6 files changed, 65 insertions, 6 deletions

diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 281c1bd..51b968d 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -285,6 +285,45 @@ static struct Qdisc_ops *qdisc_lookup_ops(struct nlattr *kind)
 	return q;
 }
 
+/* The linklayer setting were not transferred from iproute2, in older
+ * versions, and the rate tables lookup systems have been dropped in
+ * the kernel. To keep backward compatible with older iproute2 tc
+ * utils, we detect the linklayer setting by detecting if the rate
+ * table were modified.
+ *
+ * For linklayer ATM table entries, the rate table will be aligned to
+ * 48 bytes, thus some table entries will contain the same value.  The
+ * mpu (min packet unit) is also encoded into the old rate table, thus
+ * starting from the mpu, we find low and high table entries for
+ * mapping this cell.  If these entries contain the same value, when
+ * the rate tables have been modified for linklayer ATM.
+ *
+ * This is done by rounding mpu to the nearest 48 bytes cell/entry,
+ * and then roundup to the next cell, calc the table entry one below,
+ * and compare.
+ */
+static __u8 __detect_linklayer(struct tc_ratespec *r, __u32 *rtab)
+{
+	int low       = roundup(r->mpu, 48);
+	int high      = roundup(low+1, 48);
+	int cell_low  = low >> r->cell_log;
+	int cell_high = (high >> r->cell_log) - 1;
+
+	/* rtab is too inaccurate at rates > 100Mbit/s */
+	if ((r->rate > (100000000/8)) || (rtab[0] == 0)) {
+		pr_debug("TC linklayer: Giving up ATM detection\n");
+		return TC_LINKLAYER_ETHERNET;
+	}
+
+	if ((cell_high > cell_low) && (cell_high < 256)
+	    && (rtab[cell_low] == rtab[cell_high])) {
+		pr_debug("TC linklayer: Detected ATM, low(%d)=high(%d)=%u\n",
+			 cell_low, cell_high, rtab[cell_high]);
+		return TC_LINKLAYER_ATM;
+	}
+	return TC_LINKLAYER_ETHERNET;
+}
+
 static struct qdisc_rate_table *qdisc_rtab_list;
 
 struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct nlattr *tab)
@@ -308,6 +347,8 @@ struct qdisc_rate_table *qdisc_get_rtab(struct tc_ratespec *r, struct nlattr *ta
 		rtab->rate = *r;
 		rtab->refcnt = 1;
 		memcpy(rtab->data, nla_data(tab), 1024);
+		if (r->linklayer == TC_LINKLAYER_UNAWARE)
+			r->linklayer = __detect_linklayer(r, rtab->data);
 		rtab->next = qdisc_rtab_list;
 		qdisc_rtab_list = rtab;
 	}
diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c
index ca8e0a5..1f9c314 100644
--- a/net/sched/sch_atm.c
+++ b/net/sched/sch_atm.c
@@ -605,6 +605,7 @@ static int atm_tc_dump_class(struct Qdisc *sch, unsigned long cl,
 		struct sockaddr_atmpvc pvc;
 		int state;
 
+		memset(&pvc, 0, sizeof(pvc));
 		pvc.sap_family = AF_ATMPVC;
 		pvc.sap_addr.itf = flow->vcc->dev ? flow->vcc->dev->number : -1;
 		pvc.sap_addr.vpi = flow->vcc->vpi;
diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c
index 71a5688..7a42c81 100644
--- a/net/sched/sch_cbq.c
+++ b/net/sched/sch_cbq.c
@@ -1465,6 +1465,7 @@ static int cbq_dump_wrr(struct sk_buff *skb, struct cbq_class *cl)
 	unsigned char *b = skb_tail_pointer(skb);
 	struct tc_cbq_wrropt opt;
 
+	memset(&opt, 0, sizeof(opt));
 	opt.flags = 0;
 	opt.allot = cl->allot;
 	opt.priority = cl->priority + 1;
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 4626cef..48be3d5 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -25,6 +25,7 @@
 #include <linux/rcupdate.h>
 #include <linux/list.h>
 #include <linux/slab.h>
+#include <linux/if_vlan.h>
 #include <net/sch_generic.h>
 #include <net/pkt_sched.h>
 #include <net/dst.h>
@@ -207,15 +208,19 @@ void __qdisc_run(struct Qdisc *q)
 
 unsigned long dev_trans_start(struct net_device *dev)
 {
-	unsigned long val, res = dev->trans_start;
+	unsigned long val, res;
 	unsigned int i;
 
+	if (is_vlan_dev(dev))
+		dev = vlan_dev_real_dev(dev);
+	res = dev->trans_start;
 	for (i = 0; i < dev->num_tx_queues; i++) {
 		val = netdev_get_tx_queue(dev, i)->trans_start;
 		if (val && time_after(val, res))
 			res = val;
 	}
 	dev->trans_start = res;
+
 	return res;
 }
 EXPORT_SYMBOL(dev_trans_start);
@@ -904,6 +909,7 @@ void psched_ratecfg_precompute(struct psched_ratecfg *r,
 	memset(r, 0, sizeof(*r));
 	r->overhead = conf->overhead;
 	r->rate_bytes_ps = conf->rate;
+	r->linklayer = (conf->linklayer & TC_LINKLAYER_MASK);
 	r->mult = 1;
 	/*
 	 * The deal here is to replace a divide by a reciprocal one
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c
index c2124ea..c2178b1 100644
--- a/net/sched/sch_htb.c
+++ b/net/sched/sch_htb.c
@@ -100,7 +100,7 @@ struct htb_class {
 	struct psched_ratecfg	ceil;
 	s64			buffer, cbuffer;/* token bucket depth/rate */
 	s64			mbuffer;	/* max wait time */
-	int			prio;		/* these two are used only by leaves... */
+	u32			prio;		/* these two are used only by leaves... */
 	int			quantum;	/* but stored for parent-to-leaf return */
 
 	struct tcf_proto	*filter_list;	/* class attached filters */
@@ -1329,6 +1329,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 	struct htb_sched *q = qdisc_priv(sch);
 	struct htb_class *cl = (struct htb_class *)*arg, *parent;
 	struct nlattr *opt = tca[TCA_OPTIONS];
+	struct qdisc_rate_table *rtab = NULL, *ctab = NULL;
 	struct nlattr *tb[TCA_HTB_MAX + 1];
 	struct tc_htb_opt *hopt;
 
@@ -1350,6 +1351,18 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 	if (!hopt->rate.rate || !hopt->ceil.rate)
 		goto failure;
 
+	/* Keeping backward compatible with rate_table based iproute2 tc */
+	if (hopt->rate.linklayer == TC_LINKLAYER_UNAWARE) {
+		rtab = qdisc_get_rtab(&hopt->rate, tb[TCA_HTB_RTAB]);
+		if (rtab)
+			qdisc_put_rtab(rtab);
+	}
+	if (hopt->ceil.linklayer == TC_LINKLAYER_UNAWARE) {
+		ctab = qdisc_get_rtab(&hopt->ceil, tb[TCA_HTB_CTAB]);
+		if (ctab)
+			qdisc_put_rtab(ctab);
+	}
+
 	if (!cl) {		/* new class */
 		struct Qdisc *new_q;
 		int prio;
diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
index 82f6016..a6d788d 100644
--- a/net/sched/sch_netem.c
+++ b/net/sched/sch_netem.c
@@ -412,12 +412,9 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 
 	/* If a delay is expected, orphan the skb. (orphaning usually takes
 	 * place at TX completion time, so _before_ the link transit delay)
-	 * Ideally, this orphaning should be done after the rate limiting
-	 * module, because this breaks TCP Small Queue, and other mechanisms
-	 * based on socket sk_wmem_alloc.
 	 */
 	if (q->latency || q->jitter)
-		skb_orphan(skb);
+		skb_orphan_partial(skb);
 
 	/*
 	 * If we need to duplicate packet, then re-insert at top of the