct tfrc_rx_hist *h)
 {
-	struct dccp_tx_hist_entry *packet = NULL, *entry;
+	int i;
 
-	list_for_each_entry(entry, list, dccphtx_node)
-		if (entry->dccphtx_seqno == seq) {
-			packet = entry;
-			break;
+	for (i = 0; i <= TFRC_NDUPACK; ++i)
+		if (h->ring[i] != NULL) {
+			kmem_cache_free(tfrc_rx_hist_slab, h->ring[i]);
+			h->ring[i] = NULL;
 		}
-
-	return packet;
 }
 
-EXPORT_SYMBOL_GPL(dccp_tx_hist_find_entry);
-
-void dccp_tx_hist_purge_older(struct dccp_tx_hist *hist,
-			      struct list_head *list,
-			      struct dccp_tx_hist_entry *packet)
+/**
+ * tfrc_rx_hist_rtt_last_s - reference entry to compute RTT samples against
+ */
+static inline struct tfrc_rx_hist_entry *
+			tfrc_rx_hist_rtt_last_s(const struct tfrc_rx_hist *h)
 {
-	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);
-	}
+	return h->ring[0];
 }
 
-EXPORT_SYMBOL_GPL(dccp_tx_hist_purge_older);
+/**
+ * tfrc_rx_hist_rtt_prev_s - previously suitable (wrt rtt_last_s) RTT-sampling entry
+ */
+static inline struct tfrc_rx_hist_entry *
+			tfrc_rx_hist_rtt_prev_s(const struct tfrc_rx_hist *h)
+{
+	return h->ring[h->rtt_sample_prev];
+}
 
-void dccp_tx_hist_purge(struct dccp_tx_hist *hist, struct list_head *list)
+/**
+ * tfrc_rx_hist_sample_rtt  -  Sample RTT from timestamp / CCVal
+ * Based on ideas presented in RFC 4342, 8.1. Returns 0 if it was not able
+ * to compute a sample with given data - calling function should check this.
+ */
+u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, const struct sk_buff *skb)
 {
-	struct dccp_tx_hist_entry *entry, *next;
+	u32 sample = 0,
+	    delta_v = SUB16(dccp_hdr(skb)->dccph_ccval,
+			    tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
+
+	if (delta_v < 1 || delta_v > 4) {	/* unsuitable CCVal delta */
+		if (h->rtt_sample_prev == 2) {	/* previous candidate stored */
+			sample = SUB16(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval,
+				       tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
+			if (sample)
+				sample = 4 / sample *
+				         ktime_us_delta(tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_tstamp,
+							tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp);
+			else    /*
+				 * FIXME: This condition is in principle not
+				 * possible but occurs when CCID is used for
+				 * two-way data traffic. I have tried to trace
+				 * it, but the cause does not seem to be here.
+				 */
+				DCCP_BUG("please report to dccp@vger.kernel.org"
+					 " => prev = %u, last = %u",
+					 tfrc_rx_hist_rtt_prev_s(h)->tfrchrx_ccval,
+					 tfrc_rx_hist_rtt_last_s(h)->tfrchrx_ccval);
+		} else if (delta_v < 1) {
+			h->rtt_sample_prev = 1;
+			goto keep_ref_for_next_time;
+		}
 
-	list_for_each_entry_safe(entry, next, list, dccphtx_node) {
-		list_del_init(&entry->dccphtx_node);
-		dccp_tx_hist_entry_delete(hist, entry);
+	} else if (delta_v == 4) /* optimal match */
+		sample = ktime_to_us(net_timedelta(tfrc_rx_hist_rtt_last_s(h)->tfrchrx_tstamp));
+	else {			 /* suboptimal match */
+		h->rtt_sample_prev = 2;
+		goto keep_ref_for_next_time;
 	}
-}
 
-EXPORT_SYMBOL_GPL(dccp_tx_hist_purge);
+	if (unlikely(sample > DCCP_SANE_RTT_MAX)) {
+		DCCP_WARN("RTT sample %u too large, using max\n", sample);
+		sample = DCCP_SANE_RTT_MAX;
+	}
+
+	h->rtt_sample_prev = 0;	       /* use current entry as next reference */
+keep_ref_for_next_time:
 
-MODULE_AUTHOR("Ian McDonald <iam4@cs.waikato.ac.nz>, "
-	      "Arnaldo Carvalho de Melo <acme@ghostprotocols.net>");
-MODULE_DESCRIPTION("DCCP TFRC library");
-MODULE_LICENSE("GPL");
+	return sample;
+}
diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h
index 122e96737ff..ee362b0b630 100644
--- a/net/dccp/ccids/lib/packet_history.h
+++ b/net/dccp/ccids/lib/packet_history.h
@@ -1,13 +1,12 @@
 /*
- *  net/dccp/packet_history.h
+ *  Packet RX/TX history data structures and routines for TFRC-based protocols.
  *
- *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
- *
- *  An implementation of the DCCP protocol
+ *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
+ *  Copyright (c) 2005-6 The University of Waikato, Hamilton, New Zealand.
  *
  *  This code has been developed by the University of Waikato WAND
  *  research group. For further information please see http://www.wand.net.nz/
- *  or e-mail Ian McDonald - iam4@cs.waikato.ac.nz
+ *  or e-mail Ian McDonald - ian.mcdonald@jandi.co.nz
  *
  *  This code also uses code from Lulea University, rereleased as GPL by its
  *  authors:
@@ -37,164 +36,120 @@
 #ifndef _DCCP_PKT_HIST_
 #define _DCCP_PKT_HIST_
 
-#include <linux/config.h>
 #include <linux/list.h>
 #include <linux/slab.h>
-#include <linux/time.h>
-
-#include "../../dccp.h"
-
-/* Number of later packets received before one is considered lost */
-#define TFRC_RECV_NUM_LATE_LOSS	 3
-
-#define TFRC_WIN_COUNT_PER_RTT	 4
-#define TFRC_WIN_COUNT_LIMIT	16
-
-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;
-};
+#include "tfrc.h"
 
-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;
+/**
+ *  tfrc_tx_hist_entry  -  Simple singly-linked TX history list
+ *  @next:  next oldest entry (LIFO order)
+ *  @seqno: sequence number of this entry
+ *  @stamp: send time of packet with sequence number @seqno
+ */
+struct tfrc_tx_hist_entry {
+	struct tfrc_tx_hist_entry *next;
+	u64			  seqno;
+	ktime_t			  stamp;
 };
 
-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);
-
-static inline struct dccp_tx_hist_entry *
-		dccp_tx_hist_entry_new(struct dccp_tx_hist *hist,
-				       const gfp_t prio)
+static inline struct tfrc_tx_hist_entry *
+	tfrc_tx_hist_find_entry(struct tfrc_tx_hist_entry *head, u64 seqno)
 {
-	struct dccp_tx_hist_entry *entry = kmem_cache_alloc(hist->dccptxh_slab,
-							    prio);
-
-	if (entry != NULL)
-		entry->dccphtx_sent = 0;
-
-	return entry;
+	while (head != NULL && head->seqno != seqno)
+		head = head->next;
+	return head;
 }
 
-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);
-}
+int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno);
+void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp);
 
-extern struct dccp_tx_hist_entry *
-			dccp_tx_hist_find_entry(const struct list_head *list,
-						const u64 seq);
+/* Subtraction a-b modulo-16, respects circular wrap-around */
+#define SUB16(a, b) (((a) + 16 - (b)) & 0xF)
 
-static inline void dccp_tx_hist_add_entry(struct list_head *list,
-					  struct dccp_tx_hist_entry *entry)
-{
-	list_add(&entry->dccphtx_node, list);
-}
+/* Number of packets to wait after a missing packet (RFC 4342, 6.1) */
+#define TFRC_NDUPACK 3
 
-extern void dccp_tx_hist_purge_older(struct dccp_tx_hist *hist,
-				     struct list_head *list,
-				     struct dccp_tx_hist_entry *next);
+/**
+ * tfrc_rx_hist_entry - Store information about a single received packet
+ * @tfrchrx_seqno:	DCCP packet sequence number
+ * @tfrchrx_ccval:	window counter value of packet (RFC 4342, 8.1)
+ * @tfrchrx_ndp:	the NDP count (if any) of the packet
+ * @tfrchrx_tstamp:	actual receive time of packet
+ */
+struct tfrc_rx_hist_entry {
+	u64		 tfrchrx_seqno:48,
+			 tfrchrx_ccval:4,
+			 tfrchrx_type:4;
+	u64		 tfrchrx_ndp:48;
+	ktime_t		 tfrchrx_tstamp;
+};
 
-extern void dccp_tx_hist_purge(struct dccp_tx_hist *hist,
-			       struct list_head *list);
+/**
+ * tfrc_rx_hist  -  RX history structure for TFRC-based protocols
+ * @ring:		Packet history for RTT sampling and loss detection
+ * @loss_count:		Number of entries in circular history
+ * @loss_start:		Movable index (for loss detection)
+ * @rtt_sample_prev:	Used during RTT sampling, points to candidate entry
+ */
+struct tfrc_rx_hist {
+	struct tfrc_rx_hist_entry *ring[TFRC_NDUPACK + 1];
+	u8			  loss_count:2,
+				  loss_start:2;
+#define rtt_sample_prev		  loss_start
+};
 
-static inline struct dccp_tx_hist_entry *
-		dccp_tx_hist_head(struct list_head *list)
+/**
+ * tfrc_rx_hist_index - index to reach n-th entry after loss_start
+ */
+static inline u8 tfrc_rx_hist_index(const struct tfrc_rx_hist *h, const u8 n)
 {
-	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;
+	return (h->loss_start + n) & TFRC_NDUPACK;
 }
 
-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)
+/**
+ * tfrc_rx_hist_last_rcv - entry with highest-received-seqno so far
+ */
+static inline struct tfrc_rx_hist_entry *
+			tfrc_rx_hist_last_rcv(const struct tfrc_rx_hist *h)
 {
-	struct dccp_rx_hist_entry *entry = kmem_cache_alloc(hist->dccprxh_slab,
-							    prio);
-
-	if (entry != NULL) {
-		const struct dccp_hdr *dh = dccp_hdr(skb);
-
-		entry->dccphrx_seqno = DCCP_SKB_CB(skb)->dccpd_seq;
-		entry->dccphrx_ccval = dh->dccph_ccval;
-		entry->dccphrx_type  = dh->dccph_type;
-		entry->dccphrx_ndp   = ndp;
-		dccp_timestamp(sk, &entry->dccphrx_tstamp);
-	}
-
-	return entry;
+	return h->ring[tfrc_rx_hist_index(h, h->loss_count)];
 }
 
-static inline void dccp_rx_hist_entry_delete(struct dccp_rx_hist *hist,
-					     struct dccp_rx_hist_entry *entry)
+/**
+ * tfrc_rx_hist_entry - return the n-th history entry after loss_start
+ */
+static inline struct tfrc_rx_hist_entry *
+			tfrc_rx_hist_entry(const struct tfrc_rx_hist *h, const u8 n)
 {
-	if (entry != NULL)
-		kmem_cache_free(hist->dccprxh_slab, entry);
+	return h->ring[tfrc_rx_hist_index(h, n)];
 }
 
-extern void dccp_rx_hist_purge(struct dccp_rx_hist *hist,
-			       struct list_head *list);
-
-static inline void dccp_rx_hist_add_entry(struct list_head *list,
-					  struct dccp_rx_hist_entry *entry)
+/**
+ * tfrc_rx_hist_loss_prev - entry with highest-received-seqno before loss was detected
+ */
+static inline struct tfrc_rx_hist_entry *
+			tfrc_rx_hist_loss_prev(const struct tfrc_rx_hist *h)
 {
-	list_add(&entry->dccphrx_node, list);
+	return h->ring[h->loss_start];
 }
 
-static inline struct dccp_rx_hist_entry *
-		dccp_rx_hist_head(struct list_head *list)
+/* indicate whether previously a packet was detected missing */
+static inline bool tfrc_rx_hist_loss_pending(const struct tfrc_rx_hist *h)
 {
-	struct dccp_rx_hist_entry *head = NULL;
-
-	if (!list_empty(list))
-		head = list_entry(list->next, struct dccp_rx_hist_entry,
-				  dccphrx_node);
-	return head;
+	return h->loss_count > 0;
 }
 
-static inline int
-	dccp_rx_hist_entry_data_packet(const struct dccp_rx_hist_entry *entry)
-{
-	return entry->dccphrx_type == DCCP_PKT_DATA ||
-	       entry->dccphrx_type == DCCP_PKT_DATAACK;
-}
+void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, const struct sk_buff *skb,
+			     const u64 ndp);
 
-extern int 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);
+int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb);
 
-extern u64 dccp_rx_hist_detect_loss(struct list_head *rx_list,
-				    struct list_head *li_list, u8 *win_loss);
+struct tfrc_loss_hist;
+int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, struct tfrc_loss_hist *lh,
+			struct sk_buff *skb, const u64 ndp,
+			u32 (*first_li)(struct sock *sk), struct sock *sk);
+u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, const struct sk_buff *skb);
+int tfrc_rx_hist_alloc(struct tfrc_rx_hist *h);
+void tfrc_rx_hist_purge(struct tfrc_rx_hist *h);
 
 #endif /* _DCCP_PKT_HIST_ */
diff --git a/net/dccp/ccids/lib/tfrc.c b/net/dccp/ccids/lib/tfrc.c
new file mode 100644
index 00000000000..62b5828acde
--- /dev/null
+++ b/net/dccp/ccids/lib/tfrc.c
@@ -0,0 +1,45 @@
+/*
+ * TFRC library initialisation
+ *
+ * Copyright (c) 2007 The University of Aberdeen, Scotland, UK
+ * Copyright (c) 2007 Arnaldo Carvalho de Melo <acme@redhat.com>
+ */
+#include <linux/moduleparam.h>
+#include "tfrc.h"
+
+#ifdef CONFIG_IP_DCCP_TFRC_DEBUG
+bool tfrc_debug;
+module_param(tfrc_debug, bool, 0644);
+MODULE_PARM_DESC(tfrc_debug, "Enable TFRC debug messages");
+#endif
+
+int __init tfrc_lib_init(void)
+{
+	int rc = tfrc_li_init();
+
+	if (rc)
+		goto out;
+
+	rc = tfrc_tx_packet_history_init();
+	if (rc)
+		goto out_free_loss_intervals;
+
+	rc = tfrc_rx_packet_history_init();
+	if (rc)
+		goto out_free_tx_history;
+	return 0;
+
+out_free_tx_history:
+	tfrc_tx_packet_history_exit();
+out_free_loss_intervals:
+	tfrc_li_exit();
+out:
+	return rc;
+}
+
+void tfrc_lib_exit(void)
+{
+	tfrc_rx_packet_history_exit();
+	tfrc_tx_packet_history_exit();
+	tfrc_li_exit();
+}
diff --git a/net/dccp/ccids/lib/tfrc.h b/net/dccp/ccids/lib/tfrc.h
index 130c4c40cfe..40ee7d62b65 100644
--- a/net/dccp/ccids/lib/tfrc.h
+++ b/net/dccp/ccids/lib/tfrc.h
@@ -1,22 +1,77 @@
 #ifndef _TFRC_H_
 #define _TFRC_H_
 /*
- *  net/dccp/ccids/lib/tfrc.h
- *
- *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
- *  Copyright (c) 2005 Ian McDonald <iam4@cs.waikato.ac.nz>
- *  Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
- *  Copyright (c) 2003 Nils-Erik Mattsson, Joacim Haggmark, Magnus Erixzon
+ *  Copyright (c) 2007   The University of Aberdeen, Scotland, UK
+ *  Copyright (c) 2005-6 The University of Waikato, Hamilton, New Zealand.
+ *  Copyright (c) 2005-6 Ian McDonald <ian.mcdonald@jandi.co.nz>
+ *  Copyright (c) 2005   Arnaldo Carvalho de Melo <acme@conectiva.com.br>
+ *  Copyright (c) 2003   Nils-Erik Mattsson, Joacim Haggmark, Magnus Erixzon
  *
  *  This program is free software; you can redistribute it and/or modify
  *  it under the terms of the GNU General Public License as published by
  *  the Free Software Foundation; either version 2 of the License, or
  *  (at your option) any later version.
  */
-
 #include <linux/types.h>
+#include <linux/math64.h>
+#include "../../dccp.h"
+
+/* internal includes that this library exports: */
+#include "loss_interval.h"
+#include "packet_history.h"
+
+#ifdef CONFIG_IP_DCCP_TFRC_DEBUG
+extern bool tfrc_debug;
+#define tfrc_pr_debug(format, a...)	DCCP_PR_DEBUG(tfrc_debug, format, ##a)
+#else
+#define tfrc_pr_debug(format, a...)
+#endif
+
+/* integer-arithmetic divisions of type (a * 1000000)/b */
+static inline u64 scaled_div(u64 a, u64 b)
+{
+	BUG_ON(b == 0);
+	return div64_u64(a * 1000000, b);
+}
+
+static inline u32 scaled_div32(u64 a, u64 b)
+{
+	u64 result = scaled_div(a, b);
+
+	if (result > UINT_MAX) {
+		DCCP_CRIT("Overflow: %llu/%llu > UINT_MAX",
+			  (unsigned long long)a, (unsigned long long)b);
+		return UINT_MAX;
+	}
+	return result;
+}
+
+/**
+ * tfrc_ewma  -  Exponentially weighted moving average
+ * @weight: Weight to be used as damping factor, in units of 1/10
+ */
+static inline u32 tfrc_ewma(const u32 avg, const u32 newval, const u8 weight)
+{
+	return avg ? (weight * avg + (10 - weight) * newval) / 10 : newval;
+}
+
+u32 tfrc_calc_x(u16 s, u32 R, u32 p);
+u32 tfrc_calc_x_reverse_lookup(u32 fvalue);
+u32 tfrc_invert_loss_event_rate(u32 loss_event_rate);
+
+int tfrc_tx_packet_history_init(void);
+void tfrc_tx_packet_history_exit(void);
+int tfrc_rx_packet_history_init(void);
+void tfrc_rx_packet_history_exit(void);
 
-extern u32 tfrc_calc_x(u16 s, u32 R, u32 p);
-extern u32 tfrc_calc_x_reverse_lookup(u32 fvalue);
+int tfrc_li_init(void);
+void tfrc_li_exit(void);
 
+#ifdef CONFIG_IP_DCCP_TFRC_LIB
+int tfrc_lib_init(void);
+void tfrc_lib_exit(void);
+#else
+#define tfrc_lib_init() (0)
+#define tfrc_lib_exit()
+#endif
 #endif /* _TFRC_H_ */
diff --git a/net/dccp/ccids/lib/tfrc_equation.c b/net/dccp/ccids/lib/tfrc_equation.c
index add3cae65e2..88ef98285be 100644
--- a/net/dccp/ccids/lib/tfrc_equation.c
+++ b/net/dccp/ccids/lib/tfrc_equation.c
@@ -1,8 +1,6 @@
 /*
- *  net/dccp/ccids/lib/tfrc_equation.c
- *
  *  Copyright (c) 2005 The University of Waikato, Hamilton, New Zealand.
- *  Copyright (c) 2005 Ian McDonald <iam4@cs.waikato.ac.nz>
+ *  Copyright (c) 2005 Ian McDonald <ian.mcdonald@jandi.co.nz>
  *  Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@conectiva.com.br>
  *  Copyright (c) 2003 Nils-Erik Mattsson, Joacim Haggmark, Magnus Erixzon
  *
@@ -12,18 +10,84 @@
  *  (at your option) any later version.
  */
 
-#include <linux/config.h>
 #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 },
@@ -527,117 +591,115 @@ 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;
+
+	/* 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 (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 (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;
-
-	if (index < 0)
-		/* p should be 0 unless there is a bug in my code */
-		index = 0;
-
-	if (R == 0)
-		R = 1; /* RTT can't be zero or else divide by zero */
-
-	BUG_ON(index >= TFRC_CALC_X_ARRSIZE);
-
-	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 */
+	} else {				      /* 0.05   <  p <= 1.00  */
+		index = p/(1000000/TFRC_CALC_X_ARRSIZE) - 1;
 
-	return (u32)tmp1; 
+		f = tfrc_calc_x_lookup[index][0];
+	}
+
+	/*
+	 * 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)
+ *  @fvalue: function value to match, scaled by 1000000
  *
- * both fvalue and p are multiplied by 1,000,000 to use ints
+ *  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 %u smaller than resolution\n", fvalue);
+		return TFRC_SMALLEST_P;
+	}
+	if (fvalue > tfrc_calc_x_lookup[TFRC_CALC_X_ARRSIZE - 1][0]) {
+		DCCP_WARN("fvalue %u exceeds bounds!\n", fvalue);
 		return 1000000;
-	else
-		small = 0;
+	}
 
-	while (fvalue > tfrc_calc_x_lookup[ctr][small])
-		ctr++;
+	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;
+	}
 
-	if (small)
-		return TFRC_CALC_X_SPLIT * ctr / TFRC_CALC_X_ARRSIZE;
-	else
-		return 1000000 * ctr / 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);
+/**
+ * tfrc_invert_loss_event_rate  -  Compute p so that 10^6 corresponds to 100%
+ * When @loss_event_rate is large, there is a chance that p is truncated to 0.
+ * To avoid re-entering slow-start in that case, we set p = TFRC_SMALLEST_P > 0.
+ */
+u32 tfrc_invert_loss_event_rate(u32 loss_event_rate)
+{
+	if (loss_event_rate == UINT_MAX)		/* see RFC 4342, 8.5 */
+		return 0;
+	if (unlikely(loss_event_rate == 0))		/* map 1/0 into 100% */
+		return 1000000;
+	return max_t(u32, scaled_div(1, loss_event_rate), TFRC_SMALLEST_P);
+}