From: Eric Dumazet <eric.dumazet@gmail.com>
To: "David S. Miller" <davem@davemloft.net>,
Patrick McHardy <kaber@trash.net>
Cc: Linux Netdev List <netdev@vger.kernel.org>
Subject: [RFC] pkt_sched: gen_estimator: Dont report fake rate estimators
Date: Thu, 01 Oct 2009 21:07:51 +0200 [thread overview]
Message-ID: <4AC4FE07.5070204@gmail.com> (raw)
We currently send TCA_STATS_RATE_EST elements to netlink users, even if no estimator
is running.
# tc -s -d qdisc
qdisc pfifo_fast 0: dev eth0 root bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
Sent 112833764978 bytes 1495081739 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
User has no way to tell if the "rate 0bit 0pps" is a real estimation, or a fake
one (because no estimator is active)
After this patch, tc command output is :
$ tc -s -d qdisc
qdisc pfifo_fast 0: dev eth0 root bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
Sent 561075 bytes 1196 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
include/linux/gen_stats.h | 16 +++++++++++++---
net/core/gen_estimator.c | 9 +++++----
net/core/gen_stats.c | 9 ++++++---
net/sched/act_police.c | 2 +-
4 files changed, 25 insertions(+), 11 deletions(-)
diff --git a/include/linux/gen_stats.h b/include/linux/gen_stats.h
index 710e901..7678ded 100644
--- a/include/linux/gen_stats.h
+++ b/include/linux/gen_stats.h
@@ -30,17 +30,27 @@ struct gnet_stats_basic_packed
} __attribute__ ((packed));
/**
- * struct gnet_stats_rate_est - rate estimator
+ * struct gnet_stats_user_rate_est - rate estimator
* @bps: current byte rate
* @pps: current packet rate
*/
-struct gnet_stats_rate_est
-{
+struct gnet_stats_user_rate_est {
__u32 bps;
__u32 pps;
};
/**
+ * struct gnet_stats_rate_est - rate estimator with flags
+ * @est: current byte/packet rate
+ * @flags: set to one if estimation is valid
+ */
+struct gnet_stats_rate_est {
+ struct gnet_stats_user_rate_est est;
+ int flags;
+};
+#define RATE_EST_VALID 1
+
+/**
* struct gnet_stats_queue - queuing statistics
* @qlen: queue length
* @backlog: backlog size of queue
diff --git a/net/core/gen_estimator.c b/net/core/gen_estimator.c
index 493775f..5ba9d90 100644
--- a/net/core/gen_estimator.c
+++ b/net/core/gen_estimator.c
@@ -129,12 +129,13 @@ static void est_timer(unsigned long arg)
brate = (nbytes - e->last_bytes)<<(7 - idx);
e->last_bytes = nbytes;
e->avbps += (brate >> e->ewma_log) - (e->avbps >> e->ewma_log);
- e->rate_est->bps = (e->avbps+0xF)>>5;
+ e->rate_est->est.bps = (e->avbps+0xF)>>5;
rate = (npackets - e->last_packets)<<(12 - idx);
e->last_packets = npackets;
e->avpps += (rate >> e->ewma_log) - (e->avpps >> e->ewma_log);
- e->rate_est->pps = (e->avpps+0x1FF)>>10;
+ e->rate_est->est.pps = (e->avpps+0x1FF)>>10;
+ e->rate_est->flags |= RATE_EST_VALID;
skip:
read_unlock(&est_lock);
spin_unlock(e->stats_lock);
@@ -227,9 +228,9 @@ int gen_new_estimator(struct gnet_stats_basic_packed *bstats,
est->stats_lock = stats_lock;
est->ewma_log = parm->ewma_log;
est->last_bytes = bstats->bytes;
- est->avbps = rate_est->bps<<5;
+ est->avbps = rate_est->est.bps<<5;
est->last_packets = bstats->packets;
- est->avpps = rate_est->pps<<10;
+ est->avpps = rate_est->est.pps<<10;
if (!elist[idx].timer.function) {
INIT_LIST_HEAD(&elist[idx].list);
diff --git a/net/core/gen_stats.c b/net/core/gen_stats.c
index 8569310..b6f723c 100644
--- a/net/core/gen_stats.c
+++ b/net/core/gen_stats.c
@@ -138,13 +138,16 @@ gnet_stats_copy_basic(struct gnet_dump *d, struct gnet_stats_basic_packed *b)
int
gnet_stats_copy_rate_est(struct gnet_dump *d, struct gnet_stats_rate_est *r)
{
+ if (!(r->flags & RATE_EST_VALID))
+ return 0;
+
if (d->compat_tc_stats) {
- d->tc_stats.bps = r->bps;
- d->tc_stats.pps = r->pps;
+ d->tc_stats.bps = r->est.bps;
+ d->tc_stats.pps = r->est.pps;
}
if (d->tail)
- return gnet_stats_copy(d, TCA_STATS_RATE_EST, r, sizeof(*r));
+ return gnet_stats_copy(d, TCA_STATS_RATE_EST, &r->est, sizeof(r->est));
return 0;
}
diff --git a/net/sched/act_police.c b/net/sched/act_police.c
index 723964c..ba01081 100644
--- a/net/sched/act_police.c
+++ b/net/sched/act_police.c
@@ -292,7 +292,7 @@ static int tcf_act_police(struct sk_buff *skb, struct tc_action *a,
police->tcf_bstats.packets++;
if (police->tcfp_ewma_rate &&
- police->tcf_rate_est.bps >= police->tcfp_ewma_rate) {
+ police->tcf_rate_est.est.bps >= police->tcfp_ewma_rate) {
police->tcf_qstats.overlimits++;
if (police->tcf_action == TC_ACT_SHOT)
police->tcf_qstats.drops++;
next reply other threads:[~2009-10-01 19:07 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-01 19:07 Eric Dumazet [this message]
2009-10-01 19:37 ` [RFC] pkt_sched: gen_estimator: Dont report fake rate estimators David Miller
2009-10-01 21:05 ` Jarek Poplawski
2009-10-01 21:14 ` David Miller
2009-10-01 21:21 ` Jarek Poplawski
2009-10-02 7:08 ` Jarek Poplawski
2009-10-02 7:12 ` Eric Dumazet
2009-10-02 7:17 ` Jarek Poplawski
2009-10-02 7:32 ` Jarek Poplawski
2009-10-02 10:35 ` [RFC take2] " Eric Dumazet
2009-10-02 11:25 ` Jarek Poplawski
2009-10-02 12:39 ` Eric Dumazet
2009-10-02 20:11 ` Jarek Poplawski
2009-10-02 20:32 ` Eric Dumazet
2009-10-07 8:27 ` David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4AC4FE07.5070204@gmail.com \
--to=eric.dumazet@gmail.com \
--cc=davem@davemloft.net \
--cc=kaber@trash.net \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).