netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Robert Olsson <Robert.Olsson@data.slu.se>
To: sfeldma@pobox.com
Cc: Robert Olsson <Robert.Olsson@data.slu.se>,
	Lennert Buytenhek <buytenh@wantstofly.org>,
	jamal <hadi@cyberus.ca>,
	P@draigBrady.com, mellia@prezzemolo.polito.it,
	e1000-devel@lists.sourceforge.net,
	Jorge Manuel Finochietto <jorge.finochietto@polito.it>,
	Giulio Galante <galante@polito.it>,
	netdev@oss.sgi.com
Subject: Re: [E1000-devel] Transmission limit
Date: Thu, 2 Dec 2004 19:23:24 +0100	[thread overview]
Message-ID: <16815.23964.93437.411404@robur.slu.se> (raw)
In-Reply-To: <1101919791.5198.15.camel@localhost.localdomain>


Hello!

Below is little patch to clean skb at xmit. It's old jungle trick Jamal
and I used w. tulip. Note we can now even decrease the size of TX ring.

It can increase TX performance from 800 kpps to
  1125128pps 576Mb/sec (576065536bps) errors: 0
  1124946pps 575Mb/sec (575972352bps) errors: 0

But suffers from scheduling problems as the previous patch. Often we just get
  582108pps 298Mb/sec (298039296bps) errors: 0

When the sender CPU free (it's) skb's. we might get some "TX free affinity"
which are unrelated to irq affinity of course not 100% perfect.
 
And some of Scotts may still be used. 

--- drivers/net/e1000/e1000.h.orig	2004-12-01 13:59:36.000000000 +0100
+++ drivers/net/e1000/e1000.h	2004-12-02 20:11:31.000000000 +0100
@@ -103,7 +103,7 @@
 #define E1000_MAX_INTR 10
 
 /* TX/RX descriptor defines */
-#define E1000_DEFAULT_TXD                  256
+#define E1000_DEFAULT_TXD                  128
 #define E1000_MAX_TXD                      256
 #define E1000_MIN_TXD                       80
 #define E1000_MAX_82544_TXD               4096
--- drivers/net/e1000/e1000_main.c.orig	2004-12-01 13:59:36.000000000 +0100
+++ drivers/net/e1000/e1000_main.c	2004-12-02 20:37:40.000000000 +0100
@@ -1820,6 +1820,10 @@
  		return NETDEV_TX_LOCKED; 
  	} 
 
+
+	if( adapter->tx_ring.next_to_use - adapter->tx_ring.next_to_clean > 80 )
+		e1000_clean_tx_ring(adapter);
+
 	/* need: count + 2 desc gap to keep tail from touching
 	 * head, otherwise try next time */
 	if(E1000_DESC_UNUSED(&adapter->tx_ring) < count + 2) {


						  --ro

  parent reply	other threads:[~2004-12-02 18:23 UTC|newest]

Thread overview: 85+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1101467291.24742.70.camel@mellia.lipar.polito.it>
2004-11-26 14:05 ` [E1000-devel] Transmission limit P
2004-11-26 15:31   ` Marco Mellia
2004-11-26 19:56     ` jamal
2004-11-29 14:21       ` Marco Mellia
2004-11-30 13:46         ` jamal
2004-12-02 17:24           ` Marco Mellia
2004-11-26 20:06     ` jamal
2004-11-26 20:56     ` Lennert Buytenhek
2004-11-26 21:02       ` Lennert Buytenhek
2004-11-27  9:25     ` Harald Welte
     [not found]       ` <20041127111101.GC23139@xi.wantstofly.org>
2004-11-27 11:31         ` Harald Welte
2004-11-27 20:12       ` Cesar Marcondes
2004-11-29  8:53       ` Marco Mellia
2004-11-29 14:50         ` Lennert Buytenhek
2004-11-30  8:42           ` Marco Mellia
2004-12-01 12:25             ` jamal
2004-12-02 13:39               ` Marco Mellia
2004-12-03 13:07                 ` jamal
2004-11-26 15:40   ` Robert Olsson
2004-11-26 15:59     ` Marco Mellia
2004-11-26 16:57       ` P
2004-11-26 20:01         ` jamal
2004-11-29 10:19           ` P
2004-11-29 13:09           ` Robert Olsson
2004-11-29 20:16             ` David S. Miller
2004-12-01 16:47               ` Robert Olsson
2004-11-30 13:31             ` jamal
2004-11-30 13:46               ` Lennert Buytenhek
2004-11-30 14:25                 ` jamal
2004-12-01  0:11                   ` Lennert Buytenhek
2004-12-01  1:09                     ` Scott Feldman
2004-12-01 15:34                       ` Robert Olsson
2004-12-01 16:49                         ` Scott Feldman
2004-12-01 17:37                           ` Robert Olsson
2004-12-02 17:54                           ` Robert Olsson
2004-12-02 18:23                           ` Robert Olsson [this message]
2004-12-02 23:25                             ` Lennert Buytenhek
2004-12-03  5:23                             ` Scott Feldman
2004-12-10 16:24                             ` Martin Josefsson
2004-12-01 18:29                       ` Lennert Buytenhek
2004-12-01 21:35                         ` Lennert Buytenhek
2004-12-02  6:13                           ` Scott Feldman
2004-12-03 13:24                             ` jamal
2004-12-05 14:50                             ` 1.03Mpps on e1000 (was: Re: [E1000-devel] Transmission limit) Lennert Buytenhek
2004-12-05 15:03                               ` Martin Josefsson
2004-12-05 15:15                                 ` Lennert Buytenhek
2004-12-05 15:19                                   ` Martin Josefsson
2004-12-05 15:30                                     ` Martin Josefsson
2004-12-05 17:00                                       ` Lennert Buytenhek
2004-12-05 17:11                                         ` Martin Josefsson
2004-12-05 17:38                                           ` Martin Josefsson
2004-12-05 18:14                                             ` Lennert Buytenhek
2004-12-05 15:42                                 ` Martin Josefsson
2004-12-05 16:48                                   ` Martin Josefsson
2004-12-05 17:01                                     ` Martin Josefsson
2004-12-05 17:58                                     ` Lennert Buytenhek
2004-12-05 17:44                                   ` Lennert Buytenhek
2004-12-05 17:51                                     ` Lennert Buytenhek
2004-12-05 17:54                                       ` Martin Josefsson
2004-12-06 11:32                                         ` 1.03Mpps on e1000 (was: " jamal
2004-12-06 12:11                                           ` Lennert Buytenhek
2004-12-06 12:20                                             ` jamal
2004-12-06 12:23                                               ` Lennert Buytenhek
2004-12-06 12:30                                                 ` Martin Josefsson
2004-12-06 13:11                                                   ` jamal
     [not found]                                                     ` <20041206132907.GA13411@xi.wantstofly.org>
     [not found]                                                       ` <16820.37049.396306.295878@robur.slu.se>
2004-12-06 17:32                                                         ` 1.03Mpps on e1000 (was: Re: [E1000-devel] " P
2004-12-08 23:36                                   ` Ray Lehtiniemi
     [not found]                                     ` <41B825A5.2000009@draigBrady.com>
     [not found]                                       ` <20041209161825.GA32454@mail.com>
2004-12-09 17:12                                         ` 1.03Mpps on e1000 P
     [not found]                                         ` <20041209164820.GB32454@mail.com>
2004-12-09 17:19                                           ` P
2004-12-09 23:25                                             ` Ray Lehtiniemi
2004-12-05 21:12                                 ` 1.03Mpps on e1000 (was: Re: [E1000-devel] Transmission limit) Scott Feldman
2004-12-05 21:25                                   ` Lennert Buytenhek
2004-12-06  1:23                                     ` 1.03Mpps on e1000 (was: " Scott Feldman
2004-12-02 17:31                       ` [E1000-devel] Transmission limit Marco Mellia
2004-12-03 20:57                       ` Lennert Buytenhek
2004-12-04 10:36                         ` Lennert Buytenhek
2004-12-01 12:08                     ` jamal
2004-12-01 15:24                       ` Lennert Buytenhek
2004-11-26 17:58       ` Robert Olsson
2004-11-27 20:00   ` Lennert Buytenhek
2004-11-29 12:44     ` Marco Mellia
2004-11-29 15:19       ` Lennert Buytenhek
2004-11-29 17:32         ` Marco Mellia
2004-11-29 19:08           ` Lennert Buytenhek
2004-11-29 19:09             ` Lennert Buytenhek

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=16815.23964.93437.411404@robur.slu.se \
    --to=robert.olsson@data.slu.se \
    --cc=P@draigBrady.com \
    --cc=buytenh@wantstofly.org \
    --cc=e1000-devel@lists.sourceforge.net \
    --cc=galante@polito.it \
    --cc=hadi@cyberus.ca \
    --cc=jorge.finochietto@polito.it \
    --cc=mellia@prezzemolo.polito.it \
    --cc=netdev@oss.sgi.com \
    --cc=sfeldma@pobox.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).