netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Graf <tgraf@suug.ch>
To: "David S. Miller" <davem@davemloft.net>
Cc: netdev@oss.sgi.com
Subject: [PATCH 3/3] NET: Generic network statistics/estimator documentation
Date: Sun, 3 Oct 2004 23:41:46 +0200	[thread overview]
Message-ID: <20041003214146.GJ14344@postel.suug.ch> (raw)
In-Reply-To: <20041003213124.GG14344@postel.suug.ch>

Documentation of generic network statistics and estimator API.

Signed-off-by: Thomas Graf <tgraf@suug.ch>


--- linux-2.6.9-rc3.orig/Documentation/networking/gen_stats.txt	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.9-rc3/Documentation/networking/gen_stats.txt	2004-10-03 22:33:57.000000000 +0200
@@ -0,0 +1,117 @@
+Generic networking statistics for netlink users
+======================================================================
+
+Statistic counters are grouped into structs:
+
+Struct               TLV type              Description
+----------------------------------------------------------------------
+gnet_stats_basic     TCA_STATS_BASIC       Basic statistics
+gnet_stats_rate_est  TCA_STATS_RATE_EST    Rate estimator
+gnet_stats_queue     TCA_STATS_QUEUE       Queue statistics
+none                 TCA_STATS_APP         Application specific
+
+
+Collecting:
+-----------
+
+Declare the statistic structs you need:
+struct mystruct {
+	struct gnet_stats_basic	bstats;
+	struct gnet_stats_queue	qstats;
+	...
+};
+
+Update statistics:
+mystruct->tstats.packet++;
+mystruct->qstats.backlog += skb->pkt_len;
+
+
+Export to userspace (Dump):
+---------------------------
+
+my_dumping_routine(struct sk_buff *skb, ...)
+{
+	struct gnet_dump dump;
+
+	if (gnet_stats_start_copy(skb, TCA_STATS2, &mystruct->lock, &dump) < 0)
+		goto rtattr_failure;
+
+	if (gnet_stats_copy_basic(&dump, &mystruct->bstats) < 0 ||
+	    gnet_stats_copy_queue(&dump, &mystruct->qstats) < 0 ||
+		gnet_stats_copy_app(&dump, &xstats, sizeof(xstats)) < 0)
+		goto rtattr_failure;
+
+	if (gnet_stats_finish_copy(&dump) < 0)
+		goto rtattr_failure;
+	...
+}
+
+TCA_STATS/TCA_XSTATS backward compatibility:
+--------------------------------------------
+
+Prior users of struct tc_stats and xstats can maintain backward
+compatibility by calling the compat wrappers to keep providing the
+existing TLV types.
+
+my_dumping_routine(struct sk_buff *skb, ...)
+{
+    if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS,
+		TCA_XSTATS, &mystruct->lock, &dump) < 0)
+		goto rtattr_failure;
+	...
+}
+
+A struct tc_stats will be filled out during gnet_stats_copy_* calls
+and appended to the skb. TCA_XSTATS is provided if gnet_stats_copy_app
+was called.
+
+
+Locking:
+--------
+
+Locks are taken before writing and released once all statistics have
+been written. Locks are always released in case of an error. You
+are responsible for making sure that the lock is initialized.
+
+
+Rate Estimator:
+--------------
+
+0) Prepare an estimator attribute. Most likely this would be in user
+   space. The value of this TLV should contain a tc_estimator structure.
+   As usual, such a TLV nees to be 32 bit aligned and therefore the
+   length needs to be appropriately set etc. The estimator interval
+   and ewma log need to be converted to the appropriate values.
+   tc_estimator.c::tc_setup_estimator() is advisable to be used as the
+   conversion routine. It does a few clever things. It takes a time
+   interval in microsecs, a time constant also in microsecs and a struct
+   tc_estimator to  be populated. The returned tc_estimator can be
+   transported to the kernel.  Transfer such a structure in a TLV of type
+   TCA_RATE to your code in the kernel.
+
+In the kernel when setting up:
+1) make sure you have basic stats and rate stats setup first.
+2) make sure you have initialized stats lock that is used to setup such
+   stats.
+3) Now initialize a new estimator:
+
+   int ret = gen_new_estimator(my_basicstats,my_rate_est_stats,
+       mystats_lock, attr_with_tcestimator_struct);
+
+   if ret == 0
+       success
+   else
+       failed
+
+From now on, everytime you dump my_rate_est_stats it will contain
+uptodate info.
+
+Once you are done, call gen_kill_estimator(my_basicstats,
+my_rate_est_stats) Make sure that my_basicstats and my_rate_est_stats
+are still valid (i.e still exist) at the time of making this call.
+
+
+Authors:
+--------
+Thomas Graf <tgraf@suug.ch>
+Jamal Hadi Salim <hadi@cyberus.ca>

  parent reply	other threads:[~2004-10-03 21:41 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-10-03 21:31 [PATCH 0/3] NET: Generic network statistics/estimator Thomas Graf
2004-10-03 21:34 ` [PATCH 1/3] NET: Generic network statistics API Thomas Graf
2004-10-03 21:39 ` [PATCH 2/3] NET: Generic rate estimator Thomas Graf
2004-10-03 23:14   ` David S. Miller
2004-10-03 23:36     ` Thomas Graf
2004-10-04  1:16       ` jamal
2004-10-04 12:53         ` Thomas Graf
2004-10-04 13:24           ` jamal
2004-10-04 14:15             ` Thomas Graf
2004-10-04 14:59               ` jamal
2004-10-04 15:29                 ` Thomas Graf
2004-10-04 19:30                 ` [RESEND PATCH " Thomas Graf
2004-10-04 20:07                   ` jamal
2004-10-04 20:20                     ` Thomas Graf
2004-10-03 23:57     ` [PATCH " Thomas Graf
2004-10-05 21:03       ` David S. Miller
2004-10-05 22:21         ` [RFC] Replacing qdisc tc_stats with generic statistics to make it extendable Thomas Graf
2004-10-06 17:41           ` David S. Miller
2004-10-03 21:41 ` Thomas Graf [this message]
2004-10-03 21:47 ` [PATCH 0/3] NET: Generic network statistics/estimator Thomas Graf
2004-10-03 22:23   ` David S. Miller
2004-10-03 22:34     ` jamal

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=20041003214146.GJ14344@postel.suug.ch \
    --to=tgraf@suug.ch \
    --cc=davem@davemloft.net \
    --cc=netdev@oss.sgi.com \
    /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).