* Re: [RFC: 2.6 patch] remove drivers/net/eepro100.c
From: David S. Miller @ 2005-11-18 9:08 UTC (permalink / raw)
To: rmk+lkml; +Cc: bunk, saw, jgarzik, linux-kernel, netdev
In-Reply-To: <20051118090158.GA11621@flint.arm.linux.org.uk>
From: Russell King <rmk+lkml@arm.linux.org.uk>
Date: Fri, 18 Nov 2005 09:01:59 +0000
> On Fri, Nov 18, 2005 at 04:33:02AM +0100, Adrian Bunk wrote:
> > This patch removes the obsolete drivers/net/eepro100.c driver.
> >
> > Is there any reason why it should be kept?
>
> Tt's the only driver which works correctly on ARM CPUs. e100 is
> basically buggy. This has been discussed here on lkml and more
> recently on linux-netdev. If anyone has any further questions
> please read the archives of those two lists.
Yes, sadly this is still true and therefore the eepro100 driver
needs to stick around.
^ permalink raw reply
* Re: [RFC: 2.6 patch] remove drivers/net/eepro100.c
From: Russell King @ 2005-11-18 9:01 UTC (permalink / raw)
To: Adrian Bunk; +Cc: saw, jgarzik, linux-kernel, netdev
In-Reply-To: <20051118033302.GO11494@stusta.de>
On Fri, Nov 18, 2005 at 04:33:02AM +0100, Adrian Bunk wrote:
> This patch removes the obsolete drivers/net/eepro100.c driver.
>
> Is there any reason why it should be kept?
Tt's the only driver which works correctly on ARM CPUs. e100 is
basically buggy. This has been discussed here on lkml and more
recently on linux-netdev. If anyone has any further questions
please read the archives of those two lists.
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
^ permalink raw reply
* [2.6 patch] move some code to net/ipx/af_ipx.c
From: Adrian Bunk @ 2005-11-18 5:22 UTC (permalink / raw)
To: Matt Mackall, acme; +Cc: Andrew Morton, linux-kernel, netdev
In-Reply-To: <20051114015707.GB5735@stusta.de>
On Mon, Nov 14, 2005 at 02:57:07AM +0100, Adrian Bunk wrote:
> On Fri, Nov 11, 2005 at 02:35:51AM -0600, Matt Mackall wrote:
> > trivial: drop unused 802.3 code if we compile without IPX
> >
> > (originally from http://wohnheim.fh-wedel.de/~joern/software/kernel/je/25/)
> >
> > Signed-off-by: Matt Mackall <mpm@selenic.com>
> >
> > Index: tiny/net/802/Makefile
> > ===================================================================
> > --- tiny.orig/net/802/Makefile 2005-03-15 00:24:59.000000000 -0600
> > +++ tiny/net/802/Makefile 2005-03-15 00:25:48.000000000 -0600
> > @@ -2,8 +2,6 @@
> > # Makefile for the Linux 802.x protocol layers.
> > #
> >
> > -obj-y := p8023.o
> > -
> > # Check the p8022 selections against net/core/Makefile.
> > obj-$(CONFIG_SYSCTL) += sysctl_net_802.o
> > obj-$(CONFIG_LLC) += p8022.o psnap.o
> > @@ -11,5 +9,5 @@ obj-$(CONFIG_TR) += p8022.o psnap.o tr.o
> > obj-$(CONFIG_NET_FC) += fc.o
> > obj-$(CONFIG_FDDI) += fddi.o
> > obj-$(CONFIG_HIPPI) += hippi.o
> > -obj-$(CONFIG_IPX) += p8022.o psnap.o
> > +obj-$(CONFIG_IPX) += p8022.o psnap.o p8023.o
> > obj-$(CONFIG_ATALK) += p8022.o psnap.o
>
> This patch isn't bad, but looking closer we could move the contents of
> p8023.c as well as the contents of at least p8022.c and pe2.c into
> af_ipx.c.
>
> Is the contents of any of these three files expected to be used
> outside IPX (closest candidate would be appletalk)?
Below is a patch implementing what I was thinking of.
cu
Adrian
<-- snip -->
This patch moves some code only used in this file to net/ipx/af_ipx.c .
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
include/net/p8022.h | 13 -----
net/802/Makefile | 17 ++-----
net/802/p8022.c | 66 ---------------------------
net/802/p8023.c | 61 -------------------------
net/8021q/vlan.c | 1
net/8021q/vlan_dev.c | 1
net/ethernet/Makefile | 2
net/ethernet/pe2.c | 39 ----------------
net/ipx/af_ipx.c | 102 ++++++++++++++++++++++++++++++++++++++++--
9 files changed, 106 insertions(+), 196 deletions(-)
--- linux-2.6.15-rc1-mm1-full/net/802/Makefile.old 2005-11-18 02:14:35.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/net/802/Makefile 2005-11-18 02:15:06.000000000 +0100
@@ -2,14 +2,11 @@
# Makefile for the Linux 802.x protocol layers.
#
-obj-y := p8023.o
-
-# Check the p8022 selections against net/core/Makefile.
obj-$(CONFIG_SYSCTL) += sysctl_net_802.o
-obj-$(CONFIG_LLC) += p8022.o psnap.o
-obj-$(CONFIG_TR) += p8022.o psnap.o tr.o sysctl_net_802.o
-obj-$(CONFIG_NET_FC) += fc.o
-obj-$(CONFIG_FDDI) += fddi.o
-obj-$(CONFIG_HIPPI) += hippi.o
-obj-$(CONFIG_IPX) += p8022.o psnap.o
-obj-$(CONFIG_ATALK) += p8022.o psnap.o
+obj-$(CONFIG_LLC) += psnap.o
+obj-$(CONFIG_TR) += psnap.o tr.o sysctl_net_802.o
+obj-$(CONFIG_NET_FC) += fc.o
+obj-$(CONFIG_FDDI) += fddi.o
+obj-$(CONFIG_HIPPI) += hippi.o
+obj-$(CONFIG_IPX) += psnap.o
+obj-$(CONFIG_ATALK) += psnap.o
--- linux-2.6.15-rc1-mm1-full/net/ethernet/Makefile.old 2005-11-18 02:15:17.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/net/ethernet/Makefile 2005-11-18 02:15:22.000000000 +0100
@@ -4,5 +4,3 @@
obj-y += eth.o
obj-$(CONFIG_SYSCTL) += sysctl_net_ether.o
-obj-$(subst m,y,$(CONFIG_IPX)) += pe2.o
-obj-$(subst m,y,$(CONFIG_ATALK)) += pe2.o
--- linux-2.6.15-rc1-mm1-full/net/8021q/vlan.c.old 2005-11-18 02:19:40.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/net/8021q/vlan.c 2005-11-18 02:19:46.000000000 +0100
@@ -26,7 +26,6 @@
#include <linux/mm.h>
#include <linux/in.h>
#include <linux/init.h>
-#include <net/p8022.h>
#include <net/arp.h>
#include <linux/rtnetlink.h>
#include <linux/notifier.h>
--- linux-2.6.15-rc1-mm1-full/net/8021q/vlan_dev.c.old 2005-11-18 02:19:55.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/net/8021q/vlan_dev.c 2005-11-18 02:19:58.000000000 +0100
@@ -29,7 +29,6 @@
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <net/datalink.h>
-#include <net/p8022.h>
#include <net/arp.h>
#include "vlan.h"
--- linux-2.6.15-rc1-mm1-full/net/ipx/af_ipx.c.old 2005-11-18 02:17:00.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/net/ipx/af_ipx.c 2005-11-18 02:26:01.000000000 +0100
@@ -48,10 +48,10 @@
#include <linux/termios.h>
#include <net/ipx.h>
-#include <net/p8022.h>
#include <net/psnap.h>
#include <net/sock.h>
#include <net/tcp_states.h>
+#include <net/llc.h>
#include <asm/uaccess.h>
@@ -1939,8 +1939,104 @@
.notifier_call = ipxitf_device_event,
};
-extern struct datalink_proto *make_EII_client(void);
-extern void destroy_EII_client(struct datalink_proto *);
+static int p8022_request(struct datalink_proto *dl, struct sk_buff *skb,
+ unsigned char *dest)
+{
+ llc_build_and_send_ui_pkt(dl->sap, skb, dest, dl->sap->laddr.lsap);
+ return 0;
+}
+
+static struct datalink_proto *register_8022_client(unsigned char type,
+ int (*func)(struct sk_buff *skb,
+ struct net_device *dev,
+ struct packet_type *pt,
+ struct net_device *orig_dev))
+{
+ struct datalink_proto *proto;
+
+ proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
+ if (proto) {
+ proto->type[0] = type;
+ proto->header_length = 3;
+ proto->request = p8022_request;
+ proto->sap = llc_sap_open(type, func);
+ if (!proto->sap) {
+ kfree(proto);
+ proto = NULL;
+ }
+ }
+ return proto;
+}
+
+static void unregister_8022_client(struct datalink_proto *proto)
+{
+ llc_sap_put(proto->sap);
+ kfree(proto);
+}
+
+/*
+ * Place an 802.3 header on a packet. The driver will do the mac
+ * addresses, we just need to give it the buffer length.
+ */
+static int p8023_request(struct datalink_proto *dl,
+ struct sk_buff *skb, unsigned char *dest_node)
+{
+ struct net_device *dev = skb->dev;
+
+ dev->hard_header(skb, dev, ETH_P_802_3, dest_node, NULL, skb->len);
+ return dev_queue_xmit(skb);
+}
+
+/*
+ * Create an 802.3 client. Note there can be only one 802.3 client
+ */
+static struct datalink_proto *make_8023_client(void)
+{
+ struct datalink_proto *proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
+
+ if (proto) {
+ proto->header_length = 0;
+ proto->request = p8023_request;
+ }
+ return proto;
+}
+
+/*
+ * Destroy the 802.3 client.
+ */
+static void destroy_8023_client(struct datalink_proto *dl)
+{
+ kfree(dl);
+}
+
+static int pEII_request(struct datalink_proto *dl,
+ struct sk_buff *skb, unsigned char *dest_node)
+{
+ struct net_device *dev = skb->dev;
+
+ skb->protocol = htons(ETH_P_IPX);
+ if (dev->hard_header)
+ dev->hard_header(skb, dev, ETH_P_IPX,
+ dest_node, NULL, skb->len);
+ return dev_queue_xmit(skb);
+}
+
+static struct datalink_proto *make_EII_client(void)
+{
+ struct datalink_proto *proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
+
+ if (proto) {
+ proto->header_length = 0;
+ proto->request = pEII_request;
+ }
+
+ return proto;
+}
+
+static void destroy_EII_client(struct datalink_proto *dl)
+{
+ kfree(dl);
+}
static unsigned char ipx_8022_type = 0xE0;
static unsigned char ipx_snap_id[5] = { 0x0, 0x0, 0x0, 0x81, 0x37 };
--- linux-2.6.15-rc1-mm1-full/include/net/p8022.h 2005-10-28 02:02:08.000000000 +0200
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,13 +0,0 @@
-#ifndef _NET_P8022_H
-#define _NET_P8022_H
-extern struct datalink_proto *
- register_8022_client(unsigned char type,
- int (*func)(struct sk_buff *skb,
- struct net_device *dev,
- struct packet_type *pt,
- struct net_device *orig_dev));
-extern void unregister_8022_client(struct datalink_proto *proto);
-
-extern struct datalink_proto *make_8023_client(void);
-extern void destroy_8023_client(struct datalink_proto *dl);
-#endif
--- linux-2.6.15-rc1-mm1-full/net/ethernet/pe2.c 2005-11-17 21:30:56.000000000 +0100
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,39 +0,0 @@
-#include <linux/in.h>
-#include <linux/mm.h>
-#include <linux/module.h>
-#include <linux/netdevice.h>
-#include <linux/skbuff.h>
-
-#include <net/datalink.h>
-
-static int pEII_request(struct datalink_proto *dl,
- struct sk_buff *skb, unsigned char *dest_node)
-{
- struct net_device *dev = skb->dev;
-
- skb->protocol = htons(ETH_P_IPX);
- if (dev->hard_header)
- dev->hard_header(skb, dev, ETH_P_IPX,
- dest_node, NULL, skb->len);
- return dev_queue_xmit(skb);
-}
-
-struct datalink_proto *make_EII_client(void)
-{
- struct datalink_proto *proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
-
- if (proto) {
- proto->header_length = 0;
- proto->request = pEII_request;
- }
-
- return proto;
-}
-
-void destroy_EII_client(struct datalink_proto *dl)
-{
- kfree(dl);
-}
-
-EXPORT_SYMBOL(destroy_EII_client);
-EXPORT_SYMBOL(make_EII_client);
--- linux-2.6.15-rc1-mm1-full/net/802/p8022.c 2005-10-28 02:02:08.000000000 +0200
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,66 +0,0 @@
-/*
- * NET3: Support for 802.2 demultiplexing off Ethernet (Token ring
- * is kept separate see p8022tr.c)
- * 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.
- *
- * Demultiplex 802.2 encoded protocols. We match the entry by the
- * SSAP/DSAP pair and then deliver to the registered datalink that
- * matches. The control byte is ignored and handling of such items
- * is up to the routine passed the frame.
- *
- * Unlike the 802.3 datalink we have a list of 802.2 entries as
- * there are multiple protocols to demux. The list is currently
- * short (3 or 4 entries at most). The current demux assumes this.
- */
-#include <linux/module.h>
-#include <linux/netdevice.h>
-#include <linux/skbuff.h>
-#include <net/datalink.h>
-#include <linux/mm.h>
-#include <linux/in.h>
-#include <linux/init.h>
-#include <net/llc.h>
-#include <net/p8022.h>
-
-static int p8022_request(struct datalink_proto *dl, struct sk_buff *skb,
- unsigned char *dest)
-{
- llc_build_and_send_ui_pkt(dl->sap, skb, dest, dl->sap->laddr.lsap);
- return 0;
-}
-
-struct datalink_proto *register_8022_client(unsigned char type,
- int (*func)(struct sk_buff *skb,
- struct net_device *dev,
- struct packet_type *pt,
- struct net_device *orig_dev))
-{
- struct datalink_proto *proto;
-
- proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
- if (proto) {
- proto->type[0] = type;
- proto->header_length = 3;
- proto->request = p8022_request;
- proto->sap = llc_sap_open(type, func);
- if (!proto->sap) {
- kfree(proto);
- proto = NULL;
- }
- }
- return proto;
-}
-
-void unregister_8022_client(struct datalink_proto *proto)
-{
- llc_sap_put(proto->sap);
- kfree(proto);
-}
-
-EXPORT_SYMBOL(register_8022_client);
-EXPORT_SYMBOL(unregister_8022_client);
-
-MODULE_LICENSE("GPL");
--- linux-2.6.15-rc1-mm1-full/net/802/p8023.c 2005-11-17 21:30:55.000000000 +0100
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,61 +0,0 @@
-/*
- * NET3: 802.3 data link hooks used for IPX 802.3
- *
- * 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.
- *
- * 802.3 isn't really a protocol data link layer. Some old IPX stuff
- * uses it however. Note that there is only one 802.3 protocol layer
- * in the system. We don't currently support different protocols
- * running raw 802.3 on different devices. Thankfully nobody else
- * has done anything like the old IPX.
- */
-
-#include <linux/in.h>
-#include <linux/mm.h>
-#include <linux/module.h>
-#include <linux/netdevice.h>
-#include <linux/skbuff.h>
-
-#include <net/datalink.h>
-#include <net/p8022.h>
-
-/*
- * Place an 802.3 header on a packet. The driver will do the mac
- * addresses, we just need to give it the buffer length.
- */
-static int p8023_request(struct datalink_proto *dl,
- struct sk_buff *skb, unsigned char *dest_node)
-{
- struct net_device *dev = skb->dev;
-
- dev->hard_header(skb, dev, ETH_P_802_3, dest_node, NULL, skb->len);
- return dev_queue_xmit(skb);
-}
-
-/*
- * Create an 802.3 client. Note there can be only one 802.3 client
- */
-struct datalink_proto *make_8023_client(void)
-{
- struct datalink_proto *proto = kmalloc(sizeof(*proto), GFP_ATOMIC);
-
- if (proto) {
- proto->header_length = 0;
- proto->request = p8023_request;
- }
- return proto;
-}
-
-/*
- * Destroy the 802.3 client.
- */
-void destroy_8023_client(struct datalink_proto *dl)
-{
- kfree(dl);
-}
-
-EXPORT_SYMBOL(destroy_8023_client);
-EXPORT_SYMBOL(make_8023_client);
^ permalink raw reply
* Re: [Orinoco-devel] [2.6 patch] drivers/net/wireless/orinoco.h: "extern inline" -> "static inline"
From: Adrian Bunk @ 2005-11-18 4:00 UTC (permalink / raw)
To: proski, orinoco-devel, linux-kernel, jgarzik, netdev
In-Reply-To: <20051118035236.GB23760@localhost.localdomain>
On Fri, Nov 18, 2005 at 02:52:36PM +1100, David Gibson wrote:
> On Fri, Nov 18, 2005 at 04:33:29AM +0100, Adrian Bunk wrote:
> > "extern inline" doesn't make much sense.
>
> Yes it does. "extern inline" tells gcc not to fall back to out of
> line version if it can't inline the function. These functions *must*
> by inlined, or they'll break horribly on Sparc, at least.
>...
For any non-ancient gcc and !alpha, we are already telling gcc via
__attribute__((always_inline)) that it should abort compilation if it
can't inline the function.
The problem is that "extern inline" gives warnings with
-Wmissing-prototypes, and I want to add this flag to the CFLAGS since it
helps us to avoid a class of nasty runtime errors.
> David Gibson
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply
* Re: [2.6 patch] drivers/net/wireless/orinoco.h: "extern inline" -> "static inline"
From: David Gibson @ 2005-11-18 3:52 UTC (permalink / raw)
To: Adrian Bunk
Cc: proski-mXXj517/zsQ,
orinoco-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
jgarzik-e+AXbWqSrlAAvxtiuMwx3w, netdev-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20051118033329.GU11494-HeJ8Db2Gnd6zQB+pC5nmwQ@public.gmane.org>
On Fri, Nov 18, 2005 at 04:33:29AM +0100, Adrian Bunk wrote:
> "extern inline" doesn't make much sense.
Yes it does. "extern inline" tells gcc not to fall back to out of
line version if it can't inline the function. These functions *must*
by inlined, or they'll break horribly on Sparc, at least.
> --- linux-2.6.15-rc1-mm1-full/drivers/net/wireless/orinoco.h.old 2005-11-18 02:38:43.000000000 +0100
> +++ linux-2.6.15-rc1-mm1-full/drivers/net/wireless/orinoco.h 2005-11-18 02:38:47.000000000 +0100
> @@ -155,7 +155,7 @@
> * SPARC, due to its weird semantics for save/restore flags. extern
> * inline should prevent the kernel from linking or module from
> * loading if they are not inlined. */
> -extern inline int orinoco_lock(struct orinoco_private *priv,
> +static inline int orinoco_lock(struct orinoco_private *priv,
> unsigned long *flags)
> {
> spin_lock_irqsave(&priv->lock, *flags);
> @@ -168,7 +168,7 @@
> return 0;
> }
>
> -extern inline void orinoco_unlock(struct orinoco_private *priv,
> +static inline void orinoco_unlock(struct orinoco_private *priv,
> unsigned long *flags)
> {
> spin_unlock_irqrestore(&priv->lock, *flags);
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today
Register for a JBoss Training Course. Free Certification Exam
for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
^ permalink raw reply
* [2.6 patch] drivers/net/wireless/orinoco.h: "extern inline" -> "static inline"
From: Adrian Bunk @ 2005-11-18 3:33 UTC (permalink / raw)
To: proski, hermes; +Cc: orinoco-devel, linux-kernel, jgarzik, netdev
"extern inline" doesn't make much sense.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
drivers/net/wireless/orinoco.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- linux-2.6.15-rc1-mm1-full/drivers/net/wireless/orinoco.h.old 2005-11-18 02:38:43.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/wireless/orinoco.h 2005-11-18 02:38:47.000000000 +0100
@@ -155,7 +155,7 @@
* SPARC, due to its weird semantics for save/restore flags. extern
* inline should prevent the kernel from linking or module from
* loading if they are not inlined. */
-extern inline int orinoco_lock(struct orinoco_private *priv,
+static inline int orinoco_lock(struct orinoco_private *priv,
unsigned long *flags)
{
spin_lock_irqsave(&priv->lock, *flags);
@@ -168,7 +168,7 @@
return 0;
}
-extern inline void orinoco_unlock(struct orinoco_private *priv,
+static inline void orinoco_unlock(struct orinoco_private *priv,
unsigned long *flags)
{
spin_unlock_irqrestore(&priv->lock, *flags);
^ permalink raw reply
* [2.6 patch] drivers/net/gianfar.h: "extern inline" -> "static inline"
From: Adrian Bunk @ 2005-11-18 3:33 UTC (permalink / raw)
To: jgarzik; +Cc: netdev, linux-kernel
"extern inline" doesn't make much sense.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
drivers/net/gianfar.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- linux-2.6.15-rc1-mm1-full/drivers/net/gianfar.h.old 2005-11-18 02:38:02.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/gianfar.h 2005-11-18 02:38:10.000000000 +0100
@@ -711,14 +711,14 @@
uint32_t msg_enable;
};
-extern inline u32 gfar_read(volatile unsigned *addr)
+static inline u32 gfar_read(volatile unsigned *addr)
{
u32 val;
val = in_be32(addr);
return val;
}
-extern inline void gfar_write(volatile unsigned *addr, u32 val)
+static inline void gfar_write(volatile unsigned *addr, u32 val)
{
out_be32(addr, val);
}
^ permalink raw reply
* [2.6 patch] drivers/net/bonding/bonding.h: "extern inline" -> "static inline"
From: Adrian Bunk @ 2005-11-18 3:33 UTC (permalink / raw)
To: ctindel, fubar; +Cc: bonding-devel, linux-kernel, netdev
"extern inline" doesn't make much sense.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
drivers/net/bonding/bonding.h | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
--- linux-2.6.15-rc1-mm1-full/drivers/net/bonding/bonding.h.old 2005-11-18 02:36:37.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/bonding/bonding.h 2005-11-18 02:37:05.000000000 +0100
@@ -218,7 +218,7 @@
*
* Caller must hold bond lock for read
*/
-extern inline struct slave *bond_get_slave_by_dev(struct bonding *bond, struct net_device *slave_dev)
+static inline struct slave *bond_get_slave_by_dev(struct bonding *bond, struct net_device *slave_dev)
{
struct slave *slave = NULL;
int i;
@@ -232,7 +232,7 @@
return slave;
}
-extern inline struct bonding *bond_get_bond_by_slave(struct slave *slave)
+static inline struct bonding *bond_get_bond_by_slave(struct slave *slave)
{
if (!slave || !slave->dev->master) {
return NULL;
@@ -241,13 +241,13 @@
return (struct bonding *)slave->dev->master->priv;
}
-extern inline void bond_set_slave_inactive_flags(struct slave *slave)
+static inline void bond_set_slave_inactive_flags(struct slave *slave)
{
slave->state = BOND_STATE_BACKUP;
slave->dev->flags |= IFF_NOARP;
}
-extern inline void bond_set_slave_active_flags(struct slave *slave)
+static inline void bond_set_slave_active_flags(struct slave *slave)
{
slave->state = BOND_STATE_ACTIVE;
slave->dev->flags &= ~IFF_NOARP;
^ permalink raw reply
* [RFC: 2.6 patch] remove drivers/net/tulip/xircom_tulip_cb.c
From: Adrian Bunk @ 2005-11-18 3:33 UTC (permalink / raw)
To: jgarzik; +Cc: tulip-users, linux-kernel, netdev
This patch removes the obsolete drivers/net/tulip/xircom_tulip_cb.c
driver.
Is there any reason why it should be kept?
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
drivers/net/tulip/Kconfig | 16
drivers/net/tulip/Makefile | 1
drivers/net/tulip/xircom_tulip_cb.c | 1748 ----------------------------
3 files changed, 1 insertion(+), 1764 deletions(-)
--- linux-2.6.15-rc1-mm1-full/drivers/net/tulip/Kconfig.old 2005-11-18 03:45:53.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/tulip/Kconfig 2005-11-18 03:46:20.000000000 +0100
@@ -148,7 +148,7 @@
be called uli526x.
config PCMCIA_XIRCOM
- tristate "Xircom CardBus support (new driver)"
+ tristate "Xircom CardBus support"
depends on NET_TULIP && CARDBUS
---help---
This driver is for the Digital "Tulip" Ethernet CardBus adapters.
@@ -160,19 +160,5 @@
<file:Documentation/networking/net-modules.txt>. The module will
be called xircom_cb. If unsure, say N.
-config PCMCIA_XIRTULIP
- tristate "Xircom Tulip-like CardBus support (old driver)"
- depends on NET_TULIP && CARDBUS && BROKEN_ON_SMP
- select CRC32
- ---help---
- This driver is for the Digital "Tulip" Ethernet CardBus adapters.
- It should work with most DEC 21*4*-based chips/ethercards, as well
- as with work-alike chips from Lite-On (PNIC) and Macronix (MXIC) and
- ASIX.
-
- To compile this driver as a module, choose M here and read
- <file:Documentation/networking/net-modules.txt>. The module will
- be called xircom_tulip_cb. If unsure, say N.
-
endmenu
--- linux-2.6.15-rc1-mm1-full/drivers/net/tulip/Makefile.old 2005-11-18 03:46:32.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/tulip/Makefile 2005-11-18 03:46:41.000000000 +0100
@@ -2,7 +2,6 @@
# Makefile for the Linux "Tulip" family network device drivers.
#
-obj-$(CONFIG_PCMCIA_XIRTULIP) += xircom_tulip_cb.o
obj-$(CONFIG_PCMCIA_XIRCOM) += xircom_cb.o
obj-$(CONFIG_DM9102) += dmfe.o
obj-$(CONFIG_WINBOND_840) += winbond-840.o
--- linux-2.6.15-rc1-mm1-full/drivers/net/tulip/xircom_tulip_cb.c 2005-10-28 02:02:08.000000000 +0200
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,1748 +0,0 @@
-/* xircom_tulip_cb.c: A Xircom CBE-100 ethernet driver for Linux. */
-/*
- Written/copyright 1994-1999 by Donald Becker.
-
- This software may be used and distributed according to the terms
- of the GNU General Public License, incorporated herein by reference.
-
- The author may be reached as becker@scyld.com, or C/O
- Scyld Computing Corporation
- 410 Severn Ave., Suite 210
- Annapolis MD 21403
-
- -----------------------------------------------------------
-
- Linux kernel-specific changes:
-
- LK1.0 (Ion Badulescu)
- - Major cleanup
- - Use 2.4 PCI API
- - Support ethtool
- - Rewrite perfect filter/hash code
- - Use interrupts for media changes
-
- LK1.1 (Ion Badulescu)
- - Disallow negotiation of unsupported full-duplex modes
-*/
-
-#define DRV_NAME "xircom_tulip_cb"
-#define DRV_VERSION "0.91+LK1.1"
-#define DRV_RELDATE "October 11, 2001"
-
-#define CARDBUS 1
-
-/* A few user-configurable values. */
-
-#define xircom_debug debug
-#ifdef XIRCOM_DEBUG
-static int xircom_debug = XIRCOM_DEBUG;
-#else
-static int xircom_debug = 1;
-#endif
-
-/* Maximum events (Rx packets, etc.) to handle at each interrupt. */
-static int max_interrupt_work = 25;
-
-#define MAX_UNITS 4
-/* Used to pass the full-duplex flag, etc. */
-static int full_duplex[MAX_UNITS];
-static int options[MAX_UNITS];
-static int mtu[MAX_UNITS]; /* Jumbo MTU for interfaces. */
-
-/* Keep the ring sizes a power of two for efficiency.
- Making the Tx ring too large decreases the effectiveness of channel
- bonding and packet priority.
- There are no ill effects from too-large receive rings. */
-#define TX_RING_SIZE 16
-#define RX_RING_SIZE 32
-
-/* Set the copy breakpoint for the copy-only-tiny-buffer Rx structure. */
-#ifdef __alpha__
-static int rx_copybreak = 1518;
-#else
-static int rx_copybreak = 100;
-#endif
-
-/*
- Set the bus performance register.
- Typical: Set 16 longword cache alignment, no burst limit.
- Cache alignment bits 15:14 Burst length 13:8
- 0000 No alignment 0x00000000 unlimited 0800 8 longwords
- 4000 8 longwords 0100 1 longword 1000 16 longwords
- 8000 16 longwords 0200 2 longwords 2000 32 longwords
- C000 32 longwords 0400 4 longwords
- Warning: many older 486 systems are broken and require setting 0x00A04800
- 8 longword cache alignment, 8 longword burst.
- ToDo: Non-Intel setting could be better.
-*/
-
-#if defined(__alpha__) || defined(__ia64__) || defined(__x86_64__)
-static int csr0 = 0x01A00000 | 0xE000;
-#elif defined(__powerpc__)
-static int csr0 = 0x01B00000 | 0x8000;
-#elif defined(__sparc__)
-static int csr0 = 0x01B00080 | 0x8000;
-#elif defined(__i386__)
-static int csr0 = 0x01A00000 | 0x8000;
-#else
-#warning Processor architecture undefined!
-static int csr0 = 0x00A00000 | 0x4800;
-#endif
-
-/* Operational parameters that usually are not changed. */
-/* Time in jiffies before concluding the transmitter is hung. */
-#define TX_TIMEOUT (4 * HZ)
-#define PKT_BUF_SZ 1536 /* Size of each temporary Rx buffer.*/
-#define PKT_SETUP_SZ 192 /* Size of the setup frame */
-
-/* PCI registers */
-#define PCI_POWERMGMT 0x40
-
-#include <linux/config.h>
-#include <linux/module.h>
-#include <linux/moduleparam.h>
-#include <linux/kernel.h>
-#include <linux/pci.h>
-#include <linux/netdevice.h>
-#include <linux/etherdevice.h>
-#include <linux/delay.h>
-#include <linux/init.h>
-#include <linux/mii.h>
-#include <linux/ethtool.h>
-#include <linux/crc32.h>
-
-#include <asm/io.h>
-#include <asm/processor.h> /* Processor type for cache alignment. */
-#include <asm/uaccess.h>
-
-
-/* These identify the driver base version and may not be removed. */
-static char version[] __devinitdata =
-KERN_INFO DRV_NAME ".c derived from tulip.c:v0.91 4/14/99 becker@scyld.com\n"
-KERN_INFO " unofficial 2.4.x kernel port, version " DRV_VERSION ", " DRV_RELDATE "\n";
-
-MODULE_AUTHOR("Donald Becker <becker@scyld.com>");
-MODULE_DESCRIPTION("Xircom CBE-100 ethernet driver");
-MODULE_LICENSE("GPL v2");
-MODULE_VERSION(DRV_VERSION);
-
-module_param(debug, int, 0);
-module_param(max_interrupt_work, int, 0);
-module_param(rx_copybreak, int, 0);
-module_param(csr0, int, 0);
-
-module_param_array(options, int, NULL, 0);
-module_param_array(full_duplex, int, NULL, 0);
-
-#define RUN_AT(x) (jiffies + (x))
-
-/*
- Theory of Operation
-
-I. Board Compatibility
-
-This device driver was forked from the driver for the DECchip "Tulip",
-Digital's single-chip ethernet controllers for PCI. It supports Xircom's
-almost-Tulip-compatible CBE-100 CardBus adapters.
-
-II. Board-specific settings
-
-PCI bus devices are configured by the system at boot time, so no jumpers
-need to be set on the board. The system BIOS preferably should assign the
-PCI INTA signal to an otherwise unused system IRQ line.
-
-III. Driver operation
-
-IIIa. Ring buffers
-
-The Xircom can use either ring buffers or lists of Tx and Rx descriptors.
-This driver uses statically allocated rings of Rx and Tx descriptors, set at
-compile time by RX/TX_RING_SIZE. This version of the driver allocates skbuffs
-for the Rx ring buffers at open() time and passes the skb->data field to the
-Xircom as receive data buffers. When an incoming frame is less than
-RX_COPYBREAK bytes long, a fresh skbuff is allocated and the frame is
-copied to the new skbuff. When the incoming frame is larger, the skbuff is
-passed directly up the protocol stack and replaced by a newly allocated
-skbuff.
-
-The RX_COPYBREAK value is chosen to trade-off the memory wasted by
-using a full-sized skbuff for small frames vs. the copying costs of larger
-frames. For small frames the copying cost is negligible (esp. considering
-that we are pre-loading the cache with immediately useful header
-information). For large frames the copying cost is non-trivial, and the
-larger copy might flush the cache of useful data. A subtle aspect of this
-choice is that the Xircom only receives into longword aligned buffers, thus
-the IP header at offset 14 isn't longword aligned for further processing.
-Copied frames are put into the new skbuff at an offset of "+2", thus copying
-has the beneficial effect of aligning the IP header and preloading the
-cache.
-
-IIIC. Synchronization
-The driver runs as two independent, single-threaded flows of control. One
-is the send-packet routine, which enforces single-threaded use by the
-dev->tbusy flag. The other thread is the interrupt handler, which is single
-threaded by the hardware and other software.
-
-The send packet thread has partial control over the Tx ring and 'dev->tbusy'
-flag. It sets the tbusy flag whenever it's queuing a Tx packet. If the next
-queue slot is empty, it clears the tbusy flag when finished otherwise it sets
-the 'tp->tx_full' flag.
-
-The interrupt handler has exclusive control over the Rx ring and records stats
-from the Tx ring. (The Tx-done interrupt can't be selectively turned off, so
-we can't avoid the interrupt overhead by having the Tx routine reap the Tx
-stats.) After reaping the stats, it marks the queue entry as empty by setting
-the 'base' to zero. Iff the 'tp->tx_full' flag is set, it clears both the
-tx_full and tbusy flags.
-
-IV. Notes
-
-IVb. References
-
-http://cesdis.gsfc.nasa.gov/linux/misc/NWay.html
-http://www.digital.com (search for current 21*4* datasheets and "21X4 SROM")
-http://www.national.com/pf/DP/DP83840A.html
-
-IVc. Errata
-
-*/
-
-/* A full-duplex map for media types. */
-enum MediaIs {
- MediaIsFD = 1, MediaAlwaysFD=2, MediaIsMII=4, MediaIsFx=8,
- MediaIs100=16};
-static const char media_cap[] =
-{0,0,0,16, 3,19,16,24, 27,4,7,5, 0,20,23,20 };
-
-/* Offsets to the Command and Status Registers, "CSRs". All accesses
- must be longword instructions and quadword aligned. */
-enum xircom_offsets {
- CSR0=0, CSR1=0x08, CSR2=0x10, CSR3=0x18, CSR4=0x20, CSR5=0x28,
- CSR6=0x30, CSR7=0x38, CSR8=0x40, CSR9=0x48, CSR10=0x50, CSR11=0x58,
- CSR12=0x60, CSR13=0x68, CSR14=0x70, CSR15=0x78, CSR16=0x04, };
-
-/* The bits in the CSR5 status registers, mostly interrupt sources. */
-enum status_bits {
- LinkChange=0x08000000,
- NormalIntr=0x10000, NormalIntrMask=0x00014045,
- AbnormalIntr=0x8000, AbnormalIntrMask=0x0a00a5a2,
- ReservedIntrMask=0xe0001a18,
- EarlyRxIntr=0x4000, BusErrorIntr=0x2000,
- EarlyTxIntr=0x400, RxDied=0x100, RxNoBuf=0x80, RxIntr=0x40,
- TxFIFOUnderflow=0x20, TxNoBuf=0x04, TxDied=0x02, TxIntr=0x01,
-};
-
-enum csr0_control_bits {
- EnableMWI=0x01000000, EnableMRL=0x00800000,
- EnableMRM=0x00200000, EqualBusPrio=0x02,
- SoftwareReset=0x01,
-};
-
-enum csr6_control_bits {
- ReceiveAllBit=0x40000000, AllMultiBit=0x80, PromiscBit=0x40,
- HashFilterBit=0x01, FullDuplexBit=0x0200,
- TxThresh10=0x400000, TxStoreForw=0x200000,
- TxThreshMask=0xc000, TxThreshShift=14,
- EnableTx=0x2000, EnableRx=0x02,
- ReservedZeroMask=0x8d930134, ReservedOneMask=0x320c0000,
- EnableTxRx=(EnableTx | EnableRx),
-};
-
-
-enum tbl_flag {
- HAS_MII=1, HAS_ACPI=2,
-};
-static struct xircom_chip_table {
- char *chip_name;
- int valid_intrs; /* CSR7 interrupt enable settings */
- int flags;
-} xircom_tbl[] = {
- { "Xircom Cardbus Adapter",
- LinkChange | NormalIntr | AbnormalIntr | BusErrorIntr |
- RxDied | RxNoBuf | RxIntr | TxFIFOUnderflow | TxNoBuf | TxDied | TxIntr,
- HAS_MII | HAS_ACPI, },
- { NULL, },
-};
-/* This matches the table above. */
-enum chips {
- X3201_3,
-};
-
-
-/* The Xircom Rx and Tx buffer descriptors. */
-struct xircom_rx_desc {
- s32 status;
- s32 length;
- u32 buffer1, buffer2;
-};
-
-struct xircom_tx_desc {
- s32 status;
- s32 length;
- u32 buffer1, buffer2; /* We use only buffer 1. */
-};
-
-enum tx_desc0_status_bits {
- Tx0DescOwned=0x80000000, Tx0DescError=0x8000, Tx0NoCarrier=0x0800,
- Tx0LateColl=0x0200, Tx0ManyColl=0x0100, Tx0Underflow=0x02,
-};
-enum tx_desc1_status_bits {
- Tx1ComplIntr=0x80000000, Tx1LastSeg=0x40000000, Tx1FirstSeg=0x20000000,
- Tx1SetupPkt=0x08000000, Tx1DisableCRC=0x04000000, Tx1RingWrap=0x02000000,
- Tx1ChainDesc=0x01000000, Tx1NoPad=0x800000, Tx1HashSetup=0x400000,
- Tx1WholePkt=(Tx1FirstSeg | Tx1LastSeg),
-};
-enum rx_desc0_status_bits {
- Rx0DescOwned=0x80000000, Rx0DescError=0x8000, Rx0NoSpace=0x4000,
- Rx0Runt=0x0800, Rx0McastPkt=0x0400, Rx0FirstSeg=0x0200, Rx0LastSeg=0x0100,
- Rx0HugeFrame=0x80, Rx0CRCError=0x02,
- Rx0WholePkt=(Rx0FirstSeg | Rx0LastSeg),
-};
-enum rx_desc1_status_bits {
- Rx1RingWrap=0x02000000, Rx1ChainDesc=0x01000000,
-};
-
-struct xircom_private {
- struct xircom_rx_desc rx_ring[RX_RING_SIZE];
- struct xircom_tx_desc tx_ring[TX_RING_SIZE];
- /* The saved address of a sent-in-place packet/buffer, for skfree(). */
- struct sk_buff* tx_skbuff[TX_RING_SIZE];
-#ifdef CARDBUS
- /* The X3201-3 requires 4-byte aligned tx bufs */
- struct sk_buff* tx_aligned_skbuff[TX_RING_SIZE];
-#endif
- /* The addresses of receive-in-place skbuffs. */
- struct sk_buff* rx_skbuff[RX_RING_SIZE];
- u16 setup_frame[PKT_SETUP_SZ / sizeof(u16)]; /* Pseudo-Tx frame to init address table. */
- int chip_id;
- struct net_device_stats stats;
- unsigned int cur_rx, cur_tx; /* The next free ring entry */
- unsigned int dirty_rx, dirty_tx; /* The ring entries to be free()ed. */
- unsigned int tx_full:1; /* The Tx queue is full. */
- unsigned int speed100:1;
- unsigned int full_duplex:1; /* Full-duplex operation requested. */
- unsigned int autoneg:1;
- unsigned int default_port:4; /* Last dev->if_port value. */
- unsigned int open:1;
- unsigned int csr0; /* CSR0 setting. */
- unsigned int csr6; /* Current CSR6 control settings. */
- u16 to_advertise; /* NWay capabilities advertised. */
- u16 advertising[4];
- signed char phys[4], mii_cnt; /* MII device addresses. */
- int saved_if_port;
- struct pci_dev *pdev;
- spinlock_t lock;
-};
-
-static int mdio_read(struct net_device *dev, int phy_id, int location);
-static void mdio_write(struct net_device *dev, int phy_id, int location, int value);
-static void xircom_up(struct net_device *dev);
-static void xircom_down(struct net_device *dev);
-static int xircom_open(struct net_device *dev);
-static void xircom_tx_timeout(struct net_device *dev);
-static void xircom_init_ring(struct net_device *dev);
-static int xircom_start_xmit(struct sk_buff *skb, struct net_device *dev);
-static int xircom_rx(struct net_device *dev);
-static void xircom_media_change(struct net_device *dev);
-static irqreturn_t xircom_interrupt(int irq, void *dev_instance, struct pt_regs *regs);
-static int xircom_close(struct net_device *dev);
-static struct net_device_stats *xircom_get_stats(struct net_device *dev);
-static int xircom_ioctl(struct net_device *dev, struct ifreq *rq, int cmd);
-static void set_rx_mode(struct net_device *dev);
-static void check_duplex(struct net_device *dev);
-static struct ethtool_ops ops;
-
-
-/* The Xircom cards are picky about when certain bits in CSR6 can be
- manipulated. Keith Owens <kaos@ocs.com.au>. */
-static void outl_CSR6(u32 newcsr6, long ioaddr)
-{
- const int strict_bits =
- TxThresh10 | TxStoreForw | TxThreshMask | EnableTxRx | FullDuplexBit;
- int csr5, csr5_22_20, csr5_19_17, currcsr6, attempts = 200;
- unsigned long flags;
- save_flags(flags);
- cli();
- /* mask out the reserved bits that always read 0 on the Xircom cards */
- newcsr6 &= ~ReservedZeroMask;
- /* or in the reserved bits that always read 1 */
- newcsr6 |= ReservedOneMask;
- currcsr6 = inl(ioaddr + CSR6);
- if (((newcsr6 & strict_bits) == (currcsr6 & strict_bits)) ||
- ((currcsr6 & ~EnableTxRx) == 0)) {
- outl(newcsr6, ioaddr + CSR6); /* safe */
- restore_flags(flags);
- return;
- }
- /* make sure the transmitter and receiver are stopped first */
- currcsr6 &= ~EnableTxRx;
- while (1) {
- csr5 = inl(ioaddr + CSR5);
- if (csr5 == 0xffffffff)
- break; /* cannot read csr5, card removed? */
- csr5_22_20 = csr5 & 0x700000;
- csr5_19_17 = csr5 & 0x0e0000;
- if ((csr5_22_20 == 0 || csr5_22_20 == 0x600000) &&
- (csr5_19_17 == 0 || csr5_19_17 == 0x80000 || csr5_19_17 == 0xc0000))
- break; /* both are stopped or suspended */
- if (!--attempts) {
- printk(KERN_INFO DRV_NAME ": outl_CSR6 too many attempts,"
- "csr5=0x%08x\n", csr5);
- outl(newcsr6, ioaddr + CSR6); /* unsafe but do it anyway */
- restore_flags(flags);
- return;
- }
- outl(currcsr6, ioaddr + CSR6);
- udelay(1);
- }
- /* now it is safe to change csr6 */
- outl(newcsr6, ioaddr + CSR6);
- restore_flags(flags);
-}
-
-
-static void __devinit read_mac_address(struct net_device *dev)
-{
- long ioaddr = dev->base_addr;
- int i, j;
- unsigned char tuple, link, data_id, data_count;
-
- /* Xircom has its address stored in the CIS;
- * we access it through the boot rom interface for now
- * this might not work, as the CIS is not parsed but I
- * (danilo) use the offset I found on my card's CIS !!!
- *
- * Doug Ledford: I changed this routine around so that it
- * walks the CIS memory space, parsing the config items, and
- * finds the proper lan_node_id tuple and uses the data
- * stored there.
- */
- outl(1 << 12, ioaddr + CSR9); /* enable boot rom access */
- for (i = 0x100; i < 0x1f7; i += link+2) {
- outl(i, ioaddr + CSR10);
- tuple = inl(ioaddr + CSR9) & 0xff;
- outl(i + 1, ioaddr + CSR10);
- link = inl(ioaddr + CSR9) & 0xff;
- outl(i + 2, ioaddr + CSR10);
- data_id = inl(ioaddr + CSR9) & 0xff;
- outl(i + 3, ioaddr + CSR10);
- data_count = inl(ioaddr + CSR9) & 0xff;
- if ( (tuple == 0x22) &&
- (data_id == 0x04) && (data_count == 0x06) ) {
- /*
- * This is it. We have the data we want.
- */
- for (j = 0; j < 6; j++) {
- outl(i + j + 4, ioaddr + CSR10);
- dev->dev_addr[j] = inl(ioaddr + CSR9) & 0xff;
- }
- break;
- } else if (link == 0) {
- break;
- }
- }
-}
-
-
-/*
- * locate the MII interfaces and initialize them.
- * we disable full-duplex modes here,
- * because we don't know how to handle them.
- */
-static void find_mii_transceivers(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- int phy, phy_idx;
-
- if (media_cap[tp->default_port] & MediaIsMII) {
- u16 media2advert[] = { 0x20, 0x40, 0x03e0, 0x60, 0x80, 0x100, 0x200 };
- tp->to_advertise = media2advert[tp->default_port - 9];
- } else
- tp->to_advertise =
- /*ADVERTISE_100BASE4 | ADVERTISE_100FULL |*/ ADVERTISE_100HALF |
- /*ADVERTISE_10FULL |*/ ADVERTISE_10HALF | ADVERTISE_CSMA;
-
- /* Find the connected MII xcvrs.
- Doing this in open() would allow detecting external xcvrs later,
- but takes much time. */
- for (phy = 0, phy_idx = 0; phy < 32 && phy_idx < sizeof(tp->phys); phy++) {
- int mii_status = mdio_read(dev, phy, MII_BMSR);
- if ((mii_status & (BMSR_100BASE4 | BMSR_100HALF | BMSR_10HALF)) == BMSR_100BASE4 ||
- ((mii_status & BMSR_100BASE4) == 0 &&
- (mii_status & (BMSR_100FULL | BMSR_100HALF | BMSR_10FULL | BMSR_10HALF)) != 0)) {
- int mii_reg0 = mdio_read(dev, phy, MII_BMCR);
- int mii_advert = mdio_read(dev, phy, MII_ADVERTISE);
- int reg4 = ((mii_status >> 6) & tp->to_advertise) | ADVERTISE_CSMA;
- tp->phys[phy_idx] = phy;
- tp->advertising[phy_idx++] = reg4;
- printk(KERN_INFO "%s: MII transceiver #%d "
- "config %4.4x status %4.4x advertising %4.4x.\n",
- dev->name, phy, mii_reg0, mii_status, mii_advert);
- }
- }
- tp->mii_cnt = phy_idx;
- if (phy_idx == 0) {
- printk(KERN_INFO "%s: ***WARNING***: No MII transceiver found!\n",
- dev->name);
- tp->phys[0] = 0;
- }
-}
-
-
-/*
- * To quote Arjan van de Ven:
- * transceiver_voodoo() enables the external UTP plug thingy.
- * it's called voodoo as I stole this code and cannot cross-reference
- * it with the specification.
- * Actually it seems to go like this:
- * - GPIO2 enables the MII itself so we can talk to it. The MII gets reset
- * so any prior MII settings are lost.
- * - GPIO0 enables the TP port so the MII can talk to the network.
- * - a software reset will reset both GPIO pins.
- * I also moved the software reset here, because doing it in xircom_up()
- * required enabling the GPIO pins each time, which reset the MII each time.
- * Thus we couldn't control the MII -- which sucks because we don't know
- * how to handle full-duplex modes so we *must* disable them.
- */
-static void transceiver_voodoo(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
-
- /* Reset the chip, holding bit 0 set at least 50 PCI cycles. */
- outl(SoftwareReset, ioaddr + CSR0);
- udelay(2);
-
- /* Deassert reset. */
- outl(tp->csr0, ioaddr + CSR0);
-
- /* Reset the xcvr interface and turn on heartbeat. */
- outl(0x0008, ioaddr + CSR15);
- udelay(5); /* The delays are Xircom-recommended to give the
- * chipset time to reset the actual hardware
- * on the PCMCIA card
- */
- outl(0xa8050000, ioaddr + CSR15);
- udelay(5);
- outl(0xa00f0000, ioaddr + CSR15);
- udelay(5);
-
- outl_CSR6(0, ioaddr);
- //outl_CSR6(FullDuplexBit, ioaddr);
-}
-
-
-static int __devinit xircom_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
-{
- struct net_device *dev;
- struct xircom_private *tp;
- static int board_idx = -1;
- int chip_idx = id->driver_data;
- long ioaddr;
- int i;
- u8 chip_rev;
-
-/* when built into the kernel, we only print version if device is found */
-#ifndef MODULE
- static int printed_version;
- if (!printed_version++)
- printk(version);
-#endif
-
- //printk(KERN_INFO "xircom_init_one(%s)\n", pci_name(pdev));
-
- board_idx++;
-
- if (pci_enable_device(pdev))
- return -ENODEV;
-
- pci_set_master(pdev);
-
- ioaddr = pci_resource_start(pdev, 0);
- dev = alloc_etherdev(sizeof(*tp));
- if (!dev) {
- printk (KERN_ERR DRV_NAME "%d: cannot alloc etherdev, aborting\n", board_idx);
- return -ENOMEM;
- }
- SET_MODULE_OWNER(dev);
- SET_NETDEV_DEV(dev, &pdev->dev);
-
- dev->base_addr = ioaddr;
- dev->irq = pdev->irq;
-
- if (pci_request_regions(pdev, dev->name)) {
- printk (KERN_ERR DRV_NAME " %d: cannot reserve PCI resources, aborting\n", board_idx);
- goto err_out_free_netdev;
- }
-
- /* Bring the chip out of sleep mode.
- Caution: Snooze mode does not work with some boards! */
- if (xircom_tbl[chip_idx].flags & HAS_ACPI)
- pci_write_config_dword(pdev, PCI_POWERMGMT, 0);
-
- /* Stop the chip's Tx and Rx processes. */
- outl_CSR6(inl(ioaddr + CSR6) & ~EnableTxRx, ioaddr);
- /* Clear the missed-packet counter. */
- (volatile int)inl(ioaddr + CSR8);
-
- tp = netdev_priv(dev);
-
- spin_lock_init(&tp->lock);
- tp->pdev = pdev;
- tp->chip_id = chip_idx;
- /* BugFixes: The 21143-TD hangs with PCI Write-and-Invalidate cycles. */
- /* XXX: is this necessary for Xircom? */
- tp->csr0 = csr0 & ~EnableMWI;
-
- pci_set_drvdata(pdev, dev);
-
- /* The lower four bits are the media type. */
- if (board_idx >= 0 && board_idx < MAX_UNITS) {
- tp->default_port = options[board_idx] & 15;
- if ((options[board_idx] & 0x90) || full_duplex[board_idx] > 0)
- tp->full_duplex = 1;
- if (mtu[board_idx] > 0)
- dev->mtu = mtu[board_idx];
- }
- if (dev->mem_start)
- tp->default_port = dev->mem_start;
- if (tp->default_port) {
- if (media_cap[tp->default_port] & MediaAlwaysFD)
- tp->full_duplex = 1;
- }
- if (tp->full_duplex)
- tp->autoneg = 0;
- else
- tp->autoneg = 1;
- tp->speed100 = 1;
-
- /* The Xircom-specific entries in the device structure. */
- dev->open = &xircom_open;
- dev->hard_start_xmit = &xircom_start_xmit;
- dev->stop = &xircom_close;
- dev->get_stats = &xircom_get_stats;
- dev->do_ioctl = &xircom_ioctl;
-#ifdef HAVE_MULTICAST
- dev->set_multicast_list = &set_rx_mode;
-#endif
- dev->tx_timeout = xircom_tx_timeout;
- dev->watchdog_timeo = TX_TIMEOUT;
- SET_ETHTOOL_OPS(dev, &ops);
-
- transceiver_voodoo(dev);
-
- read_mac_address(dev);
-
- if (register_netdev(dev))
- goto err_out_cleardev;
-
- pci_read_config_byte(pdev, PCI_REVISION_ID, &chip_rev);
- printk(KERN_INFO "%s: %s rev %d at %#3lx,",
- dev->name, xircom_tbl[chip_idx].chip_name, chip_rev, ioaddr);
- for (i = 0; i < 6; i++)
- printk("%c%2.2X", i ? ':' : ' ', dev->dev_addr[i]);
- printk(", IRQ %d.\n", dev->irq);
-
- if (xircom_tbl[chip_idx].flags & HAS_MII) {
- find_mii_transceivers(dev);
- check_duplex(dev);
- }
-
- return 0;
-
-err_out_cleardev:
- pci_set_drvdata(pdev, NULL);
- pci_release_regions(pdev);
-err_out_free_netdev:
- free_netdev(dev);
- return -ENODEV;
-}
-
-
-/* MII transceiver control section.
- Read and write the MII registers using software-generated serial
- MDIO protocol. See the MII specifications or DP83840A data sheet
- for details. */
-
-/* The maximum data clock rate is 2.5 Mhz. The minimum timing is usually
- met by back-to-back PCI I/O cycles, but we insert a delay to avoid
- "overclocking" issues or future 66Mhz PCI. */
-#define mdio_delay() inl(mdio_addr)
-
-/* Read and write the MII registers using software-generated serial
- MDIO protocol. It is just different enough from the EEPROM protocol
- to not share code. The maxium data clock rate is 2.5 Mhz. */
-#define MDIO_SHIFT_CLK 0x10000
-#define MDIO_DATA_WRITE0 0x00000
-#define MDIO_DATA_WRITE1 0x20000
-#define MDIO_ENB 0x00000 /* Ignore the 0x02000 databook setting. */
-#define MDIO_ENB_IN 0x40000
-#define MDIO_DATA_READ 0x80000
-
-static int mdio_read(struct net_device *dev, int phy_id, int location)
-{
- int i;
- int read_cmd = (0xf6 << 10) | (phy_id << 5) | location;
- int retval = 0;
- long ioaddr = dev->base_addr;
- long mdio_addr = ioaddr + CSR9;
-
- /* Establish sync by sending at least 32 logic ones. */
- for (i = 32; i >= 0; i--) {
- outl(MDIO_ENB | MDIO_DATA_WRITE1, mdio_addr);
- mdio_delay();
- outl(MDIO_ENB | MDIO_DATA_WRITE1 | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- /* Shift the read command bits out. */
- for (i = 15; i >= 0; i--) {
- int dataval = (read_cmd & (1 << i)) ? MDIO_DATA_WRITE1 : 0;
-
- outl(MDIO_ENB | dataval, mdio_addr);
- mdio_delay();
- outl(MDIO_ENB | dataval | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- /* Read the two transition, 16 data, and wire-idle bits. */
- for (i = 19; i > 0; i--) {
- outl(MDIO_ENB_IN, mdio_addr);
- mdio_delay();
- retval = (retval << 1) | ((inl(mdio_addr) & MDIO_DATA_READ) ? 1 : 0);
- outl(MDIO_ENB_IN | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- return (retval>>1) & 0xffff;
-}
-
-
-static void mdio_write(struct net_device *dev, int phy_id, int location, int value)
-{
- int i;
- int cmd = (0x5002 << 16) | (phy_id << 23) | (location << 18) | value;
- long ioaddr = dev->base_addr;
- long mdio_addr = ioaddr + CSR9;
-
- /* Establish sync by sending 32 logic ones. */
- for (i = 32; i >= 0; i--) {
- outl(MDIO_ENB | MDIO_DATA_WRITE1, mdio_addr);
- mdio_delay();
- outl(MDIO_ENB | MDIO_DATA_WRITE1 | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- /* Shift the command bits out. */
- for (i = 31; i >= 0; i--) {
- int dataval = (cmd & (1 << i)) ? MDIO_DATA_WRITE1 : 0;
- outl(MDIO_ENB | dataval, mdio_addr);
- mdio_delay();
- outl(MDIO_ENB | dataval | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- /* Clear out extra bits. */
- for (i = 2; i > 0; i--) {
- outl(MDIO_ENB_IN, mdio_addr);
- mdio_delay();
- outl(MDIO_ENB_IN | MDIO_SHIFT_CLK, mdio_addr);
- mdio_delay();
- }
- return;
-}
-
-
-static void
-xircom_up(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
- int i;
-
- xircom_init_ring(dev);
- /* Clear the tx ring */
- for (i = 0; i < TX_RING_SIZE; i++) {
- tp->tx_skbuff[i] = NULL;
- tp->tx_ring[i].status = 0;
- }
-
- if (xircom_debug > 1)
- printk(KERN_DEBUG "%s: xircom_up() irq %d.\n", dev->name, dev->irq);
-
- outl(virt_to_bus(tp->rx_ring), ioaddr + CSR3);
- outl(virt_to_bus(tp->tx_ring), ioaddr + CSR4);
-
- tp->saved_if_port = dev->if_port;
- if (dev->if_port == 0)
- dev->if_port = tp->default_port;
-
- tp->csr6 = TxThresh10 /*| FullDuplexBit*/; /* XXX: why 10 and not 100? */
-
- set_rx_mode(dev);
-
- /* Start the chip's Tx to process setup frame. */
- outl_CSR6(tp->csr6, ioaddr);
- outl_CSR6(tp->csr6 | EnableTx, ioaddr);
-
- /* Acknowledge all outstanding interrupts sources */
- outl(xircom_tbl[tp->chip_id].valid_intrs, ioaddr + CSR5);
- /* Enable interrupts by setting the interrupt mask. */
- outl(xircom_tbl[tp->chip_id].valid_intrs, ioaddr + CSR7);
- /* Enable Rx */
- outl_CSR6(tp->csr6 | EnableTxRx, ioaddr);
- /* Rx poll demand */
- outl(0, ioaddr + CSR2);
-
- /* Tell the net layer we're ready */
- netif_start_queue (dev);
-
- /* Check current media state */
- xircom_media_change(dev);
-
- if (xircom_debug > 2) {
- printk(KERN_DEBUG "%s: Done xircom_up(), CSR0 %8.8x, CSR5 %8.8x CSR6 %8.8x.\n",
- dev->name, inl(ioaddr + CSR0), inl(ioaddr + CSR5),
- inl(ioaddr + CSR6));
- }
-}
-
-
-static int
-xircom_open(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
-
- if (request_irq(dev->irq, &xircom_interrupt, SA_SHIRQ, dev->name, dev))
- return -EAGAIN;
-
- xircom_up(dev);
- tp->open = 1;
-
- return 0;
-}
-
-
-static void xircom_tx_timeout(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
-
- if (media_cap[dev->if_port] & MediaIsMII) {
- /* Do nothing -- the media monitor should handle this. */
- if (xircom_debug > 1)
- printk(KERN_WARNING "%s: Transmit timeout using MII device.\n",
- dev->name);
- }
-
-#if defined(way_too_many_messages)
- if (xircom_debug > 3) {
- int i;
- for (i = 0; i < RX_RING_SIZE; i++) {
- u8 *buf = (u8 *)(tp->rx_ring[i].buffer1);
- int j;
- printk(KERN_DEBUG "%2d: %8.8x %8.8x %8.8x %8.8x "
- "%2.2x %2.2x %2.2x.\n",
- i, (unsigned int)tp->rx_ring[i].status,
- (unsigned int)tp->rx_ring[i].length,
- (unsigned int)tp->rx_ring[i].buffer1,
- (unsigned int)tp->rx_ring[i].buffer2,
- buf[0], buf[1], buf[2]);
- for (j = 0; buf[j] != 0xee && j < 1600; j++)
- if (j < 100) printk(" %2.2x", buf[j]);
- printk(" j=%d.\n", j);
- }
- printk(KERN_DEBUG " Rx ring %8.8x: ", (int)tp->rx_ring);
- for (i = 0; i < RX_RING_SIZE; i++)
- printk(" %8.8x", (unsigned int)tp->rx_ring[i].status);
- printk("\n" KERN_DEBUG " Tx ring %8.8x: ", (int)tp->tx_ring);
- for (i = 0; i < TX_RING_SIZE; i++)
- printk(" %8.8x", (unsigned int)tp->tx_ring[i].status);
- printk("\n");
- }
-#endif
-
- /* Stop and restart the chip's Tx/Rx processes . */
- outl_CSR6(tp->csr6 | EnableRx, ioaddr);
- outl_CSR6(tp->csr6 | EnableTxRx, ioaddr);
- /* Trigger an immediate transmit demand. */
- outl(0, ioaddr + CSR1);
-
- dev->trans_start = jiffies;
- netif_wake_queue (dev);
- tp->stats.tx_errors++;
-}
-
-
-/* Initialize the Rx and Tx rings, along with various 'dev' bits. */
-static void xircom_init_ring(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- int i;
-
- tp->tx_full = 0;
- tp->cur_rx = tp->cur_tx = 0;
- tp->dirty_rx = tp->dirty_tx = 0;
-
- for (i = 0; i < RX_RING_SIZE; i++) {
- tp->rx_ring[i].status = 0;
- tp->rx_ring[i].length = PKT_BUF_SZ;
- tp->rx_ring[i].buffer2 = virt_to_bus(&tp->rx_ring[i+1]);
- tp->rx_skbuff[i] = NULL;
- }
- /* Mark the last entry as wrapping the ring. */
- tp->rx_ring[i-1].length = PKT_BUF_SZ | Rx1RingWrap;
- tp->rx_ring[i-1].buffer2 = virt_to_bus(&tp->rx_ring[0]);
-
- for (i = 0; i < RX_RING_SIZE; i++) {
- /* Note the receive buffer must be longword aligned.
- dev_alloc_skb() provides 16 byte alignment. But do *not*
- use skb_reserve() to align the IP header! */
- struct sk_buff *skb = dev_alloc_skb(PKT_BUF_SZ);
- tp->rx_skbuff[i] = skb;
- if (skb == NULL)
- break;
- skb->dev = dev; /* Mark as being used by this device. */
- tp->rx_ring[i].status = Rx0DescOwned; /* Owned by Xircom chip */
- tp->rx_ring[i].buffer1 = virt_to_bus(skb->data);
- }
- tp->dirty_rx = (unsigned int)(i - RX_RING_SIZE);
-
- /* The Tx buffer descriptor is filled in as needed, but we
- do need to clear the ownership bit. */
- for (i = 0; i < TX_RING_SIZE; i++) {
- tp->tx_skbuff[i] = NULL;
- tp->tx_ring[i].status = 0;
- tp->tx_ring[i].buffer2 = virt_to_bus(&tp->tx_ring[i+1]);
-#ifdef CARDBUS
- if (tp->chip_id == X3201_3)
- tp->tx_aligned_skbuff[i] = dev_alloc_skb(PKT_BUF_SZ);
-#endif /* CARDBUS */
- }
- tp->tx_ring[i-1].buffer2 = virt_to_bus(&tp->tx_ring[0]);
-}
-
-
-static int
-xircom_start_xmit(struct sk_buff *skb, struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- int entry;
- u32 flag;
-
- /* Caution: the write order is important here, set the base address
- with the "ownership" bits last. */
-
- /* Calculate the next Tx descriptor entry. */
- entry = tp->cur_tx % TX_RING_SIZE;
-
- tp->tx_skbuff[entry] = skb;
-#ifdef CARDBUS
- if (tp->chip_id == X3201_3) {
- memcpy(tp->tx_aligned_skbuff[entry]->data,skb->data,skb->len);
- tp->tx_ring[entry].buffer1 = virt_to_bus(tp->tx_aligned_skbuff[entry]->data);
- } else
-#endif
- tp->tx_ring[entry].buffer1 = virt_to_bus(skb->data);
-
- if (tp->cur_tx - tp->dirty_tx < TX_RING_SIZE/2) {/* Typical path */
- flag = Tx1WholePkt; /* No interrupt */
- } else if (tp->cur_tx - tp->dirty_tx == TX_RING_SIZE/2) {
- flag = Tx1WholePkt | Tx1ComplIntr; /* Tx-done intr. */
- } else if (tp->cur_tx - tp->dirty_tx < TX_RING_SIZE - 2) {
- flag = Tx1WholePkt; /* No Tx-done intr. */
- } else {
- /* Leave room for set_rx_mode() to fill entries. */
- flag = Tx1WholePkt | Tx1ComplIntr; /* Tx-done intr. */
- tp->tx_full = 1;
- }
- if (entry == TX_RING_SIZE - 1)
- flag |= Tx1WholePkt | Tx1ComplIntr | Tx1RingWrap;
-
- tp->tx_ring[entry].length = skb->len | flag;
- tp->tx_ring[entry].status = Tx0DescOwned; /* Pass ownership to the chip. */
- tp->cur_tx++;
- if (tp->tx_full)
- netif_stop_queue (dev);
- else
- netif_wake_queue (dev);
-
- /* Trigger an immediate transmit demand. */
- outl(0, dev->base_addr + CSR1);
-
- dev->trans_start = jiffies;
-
- return 0;
-}
-
-
-static void xircom_media_change(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
- u16 reg0, reg1, reg4, reg5;
- u32 csr6 = inl(ioaddr + CSR6), newcsr6;
-
- /* reset status first */
- mdio_read(dev, tp->phys[0], MII_BMCR);
- mdio_read(dev, tp->phys[0], MII_BMSR);
-
- reg0 = mdio_read(dev, tp->phys[0], MII_BMCR);
- reg1 = mdio_read(dev, tp->phys[0], MII_BMSR);
-
- if (reg1 & BMSR_LSTATUS) {
- /* link is up */
- if (reg0 & BMCR_ANENABLE) {
- /* autonegotiation is enabled */
- reg4 = mdio_read(dev, tp->phys[0], MII_ADVERTISE);
- reg5 = mdio_read(dev, tp->phys[0], MII_LPA);
- if (reg4 & ADVERTISE_100FULL && reg5 & LPA_100FULL) {
- tp->speed100 = 1;
- tp->full_duplex = 1;
- } else if (reg4 & ADVERTISE_100HALF && reg5 & LPA_100HALF) {
- tp->speed100 = 1;
- tp->full_duplex = 0;
- } else if (reg4 & ADVERTISE_10FULL && reg5 & LPA_10FULL) {
- tp->speed100 = 0;
- tp->full_duplex = 1;
- } else {
- tp->speed100 = 0;
- tp->full_duplex = 0;
- }
- } else {
- /* autonegotiation is disabled */
- if (reg0 & BMCR_SPEED100)
- tp->speed100 = 1;
- else
- tp->speed100 = 0;
- if (reg0 & BMCR_FULLDPLX)
- tp->full_duplex = 1;
- else
- tp->full_duplex = 0;
- }
- printk(KERN_DEBUG "%s: Link is up, running at %sMbit %s-duplex\n",
- dev->name,
- tp->speed100 ? "100" : "10",
- tp->full_duplex ? "full" : "half");
- netif_carrier_on(dev);
- newcsr6 = csr6 & ~FullDuplexBit;
- if (tp->full_duplex)
- newcsr6 |= FullDuplexBit;
- if (newcsr6 != csr6)
- outl_CSR6(newcsr6, ioaddr + CSR6);
- } else {
- printk(KERN_DEBUG "%s: Link is down\n", dev->name);
- netif_carrier_off(dev);
- }
-}
-
-
-static void check_duplex(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- u16 reg0;
-
- mdio_write(dev, tp->phys[0], MII_BMCR, BMCR_RESET);
- udelay(500);
- while (mdio_read(dev, tp->phys[0], MII_BMCR) & BMCR_RESET);
-
- reg0 = mdio_read(dev, tp->phys[0], MII_BMCR);
- mdio_write(dev, tp->phys[0], MII_ADVERTISE, tp->advertising[0]);
-
- if (tp->autoneg) {
- reg0 &= ~(BMCR_SPEED100 | BMCR_FULLDPLX);
- reg0 |= BMCR_ANENABLE | BMCR_ANRESTART;
- } else {
- reg0 &= ~(BMCR_ANENABLE | BMCR_ANRESTART);
- if (tp->speed100)
- reg0 |= BMCR_SPEED100;
- if (tp->full_duplex)
- reg0 |= BMCR_FULLDPLX;
- printk(KERN_DEBUG "%s: Link forced to %sMbit %s-duplex\n",
- dev->name,
- tp->speed100 ? "100" : "10",
- tp->full_duplex ? "full" : "half");
- }
- mdio_write(dev, tp->phys[0], MII_BMCR, reg0);
-}
-
-
-/* The interrupt handler does all of the Rx thread work and cleans up
- after the Tx thread. */
-static irqreturn_t xircom_interrupt(int irq, void *dev_instance, struct pt_regs *regs)
-{
- struct net_device *dev = dev_instance;
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
- int csr5, work_budget = max_interrupt_work;
- int handled = 0;
-
- spin_lock (&tp->lock);
-
- do {
- csr5 = inl(ioaddr + CSR5);
- /* Acknowledge all of the current interrupt sources ASAP. */
- outl(csr5 & 0x0001ffff, ioaddr + CSR5);
-
- if (xircom_debug > 4)
- printk(KERN_DEBUG "%s: interrupt csr5=%#8.8x new csr5=%#8.8x.\n",
- dev->name, csr5, inl(dev->base_addr + CSR5));
-
- if (csr5 == 0xffffffff)
- break; /* all bits set, assume PCMCIA card removed */
-
- if ((csr5 & (NormalIntr|AbnormalIntr)) == 0)
- break;
-
- handled = 1;
-
- if (csr5 & (RxIntr | RxNoBuf))
- work_budget -= xircom_rx(dev);
-
- if (csr5 & (TxNoBuf | TxDied | TxIntr)) {
- unsigned int dirty_tx;
-
- for (dirty_tx = tp->dirty_tx; tp->cur_tx - dirty_tx > 0;
- dirty_tx++) {
- int entry = dirty_tx % TX_RING_SIZE;
- int status = tp->tx_ring[entry].status;
-
- if (status < 0)
- break; /* It still hasn't been Txed */
- /* Check for Rx filter setup frames. */
- if (tp->tx_skbuff[entry] == NULL)
- continue;
-
- if (status & Tx0DescError) {
- /* There was an major error, log it. */
-#ifndef final_version
- if (xircom_debug > 1)
- printk(KERN_DEBUG "%s: Transmit error, Tx status %8.8x.\n",
- dev->name, status);
-#endif
- tp->stats.tx_errors++;
- if (status & Tx0ManyColl) {
- tp->stats.tx_aborted_errors++;
- }
- if (status & Tx0NoCarrier) tp->stats.tx_carrier_errors++;
- if (status & Tx0LateColl) tp->stats.tx_window_errors++;
- if (status & Tx0Underflow) tp->stats.tx_fifo_errors++;
- } else {
- tp->stats.tx_bytes += tp->tx_ring[entry].length & 0x7ff;
- tp->stats.collisions += (status >> 3) & 15;
- tp->stats.tx_packets++;
- }
-
- /* Free the original skb. */
- dev_kfree_skb_irq(tp->tx_skbuff[entry]);
- tp->tx_skbuff[entry] = NULL;
- }
-
-#ifndef final_version
- if (tp->cur_tx - dirty_tx > TX_RING_SIZE) {
- printk(KERN_ERR "%s: Out-of-sync dirty pointer, %d vs. %d, full=%d.\n",
- dev->name, dirty_tx, tp->cur_tx, tp->tx_full);
- dirty_tx += TX_RING_SIZE;
- }
-#endif
-
- if (tp->tx_full &&
- tp->cur_tx - dirty_tx < TX_RING_SIZE - 2)
- /* The ring is no longer full */
- tp->tx_full = 0;
-
- if (tp->tx_full)
- netif_stop_queue (dev);
- else
- netif_wake_queue (dev);
-
- tp->dirty_tx = dirty_tx;
- if (csr5 & TxDied) {
- if (xircom_debug > 2)
- printk(KERN_WARNING "%s: The transmitter stopped."
- " CSR5 is %x, CSR6 %x, new CSR6 %x.\n",
- dev->name, csr5, inl(ioaddr + CSR6), tp->csr6);
- outl_CSR6(tp->csr6 | EnableRx, ioaddr);
- outl_CSR6(tp->csr6 | EnableTxRx, ioaddr);
- }
- }
-
- /* Log errors. */
- if (csr5 & AbnormalIntr) { /* Abnormal error summary bit. */
- if (csr5 & LinkChange)
- xircom_media_change(dev);
- if (csr5 & TxFIFOUnderflow) {
- if ((tp->csr6 & TxThreshMask) != TxThreshMask)
- tp->csr6 += (1 << TxThreshShift); /* Bump up the Tx threshold */
- else
- tp->csr6 |= TxStoreForw; /* Store-n-forward. */
- /* Restart the transmit process. */
- outl_CSR6(tp->csr6 | EnableRx, ioaddr);
- outl_CSR6(tp->csr6 | EnableTxRx, ioaddr);
- }
- if (csr5 & RxDied) { /* Missed a Rx frame. */
- tp->stats.rx_errors++;
- tp->stats.rx_missed_errors += inl(ioaddr + CSR8) & 0xffff;
- outl_CSR6(tp->csr6 | EnableTxRx, ioaddr);
- }
- /* Clear all error sources, included undocumented ones! */
- outl(0x0800f7ba, ioaddr + CSR5);
- }
- if (--work_budget < 0) {
- if (xircom_debug > 1)
- printk(KERN_WARNING "%s: Too much work during an interrupt, "
- "csr5=0x%8.8x.\n", dev->name, csr5);
- /* Acknowledge all interrupt sources. */
- outl(0x8001ffff, ioaddr + CSR5);
- break;
- }
- } while (1);
-
- if (xircom_debug > 3)
- printk(KERN_DEBUG "%s: exiting interrupt, csr5=%#4.4x.\n",
- dev->name, inl(ioaddr + CSR5));
-
- spin_unlock (&tp->lock);
- return IRQ_RETVAL(handled);
-}
-
-
-static int
-xircom_rx(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- int entry = tp->cur_rx % RX_RING_SIZE;
- int rx_work_limit = tp->dirty_rx + RX_RING_SIZE - tp->cur_rx;
- int work_done = 0;
-
- if (xircom_debug > 4)
- printk(KERN_DEBUG " In xircom_rx(), entry %d %8.8x.\n", entry,
- tp->rx_ring[entry].status);
- /* If we own the next entry, it's a new packet. Send it up. */
- while (tp->rx_ring[entry].status >= 0) {
- s32 status = tp->rx_ring[entry].status;
-
- if (xircom_debug > 5)
- printk(KERN_DEBUG " In xircom_rx(), entry %d %8.8x.\n", entry,
- tp->rx_ring[entry].status);
- if (--rx_work_limit < 0)
- break;
- if ((status & 0x38008300) != 0x0300) {
- if ((status & 0x38000300) != 0x0300) {
- /* Ignore earlier buffers. */
- if ((status & 0xffff) != 0x7fff) {
- if (xircom_debug > 1)
- printk(KERN_WARNING "%s: Oversized Ethernet frame "
- "spanned multiple buffers, status %8.8x!\n",
- dev->name, status);
- tp->stats.rx_length_errors++;
- }
- } else if (status & Rx0DescError) {
- /* There was a fatal error. */
- if (xircom_debug > 2)
- printk(KERN_DEBUG "%s: Receive error, Rx status %8.8x.\n",
- dev->name, status);
- tp->stats.rx_errors++; /* end of a packet.*/
- if (status & (Rx0Runt | Rx0HugeFrame)) tp->stats.rx_length_errors++;
- if (status & Rx0CRCError) tp->stats.rx_crc_errors++;
- }
- } else {
- /* Omit the four octet CRC from the length. */
- short pkt_len = ((status >> 16) & 0x7ff) - 4;
- struct sk_buff *skb;
-
-#ifndef final_version
- if (pkt_len > 1518) {
- printk(KERN_WARNING "%s: Bogus packet size of %d (%#x).\n",
- dev->name, pkt_len, pkt_len);
- pkt_len = 1518;
- tp->stats.rx_length_errors++;
- }
-#endif
- /* Check if the packet is long enough to accept without copying
- to a minimally-sized skbuff. */
- if (pkt_len < rx_copybreak
- && (skb = dev_alloc_skb(pkt_len + 2)) != NULL) {
- skb->dev = dev;
- skb_reserve(skb, 2); /* 16 byte align the IP header */
-#if ! defined(__alpha__)
- eth_copy_and_sum(skb, bus_to_virt(tp->rx_ring[entry].buffer1),
- pkt_len, 0);
- skb_put(skb, pkt_len);
-#else
- memcpy(skb_put(skb, pkt_len),
- bus_to_virt(tp->rx_ring[entry].buffer1), pkt_len);
-#endif
- work_done++;
- } else { /* Pass up the skb already on the Rx ring. */
- skb_put(skb = tp->rx_skbuff[entry], pkt_len);
- tp->rx_skbuff[entry] = NULL;
- }
- skb->protocol = eth_type_trans(skb, dev);
- netif_rx(skb);
- dev->last_rx = jiffies;
- tp->stats.rx_packets++;
- tp->stats.rx_bytes += pkt_len;
- }
- entry = (++tp->cur_rx) % RX_RING_SIZE;
- }
-
- /* Refill the Rx ring buffers. */
- for (; tp->cur_rx - tp->dirty_rx > 0; tp->dirty_rx++) {
- entry = tp->dirty_rx % RX_RING_SIZE;
- if (tp->rx_skbuff[entry] == NULL) {
- struct sk_buff *skb;
- skb = tp->rx_skbuff[entry] = dev_alloc_skb(PKT_BUF_SZ);
- if (skb == NULL)
- break;
- skb->dev = dev; /* Mark as being used by this device. */
- tp->rx_ring[entry].buffer1 = virt_to_bus(skb->data);
- work_done++;
- }
- tp->rx_ring[entry].status = Rx0DescOwned;
- }
-
- return work_done;
-}
-
-
-static void
-xircom_down(struct net_device *dev)
-{
- long ioaddr = dev->base_addr;
- struct xircom_private *tp = netdev_priv(dev);
-
- /* Disable interrupts by clearing the interrupt mask. */
- outl(0, ioaddr + CSR7);
- /* Stop the chip's Tx and Rx processes. */
- outl_CSR6(inl(ioaddr + CSR6) & ~EnableTxRx, ioaddr);
-
- if (inl(ioaddr + CSR6) != 0xffffffff)
- tp->stats.rx_missed_errors += inl(ioaddr + CSR8) & 0xffff;
-
- dev->if_port = tp->saved_if_port;
-}
-
-
-static int
-xircom_close(struct net_device *dev)
-{
- long ioaddr = dev->base_addr;
- struct xircom_private *tp = netdev_priv(dev);
- int i;
-
- if (xircom_debug > 1)
- printk(KERN_DEBUG "%s: Shutting down ethercard, status was %2.2x.\n",
- dev->name, inl(ioaddr + CSR5));
-
- netif_stop_queue(dev);
-
- if (netif_device_present(dev))
- xircom_down(dev);
-
- free_irq(dev->irq, dev);
-
- /* Free all the skbuffs in the Rx queue. */
- for (i = 0; i < RX_RING_SIZE; i++) {
- struct sk_buff *skb = tp->rx_skbuff[i];
- tp->rx_skbuff[i] = NULL;
- tp->rx_ring[i].status = 0; /* Not owned by Xircom chip. */
- tp->rx_ring[i].length = 0;
- tp->rx_ring[i].buffer1 = 0xBADF00D0; /* An invalid address. */
- if (skb) {
- dev_kfree_skb(skb);
- }
- }
- for (i = 0; i < TX_RING_SIZE; i++) {
- if (tp->tx_skbuff[i])
- dev_kfree_skb(tp->tx_skbuff[i]);
- tp->tx_skbuff[i] = NULL;
- }
-
- tp->open = 0;
- return 0;
-}
-
-
-static struct net_device_stats *xircom_get_stats(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- long ioaddr = dev->base_addr;
-
- if (netif_device_present(dev))
- tp->stats.rx_missed_errors += inl(ioaddr + CSR8) & 0xffff;
-
- return &tp->stats;
-}
-
-static int xircom_get_settings(struct net_device *dev, struct ethtool_cmd *ecmd)
-{
- struct xircom_private *tp = netdev_priv(dev);
- ecmd->supported =
- SUPPORTED_10baseT_Half |
- SUPPORTED_10baseT_Full |
- SUPPORTED_100baseT_Half |
- SUPPORTED_100baseT_Full |
- SUPPORTED_Autoneg |
- SUPPORTED_MII;
-
- ecmd->advertising = ADVERTISED_MII;
- if (tp->advertising[0] & ADVERTISE_10HALF)
- ecmd->advertising |= ADVERTISED_10baseT_Half;
- if (tp->advertising[0] & ADVERTISE_10FULL)
- ecmd->advertising |= ADVERTISED_10baseT_Full;
- if (tp->advertising[0] & ADVERTISE_100HALF)
- ecmd->advertising |= ADVERTISED_100baseT_Half;
- if (tp->advertising[0] & ADVERTISE_100FULL)
- ecmd->advertising |= ADVERTISED_100baseT_Full;
- if (tp->autoneg) {
- ecmd->advertising |= ADVERTISED_Autoneg;
- ecmd->autoneg = AUTONEG_ENABLE;
- } else
- ecmd->autoneg = AUTONEG_DISABLE;
-
- ecmd->port = PORT_MII;
- ecmd->transceiver = XCVR_INTERNAL;
- ecmd->phy_address = tp->phys[0];
- ecmd->speed = tp->speed100 ? SPEED_100 : SPEED_10;
- ecmd->duplex = tp->full_duplex ? DUPLEX_FULL : DUPLEX_HALF;
- ecmd->maxtxpkt = TX_RING_SIZE / 2;
- ecmd->maxrxpkt = 0;
- return 0;
-}
-
-static int xircom_set_settings(struct net_device *dev, struct ethtool_cmd *ecmd)
-{
- struct xircom_private *tp = netdev_priv(dev);
- u16 autoneg, speed100, full_duplex;
-
- autoneg = (ecmd->autoneg == AUTONEG_ENABLE);
- speed100 = (ecmd->speed == SPEED_100);
- full_duplex = (ecmd->duplex == DUPLEX_FULL);
-
- tp->autoneg = autoneg;
- if (speed100 != tp->speed100 ||
- full_duplex != tp->full_duplex) {
- tp->speed100 = speed100;
- tp->full_duplex = full_duplex;
- /* change advertising bits */
- tp->advertising[0] &= ~(ADVERTISE_10HALF |
- ADVERTISE_10FULL |
- ADVERTISE_100HALF |
- ADVERTISE_100FULL |
- ADVERTISE_100BASE4);
- if (speed100) {
- if (full_duplex)
- tp->advertising[0] |= ADVERTISE_100FULL;
- else
- tp->advertising[0] |= ADVERTISE_100HALF;
- } else {
- if (full_duplex)
- tp->advertising[0] |= ADVERTISE_10FULL;
- else
- tp->advertising[0] |= ADVERTISE_10HALF;
- }
- }
- check_duplex(dev);
- return 0;
-}
-
-static void xircom_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
-{
- struct xircom_private *tp = netdev_priv(dev);
- strcpy(info->driver, DRV_NAME);
- strcpy(info->version, DRV_VERSION);
- strcpy(info->bus_info, pci_name(tp->pdev));
-}
-
-static struct ethtool_ops ops = {
- .get_settings = xircom_get_settings,
- .set_settings = xircom_set_settings,
- .get_drvinfo = xircom_get_drvinfo,
-};
-
-/* Provide ioctl() calls to examine the MII xcvr state. */
-static int xircom_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
-{
- struct xircom_private *tp = netdev_priv(dev);
- u16 *data = (u16 *)&rq->ifr_ifru;
- int phy = tp->phys[0] & 0x1f;
- unsigned long flags;
-
- switch(cmd) {
- /* Legacy mii-diag interface */
- case SIOCGMIIPHY: /* Get address of MII PHY in use. */
- if (tp->mii_cnt)
- data[0] = phy;
- else
- return -ENODEV;
- return 0;
- case SIOCGMIIREG: /* Read MII PHY register. */
- save_flags(flags);
- cli();
- data[3] = mdio_read(dev, data[0] & 0x1f, data[1] & 0x1f);
- restore_flags(flags);
- return 0;
- case SIOCSMIIREG: /* Write MII PHY register. */
- if (!capable(CAP_NET_ADMIN))
- return -EPERM;
- save_flags(flags);
- cli();
- if (data[0] == tp->phys[0]) {
- u16 value = data[2];
- switch (data[1]) {
- case 0:
- if (value & (BMCR_RESET | BMCR_ANENABLE))
- /* Autonegotiation. */
- tp->autoneg = 1;
- else {
- tp->full_duplex = (value & BMCR_FULLDPLX) ? 1 : 0;
- tp->autoneg = 0;
- }
- break;
- case 4:
- tp->advertising[0] = value;
- break;
- }
- check_duplex(dev);
- }
- mdio_write(dev, data[0] & 0x1f, data[1] & 0x1f, data[2]);
- restore_flags(flags);
- return 0;
- default:
- return -EOPNOTSUPP;
- }
-
- return -EOPNOTSUPP;
-}
-
-/* Set or clear the multicast filter for this adaptor.
- Note that we only use exclusion around actually queueing the
- new frame, not around filling tp->setup_frame. This is non-deterministic
- when re-entered but still correct. */
-static void set_rx_mode(struct net_device *dev)
-{
- struct xircom_private *tp = netdev_priv(dev);
- struct dev_mc_list *mclist;
- long ioaddr = dev->base_addr;
- int csr6 = inl(ioaddr + CSR6);
- u16 *eaddrs, *setup_frm;
- u32 tx_flags;
- int i;
-
- tp->csr6 &= ~(AllMultiBit | PromiscBit | HashFilterBit);
- csr6 &= ~(AllMultiBit | PromiscBit | HashFilterBit);
- if (dev->flags & IFF_PROMISC) { /* Set promiscuous. */
- tp->csr6 |= PromiscBit;
- csr6 |= PromiscBit;
- goto out;
- }
-
- if ((dev->mc_count > 1000) || (dev->flags & IFF_ALLMULTI)) {
- /* Too many to filter well -- accept all multicasts. */
- tp->csr6 |= AllMultiBit;
- csr6 |= AllMultiBit;
- goto out;
- }
-
- tx_flags = Tx1WholePkt | Tx1SetupPkt | PKT_SETUP_SZ;
-
- /* Note that only the low-address shortword of setup_frame is valid! */
- setup_frm = tp->setup_frame;
- mclist = dev->mc_list;
-
- /* Fill the first entry with our physical address. */
- eaddrs = (u16 *)dev->dev_addr;
- *setup_frm = cpu_to_le16(eaddrs[0]); setup_frm += 2;
- *setup_frm = cpu_to_le16(eaddrs[1]); setup_frm += 2;
- *setup_frm = cpu_to_le16(eaddrs[2]); setup_frm += 2;
-
- if (dev->mc_count > 14) { /* Must use a multicast hash table. */
- u32 *hash_table = (u32 *)(tp->setup_frame + 4 * 12);
- u32 hash, hash2;
-
- tx_flags |= Tx1HashSetup;
- tp->csr6 |= HashFilterBit;
- csr6 |= HashFilterBit;
-
- /* Fill the unused 3 entries with the broadcast address.
- At least one entry *must* contain the broadcast address!!!*/
- for (i = 0; i < 3; i++) {
- *setup_frm = 0xffff; setup_frm += 2;
- *setup_frm = 0xffff; setup_frm += 2;
- *setup_frm = 0xffff; setup_frm += 2;
- }
-
- /* Truly brain-damaged hash filter layout */
- /* XXX: not sure if I should take the last or the first 9 bits */
- for (i = 0; i < dev->mc_count; i++, mclist = mclist->next) {
- u32 *hptr;
- hash = ether_crc(ETH_ALEN, mclist->dmi_addr) & 0x1ff;
- if (hash < 384) {
- hash2 = hash + ((hash >> 4) << 4) +
- ((hash >> 5) << 5);
- } else {
- hash -= 384;
- hash2 = 64 + hash + (hash >> 4) * 80;
- }
- hptr = &hash_table[hash2 & ~0x1f];
- *hptr |= cpu_to_le32(1 << (hash2 & 0x1f));
- }
- } else {
- /* We have <= 14 mcast addresses so we can use Xircom's
- wonderful 16-address perfect filter. */
- for (i = 0; i < dev->mc_count; i++, mclist = mclist->next) {
- eaddrs = (u16 *)mclist->dmi_addr;
- *setup_frm = cpu_to_le16(eaddrs[0]); setup_frm += 2;
- *setup_frm = cpu_to_le16(eaddrs[1]); setup_frm += 2;
- *setup_frm = cpu_to_le16(eaddrs[2]); setup_frm += 2;
- }
- /* Fill the unused entries with the broadcast address.
- At least one entry *must* contain the broadcast address!!!*/
- for (; i < 15; i++) {
- *setup_frm = 0xffff; setup_frm += 2;
- *setup_frm = 0xffff; setup_frm += 2;
- *setup_frm = 0xffff; setup_frm += 2;
- }
- }
-
- /* Now add this frame to the Tx list. */
- if (tp->cur_tx - tp->dirty_tx > TX_RING_SIZE - 2) {
- /* Same setup recently queued, we need not add it. */
- /* XXX: Huh? All it means is that the Tx list is full...*/
- } else {
- unsigned long flags;
- unsigned int entry;
- int dummy = -1;
-
- save_flags(flags); cli();
- entry = tp->cur_tx++ % TX_RING_SIZE;
-
- if (entry != 0) {
- /* Avoid a chip errata by prefixing a dummy entry. */
- tp->tx_skbuff[entry] = NULL;
- tp->tx_ring[entry].length =
- (entry == TX_RING_SIZE - 1) ? Tx1RingWrap : 0;
- tp->tx_ring[entry].buffer1 = 0;
- /* race with chip, set Tx0DescOwned later */
- dummy = entry;
- entry = tp->cur_tx++ % TX_RING_SIZE;
- }
-
- tp->tx_skbuff[entry] = NULL;
- /* Put the setup frame on the Tx list. */
- if (entry == TX_RING_SIZE - 1)
- tx_flags |= Tx1RingWrap; /* Wrap ring. */
- tp->tx_ring[entry].length = tx_flags;
- tp->tx_ring[entry].buffer1 = virt_to_bus(tp->setup_frame);
- tp->tx_ring[entry].status = Tx0DescOwned;
- if (tp->cur_tx - tp->dirty_tx >= TX_RING_SIZE - 2) {
- tp->tx_full = 1;
- netif_stop_queue (dev);
- }
- if (dummy >= 0)
- tp->tx_ring[dummy].status = Tx0DescOwned;
- restore_flags(flags);
- /* Trigger an immediate transmit demand. */
- outl(0, ioaddr + CSR1);
- }
-
-out:
- outl_CSR6(csr6, ioaddr);
-}
-
-
-static struct pci_device_id xircom_pci_table[] = {
- { 0x115D, 0x0003, PCI_ANY_ID, PCI_ANY_ID, 0, 0, X3201_3 },
- {0},
-};
-MODULE_DEVICE_TABLE(pci, xircom_pci_table);
-
-
-#ifdef CONFIG_PM
-static int xircom_suspend(struct pci_dev *pdev, pm_message_t state)
-{
- struct net_device *dev = pci_get_drvdata(pdev);
- struct xircom_private *tp = netdev_priv(dev);
- printk(KERN_INFO "xircom_suspend(%s)\n", dev->name);
- if (tp->open)
- xircom_down(dev);
-
- pci_save_state(pdev);
- pci_disable_device(pdev);
- pci_set_power_state(pdev, 3);
-
- return 0;
-}
-
-
-static int xircom_resume(struct pci_dev *pdev)
-{
- struct net_device *dev = pci_get_drvdata(pdev);
- struct xircom_private *tp = netdev_priv(dev);
- printk(KERN_INFO "xircom_resume(%s)\n", dev->name);
-
- pci_set_power_state(pdev,0);
- pci_enable_device(pdev);
- pci_restore_state(pdev);
-
- /* Bring the chip out of sleep mode.
- Caution: Snooze mode does not work with some boards! */
- if (xircom_tbl[tp->chip_id].flags & HAS_ACPI)
- pci_write_config_dword(tp->pdev, PCI_POWERMGMT, 0);
-
- transceiver_voodoo(dev);
- if (xircom_tbl[tp->chip_id].flags & HAS_MII)
- check_duplex(dev);
-
- if (tp->open)
- xircom_up(dev);
- return 0;
-}
-#endif /* CONFIG_PM */
-
-
-static void __devexit xircom_remove_one(struct pci_dev *pdev)
-{
- struct net_device *dev = pci_get_drvdata(pdev);
-
- printk(KERN_INFO "xircom_remove_one(%s)\n", dev->name);
- unregister_netdev(dev);
- pci_release_regions(pdev);
- free_netdev(dev);
- pci_set_drvdata(pdev, NULL);
-}
-
-
-static struct pci_driver xircom_driver = {
- .name = DRV_NAME,
- .id_table = xircom_pci_table,
- .probe = xircom_init_one,
- .remove = __devexit_p(xircom_remove_one),
-#ifdef CONFIG_PM
- .suspend = xircom_suspend,
- .resume = xircom_resume
-#endif /* CONFIG_PM */
-};
-
-
-static int __init xircom_init(void)
-{
-/* when a module, this is printed whether or not devices are found in probe */
-#ifdef MODULE
- printk(version);
-#endif
- return pci_module_init(&xircom_driver);
-}
-
-
-static void __exit xircom_exit(void)
-{
- pci_unregister_driver(&xircom_driver);
-}
-
-module_init(xircom_init)
-module_exit(xircom_exit)
-
-/*
- * Local variables:
- * c-indent-level: 4
- * c-basic-offset: 4
- * tab-width: 4
- * End:
- */
^ permalink raw reply
* [RFC: 2.6 patch] remove drivers/net/eepro100.c
From: Adrian Bunk @ 2005-11-18 3:33 UTC (permalink / raw)
To: saw; +Cc: jgarzik, linux-kernel, netdev
This patch removes the obsolete drivers/net/eepro100.c driver.
Is there any reason why it should be kept?
Signed-off-by: Adrian Bunk <bunk@stusta.de>
---
MAINTAINERS | 5
drivers/net/Kconfig | 14
drivers/net/Makefile | 1
drivers/net/eepro100.c | 2412 -----------------------------------------
4 files changed, 2432 deletions(-)
--- linux-2.6.15-rc1-mm1-full/MAINTAINERS.old 2005-11-18 03:49:44.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/MAINTAINERS 2005-11-18 03:49:50.000000000 +0100
@@ -868,11 +868,6 @@
W: http://ebtables.sourceforge.net/
S: Maintained
-EEPRO100 NETWORK DRIVER
-P: Andrey V. Savochkin
-M: saw@saw.sw.com.sg
-S: Maintained
-
EMU10K1 SOUND DRIVER
P: James Courtier-Dutton
M: James@superbug.demon.co.uk
--- linux-2.6.15-rc1-mm1-full/drivers/net/Kconfig.old 2005-11-18 03:49:59.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/Kconfig 2005-11-18 03:50:13.000000000 +0100
@@ -1405,20 +1405,6 @@
<file:Documentation/networking/net-modules.txt>. The module
will be called dgrs.
-config EEPRO100
- tristate "EtherExpressPro/100 support (eepro100, original Becker driver)"
- depends on NET_PCI && PCI
- select MII
- help
- If you have an Intel EtherExpress PRO/100 PCI network (Ethernet)
- card, say Y and read the Ethernet-HOWTO, available from
- <http://www.tldp.org/docs.html#howto>.
-
- To compile this driver as a module, choose M here and read
- <file:Documentation/networking/net-modules.txt>. The module
- will be called eepro100.
-
-
config E100
tristate "Intel(R) PRO/100+ support"
depends on NET_PCI && PCI
--- linux-2.6.15-rc1-mm1-full/drivers/net/Makefile.old 2005-11-18 03:50:22.000000000 +0100
+++ linux-2.6.15-rc1-mm1-full/drivers/net/Makefile 2005-11-18 03:50:26.000000000 +0100
@@ -40,7 +40,6 @@
obj-$(CONFIG_TYPHOON) += typhoon.o
obj-$(CONFIG_NE2K_PCI) += ne2k-pci.o 8390.o
obj-$(CONFIG_PCNET32) += pcnet32.o
-obj-$(CONFIG_EEPRO100) += eepro100.o
obj-$(CONFIG_E100) += e100.o
obj-$(CONFIG_TLAN) += tlan.o
obj-$(CONFIG_EPIC100) += epic100.o
--- linux-2.6.15-rc1-mm1-full/drivers/net/eepro100.c 2005-10-28 02:02:08.000000000 +0200
+++ /dev/null 2005-11-08 19:07:57.000000000 +0100
@@ -1,2412 +0,0 @@
-/* drivers/net/eepro100.c: An Intel i82557-559 Ethernet driver for Linux. */
-/*
- Written 1996-1999 by Donald Becker.
-
- The driver also contains updates by different kernel developers
- (see incomplete list below).
- Current maintainer is Andrey V. Savochkin <saw@saw.sw.com.sg>.
- Please use this email address and linux-kernel mailing list for bug reports.
-
- This software may be used and distributed according to the terms
- of the GNU General Public License, incorporated herein by reference.
-
- This driver is for the Intel EtherExpress Pro100 (Speedo3) design.
- It should work with all i82557/558/559 boards.
-
- Version history:
- 1998 Apr - 2000 Feb Andrey V. Savochkin <saw@saw.sw.com.sg>
- Serious fixes for multicast filter list setting, TX timeout routine;
- RX ring refilling logic; other stuff
- 2000 Feb Jeff Garzik <jgarzik@pobox.com>
- Convert to new PCI driver interface
- 2000 Mar 24 Dragan Stancevic <visitor@valinux.com>
- Disabled FC and ER, to avoid lockups when when we get FCP interrupts.
- 2000 Jul 17 Goutham Rao <goutham.rao@intel.com>
- PCI DMA API fixes, adding pci_dma_sync_single calls where neccesary
- 2000 Aug 31 David Mosberger <davidm@hpl.hp.com>
- rx_align support: enables rx DMA without causing unaligned accesses.
-*/
-
-static const char *version =
-"eepro100.c:v1.09j-t 9/29/99 Donald Becker http://www.scyld.com/network/eepro100.html\n"
-"eepro100.c: $Revision: 1.36 $ 2000/11/17 Modified by Andrey V. Savochkin <saw@saw.sw.com.sg> and others\n";
-
-/* A few user-configurable values that apply to all boards.
- First set is undocumented and spelled per Intel recommendations. */
-
-static int congenb /* = 0 */; /* Enable congestion control in the DP83840. */
-static int txfifo = 8; /* Tx FIFO threshold in 4 byte units, 0-15 */
-static int rxfifo = 8; /* Rx FIFO threshold, default 32 bytes. */
-/* Tx/Rx DMA burst length, 0-127, 0 == no preemption, tx==128 -> disabled. */
-static int txdmacount = 128;
-static int rxdmacount /* = 0 */;
-
-#if defined(__ia64__) || defined(__alpha__) || defined(__sparc__) || defined(__mips__) || \
- defined(__arm__)
- /* align rx buffers to 2 bytes so that IP header is aligned */
-# define rx_align(skb) skb_reserve((skb), 2)
-# define RxFD_ALIGNMENT __attribute__ ((aligned (2), packed))
-#else
-# define rx_align(skb)
-# define RxFD_ALIGNMENT
-#endif
-
-/* Set the copy breakpoint for the copy-only-tiny-buffer Rx method.
- Lower values use more memory, but are faster. */
-static int rx_copybreak = 200;
-
-/* Maximum events (Rx packets, etc.) to handle at each interrupt. */
-static int max_interrupt_work = 20;
-
-/* Maximum number of multicast addresses to filter (vs. rx-all-multicast) */
-static int multicast_filter_limit = 64;
-
-/* 'options' is used to pass a transceiver override or full-duplex flag
- e.g. "options=16" for FD, "options=32" for 100mbps-only. */
-static int full_duplex[] = {-1, -1, -1, -1, -1, -1, -1, -1};
-static int options[] = {-1, -1, -1, -1, -1, -1, -1, -1};
-
-/* A few values that may be tweaked. */
-/* The ring sizes should be a power of two for efficiency. */
-#define TX_RING_SIZE 64
-#define RX_RING_SIZE 64
-/* How much slots multicast filter setup may take.
- Do not descrease without changing set_rx_mode() implementaion. */
-#define TX_MULTICAST_SIZE 2
-#define TX_MULTICAST_RESERV (TX_MULTICAST_SIZE*2)
-/* Actual number of TX packets queued, must be
- <= TX_RING_SIZE-TX_MULTICAST_RESERV. */
-#define TX_QUEUE_LIMIT (TX_RING_SIZE-TX_MULTICAST_RESERV)
-/* Hysteresis marking queue as no longer full. */
-#define TX_QUEUE_UNFULL (TX_QUEUE_LIMIT-4)
-
-/* Operational parameters that usually are not changed. */
-
-/* Time in jiffies before concluding the transmitter is hung. */
-#define TX_TIMEOUT (2*HZ)
-/* Size of an pre-allocated Rx buffer: <Ethernet MTU> + slack.*/
-#define PKT_BUF_SZ 1536
-
-#include <linux/config.h>
-#include <linux/module.h>
-
-#include <linux/kernel.h>
-#include <linux/string.h>
-#include <linux/errno.h>
-#include <linux/ioport.h>
-#include <linux/slab.h>
-#include <linux/interrupt.h>
-#include <linux/timer.h>
-#include <linux/pci.h>
-#include <linux/spinlock.h>
-#include <linux/init.h>
-#include <linux/mii.h>
-#include <linux/delay.h>
-#include <linux/bitops.h>
-
-#include <asm/io.h>
-#include <asm/uaccess.h>
-#include <asm/irq.h>
-
-#include <linux/netdevice.h>
-#include <linux/etherdevice.h>
-#include <linux/rtnetlink.h>
-#include <linux/skbuff.h>
-#include <linux/ethtool.h>
-
-static int use_io;
-static int debug = -1;
-#define DEBUG_DEFAULT (NETIF_MSG_DRV | \
- NETIF_MSG_HW | \
- NETIF_MSG_RX_ERR | \
- NETIF_MSG_TX_ERR)
-#define DEBUG ((debug >= 0) ? (1<<debug)-1 : DEBUG_DEFAULT)
-
-
-MODULE_AUTHOR("Maintainer: Andrey V. Savochkin <saw@saw.sw.com.sg>");
-MODULE_DESCRIPTION("Intel i82557/i82558/i82559 PCI EtherExpressPro driver");
-MODULE_LICENSE("GPL");
-module_param(use_io, int, 0);
-module_param(debug, int, 0);
-module_param_array(options, int, NULL, 0);
-module_param_array(full_duplex, int, NULL, 0);
-module_param(congenb, int, 0);
-module_param(txfifo, int, 0);
-module_param(rxfifo, int, 0);
-module_param(txdmacount, int, 0);
-module_param(rxdmacount, int, 0);
-module_param(rx_copybreak, int, 0);
-module_param(max_interrupt_work, int, 0);
-module_param(multicast_filter_limit, int, 0);
-MODULE_PARM_DESC(debug, "debug level (0-6)");
-MODULE_PARM_DESC(options, "Bits 0-3: transceiver type, bit 4: full duplex, bit 5: 100Mbps");
-MODULE_PARM_DESC(full_duplex, "full duplex setting(s) (1)");
-MODULE_PARM_DESC(congenb, "Enable congestion control (1)");
-MODULE_PARM_DESC(txfifo, "Tx FIFO threshold in 4 byte units, (0-15)");
-MODULE_PARM_DESC(rxfifo, "Rx FIFO threshold in 4 byte units, (0-15)");
-MODULE_PARM_DESC(txdmacount, "Tx DMA burst length; 128 - disable (0-128)");
-MODULE_PARM_DESC(rxdmacount, "Rx DMA burst length; 128 - disable (0-128)");
-MODULE_PARM_DESC(rx_copybreak, "copy breakpoint for copy-only-tiny-frames");
-MODULE_PARM_DESC(max_interrupt_work, "maximum events handled per interrupt");
-MODULE_PARM_DESC(multicast_filter_limit, "maximum number of filtered multicast addresses");
-
-#define RUN_AT(x) (jiffies + (x))
-
-#define netdevice_start(dev)
-#define netdevice_stop(dev)
-#define netif_set_tx_timeout(dev, tf, tm) \
- do { \
- (dev)->tx_timeout = (tf); \
- (dev)->watchdog_timeo = (tm); \
- } while(0)
-
-
-
-/*
- Theory of Operation
-
-I. Board Compatibility
-
-This device driver is designed for the Intel i82557 "Speedo3" chip, Intel's
-single-chip fast Ethernet controller for PCI, as used on the Intel
-EtherExpress Pro 100 adapter.
-
-II. Board-specific settings
-
-PCI bus devices are configured by the system at boot time, so no jumpers
-need to be set on the board. The system BIOS should be set to assign the
-PCI INTA signal to an otherwise unused system IRQ line. While it's
-possible to share PCI interrupt lines, it negatively impacts performance and
-only recent kernels support it.
-
-III. Driver operation
-
-IIIA. General
-The Speedo3 is very similar to other Intel network chips, that is to say
-"apparently designed on a different planet". This chips retains the complex
-Rx and Tx descriptors and multiple buffers pointers as previous chips, but
-also has simplified Tx and Rx buffer modes. This driver uses the "flexible"
-Tx mode, but in a simplified lower-overhead manner: it associates only a
-single buffer descriptor with each frame descriptor.
-
-Despite the extra space overhead in each receive skbuff, the driver must use
-the simplified Rx buffer mode to assure that only a single data buffer is
-associated with each RxFD. The driver implements this by reserving space
-for the Rx descriptor at the head of each Rx skbuff.
-
-The Speedo-3 has receive and command unit base addresses that are added to
-almost all descriptor pointers. The driver sets these to zero, so that all
-pointer fields are absolute addresses.
-
-The System Control Block (SCB) of some previous Intel chips exists on the
-chip in both PCI I/O and memory space. This driver uses the I/O space
-registers, but might switch to memory mapped mode to better support non-x86
-processors.
-
-IIIB. Transmit structure
-
-The driver must use the complex Tx command+descriptor mode in order to
-have a indirect pointer to the skbuff data section. Each Tx command block
-(TxCB) is associated with two immediately appended Tx Buffer Descriptor
-(TxBD). A fixed ring of these TxCB+TxBD pairs are kept as part of the
-speedo_private data structure for each adapter instance.
-
-The newer i82558 explicitly supports this structure, and can read the two
-TxBDs in the same PCI burst as the TxCB.
-
-This ring structure is used for all normal transmit packets, but the
-transmit packet descriptors aren't long enough for most non-Tx commands such
-as CmdConfigure. This is complicated by the possibility that the chip has
-already loaded the link address in the previous descriptor. So for these
-commands we convert the next free descriptor on the ring to a NoOp, and point
-that descriptor's link to the complex command.
-
-An additional complexity of these non-transmit commands are that they may be
-added asynchronous to the normal transmit queue, so we disable interrupts
-whenever the Tx descriptor ring is manipulated.
-
-A notable aspect of these special configure commands is that they do
-work with the normal Tx ring entry scavenge method. The Tx ring scavenge
-is done at interrupt time using the 'dirty_tx' index, and checking for the
-command-complete bit. While the setup frames may have the NoOp command on the
-Tx ring marked as complete, but not have completed the setup command, this
-is not a problem. The tx_ring entry can be still safely reused, as the
-tx_skbuff[] entry is always empty for config_cmd and mc_setup frames.
-
-Commands may have bits set e.g. CmdSuspend in the command word to either
-suspend or stop the transmit/command unit. This driver always flags the last
-command with CmdSuspend, erases the CmdSuspend in the previous command, and
-then issues a CU_RESUME.
-Note: Watch out for the potential race condition here: imagine
- erasing the previous suspend
- the chip processes the previous command
- the chip processes the final command, and suspends
- doing the CU_RESUME
- the chip processes the next-yet-valid post-final-command.
-So blindly sending a CU_RESUME is only safe if we do it immediately after
-after erasing the previous CmdSuspend, without the possibility of an
-intervening delay. Thus the resume command is always within the
-interrupts-disabled region. This is a timing dependence, but handling this
-condition in a timing-independent way would considerably complicate the code.
-
-Note: In previous generation Intel chips, restarting the command unit was a
-notoriously slow process. This is presumably no longer true.
-
-IIIC. Receive structure
-
-Because of the bus-master support on the Speedo3 this driver uses the new
-SKBUFF_RX_COPYBREAK scheme, rather than a fixed intermediate receive buffer.
-This scheme allocates full-sized skbuffs as receive buffers. The value
-SKBUFF_RX_COPYBREAK is used as the copying breakpoint: it is chosen to
-trade-off the memory wasted by passing the full-sized skbuff to the queue
-layer for all frames vs. the copying cost of copying a frame to a
-correctly-sized skbuff.
-
-For small frames the copying cost is negligible (esp. considering that we
-are pre-loading the cache with immediately useful header information), so we
-allocate a new, minimally-sized skbuff. For large frames the copying cost
-is non-trivial, and the larger copy might flush the cache of useful data, so
-we pass up the skbuff the packet was received into.
-
-IV. Notes
-
-Thanks to Steve Williams of Intel for arranging the non-disclosure agreement
-that stated that I could disclose the information. But I still resent
-having to sign an Intel NDA when I'm helping Intel sell their own product!
-
-*/
-
-static int speedo_found1(struct pci_dev *pdev, void __iomem *ioaddr, int fnd_cnt, int acpi_idle_state);
-
-enum pci_flags_bit {
- PCI_USES_IO=1, PCI_USES_MEM=2, PCI_USES_MASTER=4,
- PCI_ADDR0=0x10<<0, PCI_ADDR1=0x10<<1, PCI_ADDR2=0x10<<2, PCI_ADDR3=0x10<<3,
-};
-
-/* Offsets to the various registers.
- All accesses need not be longword aligned. */
-enum speedo_offsets {
- SCBStatus = 0, SCBCmd = 2, /* Rx/Command Unit command and status. */
- SCBIntmask = 3,
- SCBPointer = 4, /* General purpose pointer. */
- SCBPort = 8, /* Misc. commands and operands. */
- SCBflash = 12, SCBeeprom = 14, /* EEPROM and flash memory control. */
- SCBCtrlMDI = 16, /* MDI interface control. */
- SCBEarlyRx = 20, /* Early receive byte count. */
-};
-/* Commands that can be put in a command list entry. */
-enum commands {
- CmdNOp = 0, CmdIASetup = 0x10000, CmdConfigure = 0x20000,
- CmdMulticastList = 0x30000, CmdTx = 0x40000, CmdTDR = 0x50000,
- CmdDump = 0x60000, CmdDiagnose = 0x70000,
- CmdSuspend = 0x40000000, /* Suspend after completion. */
- CmdIntr = 0x20000000, /* Interrupt after completion. */
- CmdTxFlex = 0x00080000, /* Use "Flexible mode" for CmdTx command. */
-};
-/* Clear CmdSuspend (1<<30) avoiding interference with the card access to the
- status bits. Previous driver versions used separate 16 bit fields for
- commands and statuses. --SAW
- */
-#if defined(__alpha__)
-# define clear_suspend(cmd) clear_bit(30, &(cmd)->cmd_status);
-#else
-# if defined(__LITTLE_ENDIAN)
-# define clear_suspend(cmd) ((__u16 *)&(cmd)->cmd_status)[1] &= ~0x4000
-# elif defined(__BIG_ENDIAN)
-# define clear_suspend(cmd) ((__u16 *)&(cmd)->cmd_status)[1] &= ~0x0040
-# else
-# error Unsupported byteorder
-# endif
-#endif
-
-enum SCBCmdBits {
- SCBMaskCmdDone=0x8000, SCBMaskRxDone=0x4000, SCBMaskCmdIdle=0x2000,
- SCBMaskRxSuspend=0x1000, SCBMaskEarlyRx=0x0800, SCBMaskFlowCtl=0x0400,
- SCBTriggerIntr=0x0200, SCBMaskAll=0x0100,
- /* The rest are Rx and Tx commands. */
- CUStart=0x0010, CUResume=0x0020, CUStatsAddr=0x0040, CUShowStats=0x0050,
- CUCmdBase=0x0060, /* CU Base address (set to zero) . */
- CUDumpStats=0x0070, /* Dump then reset stats counters. */
- RxStart=0x0001, RxResume=0x0002, RxAbort=0x0004, RxAddrLoad=0x0006,
- RxResumeNoResources=0x0007,
-};
-
-enum SCBPort_cmds {
- PortReset=0, PortSelfTest=1, PortPartialReset=2, PortDump=3,
-};
-
-/* The Speedo3 Rx and Tx frame/buffer descriptors. */
-struct descriptor { /* A generic descriptor. */
- volatile s32 cmd_status; /* All command and status fields. */
- u32 link; /* struct descriptor * */
- unsigned char params[0];
-};
-
-/* The Speedo3 Rx and Tx buffer descriptors. */
-struct RxFD { /* Receive frame descriptor. */
- volatile s32 status;
- u32 link; /* struct RxFD * */
- u32 rx_buf_addr; /* void * */
- u32 count;
-} RxFD_ALIGNMENT;
-
-/* Selected elements of the Tx/RxFD.status word. */
-enum RxFD_bits {
- RxComplete=0x8000, RxOK=0x2000,
- RxErrCRC=0x0800, RxErrAlign=0x0400, RxErrTooBig=0x0200, RxErrSymbol=0x0010,
- RxEth2Type=0x0020, RxNoMatch=0x0004, RxNoIAMatch=0x0002,
- TxUnderrun=0x1000, StatusComplete=0x8000,
-};
-
-#define CONFIG_DATA_SIZE 22
-struct TxFD { /* Transmit frame descriptor set. */
- s32 status;
- u32 link; /* void * */
- u32 tx_desc_addr; /* Always points to the tx_buf_addr element. */
- s32 count; /* # of TBD (=1), Tx start thresh., etc. */
- /* This constitutes two "TBD" entries -- we only use one. */
-#define TX_DESCR_BUF_OFFSET 16
- u32 tx_buf_addr0; /* void *, frame to be transmitted. */
- s32 tx_buf_size0; /* Length of Tx frame. */
- u32 tx_buf_addr1; /* void *, frame to be transmitted. */
- s32 tx_buf_size1; /* Length of Tx frame. */
- /* the structure must have space for at least CONFIG_DATA_SIZE starting
- * from tx_desc_addr field */
-};
-
-/* Multicast filter setting block. --SAW */
-struct speedo_mc_block {
- struct speedo_mc_block *next;
- unsigned int tx;
- dma_addr_t frame_dma;
- unsigned int len;
- struct descriptor frame __attribute__ ((__aligned__(16)));
-};
-
-/* Elements of the dump_statistics block. This block must be lword aligned. */
-struct speedo_stats {
- u32 tx_good_frames;
- u32 tx_coll16_errs;
- u32 tx_late_colls;
- u32 tx_underruns;
- u32 tx_lost_carrier;
- u32 tx_deferred;
- u32 tx_one_colls;
- u32 tx_multi_colls;
- u32 tx_total_colls;
- u32 rx_good_frames;
- u32 rx_crc_errs;
- u32 rx_align_errs;
- u32 rx_resource_errs;
- u32 rx_overrun_errs;
- u32 rx_colls_errs;
- u32 rx_runt_errs;
- u32 done_marker;
-};
-
-enum Rx_ring_state_bits {
- RrNoMem=1, RrPostponed=2, RrNoResources=4, RrOOMReported=8,
-};
-
-/* Do not change the position (alignment) of the first few elements!
- The later elements are grouped for cache locality.
-
- Unfortunately, all the positions have been shifted since there.
- A new re-alignment is required. 2000/03/06 SAW */
-struct speedo_private {
- void __iomem *regs;
- struct TxFD *tx_ring; /* Commands (usually CmdTxPacket). */
- struct RxFD *rx_ringp[RX_RING_SIZE]; /* Rx descriptor, used as ring. */
- /* The addresses of a Tx/Rx-in-place packets/buffers. */
- struct sk_buff *tx_skbuff[TX_RING_SIZE];
- struct sk_buff *rx_skbuff[RX_RING_SIZE];
- /* Mapped addresses of the rings. */
- dma_addr_t tx_ring_dma;
-#define TX_RING_ELEM_DMA(sp, n) ((sp)->tx_ring_dma + (n)*sizeof(struct TxFD))
- dma_addr_t rx_ring_dma[RX_RING_SIZE];
- struct descriptor *last_cmd; /* Last command sent. */
- unsigned int cur_tx, dirty_tx; /* The ring entries to be free()ed. */
- spinlock_t lock; /* Group with Tx control cache line. */
- u32 tx_threshold; /* The value for txdesc.count. */
- struct RxFD *last_rxf; /* Last filled RX buffer. */
- dma_addr_t last_rxf_dma;
- unsigned int cur_rx, dirty_rx; /* The next free ring entry */
- long last_rx_time; /* Last Rx, in jiffies, to handle Rx hang. */
- struct net_device_stats stats;
- struct speedo_stats *lstats;
- dma_addr_t lstats_dma;
- int chip_id;
- struct pci_dev *pdev;
- struct timer_list timer; /* Media selection timer. */
- struct speedo_mc_block *mc_setup_head; /* Multicast setup frame list head. */
- struct speedo_mc_block *mc_setup_tail; /* Multicast setup frame list tail. */
- long in_interrupt; /* Word-aligned dev->interrupt */
- unsigned char acpi_pwr;
- signed char rx_mode; /* Current PROMISC/ALLMULTI setting. */
- unsigned int tx_full:1; /* The Tx queue is full. */
- unsigned int flow_ctrl:1; /* Use 802.3x flow control. */
- unsigned int rx_bug:1; /* Work around receiver hang errata. */
- unsigned char default_port:8; /* Last dev->if_port value. */
- unsigned char rx_ring_state; /* RX ring status flags. */
- unsigned short phy[2]; /* PHY media interfaces available. */
- unsigned short partner; /* Link partner caps. */
- struct mii_if_info mii_if; /* MII API hooks, info */
- u32 msg_enable; /* debug message level */
-};
-
-/* The parameters for a CmdConfigure operation.
- There are so many options that it would be difficult to document each bit.
- We mostly use the default or recommended settings. */
-static const char i82557_config_cmd[CONFIG_DATA_SIZE] = {
- 22, 0x08, 0, 0, 0, 0, 0x32, 0x03, 1, /* 1=Use MII 0=Use AUI */
- 0, 0x2E, 0, 0x60, 0,
- 0xf2, 0x48, 0, 0x40, 0xf2, 0x80, /* 0x40=Force full-duplex */
- 0x3f, 0x05, };
-static const char i82558_config_cmd[CONFIG_DATA_SIZE] = {
- 22, 0x08, 0, 1, 0, 0, 0x22, 0x03, 1, /* 1=Use MII 0=Use AUI */
- 0, 0x2E, 0, 0x60, 0x08, 0x88,
- 0x68, 0, 0x40, 0xf2, 0x84, /* Disable FC */
- 0x31, 0x05, };
-
-/* PHY media interface chips. */
-static const char *phys[] = {
- "None", "i82553-A/B", "i82553-C", "i82503",
- "DP83840", "80c240", "80c24", "i82555",
- "unknown-8", "unknown-9", "DP83840A", "unknown-11",
- "unknown-12", "unknown-13", "unknown-14", "unknown-15", };
-enum phy_chips { NonSuchPhy=0, I82553AB, I82553C, I82503, DP83840, S80C240,
- S80C24, I82555, DP83840A=10, };
-static const char is_mii[] = { 0, 1, 1, 0, 1, 1, 0, 1 };
-#define EE_READ_CMD (6)
-
-static int eepro100_init_one(struct pci_dev *pdev,
- const struct pci_device_id *ent);
-
-static int do_eeprom_cmd(void __iomem *ioaddr, int cmd, int cmd_len);
-static int mdio_read(struct net_device *dev, int phy_id, int location);
-static void mdio_write(struct net_device *dev, int phy_id, int location, int value);
-static int speedo_open(struct net_device *dev);
-static void speedo_resume(struct net_device *dev);
-static void speedo_timer(unsigned long data);
-static void speedo_init_rx_ring(struct net_device *dev);
-static void speedo_tx_timeout(struct net_device *dev);
-static int speedo_start_xmit(struct sk_buff *skb, struct net_device *dev);
-static void speedo_refill_rx_buffers(struct net_device *dev, int force);
-static int speedo_rx(struct net_device *dev);
-static void speedo_tx_buffer_gc(struct net_device *dev);
-static irqreturn_t speedo_interrupt(int irq, void *dev_instance, struct pt_regs *regs);
-static int speedo_close(struct net_device *dev);
-static struct net_device_stats *speedo_get_stats(struct net_device *dev);
-static int speedo_ioctl(struct net_device *dev, struct ifreq *rq, int cmd);
-static void set_rx_mode(struct net_device *dev);
-static void speedo_show_state(struct net_device *dev);
-static struct ethtool_ops ethtool_ops;
-
-\f
-
-#ifdef honor_default_port
-/* Optional driver feature to allow forcing the transceiver setting.
- Not recommended. */
-static int mii_ctrl[8] = { 0x3300, 0x3100, 0x0000, 0x0100,
- 0x2000, 0x2100, 0x0400, 0x3100};
-#endif
-
-/* How to wait for the command unit to accept a command.
- Typically this takes 0 ticks. */
-static inline unsigned char wait_for_cmd_done(struct net_device *dev,
- struct speedo_private *sp)
-{
- int wait = 1000;
- void __iomem *cmd_ioaddr = sp->regs + SCBCmd;
- unsigned char r;
-
- do {
- udelay(1);
- r = ioread8(cmd_ioaddr);
- } while(r && --wait >= 0);
-
- if (wait < 0)
- printk(KERN_ALERT "%s: wait_for_cmd_done timeout!\n", dev->name);
- return r;
-}
-
-static int __devinit eepro100_init_one (struct pci_dev *pdev,
- const struct pci_device_id *ent)
-{
- void __iomem *ioaddr;
- int irq, pci_bar;
- int acpi_idle_state = 0, pm;
- static int cards_found /* = 0 */;
- unsigned long pci_base;
-
-#ifndef MODULE
- /* when built-in, we only print version if device is found */
- static int did_version;
- if (did_version++ == 0)
- printk(version);
-#endif
-
- /* save power state before pci_enable_device overwrites it */
- pm = pci_find_capability(pdev, PCI_CAP_ID_PM);
- if (pm) {
- u16 pwr_command;
- pci_read_config_word(pdev, pm + PCI_PM_CTRL, &pwr_command);
- acpi_idle_state = pwr_command & PCI_PM_CTRL_STATE_MASK;
- }
-
- if (pci_enable_device(pdev))
- goto err_out_free_mmio_region;
-
- pci_set_master(pdev);
-
- if (!request_region(pci_resource_start(pdev, 1),
- pci_resource_len(pdev, 1), "eepro100")) {
- printk (KERN_ERR "eepro100: cannot reserve I/O ports\n");
- goto err_out_none;
- }
- if (!request_mem_region(pci_resource_start(pdev, 0),
- pci_resource_len(pdev, 0), "eepro100")) {
- printk (KERN_ERR "eepro100: cannot reserve MMIO region\n");
- goto err_out_free_pio_region;
- }
-
- irq = pdev->irq;
- pci_bar = use_io ? 1 : 0;
- pci_base = pci_resource_start(pdev, pci_bar);
- if (DEBUG & NETIF_MSG_PROBE)
- printk("Found Intel i82557 PCI Speedo at %#lx, IRQ %d.\n",
- pci_base, irq);
-
- ioaddr = pci_iomap(pdev, pci_bar, 0);
- if (!ioaddr) {
- printk (KERN_ERR "eepro100: cannot remap IO\n");
- goto err_out_free_mmio_region;
- }
-
- if (speedo_found1(pdev, ioaddr, cards_found, acpi_idle_state) == 0)
- cards_found++;
- else
- goto err_out_iounmap;
-
- return 0;
-
-err_out_iounmap: ;
- pci_iounmap(pdev, ioaddr);
-err_out_free_mmio_region:
- release_mem_region(pci_resource_start(pdev, 0), pci_resource_len(pdev, 0));
-err_out_free_pio_region:
- release_region(pci_resource_start(pdev, 1), pci_resource_len(pdev, 1));
-err_out_none:
- return -ENODEV;
-}
-
-#ifdef CONFIG_NET_POLL_CONTROLLER
-/*
- * Polling 'interrupt' - used by things like netconsole to send skbs
- * without having to re-enable interrupts. It's not called while
- * the interrupt routine is executing.
- */
-
-static void poll_speedo (struct net_device *dev)
-{
- /* disable_irq is not very nice, but with the funny lockless design
- we have no other choice. */
- disable_irq(dev->irq);
- speedo_interrupt (dev->irq, dev, NULL);
- enable_irq(dev->irq);
-}
-#endif
-
-static int __devinit speedo_found1(struct pci_dev *pdev,
- void __iomem *ioaddr, int card_idx, int acpi_idle_state)
-{
- struct net_device *dev;
- struct speedo_private *sp;
- const char *product;
- int i, option;
- u16 eeprom[0x100];
- int size;
- void *tx_ring_space;
- dma_addr_t tx_ring_dma;
-
- size = TX_RING_SIZE * sizeof(struct TxFD) + sizeof(struct speedo_stats);
- tx_ring_space = pci_alloc_consistent(pdev, size, &tx_ring_dma);
- if (tx_ring_space == NULL)
- return -1;
-
- dev = alloc_etherdev(sizeof(struct speedo_private));
- if (dev == NULL) {
- printk(KERN_ERR "eepro100: Could not allocate ethernet device.\n");
- pci_free_consistent(pdev, size, tx_ring_space, tx_ring_dma);
- return -1;
- }
-
- SET_MODULE_OWNER(dev);
- SET_NETDEV_DEV(dev, &pdev->dev);
-
- if (dev->mem_start > 0)
- option = dev->mem_start;
- else if (card_idx >= 0 && options[card_idx] >= 0)
- option = options[card_idx];
- else
- option = 0;
-
- rtnl_lock();
- if (dev_alloc_name(dev, dev->name) < 0)
- goto err_free_unlock;
-
- /* Read the station address EEPROM before doing the reset.
- Nominally his should even be done before accepting the device, but
- then we wouldn't have a device name with which to report the error.
- The size test is for 6 bit vs. 8 bit address serial EEPROMs.
- */
- {
- void __iomem *iobase;
- int read_cmd, ee_size;
- u16 sum;
- int j;
-
- /* Use IO only to avoid postponed writes and satisfy EEPROM timing
- requirements. */
- iobase = pci_iomap(pdev, 1, pci_resource_len(pdev, 1));
- if (!iobase)
- goto err_free_unlock;
- if ((do_eeprom_cmd(iobase, EE_READ_CMD << 24, 27) & 0xffe0000)
- == 0xffe0000) {
- ee_size = 0x100;
- read_cmd = EE_READ_CMD << 24;
- } else {
- ee_size = 0x40;
- read_cmd = EE_READ_CMD << 22;
- }
-
- for (j = 0, i = 0, sum = 0; i < ee_size; i++) {
- u16 value = do_eeprom_cmd(iobase, read_cmd | (i << 16), 27);
- eeprom[i] = value;
- sum += value;
- if (i < 3) {
- dev->dev_addr[j++] = value;
- dev->dev_addr[j++] = value >> 8;
- }
- }
- if (sum != 0xBABA)
- printk(KERN_WARNING "%s: Invalid EEPROM checksum %#4.4x, "
- "check settings before activating this device!\n",
- dev->name, sum);
- /* Don't unregister_netdev(dev); as the EEPro may actually be
- usable, especially if the MAC address is set later.
- On the other hand, it may be unusable if MDI data is corrupted. */
-
- pci_iounmap(pdev, iobase);
- }
-
- /* Reset the chip: stop Tx and Rx processes and clear counters.
- This takes less than 10usec and will easily finish before the next
- action. */
- iowrite32(PortReset, ioaddr + SCBPort);
- ioread32(ioaddr + SCBPort);
- udelay(10);
-
- if (eeprom[3] & 0x0100)
- product = "OEM i82557/i82558 10/100 Ethernet";
- else
- product = pci_name(pdev);
-
- printk(KERN_INFO "%s: %s, ", dev->name, product);
-
- for (i = 0; i < 5; i++)
- printk("%2.2X:", dev->dev_addr[i]);
- printk("%2.2X, ", dev->dev_addr[i]);
- printk("IRQ %d.\n", pdev->irq);
-
- sp = netdev_priv(dev);
-
- /* we must initialize this early, for mdio_{read,write} */
- sp->regs = ioaddr;
-
-#if 1 || defined(kernel_bloat)
- /* OK, this is pure kernel bloat. I don't like it when other drivers
- waste non-pageable kernel space to emit similar messages, but I need
- them for bug reports. */
- {
- const char *connectors[] = {" RJ45", " BNC", " AUI", " MII"};
- /* The self-test results must be paragraph aligned. */
- volatile s32 *self_test_results;
- int boguscnt = 16000; /* Timeout for set-test. */
- if ((eeprom[3] & 0x03) != 0x03)
- printk(KERN_INFO " Receiver lock-up bug exists -- enabling"
- " work-around.\n");
- printk(KERN_INFO " Board assembly %4.4x%2.2x-%3.3d, Physical"
- " connectors present:",
- eeprom[8], eeprom[9]>>8, eeprom[9] & 0xff);
- for (i = 0; i < 4; i++)
- if (eeprom[5] & (1<<i))
- printk(connectors[i]);
- printk("\n"KERN_INFO" Primary interface chip %s PHY #%d.\n",
- phys[(eeprom[6]>>8)&15], eeprom[6] & 0x1f);
- if (eeprom[7] & 0x0700)
- printk(KERN_INFO " Secondary interface chip %s.\n",
- phys[(eeprom[7]>>8)&7]);
- if (((eeprom[6]>>8) & 0x3f) == DP83840
- || ((eeprom[6]>>8) & 0x3f) == DP83840A) {
- int mdi_reg23 = mdio_read(dev, eeprom[6] & 0x1f, 23) | 0x0422;
- if (congenb)
- mdi_reg23 |= 0x0100;
- printk(KERN_INFO" DP83840 specific setup, setting register 23 to %4.4x.\n",
- mdi_reg23);
- mdio_write(dev, eeprom[6] & 0x1f, 23, mdi_reg23);
- }
- if ((option >= 0) && (option & 0x70)) {
- printk(KERN_INFO " Forcing %dMbs %s-duplex operation.\n",
- (option & 0x20 ? 100 : 10),
- (option & 0x10 ? "full" : "half"));
- mdio_write(dev, eeprom[6] & 0x1f, MII_BMCR,
- ((option & 0x20) ? 0x2000 : 0) | /* 100mbps? */
- ((option & 0x10) ? 0x0100 : 0)); /* Full duplex? */
- }
-
- /* Perform a system self-test. */
- self_test_results = (s32*) ((((long) tx_ring_space) + 15) & ~0xf);
- self_test_results[0] = 0;
- self_test_results[1] = -1;
- iowrite32(tx_ring_dma | PortSelfTest, ioaddr + SCBPort);
- do {
- udelay(10);
- } while (self_test_results[1] == -1 && --boguscnt >= 0);
-
- if (boguscnt < 0) { /* Test optimized out. */
- printk(KERN_ERR "Self test failed, status %8.8x:\n"
- KERN_ERR " Failure to initialize the i82557.\n"
- KERN_ERR " Verify that the card is a bus-master"
- " capable slot.\n",
- self_test_results[1]);
- } else
- printk(KERN_INFO " General self-test: %s.\n"
- KERN_INFO " Serial sub-system self-test: %s.\n"
- KERN_INFO " Internal registers self-test: %s.\n"
- KERN_INFO " ROM checksum self-test: %s (%#8.8x).\n",
- self_test_results[1] & 0x1000 ? "failed" : "passed",
- self_test_results[1] & 0x0020 ? "failed" : "passed",
- self_test_results[1] & 0x0008 ? "failed" : "passed",
- self_test_results[1] & 0x0004 ? "failed" : "passed",
- self_test_results[0]);
- }
-#endif /* kernel_bloat */
-
- iowrite32(PortReset, ioaddr + SCBPort);
- ioread32(ioaddr + SCBPort);
- udelay(10);
-
- /* Return the chip to its original power state. */
- pci_set_power_state(pdev, acpi_idle_state);
-
- pci_set_drvdata (pdev, dev);
- SET_NETDEV_DEV(dev, &pdev->dev);
-
- dev->irq = pdev->irq;
-
- sp->pdev = pdev;
- sp->msg_enable = DEBUG;
- sp->acpi_pwr = acpi_idle_state;
- sp->tx_ring = tx_ring_space;
- sp->tx_ring_dma = tx_ring_dma;
- sp->lstats = (struct speedo_stats *)(sp->tx_ring + TX_RING_SIZE);
- sp->lstats_dma = TX_RING_ELEM_DMA(sp, TX_RING_SIZE);
- init_timer(&sp->timer); /* used in ioctl() */
- spin_lock_init(&sp->lock);
-
- sp->mii_if.full_duplex = option >= 0 && (option & 0x10) ? 1 : 0;
- if (card_idx >= 0) {
- if (full_duplex[card_idx] >= 0)
- sp->mii_if.full_duplex = full_duplex[card_idx];
- }
- sp->default_port = option >= 0 ? (option & 0x0f) : 0;
-
- sp->phy[0] = eeprom[6];
- sp->phy[1] = eeprom[7];
-
- sp->mii_if.phy_id = eeprom[6] & 0x1f;
- sp->mii_if.phy_id_mask = 0x1f;
- sp->mii_if.reg_num_mask = 0x1f;
- sp->mii_if.dev = dev;
- sp->mii_if.mdio_read = mdio_read;
- sp->mii_if.mdio_write = mdio_write;
-
- sp->rx_bug = (eeprom[3] & 0x03) == 3 ? 0 : 1;
- if (((pdev->device > 0x1030 && (pdev->device < 0x103F)))
- || (pdev->device == 0x2449) || (pdev->device == 0x2459)
- || (pdev->device == 0x245D)) {
- sp->chip_id = 1;
- }
-
- if (sp->rx_bug)
- printk(KERN_INFO " Receiver lock-up workaround activated.\n");
-
- /* The Speedo-specific entries in the device structure. */
- dev->open = &speedo_open;
- dev->hard_start_xmit = &speedo_start_xmit;
- netif_set_tx_timeout(dev, &speedo_tx_timeout, TX_TIMEOUT);
- dev->stop = &speedo_close;
- dev->get_stats = &speedo_get_stats;
- dev->set_multicast_list = &set_rx_mode;
- dev->do_ioctl = &speedo_ioctl;
- SET_ETHTOOL_OPS(dev, ðtool_ops);
-#ifdef CONFIG_NET_POLL_CONTROLLER
- dev->poll_controller = &poll_speedo;
-#endif
-
- if (register_netdevice(dev))
- goto err_free_unlock;
- rtnl_unlock();
-
- return 0;
-
- err_free_unlock:
- rtnl_unlock();
- free_netdev(dev);
- return -1;
-}
-
-static void do_slow_command(struct net_device *dev, struct speedo_private *sp, int cmd)
-{
- void __iomem *cmd_ioaddr = sp->regs + SCBCmd;
- int wait = 0;
- do
- if (ioread8(cmd_ioaddr) == 0) break;
- while(++wait <= 200);
- if (wait > 100)
- printk(KERN_ERR "Command %4.4x never accepted (%d polls)!\n",
- ioread8(cmd_ioaddr), wait);
-
- iowrite8(cmd, cmd_ioaddr);
-
- for (wait = 0; wait <= 100; wait++)
- if (ioread8(cmd_ioaddr) == 0) return;
- for (; wait <= 20000; wait++)
- if (ioread8(cmd_ioaddr) == 0) return;
- else udelay(1);
- printk(KERN_ERR "Command %4.4x was not accepted after %d polls!"
- " Current status %8.8x.\n",
- cmd, wait, ioread32(sp->regs + SCBStatus));
-}
-
-/* Serial EEPROM section.
- A "bit" grungy, but we work our way through bit-by-bit :->. */
-/* EEPROM_Ctrl bits. */
-#define EE_SHIFT_CLK 0x01 /* EEPROM shift clock. */
-#define EE_CS 0x02 /* EEPROM chip select. */
-#define EE_DATA_WRITE 0x04 /* EEPROM chip data in. */
-#define EE_DATA_READ 0x08 /* EEPROM chip data out. */
-#define EE_ENB (0x4800 | EE_CS)
-#define EE_WRITE_0 0x4802
-#define EE_WRITE_1 0x4806
-#define EE_OFFSET SCBeeprom
-
-/* The fixes for the code were kindly provided by Dragan Stancevic
- <visitor@valinux.com> to strictly follow Intel specifications of EEPROM
- access timing.
- The publicly available sheet 64486302 (sec. 3.1) specifies 1us access
- interval for serial EEPROM. However, it looks like that there is an
- additional requirement dictating larger udelay's in the code below.
- 2000/05/24 SAW */
-static int __devinit do_eeprom_cmd(void __iomem *ioaddr, int cmd, int cmd_len)
-{
- unsigned retval = 0;
- void __iomem *ee_addr = ioaddr + SCBeeprom;
-
- iowrite16(EE_ENB, ee_addr); udelay(2);
- iowrite16(EE_ENB | EE_SHIFT_CLK, ee_addr); udelay(2);
-
- /* Shift the command bits out. */
- do {
- short dataval = (cmd & (1 << cmd_len)) ? EE_WRITE_1 : EE_WRITE_0;
- iowrite16(dataval, ee_addr); udelay(2);
- iowrite16(dataval | EE_SHIFT_CLK, ee_addr); udelay(2);
- retval = (retval << 1) | ((ioread16(ee_addr) & EE_DATA_READ) ? 1 : 0);
- } while (--cmd_len >= 0);
- iowrite16(EE_ENB, ee_addr); udelay(2);
-
- /* Terminate the EEPROM access. */
- iowrite16(EE_ENB & ~EE_CS, ee_addr);
- return retval;
-}
-
-static int mdio_read(struct net_device *dev, int phy_id, int location)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int val, boguscnt = 64*10; /* <64 usec. to complete, typ 27 ticks */
- iowrite32(0x08000000 | (location<<16) | (phy_id<<21), ioaddr + SCBCtrlMDI);
- do {
- val = ioread32(ioaddr + SCBCtrlMDI);
- if (--boguscnt < 0) {
- printk(KERN_ERR " mdio_read() timed out with val = %8.8x.\n", val);
- break;
- }
- } while (! (val & 0x10000000));
- return val & 0xffff;
-}
-
-static void mdio_write(struct net_device *dev, int phy_id, int location, int value)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int val, boguscnt = 64*10; /* <64 usec. to complete, typ 27 ticks */
- iowrite32(0x04000000 | (location<<16) | (phy_id<<21) | value,
- ioaddr + SCBCtrlMDI);
- do {
- val = ioread32(ioaddr + SCBCtrlMDI);
- if (--boguscnt < 0) {
- printk(KERN_ERR" mdio_write() timed out with val = %8.8x.\n", val);
- break;
- }
- } while (! (val & 0x10000000));
-}
-
-static int
-speedo_open(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int retval;
-
- if (netif_msg_ifup(sp))
- printk(KERN_DEBUG "%s: speedo_open() irq %d.\n", dev->name, dev->irq);
-
- pci_set_power_state(sp->pdev, PCI_D0);
-
- /* Set up the Tx queue early.. */
- sp->cur_tx = 0;
- sp->dirty_tx = 0;
- sp->last_cmd = NULL;
- sp->tx_full = 0;
- sp->in_interrupt = 0;
-
- /* .. we can safely take handler calls during init. */
- retval = request_irq(dev->irq, &speedo_interrupt, SA_SHIRQ, dev->name, dev);
- if (retval) {
- return retval;
- }
-
- dev->if_port = sp->default_port;
-
-#ifdef oh_no_you_dont_unless_you_honour_the_options_passed_in_to_us
- /* Retrigger negotiation to reset previous errors. */
- if ((sp->phy[0] & 0x8000) == 0) {
- int phy_addr = sp->phy[0] & 0x1f ;
- /* Use 0x3300 for restarting NWay, other values to force xcvr:
- 0x0000 10-HD
- 0x0100 10-FD
- 0x2000 100-HD
- 0x2100 100-FD
- */
-#ifdef honor_default_port
- mdio_write(dev, phy_addr, MII_BMCR, mii_ctrl[dev->default_port & 7]);
-#else
- mdio_write(dev, phy_addr, MII_BMCR, 0x3300);
-#endif
- }
-#endif
-
- speedo_init_rx_ring(dev);
-
- /* Fire up the hardware. */
- iowrite16(SCBMaskAll, ioaddr + SCBCmd);
- speedo_resume(dev);
-
- netdevice_start(dev);
- netif_start_queue(dev);
-
- /* Setup the chip and configure the multicast list. */
- sp->mc_setup_head = NULL;
- sp->mc_setup_tail = NULL;
- sp->flow_ctrl = sp->partner = 0;
- sp->rx_mode = -1; /* Invalid -> always reset the mode. */
- set_rx_mode(dev);
- if ((sp->phy[0] & 0x8000) == 0)
- sp->mii_if.advertising = mdio_read(dev, sp->phy[0] & 0x1f, MII_ADVERTISE);
-
- mii_check_link(&sp->mii_if);
-
- if (netif_msg_ifup(sp)) {
- printk(KERN_DEBUG "%s: Done speedo_open(), status %8.8x.\n",
- dev->name, ioread16(ioaddr + SCBStatus));
- }
-
- /* Set the timer. The timer serves a dual purpose:
- 1) to monitor the media interface (e.g. link beat) and perhaps switch
- to an alternate media type
- 2) to monitor Rx activity, and restart the Rx process if the receiver
- hangs. */
- sp->timer.expires = RUN_AT((24*HZ)/10); /* 2.4 sec. */
- sp->timer.data = (unsigned long)dev;
- sp->timer.function = &speedo_timer; /* timer handler */
- add_timer(&sp->timer);
-
- /* No need to wait for the command unit to accept here. */
- if ((sp->phy[0] & 0x8000) == 0)
- mdio_read(dev, sp->phy[0] & 0x1f, MII_BMCR);
-
- return 0;
-}
-
-/* Start the chip hardware after a full reset. */
-static void speedo_resume(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
-
- /* Start with a Tx threshold of 256 (0x..20.... 8 byte units). */
- sp->tx_threshold = 0x01208000;
-
- /* Set the segment registers to '0'. */
- if (wait_for_cmd_done(dev, sp) != 0) {
- iowrite32(PortPartialReset, ioaddr + SCBPort);
- udelay(10);
- }
-
- iowrite32(0, ioaddr + SCBPointer);
- ioread32(ioaddr + SCBPointer); /* Flush to PCI. */
- udelay(10); /* Bogus, but it avoids the bug. */
-
- /* Note: these next two operations can take a while. */
- do_slow_command(dev, sp, RxAddrLoad);
- do_slow_command(dev, sp, CUCmdBase);
-
- /* Load the statistics block and rx ring addresses. */
- iowrite32(sp->lstats_dma, ioaddr + SCBPointer);
- ioread32(ioaddr + SCBPointer); /* Flush to PCI */
-
- iowrite8(CUStatsAddr, ioaddr + SCBCmd);
- sp->lstats->done_marker = 0;
- wait_for_cmd_done(dev, sp);
-
- if (sp->rx_ringp[sp->cur_rx % RX_RING_SIZE] == NULL) {
- if (netif_msg_rx_err(sp))
- printk(KERN_DEBUG "%s: NULL cur_rx in speedo_resume().\n",
- dev->name);
- } else {
- iowrite32(sp->rx_ring_dma[sp->cur_rx % RX_RING_SIZE],
- ioaddr + SCBPointer);
- ioread32(ioaddr + SCBPointer); /* Flush to PCI */
- }
-
- /* Note: RxStart should complete instantly. */
- do_slow_command(dev, sp, RxStart);
- do_slow_command(dev, sp, CUDumpStats);
-
- /* Fill the first command with our physical address. */
- {
- struct descriptor *ias_cmd;
-
- ias_cmd =
- (struct descriptor *)&sp->tx_ring[sp->cur_tx++ % TX_RING_SIZE];
- /* Avoid a bug(?!) here by marking the command already completed. */
- ias_cmd->cmd_status = cpu_to_le32((CmdSuspend | CmdIASetup) | 0xa000);
- ias_cmd->link =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, sp->cur_tx % TX_RING_SIZE));
- memcpy(ias_cmd->params, dev->dev_addr, 6);
- if (sp->last_cmd)
- clear_suspend(sp->last_cmd);
- sp->last_cmd = ias_cmd;
- }
-
- /* Start the chip's Tx process and unmask interrupts. */
- iowrite32(TX_RING_ELEM_DMA(sp, sp->dirty_tx % TX_RING_SIZE),
- ioaddr + SCBPointer);
- /* We are not ACK-ing FCP and ER in the interrupt handler yet so they should
- remain masked --Dragan */
- iowrite16(CUStart | SCBMaskEarlyRx | SCBMaskFlowCtl, ioaddr + SCBCmd);
-}
-
-/*
- * Sometimes the receiver stops making progress. This routine knows how to
- * get it going again, without losing packets or being otherwise nasty like
- * a chip reset would be. Previously the driver had a whole sequence
- * of if RxSuspended, if it's no buffers do one thing, if it's no resources,
- * do another, etc. But those things don't really matter. Separate logic
- * in the ISR provides for allocating buffers--the other half of operation
- * is just making sure the receiver is active. speedo_rx_soft_reset does that.
- * This problem with the old, more involved algorithm is shown up under
- * ping floods on the order of 60K packets/second on a 100Mbps fdx network.
- */
-static void
-speedo_rx_soft_reset(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- struct RxFD *rfd;
- void __iomem *ioaddr;
-
- ioaddr = sp->regs;
- if (wait_for_cmd_done(dev, sp) != 0) {
- printk("%s: previous command stalled\n", dev->name);
- return;
- }
- /*
- * Put the hardware into a known state.
- */
- iowrite8(RxAbort, ioaddr + SCBCmd);
-
- rfd = sp->rx_ringp[sp->cur_rx % RX_RING_SIZE];
-
- rfd->rx_buf_addr = 0xffffffff;
-
- if (wait_for_cmd_done(dev, sp) != 0) {
- printk("%s: RxAbort command stalled\n", dev->name);
- return;
- }
- iowrite32(sp->rx_ring_dma[sp->cur_rx % RX_RING_SIZE],
- ioaddr + SCBPointer);
- iowrite8(RxStart, ioaddr + SCBCmd);
-}
-
-
-/* Media monitoring and control. */
-static void speedo_timer(unsigned long data)
-{
- struct net_device *dev = (struct net_device *)data;
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int phy_num = sp->phy[0] & 0x1f;
-
- /* We have MII and lost link beat. */
- if ((sp->phy[0] & 0x8000) == 0) {
- int partner = mdio_read(dev, phy_num, MII_LPA);
- if (partner != sp->partner) {
- int flow_ctrl = sp->mii_if.advertising & partner & 0x0400 ? 1 : 0;
- if (netif_msg_link(sp)) {
- printk(KERN_DEBUG "%s: Link status change.\n", dev->name);
- printk(KERN_DEBUG "%s: Old partner %x, new %x, adv %x.\n",
- dev->name, sp->partner, partner, sp->mii_if.advertising);
- }
- sp->partner = partner;
- if (flow_ctrl != sp->flow_ctrl) {
- sp->flow_ctrl = flow_ctrl;
- sp->rx_mode = -1; /* Trigger a reload. */
- }
- }
- }
- mii_check_link(&sp->mii_if);
- if (netif_msg_timer(sp)) {
- printk(KERN_DEBUG "%s: Media control tick, status %4.4x.\n",
- dev->name, ioread16(ioaddr + SCBStatus));
- }
- if (sp->rx_mode < 0 ||
- (sp->rx_bug && jiffies - sp->last_rx_time > 2*HZ)) {
- /* We haven't received a packet in a Long Time. We might have been
- bitten by the receiver hang bug. This can be cleared by sending
- a set multicast list command. */
- if (netif_msg_timer(sp))
- printk(KERN_DEBUG "%s: Sending a multicast list set command"
- " from a timer routine,"
- " m=%d, j=%ld, l=%ld.\n",
- dev->name, sp->rx_mode, jiffies, sp->last_rx_time);
- set_rx_mode(dev);
- }
- /* We must continue to monitor the media. */
- sp->timer.expires = RUN_AT(2*HZ); /* 2.0 sec. */
- add_timer(&sp->timer);
-}
-
-static void speedo_show_state(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- int i;
-
- if (netif_msg_pktdata(sp)) {
- printk(KERN_DEBUG "%s: Tx ring dump, Tx queue %u / %u:\n",
- dev->name, sp->cur_tx, sp->dirty_tx);
- for (i = 0; i < TX_RING_SIZE; i++)
- printk(KERN_DEBUG "%s: %c%c%2d %8.8x.\n", dev->name,
- i == sp->dirty_tx % TX_RING_SIZE ? '*' : ' ',
- i == sp->cur_tx % TX_RING_SIZE ? '=' : ' ',
- i, sp->tx_ring[i].status);
-
- printk(KERN_DEBUG "%s: Printing Rx ring"
- " (next to receive into %u, dirty index %u).\n",
- dev->name, sp->cur_rx, sp->dirty_rx);
- for (i = 0; i < RX_RING_SIZE; i++)
- printk(KERN_DEBUG "%s: %c%c%c%2d %8.8x.\n", dev->name,
- sp->rx_ringp[i] == sp->last_rxf ? 'l' : ' ',
- i == sp->dirty_rx % RX_RING_SIZE ? '*' : ' ',
- i == sp->cur_rx % RX_RING_SIZE ? '=' : ' ',
- i, (sp->rx_ringp[i] != NULL) ?
- (unsigned)sp->rx_ringp[i]->status : 0);
- }
-
-#if 0
- {
- void __iomem *ioaddr = sp->regs;
- int phy_num = sp->phy[0] & 0x1f;
- for (i = 0; i < 16; i++) {
- /* FIXME: what does it mean? --SAW */
- if (i == 6) i = 21;
- printk(KERN_DEBUG "%s: PHY index %d register %d is %4.4x.\n",
- dev->name, phy_num, i, mdio_read(dev, phy_num, i));
- }
- }
-#endif
-
-}
-
-/* Initialize the Rx and Tx rings, along with various 'dev' bits. */
-static void
-speedo_init_rx_ring(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- struct RxFD *rxf, *last_rxf = NULL;
- dma_addr_t last_rxf_dma = 0 /* to shut up the compiler */;
- int i;
-
- sp->cur_rx = 0;
-
- for (i = 0; i < RX_RING_SIZE; i++) {
- struct sk_buff *skb;
- skb = dev_alloc_skb(PKT_BUF_SZ + sizeof(struct RxFD));
- if (skb)
- rx_align(skb); /* Align IP on 16 byte boundary */
- sp->rx_skbuff[i] = skb;
- if (skb == NULL)
- break; /* OK. Just initially short of Rx bufs. */
- skb->dev = dev; /* Mark as being used by this device. */
- rxf = (struct RxFD *)skb->data;
- sp->rx_ringp[i] = rxf;
- sp->rx_ring_dma[i] =
- pci_map_single(sp->pdev, rxf,
- PKT_BUF_SZ + sizeof(struct RxFD), PCI_DMA_BIDIRECTIONAL);
- skb_reserve(skb, sizeof(struct RxFD));
- if (last_rxf) {
- last_rxf->link = cpu_to_le32(sp->rx_ring_dma[i]);
- pci_dma_sync_single_for_device(sp->pdev, last_rxf_dma,
- sizeof(struct RxFD), PCI_DMA_TODEVICE);
- }
- last_rxf = rxf;
- last_rxf_dma = sp->rx_ring_dma[i];
- rxf->status = cpu_to_le32(0x00000001); /* '1' is flag value only. */
- rxf->link = 0; /* None yet. */
- /* This field unused by i82557. */
- rxf->rx_buf_addr = 0xffffffff;
- rxf->count = cpu_to_le32(PKT_BUF_SZ << 16);
- pci_dma_sync_single_for_device(sp->pdev, sp->rx_ring_dma[i],
- sizeof(struct RxFD), PCI_DMA_TODEVICE);
- }
- sp->dirty_rx = (unsigned int)(i - RX_RING_SIZE);
- /* Mark the last entry as end-of-list. */
- last_rxf->status = cpu_to_le32(0xC0000002); /* '2' is flag value only. */
- pci_dma_sync_single_for_device(sp->pdev, sp->rx_ring_dma[RX_RING_SIZE-1],
- sizeof(struct RxFD), PCI_DMA_TODEVICE);
- sp->last_rxf = last_rxf;
- sp->last_rxf_dma = last_rxf_dma;
-}
-
-static void speedo_purge_tx(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- int entry;
-
- while ((int)(sp->cur_tx - sp->dirty_tx) > 0) {
- entry = sp->dirty_tx % TX_RING_SIZE;
- if (sp->tx_skbuff[entry]) {
- sp->stats.tx_errors++;
- pci_unmap_single(sp->pdev,
- le32_to_cpu(sp->tx_ring[entry].tx_buf_addr0),
- sp->tx_skbuff[entry]->len, PCI_DMA_TODEVICE);
- dev_kfree_skb_irq(sp->tx_skbuff[entry]);
- sp->tx_skbuff[entry] = NULL;
- }
- sp->dirty_tx++;
- }
- while (sp->mc_setup_head != NULL) {
- struct speedo_mc_block *t;
- if (netif_msg_tx_err(sp))
- printk(KERN_DEBUG "%s: freeing mc frame.\n", dev->name);
- pci_unmap_single(sp->pdev, sp->mc_setup_head->frame_dma,
- sp->mc_setup_head->len, PCI_DMA_TODEVICE);
- t = sp->mc_setup_head->next;
- kfree(sp->mc_setup_head);
- sp->mc_setup_head = t;
- }
- sp->mc_setup_tail = NULL;
- sp->tx_full = 0;
- netif_wake_queue(dev);
-}
-
-static void reset_mii(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
-
- /* Reset the MII transceiver, suggested by Fred Young @ scalable.com. */
- if ((sp->phy[0] & 0x8000) == 0) {
- int phy_addr = sp->phy[0] & 0x1f;
- int advertising = mdio_read(dev, phy_addr, MII_ADVERTISE);
- int mii_bmcr = mdio_read(dev, phy_addr, MII_BMCR);
- mdio_write(dev, phy_addr, MII_BMCR, 0x0400);
- mdio_write(dev, phy_addr, MII_BMSR, 0x0000);
- mdio_write(dev, phy_addr, MII_ADVERTISE, 0x0000);
- mdio_write(dev, phy_addr, MII_BMCR, 0x8000);
-#ifdef honor_default_port
- mdio_write(dev, phy_addr, MII_BMCR, mii_ctrl[dev->default_port & 7]);
-#else
- mdio_read(dev, phy_addr, MII_BMCR);
- mdio_write(dev, phy_addr, MII_BMCR, mii_bmcr);
- mdio_write(dev, phy_addr, MII_ADVERTISE, advertising);
-#endif
- }
-}
-
-static void speedo_tx_timeout(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int status = ioread16(ioaddr + SCBStatus);
- unsigned long flags;
-
- if (netif_msg_tx_err(sp)) {
- printk(KERN_WARNING "%s: Transmit timed out: status %4.4x "
- " %4.4x at %d/%d command %8.8x.\n",
- dev->name, status, ioread16(ioaddr + SCBCmd),
- sp->dirty_tx, sp->cur_tx,
- sp->tx_ring[sp->dirty_tx % TX_RING_SIZE].status);
-
- }
- speedo_show_state(dev);
-#if 0
- if ((status & 0x00C0) != 0x0080
- && (status & 0x003C) == 0x0010) {
- /* Only the command unit has stopped. */
- printk(KERN_WARNING "%s: Trying to restart the transmitter...\n",
- dev->name);
- iowrite32(TX_RING_ELEM_DMA(sp, dirty_tx % TX_RING_SIZE]),
- ioaddr + SCBPointer);
- iowrite16(CUStart, ioaddr + SCBCmd);
- reset_mii(dev);
- } else {
-#else
- {
-#endif
- del_timer_sync(&sp->timer);
- /* Reset the Tx and Rx units. */
- iowrite32(PortReset, ioaddr + SCBPort);
- /* We may get spurious interrupts here. But I don't think that they
- may do much harm. 1999/12/09 SAW */
- udelay(10);
- /* Disable interrupts. */
- iowrite16(SCBMaskAll, ioaddr + SCBCmd);
- synchronize_irq(dev->irq);
- speedo_tx_buffer_gc(dev);
- /* Free as much as possible.
- It helps to recover from a hang because of out-of-memory.
- It also simplifies speedo_resume() in case TX ring is full or
- close-to-be full. */
- speedo_purge_tx(dev);
- speedo_refill_rx_buffers(dev, 1);
- spin_lock_irqsave(&sp->lock, flags);
- speedo_resume(dev);
- sp->rx_mode = -1;
- dev->trans_start = jiffies;
- spin_unlock_irqrestore(&sp->lock, flags);
- set_rx_mode(dev); /* it takes the spinlock itself --SAW */
- /* Reset MII transceiver. Do it before starting the timer to serialize
- mdio_xxx operations. Yes, it's a paranoya :-) 2000/05/09 SAW */
- reset_mii(dev);
- sp->timer.expires = RUN_AT(2*HZ);
- add_timer(&sp->timer);
- }
- return;
-}
-
-static int
-speedo_start_xmit(struct sk_buff *skb, struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int entry;
-
- /* Prevent interrupts from changing the Tx ring from underneath us. */
- unsigned long flags;
-
- spin_lock_irqsave(&sp->lock, flags);
-
- /* Check if there are enough space. */
- if ((int)(sp->cur_tx - sp->dirty_tx) >= TX_QUEUE_LIMIT) {
- printk(KERN_ERR "%s: incorrect tbusy state, fixed.\n", dev->name);
- netif_stop_queue(dev);
- sp->tx_full = 1;
- spin_unlock_irqrestore(&sp->lock, flags);
- return 1;
- }
-
- /* Calculate the Tx descriptor entry. */
- entry = sp->cur_tx++ % TX_RING_SIZE;
-
- sp->tx_skbuff[entry] = skb;
- sp->tx_ring[entry].status =
- cpu_to_le32(CmdSuspend | CmdTx | CmdTxFlex);
- if (!(entry & ((TX_RING_SIZE>>2)-1)))
- sp->tx_ring[entry].status |= cpu_to_le32(CmdIntr);
- sp->tx_ring[entry].link =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, sp->cur_tx % TX_RING_SIZE));
- sp->tx_ring[entry].tx_desc_addr =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, entry) + TX_DESCR_BUF_OFFSET);
- /* The data region is always in one buffer descriptor. */
- sp->tx_ring[entry].count = cpu_to_le32(sp->tx_threshold);
- sp->tx_ring[entry].tx_buf_addr0 =
- cpu_to_le32(pci_map_single(sp->pdev, skb->data,
- skb->len, PCI_DMA_TODEVICE));
- sp->tx_ring[entry].tx_buf_size0 = cpu_to_le32(skb->len);
-
- /* workaround for hardware bug on 10 mbit half duplex */
-
- if ((sp->partner == 0) && (sp->chip_id == 1)) {
- wait_for_cmd_done(dev, sp);
- iowrite8(0 , ioaddr + SCBCmd);
- udelay(1);
- }
-
- /* Trigger the command unit resume. */
- wait_for_cmd_done(dev, sp);
- clear_suspend(sp->last_cmd);
- /* We want the time window between clearing suspend flag on the previous
- command and resuming CU to be as small as possible.
- Interrupts in between are very undesired. --SAW */
- iowrite8(CUResume, ioaddr + SCBCmd);
- sp->last_cmd = (struct descriptor *)&sp->tx_ring[entry];
-
- /* Leave room for set_rx_mode(). If there is no more space than reserved
- for multicast filter mark the ring as full. */
- if ((int)(sp->cur_tx - sp->dirty_tx) >= TX_QUEUE_LIMIT) {
- netif_stop_queue(dev);
- sp->tx_full = 1;
- }
-
- spin_unlock_irqrestore(&sp->lock, flags);
-
- dev->trans_start = jiffies;
-
- return 0;
-}
-
-static void speedo_tx_buffer_gc(struct net_device *dev)
-{
- unsigned int dirty_tx;
- struct speedo_private *sp = netdev_priv(dev);
-
- dirty_tx = sp->dirty_tx;
- while ((int)(sp->cur_tx - dirty_tx) > 0) {
- int entry = dirty_tx % TX_RING_SIZE;
- int status = le32_to_cpu(sp->tx_ring[entry].status);
-
- if (netif_msg_tx_done(sp))
- printk(KERN_DEBUG " scavenge candidate %d status %4.4x.\n",
- entry, status);
- if ((status & StatusComplete) == 0)
- break; /* It still hasn't been processed. */
- if (status & TxUnderrun)
- if (sp->tx_threshold < 0x01e08000) {
- if (netif_msg_tx_err(sp))
- printk(KERN_DEBUG "%s: TX underrun, threshold adjusted.\n",
- dev->name);
- sp->tx_threshold += 0x00040000;
- }
- /* Free the original skb. */
- if (sp->tx_skbuff[entry]) {
- sp->stats.tx_packets++; /* Count only user packets. */
- sp->stats.tx_bytes += sp->tx_skbuff[entry]->len;
- pci_unmap_single(sp->pdev,
- le32_to_cpu(sp->tx_ring[entry].tx_buf_addr0),
- sp->tx_skbuff[entry]->len, PCI_DMA_TODEVICE);
- dev_kfree_skb_irq(sp->tx_skbuff[entry]);
- sp->tx_skbuff[entry] = NULL;
- }
- dirty_tx++;
- }
-
- if (netif_msg_tx_err(sp) && (int)(sp->cur_tx - dirty_tx) > TX_RING_SIZE) {
- printk(KERN_ERR "out-of-sync dirty pointer, %d vs. %d,"
- " full=%d.\n",
- dirty_tx, sp->cur_tx, sp->tx_full);
- dirty_tx += TX_RING_SIZE;
- }
-
- while (sp->mc_setup_head != NULL
- && (int)(dirty_tx - sp->mc_setup_head->tx - 1) > 0) {
- struct speedo_mc_block *t;
- if (netif_msg_tx_err(sp))
- printk(KERN_DEBUG "%s: freeing mc frame.\n", dev->name);
- pci_unmap_single(sp->pdev, sp->mc_setup_head->frame_dma,
- sp->mc_setup_head->len, PCI_DMA_TODEVICE);
- t = sp->mc_setup_head->next;
- kfree(sp->mc_setup_head);
- sp->mc_setup_head = t;
- }
- if (sp->mc_setup_head == NULL)
- sp->mc_setup_tail = NULL;
-
- sp->dirty_tx = dirty_tx;
-}
-
-/* The interrupt handler does all of the Rx thread work and cleans up
- after the Tx thread. */
-static irqreturn_t speedo_interrupt(int irq, void *dev_instance, struct pt_regs *regs)
-{
- struct net_device *dev = (struct net_device *)dev_instance;
- struct speedo_private *sp;
- void __iomem *ioaddr;
- long boguscnt = max_interrupt_work;
- unsigned short status;
- unsigned int handled = 0;
-
- sp = netdev_priv(dev);
- ioaddr = sp->regs;
-
-#ifndef final_version
- /* A lock to prevent simultaneous entry on SMP machines. */
- if (test_and_set_bit(0, (void*)&sp->in_interrupt)) {
- printk(KERN_ERR"%s: SMP simultaneous entry of an interrupt handler.\n",
- dev->name);
- sp->in_interrupt = 0; /* Avoid halting machine. */
- return IRQ_NONE;
- }
-#endif
-
- do {
- status = ioread16(ioaddr + SCBStatus);
- /* Acknowledge all of the current interrupt sources ASAP. */
- /* Will change from 0xfc00 to 0xff00 when we start handling
- FCP and ER interrupts --Dragan */
- iowrite16(status & 0xfc00, ioaddr + SCBStatus);
-
- if (netif_msg_intr(sp))
- printk(KERN_DEBUG "%s: interrupt status=%#4.4x.\n",
- dev->name, status);
-
- if ((status & 0xfc00) == 0)
- break;
- handled = 1;
-
-
- if ((status & 0x5000) || /* Packet received, or Rx error. */
- (sp->rx_ring_state&(RrNoMem|RrPostponed)) == RrPostponed)
- /* Need to gather the postponed packet. */
- speedo_rx(dev);
-
- /* Always check if all rx buffers are allocated. --SAW */
- speedo_refill_rx_buffers(dev, 0);
-
- spin_lock(&sp->lock);
- /*
- * The chip may have suspended reception for various reasons.
- * Check for that, and re-prime it should this be the case.
- */
- switch ((status >> 2) & 0xf) {
- case 0: /* Idle */
- break;
- case 1: /* Suspended */
- case 2: /* No resources (RxFDs) */
- case 9: /* Suspended with no more RBDs */
- case 10: /* No resources due to no RBDs */
- case 12: /* Ready with no RBDs */
- speedo_rx_soft_reset(dev);
- break;
- case 3: case 5: case 6: case 7: case 8:
- case 11: case 13: case 14: case 15:
- /* these are all reserved values */
- break;
- }
-
-
- /* User interrupt, Command/Tx unit interrupt or CU not active. */
- if (status & 0xA400) {
- speedo_tx_buffer_gc(dev);
- if (sp->tx_full
- && (int)(sp->cur_tx - sp->dirty_tx) < TX_QUEUE_UNFULL) {
- /* The ring is no longer full. */
- sp->tx_full = 0;
- netif_wake_queue(dev); /* Attention: under a spinlock. --SAW */
- }
- }
-
- spin_unlock(&sp->lock);
-
- if (--boguscnt < 0) {
- printk(KERN_ERR "%s: Too much work at interrupt, status=0x%4.4x.\n",
- dev->name, status);
- /* Clear all interrupt sources. */
- /* Will change from 0xfc00 to 0xff00 when we start handling
- FCP and ER interrupts --Dragan */
- iowrite16(0xfc00, ioaddr + SCBStatus);
- break;
- }
- } while (1);
-
- if (netif_msg_intr(sp))
- printk(KERN_DEBUG "%s: exiting interrupt, status=%#4.4x.\n",
- dev->name, ioread16(ioaddr + SCBStatus));
-
- clear_bit(0, (void*)&sp->in_interrupt);
- return IRQ_RETVAL(handled);
-}
-
-static inline struct RxFD *speedo_rx_alloc(struct net_device *dev, int entry)
-{
- struct speedo_private *sp = netdev_priv(dev);
- struct RxFD *rxf;
- struct sk_buff *skb;
- /* Get a fresh skbuff to replace the consumed one. */
- skb = dev_alloc_skb(PKT_BUF_SZ + sizeof(struct RxFD));
- if (skb)
- rx_align(skb); /* Align IP on 16 byte boundary */
- sp->rx_skbuff[entry] = skb;
- if (skb == NULL) {
- sp->rx_ringp[entry] = NULL;
- return NULL;
- }
- rxf = sp->rx_ringp[entry] = (struct RxFD *)skb->data;
- sp->rx_ring_dma[entry] =
- pci_map_single(sp->pdev, rxf,
- PKT_BUF_SZ + sizeof(struct RxFD), PCI_DMA_FROMDEVICE);
- skb->dev = dev;
- skb_reserve(skb, sizeof(struct RxFD));
- rxf->rx_buf_addr = 0xffffffff;
- pci_dma_sync_single_for_device(sp->pdev, sp->rx_ring_dma[entry],
- sizeof(struct RxFD), PCI_DMA_TODEVICE);
- return rxf;
-}
-
-static inline void speedo_rx_link(struct net_device *dev, int entry,
- struct RxFD *rxf, dma_addr_t rxf_dma)
-{
- struct speedo_private *sp = netdev_priv(dev);
- rxf->status = cpu_to_le32(0xC0000001); /* '1' for driver use only. */
- rxf->link = 0; /* None yet. */
- rxf->count = cpu_to_le32(PKT_BUF_SZ << 16);
- sp->last_rxf->link = cpu_to_le32(rxf_dma);
- sp->last_rxf->status &= cpu_to_le32(~0xC0000000);
- pci_dma_sync_single_for_device(sp->pdev, sp->last_rxf_dma,
- sizeof(struct RxFD), PCI_DMA_TODEVICE);
- sp->last_rxf = rxf;
- sp->last_rxf_dma = rxf_dma;
-}
-
-static int speedo_refill_rx_buf(struct net_device *dev, int force)
-{
- struct speedo_private *sp = netdev_priv(dev);
- int entry;
- struct RxFD *rxf;
-
- entry = sp->dirty_rx % RX_RING_SIZE;
- if (sp->rx_skbuff[entry] == NULL) {
- rxf = speedo_rx_alloc(dev, entry);
- if (rxf == NULL) {
- unsigned int forw;
- int forw_entry;
- if (netif_msg_rx_err(sp) || !(sp->rx_ring_state & RrOOMReported)) {
- printk(KERN_WARNING "%s: can't fill rx buffer (force %d)!\n",
- dev->name, force);
- sp->rx_ring_state |= RrOOMReported;
- }
- speedo_show_state(dev);
- if (!force)
- return -1; /* Better luck next time! */
- /* Borrow an skb from one of next entries. */
- for (forw = sp->dirty_rx + 1; forw != sp->cur_rx; forw++)
- if (sp->rx_skbuff[forw % RX_RING_SIZE] != NULL)
- break;
- if (forw == sp->cur_rx)
- return -1;
- forw_entry = forw % RX_RING_SIZE;
- sp->rx_skbuff[entry] = sp->rx_skbuff[forw_entry];
- sp->rx_skbuff[forw_entry] = NULL;
- rxf = sp->rx_ringp[forw_entry];
- sp->rx_ringp[forw_entry] = NULL;
- sp->rx_ringp[entry] = rxf;
- }
- } else {
- rxf = sp->rx_ringp[entry];
- }
- speedo_rx_link(dev, entry, rxf, sp->rx_ring_dma[entry]);
- sp->dirty_rx++;
- sp->rx_ring_state &= ~(RrNoMem|RrOOMReported); /* Mark the progress. */
- return 0;
-}
-
-static void speedo_refill_rx_buffers(struct net_device *dev, int force)
-{
- struct speedo_private *sp = netdev_priv(dev);
-
- /* Refill the RX ring. */
- while ((int)(sp->cur_rx - sp->dirty_rx) > 0 &&
- speedo_refill_rx_buf(dev, force) != -1);
-}
-
-static int
-speedo_rx(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- int entry = sp->cur_rx % RX_RING_SIZE;
- int rx_work_limit = sp->dirty_rx + RX_RING_SIZE - sp->cur_rx;
- int alloc_ok = 1;
- int npkts = 0;
-
- if (netif_msg_intr(sp))
- printk(KERN_DEBUG " In speedo_rx().\n");
- /* If we own the next entry, it's a new packet. Send it up. */
- while (sp->rx_ringp[entry] != NULL) {
- int status;
- int pkt_len;
-
- pci_dma_sync_single_for_cpu(sp->pdev, sp->rx_ring_dma[entry],
- sizeof(struct RxFD), PCI_DMA_FROMDEVICE);
- status = le32_to_cpu(sp->rx_ringp[entry]->status);
- pkt_len = le32_to_cpu(sp->rx_ringp[entry]->count) & 0x3fff;
-
- if (!(status & RxComplete))
- break;
-
- if (--rx_work_limit < 0)
- break;
-
- /* Check for a rare out-of-memory case: the current buffer is
- the last buffer allocated in the RX ring. --SAW */
- if (sp->last_rxf == sp->rx_ringp[entry]) {
- /* Postpone the packet. It'll be reaped at an interrupt when this
- packet is no longer the last packet in the ring. */
- if (netif_msg_rx_err(sp))
- printk(KERN_DEBUG "%s: RX packet postponed!\n",
- dev->name);
- sp->rx_ring_state |= RrPostponed;
- break;
- }
-
- if (netif_msg_rx_status(sp))
- printk(KERN_DEBUG " speedo_rx() status %8.8x len %d.\n", status,
- pkt_len);
- if ((status & (RxErrTooBig|RxOK|0x0f90)) != RxOK) {
- if (status & RxErrTooBig)
- printk(KERN_ERR "%s: Ethernet frame overran the Rx buffer, "
- "status %8.8x!\n", dev->name, status);
- else if (! (status & RxOK)) {
- /* There was a fatal error. This *should* be impossible. */
- sp->stats.rx_errors++;
- printk(KERN_ERR "%s: Anomalous event in speedo_rx(), "
- "status %8.8x.\n",
- dev->name, status);
- }
- } else {
- struct sk_buff *skb;
-
- /* Check if the packet is long enough to just accept without
- copying to a properly sized skbuff. */
- if (pkt_len < rx_copybreak
- && (skb = dev_alloc_skb(pkt_len + 2)) != 0) {
- skb->dev = dev;
- skb_reserve(skb, 2); /* Align IP on 16 byte boundaries */
- /* 'skb_put()' points to the start of sk_buff data area. */
- pci_dma_sync_single_for_cpu(sp->pdev, sp->rx_ring_dma[entry],
- sizeof(struct RxFD) + pkt_len,
- PCI_DMA_FROMDEVICE);
-
-#if 1 || USE_IP_CSUM
- /* Packet is in one chunk -- we can copy + cksum. */
- eth_copy_and_sum(skb, sp->rx_skbuff[entry]->data, pkt_len, 0);
- skb_put(skb, pkt_len);
-#else
- memcpy(skb_put(skb, pkt_len), sp->rx_skbuff[entry]->data,
- pkt_len);
-#endif
- pci_dma_sync_single_for_device(sp->pdev, sp->rx_ring_dma[entry],
- sizeof(struct RxFD) + pkt_len,
- PCI_DMA_FROMDEVICE);
- npkts++;
- } else {
- /* Pass up the already-filled skbuff. */
- skb = sp->rx_skbuff[entry];
- if (skb == NULL) {
- printk(KERN_ERR "%s: Inconsistent Rx descriptor chain.\n",
- dev->name);
- break;
- }
- sp->rx_skbuff[entry] = NULL;
- skb_put(skb, pkt_len);
- npkts++;
- sp->rx_ringp[entry] = NULL;
- pci_unmap_single(sp->pdev, sp->rx_ring_dma[entry],
- PKT_BUF_SZ + sizeof(struct RxFD),
- PCI_DMA_FROMDEVICE);
- }
- skb->protocol = eth_type_trans(skb, dev);
- netif_rx(skb);
- dev->last_rx = jiffies;
- sp->stats.rx_packets++;
- sp->stats.rx_bytes += pkt_len;
- }
- entry = (++sp->cur_rx) % RX_RING_SIZE;
- sp->rx_ring_state &= ~RrPostponed;
- /* Refill the recently taken buffers.
- Do it one-by-one to handle traffic bursts better. */
- if (alloc_ok && speedo_refill_rx_buf(dev, 0) == -1)
- alloc_ok = 0;
- }
-
- /* Try hard to refill the recently taken buffers. */
- speedo_refill_rx_buffers(dev, 1);
-
- if (npkts)
- sp->last_rx_time = jiffies;
-
- return 0;
-}
-
-static int
-speedo_close(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- int i;
-
- netdevice_stop(dev);
- netif_stop_queue(dev);
-
- if (netif_msg_ifdown(sp))
- printk(KERN_DEBUG "%s: Shutting down ethercard, status was %4.4x.\n",
- dev->name, ioread16(ioaddr + SCBStatus));
-
- /* Shut off the media monitoring timer. */
- del_timer_sync(&sp->timer);
-
- iowrite16(SCBMaskAll, ioaddr + SCBCmd);
-
- /* Shutting down the chip nicely fails to disable flow control. So.. */
- iowrite32(PortPartialReset, ioaddr + SCBPort);
- ioread32(ioaddr + SCBPort); /* flush posted write */
- /*
- * The chip requires a 10 microsecond quiet period. Wait here!
- */
- udelay(10);
-
- free_irq(dev->irq, dev);
- speedo_show_state(dev);
-
- /* Free all the skbuffs in the Rx and Tx queues. */
- for (i = 0; i < RX_RING_SIZE; i++) {
- struct sk_buff *skb = sp->rx_skbuff[i];
- sp->rx_skbuff[i] = NULL;
- /* Clear the Rx descriptors. */
- if (skb) {
- pci_unmap_single(sp->pdev,
- sp->rx_ring_dma[i],
- PKT_BUF_SZ + sizeof(struct RxFD), PCI_DMA_FROMDEVICE);
- dev_kfree_skb(skb);
- }
- }
-
- for (i = 0; i < TX_RING_SIZE; i++) {
- struct sk_buff *skb = sp->tx_skbuff[i];
- sp->tx_skbuff[i] = NULL;
- /* Clear the Tx descriptors. */
- if (skb) {
- pci_unmap_single(sp->pdev,
- le32_to_cpu(sp->tx_ring[i].tx_buf_addr0),
- skb->len, PCI_DMA_TODEVICE);
- dev_kfree_skb(skb);
- }
- }
-
- /* Free multicast setting blocks. */
- for (i = 0; sp->mc_setup_head != NULL; i++) {
- struct speedo_mc_block *t;
- t = sp->mc_setup_head->next;
- kfree(sp->mc_setup_head);
- sp->mc_setup_head = t;
- }
- sp->mc_setup_tail = NULL;
- if (netif_msg_ifdown(sp))
- printk(KERN_DEBUG "%s: %d multicast blocks dropped.\n", dev->name, i);
-
- pci_set_power_state(sp->pdev, PCI_D2);
-
- return 0;
-}
-
-/* The Speedo-3 has an especially awkward and unusable method of getting
- statistics out of the chip. It takes an unpredictable length of time
- for the dump-stats command to complete. To avoid a busy-wait loop we
- update the stats with the previous dump results, and then trigger a
- new dump.
-
- Oh, and incoming frames are dropped while executing dump-stats!
- */
-static struct net_device_stats *
-speedo_get_stats(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
-
- /* Update only if the previous dump finished. */
- if (sp->lstats->done_marker == le32_to_cpu(0xA007)) {
- sp->stats.tx_aborted_errors += le32_to_cpu(sp->lstats->tx_coll16_errs);
- sp->stats.tx_window_errors += le32_to_cpu(sp->lstats->tx_late_colls);
- sp->stats.tx_fifo_errors += le32_to_cpu(sp->lstats->tx_underruns);
- sp->stats.tx_fifo_errors += le32_to_cpu(sp->lstats->tx_lost_carrier);
- /*sp->stats.tx_deferred += le32_to_cpu(sp->lstats->tx_deferred);*/
- sp->stats.collisions += le32_to_cpu(sp->lstats->tx_total_colls);
- sp->stats.rx_crc_errors += le32_to_cpu(sp->lstats->rx_crc_errs);
- sp->stats.rx_frame_errors += le32_to_cpu(sp->lstats->rx_align_errs);
- sp->stats.rx_over_errors += le32_to_cpu(sp->lstats->rx_resource_errs);
- sp->stats.rx_fifo_errors += le32_to_cpu(sp->lstats->rx_overrun_errs);
- sp->stats.rx_length_errors += le32_to_cpu(sp->lstats->rx_runt_errs);
- sp->lstats->done_marker = 0x0000;
- if (netif_running(dev)) {
- unsigned long flags;
- /* Take a spinlock to make wait_for_cmd_done and sending the
- command atomic. --SAW */
- spin_lock_irqsave(&sp->lock, flags);
- wait_for_cmd_done(dev, sp);
- iowrite8(CUDumpStats, ioaddr + SCBCmd);
- spin_unlock_irqrestore(&sp->lock, flags);
- }
- }
- return &sp->stats;
-}
-
-static void speedo_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
-{
- struct speedo_private *sp = netdev_priv(dev);
- strncpy(info->driver, "eepro100", sizeof(info->driver)-1);
- strncpy(info->version, version, sizeof(info->version)-1);
- if (sp->pdev)
- strcpy(info->bus_info, pci_name(sp->pdev));
-}
-
-static int speedo_get_settings(struct net_device *dev, struct ethtool_cmd *ecmd)
-{
- struct speedo_private *sp = netdev_priv(dev);
- spin_lock_irq(&sp->lock);
- mii_ethtool_gset(&sp->mii_if, ecmd);
- spin_unlock_irq(&sp->lock);
- return 0;
-}
-
-static int speedo_set_settings(struct net_device *dev, struct ethtool_cmd *ecmd)
-{
- struct speedo_private *sp = netdev_priv(dev);
- int res;
- spin_lock_irq(&sp->lock);
- res = mii_ethtool_sset(&sp->mii_if, ecmd);
- spin_unlock_irq(&sp->lock);
- return res;
-}
-
-static int speedo_nway_reset(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- return mii_nway_restart(&sp->mii_if);
-}
-
-static u32 speedo_get_link(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- return mii_link_ok(&sp->mii_if);
-}
-
-static u32 speedo_get_msglevel(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- return sp->msg_enable;
-}
-
-static void speedo_set_msglevel(struct net_device *dev, u32 v)
-{
- struct speedo_private *sp = netdev_priv(dev);
- sp->msg_enable = v;
-}
-
-static struct ethtool_ops ethtool_ops = {
- .get_drvinfo = speedo_get_drvinfo,
- .get_settings = speedo_get_settings,
- .set_settings = speedo_set_settings,
- .nway_reset = speedo_nway_reset,
- .get_link = speedo_get_link,
- .get_msglevel = speedo_get_msglevel,
- .set_msglevel = speedo_set_msglevel,
-};
-
-static int speedo_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
-{
- struct speedo_private *sp = netdev_priv(dev);
- struct mii_ioctl_data *data = if_mii(rq);
- int phy = sp->phy[0] & 0x1f;
- int saved_acpi;
- int t;
-
- switch(cmd) {
- case SIOCGMIIPHY: /* Get address of MII PHY in use. */
- data->phy_id = phy;
-
- case SIOCGMIIREG: /* Read MII PHY register. */
- /* FIXME: these operations need to be serialized with MDIO
- access from the timeout handler.
- They are currently serialized only with MDIO access from the
- timer routine. 2000/05/09 SAW */
- saved_acpi = pci_set_power_state(sp->pdev, PCI_D0);
- t = del_timer_sync(&sp->timer);
- data->val_out = mdio_read(dev, data->phy_id & 0x1f, data->reg_num & 0x1f);
- if (t)
- add_timer(&sp->timer); /* may be set to the past --SAW */
- pci_set_power_state(sp->pdev, saved_acpi);
- return 0;
-
- case SIOCSMIIREG: /* Write MII PHY register. */
- if (!capable(CAP_NET_ADMIN))
- return -EPERM;
- saved_acpi = pci_set_power_state(sp->pdev, PCI_D0);
- t = del_timer_sync(&sp->timer);
- mdio_write(dev, data->phy_id, data->reg_num, data->val_in);
- if (t)
- add_timer(&sp->timer); /* may be set to the past --SAW */
- pci_set_power_state(sp->pdev, saved_acpi);
- return 0;
- default:
- return -EOPNOTSUPP;
- }
-}
-
-/* Set or clear the multicast filter for this adaptor.
- This is very ugly with Intel chips -- we usually have to execute an
- entire configuration command, plus process a multicast command.
- This is complicated. We must put a large configuration command and
- an arbitrarily-sized multicast command in the transmit list.
- To minimize the disruption -- the previous command might have already
- loaded the link -- we convert the current command block, normally a Tx
- command, into a no-op and link it to the new command.
-*/
-static void set_rx_mode(struct net_device *dev)
-{
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
- struct descriptor *last_cmd;
- char new_rx_mode;
- unsigned long flags;
- int entry, i;
-
- if (dev->flags & IFF_PROMISC) { /* Set promiscuous. */
- new_rx_mode = 3;
- } else if ((dev->flags & IFF_ALLMULTI) ||
- dev->mc_count > multicast_filter_limit) {
- new_rx_mode = 1;
- } else
- new_rx_mode = 0;
-
- if (netif_msg_rx_status(sp))
- printk(KERN_DEBUG "%s: set_rx_mode %d -> %d\n", dev->name,
- sp->rx_mode, new_rx_mode);
-
- if ((int)(sp->cur_tx - sp->dirty_tx) > TX_RING_SIZE - TX_MULTICAST_SIZE) {
- /* The Tx ring is full -- don't add anything! Hope the mode will be
- * set again later. */
- sp->rx_mode = -1;
- return;
- }
-
- if (new_rx_mode != sp->rx_mode) {
- u8 *config_cmd_data;
-
- spin_lock_irqsave(&sp->lock, flags);
- entry = sp->cur_tx++ % TX_RING_SIZE;
- last_cmd = sp->last_cmd;
- sp->last_cmd = (struct descriptor *)&sp->tx_ring[entry];
-
- sp->tx_skbuff[entry] = NULL; /* Redundant. */
- sp->tx_ring[entry].status = cpu_to_le32(CmdSuspend | CmdConfigure);
- sp->tx_ring[entry].link =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, (entry + 1) % TX_RING_SIZE));
- config_cmd_data = (void *)&sp->tx_ring[entry].tx_desc_addr;
- /* Construct a full CmdConfig frame. */
- memcpy(config_cmd_data, i82558_config_cmd, CONFIG_DATA_SIZE);
- config_cmd_data[1] = (txfifo << 4) | rxfifo;
- config_cmd_data[4] = rxdmacount;
- config_cmd_data[5] = txdmacount + 0x80;
- config_cmd_data[15] |= (new_rx_mode & 2) ? 1 : 0;
- /* 0x80 doesn't disable FC 0x84 does.
- Disable Flow control since we are not ACK-ing any FC interrupts
- for now. --Dragan */
- config_cmd_data[19] = 0x84;
- config_cmd_data[19] |= sp->mii_if.full_duplex ? 0x40 : 0;
- config_cmd_data[21] = (new_rx_mode & 1) ? 0x0D : 0x05;
- if (sp->phy[0] & 0x8000) { /* Use the AUI port instead. */
- config_cmd_data[15] |= 0x80;
- config_cmd_data[8] = 0;
- }
- /* Trigger the command unit resume. */
- wait_for_cmd_done(dev, sp);
- clear_suspend(last_cmd);
- iowrite8(CUResume, ioaddr + SCBCmd);
- if ((int)(sp->cur_tx - sp->dirty_tx) >= TX_QUEUE_LIMIT) {
- netif_stop_queue(dev);
- sp->tx_full = 1;
- }
- spin_unlock_irqrestore(&sp->lock, flags);
- }
-
- if (new_rx_mode == 0 && dev->mc_count < 4) {
- /* The simple case of 0-3 multicast list entries occurs often, and
- fits within one tx_ring[] entry. */
- struct dev_mc_list *mclist;
- u16 *setup_params, *eaddrs;
-
- spin_lock_irqsave(&sp->lock, flags);
- entry = sp->cur_tx++ % TX_RING_SIZE;
- last_cmd = sp->last_cmd;
- sp->last_cmd = (struct descriptor *)&sp->tx_ring[entry];
-
- sp->tx_skbuff[entry] = NULL;
- sp->tx_ring[entry].status = cpu_to_le32(CmdSuspend | CmdMulticastList);
- sp->tx_ring[entry].link =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, (entry + 1) % TX_RING_SIZE));
- sp->tx_ring[entry].tx_desc_addr = 0; /* Really MC list count. */
- setup_params = (u16 *)&sp->tx_ring[entry].tx_desc_addr;
- *setup_params++ = cpu_to_le16(dev->mc_count*6);
- /* Fill in the multicast addresses. */
- for (i = 0, mclist = dev->mc_list; i < dev->mc_count;
- i++, mclist = mclist->next) {
- eaddrs = (u16 *)mclist->dmi_addr;
- *setup_params++ = *eaddrs++;
- *setup_params++ = *eaddrs++;
- *setup_params++ = *eaddrs++;
- }
-
- wait_for_cmd_done(dev, sp);
- clear_suspend(last_cmd);
- /* Immediately trigger the command unit resume. */
- iowrite8(CUResume, ioaddr + SCBCmd);
-
- if ((int)(sp->cur_tx - sp->dirty_tx) >= TX_QUEUE_LIMIT) {
- netif_stop_queue(dev);
- sp->tx_full = 1;
- }
- spin_unlock_irqrestore(&sp->lock, flags);
- } else if (new_rx_mode == 0) {
- struct dev_mc_list *mclist;
- u16 *setup_params, *eaddrs;
- struct speedo_mc_block *mc_blk;
- struct descriptor *mc_setup_frm;
- int i;
-
- mc_blk = kmalloc(sizeof(*mc_blk) + 2 + multicast_filter_limit*6,
- GFP_ATOMIC);
- if (mc_blk == NULL) {
- printk(KERN_ERR "%s: Failed to allocate a setup frame.\n",
- dev->name);
- sp->rx_mode = -1; /* We failed, try again. */
- return;
- }
- mc_blk->next = NULL;
- mc_blk->len = 2 + multicast_filter_limit*6;
- mc_blk->frame_dma =
- pci_map_single(sp->pdev, &mc_blk->frame, mc_blk->len,
- PCI_DMA_TODEVICE);
- mc_setup_frm = &mc_blk->frame;
-
- /* Fill the setup frame. */
- if (netif_msg_ifup(sp))
- printk(KERN_DEBUG "%s: Constructing a setup frame at %p.\n",
- dev->name, mc_setup_frm);
- mc_setup_frm->cmd_status =
- cpu_to_le32(CmdSuspend | CmdIntr | CmdMulticastList);
- /* Link set below. */
- setup_params = (u16 *)&mc_setup_frm->params;
- *setup_params++ = cpu_to_le16(dev->mc_count*6);
- /* Fill in the multicast addresses. */
- for (i = 0, mclist = dev->mc_list; i < dev->mc_count;
- i++, mclist = mclist->next) {
- eaddrs = (u16 *)mclist->dmi_addr;
- *setup_params++ = *eaddrs++;
- *setup_params++ = *eaddrs++;
- *setup_params++ = *eaddrs++;
- }
-
- /* Disable interrupts while playing with the Tx Cmd list. */
- spin_lock_irqsave(&sp->lock, flags);
-
- if (sp->mc_setup_tail)
- sp->mc_setup_tail->next = mc_blk;
- else
- sp->mc_setup_head = mc_blk;
- sp->mc_setup_tail = mc_blk;
- mc_blk->tx = sp->cur_tx;
-
- entry = sp->cur_tx++ % TX_RING_SIZE;
- last_cmd = sp->last_cmd;
- sp->last_cmd = mc_setup_frm;
-
- /* Change the command to a NoOp, pointing to the CmdMulti command. */
- sp->tx_skbuff[entry] = NULL;
- sp->tx_ring[entry].status = cpu_to_le32(CmdNOp);
- sp->tx_ring[entry].link = cpu_to_le32(mc_blk->frame_dma);
-
- /* Set the link in the setup frame. */
- mc_setup_frm->link =
- cpu_to_le32(TX_RING_ELEM_DMA(sp, (entry + 1) % TX_RING_SIZE));
-
- pci_dma_sync_single_for_device(sp->pdev, mc_blk->frame_dma,
- mc_blk->len, PCI_DMA_TODEVICE);
-
- wait_for_cmd_done(dev, sp);
- clear_suspend(last_cmd);
- /* Immediately trigger the command unit resume. */
- iowrite8(CUResume, ioaddr + SCBCmd);
-
- if ((int)(sp->cur_tx - sp->dirty_tx) >= TX_QUEUE_LIMIT) {
- netif_stop_queue(dev);
- sp->tx_full = 1;
- }
- spin_unlock_irqrestore(&sp->lock, flags);
-
- if (netif_msg_rx_status(sp))
- printk(" CmdMCSetup frame length %d in entry %d.\n",
- dev->mc_count, entry);
- }
-
- sp->rx_mode = new_rx_mode;
-}
-\f
-#ifdef CONFIG_PM
-static int eepro100_suspend(struct pci_dev *pdev, pm_message_t state)
-{
- struct net_device *dev = pci_get_drvdata (pdev);
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
-
- pci_save_state(pdev);
-
- if (!netif_running(dev))
- return 0;
-
- del_timer_sync(&sp->timer);
-
- netif_device_detach(dev);
- iowrite32(PortPartialReset, ioaddr + SCBPort);
-
- /* XXX call pci_set_power_state ()? */
- pci_disable_device(pdev);
- pci_set_power_state (pdev, PCI_D3hot);
- return 0;
-}
-
-static int eepro100_resume(struct pci_dev *pdev)
-{
- struct net_device *dev = pci_get_drvdata (pdev);
- struct speedo_private *sp = netdev_priv(dev);
- void __iomem *ioaddr = sp->regs;
-
- pci_set_power_state(pdev, PCI_D0);
- pci_restore_state(pdev);
- pci_enable_device(pdev);
- pci_set_master(pdev);
-
- if (!netif_running(dev))
- return 0;
-
- /* I'm absolutely uncertain if this part of code may work.
- The problems are:
- - correct hardware reinitialization;
- - correct driver behavior between different steps of the
- reinitialization;
- - serialization with other driver calls.
- 2000/03/08 SAW */
- iowrite16(SCBMaskAll, ioaddr + SCBCmd);
- speedo_resume(dev);
- netif_device_attach(dev);
- sp->rx_mode = -1;
- sp->flow_ctrl = sp->partner = 0;
- set_rx_mode(dev);
- sp->timer.expires = RUN_AT(2*HZ);
- add_timer(&sp->timer);
- return 0;
-}
-#endif /* CONFIG_PM */
-
-static void __devexit eepro100_remove_one (struct pci_dev *pdev)
-{
- struct net_device *dev = pci_get_drvdata (pdev);
- struct speedo_private *sp = netdev_priv(dev);
-
- unregister_netdev(dev);
-
- release_region(pci_resource_start(pdev, 1), pci_resource_len(pdev, 1));
- release_mem_region(pci_resource_start(pdev, 0), pci_resource_len(pdev, 0));
-
- pci_iounmap(pdev, sp->regs);
- pci_free_consistent(pdev, TX_RING_SIZE * sizeof(struct TxFD)
- + sizeof(struct speedo_stats),
- sp->tx_ring, sp->tx_ring_dma);
- pci_disable_device(pdev);
- free_netdev(dev);
-}
-\f
-static struct pci_device_id eepro100_pci_tbl[] = {
- { PCI_VENDOR_ID_INTEL, 0x1229, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1209, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1029, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1030, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1031, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1032, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1033, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1034, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1035, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1036, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1037, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1038, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1039, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x103A, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x103B, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x103C, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x103D, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x103E, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1050, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1059, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x1227, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x2449, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x2459, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x245D, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x5200, PCI_ANY_ID, PCI_ANY_ID, },
- { PCI_VENDOR_ID_INTEL, 0x5201, PCI_ANY_ID, PCI_ANY_ID, },
- { 0,}
-};
-MODULE_DEVICE_TABLE(pci, eepro100_pci_tbl);
-
-static struct pci_driver eepro100_driver = {
- .name = "eepro100",
- .id_table = eepro100_pci_tbl,
- .probe = eepro100_init_one,
- .remove = __devexit_p(eepro100_remove_one),
-#ifdef CONFIG_PM
- .suspend = eepro100_suspend,
- .resume = eepro100_resume,
-#endif /* CONFIG_PM */
-};
-
-static int __init eepro100_init_module(void)
-{
-#ifdef MODULE
- printk(version);
-#endif
- return pci_module_init(&eepro100_driver);
-}
-
-static void __exit eepro100_cleanup_module(void)
-{
- pci_unregister_driver(&eepro100_driver);
-}
-
-module_init(eepro100_init_module);
-module_exit(eepro100_cleanup_module);
-\f
-/*
- * Local variables:
- * compile-command: "gcc -DMODULE -D__KERNEL__ -I/usr/src/linux/net/inet -Wall -Wstrict-prototypes -O6 -c eepro100.c `[ -f /usr/include/linux/modversions.h ] && echo -DMODVERSIONS`"
- * c-indent-level: 4
- * c-basic-offset: 4
- * tab-width: 4
- * End:
- */
^ permalink raw reply
* Re: [DEBUG INFO]IPv6: sleeping function called from invalid context.
From: YOSHIFUJI Hideaki / 吉藤英明 @ 2005-11-18 0:47 UTC (permalink / raw)
To: yanzheng; +Cc: netdev, linux-kernel, yoshfuji
In-Reply-To: <437D23EB.4020204@21cn.com>
In article <437D23EB.4020204@21cn.com> (at Fri, 18 Nov 2005 08:44:27 +0800), Yan Zheng <yanzheng@21cn.com> says:
> I get follow message when switch to single user mode and the kernel version is 2.6.15-rc1-git5.
:
> Nov 18 08:26:23 localhost kernel: Debug: sleeping function called from invalid context at mm/slab.c:2472
> Nov 18 08:26:23 localhost kernel: in_atomic():1, irqs_disabled():0
> Nov 18 08:26:23 localhost kernel: [<c0149d5a>] kmem_cache_alloc+0x5a/0x70
> Nov 18 08:26:23 localhost kernel: [<e0f47336>] inet6_dump_fib+0xb6/0x110 [ipv6]
I remember someone replaced GFP_ATOMIC with GFP_KERNEL...
--yoshfuji
^ permalink raw reply
* [DEBUG INFO]IPv6: sleeping function called from invalid context.
From: Yan Zheng @ 2005-11-18 0:44 UTC (permalink / raw)
To: netdev; +Cc: linux-kernel, yoshfuji
I get follow message when switch to single user mode and the kernel version is 2.6.15-rc1-git5.
Regards
Nov 18 08:26:23 localhost kernel: Debug: sleeping function called from invalid context at mm/slab.c:2472
Nov 18 08:26:23 localhost kernel: in_atomic():1, irqs_disabled():0
Nov 18 08:26:23 localhost kernel: [<c0149d5a>] kmem_cache_alloc+0x5a/0x70
Nov 18 08:26:23 localhost kernel: [<e0f47336>] inet6_dump_fib+0xb6/0x110 [ipv6]
Nov 18 08:26:23 localhost kernel: [<c02e129c>] netlink_dump+0x4c/0x1e0
Nov 18 08:26:23 localhost kernel: [<c02e150a>] netlink_dump_start+0xda/0x170
Nov 18 08:26:23 localhost kernel: [<c02d2d05>] rtnetlink_rcv_msg+0x1d5/0x250
Nov 18 08:26:23 localhost kernel: [<e0f47280>] inet6_dump_fib+0x0/0x110 [ipv6]
Nov 18 08:26:23 localhost kernel: [<c02d2b30>] rtnetlink_rcv_msg+0x0/0x250
Nov 18 08:26:23 localhost kernel: [<c02e17ed>] netlink_rcv_skb+0x4d/0x90
Nov 18 08:26:23 localhost kernel: [<c02d2b30>] rtnetlink_rcv_msg+0x0/0x250
Nov 18 08:26:23 localhost kernel: [<c02e1860>] netlink_run_queue+0x30/0xc0
Nov 18 08:26:23 localhost kernel: [<c02d2b03>] rtnetlink_rcv+0x23/0x50
Nov 18 08:26:23 localhost kernel: [<c02e1092>] netlink_data_ready+0x12/0x60
Nov 18 08:26:23 localhost kernel: [<c0335414>] _spin_unlock_irqrestore+0x14/0x30
Nov 18 08:26:23 localhost kernel: [<c02e0284>] netlink_sendskb+0x24/0x50
Nov 18 08:26:23 localhost kernel: [<c02e0d3c>] netlink_sendmsg+0x29c/0x350
Nov 18 08:26:23 localhost kernel: [<c02be551>] sock_sendmsg+0x111/0x150
Nov 18 08:26:23 localhost kernel: [<c01c56f0>] avc_lookup+0xc0/0x120
Nov 18 08:26:23 localhost kernel: [<c0131f10>] autoremove_wake_function+0x0/0x50
Nov 18 08:26:23 localhost kernel: [<c02bdfff>] move_addr_to_user+0x5f/0x70
Nov 18 08:26:23 localhost kernel: [<c02c0628>] sys_recvmsg+0x1b8/0x230
Nov 18 08:26:23 localhost kernel: [<c013ef94>] audit_sockaddr+0x54/0xc0
Nov 18 08:26:23 localhost kernel: [<c02bfded>] sys_sendto+0x10d/0x150
Nov 18 08:26:23 localhost kernel: [<c0141b6d>] filemap_nopage+0x30d/0x3e0
Nov 18 08:26:23 localhost kernel: [<c0145424>] __alloc_pages+0x64/0x310
Nov 18 08:26:23 localhost kernel: [<c013db8e>] audit_filter_syscall+0x4e/0x130
Nov 18 08:26:23 localhost kernel: [<c013db8e>] audit_filter_syscall+0x4e/0x130
Nov 18 08:26:23 localhost kernel: [<c02c0865>] sys_socketcall+0x1c5/0x2a0
Nov 18 08:26:23 localhost kernel: [<c0103261>] syscall_call+0x7/0xb
Nov 18 08:26:23 localhost kernel: Removing netfilter NETLINK layer.
^ permalink raw reply
* Re: 2.6.15-rc1: NET_CLS_U32 not working?
From: Christian @ 2005-11-17 15:57 UTC (permalink / raw)
To: linux-kernel; +Cc: netdev
In-Reply-To: <437BC98E.9020002@g-house.de>
Christian schrieb:
> Adrian Bunk schrieb:
>>
>> I'm assuming you are trying to insert the new module in your old kernel?
yes, rebuilding the whole kernel (after just enabling NET_CLS_U32 as a
module) makes the warnings go away.
>> This is one of the unfortunate but hardly avoidable cases where adding
>> a module requires installing a new kernel.
i wonder why/if this is really needed. although not critical, this
behaviour is pretty annoying....
thanks,
Christian.
--
BOFH excuse #24:
network packets travelling uphill (use a carrier pigeon)
^ permalink raw reply
* Re: [PATCH] [RFT] ip_tables NUMA optimization
From: Eric Dumazet @ 2005-11-17 15:21 UTC (permalink / raw)
To: Harald Welte; +Cc: Linux Netdev List, Netfilter Development Mailinglist
In-Reply-To: <20051117144328.GG14201@sunbeam.de.gnumonks.org>
Harald Welte a écrit :
> I've hand-merged Eric Dumazet's original ip_tables numa optimization
> patch to current git head. Also, I've ported it to ip6_tables and
> arp_tables.
>
> Since this is 2.6.16 stuff, I don't want to have it applied at this
> time, but merely request testers (esp. for the non-ipv4 part).
>
> Thanks,
>
Hi Harald
Thank you for doing this, since I'm currently too busy with my day job to
update the patch myself.
However I did read your patch and caught one vfree() call wrongly replaced by
a kfree() call in arp_tables.c
>
> @@ -911,6 +930,47 @@ static int get_entries(const struct arpt
> return ret;
> }
>
> +static void free_table_info(struct arpt_table_info *info)
> +{
> + int cpu;
> + for_each_cpu(cpu) {
> + if (info->size <= PAGE_SIZE)
> + kfree(info->entries[cpu]);
> + else
> + kfree(info->entries[cpu]); !!!!
> + }
> + kfree(info);
It should probably use vfree() like :
> + for_each_cpu(cpu) {
> + if (info->size <= PAGE_SIZE)
> + kfree(info->entries[cpu]);
> + else
> + vfree(info->entries[cpu]);
> + }
See you
Eric Dumazet
^ permalink raw reply
* [PATCH] [RFT] ip_tables NUMA optimization
From: Harald Welte @ 2005-11-17 14:43 UTC (permalink / raw)
To: Netfilter Development Mailinglist; +Cc: Linux Netdev List, Eric dumazet
[-- Attachment #1.1: Type: text/plain, Size: 670 bytes --]
I've hand-merged Eric Dumazet's original ip_tables numa optimization
patch to current git head. Also, I've ported it to ip6_tables and
arp_tables.
Since this is 2.6.16 stuff, I don't want to have it applied at this
time, but merely request testers (esp. for the non-ipv4 part).
Thanks,
--
- Harald Welte <laforge@netfilter.org> http://netfilter.org/
============================================================================
"Fragmentation is like classful addressing -- an interesting early
architectural error that shows how much experimentation was going
on while IP was being designed." -- Paul Vixie
[-- Attachment #1.2: vmalloc_node.patch --]
[-- Type: text/plain, Size: 41093 bytes --]
[NETFILTER] ip_tables: NUMA-aware allocation
Part of a performance problem with ip_tables is that memory allocation is not
NUMA aware, but 'only' SMP aware (ie each CPU normally touch separate cache
lines)
Even with small iptables rules, the cost of this misplacement can be high on
common workloads. Instead of using one vmalloc() area (located in the node of
the iptables process), we now allocate an area for each possible CPU, using
vmalloc_node() so that memory should be allocated in the CPU's node if
possible.
Port to arp_tables and ip6_tables by Harald Welte.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
---
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -68,19 +68,14 @@ struct arpt_table_info {
unsigned int initial_entries;
unsigned int hook_entry[NF_ARP_NUMHOOKS];
unsigned int underflow[NF_ARP_NUMHOOKS];
- char entries[0] __attribute__((aligned(SMP_CACHE_BYTES)));
+ void *entries[NR_CPUS];
};
static LIST_HEAD(arpt_target);
static LIST_HEAD(arpt_tables);
+#define SET_COUNTER(c,b,p) do { (c).bcnt = (b); (c).pcnt = (p); } while(0)
#define ADD_COUNTER(c,b,p) do { (c).bcnt += (b); (c).pcnt += (p); } while(0)
-#ifdef CONFIG_SMP
-#define TABLE_OFFSET(t,p) (SMP_ALIGN((t)->size)*(p))
-#else
-#define TABLE_OFFSET(t,p) 0
-#endif
-
static inline int arp_devaddr_compare(const struct arpt_devaddr_info *ap,
char *hdr_addr, int len)
{
@@ -269,9 +264,7 @@ unsigned int arpt_do_table(struct sk_buf
outdev = out ? out->name : nulldevname;
read_lock_bh(&table->lock);
- table_base = (void *)table->private->entries
- + TABLE_OFFSET(table->private,
- smp_processor_id());
+ table_base = (void *)table->private->entries[smp_processor_id()];
e = get_entry(table_base, table->private->hook_entry[hook]);
back = get_entry(table_base, table->private->underflow[hook]);
@@ -462,7 +455,8 @@ static inline int unconditional(const st
/* Figures out from what hook each rule can be called: returns 0 if
* there are loops. Puts hook bitmask in comefrom.
*/
-static int mark_source_chains(struct arpt_table_info *newinfo, unsigned int valid_hooks)
+static int mark_source_chains(struct arpt_table_info *newinfo,
+ unsigned int valid_hooks, void *entry0)
{
unsigned int hook;
@@ -472,7 +466,7 @@ static int mark_source_chains(struct arp
for (hook = 0; hook < NF_ARP_NUMHOOKS; hook++) {
unsigned int pos = newinfo->hook_entry[hook];
struct arpt_entry *e
- = (struct arpt_entry *)(newinfo->entries + pos);
+ = (struct arpt_entry *)(entry0 + pos);
if (!(valid_hooks & (1 << hook)))
continue;
@@ -514,13 +508,13 @@ static int mark_source_chains(struct arp
goto next;
e = (struct arpt_entry *)
- (newinfo->entries + pos);
+ (entry0 + pos);
} while (oldpos == pos + e->next_offset);
/* Move along one */
size = e->next_offset;
e = (struct arpt_entry *)
- (newinfo->entries + pos + size);
+ (entry0 + pos + size);
e->counters.pcnt = pos;
pos += size;
} else {
@@ -537,7 +531,7 @@ static int mark_source_chains(struct arp
newpos = pos + e->next_offset;
}
e = (struct arpt_entry *)
- (newinfo->entries + newpos);
+ (entry0 + newpos);
e->counters.pcnt = pos;
pos = newpos;
}
@@ -689,6 +683,7 @@ static inline int cleanup_entry(struct a
static int translate_table(const char *name,
unsigned int valid_hooks,
struct arpt_table_info *newinfo,
+ void *entry0,
unsigned int size,
unsigned int number,
const unsigned int *hook_entries,
@@ -713,8 +708,8 @@ static int translate_table(const char *n
ret = ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
check_entry_size_and_hooks,
newinfo,
- newinfo->entries,
- newinfo->entries + size,
+ entry0,
+ entry0 + size,
hook_entries, underflows, &i);
duprintf("translate_table: ARPT_ENTRY_ITERATE gives %d\n", ret);
if (ret != 0)
@@ -743,29 +738,26 @@ static int translate_table(const char *n
}
}
- if (!mark_source_chains(newinfo, valid_hooks)) {
+ if (!mark_source_chains(newinfo, valid_hooks, entry0)) {
duprintf("Looping hook\n");
return -ELOOP;
}
/* Finally, each sanity check must pass */
i = 0;
- ret = ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ ret = ARPT_ENTRY_ITERATE(entry0, newinfo->size,
check_entry, name, size, &i);
if (ret != 0) {
- ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ ARPT_ENTRY_ITERATE(entry0, newinfo->size,
cleanup_entry, &i);
return ret;
}
/* And one copy for every other CPU */
for_each_cpu(i) {
- if (i == 0)
- continue;
- memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
- newinfo->entries,
- SMP_ALIGN(newinfo->size));
+ if (newinfo->entries[i] && newinfo->entries[i] != entry0)
+ memcpy(newinfo->entries[smp_processor_id()], entry0, newinfo->size);
}
return ret;
@@ -807,20 +799,45 @@ static inline int add_entry_to_counter(c
return 0;
}
+static inline int set_entry_to_counter(const struct arpt_entry *e,
+ struct arpt_counters total[],
+ unsigned int *i)
+{
+ SET_COUNTER(total[*i], e->counters.bcnt, e->counters.pcnt);
+
+ (*i)++;
+ return 0;
+}
+
static void get_counters(const struct arpt_table_info *t,
struct arpt_counters counters[])
{
unsigned int cpu;
unsigned int i;
+ unsigned int curcpu = get_cpu();
+
+ /* Instead of clearing (by a previous call to memset())
+ * the counters and using adds, we set the counters
+ * with data used by current CPU */
+
+ i = 0;
+ ARPT_ENTRY_ITERATE(t->entries[curcpu],
+ t->size,
+ set_entry_to_counter,
+ counters,
+ &i);
for_each_cpu(cpu) {
+ if (cpu == curcpu)
+ continue;
i = 0;
- ARPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
+ ARPT_ENTRY_ITERATE(t->entries[cpu],
t->size,
add_entry_to_counter,
counters,
&i);
}
+ put_cpu();
}
static int copy_entries_to_user(unsigned int total_size,
@@ -831,6 +848,7 @@ static int copy_entries_to_user(unsigned
struct arpt_entry *e;
struct arpt_counters *counters;
int ret = 0;
+ void *loc_cpu_entry;
/* We need atomic snapshot of counters: rest doesn't change
* (other than comefrom, which userspace doesn't care
@@ -843,13 +861,13 @@ static int copy_entries_to_user(unsigned
return -ENOMEM;
/* First, sum counters... */
- memset(counters, 0, countersize);
write_lock_bh(&table->lock);
get_counters(table->private, counters);
write_unlock_bh(&table->lock);
- /* ... then copy entire thing from CPU 0... */
- if (copy_to_user(userptr, table->private->entries, total_size) != 0) {
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ /* ... then copy entire thing ... */
+ if (copy_to_user(userptr, loc_cpu_entry, total_size) != 0) {
ret = -EFAULT;
goto free_counters;
}
@@ -859,7 +877,7 @@ static int copy_entries_to_user(unsigned
for (off = 0, num = 0; off < total_size; off += e->next_offset, num++){
struct arpt_entry_target *t;
- e = (struct arpt_entry *)(table->private->entries + off);
+ e = (struct arpt_entry *)(loc_cpu_entry + off);
if (copy_to_user(userptr + off
+ offsetof(struct arpt_entry, counters),
&counters[num],
@@ -880,6 +898,7 @@ static int copy_entries_to_user(unsigned
}
free_counters:
+ put_cpu();
vfree(counters);
return ret;
}
@@ -911,6 +930,47 @@ static int get_entries(const struct arpt
return ret;
}
+static void free_table_info(struct arpt_table_info *info)
+{
+ int cpu;
+ for_each_cpu(cpu) {
+ if (info->size <= PAGE_SIZE)
+ kfree(info->entries[cpu]);
+ else
+ kfree(info->entries[cpu]);
+ }
+ kfree(info);
+}
+
+static struct arpt_table_info *alloc_table_info(unsigned int size)
+{
+ struct arpt_table_info *newinfo;
+ int cpu;
+
+ newinfo = kzalloc(sizeof(struct arpt_table_info), GFP_KERNEL);
+ if (!newinfo)
+ return NULL;
+
+ newinfo->size = size;
+
+ for_each_cpu(cpu) {
+ if (size <= PAGE_SIZE)
+ newinfo->entries[cpu] = kmalloc_node(size,
+ GFP_KERNEL,
+ cpu_to_node(cpu));
+ else
+ newinfo->entries[cpu] = vmalloc_node(size,
+ cpu_to_node(cpu));
+
+ if (newinfo->entries[cpu] == NULL) {
+ free_table_info(newinfo);
+ return NULL;
+ }
+ }
+
+ return newinfo;
+}
+
static int do_replace(void __user *user, unsigned int len)
{
int ret;
@@ -918,6 +978,7 @@ static int do_replace(void __user *user,
struct arpt_table *t;
struct arpt_table_info *newinfo, *oldinfo;
struct arpt_counters *counters;
+ void *loc_cpu_entry, *loc_cpu_old_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -930,13 +991,13 @@ static int do_replace(void __user *user,
if ((SMP_ALIGN(tmp.size) >> PAGE_SHIFT) + 2 > num_physpages)
return -ENOMEM;
- newinfo = vmalloc(sizeof(struct arpt_table_info)
- + SMP_ALIGN(tmp.size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(tmp.size);
if (!newinfo)
return -ENOMEM;
- if (copy_from_user(newinfo->entries, user + sizeof(tmp),
+ /* choose the copy that is on our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ if (copy_from_user(loc_cpu_entry, user + sizeof(tmp),
tmp.size) != 0) {
ret = -EFAULT;
goto free_newinfo;
@@ -947,10 +1008,9 @@ static int do_replace(void __user *user,
ret = -ENOMEM;
goto free_newinfo;
}
- memset(counters, 0, tmp.num_counters * sizeof(struct arpt_counters));
ret = translate_table(tmp.name, tmp.valid_hooks,
- newinfo, tmp.size, tmp.num_entries,
+ newinfo, loc_cpu_entry, tmp.size, tmp.num_entries,
tmp.hook_entry, tmp.underflow);
if (ret != 0)
goto free_newinfo_counters;
@@ -989,8 +1049,10 @@ static int do_replace(void __user *user,
/* Get the old counters. */
get_counters(oldinfo, counters);
/* Decrease module usage counts and free resource */
- ARPT_ENTRY_ITERATE(oldinfo->entries, oldinfo->size, cleanup_entry,NULL);
- vfree(oldinfo);
+ loc_cpu_old_entry = oldinfo->entries[smp_processor_id()];
+ ARPT_ENTRY_ITERATE(loc_cpu_old_entry, oldinfo->size, cleanup_entry,NULL);
+ put_cpu();
+ free_table_info(oldinfo);
if (copy_to_user(tmp.counters, counters,
sizeof(struct arpt_counters) * tmp.num_counters) != 0)
ret = -EFAULT;
@@ -1002,11 +1064,12 @@ static int do_replace(void __user *user,
module_put(t->me);
up(&arpt_mutex);
free_newinfo_counters_untrans:
- ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry, NULL);
+ ARPT_ENTRY_ITERATE(loc_cpu_entry, newinfo->size, cleanup_entry, NULL);
free_newinfo_counters:
vfree(counters);
free_newinfo:
- vfree(newinfo);
+ put_cpu();
+ free_table_info(newinfo);
return ret;
}
@@ -1030,6 +1093,7 @@ static int do_add_counters(void __user *
struct arpt_counters_info tmp, *paddc;
struct arpt_table *t;
int ret = 0;
+ void *loc_cpu_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -1059,7 +1123,9 @@ static int do_add_counters(void __user *
}
i = 0;
- ARPT_ENTRY_ITERATE(t->private->entries,
+ /* Choose the copy that is on our node */
+ loc_cpu_entry = t->private->entries[smp_processor_id()];
+ ARPT_ENTRY_ITERATE(loc_cpu_entry,
t->private->size,
add_counter_to_entry,
paddc->counters,
@@ -1220,30 +1286,33 @@ int arpt_register_table(struct arpt_tabl
struct arpt_table_info *newinfo;
static struct arpt_table_info bootstrap
= { 0, 0, 0, { 0 }, { 0 }, { } };
+ void *loc_cpu_entry;
- newinfo = vmalloc(sizeof(struct arpt_table_info)
- + SMP_ALIGN(repl->size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(repl->size);
if (!newinfo) {
ret = -ENOMEM;
return ret;
}
- memcpy(newinfo->entries, repl->entries, repl->size);
+
+ /* choose the copy on our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ memcpy(loc_cpu_entry, repl->entries, repl->size);
ret = translate_table(table->name, table->valid_hooks,
- newinfo, repl->size,
+ newinfo, loc_cpu_entry, repl->size,
repl->num_entries,
repl->hook_entry,
repl->underflow);
+ put_cpu();
duprintf("arpt_register_table: translate table gives %d\n", ret);
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
ret = down_interruptible(&arpt_mutex);
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
@@ -1272,20 +1341,24 @@ int arpt_register_table(struct arpt_tabl
return ret;
free_unlock:
- vfree(newinfo);
+ free_table_info(newinfo);
goto unlock;
}
void arpt_unregister_table(struct arpt_table *table)
{
+ void *loc_cpu_entry;
+
down(&arpt_mutex);
LIST_DELETE(&arpt_tables, table);
up(&arpt_mutex);
/* Decrease module usage counts and free resources */
- ARPT_ENTRY_ITERATE(table->private->entries, table->private->size,
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ ARPT_ENTRY_ITERATE(loc_cpu_entry, table->private->size,
cleanup_entry, NULL);
- vfree(table->private);
+ put_cpu();
+ free_table_info(table->private);
}
/* The built-in targets: standard (NULL) and error. */
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -83,11 +83,6 @@ static DECLARE_MUTEX(ipt_mutex);
context stops packets coming through and allows user context to read
the counters or update the rules.
- To be cache friendly on SMP, we arrange them like so:
- [ n-entries ]
- ... cache-align padding ...
- [ n-entries ]
-
Hence the start of any table is given by get_table() below. */
/* The table itself */
@@ -105,20 +100,15 @@ struct ipt_table_info
unsigned int underflow[NF_IP_NUMHOOKS];
/* ipt_entry tables: one per CPU */
- char entries[0] ____cacheline_aligned;
+ void *entries[NR_CPUS];
};
static LIST_HEAD(ipt_target);
static LIST_HEAD(ipt_match);
static LIST_HEAD(ipt_tables);
+#define SET_COUNTER(c,b,p) do { (c).bcnt = (b); (c).pcnt = (p); } while(0)
#define ADD_COUNTER(c,b,p) do { (c).bcnt += (b); (c).pcnt += (p); } while(0)
-#ifdef CONFIG_SMP
-#define TABLE_OFFSET(t,p) (SMP_ALIGN((t)->size)*(p))
-#else
-#define TABLE_OFFSET(t,p) 0
-#endif
-
#if 0
#define down(x) do { printk("DOWN:%u:" #x "\n", __LINE__); down(x); } while(0)
#define down_interruptible(x) ({ int __r; printk("DOWNi:%u:" #x "\n", __LINE__); __r = down_interruptible(x); if (__r != 0) printk("ABORT-DOWNi:%u\n", __LINE__); __r; })
@@ -290,8 +280,7 @@ ipt_do_table(struct sk_buff **pskb,
read_lock_bh(&table->lock);
IP_NF_ASSERT(table->valid_hooks & (1 << hook));
- table_base = (void *)table->private->entries
- + TABLE_OFFSET(table->private, smp_processor_id());
+ table_base = (void *)table->private->entries[smp_processor_id()];
e = get_entry(table_base, table->private->hook_entry[hook]);
#ifdef CONFIG_NETFILTER_DEBUG
@@ -563,7 +552,8 @@ unconditional(const struct ipt_ip *ip)
/* Figures out from what hook each rule can be called: returns 0 if
there are loops. Puts hook bitmask in comefrom. */
static int
-mark_source_chains(struct ipt_table_info *newinfo, unsigned int valid_hooks)
+mark_source_chains(struct ipt_table_info *newinfo,
+ unsigned int valid_hooks, void *entry0)
{
unsigned int hook;
@@ -572,7 +562,7 @@ mark_source_chains(struct ipt_table_info
for (hook = 0; hook < NF_IP_NUMHOOKS; hook++) {
unsigned int pos = newinfo->hook_entry[hook];
struct ipt_entry *e
- = (struct ipt_entry *)(newinfo->entries + pos);
+ = (struct ipt_entry *)(entry0 + pos);
if (!(valid_hooks & (1 << hook)))
continue;
@@ -622,13 +612,13 @@ mark_source_chains(struct ipt_table_info
goto next;
e = (struct ipt_entry *)
- (newinfo->entries + pos);
+ (entry0 + pos);
} while (oldpos == pos + e->next_offset);
/* Move along one */
size = e->next_offset;
e = (struct ipt_entry *)
- (newinfo->entries + pos + size);
+ (entry0 + pos + size);
e->counters.pcnt = pos;
pos += size;
} else {
@@ -645,7 +635,7 @@ mark_source_chains(struct ipt_table_info
newpos = pos + e->next_offset;
}
e = (struct ipt_entry *)
- (newinfo->entries + newpos);
+ (entry0 + newpos);
e->counters.pcnt = pos;
pos = newpos;
}
@@ -855,6 +845,7 @@ static int
translate_table(const char *name,
unsigned int valid_hooks,
struct ipt_table_info *newinfo,
+ void *entry0,
unsigned int size,
unsigned int number,
const unsigned int *hook_entries,
@@ -875,11 +866,11 @@ translate_table(const char *name,
duprintf("translate_table: size %u\n", newinfo->size);
i = 0;
/* Walk through entries, checking offsets. */
- ret = IPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ ret = IPT_ENTRY_ITERATE(entry0, newinfo->size,
check_entry_size_and_hooks,
newinfo,
- newinfo->entries,
- newinfo->entries + size,
+ entry0,
+ entry0 + size,
hook_entries, underflows, &i);
if (ret != 0)
return ret;
@@ -907,27 +898,24 @@ translate_table(const char *name,
}
}
- if (!mark_source_chains(newinfo, valid_hooks))
+ if (!mark_source_chains(newinfo, valid_hooks, entry0))
return -ELOOP;
/* Finally, each sanity check must pass */
i = 0;
- ret = IPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ ret = IPT_ENTRY_ITERATE(entry0, newinfo->size,
check_entry, name, size, &i);
if (ret != 0) {
- IPT_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ IPT_ENTRY_ITERATE(entry0, newinfo->size,
cleanup_entry, &i);
return ret;
}
/* And one copy for every other CPU */
for_each_cpu(i) {
- if (i == 0)
- continue;
- memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
- newinfo->entries,
- SMP_ALIGN(newinfo->size));
+ if (newinfo->entries[i] && newinfo->entries[i] != entry0)
+ memcpy(newinfo->entries[i], entry0, newinfo->size);
}
return ret;
@@ -943,15 +931,12 @@ replace_table(struct ipt_table *table,
#ifdef CONFIG_NETFILTER_DEBUG
{
- struct ipt_entry *table_base;
- unsigned int i;
+ int cpu;
- for_each_cpu(i) {
- table_base =
- (void *)newinfo->entries
- + TABLE_OFFSET(newinfo, i);
-
- table_base->comefrom = 0xdead57ac;
+ for_each_cpu(cpu) {
+ struct ipt_entry *table_base = newinfo->entries[cpu];
+ if (table_base)
+ table_base->comefrom = 0xdead57ac;
}
}
#endif
@@ -986,21 +971,47 @@ add_entry_to_counter(const struct ipt_en
return 0;
}
+static inline int
+set_entry_to_counter(const struct ipt_entry *e,
+ struct ipt_counters total[],
+ unsigned int *i)
+{
+ SET_COUNTER(total[*i], e->counters.bcnt, e->counters.pcnt);
+
+ (*i)++;
+ return 0;
+}
+
static void
get_counters(const struct ipt_table_info *t,
struct ipt_counters counters[])
{
unsigned int cpu;
unsigned int i;
+ unsigned int curcpu = get_cpu();
+
+ /* Instead of clearing (by a previous call to memset())
+ * the counters and using adds, we set the counters
+ * with data used by current CPU */
+
+ i = 0;
+ IPT_ENTRY_ITERATE(t->entries[curcpu],
+ t->size,
+ set_entry_to_counter,
+ counters,
+ &i);
for_each_cpu(cpu) {
+ if (cpu == curcpu)
+ continue;
i = 0;
- IPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
+ IPT_ENTRY_ITERATE(t->entries[cpu],
t->size,
add_entry_to_counter,
counters,
&i);
}
+ put_cpu();
}
static int
@@ -1012,6 +1023,7 @@ copy_entries_to_user(unsigned int total_
struct ipt_entry *e;
struct ipt_counters *counters;
int ret = 0;
+ void *loc_cpu_entry;
/* We need atomic snapshot of counters: rest doesn't change
(other than comefrom, which userspace doesn't care
@@ -1023,13 +1035,14 @@ copy_entries_to_user(unsigned int total_
return -ENOMEM;
/* First, sum counters... */
- memset(counters, 0, countersize);
write_lock_bh(&table->lock);
get_counters(table->private, counters);
write_unlock_bh(&table->lock);
- /* ... then copy entire thing from CPU 0... */
- if (copy_to_user(userptr, table->private->entries, total_size) != 0) {
+ /* choose the copy that is on our node/cpu, ... */
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ /* ... then copy entire thing ... */
+ if (copy_to_user(userptr, loc_cpu_entry, total_size) != 0) {
ret = -EFAULT;
goto free_counters;
}
@@ -1041,7 +1054,7 @@ copy_entries_to_user(unsigned int total_
struct ipt_entry_match *m;
struct ipt_entry_target *t;
- e = (struct ipt_entry *)(table->private->entries + off);
+ e = (struct ipt_entry *)(loc_cpu_entry + off);
if (copy_to_user(userptr + off
+ offsetof(struct ipt_entry, counters),
&counters[num],
@@ -1078,6 +1091,7 @@ copy_entries_to_user(unsigned int total_
}
free_counters:
+ put_cpu();
vfree(counters);
return ret;
}
@@ -1110,6 +1124,45 @@ get_entries(const struct ipt_get_entries
return ret;
}
+static void free_table_info(struct ipt_table_info *info)
+{
+ int cpu;
+ for_each_cpu(cpu) {
+ if (info->size <= PAGE_SIZE)
+ kfree(info->entries[cpu]);
+ else
+ vfree(info->entries[cpu]);
+ }
+ kfree(info);
+}
+
+static struct ipt_table_info *alloc_table_info(unsigned int size)
+{
+ struct ipt_table_info *newinfo;
+ int cpu;
+
+ newinfo = kzalloc(sizeof(struct ipt_table_info), GFP_KERNEL);
+ if (!newinfo)
+ return NULL;
+
+ newinfo->size = size;
+
+ for_each_cpu(cpu) {
+ if (size <= PAGE_SIZE)
+ newinfo->entries[cpu] = kmalloc_node(size,
+ GFP_KERNEL,
+ cpu_to_node(cpu));
+ else
+ newinfo->entries[cpu] = vmalloc_node(size, cpu_to_node(cpu));
+ if (newinfo->entries[cpu] == 0) {
+ free_table_info(newinfo);
+ return NULL;
+ }
+ }
+
+ return newinfo;
+}
+
static int
do_replace(void __user *user, unsigned int len)
{
@@ -1118,6 +1171,7 @@ do_replace(void __user *user, unsigned i
struct ipt_table *t;
struct ipt_table_info *newinfo, *oldinfo;
struct ipt_counters *counters;
+ void *loc_cpu_entry, *loc_cpu_old_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -1130,13 +1184,13 @@ do_replace(void __user *user, unsigned i
if ((SMP_ALIGN(tmp.size) >> PAGE_SHIFT) + 2 > num_physpages)
return -ENOMEM;
- newinfo = vmalloc(sizeof(struct ipt_table_info)
- + SMP_ALIGN(tmp.size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(tmp.size);
if (!newinfo)
return -ENOMEM;
- if (copy_from_user(newinfo->entries, user + sizeof(tmp),
+ /* choose the copy that is our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ if (copy_from_user(loc_cpu_entry, user + sizeof(tmp),
tmp.size) != 0) {
ret = -EFAULT;
goto free_newinfo;
@@ -1147,10 +1201,9 @@ do_replace(void __user *user, unsigned i
ret = -ENOMEM;
goto free_newinfo;
}
- memset(counters, 0, tmp.num_counters * sizeof(struct ipt_counters));
ret = translate_table(tmp.name, tmp.valid_hooks,
- newinfo, tmp.size, tmp.num_entries,
+ newinfo, loc_cpu_entry, tmp.size, tmp.num_entries,
tmp.hook_entry, tmp.underflow);
if (ret != 0)
goto free_newinfo_counters;
@@ -1189,8 +1242,10 @@ do_replace(void __user *user, unsigned i
/* Get the old counters. */
get_counters(oldinfo, counters);
/* Decrease module usage counts and free resource */
- IPT_ENTRY_ITERATE(oldinfo->entries, oldinfo->size, cleanup_entry,NULL);
- vfree(oldinfo);
+ loc_cpu_old_entry = oldinfo->entries[smp_processor_id()];
+ IPT_ENTRY_ITERATE(loc_cpu_old_entry, oldinfo->size, cleanup_entry,NULL);
+ put_cpu();
+ free_table_info(oldinfo);
if (copy_to_user(tmp.counters, counters,
sizeof(struct ipt_counters) * tmp.num_counters) != 0)
ret = -EFAULT;
@@ -1202,11 +1257,12 @@ do_replace(void __user *user, unsigned i
module_put(t->me);
up(&ipt_mutex);
free_newinfo_counters_untrans:
- IPT_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry,NULL);
+ IPT_ENTRY_ITERATE(loc_cpu_entry, newinfo->size, cleanup_entry,NULL);
free_newinfo_counters:
vfree(counters);
free_newinfo:
- vfree(newinfo);
+ put_cpu();
+ free_table_info(newinfo);
return ret;
}
@@ -1239,6 +1295,7 @@ do_add_counters(void __user *user, unsig
struct ipt_counters_info tmp, *paddc;
struct ipt_table *t;
int ret = 0;
+ void *loc_cpu_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -1268,7 +1325,9 @@ do_add_counters(void __user *user, unsig
}
i = 0;
- IPT_ENTRY_ITERATE(t->private->entries,
+ /* Choose the copy that is on our node */
+ loc_cpu_entry = t->private->entries[smp_processor_id()];
+ IPT_ENTRY_ITERATE(loc_cpu_entry,
t->private->size,
add_counter_to_entry,
paddc->counters,
@@ -1460,28 +1519,30 @@ int ipt_register_table(struct ipt_table
struct ipt_table_info *newinfo;
static struct ipt_table_info bootstrap
= { 0, 0, 0, { 0 }, { 0 }, { } };
+ void *loc_cpu_entry;
- newinfo = vmalloc(sizeof(struct ipt_table_info)
- + SMP_ALIGN(repl->size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(repl->size);
if (!newinfo)
return -ENOMEM;
- memcpy(newinfo->entries, repl->entries, repl->size);
+ /* choose the copy on our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ memcpy(loc_cpu_entry, repl->entries, repl->size);
ret = translate_table(table->name, table->valid_hooks,
- newinfo, repl->size,
+ newinfo, loc_cpu_entry, repl->size,
repl->num_entries,
repl->hook_entry,
repl->underflow);
+ put_cpu();
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
ret = down_interruptible(&ipt_mutex);
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
@@ -1510,20 +1571,24 @@ int ipt_register_table(struct ipt_table
return ret;
free_unlock:
- vfree(newinfo);
+ free_table_info(newinfo);
goto unlock;
}
void ipt_unregister_table(struct ipt_table *table)
{
+ void *loc_cpu_entry;
+
down(&ipt_mutex);
LIST_DELETE(&ipt_tables, table);
up(&ipt_mutex);
/* Decrease module usage counts and free resources */
- IPT_ENTRY_ITERATE(table->private->entries, table->private->size,
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ IPT_ENTRY_ITERATE(loc_cpu_entry, table->private->size,
cleanup_entry, NULL);
- vfree(table->private);
+ put_cpu();
+ free_table_info(table->private);
}
/* Returns 1 if the port is matched by the range, 0 otherwise */
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -86,11 +86,6 @@ static DECLARE_MUTEX(ip6t_mutex);
context stops packets coming through and allows user context to read
the counters or update the rules.
- To be cache friendly on SMP, we arrange them like so:
- [ n-entries ]
- ... cache-align padding ...
- [ n-entries ]
-
Hence the start of any table is given by get_table() below. */
/* The table itself */
@@ -108,20 +103,15 @@ struct ip6t_table_info
unsigned int underflow[NF_IP6_NUMHOOKS];
/* ip6t_entry tables: one per CPU */
- char entries[0] ____cacheline_aligned;
+ void *entries[NR_CPUS];
};
static LIST_HEAD(ip6t_target);
static LIST_HEAD(ip6t_match);
static LIST_HEAD(ip6t_tables);
+#define SET_COUNTER(c,b,p) do { (c).bcnt = (b); (c).pcnt = (p); } while(0)
#define ADD_COUNTER(c,b,p) do { (c).bcnt += (b); (c).pcnt += (p); } while(0)
-#ifdef CONFIG_SMP
-#define TABLE_OFFSET(t,p) (SMP_ALIGN((t)->size)*(p))
-#else
-#define TABLE_OFFSET(t,p) 0
-#endif
-
#if 0
#define down(x) do { printk("DOWN:%u:" #x "\n", __LINE__); down(x); } while(0)
#define down_interruptible(x) ({ int __r; printk("DOWNi:%u:" #x "\n", __LINE__); __r = down_interruptible(x); if (__r != 0) printk("ABORT-DOWNi:%u\n", __LINE__); __r; })
@@ -376,8 +366,7 @@ ip6t_do_table(struct sk_buff **pskb,
read_lock_bh(&table->lock);
IP_NF_ASSERT(table->valid_hooks & (1 << hook));
- table_base = (void *)table->private->entries
- + TABLE_OFFSET(table->private, smp_processor_id());
+ table_base = (void *)table->private->entries[smp_processor_id()];
e = get_entry(table_base, table->private->hook_entry[hook]);
#ifdef CONFIG_NETFILTER_DEBUG
@@ -649,7 +638,8 @@ unconditional(const struct ip6t_ip6 *ipv
/* Figures out from what hook each rule can be called: returns 0 if
there are loops. Puts hook bitmask in comefrom. */
static int
-mark_source_chains(struct ip6t_table_info *newinfo, unsigned int valid_hooks)
+mark_source_chains(struct ip6t_table_info *newinfo,
+ unsigned int valid_hooks, void *entry0)
{
unsigned int hook;
@@ -658,7 +648,7 @@ mark_source_chains(struct ip6t_table_inf
for (hook = 0; hook < NF_IP6_NUMHOOKS; hook++) {
unsigned int pos = newinfo->hook_entry[hook];
struct ip6t_entry *e
- = (struct ip6t_entry *)(newinfo->entries + pos);
+ = (struct ip6t_entry *)(entry0 + pos);
if (!(valid_hooks & (1 << hook)))
continue;
@@ -708,13 +698,13 @@ mark_source_chains(struct ip6t_table_inf
goto next;
e = (struct ip6t_entry *)
- (newinfo->entries + pos);
+ (entry0 + pos);
} while (oldpos == pos + e->next_offset);
/* Move along one */
size = e->next_offset;
e = (struct ip6t_entry *)
- (newinfo->entries + pos + size);
+ (entry0 + pos + size);
e->counters.pcnt = pos;
pos += size;
} else {
@@ -731,7 +721,7 @@ mark_source_chains(struct ip6t_table_inf
newpos = pos + e->next_offset;
}
e = (struct ip6t_entry *)
- (newinfo->entries + newpos);
+ (entry0 + newpos);
e->counters.pcnt = pos;
pos = newpos;
}
@@ -941,6 +931,7 @@ static int
translate_table(const char *name,
unsigned int valid_hooks,
struct ip6t_table_info *newinfo,
+ void *entry0,
unsigned int size,
unsigned int number,
const unsigned int *hook_entries,
@@ -964,8 +955,8 @@ translate_table(const char *name,
ret = IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size,
check_entry_size_and_hooks,
newinfo,
- newinfo->entries,
- newinfo->entries + size,
+ entry0,
+ entry0 + size,
hook_entries, underflows, &i);
if (ret != 0)
return ret;
@@ -993,27 +984,24 @@ translate_table(const char *name,
}
}
- if (!mark_source_chains(newinfo, valid_hooks))
+ if (!mark_source_chains(newinfo, valid_hooks, entry0))
return -ELOOP;
/* Finally, each sanity check must pass */
i = 0;
- ret = IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ ret = IP6T_ENTRY_ITERATE(entry0, newinfo->size,
check_entry, name, size, &i);
if (ret != 0) {
- IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size,
+ IP6T_ENTRY_ITERATE(entry0, newinfo->size,
cleanup_entry, &i);
return ret;
}
/* And one copy for every other CPU */
for_each_cpu(i) {
- if (i == 0)
- continue;
- memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
- newinfo->entries,
- SMP_ALIGN(newinfo->size));
+ if (newinfo->entries[i] && newinfo->entries[i] != entry0)
+ memcpy(newinfo->entries[i], entry0, newinfo->size);
}
return ret;
@@ -1029,15 +1017,12 @@ replace_table(struct ip6t_table *table,
#ifdef CONFIG_NETFILTER_DEBUG
{
- struct ip6t_entry *table_base;
- unsigned int i;
-
- for_each_cpu(i) {
- table_base =
- (void *)newinfo->entries
- + TABLE_OFFSET(newinfo, i);
+ int cpu;
- table_base->comefrom = 0xdead57ac;
+ for_each_cpu(cpu) {
+ struct ip6t_entry *table_base = newinfo->entries[cpu];
+ if (table_base)
+ table_base->comefrom = 0xdead57ac;
}
}
#endif
@@ -1072,21 +1057,47 @@ add_entry_to_counter(const struct ip6t_e
return 0;
}
+static inline int
+set_entry_to_counter(const struct ip6t_entry *e,
+ struct ip6t_counters total[],
+ unsigned int *i)
+{
+ SET_COUNTER(total[*i], e->counters.bcnt, e->counters.pcnt);
+
+ (*i)++;
+ return 0;
+}
+
static void
get_counters(const struct ip6t_table_info *t,
struct ip6t_counters counters[])
{
unsigned int cpu;
unsigned int i;
+ unsigned int curcpu = get_cpu();
+
+ /* Instead of clearing (by a previous call to memset())
+ * the counters and using adds, we set the counters
+ * with data used by current CPU */
+
+ i = 0;
+ IP6T_ENTRY_ITERATE(t->entries[curcpu],
+ t->size,
+ set_entry_to_counter,
+ counters,
+ &i);
for_each_cpu(cpu) {
+ if (cpu == curcpu)
+ continue;
i = 0;
- IP6T_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
+ IP6T_ENTRY_ITERATE(t->entries[cpu],
t->size,
add_entry_to_counter,
counters,
&i);
}
+ put_cpu();
}
static int
@@ -1098,6 +1109,7 @@ copy_entries_to_user(unsigned int total_
struct ip6t_entry *e;
struct ip6t_counters *counters;
int ret = 0;
+ void *loc_cpu_entry;
/* We need atomic snapshot of counters: rest doesn't change
(other than comefrom, which userspace doesn't care
@@ -1109,13 +1121,13 @@ copy_entries_to_user(unsigned int total_
return -ENOMEM;
/* First, sum counters... */
- memset(counters, 0, countersize);
write_lock_bh(&table->lock);
get_counters(table->private, counters);
write_unlock_bh(&table->lock);
- /* ... then copy entire thing from CPU 0... */
- if (copy_to_user(userptr, table->private->entries, total_size) != 0) {
+ /* choose the copy that is on ourc node/cpu */
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ if (copy_to_user(userptr, loc_cpu_entry, total_size) != 0) {
ret = -EFAULT;
goto free_counters;
}
@@ -1127,7 +1139,7 @@ copy_entries_to_user(unsigned int total_
struct ip6t_entry_match *m;
struct ip6t_entry_target *t;
- e = (struct ip6t_entry *)(table->private->entries + off);
+ e = (struct ip6t_entry *)(loc_cpu_entry + off);
if (copy_to_user(userptr + off
+ offsetof(struct ip6t_entry, counters),
&counters[num],
@@ -1164,6 +1176,7 @@ copy_entries_to_user(unsigned int total_
}
free_counters:
+ put_cpu();
vfree(counters);
return ret;
}
@@ -1196,6 +1209,46 @@ get_entries(const struct ip6t_get_entrie
return ret;
}
+static void free_table_info(struct ip6t_table_info *info)
+{
+ int cpu;
+ for_each_cpu(cpu) {
+ if (info->size <= PAGE_SIZE)
+ kfree(info->entries[cpu]);
+ else
+ vfree(info->entries[cpu]);
+ }
+ kfree(info);
+}
+
+static struct ip6t_table_info *alloc_table_info(unsigned int size)
+{
+ struct ip6t_table_info *newinfo;
+ int cpu;
+
+ newinfo = kzalloc(sizeof(struct ip6t_table_info), GFP_KERNEL);
+ if (!newinfo)
+ return NULL;
+
+ newinfo->size = size;
+
+ for_each_cpu(cpu) {
+ if (size <= PAGE_SIZE)
+ newinfo->entries[cpu] = kmalloc_node(size,
+ GFP_KERNEL,
+ cpu_to_node(cpu));
+ else
+ newinfo->entries[cpu] = vmalloc_node(size,
+ cpu_to_node(cpu));
+ if (newinfo->entries[cpu] == NULL) {
+ free_table_info(newinfo);
+ return NULL;
+ }
+ }
+
+ return newinfo;
+}
+
static int
do_replace(void __user *user, unsigned int len)
{
@@ -1204,6 +1257,7 @@ do_replace(void __user *user, unsigned i
struct ip6t_table *t;
struct ip6t_table_info *newinfo, *oldinfo;
struct ip6t_counters *counters;
+ void *loc_cpu_entry, *loc_cpu_old_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -1212,13 +1266,13 @@ do_replace(void __user *user, unsigned i
if ((SMP_ALIGN(tmp.size) >> PAGE_SHIFT) + 2 > num_physpages)
return -ENOMEM;
- newinfo = vmalloc(sizeof(struct ip6t_table_info)
- + SMP_ALIGN(tmp.size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(tmp.size);
if (!newinfo)
return -ENOMEM;
- if (copy_from_user(newinfo->entries, user + sizeof(tmp),
+ /* choose the copy that is on our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ if (copy_from_user(loc_cpu_entry, user + sizeof(tmp),
tmp.size) != 0) {
ret = -EFAULT;
goto free_newinfo;
@@ -1229,10 +1283,9 @@ do_replace(void __user *user, unsigned i
ret = -ENOMEM;
goto free_newinfo;
}
- memset(counters, 0, tmp.num_counters * sizeof(struct ip6t_counters));
ret = translate_table(tmp.name, tmp.valid_hooks,
- newinfo, tmp.size, tmp.num_entries,
+ newinfo, loc_cpu_entry, tmp.size, tmp.num_entries,
tmp.hook_entry, tmp.underflow);
if (ret != 0)
goto free_newinfo_counters;
@@ -1271,8 +1324,10 @@ do_replace(void __user *user, unsigned i
/* Get the old counters. */
get_counters(oldinfo, counters);
/* Decrease module usage counts and free resource */
- IP6T_ENTRY_ITERATE(oldinfo->entries, oldinfo->size, cleanup_entry,NULL);
- vfree(oldinfo);
+ loc_cpu_old_entry = oldinfo->entries[smp_processor_id()];
+ IP6T_ENTRY_ITERATE(loc_cpu_old_entry, oldinfo->size, cleanup_entry,NULL);
+ put_cpu();
+ free_table_info(oldinfo);
if (copy_to_user(tmp.counters, counters,
sizeof(struct ip6t_counters) * tmp.num_counters) != 0)
ret = -EFAULT;
@@ -1284,11 +1339,12 @@ do_replace(void __user *user, unsigned i
module_put(t->me);
up(&ip6t_mutex);
free_newinfo_counters_untrans:
- IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry,NULL);
+ IP6T_ENTRY_ITERATE(loc_cpu_entry, newinfo->size, cleanup_entry,NULL);
free_newinfo_counters:
vfree(counters);
free_newinfo:
- vfree(newinfo);
+ put_cpu();
+ free_table_info(newinfo);
return ret;
}
@@ -1321,6 +1377,7 @@ do_add_counters(void __user *user, unsig
struct ip6t_counters_info tmp, *paddc;
struct ip6t_table *t;
int ret = 0;
+ void *loc_cpu_entry;
if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
return -EFAULT;
@@ -1350,7 +1407,9 @@ do_add_counters(void __user *user, unsig
}
i = 0;
- IP6T_ENTRY_ITERATE(t->private->entries,
+ /* Choose the copy that is on our node */
+ loc_cpu_entry = t->private->entries[smp_processor_id()];
+ IP6T_ENTRY_ITERATE(loc_cpu_entry,
t->private->size,
add_counter_to_entry,
paddc->counters,
@@ -1543,28 +1602,30 @@ int ip6t_register_table(struct ip6t_tabl
struct ip6t_table_info *newinfo;
static struct ip6t_table_info bootstrap
= { 0, 0, 0, { 0 }, { 0 }, { } };
+ void *loc_cpu_entry;
- newinfo = vmalloc(sizeof(struct ip6t_table_info)
- + SMP_ALIGN(repl->size) *
- (highest_possible_processor_id()+1));
+ newinfo = alloc_table_info(repl->size);
if (!newinfo)
return -ENOMEM;
- memcpy(newinfo->entries, repl->entries, repl->size);
+ /* choose the copy on our node/cpu */
+ loc_cpu_entry = newinfo->entries[get_cpu()];
+ memcpy(loc_cpu_entry, repl->entries, repl->size);
ret = translate_table(table->name, table->valid_hooks,
- newinfo, repl->size,
+ newinfo, loc_cpu_entry, repl->size,
repl->num_entries,
repl->hook_entry,
repl->underflow);
+ put_cpu();
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
ret = down_interruptible(&ip6t_mutex);
if (ret != 0) {
- vfree(newinfo);
+ free_table_info(newinfo);
return ret;
}
@@ -1593,20 +1654,24 @@ int ip6t_register_table(struct ip6t_tabl
return ret;
free_unlock:
- vfree(newinfo);
+ free_table_info(newinfo);
goto unlock;
}
void ip6t_unregister_table(struct ip6t_table *table)
{
+ void *loc_cpu_entry;
+
down(&ip6t_mutex);
LIST_DELETE(&ip6t_tables, table);
up(&ip6t_mutex);
/* Decrease module usage counts and free resources */
- IP6T_ENTRY_ITERATE(table->private->entries, table->private->size,
+ loc_cpu_entry = table->private->entries[get_cpu()];
+ IP6T_ENTRY_ITERATE(loc_cpu_entry, table->private->size,
cleanup_entry, NULL);
- vfree(table->private);
+ put_cpu();
+ free_table_info(table->private);
}
/* Returns 1 if the port is matched by the range, 0 otherwise */
[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply
* Do away with everything you are indebted for with out sending an other cent
From: reva bradley @ 2005-11-17 9:05 UTC (permalink / raw)
To: Elin Lewis
End their harassment and tell them no more calls.
We have pioneered an advanced system of proven strategies
that will get the creditors and debt collectors off your back for good
Our debt termination program has legally stopped millions of dollars worth
of debt from being collected.
check out our Elimination Program here
http://it.geocities.com/henrys_chrissy/?s=d
take us out, po box above in site
You are right, and I must wait--wait--wait--patiently and silently--until
my bonds are loosed by intelligence rather than chance! It is a dreary fate.
But I must wait--I must wait--I must wait! I'm glad you've come to your
senses, remarked Rob, drily
So, if you've nothing more to say-- No! I have nothing more to say
^ permalink raw reply
* Re: [PATCH 00/10]: Netfilter IPsec support
From: Patrick McHardy @ 2005-11-17 2:35 UTC (permalink / raw)
To: Marco Berizzi; +Cc: netdev, netfilter-devel
In-Reply-To: <BAY103-F240FCD8D8AE36A95868233B25D0@phx.gbl>
Marco Berizzi wrote:
> How are handled NAT-T packets (udp/4500) with these patches?
Instead of ESP packets you see the encapsulated UDP packets
on the netfilter hooks:
(none):~# ping 10.0.0.1 -c 1
PING 10.0.0.1 (10.0.0.1): 56 data bytes
OUTPUT IN= OUT=eth0 SRC=10.0.0.2 DST=10.0.0.1 LEN=84 TOS=0x00
PREC=0x00 TTL=64 ID=0 DF PROTO=ICMP TYPE=8 CODE=0 ID=16899 SEQ=0
POSTROUTING IN= OUT=eth0 SRC=10.0.0.2 DST=10.0.0.1 LEN=84 TOS=0x00
PREC=0x00 TTL=64 ID=0 DF PROTO=ICMP TYPE=8 CODE=0 ID=16899 SEQ=0
OUTPUT IN= OUT=eth0 SRC=10.0.0.2 DST=10.0.0.1 LEN=160 TOS=0x00
PREC=0x00 TTL=64 ID=256 DF PROTO=UDP SPT=4500 DPT=4500 LEN=140
POSTROUTING IN= OUT=eth0 SRC=10.0.0.2 DST=10.0.0.1 LEN=160 TOS=0x00
PREC=0x00 TTL=64 ID=256 DF PROTO=UDP SPT=4500 DPT=4500 LEN=140
PREROUTING IN=eth0 OUT= MAC=fe:fd:0a:00:00:02:36:ec:4f:25:dc:68:08:00
SRC=10.0.0.1 DST=10.0.0.2 LEN=160 TOS=0x00 PREC=0x00 TTL=64 ID=19709
PROTO=UDP SPT=4500 DPT=4500 LEN=140
INPUT IN=eth0 OUT= MAC=fe:fd:0a:00:00:02:36:ec:4f:25:dc:68:08:00
SRC=10.0.0.1 DST=10.0.0.2 LEN=160 TOS=0x00 PREC=0x00 TTL=64 ID=19709
PROTO=UDP SPT=4500 DPT=4500 LEN=140
PREROUTING IN=eth0 OUT= MAC=fe:fd:0a:00:00:02:36:ec:4f:25:dc:68:08:00
SRC=10.0.0.1 DST=10.0.0.2 LEN=84 TOS=0x00 PREC=0x00 TTL=64 ID=19708
PROTO=ICMP TYPE=0 CODE=0 ID=16899 SEQ=0
INPUT IN=eth0 OUT= MAC=fe:fd:0a:00:00:02:36:ec:4f:25:dc:68:08:00
SRC=10.0.0.1 DST=10.0.0.2 LEN=84 TOS=0x00 PREC=0x00 TTL=64 ID=19708
PROTO=ICMP TYPE=0 CODE=0 ID=16899 SEQ=0
64 bytes from 10.0.0.1: icmp_seq=0 ttl=64 time=25.9 ms
^ permalink raw reply
* Re: [PATCH 02/10]: [NETFILTER]: Defer fragmentation in ip_output when connection tracking is used
From: Patrick McHardy @ 2005-11-17 2:28 UTC (permalink / raw)
To: Herbert Xu
Cc: Kernel Netdev Mailing List, Netfilter Development Mailinglist,
Harald Welte
In-Reply-To: <20051115104425.GA31719@gondor.apana.org.au>
Herbert Xu wrote:
> On Fri, Nov 11, 2005 at 03:19:17AM +0000, Patrick McHardy wrote:
>
>>[NETFILTER]: Defer fragmentation in ip_output when connection tracking is used
>>
>>This allows to get rid of the okfn use in ip_refrag and save the useless
>>fragmentation/defragmentation step when NAT is used.
>
>
> I'm slightly uneasy about this change because for POST_ROUTING, the
> defragmentation occurs in the middle of the hook, NF_IP_PRI_NAT_SRC.
>
> This means that things like the mangle table currently sees the
> fragments as opposed to the whole packet. This patch will change
> that.
>
> Now I'm not saying that this is necessarily a bad thing. In fact,
> for all I know it might make more sense to do this. But we should
> consider the possible implications before embarking on it.
Good point. I would also prefer to have fragmentation occur after
POST_ROUTING in all cases. Looking at the in-tree targets, it means
loosing the ability to do a couple of things:
- CLASSIFY fragments differently
- MARK fragments differently
- DSCP/ECN/TOS mark fragments differently
- Change TTLs of fragments to differently values
None of them seems very important. The DSCP and ECN targets were
broken until not long ago without anyone noticing, the TTL target is
relatively new. So it comes down to loosing the ability to classify
fragments of one packet differently using iptables, which doesn't
make much sense too me. In fact I think it would make classification
easier if mangle would see the whole packet.
I've CCed Harald for his opinion in case I missed something.
Regards
Patrick
^ permalink raw reply
* [PATCH]IPv6: Acquire addrconf_hash_lock for reading instead of writing in addrconf_verify(...)
From: Yan Zheng @ 2005-11-17 2:14 UTC (permalink / raw)
To: netdev; +Cc: linux-kernel, yoshfuji
Hi.
addrconf_verify(...) only traverse address hash table when addrconf_hash_lock is held
for writing, and it may hold addrconf_hash_lock for a long time. So I think it's better
to acquire addrconf_hash_lock for reading instead of writing
Signed-off-by: Yan Zheng <yanzheng@21cn.com>
============================================================
--- a/net/ipv6/addrconf.c 2005-11-17 09:27:54.000000000 +0800
+++ b/net/ipv6/addrconf.c 2005-11-17 09:25:42.000000000 +0800
@@ -2627,7 +2627,7 @@ static void addrconf_verify(unsigned lon
for (i=0; i < IN6_ADDR_HSIZE; i++) {
restart:
- write_lock(&addrconf_hash_lock);
+ read_lock(&addrconf_hash_lock);
for (ifp=inet6_addr_lst[i]; ifp; ifp=ifp->lst_next) {
unsigned long age;
#ifdef CONFIG_IPV6_PRIVACY
@@ -2649,7 +2649,7 @@ restart:
if (age >= ifp->valid_lft) {
spin_unlock(&ifp->lock);
in6_ifa_hold(ifp);
- write_unlock(&addrconf_hash_lock);
+ read_unlock(&addrconf_hash_lock);
ipv6_del_addr(ifp);
goto restart;
} else if (age >= ifp->prefered_lft) {
@@ -2668,7 +2668,7 @@ restart:
if (deprecate) {
in6_ifa_hold(ifp);
- write_unlock(&addrconf_hash_lock);
+ read_unlock(&addrconf_hash_lock);
ipv6_ifa_notify(0, ifp);
in6_ifa_put(ifp);
@@ -2686,7 +2686,7 @@ restart:
in6_ifa_hold(ifp);
in6_ifa_hold(ifpub);
spin_unlock(&ifp->lock);
- write_unlock(&addrconf_hash_lock);
+ read_unlock(&addrconf_hash_lock);
ipv6_create_tempaddr(ifpub, ifp);
in6_ifa_put(ifpub);
in6_ifa_put(ifp);
@@ -2703,7 +2703,7 @@ restart:
spin_unlock(&ifp->lock);
}
}
- write_unlock(&addrconf_hash_lock);
+ read_unlock(&addrconf_hash_lock);
}
addr_chk_timer.expires = time_before(next, jiffies + HZ) ? jiffies + HZ : next;
^ permalink raw reply
* Re: 2.6.15-rc1: NET_CLS_U32 not working?
From: Christian @ 2005-11-17 0:06 UTC (permalink / raw)
To: linux-kernel; +Cc: netdev
In-Reply-To: <20051116235813.GS5735@stusta.de>
Adrian Bunk schrieb:
>
> I'm assuming you are trying to insert the new module in your old kernel?
yes, i tried to "modprobe" the compiled cls_u32 module. but the "make
modules" errors are there anyway. i tried to compile a fresh 2.6.15-rc1
on a different machine (where i can't do "modprobe") and the errors were
there too: http://nerdbynature.de/bits/sheep/2.6.15-rc1/make-modules.log
> This is one of the unfortunate but hardly avoidable cases where adding a
> module requires installing a new kernel.
despite of the errors on "make modules" i'll reboot with the "new"
kernel asap.
> BTW: Please Cc netdev@vger.kernel.org on networking issues.
ok, will do that.
thank you,
Christian.
--
BOFH excuse #442:
Trojan horse ran out of hay
^ permalink raw reply
* Re: 2.6.15-rc1: NET_CLS_U32 not working?
From: Adrian Bunk @ 2005-11-16 23:58 UTC (permalink / raw)
To: Christian; +Cc: linux-kernel, netdev
In-Reply-To: <437BBC59.70301@g-house.de>
On Thu, Nov 17, 2005 at 12:10:17AM +0100, Christian wrote:
> hi,
>
> i noticed that some of my QoS rules are not working any more. oh, i
> forgot to enable CONFIG_NET_CLS_U32. but when enabled, i got the
> following errors when compiling / installing the module:
>
> * Warning: "unregister_tcf_proto_ops" [net/sched/cls_u32.ko] undefined!
> * Warning: "register_tcf_proto_ops" [net/sched/cls_u32.ko] undefined!
> * Warning: "tcf_exts_dump" [net/sched/cls_u32.ko] undefined!
> * Warning: "tcf_exts_dump_stats" [net/sched/cls_u32.ko] undefined!
> * Warning: "tcf_exts_change" [net/sched/cls_u32.ko] undefined!
> * Warning: "tcf_exts_validate" [net/sched/cls_u32.ko] undefined!
> * Warning: "tcf_exts_destroy" [net/sched/cls_u32.ko] undefined!
>...
> when i disabled CONFIG_NET_CLS_U32, everything compiles fine, but
> cls_u32 is missing of course :-(
>
> all the missing symbols seem to be defined in include/net/pkt_cls.h. but
> this file is #included by net/sched/cls_u32.c and other too, so i
> don't really know, why it doesn't work.
>
> FWIW, i see EXPORT_SYMBOLs at the very end of net/sched/cls_api.c, but i
> can't see if/when cls_api.c is used (included?) at all.
>...
I'm assuming you are trying to insert the new module in your old kernel?
This is one of the unfortunate but hardly avoidable cases where adding a
module requires installing a new kernel.
But there's a change in 2.6.15-rc1 that makes this issue much worse:
It is no longer user-visible.
tristate's select'ing bool's that do not change parts of the (modular)
driver but compile additional code into the kernel are simply wrong.
> thanks for looking into that,
> Christian.
>...
cu
Adrian
BTW: Please Cc netdev@vger.kernel.org on networking issues.
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply
* Re: [PATCH] [IPV4] Fix secondary IP addresses after promotion
From: Brian Pomerantz @ 2005-11-16 19:21 UTC (permalink / raw)
To: Patrick McHardy
Cc: Thomas Graf, netdev, davem, kuznet, pekkas, jmorris, yoshfuji,
kaber, linux-kernel
In-Reply-To: <4370B203.8070501@trash.net>
On Tue, Nov 08, 2005 at 03:11:15PM +0100, Patrick McHardy wrote:
>
> Yes, fixing it correctly looks very hard. Just changing the routes
> doesn't seem right to me, someone might have added it with exactly
> this prefsrc and doesn't want it to change, its also not clear how
> to notify on this. Taking care of correct ordering of the ifa_list
> is also more complicated without just deleting and readding them.
>
> I have a patch to do this, but it needs some debugging, for some
> unknown reason it crashes sometimes if I remove addresses without
> specifying the mask.
Looks like I'm back on this one because just sending the NETDEV_UP for
the secondaries didn't work if a primary other than the first one is
removed. If you have anything that you need help testing/debugging,
I'm stuck with this until it is fixed. I'd prefer not to duplicate
effort on this if you're close to a fix. If not, then I'll try to
come up with something and toss it out for comment.
BAPper
^ permalink raw reply
* kernel-2.6.14-1.1637_FC4.netdev.2 now available
From: John W. Linville @ 2005-11-16 14:33 UTC (permalink / raw)
To: fedora-list, fedora-devel-list, netdev
In-Reply-To: <20051114205110.GK25755@redhat.com>
The second FC4 fedora-netdev kernel is now available. If you are
already a fedora-netdev user, a simple 'yum update' should retrieve
the new kernels for you.
I now have a fedora-netdev-release package available. This simplifies
the process of configuring yum for the fedora-netdev repository.
The FC4 fedora-netdev-release package is available here:
http://people.redhat.com/linville/kernels/fedora-netdev/4/fedora-netdev-release-4-0.noarch.rpm
Download and install that, then you will be able to use yum to get
the FC4 fedora-netdev kernels.
Thanks, and good luck! :-)
John
P.S. What is fedora-netdev?
The purpose of this repository is two-fold: 1) to make bleeding-edge
linux kernel networking developments available to Fedora users who
need or want access to them; and, 2) to open-up the Fedora user
base as a better testing resource for the kernel netdev community.
I hope this will prove to be a win-win situation for both camps.
If you are a Fedora user with an interest or need for the latest
developments in Linux kernel networking, then _please_ try the
kernels from this repository. Your testing and feedback is greatly
appreciated, desperately requested, and graciously accepted.
Thanks in advance!
P.P.S. What netdev patches are in it?
- sky2: new experimental Marvell Yukon2 driver
- 8139cp: support ETHTOOL_GPERMADDR
- 8139too: support ETHTOOL_GPERMADDR
- b44: support ETHTOOL_GPERMADDR
- e1000: support ETHTOOL_GPERMADDR
- e100: support ETHTOOL_GPERMADDR
- forcedeth: support ETHTOOL_GPERMADDR
- ixgb: support ETHTOOL_GPERMADDR
- ne2k-pci: support ETHTOOL_GPERMADDR
- pcnet32: support ETHTOOL_GPERMADDR
- r8169: support ETHTOOL_GPERMADDR
- skge: support ETHTOOL_GPERMADDR
- sundance: support ETHTOOL_GPERMADDR
- via-rhine: support ETHTOOL_GPERMADDR
- drivers/net: fix-up schedule_timeout() usage
- Replace drivers/net/wan custom ctype macros with standard ones
- drivers/net/wan/: possible cleanups
- lne390 bogus casts
- C99 initializers in ray_cs.c
- mii: Add test for GigE support
- Add rapidio net driver
- pcnet32: set_ringparam implementation
- pcnet32: set min ring size to 4
- sky2: driver update.
- orinoco: Remove conditionals that are useless in the kernel drivers.
- orinoco: Don't include <net/ieee80211.h> twice.
- orinoco: Update PCMCIA ID's.
- Fixed some endian issues with 802.11 header usage in ieee80211_rx.c
- ieee80211 quality scaling algorithm extension handler
- ieee80211 Added wireless spy support
- Changed 802.11 headers to use ieee80211_info_element[0]
- ieee80211 Removed ieee80211_info_element_hdr
- ieee80211 Cleanup memcpy parameters.
- ieee80211 Switched to sscanf in store_debug_level
- ieee80211 Fixed type-o of abg_ture -> abg_true
- Updated ipw2200 to compile with ieee80211 abg_ture to abg_true change
- sky2: fix FIFO DMA alignment problems
- sky2: allow ethtool debug access to all of PCI space
- sky2: version 0.5
- ieee80211: Updated ipw2100 to be compatible with ieee80211_hdr changes
- ieee80211: Updated ipw2100 to be compatible with ieee80211's hard_start_xmit change
- ieee80211: Updated ipw2200 to be compatible with ieee80211_hdr changes
- ieee80211: Updated ipw2200 to be compatible with ieee80211's hard_start_xmit change.
- ieee80211: Updated atmel to be compatible with ieee80211_hdr changes
- ieee80211: Fixed a kernel oops on module unload
- ieee80211: Hardware crypto and fragmentation offload support
- ieee80211: Fix time calculation, switching to use jiffies_to_msecs
- ieee80211: Fix kernel Oops when module unload
- ieee80211: Allow drivers to fix an issue when using wpa_supplicant with WEP
- ieee82011: Added WE-18 support to default wireless extension handler
- ieee80211: Renamed ieee80211_hdr to ieee80211_hdr_3addr
- ieee80211: adds support for the creation of RTS packets
- ieee82011: Added ieee80211_tx_frame to convert generic 802.11 data frames, and callbacks
- ieee80211: Fix TKIP, repeated fragmentation problem, and payload_size reporting
- ieee80211: Return NETDEV_TX_BUSY when QoS buffer full
- ieee80211: Add QoS (WME) support to the ieee80211 subsystem
- ieee80211: Added ieee80211_geo to provide helper functions
- ieee80211: Added ieee80211_radiotap.h
- ieee80211: Additional fixes for endian-aware types
- ieee80211: "extern inline" to "static inline"
- ieee80211: Type-o, capbility definition for QoS, and ERP parsing
- ieee80211: Mixed PTK/GTK CCMP/TKIP support
- ieee80211: Keep auth mode unchanged after iwconfig key off/on cycle
- ieee80211: Updated copyright dates
- ieee80211: Updated hostap to be compatible with ieee80211_hdr changes
- ieee80211: Updated hostap to be compatible with extra_prefix_len changes
- ieee82011: Remove WIRELESS_EXT ifdefs
- forcedeth: add hardware tx checksumming
- ieee80211: Added subsystem version string and reporting via MODULE_VERSION
- ieee80211: Added handle_deauth() callback, enhanced tkip/ccmp support of varying hw/sw offload
- ieee80211: added IE comments, reason_code to reason, removed info_element from ieee80211_disassoc
- ieee80211: in-tree driver updates to sync with latest ieee80211 series
- ieee80211: update orinoco, wl3501 drivers for latest struct naming
- orinoco: Remove inneeded system includes.
- orinoco: Make nortel_pci_hw_init() static.
- orinoco: Fix memory leak and unneeded unlock in orinoco_join_ap()
- orinoco: orinoco_send_wevents() could return without unlocking.
- orinoco: Remove unneeded forward declarations.
- orinoco: Annotate endianess of variables and structure members.
- orinoco: Read only needed data in __orinoco_ev_txexc().
- orinoco: Bump version to 0.15rc3.
- RPC: Report connection errors properly when mounting with "soft"
- RPC: proper soft timeout behavior for rpcbind
- NFS: use a constant value for TCP retransmit timeouts
- RPC: portmapper doesn't need a reserved port
- RPC: extract socket logic common to both client and server
- RPC: introduce client-side transport switch
- RPC: transport switch function naming
- RPC: Reduce stack utilization in xs_sendpages
- RPC: Rename sock_lock
- RPC: Rename xprt_lock
- RPC: rename the sockstate field
- RPC: Eliminate socket.h includes in RPC client
- RPC: Add helper for waking tasks pending on a transport
- RPC: client-side transport switch cleanup
- RPC: separate TCP and UDP write space callbacks
- RPC: separate TCP and UDP transport connection logic
- RPC: separate TCP and UDP socket write paths
- RPC: skip over transport-specific heads automatically
- RPC: get rid of xprt->stream
- RPC: add API to set transport-specific timeouts
- RPC: expose API for serializing access to RPC transports
- RPC: expose API for serializing access to RPC transports
- RPC: separate xprt_timer implementations
- RPC: add generic interface for adjusting the congestion window
- RPC: add a release_rqst callout to the RPC transport switch
- RPC: remove xprt->nocong
- RPC: clean up after nocong was removed
- RPC: allow RPC client's port range to be adjustable
- RPC: make sure to get the same local port number when reconnecting
- RPC: parametrize various transport connect timeouts
- RPC: rationalize set_buffer_size
- RPC,NFS: new rpc_pipefs patch
- Revert "[PATCH] RPC,NFS: new rpc_pipefs patch"
- SUNRPC: fix bug in patch "portmapper doesn't need a reserved port"
- [netdrvr gianfar] use new phy layer
- [netdrvr] delete CONFIG_PHYCONTROL
- hostap: Fix pci_driver name for hostap_plx and hostap_pci
- hostap: Add support for WE-19
- hostap: Use GFP_ATOMIC to get rid of weird might_sleep issue
- hostap: Remove iwe_stream_add_event kludge
- Remove WIRELESS_EXT ifdefs from several wireless drivers.
- [wireless airo] remove needed dma_addr_t obfuscation
- sky2: changing mtu doesn't have to reset link
- sky2: cleanup interrupt processing
- sky2: add hardware VLAN acceleration support
- sky2: explicit set power state
- sky2: version 0.6
- sky2: nway reset (BONUS FEATURE)
- This patch fixes a typo in ieee80211.h: ieee82011_deauth -> ieee80211_deauth
- This will move the ieee80211_is_ofdm_rate function to the ieee80211.h
- Currently the info_element is parsed by 2 seperate functions, this
- When an assoc_resp is received the network structure is not completely
- Lindent and trailing whitespace script executed ieee80211 subsystem
- hostap: Remove hw specific dev_open/close handlers
- hostap: Fix hostap_pci build with PRISM2_IO_DEBUG
- hostap: Do not free local->hw_priv before unregistering netdev
- hostap: Unregister netdevs before freeing local data
- S2io: MSI/MSI-X support (runtime configurable)
- e1000: Support for 82571 and 82572 controllers
- e1000: multi-queue defines/modification to data structures
- e1000: implementation of the multi-queue feature
- e1000: Enable custom configuration bits for 82571/2 controllers
- e1000: Fixes for packet split related issues
- e1000: Added msleep_interruptible delay
- e1000: Flush shadow RAM
- e1000: fix warnings
- AX.25: Delete debug printk from mkiss driver
- AX.25: Convert mkiss.c to DEFINE_RWLOCK
- airo: fix resume
- s2io: change strncpy length arg to use size of target
- [netdrvr s2io] Add a MODULE_VERSION entry
- bonding: replicate IGMP traffic in activebackup mode
- sky2: add permanent address support.
- [wireless ipw2200] remove redundant return statement
- S2io: Offline diagnostics fixes
- rcu in bpqether driver.
- SMACK support for mkiss
- Initialize the .owner field the tty_ldisc structure.
- SUNRPC: Retry rpcbind requests if the server's portmapper isn't up
- RPC: allow call_encode() to delay transmission of an RPC call.
- ieee80211: division by zero fix
- sb1250-mac: Check the actual setting for reporting hw checksumming.
- sb1250-mac: Ensure 16-byte alignment of the descriptor ring.
- au1000_eth: Misc Au1000 net driver fixes.
- de2104x: Resurrect Cobalt support for 2.6.
- sgiseeq: Fix resource handling.
- sgiseeq: Configure PIO and DMA timing requests.
- declance: Convert to irqreturn_t.
- declance: Fix mapping of device.
- declance: Deal with the bloody KSEG vs CKSEG horror...
- declance: Use physical addresses at the interface level.
- ne: Support for RBHMA4500 eval board.
- mipsnet: Virtual ethernet driver for MIPSsim.
- e1000_intr build fix
- s2io build fix
- via-rhine: change mdelay to msleep and remove from ISR path
- epic100: fix counting of work_done in epic_poll
- bonding: cleanup comment for mode 1 IGMP xmit hack
- b44: alternate allocation option for DMA descriptors
- orinoco: remove redundance skb length check before padding
- sundance: remove if (1) { ... } block in sundance_probe1
- sundance: expand reset mask
- e1000 build fix
- RPC: stops the release_pipe() funtion from being called twice
- SUNRPC: Add support for privacy to generic gss-api code.
- SUNRPC: Provide a callback to allow free pages allocated during xdr encoding
- SUNRPC: Retry wrap in case of memory allocation failure.
- RPCSEC_GSS: cleanup au_rslack calculation
- RPCSEC_GSS: client-side privacy support
- RPCSEC_GSS: Simplify rpcsec_gss crypto code
- RPCSEC_GSS: krb5 pre-privacy cleanup
- RPCSEC_GSS: Add support for privacy to krb5 rpcsec_gss mechanism.
- RPCSEC_GSS remove all qop parameters
- RPCSEC_GSS: krb5 cleanup
- Fixed problem with not being able to decrypt/encrypt broadcast packets.
- sb1250-mac: Get rid of all the funny SBMAC_WRITECSR and SBMAC_READCSR macros.
- sb1250-mac: Whitespace cleanup.
- sundance: include MII address 0 in PHY probe
- e1000: Driver version, white space, comments, device id & other
- Fixed oops if an uninitialized key is used for encryption.
- sb1250-mac: PHY probing fixes.
- ieee80211 subsystem:
- Update version ieee80211 stamp to 1.1.6
- [PARISC] Change the driver names so /sys/bus/parisc/drivers/ looks better
- [PARISC] Convert parisc_device to use struct resource for hpa
- [PARISC] Add NETPOLL support to lasi_82596
- [DECNET]: Remove some redundant ifdeffed code
- [NET]: Wider use of for_each_*cpu()
- [PKTGEN]: Sleeping function called under lock
- [PKTGEN]: Use kzalloc
- [PKTGEN]: Spelling and white space
- [PKTGEN]: proc interface revision
- [NETFILTER] ip_conntrack: Make "hashsize" conntrack parameter writable
- [IPV4]: Kill redundant rcu_dereference on fa_info
- [IPSEC]: Kill obsolete get_mss function
- [NETLINK]: Remove dead code in af_netlink.c
- [IPV4]: Remove dead code from ip_output.c
- [SK_BUFF] kernel-doc: fix skbuff warnings
- [AX.25]: Use constant instead of magic number
- [IPV4]: Fix setting broadcast for SIOCSIFNETMASK
- [netdrvr forcedeth] scatter gather and segmentation offload support
- ieee80211 build fix
- Revert "RPC: stops the release_pipe() funtion from being called twice"
- RPC: Ensure that nobody can queue up new upcalls after rpc_close_pipes()
- gfp_t: net/*
- gfp_t: drivers/net
- [ARM] 2919/1: CS8900A ethernet driver modifications for the Comdial MP1000
- [ARM] 2897/2: PXA2xx IRDA support
- sky2: remove unused definitions
- sky2: use kzalloc
- sky2: spelling fixes
- sky2: fix NAPI and receive handling
- sky2: version 0.7
- DRIVER MODEL: Get rid of the obsolete tri-level suspend/resume callbacks
- [Bluetooth] Make more functions static
- [Bluetooth] Update security filter for Extended Inquiry Response
- [IPv4/IPv6]: UFO Scatter-gather approach
- [MCAST] IPv6: Fix algorithm to compute Querier's Query Interval
- tg3: add 5714/5715 support
- tg3: fix ASF heartbeat
- tg3: update version and minor fixes
- ibmveth fix bonding
- ibmveth fix buffer pool management
- ibmveth fix buffer replenishing
- ibmveth lockless TX
- ibmveth fix failed addbuf
- pcnet_cs: fix mii init code for older DL10019 based cards
- s2io: kconfig help fix
- b44 reports wrong advertised flags
- sis190.c: fix multicast MAC filter
- smc91x: shut down power after probing
- starfire: free_irq() on error path of netdev_open()
- [netdrvr b44] include linux/dma-mapping.h to eliminate warning
- sundance: fix DFE-580TX Tx Underrun
- New PowerPC 4xx on-chip ethernet controller driver
- sis900: come alive after temporary memory shortage
- Add Wake on LAN support to sis900 (2)
- drivers/net: Remove pointless checks for NULL prior to calling kfree()
- [netdrvr] ne2k-pci based card does not support bus-mastering.
- ipw2200: Missing kmalloc check
- [SCTP] Rename SCTP specific control message flags.
- [SCTP] Fix SCTP_SETADAPTION sockopt to use the correct structure.
- [SCTP] Allow SCTP_MAXSEG to revert to default frag point with a '0' value.
- [SCTP] Do not allow unprivileged programs initiating new associations on
- e1000: remove warning about e1000_suspend
- eepro.c: module_param_array cleanup
- b44: fix suspend/resume
- e1000: use vmalloc_node()
- revert "orinoco: Information leakage due to incorrect padding"
- Better fixup for the orinoco driver
- e1000: Fixes e1000_suspend warning when CONFIG_PM is not enabled
- [ETH]: ether address compare
- Add modalias for pmac network drivers
- mv643xx_eth_showsram: Added information message when using the SRAM
- [IPV4]: Fix issue reported by Coverity in ipv4/fib_frontend.c
- s2io iomem annotations
- bluetooth hidp is broken on s390
- drivers/net/tg3: Use the DMA_{32,64}BIT_MASK constants
- prism54: Free skb after disabling interrupts
- [DRIVER MODEL] Add missing platform_device.h header.
- PPC 44x EMAC driver: add 440SPe support
- PPC 44x EMAC driver: add 440GR support
- PPC 4xx EMAC driver: fix VSC8201 PHY initialization
- fec_8xx: Remove dependency on NETTA & NETPHONE
- fec_8xx: Add support for Intel PHY LX971
- vmalloc_node
- [ARM] 3066/1: Fix PXA irda driver suspend/resume functions
- m32r: SMC91x driver update
- smsc-ircc2: PM cleanup - do not close device when suspending
- remove some more check_region stuff
- Typo fix: dot after newline in printk strings
- sparse cleanups: NULL pointers, C99 struct init.
- [netdrvr 8139too] replace hand-crafted kernel thread with workqueue
- [BRIDGE]: Use ether_compare
- [NETFILTER]: Add "revision" support to arp_tables and ip6_tables
- [ROSE]: rose_heartbeat_expiry() locking fix
- [IPV6]: Fix behavior of ip6_route_input() for link local address
- [DCCP]: Simplify skb_set_owner_w semantics
- [DCCP]: Set socket owner iff packet is not data
- [MCAST] IPv6: Check packet size when process Multicast
- ibmveth fix panic in initial replenish cycle
- [MCAST]: ip[6]_mc_add_src should be called when number of sources is zero
- [IPV6]: inet6_ifinfo_notify should use RTM_DELLINK in addrconf_ifdown
- [PKT_SCHED]: Rework QoS and/or fair queueing configuration
- ARM: Reverted 2919/1: CS8900A ethernet driver modifications for the Comdial MP1000
- SUNRPC: allow sunrpc.o to link when CONFIG_SYSCTL is disabled
- NFS,SUNRPC,NLM: fix unused variable warnings when CONFIG_SYSCTL is disabled
- [NETFILTER] PPTP helper: Fix compilation of conntrack helper without NAT
- [NETFILTER] PPTP helper: Fix endianness bug in GRE key / CallID NAT
- [NETFILTER] NAT: Fix module refcount dropping too far
- [netdrvr 8139too] use cancel_rearming_delayed_work() to cancel thread
- [netdrvr 8139too] use rtnl_shlock_nowait() rather than rtnl_lock_interruptible()
- [NETFILTER]: Fix double free after netlink_unicast() in ctnetlink
- [NETFILTER] nfnetlink: Use kzalloc
- [NETFILTER]: CONNMARK target needs ip_conntrack
- [NETFILTER] nf_queue: Fix Ooops when no queue handler registered
- [NETEM]: use PSCHED_LESS
- drivers/net/wireless/airo.c unsigned comparason
- S2io: Multi buffer mode support
- pcnet32: show name of failing device
- pcnet32: AT2700/2701 and Bugzilla 2699 & 4551
- pcnet32: Prevent hang with 79c976
- phy address mask support for generic phy layer
- [PKT_SCHED]: Generic RED layer
- [NET]: Introduce INET_ECN_set_ce() function
- [PKT_SCHED]: RED: Use new generic red interface
- [PKT_SCHED]: RED: Use generic queue management interface
- [PKT_SCHED]: RED: Dont start idle periods while already idling
- [PKT_SCHED]: RED: Cleanup and remove unnecessary code
- [PKT_SCHED]: GRED: Cleanup equalize flag and add new WRED mode detection
- [PKT_SCHED]: GRED: Transform grio to GRED_RIO_MODE
- [PKT_SCHED]: GRED: Cleanup dumping
- [PKT_SCHED]: GRED: Dump table definition
- [PKT_SCHED]: GRED: Use a central table definition change procedure
- [PKT_SCHED]: GRED: Report out-of-bound DPs as illegal
- [PKT_SCHED]: GRED: Use central VQ change procedure
- [PKT_SCHED]: GRED: Use new generic red interface
- [PKT_SCHED]: GRED: Do not reset statistics in gred_reset/gred_change
- [PKT_SCHED]: GRED: Report congestion related drops as NET_XMIT_CN
- [PKT_SCHED]: GRED: Use generic queue management interface
- [PKT_SCHED]: GRED: Introduce tc_index_to_dp()
- [PKT_SCHED]: GRED: Improve error handling and messages
- [PKT_SCHED]: GRED: Remove initd flag
- [PKT_SCHED]: GRED: Dont abuse default VQ for equalizing
- [PKT_SCHED]: GRED: Remove auto-creation of default VQ
- [PKT_SCHED]: GRED: Cleanup and remove unnecessary code
- [PKT_SCHED]: GRED: Fix restart of idle period in WRED mode upon dequeue and drop
- [PKT_SCHED]: GRED: Support ECN marking
- [PKT_SCHED]: (G)RED: Introduce hard dropping
- [DRIVER MODEL] Improved dynamically allocated platform_device interface
- [DRIVER MODEL] Fix depca
- [DRIVER MODEL] Fix jazzsonic
- [DRIVER MODEL] Fix macsonic
- [NETEM]: Support time based reordering
- [NETEM]: Add version string
- [NET]: Fix race condition in sk_stream_wait_connect
- [TCP/DCCP]: Randomize port selection
- drivers/net/ixgb/: make some code static
- drivers/net/e1000/: possible cleanups
- drivers/net/hamradio/dmascc.c: remove dmascc_setup()
- prism54: Remove redundant assignment
- bnx2: add 5708 support
- bnx2: update firmware for 5708
- bnx2: update nvram code for 5708
- bnx2: update firmware handshake for 5708
- bnx2: refine bnx2_poll
- bnx2: update version and minor fixes
- Remove linux/version.h include from drivers/net/phy/* and net/ieee80211/*.
- [netdrvr] fac_8xx build fix
- [netdrvr s2io] warning fixes
- b44: b44_start_xmit returns with a lock held when it fails allocating
- b44: miscellaneous cleanup
- b44: expose counters through ethtool
- b44: s/spin_lock_irqsave/spin_lock/ in b44_interrupt
- b44: late request_irq in b44_open
- 3c59x: convert to use of pci_iomap API
- 3c59x: cleanup of mdio_read routines to use MII_* macros
- 3c59x: avoid blindly reading link status twice
- 3c59x: bounds checking for hw_checksums
- 3c59x: cleanup init of module parameter arrays
- 3c59x: fix some grammar in module parameter descriptions
- 3c59x: support ETHTOOL_GPERMADDR
- 3c59x: correct rx_dropped counting
- 3c59x: enable use of memory-mapped PCI I/O
- 3c59x: don't enable scatter/gather w/o checksum support
- knfsd: make sure svc_process call the correct pg_authenticate for multi-service port
- kernel-doc: fix warnings in vmalloc.c
- m68knommu: FEC ethernet header support for the ColdFire 5208
- m68knommu: FEC ethernet support for the ColdFire 5208
- scripts/Lindent on ieee80211 subsystem.
- Fix problem with WEP unicast key > index 0
- Update version ieee80211 stamp to 1.1.7
- Ran scripts/Lindent on drivers/net/wireless/ipw2{1,2}00.{c,h}
- IPW_DEBUG has already included DRV_NAME, remove double prefix print.
- Catch ipw2200 up to equivelancy with v1.0.1
- Catch ipw2200 up to equivelancy with v1.0.2
- Catch ipw2200 up to equivelancy with v1.0.3
- Catch ipw2200 up to equivelancy with v1.0.4
- Catch ipw2100 up to equivelancy with v1.1.1
- Fixed WEP on ipw2100 (priv->sec was being used instead of
- [Bug 339] Fix ipw2100 iwconfig set/get txpower.
- Move code from ipw2100_wpa_enable to IPW2100_PARAM_DROP_UNENCRYPTED to
- Catch ipw2200 up to equivelancy with v1.0.5
- Fix hardware encryption (both WEP and AES) doesn't work with fragmentation.
- Fix is_duplicate_packet() bug for fragmentation number setting.
- [bug 667] Fix the notorious "No space for Tx" bug.
- [Bug 637] Set tx power for A band.
- Changed default # of missed beacons to miss before disassociation to 24
- Updated to support ieee80211 callback to is_queue_full for 802.11e
- Fixed some compiler issues if CONFIG_IPW2200_QOS is enabled.
- Added more useful geography encoding so people's experience with
- Workaround kernel BUG_ON panic caused by unexpected duplicate packets.
- Disable host fragmentation in open mode since IPW2200/2915 hardware
- [Bug 792] Fix WPA-PSK AES both for -Dipw and -Dwext.
- Fixes the ad-hoc network WEP key list issue.
- [Bug 701] Fix a misuse of ieee->mode with ieee->iw_mode.
- Fix ipw_wx_get_txpow shows wrong disabled value.
- Fix firmware error when setting tx_power.
- Modified ipw_config and STATUS_INIT setting to correct race condition
- Switched firmware error dumping so that it will capture a log available
- Changed all of the ipw_send_cmd() calls to return any ipw_send_cmd error
- Added cmdlog in non-debug systems.
- Migrated some of the channel verification code back into the driver to
- Updated ipw2200 to use the new ieee80211 callbacks
- Added wait_state wakeup on scan completion.
- [Bug 455] Fix frequent channel change generates firmware fatal error.
- [Bug 760] Fix setting WEP key in monitor mode causes IV lost.
- Don't set hardware WEP if we are actually using TKIP/AES.
- Make all the places the firmware fails to load showerrors (in decimal,
- Adds radiotap support to ipw2200 in monitor mode..
- Fixed is_network_packet() to include checking for broadcast packets.
- Mixed PTK/GTK CCMP/TKIP support.
- Card with WEP enabled and using shared-key auth will have firmware
- Fixed problem with get_cmd_string not existing if CONFIG_IPW_DEBUG disabled.
- Removed PF_SYNCTHREAD legacy.
- Fixes problem with WEP not working (association succeeds, but no Tx/Rx)
- [Fix bug# 771] Too many (8) bytes recieved when using AES/hwcrypto
- Fixes WEP firmware error condition.
- Updated driver version stamps for ipw2100 (1.1.3) and ipw2200 (1.0.7)
- Pulled out a stray KERNEL_VERSION check around the suspend handler.
- Fix 'Driver using old /proc/net/wireless support, please fix driver !' message.
- Removed legacy WIRELESS_EXT checks from ipw2200.c
- Fixes missed beacon logic in relation to on-network AP roaming.
- Removed warning about TKIP not being configured if countermeasures are
- Added channel support for ipw2200 cards identified as 'ZZR'
- Fixed problem with not being able to send broadcast packets.
- Fixed parameter reordering in firmware log routine.
- Updated firmware version stamp to 2.4 from 2.3 so it will use the latest firmware.
- Update version ipw2200 stamp to 1.0.8
- fix NET_RADIO=n, IEEE80211=y compile
- bonding: fix feature consolidation
- kill include/linux/eeprom.h
- drivers/net/s2io.c: make functions static
- prism54 : Unused variable / extraneous udelay
- prism54 : Transmit stats updated in wrong place
- Fix sparse warning in e100 driver.
- atmel: memset correct range
- [IPV6]: Put addr_diff() into common header for future use.
- [IPV6]: Make ipv6_addr_type() more generic so that we can use it for source address selection.
- [IPV6]: RFC3484 compliant source address selection
- [PKT_SCHED]: Correctly handle empty ematch trees
- [NET]: sk_add_backlog convert from macro to inline
- [PPP]: handle misaligned accesses
- [PPP]: add PPP MPPE encryption module
- [IRDA] donauboe: locking fix
- [NET]: kfree cleanup
- [IPV4]: Fix ip_queue_xmit identity increment for TSO packets
- [Bluetooth]: Add endian annotations to the core
- [Bluetooth]: Remove the usage of /proc completely
- [SERIAL] IOC3: Update 8250 driver bits
- skge: clear PCI PHY COMA mode on boot
- skge: use kzalloc
- skge: add mii ioctl support
- skge: goto low power mode on shutdown
- skge: use prefetch on receive
- skge: spelling fixes
- skge: increase version number
- [wireless ipw2100] kill unused-var warnings for debug-disabled code
- ieee80211: cleanup crypto list handling, other minor cleanups.
- b44: replace B44_FLAG_INIT_COMPLETE with netif_running()
- b44: race on device closing
- b44: increase version number
- dgrs: fix warnings when CONFIG_ISA and CONFIG_PCI are not enabled
- IOC: And don't mark the things as broken Cowboy.
- add a vfs_permission helper
- sanitize lookup_hash prototype
- [NETFILTER]: packet counter of conntrack is 32bits
- [NETFILTER]: refcount leak of proto when ctnetlink dumping tuple
- [NETFILTER] nfnetlink: nfattr_parse() can never fail, make it void
- [NETFILTER] ctnetlink: check if protoinfo is present
- [NETFILTER] ctnetlink: add marking support from userspace
- [NETFILTER] ctnetlink: add module alias to fix autoloading
- [NETFILTER] ctnetlink: kill unused includes
- [NETFILTER] ctnetlink: get_conntrack can use GFP_KERNEL
- [NETFILTER] PPTP helper: fix PNS-PAC expectation call id
- [NETFILTER] nfnetlink: only load subsystems if CAP_NET_ADMIN is set
- [NETFILTER]: stop tracking ICMP error at early point
- [NETFILTER] ctnetlink: return -EINVAL if size is wrong
- [NETFILTER] ctnetlink: propagate error instaed of returning -EPERM
- [NETFILTER] ctnetlink: Add support to identify expectations by ID's
- [NETFILTER] ctnetlink: Fix oops when no ICMP ID info in message
- [NETFILTER] ctnetlink: ICMP_ID is u_int16_t not u_int8_t.
- [IPV6]: Fix fallout from CONFIG_IPV6_PRIVACY
- [IPV6]: ip6ip6_lock is not unlocked in error path.
- [NETFILTER]: Add nf_conntrack subsystem.
- [NETLINK]: Type-safe netlink messages/attributes interface
- [NETLINK]: Make netlink_callback->done() optional
- [NETLINK]: Generic netlink receive queue processor
- [XFRM]: Use generic netlink receive queue processor
- [RTNETLINK]: Use generic netlink receive queue processor
- [NETLINK]: Generic netlink family
- SUNRPC: don't reencode when looping in call transmit.
- [netdrvr 8139too] fast poll for thread, if an unlikely race occurs
- [BNX2]: output driver name as prefix in error message
- [BNX2]: check return of dev_alloc_skb in bnx2_test_loopback
- [BNX2]: simplify parameter checks in bnx2_{get,set}_eeprom
- [NET]: Detect hardware rx checksum faults correctly
- [TCP]: fix congestion window update when using TSO deferal
- [TCP]: simplify microsecond rtt sampling
- [TCP]: add tcp_slow_start helper
- [TCP]: Appropriate Byte Count support
- [TCP]: receive buffer growth limiting with mixed MTU
- [TCP]: spelling fixes
- [TCP]: speed up SACK processing
- disable DEBUG in ibmveth
- sky2 needs dma_mapping.h
- gianfar mii needs to zero out the mii_bus structure
- [netdrvr forcedeth] remove superfluous rx engine stop/start
- [netdrvr forcedeth] support for irq mitigation
- [netdrvr forcedeth] phy address scan range
- SAA9730: Whitespace cleanup.
- SAA9730: Driver overhaul
- smc91x: DB1200 support.
- gt96100eth.c: Don't concatenate __FUNCTION__ with strings.
- TCP: fix vegas build
- [DECNET]: fix SIGPIPE
- [IPV6]: Fix inet6_init missing unregister.
- [SCTP]: Fix potential NULL pointer dereference in sctp_v4_get_saddr
- [SCTP]: Remove timeouts[] array from sctp_endpoint.
- [SCTP]: Fix ia64 NaT consumption fault with sctp_sideffect commands.
- [SCTP]: Include ulpevents in socket receive buffer accounting.
--
John W. Linville
linville@redhat.com
--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
^ permalink raw reply
* Re: fedora-netdev.1 IPv6 freeze [Re: [ANNOUNCE] fedora-netdev kernel repository]
From: John W. Linville @ 2005-11-16 14:25 UTC (permalink / raw)
To: Thomas Graf
Cc: Development discussions related to Fedora Core, netdev,
Pekka Savola, linville
In-Reply-To: <20051116114224.GB20395@postel.suug.ch>
On Wed, Nov 16, 2005 at 12:42:24PM +0100, Thomas Graf wrote:
> * Pekka Savola <pekkas@netcore.fi> 2005-11-16 12:46
> > On Mon, 14 Nov 2005, John W. Linville wrote:
> > > http://people.redhat.com/linville/kernels/fedora-netdev/
> >
> > I guess the test can be termed a 'success' because after updating from
> > 2.6.14-1.1637_FC4 to 2.6.14-1.1637_FC4.netdev.1, I get 100%
> > reproducible kernel hang (everything just freezes as it is, no message
> > to /var/log/messages or anywhere) after I run '/sbin/ip -6 r l' or try
> > to use IPv6 in basically any other way on my ThinkPad laptop with
> > external orinoco_cs WLAN card.
> >
> > Any thoughts for the next steps?
>
> It's probably missing this patch:
It was...
Pekka, I have included the patch Thomas identified as part of the
FC4.netdev.2 build. You may want to do a 'yum update' and try it out.
Thanks!
John
--
John W. Linville
linville@redhat.com
--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
^ permalink raw reply
* fedora-netdev.1 IPv6 freeze [Re: [ANNOUNCE] fedora-netdev kernel repository]
From: Pekka Savola @ 2005-11-16 10:46 UTC (permalink / raw)
To: linville, linville,
Development discussions related to Fedora Core; +Cc: netdev
In-Reply-To: <20051114205110.GK25755@redhat.com>
On Mon, 14 Nov 2005, John W. Linville wrote:
> http://people.redhat.com/linville/kernels/fedora-netdev/
I guess the test can be termed a 'success' because after updating from
2.6.14-1.1637_FC4 to 2.6.14-1.1637_FC4.netdev.1, I get 100%
reproducible kernel hang (everything just freezes as it is, no message
to /var/log/messages or anywhere) after I run '/sbin/ip -6 r l' or try
to use IPv6 in basically any other way on my ThinkPad laptop with
external orinoco_cs WLAN card.
Any thoughts for the next steps?
--
Pekka Savola "You each name yourselves king, yet the
Netcore Oy kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox