* Re: [net-next-2.6 V7 PATCH 1/2] Add netlink support for virtual port management (was iovnl)
From: Patrick McHardy @ 2010-05-14 10:58 UTC (permalink / raw)
To: Scott Feldman; +Cc: davem, netdev, chrisw, arnd
In-Reply-To: <20100514013526.1816.45104.stgit@savbu-pc100.cisco.com>
Scott Feldman wrote:
> --- a/net/core/rtnetlink.c
> +++ b/net/core/rtnetlink.c
> @@ -653,6 +653,26 @@ static inline int rtnl_vfinfo_size(const struct net_device *dev)
> return 0;
> }
>
> +static size_t rtnl_vf_port_size(const struct net_device *dev)
> +{
> + size_t vf_port_size = nla_total_size(sizeof(struct nlattr))
> + /* VF_PORT_VF */
> + + nla_total_size(VF_PORT_PROFILE_MAX)/* VF_PORT_PROFILE */
> + + nla_total_size(sizeof(struct ifla_vf_port_vsi))
> + /* VF_PORT_VSI_TYPE */
> + + nla_total_size(VF_PORT_UUID_MAX) /* VF_PORT_VSI_INSTANCE */
> + + nla_total_size(VF_PORT_UUID_MAX) /* VF_PORT_HOST_UUID */
> + + nla_total_size(1) /* VF_PROT_VDP_REQUEST */
Do messages generated by the kernel really contain a request?
> + + nla_total_size(2); /* VF_PORT_VDP_RESPONSE */
> +
> + if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
> + return 0;
> + if (dev_num_vf(dev->dev.parent))
> + return vf_port_size * dev_num_vf(dev->dev.parent);
> + else
> + return vf_port_size;
> +}
> +
> +static int rtnl_vf_port_fill_nest(struct sk_buff *skb, struct net_device *dev,
> + int vf)
> +{
> + struct nlattr *data;
> + int err;
> +
> + data = nla_nest_start(skb, IFLA_VF_PORT);
We usually use a top-level attribute to encapsulate lists of identical
attributes. The other iflink attributes may only occur once and are
usually parsed using nla_parse_nested(), which will parse all
IFLA_VF_PORT attributes, but only return the last one.
Something like:
iflink message:
...
[IFLA_VF_PORTS]
[IFLA_VF_PORT]
[IFLA_VF_PORT_*], ...
[IFLA_VF_PORT]
[IFLA_VF_PORT_*], ...
...
> + if (!data)
> + return -EMSGSIZE;
> +
> + if (vf != VF_PORT_VF_NOT_USED)
> + nla_put_u32(skb, IFLA_VF_PORT_VF, vf);
This should be checking for errors or use NLA_PUT_U32.
> +
> + err = dev->netdev_ops->ndo_get_vf_port(dev, vf, skb);
> + if (err) {
> + nla_nest_cancel(skb, data);
> + return err;
> + }
> +
> + nla_nest_end(skb, data);
> +
> + return 0;
> +}
> +
> static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
> int type, u32 pid, u32 seq, u32 change,
> unsigned int flags)
> @@ -747,17 +819,23 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
> goto nla_put_failure;
> copy_rtnl_link_stats64(nla_data(attr), stats);
>
> + if (dev->dev.parent)
> + NLA_PUT_U32(skb, IFLA_NUM_VF, dev_num_vf(dev->dev.parent));
Just wondering, is the only case where dev.parent is non-NULL
really when virtual ports are present?
> +
> if (dev->netdev_ops->ndo_get_vf_config && dev->dev.parent) {
> int i;
> struct ifla_vf_info ivi;
>
> - NLA_PUT_U32(skb, IFLA_NUM_VF, dev_num_vf(dev->dev.parent));
> for (i = 0; i < dev_num_vf(dev->dev.parent); i++) {
> if (dev->netdev_ops->ndo_get_vf_config(dev, i, &ivi))
> break;
> NLA_PUT(skb, IFLA_VFINFO, sizeof(ivi), &ivi);
> }
> }
> +
> + if (rtnl_vf_port_fill(skb, dev))
> + goto nla_put_failure;
> +
> if (dev->rtnl_link_ops) {
> if (rtnl_link_fill(skb, dev) < 0)
> goto nla_put_failure;
> @@ -824,6 +902,7 @@ const struct nla_policy ifla_policy[IFLA_MAX+1] = {
> .len = sizeof(struct ifla_vf_vlan) },
> [IFLA_VF_TX_RATE] = { .type = NLA_BINARY,
> .len = sizeof(struct ifla_vf_tx_rate) },
> + [IFLA_VF_PORT] = { .type = NLA_NESTED },
> };
> EXPORT_SYMBOL(ifla_policy);
>
> @@ -832,6 +911,20 @@ static const struct nla_policy ifla_info_policy[IFLA_INFO_MAX+1] = {
> [IFLA_INFO_DATA] = { .type = NLA_NESTED },
> };
>
> +static const struct nla_policy ifla_vf_port_policy[IFLA_VF_PORT_MAX+1] = {
> + [IFLA_VF_PORT_VF] = { .type = NLA_U32 },
> + [IFLA_VF_PORT_PROFILE] = { .type = NLA_STRING,
> + .len = VF_PORT_PROFILE_MAX },
> + [IFLA_VF_PORT_VSI_TYPE] = { .type = NLA_BINARY,
> + .len = sizeof(struct ifla_vf_port_vsi)},
> + [IFLA_VF_PORT_INSTANCE_UUID]= { .type = NLA_BINARY,
> + .len = VF_PORT_UUID_MAX },
> + [IFLA_VF_PORT_HOST_UUID] = { .type = NLA_STRING,
> + .len = VF_PORT_UUID_MAX },
> + [IFLA_VF_PORT_REQUEST] = { .type = NLA_U8, },
> + [IFLA_VF_PORT_RESPONSE] = { .type = NLA_U16, },
> +};
> +
> struct net *rtnl_link_get_net(struct net *src_net, struct nlattr *tb[])
> {
> struct net *net;
> @@ -1028,6 +1121,27 @@ static int do_setlink(struct net_device *dev, struct ifinfomsg *ifm,
> }
> err = 0;
>
> + if (tb[IFLA_VF_PORT]) {
> + struct nlattr *vf_port[IFLA_VF_PORT_MAX+1];
> + int vf = VF_PORT_VF_NOT_USED;
> +
> + err = nla_parse_nested(vf_port, IFLA_VF_PORT_MAX,
> + tb[IFLA_VF_PORT], ifla_vf_port_policy);
> + if (err < 0)
> + goto errout;
> +
> + if (vf_port[IFLA_VF_PORT_VF])
> + vf = nla_get_u32(vf_port[IFLA_VF_PORT_VF]);
> +
> + err = -EOPNOTSUPP;
> + if (ops->ndo_set_vf_port)
> + err = ops->ndo_set_vf_port(dev, vf, vf_port);
> + if (err < 0)
> + goto errout;
> + modified = 1;
> + }
> + err = 0;
> +
> errout:
> if (err < 0 && modified && net_ratelimit())
> printk(KERN_WARNING "A link change request failed with "
>
^ permalink raw reply
* Re: [GIT PULL] last minute vhost-net fix
From: David Miller @ 2010-05-14 11:04 UTC (permalink / raw)
To: mst; +Cc: kvm, virtualization, netdev, linux-kernel
In-Reply-To: <20100513084433.GA23082@redhat.com>
From: "Michael S. Tsirkin" <mst@redhat.com>
Date: Thu, 13 May 2010 11:44:34 +0300
> David, if it's not too late, please pull the following
> last minute fix into 2.6.34.
Pulled, thanks.
^ permalink raw reply
* [GIT] Networking
From: David Miller @ 2010-05-14 11:06 UTC (permalink / raw)
To: torvalds; +Cc: akpm, netdev, linux-kernel
One small last minute fix from the VHOST folks to deal with some
memory barrier issues.
Please pull, thanks a lot!
The following changes since commit cea0d767c29669bf89f86e4aee46ef462d2ebae8:
Linus Torvalds (1):
Merge branch 'hwmon-for-linus' of git://git.kernel.org/.../jdelvare/staging
are available in the git repository at:
master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master
David S. Miller (1):
Merge branch 'net-2.6' of git://git.kernel.org/.../mst/vhost
Michael S. Tsirkin (1):
vhost: fix barrier pairing
drivers/vhost/vhost.c | 7 ++++++-
1 files changed, 6 insertions(+), 1 deletions(-)
^ permalink raw reply
* Re: [net-next-2.6 V7 PATCH 1/2] Add netlink support for virtual port management (was iovnl)
From: Arnd Bergmann @ 2010-05-14 12:12 UTC (permalink / raw)
To: Patrick McHardy; +Cc: Scott Feldman, davem, netdev, chrisw
In-Reply-To: <4BED2CD8.4020209@trash.net>
On Friday 14 May 2010, Patrick McHardy wrote:
> Scott Feldman wrote:
> > --- a/net/core/rtnetlink.c
> > +++ b/net/core/rtnetlink.c
> > @@ -653,6 +653,26 @@ static inline int rtnl_vfinfo_size(const struct net_device *dev)
> > return 0;
> > }
> >
> > +static size_t rtnl_vf_port_size(const struct net_device *dev)
> > +{
> > + size_t vf_port_size = nla_total_size(sizeof(struct nlattr))
> > + /* VF_PORT_VF */
> > + + nla_total_size(VF_PORT_PROFILE_MAX)/* VF_PORT_PROFILE */
> > + + nla_total_size(sizeof(struct ifla_vf_port_vsi))
> > + /* VF_PORT_VSI_TYPE */
> > + + nla_total_size(VF_PORT_UUID_MAX) /* VF_PORT_VSI_INSTANCE */
> > + + nla_total_size(VF_PORT_UUID_MAX) /* VF_PORT_HOST_UUID */
> > + + nla_total_size(1) /* VF_PROT_VDP_REQUEST */
>
> Do messages generated by the kernel really contain a request?
Yes, the request field of the VDP message shows the status (e.g. associated or
disassociated).
> > +static int rtnl_vf_port_fill_nest(struct sk_buff *skb, struct net_device *dev,
> > + int vf)
> > +{
> > + struct nlattr *data;
> > + int err;
> > +
> > + data = nla_nest_start(skb, IFLA_VF_PORT);
>
> We usually use a top-level attribute to encapsulate lists of identical
> attributes. The other iflink attributes may only occur once and are
> usually parsed using nla_parse_nested(), which will parse all
> IFLA_VF_PORT attributes, but only return the last one.
>
> Something like:
>
> iflink message:
> ...
> [IFLA_VF_PORTS]
> [IFLA_VF_PORT]
> [IFLA_VF_PORT_*], ...
> [IFLA_VF_PORT]
> [IFLA_VF_PORT_*], ...
> ...
Ah, I was wondering about this already. Does this mean that IFLA_VFINFO
does this incorrectly as well?
> > static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
> > int type, u32 pid, u32 seq, u32 change,
> > unsigned int flags)
> > @@ -747,17 +819,23 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
> > goto nla_put_failure;
> > copy_rtnl_link_stats64(nla_data(attr), stats);
> >
> > + if (dev->dev.parent)
> > + NLA_PUT_U32(skb, IFLA_NUM_VF, dev_num_vf(dev->dev.parent));
>
> Just wondering, is the only case where dev.parent is non-NULL
> really when virtual ports are present?
No, but if parent is NULL, we must not call dev_num_vf(). The way that enic
needs the attributes, they can be either for the VF of dev->dev.parent (the
PCI PF), or for the PF itself, even if it does not have VFs, in which case
it would be interesting to have IFLA_NUM_VF = 0 in the output.
Maybe a better structure would be to separate the two cases, also allowing
a port profile to be associated with both the PF and with each of its VFs?
Something like this:
[IFLA_NUM_VF]
[IFLA_VF_PORTS]
[IFLA_VF_PORT]
[IFLA_VF_PORT_*], ...
[IFLA_VF_PORT]
[IFLA_VF_PORT_*], ...
[IFLA_PORT_SELF]
[IFLA_VF_PORT_*], ...
Arnd
^ permalink raw reply
* [RFC] NF: IP tables idletimer target implementation
From: Luciano Coelho @ 2010-05-14 12:50 UTC (permalink / raw)
To: netdev; +Cc: Timo Teras
This patch implements an idletimer IP tables target that can be used to
identify when interfaces have been idle for a certain period of time.
It adds a file to the sysfs for each interface that is brought up. The file
contains the time remaining before the event is triggered. This file can
also be used to set the timer manually.
The default timeout should be set when the IP table rule is defined with the
--timeout parameter set.
This implementation was originally done by Timo Teras and a few other people
who have sent patches with updates and fixes. It has lived for a while in
the linux-omap tree, but has been removed when linux-omap was aligned with
upstream. Now the patch has been forward-ported, which includes a few
changes related to net namespaces, x_tables etc.
While this is not the best approach for interface idle time monitoring, it is
non-intrusive and fits well in the existing architecture without any major
changes to the networking subsystem.
Cc: Timo Teras <timo.teras@iki.fi>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
---
include/linux/netfilter_ipv4/ipt_IDLETIMER.h | 22 ++
net/ipv4/netfilter/Kconfig | 17 ++
net/ipv4/netfilter/Makefile | 1 +
net/ipv4/netfilter/ipt_IDLETIMER.c | 320 ++++++++++++++++++++++++++
4 files changed, 360 insertions(+), 0 deletions(-)
create mode 100644 include/linux/netfilter_ipv4/ipt_IDLETIMER.h
create mode 100644 net/ipv4/netfilter/ipt_IDLETIMER.c
diff --git a/include/linux/netfilter_ipv4/ipt_IDLETIMER.h b/include/linux/netfilter_ipv4/ipt_IDLETIMER.h
new file mode 100644
index 0000000..89993e2
--- /dev/null
+++ b/include/linux/netfilter_ipv4/ipt_IDLETIMER.h
@@ -0,0 +1,22 @@
+/*
+ * linux/include/linux/netfilter_ipv4/ipt_IDLETIMER.h
+ *
+ * Header file for IP tables timer target module.
+ *
+ * Copyright (C) 2004 Nokia Corporation
+ * Written by Timo Teräs <ext-timo.teras@nokia.com>
+ *
+ * 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.
+ */
+
+#ifndef _IPT_TIMER_H
+#define _IPT_TIMER_H
+
+struct ipt_idletimer_info {
+ unsigned int timeout;
+};
+
+#endif
diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig
index 1833bdb..91fba9a 100644
--- a/net/ipv4/netfilter/Kconfig
+++ b/net/ipv4/netfilter/Kconfig
@@ -204,6 +204,23 @@ config IP_NF_TARGET_REDIRECT
To compile it as a module, choose M here. If unsure, say N.
+config IP_NF_TARGET_IDLETIMER
+ tristate "IDLETIMER target support"
+ depends on IP_NF_IPTABLES
+ help
+ This option adds a `IDLETIMER' target. Each matching packet resets
+ the timer associated with input and/or output interfaces. Timer
+ expiry causes kobject uevent. Idle timer can be read via sysfs.
+
+ To compile it as a module, choose M here. If unsure, say N.
+
+config IP_NF_TARGET_IDLETIMER_DEBUG
+ bool "IDLETIMER target debugging"
+ help
+ Say Y here if you want to get debugging information when using the
+ IDLETIMER target. If unsure, say N.
+
+
config NF_NAT_SNMP_BASIC
tristate "Basic SNMP-ALG support"
depends on NF_NAT
diff --git a/net/ipv4/netfilter/Makefile b/net/ipv4/netfilter/Makefile
index 4811159..60bdaf1 100644
--- a/net/ipv4/netfilter/Makefile
+++ b/net/ipv4/netfilter/Makefile
@@ -60,6 +60,7 @@ obj-$(CONFIG_IP_NF_TARGET_MASQUERADE) += ipt_MASQUERADE.o
obj-$(CONFIG_IP_NF_TARGET_NETMAP) += ipt_NETMAP.o
obj-$(CONFIG_IP_NF_TARGET_REDIRECT) += ipt_REDIRECT.o
obj-$(CONFIG_IP_NF_TARGET_REJECT) += ipt_REJECT.o
+obj-$(CONFIG_IP_NF_TARGET_IDLETIMER) += ipt_IDLETIMER.o
obj-$(CONFIG_IP_NF_TARGET_ULOG) += ipt_ULOG.o
# generic ARP tables
diff --git a/net/ipv4/netfilter/ipt_IDLETIMER.c b/net/ipv4/netfilter/ipt_IDLETIMER.c
new file mode 100644
index 0000000..2c5b465
--- /dev/null
+++ b/net/ipv4/netfilter/ipt_IDLETIMER.c
@@ -0,0 +1,320 @@
+/*
+ * linux/net/ipv4/netfilter/ipt_IDLETIMER.c
+ *
+ * Netfilter module to trigger a timer when packet matches.
+ * After timer expires a kevent will be sent.
+ *
+ * Copyright (C) 2004, 2010 Nokia Corporation
+ * Written by Timo Teras <ext-timo.teras@nokia.com>
+ *
+ * Contact: Luciano Coelho <luciano.coelho@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ */
+
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/timer.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/notifier.h>
+#include <linux/netfilter.h>
+#include <linux/rtnetlink.h>
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter_ipv4/ipt_IDLETIMER.h>
+#include <linux/kobject.h>
+#include <linux/workqueue.h>
+
+#ifdef CONFIG_IP_NF_TARGET_IDLETIMER_DEBUG
+#define DEBUGP(format, args...) printk(KERN_DEBUG \
+ "ipt_IDLETIMER:%s:" format "\n", \
+ __func__ , ## args)
+#else
+#define DEBUGP(format, args...)
+#endif
+
+/*
+ * Internal timer management.
+ */
+static ssize_t utimer_attr_show(struct device *dev,
+ struct device_attribute *attr, char *buf);
+static ssize_t utimer_attr_store(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t count);
+
+struct utimer_t {
+ char name[IFNAMSIZ];
+ struct list_head entry;
+ struct timer_list timer;
+ struct work_struct work;
+ struct net *net;
+};
+
+static LIST_HEAD(active_utimer_head);
+static DEFINE_SPINLOCK(list_lock);
+static DEVICE_ATTR(idletimer, 0644, utimer_attr_show, utimer_attr_store);
+
+static void utimer_delete(struct utimer_t *timer)
+{
+ DEBUGP("Deleting timer '%s'\n", timer->name);
+
+ list_del(&timer->entry);
+ del_timer_sync(&timer->timer);
+ put_net(timer->net);
+ kfree(timer);
+}
+
+static void utimer_work(struct work_struct *work)
+{
+ struct utimer_t *timer = container_of(work, struct utimer_t, work);
+ struct net_device *netdev = NULL;
+
+ netdev = dev_get_by_name(timer->net, timer->name);
+
+ if (netdev != NULL) {
+ sysfs_notify(&netdev->dev.kobj, NULL,
+ "idletimer");
+ dev_put(netdev);
+ }
+}
+
+static void utimer_expired(unsigned long data)
+{
+ struct utimer_t *timer = (struct utimer_t *) data;
+
+ DEBUGP("Timer '%s' expired\n", timer->name);
+
+ spin_lock_bh(&list_lock);
+ utimer_delete(timer);
+ spin_unlock_bh(&list_lock);
+
+ schedule_work(&timer->work);
+}
+
+static struct utimer_t *utimer_create(const char *name,
+ struct net *net)
+{
+ struct utimer_t *timer;
+
+ timer = kmalloc(sizeof(struct utimer_t), GFP_ATOMIC);
+ if (timer == NULL)
+ return NULL;
+
+ list_add(&timer->entry, &active_utimer_head);
+ strlcpy(timer->name, name, sizeof(timer->name));
+ timer->net = get_net(net);
+
+ init_timer(&timer->timer);
+ timer->timer.function = utimer_expired;
+ timer->timer.data = (unsigned long) timer;
+
+ INIT_WORK(&timer->work, utimer_work);
+
+ DEBUGP("Created timer '%s'\n", timer->name);
+
+ return timer;
+}
+
+static struct utimer_t *__utimer_find(const char *name, const struct net *net)
+{
+ struct utimer_t *entry;
+
+ list_for_each_entry(entry, &active_utimer_head, entry) {
+ if (!strcmp(name, entry->name) && net == entry->net)
+ return entry;
+ }
+
+ return NULL;
+}
+
+static void utimer_modify(const char *name,
+ struct net *net,
+ unsigned long expires)
+{
+ struct utimer_t *timer;
+
+ DEBUGP("Modifying timer '%s'\n", name);
+ spin_lock_bh(&list_lock);
+ timer = __utimer_find(name, net);
+ if (timer == NULL)
+ timer = utimer_create(name, net);
+ mod_timer(&timer->timer, expires);
+ spin_unlock_bh(&list_lock);
+}
+
+static ssize_t utimer_attr_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct utimer_t *timer;
+ struct net_device *netdev = to_net_dev(dev);
+ unsigned long expires = 0;
+
+ spin_lock_bh(&list_lock);
+ timer = __utimer_find(netdev->name, dev_net(netdev));
+ if (timer)
+ expires = timer->timer.expires;
+ spin_unlock_bh(&list_lock);
+
+ if (expires)
+ return sprintf(buf, "%lu\n", (expires-jiffies) / HZ);
+
+ return sprintf(buf, "0\n");
+}
+
+static ssize_t utimer_attr_store(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t count)
+{
+ int expires;
+ struct net_device *netdev = to_net_dev(dev);
+
+ if (sscanf(buf, "%d", &expires) == 1) {
+ if (expires > 0)
+ utimer_modify(netdev->name,
+ dev_net(netdev),
+ jiffies+HZ*(unsigned long)expires);
+ }
+
+ return count;
+}
+
+static int utimer_notifier_call(struct notifier_block *this,
+ unsigned long event, void *ptr)
+{
+ struct net_device *netdev = ptr;
+ int ret;
+
+ switch (event) {
+ case NETDEV_UP:
+ DEBUGP("NETDEV_UP: %s\n", netdev->name);
+ ret = device_create_file(&netdev->dev,
+ &dev_attr_idletimer);
+ WARN_ON(ret);
+
+ break;
+ case NETDEV_DOWN:
+ DEBUGP("NETDEV_DOWN: %s\n", netdev->name);
+ device_remove_file(&netdev->dev,
+ &dev_attr_idletimer);
+ break;
+ }
+
+ return NOTIFY_DONE;
+}
+
+static struct notifier_block utimer_notifier_block = {
+ .notifier_call = utimer_notifier_call,
+};
+
+
+static int utimer_init(void)
+{
+ return register_netdevice_notifier(&utimer_notifier_block);
+}
+
+static void utimer_fini(void)
+{
+ struct utimer_t *entry, *next;
+ struct net_device *dev;
+ struct net *net;
+
+ list_for_each_entry_safe(entry, next, &active_utimer_head, entry)
+ utimer_delete(entry);
+
+ rtnl_lock();
+ unregister_netdevice_notifier(&utimer_notifier_block);
+ for_each_net(net) {
+ for_each_netdev(net, dev) {
+ utimer_notifier_call(&utimer_notifier_block,
+ NETDEV_DOWN, dev);
+ }
+ }
+ rtnl_unlock();
+}
+
+/*
+ * The actual iptables plugin.
+ */
+static unsigned int ipt_idletimer_target(struct sk_buff *skb,
+ const struct xt_action_param *par)
+{
+ const struct ipt_idletimer_info *target = par->targinfo;
+ unsigned long expires;
+
+ expires = jiffies + HZ*target->timeout;
+
+ if (par->in != NULL)
+ utimer_modify(par->in->name,
+ dev_net(par->in),
+ expires);
+
+ if (par->out != NULL)
+ utimer_modify(par->out->name,
+ dev_net(par->out),
+ expires);
+
+ return XT_CONTINUE;
+}
+
+static int ipt_idletimer_checkentry(const struct xt_tgchk_param *par)
+{
+ const struct ipt_idletimer_info *info = par->targinfo;
+
+ if (info->timeout == 0) {
+ DEBUGP("timeout value is zero\n");
+ return false;
+ }
+
+ return true;
+}
+
+static struct xt_target ipt_idletimer = {
+ .name = "IDLETIMER",
+ .family = NFPROTO_IPV4,
+ .target = ipt_idletimer_target,
+ .targetsize = sizeof(struct ipt_idletimer_info),
+ .checkentry = ipt_idletimer_checkentry,
+ .me = THIS_MODULE,
+};
+
+static int __init init(void)
+{
+ int ret;
+
+ ret = utimer_init();
+ if (ret)
+ return ret;
+
+ ret = xt_register_target(&ipt_idletimer);
+ if (ret < 0) {
+ utimer_fini();
+ return ret;
+ }
+
+ return 0;
+}
+
+static void __exit fini(void)
+{
+ xt_unregister_target(&ipt_idletimer);
+ utimer_fini();
+}
+
+module_init(init);
+module_exit(fini);
+
+MODULE_AUTHOR("Timo Teras <ext-timo.teras@nokia.com>");
+MODULE_DESCRIPTION("iptables idletimer target module");
+MODULE_LICENSE("GPL");
--
1.6.3.3
^ permalink raw reply related
* loosing IPMI-card by loading netconsole
From: Henning Fehrmann @ 2010-05-14 13:45 UTC (permalink / raw)
To: Jeff Kirsher
Cc: Jesse Brandeburg, Bruce Allan, PJ Waskiewicz, John Ronciak,
netdev, Matt Mackall, Carsten Aulbert, Tejun Heo
Hello,
We have SuperMicro PDSM 2+ boards together with 82573E and 82573L Intel NICs.
Additional we have IPMI cards which are tunneled via the 82573E NIC.
We are using the e1000e driver for NICs together with the netconsole driver.
(netconsole netconsole=4444@client_IP/eth1,514@server_IP/server_MAC). Netconsole
is using the NIC which is NOT used by the IPMI card.
Usually we are able to access the IPMI card remotely with ipmitools.
Having the kernel 2.6.27.39 installed everything worked together, the NICs, remotely
accessing the IPMI cards and netconsole.
The driver version of e1000e is 0.3.3.3-k6.
Using a more recent kernel: 2.6.32.7 we lost the ability of remotely accessing the IPMI card
when the netconsole driver is loaded.
The IPMI card is accessible before the netconsole driver is loaded and disappears
once we use netconsole. Even unloading netconsole does not help then.
We get the IPMI card back when 'removing' eth0:
echo 1 > /sys/devices/pci0000:00/0000:00:1c.4/0000:0d:00.0/remove
0d:00.0 is eth0.
The version of the e1000e driver is 1.0.2-k2.
I compiled and loaded a later version of this driver (1.1.19) without solving this problem.
For eth0 (82573E) we have a firmware version 0.15-4 installed and for eth1 we use the firmware
version 0.5-7. But this is the same for both kernel versions.
Do you have an idea?
Thank you and cheers,
Henning
^ permalink raw reply
* [PATCH] fsl_pq_mdio: Fix mdiobus allocation handling
From: Anton Vorontsov @ 2010-05-14 14:27 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linuxppc-dev
The driver could return success code even if mdiobus_alloc() failed.
This patch fixes the issue.
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
---
drivers/net/fsl_pq_mdio.c | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/fsl_pq_mdio.c b/drivers/net/fsl_pq_mdio.c
index 3acac5f..ff028f5 100644
--- a/drivers/net/fsl_pq_mdio.c
+++ b/drivers/net/fsl_pq_mdio.c
@@ -277,15 +277,17 @@ static int fsl_pq_mdio_probe(struct of_device *ofdev,
int tbiaddr = -1;
const u32 *addrp;
u64 addr = 0, size = 0;
- int err = 0;
+ int err;
priv = kzalloc(sizeof(*priv), GFP_KERNEL);
if (!priv)
return -ENOMEM;
new_bus = mdiobus_alloc();
- if (NULL == new_bus)
+ if (!new_bus) {
+ err = -ENOMEM;
goto err_free_priv;
+ }
new_bus->name = "Freescale PowerQUICC MII Bus",
new_bus->read = &fsl_pq_mdio_read,
--
1.7.0.5
^ permalink raw reply related
* [PATCH] gianfar: Remove legacy PM callbacks
From: Anton Vorontsov @ 2010-05-14 14:27 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linuxppc-dev
These callbacks were needed because dev_pm_ops support for OF
platform devices was in the powerpc tree, and the patch that
added dev_pm_ops for gianfar driver was in the netdev tree. Now
that netdev and powerpc trees have merged into Linus' tree, we
can remove the legacy hooks.
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
---
drivers/net/gianfar.c | 14 --------------
1 files changed, 0 insertions(+), 14 deletions(-)
diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c
index 5d3763f..fb23f04 100644
--- a/drivers/net/gianfar.c
+++ b/drivers/net/gianfar.c
@@ -1288,21 +1288,9 @@ static struct dev_pm_ops gfar_pm_ops = {
#define GFAR_PM_OPS (&gfar_pm_ops)
-static int gfar_legacy_suspend(struct of_device *ofdev, pm_message_t state)
-{
- return gfar_suspend(&ofdev->dev);
-}
-
-static int gfar_legacy_resume(struct of_device *ofdev)
-{
- return gfar_resume(&ofdev->dev);
-}
-
#else
#define GFAR_PM_OPS NULL
-#define gfar_legacy_suspend NULL
-#define gfar_legacy_resume NULL
#endif
@@ -3055,8 +3043,6 @@ static struct of_platform_driver gfar_driver = {
.probe = gfar_probe,
.remove = gfar_remove,
- .suspend = gfar_legacy_suspend,
- .resume = gfar_legacy_resume,
.driver.pm = GFAR_PM_OPS,
};
--
1.7.0.5
^ permalink raw reply related
* RE: loosing IPMI-card by loading netconsole
From: Ronciak, John @ 2010-05-14 14:51 UTC (permalink / raw)
To: Henning Fehrmann, Kirsher, Jeffrey T
Cc: Brandeburg, Jesse, Allan, Bruce W, Waskiewicz Jr, Peter P,
netdev@vger.kernel.org, Matt Mackall, Carsten Aulbert, Tejun Heo
In-Reply-To: <20100514134544.GA26674@gretchen.aei.mpg.de>
Sorry to hear about the problem you are having Henning. What do you mean when you say "it disappears"? Can both eth0 and eth1 ping (or be pinged)? Do all the networking devices still show up in the system when you do an 'lspci'? What happens if you down and then up the interface you are having problems with? Does 'rmmod' do the same thing as your removal method? Is there anything in the system logs saying anything about the interfaces?
We have not had reports of this so this is a bit unusual. Please let us know.
Does this happen on other systems as well or just one particular system?
Cheers,
John
> -----Original Message-----
> From: Henning Fehrmann [mailto:henning.fehrmann@aei.mpg.de]
> Sent: Friday, May 14, 2010 6:46 AM
> To: Kirsher, Jeffrey T
> Cc: Brandeburg, Jesse; Allan, Bruce W; Waskiewicz Jr, Peter P; Ronciak,
> John; netdev@vger.kernel.org; Matt Mackall; Carsten Aulbert; Tejun Heo
> Subject: loosing IPMI-card by loading netconsole
>
> Hello,
>
> We have SuperMicro PDSM 2+ boards together with 82573E and 82573L Intel
> NICs.
> Additional we have IPMI cards which are tunneled via the 82573E NIC.
>
> We are using the e1000e driver for NICs together with the netconsole
> driver.
> (netconsole netconsole=4444@client_IP/eth1,514@server_IP/server_MAC).
> Netconsole is using the NIC which is NOT used by the IPMI card.
>
> Usually we are able to access the IPMI card remotely with ipmitools.
>
> Having the kernel 2.6.27.39 installed everything worked together, the
> NICs, remotely accessing the IPMI cards and netconsole.
>
> The driver version of e1000e is 0.3.3.3-k6.
>
> Using a more recent kernel: 2.6.32.7 we lost the ability of remotely
> accessing the IPMI card when the netconsole driver is loaded.
> The IPMI card is accessible before the netconsole driver is loaded and
> disappears once we use netconsole. Even unloading netconsole does not
> help then.
>
> We get the IPMI card back when 'removing' eth0:
> echo 1 > /sys/devices/pci0000:00/0000:00:1c.4/0000:0d:00.0/remove
>
> 0d:00.0 is eth0.
>
>
> The version of the e1000e driver is 1.0.2-k2.
>
> I compiled and loaded a later version of this driver (1.1.19) without
> solving this problem.
>
>
> For eth0 (82573E) we have a firmware version 0.15-4 installed and for
> eth1 we use the firmware version 0.5-7. But this is the same for both
> kernel versions.
>
> Do you have an idea?
>
> Thank you and cheers,
> Henning
^ permalink raw reply
* Re: [PATCH 0/6] sky2: update
From: Stephen Hemminger @ 2010-05-14 15:19 UTC (permalink / raw)
To: David Miller; +Cc: mikem, netdev
In-Reply-To: <20100514.031501.57481177.davem@davemloft.net>
On Fri, 14 May 2010 03:15:01 -0700 (PDT)
David Miller <davem@davemloft.net> wrote:
> From: Stephen Hemminger <shemminger@vyatta.com>
> Date: Thu, 13 May 2010 09:12:47 -0700
>
> > Bunch of patches from Mike, with some additional comments.
>
> All applied to net-next-2.6, thanks.
The first one needs to go to net-2.6 because it a regression:
Current code will lose multicast addresses when the automatic
recovery from stuck chip happens. Auto recovery happens a lot
under load on some configurations.
--
^ permalink raw reply
* Fw: [Bug 15974] New: kernel panic when squid in bridge mode
From: Stephen Hemminger @ 2010-05-14 15:24 UTC (permalink / raw)
To: Bart De Schuymer, Patrick McHardy; +Cc: netdev
Begin forwarded message:
Date: Fri, 14 May 2010 08:52:07 GMT
From: bugzilla-daemon@bugzilla.kernel.org
To: shemminger@linux-foundation.org
Subject: [Bug 15974] New: kernel panic when squid in bridge mode
https://bugzilla.kernel.org/show_bug.cgi?id=15974
Summary: kernel panic when squid in bridge mode
Product: Networking
Version: 2.5
Kernel Version: 2.6.30.5
Platform: All
OS/Version: Linux
Tree: Mainline
Status: NEW
Severity: high
Priority: P1
Component: IPV4
AssignedTo: shemminger@linux-foundation.org
ReportedBy: senthilkumaar2021@gmail.com
Regression: No
Hi we are using squid tproxy in bridge mode .The kernel version used is
2.6.30.5 once in 10-15 hours we are getting kernel panic message in he screen
.We are passing traffic of 100Mbps through bridge.The iptables and ebtables are
used for squid
ptables -t mangle -N DIVERT
iptables -t mangle -A DIVERT -j MARK --set-mark 1
iptables -t mangle -A DIVERT -j ACCEPT
iptables -t mangle -A PREROUTING -p tcp -m socket -j DIVERT
iptables -t mangle -A PREROUTING -p tcp --dport 80 -j TPROXY --tproxy-mark
0x1/0x1 --on-port 3129
ebtables -t broute -A BROUTING -i $CLIENT_IFACE -p ipv4 --ip-proto tcp
--ip-dport 80 -j redirect --redirect-target DROP
ebtables -t broute -A BROUTING -i $INET_IFACE -p ipv4 --ip-proto tcp --ip-sport
80 -j redirect --redirect-target DROP
we have got kernel panic in kernel 2.6.28.5 also
the error is
<ffffffffa03933c2>] ? nf_nat_fn+0x138/0x14e [iptable_nat]
[<ffffffffa0393585>] ? nf_nat_in+0x2f/0x6e [iptable_nat]
[<ffffffffa027edaa>] ? br_nf_pre_routing_finish+0x0/0x2c4 [bridge]
[<ffffffffa027edfa>] br_nf_pre_routing_finish+0x50/0x2c4 [bridge]
[<ffffffffa027edaa>] ? br_nf_pre_routing_finish+0x0/0x2c4 [bridge]
[<ffffffff81339a50>] ? nf_hook_slow+0x68/0xc8
[<ffffffffa027edaa>] ? br_nf_pre_routing_finish+0x0/0x2c4 [bridge]
[<ffffffffa027f616>] br_nf_pre_routing+0x5a8/0x5c7 [bridge]
[<ffffffff813399ab>] nf_iterate+0x48/0x85
[<ffffffffa027a931>] ? br_handle_frame_finish+0x0/0x154 [bridge]
[<ffffffff81339a50>] nf_hook_slow+0x68/0xc8
[<ffffffffa027a931>] ? br_handle_frame_finish+0x0/0x154 [bridge]
[<ffffffffa027ac36>] br_handle_frame+0x1b1/0x1db [bridge]
[<ffffffff8131d54b>] netif_receive_skb+0x316/0x434
[<ffffffff8131dbfb>] napi_gro_receive+0x6e/0x83
[<ffffffffa0125bfe>] e1000_receive_skb+0x5c/0x65 [e1000e]
[<ffffffffa0125de8>] e1000_clean_rx_irq+0x1e1/0x28f [e1000e]
[<ffffffffa012730e>] e1000_clean+0x99/0x24a [e1000e]
[<ffffffff813bcfc5>] ? _spin_unlock_irqrestore+0x2c/0x43
[<ffffffff8131ba62>] net_rx_action+0xb8/0x1b4
[<ffffffff8104ed43>] __do_softirq+0x99/0x152
[<ffffffff8101284c>] call_softirq+0x1c/0x30
[<ffffffff81013a02>] do_softirq+0x52/0xb9
[<ffffffff8104e969>] irq_exit+0x53/0x8d
[<ffffffff81013d1a>] do_IRQ+0x135/0x157
[<ffffffff81011f93>] ret_from_intr+0x0/0x2e
<EOI> [<ffffffff81017e20>] ? mwait_idle+0x9e/0xc7
[<ffffffff81017e17>] ? mwait_idle+0x95/0xc7
[<ffffffff813bfd20>] ? atomic_notifier_call_chain+0x13/0x15
[<ffffffff810102f4>] ? enter_idle+0x27/0x29
Please help me in fixing the issue
--
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
--
^ permalink raw reply
* Re: [RFC] NF: IP tables idletimer target implementation
From: Patrick McHardy @ 2010-05-14 16:03 UTC (permalink / raw)
To: Luciano Coelho; +Cc: netdev, Timo Teras, Netfilter Development Mailinglist
In-Reply-To: <1273841458-10443-1-git-send-email-luciano.coelho@nokia.com>
Please CC netfilter-devel on future submissions.
Luciano Coelho wrote:
> It adds a file to the sysfs for each interface that is brought up. The file
> contains the time remaining before the event is triggered. This file can
> also be used to set the timer manually.
What is this used for? It doesn't seem to smart to poll manually
if you get an event anyways, and the timeout can already be set
per rule.
> diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig
> index 1833bdb..91fba9a 100644
> --- a/net/ipv4/netfilter/Kconfig
> +++ b/net/ipv4/netfilter/Kconfig
> @@ -204,6 +204,23 @@ config IP_NF_TARGET_REDIRECT
>
> To compile it as a module, choose M here. If unsure, say N.
>
> +config IP_NF_TARGET_IDLETIMER
This should be a x_tables target, there's nothing IPv4-specific
about it.
> diff --git a/net/ipv4/netfilter/ipt_IDLETIMER.c b/net/ipv4/netfilter/ipt_IDLETIMER.c
> new file mode 100644
> index 0000000..2c5b465
> --- /dev/null
> +++ b/net/ipv4/netfilter/ipt_IDLETIMER.c
> +
> +#ifdef CONFIG_IP_NF_TARGET_IDLETIMER_DEBUG
> +#define DEBUGP(format, args...) printk(KERN_DEBUG \
> + "ipt_IDLETIMER:%s:" format "\n", \
> + __func__ , ## args)
> +#else
> +#define DEBUGP(format, args...)
> +#endif
Please use pr_debug and get rid of the config option.
> +
> +/*
> + * Internal timer management.
> + */
> +static ssize_t utimer_attr_show(struct device *dev,
> + struct device_attribute *attr, char *buf);
> +static ssize_t utimer_attr_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf, size_t count);
> +
> +struct utimer_t {
> + char name[IFNAMSIZ];
> + struct list_head entry;
> + struct timer_list timer;
> + struct work_struct work;
> + struct net *net;
> +};
> +
> +static LIST_HEAD(active_utimer_head);
> +static DEFINE_SPINLOCK(list_lock);
> +static DEVICE_ATTR(idletimer, 0644, utimer_attr_show, utimer_attr_store);
> +
> +static void utimer_delete(struct utimer_t *timer)
> +{
> + DEBUGP("Deleting timer '%s'\n", timer->name);
> +
> + list_del(&timer->entry);
> + del_timer_sync(&timer->timer);
> + put_net(timer->net);
> + kfree(timer);
> +}
> +
> +static void utimer_work(struct work_struct *work)
> +{
> + struct utimer_t *timer = container_of(work, struct utimer_t, work);
> + struct net_device *netdev = NULL;
Unnecessary initialization.
> +
> + netdev = dev_get_by_name(timer->net, timer->name);
> +
> + if (netdev != NULL) {
> + sysfs_notify(&netdev->dev.kobj, NULL,
> + "idletimer");
> + dev_put(netdev);
> + }
> +}
> +
> +static void utimer_expired(unsigned long data)
> +{
> + struct utimer_t *timer = (struct utimer_t *) data;
> +
> + DEBUGP("Timer '%s' expired\n", timer->name);
> +
> + spin_lock_bh(&list_lock);
> + utimer_delete(timer);
> + spin_unlock_bh(&list_lock);
> +
> + schedule_work(&timer->work);
Use after free, utimer_delete() frees the timer.
> +}
> +
> +static struct utimer_t *utimer_create(const char *name,
> + struct net *net)
> +{
> + struct utimer_t *timer;
> +
> + timer = kmalloc(sizeof(struct utimer_t), GFP_ATOMIC);
> + if (timer == NULL)
> + return NULL;
> +
> + list_add(&timer->entry, &active_utimer_head);
> + strlcpy(timer->name, name, sizeof(timer->name));
> + timer->net = get_net(net);
How does this handle namespace exit?
> +
> + init_timer(&timer->timer);
> + timer->timer.function = utimer_expired;
> + timer->timer.data = (unsigned long) timer;
setup_timer()
> +
> + INIT_WORK(&timer->work, utimer_work);
> +
> + DEBUGP("Created timer '%s'\n", timer->name);
> +
> + return timer;
> +}
> +
> +static struct utimer_t *__utimer_find(const char *name, const struct net *net)
> +{
> + struct utimer_t *entry;
> +
> + list_for_each_entry(entry, &active_utimer_head, entry) {
> + if (!strcmp(name, entry->name) && net == entry->net)
> + return entry;
> + }
> +
> + return NULL;
> +}
> +
> +static void utimer_modify(const char *name,
> + struct net *net,
> + unsigned long expires)
> +{
> + struct utimer_t *timer;
> +
> + DEBUGP("Modifying timer '%s'\n", name);
> + spin_lock_bh(&list_lock);
> + timer = __utimer_find(name, net);
So you're scanning the list up to twice per packet? That seems
highly suboptimal, why not create the timer when the rule is
created and only update the timeout? You could use the interfaces
specified in struct ipt_ip.
> + if (timer == NULL)
> + timer = utimer_create(name, net);
> + mod_timer(&timer->timer, expires);
> + spin_unlock_bh(&list_lock);
> +}
> +
> +static ssize_t utimer_attr_show(struct device *dev,
> + struct device_attribute *attr, char *buf)
> +{
> + struct utimer_t *timer;
> + struct net_device *netdev = to_net_dev(dev);
> + unsigned long expires = 0;
> +
> + spin_lock_bh(&list_lock);
> + timer = __utimer_find(netdev->name, dev_net(netdev));
> + if (timer)
> + expires = timer->timer.expires;
> + spin_unlock_bh(&list_lock);
> +
> + if (expires)
> + return sprintf(buf, "%lu\n", (expires-jiffies) / HZ);
> +
> + return sprintf(buf, "0\n");
> +}
> +
> +static ssize_t utimer_attr_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf, size_t count)
> +{
> + int expires;
> + struct net_device *netdev = to_net_dev(dev);
> +
> + if (sscanf(buf, "%d", &expires) == 1) {
> + if (expires > 0)
Using %u seems better.
> + utimer_modify(netdev->name,
> + dev_net(netdev),
> + jiffies+HZ*(unsigned long)expires);
> + }
> +
> + return count;
> +}
> +
> +static int utimer_notifier_call(struct notifier_block *this,
> + unsigned long event, void *ptr)
> +{
> + struct net_device *netdev = ptr;
> + int ret;
> +
> + switch (event) {
> + case NETDEV_UP:
> + DEBUGP("NETDEV_UP: %s\n", netdev->name);
> + ret = device_create_file(&netdev->dev,
> + &dev_attr_idletimer);
> + WARN_ON(ret);
> +
> + break;
> + case NETDEV_DOWN:
> + DEBUGP("NETDEV_DOWN: %s\n", netdev->name);
> + device_remove_file(&netdev->dev,
> + &dev_attr_idletimer);
> + break;
> + }
> +
> + return NOTIFY_DONE;
> +}
> +
> +static struct notifier_block utimer_notifier_block = {
> + .notifier_call = utimer_notifier_call,
> +};
> +
> +
> +static int utimer_init(void)
> +{
> + return register_netdevice_notifier(&utimer_notifier_block);
> +}
> +
> +static void utimer_fini(void)
> +{
> + struct utimer_t *entry, *next;
> + struct net_device *dev;
> + struct net *net;
> +
> + list_for_each_entry_safe(entry, next, &active_utimer_head, entry)
> + utimer_delete(entry);
> +
> + rtnl_lock();
deadlock? unregister_netdevice_notifier() already takes the RTNL.
> + unregister_netdevice_notifier(&utimer_notifier_block);
> + for_each_net(net) {
> + for_each_netdev(net, dev) {
> + utimer_notifier_call(&utimer_notifier_block,
> + NETDEV_DOWN, dev);
> + }
> + }
> + rtnl_unlock();
> +}
> +
> +/*
> + * The actual iptables plugin.
> + */
> +static unsigned int ipt_idletimer_target(struct sk_buff *skb,
> + const struct xt_action_param *par)
> +{
> + const struct ipt_idletimer_info *target = par->targinfo;
> + unsigned long expires;
> +
> + expires = jiffies + HZ*target->timeout;
> +
> + if (par->in != NULL)
> + utimer_modify(par->in->name,
> + dev_net(par->in),
> + expires);
> +
> + if (par->out != NULL)
> + utimer_modify(par->out->name,
> + dev_net(par->out),
> + expires);
> +
> + return XT_CONTINUE;
> +}
> +
> +static int ipt_idletimer_checkentry(const struct xt_tgchk_param *par)
> +{
> + const struct ipt_idletimer_info *info = par->targinfo;
> +
> + if (info->timeout == 0) {
> + DEBUGP("timeout value is zero\n");
> + return false;
> + }
> +
> + return true;
The return convention in the current net-next tree is 0 for
no error or an errno code otherwise.
> +}
^ permalink raw reply
* Re: loosing IPMI-card by loading netconsole
From: Tejun Heo @ 2010-05-14 16:27 UTC (permalink / raw)
To: Ronciak, John
Cc: Henning Fehrmann, Kirsher, Jeffrey T, Brandeburg, Jesse,
Allan, Bruce W, Waskiewicz Jr, Peter P, netdev@vger.kernel.org,
Matt Mackall, Carsten Aulbert
In-Reply-To: <DDC57477F5D6F845A0DDCB99D3C4812D0CA63BB9E4@orsmsx510.amr.corp.intel.com>
Hello, John.
As Henning seems offline, I'll try to fill in.
On 05/14/2010 04:51 PM, Ronciak, John wrote:
> Sorry to hear about the problem you are having Henning. What do you
> mean when you say "it disappears"?
It stops responding to IPMI requests.
> Can both eth0 and eth1 ping (or be pinged)? Do all the networking
> devices still show up in the system when you do an 'lspci'?
Yeah, everything other than IPMI works just fine.
> What happens if you down and then up the interface you are having
> problems with? Does 'rmmod' do the same thing as your removal
> method?
Haven't tried these but well I think rmmoding should achieve about the
same thing.
> Is there anything in the system logs saying anything about the
> interfaces?
Nope.
> We have not had reports of this so this is a bit unusual. Please let us know.
>
> Does this happen on other systems as well or just one particular system?
Yeah, it happens on at least several hundred machines, so not an
isolated hardware issue at all.
To sum up.
On 2.6.27.39, netconsole + IPMI works fine. On 2.6.32.7, as soon as
netconsole is loaded, IPMI stops working. Unloading netconsole
doesn't revive IPMI but detaching the driver from the controller does.
In both cases, usual networking works fine.
Thanks.
--
tejun
^ permalink raw reply
* SR-IOV PCI quirk for 82599?
From: Fischer, Anna @ 2010-05-14 16:26 UTC (permalink / raw)
To: netdev@vger.kernel.org, e1000-devel@lists.sourceforge.net
There is a PCI quirk for the 82576 controller that programs the PCI BARs to use Flash memory if the BIOS has not allocated resources for the SR-IOV VF BARs.
Is there a similar quirk for the 82599, or can even the same one be used for that device?
Thanks,
Anna
^ permalink raw reply
* RE: loosing IPMI-card by loading netconsole
From: Ronciak, John @ 2010-05-14 16:39 UTC (permalink / raw)
To: Tejun Heo
Cc: Henning Fehrmann, Kirsher, Jeffrey T, Brandeburg, Jesse,
Allan, Bruce W, Waskiewicz Jr, Peter P, netdev@vger.kernel.org,
Matt Mackall, Carsten Aulbert
In-Reply-To: <4BED79EB.1000204@kernel.org>
> -----Original Message-----
> From: Tejun Heo [mailto:tj@kernel.org]
> Sent: Friday, May 14, 2010 9:27 AM
> To: Ronciak, John
> Cc: Henning Fehrmann; Kirsher, Jeffrey T; Brandeburg, Jesse; Allan,
> Bruce W; Waskiewicz Jr, Peter P; netdev@vger.kernel.org; Matt Mackall;
> Carsten Aulbert
> Subject: Re: loosing IPMI-card by loading netconsole
>
> Hello, John.
>
> As Henning seems offline, I'll try to fill in.
>
> On 05/14/2010 04:51 PM, Ronciak, John wrote:
> > Sorry to hear about the problem you are having Henning. What do you
> > mean when you say "it disappears"?
>
> It stops responding to IPMI requests.
>
> > Can both eth0 and eth1 ping (or be pinged)? Do all the networking
> > devices still show up in the system when you do an 'lspci'?
>
> Yeah, everything other than IPMI works just fine.
>
> > What happens if you down and then up the interface you are having
> > problems with? Does 'rmmod' do the same thing as your removal
> method?
>
> Haven't tried these but well I think rmmoding should achieve about the
> same thing.
>
> > Is there anything in the system logs saying anything about the
> > interfaces?
>
> Nope.
>
> > We have not had reports of this so this is a bit unusual. Please let
> us know.
> >
> > Does this happen on other systems as well or just one particular
> system?
>
> Yeah, it happens on at least several hundred machines, so not an
> isolated hardware issue at all.
>
> To sum up.
>
> On 2.6.27.39, netconsole + IPMI works fine. On 2.6.32.7, as soon as
> netconsole is loaded, IPMI stops working. Unloading netconsole doesn't
> revive IPMI but detaching the driver from the controller does.
> In both cases, usual networking works fine.
>
> Thanks.
>
> --
> Tejun
Thanks Tejun.
Since the networking things seem to be operational could it be that with the newer kernel the tunnels you had setup somehow no longer exist or have been disabled somehow? When this interface is removed and setup again it gets things fixed up again?
Cheers,
John
^ permalink raw reply
* Re: [net-next-2.6 V7 PATCH 1/2] Add netlink support for virtual port management (was iovnl)
From: Patrick McHardy @ 2010-05-14 16:42 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: Scott Feldman, davem, netdev, chrisw
In-Reply-To: <201005141412.01578.arnd@arndb.de>
Arnd Bergmann wrote:
> On Friday 14 May 2010, Patrick McHardy wrote:
>>> +static int rtnl_vf_port_fill_nest(struct sk_buff *skb, struct net_device *dev,
>>> + int vf)
>>> +{
>>> + struct nlattr *data;
>>> + int err;
>>> +
>>> + data = nla_nest_start(skb, IFLA_VF_PORT);
>> We usually use a top-level attribute to encapsulate lists of identical
>> attributes. The other iflink attributes may only occur once and are
>> usually parsed using nla_parse_nested(), which will parse all
>> IFLA_VF_PORT attributes, but only return the last one.
>>
>> Something like:
>>
>> iflink message:
>> ...
>> [IFLA_VF_PORTS]
>> [IFLA_VF_PORT]
>> [IFLA_VF_PORT_*], ...
>> [IFLA_VF_PORT]
>> [IFLA_VF_PORT_*], ...
>> ...
>
> Ah, I was wondering about this already. Does this mean that IFLA_VFINFO
> does this incorrectly as well?
Yes.
>>> static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
>>> int type, u32 pid, u32 seq, u32 change,
>>> unsigned int flags)
>>> @@ -747,17 +819,23 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
>>> goto nla_put_failure;
>>> copy_rtnl_link_stats64(nla_data(attr), stats);
>>>
>>> + if (dev->dev.parent)
>>> + NLA_PUT_U32(skb, IFLA_NUM_VF, dev_num_vf(dev->dev.parent));
>> Just wondering, is the only case where dev.parent is non-NULL
>> really when virtual ports are present?
>
> No, but if parent is NULL, we must not call dev_num_vf(). The way that enic
> needs the attributes, they can be either for the VF of dev->dev.parent (the
> PCI PF), or for the PF itself, even if it does not have VFs, in which case
> it would be interesting to have IFLA_NUM_VF = 0 in the output.
I see. I was mainly wondering about completely different types of
devices.
> Maybe a better structure would be to separate the two cases, also allowing
> a port profile to be associated with both the PF and with each of its VFs?
>
> Something like this:
>
> [IFLA_NUM_VF]
> [IFLA_VF_PORTS]
> [IFLA_VF_PORT]
> [IFLA_VF_PORT_*], ...
> [IFLA_VF_PORT]
> [IFLA_VF_PORT_*], ...
> [IFLA_PORT_SELF]
> [IFLA_VF_PORT_*], ...
That would also be fine.
^ permalink raw reply
* [PATCH v3 0/3] ptp: IEEE 1588 clock support
From: Richard Cochran @ 2010-05-14 16:44 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss
Now and again there has been talk on this list of adding PTP support
into Linux. One part of the picture is already in place, the
SO_TIMESTAMPING API for hardware time stamping. This patch set offers
the missing second part needed for complete IEEE 1588 support.
The only feature still to be implemented is the hook into the PPS
subsystem, to synchronize the Linux clock to the PTP clock.
Enjoy,
Richard
* Patch ChangeLog
** v3
*** general
- Added documentation on writing clock drivers.
- Added the ioctls for the ancillary clock features.
- Changed wrong subsys_initcall() to module_init() in clock drivers.
- Removed the (too coarse) character device mutex.
- Setting the clock now requires CAP_SYS_TIME.
*** gianfar
- Added alarm feature.
- Added device tree node binding description.
- Added fine grain locking of the clock registers.
- Added the external time stamp feature.
- Added white space for better style.
- Coverted base+offset to structure pointers for register access.
- When removing the driver, we now disable all PTP functions.
** v2
- Changed clock list from a static array into a dynamic list. Also,
use a bitmap to manage the clock's minor numbers.
- Replaced character device semaphore with a mutex.
- Drop .ko from module names in Kbuild help.
- Replace deprecated unifdef-y with header-y for user space header file.
- Added links to both of the ptpd patches on sourceforge.
- Gianfar driver now gets parameters from device tree.
- Added API documentation to Documentation/ptp/ptp.txt
Richard Cochran (3):
ptp: Added a brand new class driver for ptp clocks.
ptp: Added a clock that uses the Linux system time.
ptp: Added a clock that uses the eTSEC found on the MPC85xx.
Documentation/powerpc/dts-bindings/fsl/tsec.txt | 56 +++
Documentation/ptp/ptp.txt | 95 ++++
Documentation/ptp/testptp.c | 245 +++++++++++
Documentation/ptp/testptp.mk | 33 ++
arch/powerpc/boot/dts/mpc8313erdb.dts | 14 +
arch/powerpc/boot/dts/p2020ds.dts | 14 +
arch/powerpc/boot/dts/p2020rdb.dts | 14 +
drivers/Kconfig | 2 +
drivers/Makefile | 1 +
drivers/net/Makefile | 1 +
drivers/net/gianfar_ptp.c | 521 +++++++++++++++++++++++
drivers/net/gianfar_ptp_reg.h | 113 +++++
drivers/ptp/Kconfig | 51 +++
drivers/ptp/Makefile | 6 +
drivers/ptp/ptp_clock.c | 512 ++++++++++++++++++++++
drivers/ptp/ptp_linux.c | 136 ++++++
include/linux/Kbuild | 1 +
include/linux/ptp_clock.h | 79 ++++
include/linux/ptp_clock_kernel.h | 137 ++++++
kernel/time/ntp.c | 2 +
20 files changed, 2033 insertions(+), 0 deletions(-)
create mode 100644 Documentation/ptp/ptp.txt
create mode 100644 Documentation/ptp/testptp.c
create mode 100644 Documentation/ptp/testptp.mk
create mode 100644 drivers/net/gianfar_ptp.c
create mode 100644 drivers/net/gianfar_ptp_reg.h
create mode 100644 drivers/ptp/Kconfig
create mode 100644 drivers/ptp/Makefile
create mode 100644 drivers/ptp/ptp_clock.c
create mode 100644 drivers/ptp/ptp_linux.c
create mode 100644 include/linux/ptp_clock.h
create mode 100644 include/linux/ptp_clock_kernel.h
^ permalink raw reply
* [PATCH v3 1/3] ptp: Added a brand new class driver for ptp clocks.
From: Richard Cochran @ 2010-05-14 16:45 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss
In-Reply-To: <cover.1273855016.git.richard.cochran@omicron.at>
This patch adds an infrastructure for hardware clocks that implement
IEEE 1588, the Precision Time Protocol (PTP). A class driver offers a
registration method to particular hardware clock drivers. Each clock is
exposed to user space as a character device with ioctls that allow tuning
of the PTP clock.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
Documentation/ptp/ptp.txt | 95 +++++++
Documentation/ptp/testptp.c | 245 ++++++++++++++++++
Documentation/ptp/testptp.mk | 33 +++
drivers/Kconfig | 2 +
drivers/Makefile | 1 +
drivers/ptp/Kconfig | 26 ++
drivers/ptp/Makefile | 5 +
drivers/ptp/ptp_clock.c | 512 ++++++++++++++++++++++++++++++++++++++
include/linux/Kbuild | 1 +
include/linux/ptp_clock.h | 79 ++++++
include/linux/ptp_clock_kernel.h | 137 ++++++++++
11 files changed, 1136 insertions(+), 0 deletions(-)
create mode 100644 Documentation/ptp/ptp.txt
create mode 100644 Documentation/ptp/testptp.c
create mode 100644 Documentation/ptp/testptp.mk
create mode 100644 drivers/ptp/Kconfig
create mode 100644 drivers/ptp/Makefile
create mode 100644 drivers/ptp/ptp_clock.c
create mode 100644 include/linux/ptp_clock.h
create mode 100644 include/linux/ptp_clock_kernel.h
diff --git a/Documentation/ptp/ptp.txt b/Documentation/ptp/ptp.txt
new file mode 100644
index 0000000..46858b3
--- /dev/null
+++ b/Documentation/ptp/ptp.txt
@@ -0,0 +1,95 @@
+
+* PTP infrastructure for Linux
+
+ This patch set introduces support for IEEE 1588 PTP clocks in
+ Linux. Together with the SO_TIMESTAMPING socket options, this
+ presents a standardized method for developing PTP user space
+ programs, synchronizing Linux with external clocks, and using the
+ ancillary features of PTP hardware clocks.
+
+ A new class driver exports a kernel interface for specific clock
+ drivers and a user space interface. The infrastructure supports a
+ complete set of PTP functionality.
+
+ + Basic clock operations
+ - Set time
+ - Get time
+ - Shift the clock by a given offset atomically
+ - Adjust clock frequency
+
+ + Ancillary clock features
+ - One short or periodic alarms, with signal delivery to user program
+ - Time stamp external events
+ - Period output signals configurable from user space
+ - Synchronization of the Linux system time via the PPS subsystem
+
+** PTP kernel API
+
+ A PTP clock driver registers itself with the class driver. The
+ class driver handles all of the dealings with user space. The
+ author of a clock driver need only implement the details of
+ programming the clock hardware. The clock driver notifies the class
+ driver of asynchronous events (alarms and external time stamps) via
+ a simple message passing interface.
+
+ The class driver supports multiple PTP clock drivers. In normal use
+ cases, only one PTP clock is needed. However, for testing and
+ development, it can be useful to have more than one clock in a
+ single system, in order to allow performance comparisons.
+
+** PTP user space API
+
+ The class driver creates a character device for each registered PTP
+ clock. User space programs may control the clock using standardized
+ ioctls. A program may query, enable, configure, and disable the
+ ancillary clock features. User space can receive time stamped
+ events via blocking read() and poll(). One shot and periodic
+ signals may be configured via an ioctl API with semantics similar
+ to the POSIX timer_settime() system call.
+
+ As an real life example, the following two patches for ptpd version
+ 1.0.0 demonstrate how the API works.
+
+ https://sourceforge.net/tracker/?func=detail&aid=2992845&group_id=139814&atid=744634
+
+ https://sourceforge.net/tracker/?func=detail&aid=2992847&group_id=139814&atid=744634
+
+** Writing clock drivers
+
+ Clock drivers include include/linux/ptp_clock_kernel.h and register
+ themselves by presenting a 'struct ptp_clock_info' to the
+ registration method. Clock drivers must implement all of the
+ functions in the interface. If a clock does not offer a particular
+ ancillary feature, then the driver should just return -EOPNOTSUPP
+ from those functions.
+
+ Drivers must ensure that all of the methods in interface are
+ reentrant. Since most hardware implementations treat the time value
+ as a 64 bit integer accessed as two 32 bit registers, drivers
+ should use spin_lock_irqsave/spin_unlock_irqrestore to protect
+ against concurrent access. This locking cannot be accomplished in
+ class driver, since the lock may also be needed by the clock
+ driver's interrupt service routine.
+
+** Supported hardware
+
+ + Standard Linux system timer
+ - No special PTP features
+ - For use with software time stamping
+
+ + Freescale eTSEC gianfar
+ - 2 Time stamp external triggers, programmable polarity (opt. interrupt)
+ - 2 Alarm registers (optional interrupt)
+ - 3 Periodic signals (optional interrupt)
+
+ + National DP83640
+ - 6 GPIOs programmable as inputs or outputs
+ - 6 GPIOs with dedicated functions (LED/JTAG/clock) can also be
+ used as general inputs or outputs
+ - GPIO inputs can time stamp external triggers
+ - GPIO outputs can produce periodic signals
+ - 1 interrupt pin
+
+ + Intel IXP465
+ - Auxiliary Slave/Master Mode Snapshot (optional interrupt)
+ - Target Time (optional interrupt)
diff --git a/Documentation/ptp/testptp.c b/Documentation/ptp/testptp.c
new file mode 100644
index 0000000..ed2ceea
--- /dev/null
+++ b/Documentation/ptp/testptp.c
@@ -0,0 +1,245 @@
+/*
+ * PTP 1588 clock support - User space test program
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <errno.h>
+#include <fcntl.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <time.h>
+#include <unistd.h>
+
+#include <linux/ptp_clock.h>
+
+static void handle_alarm(int s)
+{
+ printf("received signal %d \n", s);
+}
+
+static int install_handler(int signum, void (*handler)(int))
+{
+ struct sigaction action;
+ sigset_t mask;
+
+ /* Unblock the signal. */
+ sigemptyset(&mask);
+ sigaddset(&mask, signum);
+ sigprocmask(SIG_UNBLOCK, &mask, NULL);
+
+ /* Install the signal handler. */
+ action.sa_handler = handler;
+ action.sa_flags = 0;
+ sigemptyset(&action.sa_mask);
+ sigaction(signum, &action, NULL);
+
+ return 0;
+}
+
+static void usage (char* progname)
+{
+ fprintf(stderr,
+ "usage: %s [options] \n"
+ " -a val request a one-shot alarm after 'val' seconds \n"
+ " -A val request a periodic alarm every 'val' seconds \n"
+ " -c query the ptp clock's capabilities \n"
+ " -d name device to open \n"
+ " -e val read 'val' external time stamp events \n"
+ " -f val adjust the ptp clock frequency by 'val' PPB \n"
+ " -g get the ptp clock time \n"
+ " -h prints this message \n"
+ " -s set the ptp clock time from the system time \n"
+ " -t val shift the ptp clock time by 'val' seconds \n"
+ " -v query the ptp clock api version \n"
+ ,progname);
+}
+
+int main(int argc,char *argv[])
+{
+ struct ptp_clock_caps caps;
+ struct ptp_clock_timer timer;
+ struct ptp_extts_event event;
+ struct ptp_clock_request request;
+ struct timespec ts;
+ char *progname;
+ int c, cnt, fd, val=0;
+
+ char *device = "/dev/ptp_clock_0";
+ int adjfreq=0x7fffffff;
+ int adjtime=0;
+ int capabilities=0;
+ int extts=0;
+ int gettime=0;
+ int oneshot=0;
+ int periodic=0;
+ int settime=0;
+ int version=0;
+
+ progname = strrchr(argv[0],'/');
+ progname = progname ? 1+progname : argv[0];
+ while (EOF != (c = getopt(argc, argv, "a:A:cd:e:f:ghst:v"))) {
+ switch (c) {
+ case 'a': oneshot = atoi(optarg); break;
+ case 'A': periodic = atoi(optarg); break;
+ case 'c': capabilities = 1; break;
+ case 'd': device = optarg; break;
+ case 'e': extts = atoi(optarg); break;
+ case 'f': adjfreq = atoi(optarg); break;
+ case 'g': gettime = 1; break;
+ case 's': settime = 1; break;
+ case 't': adjtime = atoi(optarg); break;
+ case 'v': version = 1; break;
+ case 'h': usage(progname); return 0;
+ case '?': usage(progname); return -1;
+ default: usage(progname); return -1;
+ }
+ }
+
+ fd = open(device, O_RDWR);
+ if (fd < 0) {
+ fprintf(stderr,"cannot open %s: %s", device, strerror(errno));
+ return -1;
+ }
+
+ if (version) {
+ if (ioctl(fd, PTP_CLOCK_APIVERS, &val)) {
+ perror("PTP_CLOCK_APIVERS");
+ } else {
+ printf("version = 0x%08x \n",val);
+ }
+ }
+
+ if (capabilities) {
+ if (ioctl(fd, PTP_CLOCK_GETCAPS, &caps)) {
+ perror("PTP_CLOCK_GETCAPS");
+ } else {
+ printf("capabilities: \n"
+ " %d maximum frequency adjustment (PPB) \n"
+ " %d programmable alarms \n"
+ " %d external time stamp channels \n"
+ " %d programmable periodic signals \n"
+ " %d pulse per second \n",
+ caps.max_adj,
+ caps.n_alarm,
+ caps.n_ext_ts,
+ caps.n_per_out,
+ caps.pps);
+ }
+ }
+
+ if (0x7fffffff != adjfreq) {
+ if (ioctl(fd, PTP_CLOCK_ADJFREQ, adjfreq)) {
+ perror("PTP_CLOCK_ADJFREQ");
+ } else {
+ puts("frequency adjustment okay");
+ }
+ }
+
+ if (adjtime) {
+ ts.tv_sec = adjtime;
+ ts.tv_nsec = 0;
+ if (ioctl(fd, PTP_CLOCK_ADJTIME, &ts)) {
+ perror("PTP_CLOCK_ADJTIME");
+ } else {
+ puts("time shift okay");
+ }
+ }
+
+ if (gettime) {
+ if (ioctl(fd, PTP_CLOCK_GETTIME, &ts)) {
+ perror("PTP_CLOCK_GETTIME");
+ } else {
+ printf("clock time: %ld.%09ld or %s",
+ ts.tv_sec, ts.tv_nsec, ctime(&ts.tv_sec));
+ }
+ }
+
+ if (settime) {
+ clock_gettime(CLOCK_REALTIME, &ts);
+ if (ioctl(fd, PTP_CLOCK_SETTIME, &ts)) {
+ perror("PTP_CLOCK_SETTIME");
+ } else {
+ puts("set time okay");
+ }
+ }
+
+ if (extts) {
+ memset(&request, 0, sizeof(request));
+ request.type = PTP_REQUEST_EXTTS;
+ request.index = 0;
+ request.flags = PTP_ENABLE_FEATURE;
+ if (ioctl(fd, PTP_FEATURE_REQUEST, &request)) {
+ perror("PTP_FEATURE_REQUEST");
+ extts = 0;
+ } else {
+ puts("set timer okay");
+ }
+ for (; extts; extts--) {
+ cnt = read(fd, &event, sizeof(event));
+ if (cnt != sizeof(event)) {
+ perror("read");
+ break;
+ }
+ printf("event index %d at %ld.%09ld \n", event.index,
+ event.ts.tv_sec, event.ts.tv_nsec);
+ }
+ /* Disable the feature again. */
+ request.flags = 0;
+ if (ioctl(fd, PTP_FEATURE_REQUEST, &request)) {
+ perror("PTP_FEATURE_REQUEST");
+ }
+ }
+
+ if (oneshot) {
+ install_handler(SIGALRM, handle_alarm);
+ memset(&timer, 0, sizeof(timer));
+ timer.signum = SIGALRM;
+ timer.tsp.it_value.tv_sec = oneshot;
+ if (ioctl(fd, PTP_CLOCK_SETTIMER, &timer)) {
+ perror("PTP_CLOCK_SETTIMER");
+ } else {
+ puts("set timer okay");
+ }
+ pause();
+ }
+
+ if (periodic) {
+ install_handler(SIGALRM, handle_alarm);
+ memset(&timer, 0, sizeof(timer));
+ timer.signum = SIGALRM;
+ timer.tsp.it_value.tv_sec = periodic;
+ timer.tsp.it_interval.tv_sec = periodic;
+ if (ioctl(fd, PTP_CLOCK_SETTIMER, &timer)) {
+ perror("PTP_CLOCK_SETTIMER");
+ } else {
+ puts("set timer okay");
+ }
+ while (1) {
+ pause();
+ }
+ }
+
+ close(fd);
+ return 0;
+}
diff --git a/Documentation/ptp/testptp.mk b/Documentation/ptp/testptp.mk
new file mode 100644
index 0000000..4ef2d97
--- /dev/null
+++ b/Documentation/ptp/testptp.mk
@@ -0,0 +1,33 @@
+# PTP 1588 clock support - User space test program
+#
+# Copyright (C) 2010 OMICRON electronics GmbH
+#
+# 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.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+CC = $(CROSS_COMPILE)gcc
+INC = -I$(KBUILD_OUTPUT)/usr/include
+CFLAGS = -Wall $(INC)
+LDLIBS = -lrt
+PROGS = testptp
+
+all: $(PROGS)
+
+testptp: testptp.o
+
+clean:
+ rm -f testptp.o
+
+distclean: clean
+ rm -f $(PROGS)
diff --git a/drivers/Kconfig b/drivers/Kconfig
index a2b902f..774fbd7 100644
--- a/drivers/Kconfig
+++ b/drivers/Kconfig
@@ -52,6 +52,8 @@ source "drivers/spi/Kconfig"
source "drivers/pps/Kconfig"
+source "drivers/ptp/Kconfig"
+
source "drivers/gpio/Kconfig"
source "drivers/w1/Kconfig"
diff --git a/drivers/Makefile b/drivers/Makefile
index f42a030..84228bc 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -75,6 +75,7 @@ obj-$(CONFIG_I2O) += message/
obj-$(CONFIG_RTC_LIB) += rtc/
obj-y += i2c/ media/
obj-$(CONFIG_PPS) += pps/
+obj-$(CONFIG_PTP_1588_CLOCK) += ptp/
obj-$(CONFIG_W1) += w1/
obj-$(CONFIG_POWER_SUPPLY) += power/
obj-$(CONFIG_HWMON) += hwmon/
diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
new file mode 100644
index 0000000..c80a25b
--- /dev/null
+++ b/drivers/ptp/Kconfig
@@ -0,0 +1,26 @@
+#
+# PTP clock support configuration
+#
+
+menu "PTP clock support"
+
+config PTP_1588_CLOCK
+ tristate "PTP clock support"
+ depends on EXPERIMENTAL
+ help
+ The IEEE 1588 standard defines a method to precisely
+ synchronize distributed clocks over Ethernet networks. The
+ standard defines a Precision Time Protocol (PTP), which can
+ be used to achieve synchronization within a few dozen
+ microseconds. In addition, with the help of special hardware
+ time stamping units, it can be possible to achieve
+ synchronization to within a few hundred nanoseconds.
+
+ This driver adds support for PTP clocks as character
+ devices. If you want to use a PTP clock, then you should
+ also enable at least one clock driver as well.
+
+ To compile this driver as a module, choose M here: the module
+ will be called ptp_clock.
+
+endmenu
diff --git a/drivers/ptp/Makefile b/drivers/ptp/Makefile
new file mode 100644
index 0000000..b86695c
--- /dev/null
+++ b/drivers/ptp/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for PTP 1588 clock support.
+#
+
+obj-$(CONFIG_PTP_1588_CLOCK) += ptp_clock.o
diff --git a/drivers/ptp/ptp_clock.c b/drivers/ptp/ptp_clock.c
new file mode 100644
index 0000000..1d94d31
--- /dev/null
+++ b/drivers/ptp/ptp_clock.c
@@ -0,0 +1,512 @@
+/*
+ * PTP 1588 clock support
+ *
+ * Partially adapted from the Linux PPS driver.
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/bitops.h>
+#include <linux/cdev.h>
+#include <linux/device.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/poll.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+
+#include <linux/ptp_clock_kernel.h>
+#include <linux/ptp_clock.h>
+
+#define PTP_MAX_ALARMS 4
+#define PTP_MAX_CLOCKS BITS_PER_LONG
+#define PTP_MAX_TIMESTAMPS 128
+
+struct alarm {
+ struct pid *pid;
+ int sig;
+};
+
+struct timestamp_event_queue {
+ struct ptp_extts_event buf[PTP_MAX_TIMESTAMPS];
+ int head;
+ int tail;
+ int overflow;
+};
+
+struct ptp_clock {
+ struct list_head list;
+ struct cdev cdev;
+ struct device *dev;
+ struct ptp_clock_info *info;
+ dev_t devid;
+ int index; /* index into clocks.map, also the minor number */
+
+ struct alarm alarm[PTP_MAX_ALARMS];
+ struct mutex alarm_mux; /* one process at a time setting an alarm */
+
+ struct timestamp_event_queue tsevq; /* simple fifo for time stamps */
+ struct mutex tsevq_mux; /* one process at a time reading the fifo */
+ wait_queue_head_t tsev_wq;
+};
+
+/* private globals */
+
+static const struct file_operations ptp_fops;
+static dev_t ptp_devt;
+static struct class *ptp_class;
+
+static struct {
+ struct list_head list;
+ DECLARE_BITMAP(map, PTP_MAX_CLOCKS);
+} clocks;
+static DEFINE_SPINLOCK(clocks_lock); /* protects 'clocks' */
+
+/* time stamp event queue operations */
+
+static inline int queue_cnt(struct timestamp_event_queue *q)
+{
+ int cnt = q->tail - q->head;
+ return cnt < 0 ? PTP_MAX_TIMESTAMPS + cnt : cnt;
+}
+
+static inline int queue_free(struct timestamp_event_queue *q)
+{
+ return PTP_MAX_TIMESTAMPS - queue_cnt(q) - 1;
+}
+
+static void enqueue_external_timestamp(struct timestamp_event_queue *queue,
+ struct ptp_clock_event *src)
+{
+ struct ptp_extts_event *dst;
+ u32 remainder;
+
+ dst = &queue->buf[ queue->tail ];
+
+ dst->index = src->index;
+ dst->ts.tv_sec = div_u64_rem(src->timestamp, 1000000000, &remainder);
+ dst->ts.tv_nsec = remainder;
+
+ if (!queue_free(queue))
+ queue->overflow++;
+
+ queue->tail = (queue->tail + 1) % PTP_MAX_TIMESTAMPS;
+}
+
+/* public interface */
+
+struct ptp_clock* ptp_clock_register(struct ptp_clock_info *info)
+{
+ struct ptp_clock *ptp;
+ int err = 0, index, major = MAJOR(ptp_devt);
+ unsigned long flags;
+
+ if (info->n_alarm > PTP_MAX_ALARMS)
+ return ERR_PTR(-EINVAL);
+
+ /* Find a free clock slot and reserve it. */
+ err = -EBUSY;
+ spin_lock_irqsave(&clocks_lock, flags);
+ index = find_first_zero_bit(clocks.map, PTP_MAX_CLOCKS);
+ if (index < PTP_MAX_CLOCKS) {
+ set_bit(index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+ } else {
+ spin_unlock_irqrestore(&clocks_lock, flags);
+ goto no_clock;
+ }
+
+ /* Initialize a clock structure. */
+ err = -ENOMEM;
+ ptp = kzalloc(sizeof(struct ptp_clock), GFP_KERNEL);
+ if (ptp == NULL)
+ goto no_memory;
+
+ ptp->info = info;
+ ptp->devid = MKDEV(major, index);
+ ptp->index = index;
+ mutex_init(&ptp->alarm_mux);
+ mutex_init(&ptp->tsevq_mux);
+ init_waitqueue_head(&ptp->tsev_wq);
+
+ /* Create a new device in our class. */
+ ptp->dev = device_create(ptp_class, NULL, ptp->devid, ptp,
+ "ptp_clock_%d", ptp->index);
+ if (IS_ERR(ptp->dev))
+ goto no_device;
+
+ dev_set_drvdata(ptp->dev, ptp);
+
+ /* Register a character device. */
+ cdev_init(&ptp->cdev, &ptp_fops);
+ ptp->cdev.owner = info->owner;
+ err = cdev_add(&ptp->cdev, ptp->devid, 1);
+ if (err)
+ goto no_cdev;
+
+ /* Clock is ready, add it into the list. */
+ spin_lock_irqsave(&clocks_lock, flags);
+ list_add(&ptp->list, &clocks.list);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+
+ return ptp;
+
+no_cdev:
+ device_destroy(ptp_class, ptp->devid);
+no_device:
+ mutex_destroy(&ptp->alarm_mux);
+ mutex_destroy(&ptp->tsevq_mux);
+ kfree(ptp);
+no_memory:
+ spin_lock_irqsave(&clocks_lock, flags);
+ clear_bit(index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+no_clock:
+ return ERR_PTR(err);
+}
+EXPORT_SYMBOL(ptp_clock_register);
+
+int ptp_clock_unregister(struct ptp_clock *ptp)
+{
+ unsigned long flags;
+
+ /* Release the clock's resources. */
+ cdev_del(&ptp->cdev);
+ device_destroy(ptp_class, ptp->devid);
+ mutex_destroy(&ptp->alarm_mux);
+ mutex_destroy(&ptp->tsevq_mux);
+
+ /* Remove the clock from the list. */
+ spin_lock_irqsave(&clocks_lock, flags);
+ list_del(&ptp->list);
+ clear_bit(ptp->index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+
+ kfree(ptp);
+
+ return 0;
+}
+EXPORT_SYMBOL(ptp_clock_unregister);
+
+void ptp_clock_event(struct ptp_clock *ptp, struct ptp_clock_event *event)
+{
+ switch (event->type) {
+
+ case PTP_CLOCK_ALARM:
+ kill_pid(ptp->alarm[ event->index ].pid,
+ ptp->alarm[ event->index ].sig, 1);
+ break;
+
+ case PTP_CLOCK_EXTTS:
+ enqueue_external_timestamp(&ptp->tsevq, event);
+ wake_up_interruptible(&ptp->tsev_wq);
+ break;
+
+ case PTP_CLOCK_PPS:
+ break;
+ }
+}
+EXPORT_SYMBOL(ptp_clock_event);
+
+/* character device operations */
+
+static int ptp_ioctl(struct inode *node, struct file *fp,
+ unsigned int cmd, unsigned long arg)
+{
+ struct ptp_clock_caps caps;
+ struct ptp_clock_request req;
+ struct ptp_clock_timer timer;
+ struct ptp_clock *ptp = fp->private_data;
+ struct ptp_clock_info *ops = ptp->info;
+ void *priv = ops->priv;
+ struct timespec ts;
+ int flags, index;
+ int err = 0;
+
+ switch (cmd) {
+
+ case PTP_CLOCK_APIVERS:
+ err = put_user(PTP_CLOCK_VERSION, (u32 __user*)arg);
+ break;
+
+ case PTP_CLOCK_ADJFREQ:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ err = ops->adjfreq(priv, arg);
+ break;
+
+ case PTP_CLOCK_ADJTIME:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ if (copy_from_user(&ts, (void __user*)arg, sizeof(ts)))
+ err = -EFAULT;
+ else
+ err = ops->adjtime(priv, &ts);
+ break;
+
+ case PTP_CLOCK_GETTIME:
+ err = ops->gettime(priv, &ts);
+ if (err)
+ break;
+ err = copy_to_user((void __user*)arg, &ts, sizeof(ts));
+ break;
+
+ case PTP_CLOCK_SETTIME:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ if (copy_from_user(&ts, (void __user*)arg, sizeof(ts)))
+ err = -EFAULT;
+ else
+ err = ops->settime(priv, &ts);
+ break;
+
+ case PTP_CLOCK_GETCAPS:
+ memset(&caps, 0, sizeof(caps));
+ caps.max_adj = ptp->info->max_adj;
+ caps.n_alarm = ptp->info->n_alarm;
+ caps.n_ext_ts = ptp->info->n_ext_ts;
+ caps.n_per_out = ptp->info->n_per_out;
+ caps.pps = ptp->info->pps;
+ err = copy_to_user((void __user*)arg, &caps, sizeof(caps));
+ break;
+
+ case PTP_CLOCK_GETTIMER:
+ if (copy_from_user(&timer, (void __user*)arg, sizeof(timer))) {
+ err = -EFAULT;
+ break;
+ }
+ index = timer.alarm_index;
+ if (index < 0 || index >= ptp->info->n_alarm) {
+ err = -EINVAL;
+ break;
+ }
+ err = ops->gettimer(priv, index, &timer.tsp);
+ if (err)
+ break;
+ err = copy_to_user((void __user*)arg, &timer, sizeof(timer));
+ break;
+
+ case PTP_CLOCK_SETTIMER:
+ if (copy_from_user(&timer, (void __user*)arg, sizeof(timer))) {
+ err = -EFAULT;
+ break;
+ }
+ index = timer.alarm_index;
+ if (index < 0 || index >= ptp->info->n_alarm) {
+ err = -EINVAL;
+ break;
+ }
+ if (!valid_signal(timer.signum))
+ return -EINVAL;
+ flags = timer.flags;
+ if (flags & (flags != TIMER_ABSTIME)) {
+ err = -EINVAL;
+ break;
+ }
+ if (mutex_lock_interruptible(&ptp->alarm_mux))
+ return -ERESTARTSYS;
+
+ if (ptp->alarm[index].pid)
+ put_pid(ptp->alarm[index].pid);
+
+ ptp->alarm[index].pid = get_pid(task_pid(current));
+ ptp->alarm[index].sig = timer.signum;
+ err = ops->settimer(priv, index, flags, &timer.tsp);
+
+ mutex_unlock(&ptp->alarm_mux);
+ break;
+
+ case PTP_FEATURE_REQUEST:
+ if (copy_from_user(&req, (void __user*)arg, sizeof(req))) {
+ err = -EFAULT;
+ break;
+ }
+ switch (req.type) {
+ case PTP_REQUEST_EXTTS:
+ case PTP_REQUEST_PEROUT:
+ break;
+ case PTP_REQUEST_PPS:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ break;
+ default:
+ err = -EINVAL;
+ break;
+ }
+ if (err)
+ break;
+ err = ops->enable(priv, &req,
+ req.flags & PTP_ENABLE_FEATURE ? 1 : 0);
+ break;
+
+ default:
+ err = -ENOTTY;
+ break;
+ }
+ return err;
+}
+
+static int ptp_open(struct inode *inode, struct file *fp)
+{
+ struct ptp_clock *ptp;
+ ptp = container_of(inode->i_cdev, struct ptp_clock, cdev);
+
+ fp->private_data = ptp;
+
+ return 0;
+}
+
+static unsigned int ptp_poll(struct file *fp, poll_table *wait)
+{
+ struct ptp_clock *ptp = fp->private_data;
+
+ poll_wait(fp, &ptp->tsev_wq, wait);
+
+ return queue_cnt(&ptp->tsevq) ? POLLIN : 0;
+}
+
+static ssize_t ptp_read(struct file *fp, char __user *buf,
+ size_t cnt, loff_t *off)
+{
+ struct ptp_clock *ptp = fp->private_data;
+ struct timestamp_event_queue *queue = &ptp->tsevq;
+ struct ptp_extts_event *event;
+ size_t qcnt;
+
+ if (mutex_lock_interruptible(&ptp->tsevq_mux))
+ return -ERESTARTSYS;
+
+ cnt = cnt / sizeof(struct ptp_extts_event);
+
+ if (wait_event_interruptible(ptp->tsev_wq,
+ (qcnt = queue_cnt(&ptp->tsevq)))) {
+ mutex_unlock(&ptp->tsevq_mux);
+ return -ERESTARTSYS;
+ }
+
+ if (cnt > qcnt)
+ cnt = qcnt;
+
+ event = &queue->buf[ queue->head ];
+
+ if (copy_to_user(buf, event, cnt * sizeof(struct ptp_extts_event))) {
+ mutex_unlock(&ptp->tsevq_mux);
+ return -EFAULT;
+ }
+ queue->head = (queue->head + cnt) % PTP_MAX_TIMESTAMPS;
+
+ mutex_unlock(&ptp->tsevq_mux);
+
+ return cnt * sizeof(struct ptp_extts_event);
+}
+
+static int ptp_release(struct inode *inode, struct file *fp)
+{
+ struct ptp_clock *ptp;
+ struct itimerspec ts = {{0,0},{0,0}};
+ int i;
+
+ ptp = container_of(inode->i_cdev, struct ptp_clock, cdev);
+
+ for (i = 0; i < ptp->info->n_alarm; i++) {
+ if (ptp->alarm[i].pid) {
+ ptp->info->settimer(ptp->info->priv, i, 0, &ts);
+ put_pid(ptp->alarm[i].pid);
+ ptp->alarm[i].pid = NULL;
+ }
+ }
+ return 0;
+}
+
+static const struct file_operations ptp_fops = {
+ .owner = THIS_MODULE,
+ .ioctl = ptp_ioctl,
+ .open = ptp_open,
+ .poll = ptp_poll,
+ .read = ptp_read,
+ .release = ptp_release,
+};
+
+/* sysfs */
+
+static ssize_t ptp_show_status(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct ptp_clock *ptp = dev_get_drvdata(dev);
+ return sprintf(buf,
+ "maximum adjustment: %d\n"
+ "programmable alarms: %d\n"
+ "external timestamps: %d\n"
+ "periodic outputs: %d\n"
+ "has pps: %d\n"
+ "device index: %d\n"
+ ,ptp->info->max_adj
+ ,ptp->info->n_alarm
+ ,ptp->info->n_ext_ts
+ ,ptp->info->n_per_out
+ ,ptp->info->pps
+ ,ptp->index);
+}
+
+struct device_attribute ptp_attrs[] = {
+ __ATTR(capabilities, S_IRUGO, ptp_show_status, NULL),
+ __ATTR_NULL,
+};
+
+/* module operations */
+
+static void __exit ptp_exit(void)
+{
+ class_destroy(ptp_class);
+ unregister_chrdev_region(ptp_devt, PTP_MAX_CLOCKS);
+}
+
+static int __init ptp_init(void)
+{
+ int err;
+
+ INIT_LIST_HEAD(&clocks.list);
+
+ ptp_class = class_create(THIS_MODULE, "ptp");
+ if (!ptp_class) {
+ printk(KERN_ERR "ptp: failed to allocate class\n");
+ return -ENOMEM;
+ }
+ ptp_class->dev_attrs = ptp_attrs;
+
+ err = alloc_chrdev_region(&ptp_devt, 0, PTP_MAX_CLOCKS, "ptp");
+ if (err < 0) {
+ printk(KERN_ERR "ptp: failed to allocate char device region\n");
+ goto no_region;
+ }
+
+ pr_info("PTP clock support registered\n");
+ return 0;
+
+no_region:
+ class_destroy(ptp_class);
+ return err;
+}
+
+subsys_initcall(ptp_init);
+module_exit(ptp_exit);
+
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_DESCRIPTION("PTP clocks support");
+MODULE_LICENSE("GPL");
diff --git a/include/linux/Kbuild b/include/linux/Kbuild
index 2fc8e14..9959fe4 100644
--- a/include/linux/Kbuild
+++ b/include/linux/Kbuild
@@ -140,6 +140,7 @@ header-y += pkt_sched.h
header-y += posix_types.h
header-y += ppdev.h
header-y += prctl.h
+header-y += ptp_clock.h
header-y += qnxtypes.h
header-y += qnx4_fs.h
header-y += radeonfb.h
diff --git a/include/linux/ptp_clock.h b/include/linux/ptp_clock.h
new file mode 100644
index 0000000..5a509c5
--- /dev/null
+++ b/include/linux/ptp_clock.h
@@ -0,0 +1,79 @@
+/*
+ * PTP 1588 clock support - user space interface
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _PTP_CLOCK_H_
+#define _PTP_CLOCK_H_
+
+#include <linux/ioctl.h>
+#include <linux/types.h>
+
+#define PTP_ENABLE_FEATURE (1<<0)
+#define PTP_RISING_EDGE (1<<1)
+#define PTP_FALLING_EDGE (1<<2)
+
+enum ptp_request_types {
+ PTP_REQUEST_EXTTS,
+ PTP_REQUEST_PEROUT,
+ PTP_REQUEST_PPS,
+};
+
+struct ptp_clock_caps {
+ __s32 max_adj; /* Maximum frequency adjustment, parts per billon. */
+ int n_alarm; /* Number of programmable alarms. */
+ int n_ext_ts; /* Number of external time stamp channels. */
+ int n_per_out; /* Number of programmable periodic signals. */
+ int pps; /* Whether the clock supports a PPS callback. */
+};
+
+struct ptp_clock_timer {
+ int alarm_index; /* Which alarm to query or configure. */
+ int signum; /* Requested signal. */
+ int flags; /* Zero or TIMER_ABSTIME, see TIMER_SETTIME(2) */
+ struct itimerspec tsp; /* See TIMER_SETTIME(2) */
+};
+
+struct ptp_clock_request {
+ int type; /* One of the ptp_request_types enumeration values. */
+ int index; /* Which channel to configure. */
+ struct timespec ts; /* For period signals, the desired period. */
+ int flags; /* Bit field for PTP_ENABLE_FEATURE or other flags. */
+};
+
+struct ptp_extts_event {
+ int index;
+ struct timespec ts;
+};
+
+#define PTP_CLOCK_VERSION 0x00000001
+
+#define PTP_CLK_MAGIC '='
+
+#define PTP_CLOCK_APIVERS _IOR (PTP_CLK_MAGIC, 1, __u32)
+#define PTP_CLOCK_ADJFREQ _IO (PTP_CLK_MAGIC, 2)
+#define PTP_CLOCK_ADJTIME _IOW (PTP_CLK_MAGIC, 3, struct timespec)
+#define PTP_CLOCK_GETTIME _IOR (PTP_CLK_MAGIC, 4, struct timespec)
+#define PTP_CLOCK_SETTIME _IOW (PTP_CLK_MAGIC, 5, struct timespec)
+
+#define PTP_CLOCK_GETCAPS _IOR (PTP_CLK_MAGIC, 6, struct ptp_clock_caps)
+#define PTP_CLOCK_GETTIMER _IOWR (PTP_CLK_MAGIC, 7, struct ptp_clock_timer)
+#define PTP_CLOCK_SETTIMER _IOW (PTP_CLK_MAGIC, 8, struct ptp_clock_timer)
+#define PTP_FEATURE_REQUEST _IOW (PTP_CLK_MAGIC, 9, struct ptp_clock_request)
+
+#endif
diff --git a/include/linux/ptp_clock_kernel.h b/include/linux/ptp_clock_kernel.h
new file mode 100644
index 0000000..b1fb2a7
--- /dev/null
+++ b/include/linux/ptp_clock_kernel.h
@@ -0,0 +1,137 @@
+/*
+ * PTP 1588 clock support
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _PTP_CLOCK_KERNEL_H_
+#define _PTP_CLOCK_KERNEL_H_
+
+#include <linux/ptp_clock.h>
+
+/**
+ * struct ptp_clock_info - decribes a PTP hardware clock
+ *
+ * @owner: The clock driver should set to THIS_MODULE.
+ * @name: A short name to identify the clock.
+ * @max_adj: The maximum possible frequency adjustment, in parts per billon.
+ * @n_alarm: The number of programmable alarms.
+ * @n_ext_ts: The number of external time stamp channels.
+ * @n_per_out: The number of programmable periodic signals.
+ * @pps: Indicates whether the clock supports a PPS callback.
+ * @priv: Passed to the clock operations, for the driver's private use.
+ *
+ * clock operations
+ *
+ * @adjfreq: Adjusts the frequency of the hardware clock.
+ * parameter delta: Desired period change in parts per billion.
+ *
+ * @adjtime: Shifts the time of the hardware clock.
+ * parameter ts: Desired change in seconds and nanoseconds.
+ *
+ * @gettime: Reads the current time from the hardware clock.
+ * parameter ts: Holds the result.
+ *
+ * @settime: Set the current time on the hardware clock.
+ * parameter ts: Time value to set.
+ *
+ * @gettimer: Reads the time remaining from the given timer.
+ * parameter index: Which alarm to query.
+ * parameter ts: Holds the result.
+ *
+ * @settimer: Arms the given timer for periodic or one shot operation.
+ * parameter index: Which alarm to set.
+ * parameter abs: TIMER_ABSTIME, or zero for relative timer.
+ * parameter ts: Alarm time and period to set.
+ *
+ * @enable: Request driver to enable or disable an ancillary feature.
+ * parameter request: Desired resource to enable or disable.
+ * parameter on: Caller passes one to enable or zero to disable.
+ *
+ * The callbacks must all return zero on success, non-zero otherwise.
+ */
+
+struct ptp_clock_info {
+ struct module *owner;
+ char name[16];
+ s32 max_adj;
+ int n_alarm;
+ int n_ext_ts;
+ int n_per_out;
+ int pps;
+ void *priv;
+ int (*adjfreq)(void *priv, s32 delta);
+ int (*adjtime)(void *priv, struct timespec *ts);
+ int (*gettime)(void *priv, struct timespec *ts);
+ int (*settime)(void *priv, struct timespec *ts);
+ int (*gettimer)(void *priv, int index, struct itimerspec *ts);
+ int (*settimer)(void *priv, int index, int abs, struct itimerspec *ts);
+ int (*enable)(void *priv, struct ptp_clock_request *request, int on);
+};
+
+struct ptp_clock;
+
+/**
+ * ptp_clock_register() - register a PTP hardware clock driver
+ *
+ * @info: Structure describing the new clock.
+ */
+
+extern struct ptp_clock* ptp_clock_register(struct ptp_clock_info *info);
+
+/**
+ * ptp_clock_unregister() - unregister a PTP hardware clock driver
+ *
+ * @ptp: The clock to remove from service.
+ */
+
+extern int ptp_clock_unregister(struct ptp_clock *ptp);
+
+
+enum ptp_clock_events {
+ PTP_CLOCK_ALARM,
+ PTP_CLOCK_EXTTS,
+ PTP_CLOCK_PPS,
+};
+
+/**
+ * struct ptp_clock_event - decribes a PTP hardware clock event
+ *
+ * @type: One of the ptp_clock_events enumeration values.
+ * @index: Identifies the source of the event.
+ * @timestamp: When the event occured.
+ */
+
+struct ptp_clock_event {
+ int type;
+ int index;
+ u64 timestamp;
+};
+
+/**
+ * ptp_clock_event() - notify the PTP layer about an event
+ *
+ * This function should only be called from interrupt context.
+ *
+ * @ptp: The clock obtained from ptp_clock_register().
+ * @event: Message structure describing the event.
+ */
+
+extern void ptp_clock_event(struct ptp_clock *ptp,
+ struct ptp_clock_event *event);
+
+#endif
--
1.6.3.3
^ permalink raw reply related
* [PATCH v3 2/3] ptp: Added a clock that uses the Linux system time.
From: Richard Cochran @ 2010-05-14 16:46 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss
In-Reply-To: <cover.1273855016.git.richard.cochran@omicron.at>
This PTP clock simply uses the NTP time adjustment system calls. The
driver can be used in systems that lack a special hardware PTP clock.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/ptp/Kconfig | 12 ++++
drivers/ptp/Makefile | 1 +
drivers/ptp/ptp_linux.c | 136 +++++++++++++++++++++++++++++++++++++++++++++++
kernel/time/ntp.c | 2 +
4 files changed, 151 insertions(+), 0 deletions(-)
create mode 100644 drivers/ptp/ptp_linux.c
diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
index c80a25b..9390d44 100644
--- a/drivers/ptp/Kconfig
+++ b/drivers/ptp/Kconfig
@@ -23,4 +23,16 @@ config PTP_1588_CLOCK
To compile this driver as a module, choose M here: the module
will be called ptp_clock.
+config PTP_1588_CLOCK_LINUX
+ tristate "Linux system timer as PTP clock"
+ depends on PTP_1588_CLOCK
+ help
+ This driver adds support for using the standard Linux time
+ source as a PTP clock. This clock is only useful if your PTP
+ programs are using software time stamps for the PTP Ethernet
+ packets.
+
+ To compile this driver as a module, choose M here: the module
+ will be called ptp_linux.
+
endmenu
diff --git a/drivers/ptp/Makefile b/drivers/ptp/Makefile
index b86695c..1651d52 100644
--- a/drivers/ptp/Makefile
+++ b/drivers/ptp/Makefile
@@ -3,3 +3,4 @@
#
obj-$(CONFIG_PTP_1588_CLOCK) += ptp_clock.o
+obj-$(CONFIG_PTP_1588_CLOCK_LINUX) += ptp_linux.o
diff --git a/drivers/ptp/ptp_linux.c b/drivers/ptp/ptp_linux.c
new file mode 100644
index 0000000..f0dbfc6
--- /dev/null
+++ b/drivers/ptp/ptp_linux.c
@@ -0,0 +1,136 @@
+/*
+ * PTP 1588 clock using the Linux system clock
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/hrtimer.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/timex.h>
+
+#include <linux/ptp_clock_kernel.h>
+
+static struct ptp_clock *linux_clock;
+
+DEFINE_SPINLOCK(adjtime_lock);
+
+static int ptp_linux_adjfreq(void *priv, s32 delta)
+{
+ struct timex txc;
+ s64 tmp = delta;
+ int err;
+ txc.freq = div_s64(tmp<<16, 1000);
+ txc.modes = ADJ_FREQUENCY;
+ err = do_adjtimex(&txc);
+ return err < 0 ? err : 0;
+}
+
+static int ptp_linux_adjtime(void *priv, struct timespec *ts)
+{
+ s64 delta;
+ ktime_t now;
+ struct timespec t2;
+ unsigned long flags;
+ int err;
+
+ delta = 1000000000LL * ts->tv_sec + ts->tv_nsec;
+
+ spin_lock_irqsave(&adjtime_lock, flags);
+
+ now = ktime_get_real();
+
+ now = delta < 0 ? ktime_sub_ns(now,-delta) : ktime_add_ns(now,delta);
+
+ t2 = ktime_to_timespec(now);
+
+ err = do_settimeofday(&t2);
+
+ spin_unlock_irqrestore(&adjtime_lock, flags);
+
+ return err;
+}
+
+static int ptp_linux_gettime(void *priv, struct timespec *ts)
+{
+ getnstimeofday(ts);
+ return 0;
+}
+
+static int ptp_linux_settime(void *priv, struct timespec *ts)
+{
+ return do_settimeofday(ts);
+}
+
+static int ptp_linux_gettimer(void *priv, int index, struct itimerspec *ts)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static int ptp_linux_settimer(void *p, int i, int abs, struct itimerspec *ts)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static int ptp_linux_enable(void *priv, struct ptp_clock_request *rq, int on)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static struct ptp_clock_info ptp_linux_caps = {
+ .owner = THIS_MODULE,
+ .name = "Linux timer",
+ .max_adj = 512000,
+ .n_alarm = 0,
+ .n_ext_ts = 0,
+ .n_per_out = 0,
+ .pps = 0,
+ .priv = NULL,
+ .adjfreq = ptp_linux_adjfreq,
+ .adjtime = ptp_linux_adjtime,
+ .gettime = ptp_linux_gettime,
+ .settime = ptp_linux_settime,
+ .gettimer = ptp_linux_gettimer,
+ .settimer = ptp_linux_settimer,
+ .enable = ptp_linux_enable,
+};
+
+/* module operations */
+
+static void __exit ptp_linux_exit(void)
+{
+ ptp_clock_unregister(linux_clock);
+}
+
+static int __init ptp_linux_init(void)
+{
+ linux_clock = ptp_clock_register(&ptp_linux_caps);
+
+ return IS_ERR(linux_clock) ? PTR_ERR(linux_clock) : 0;
+}
+
+module_init(ptp_linux_init);
+module_exit(ptp_linux_exit);
+
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_DESCRIPTION("PTP clock using the Linux system timer");
+MODULE_LICENSE("GPL");
diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c
index 7c0f180..efd1ff8 100644
--- a/kernel/time/ntp.c
+++ b/kernel/time/ntp.c
@@ -14,6 +14,7 @@
#include <linux/timex.h>
#include <linux/time.h>
#include <linux/mm.h>
+#include <linux/module.h>
/*
* NTP timekeeping variables:
@@ -535,6 +536,7 @@ int do_adjtimex(struct timex *txc)
return result;
}
+EXPORT_SYMBOL(do_adjtimex);
static int __init ntp_tick_adj_setup(char *str)
{
--
1.6.3.3
^ permalink raw reply related
* [PATCH v3 3/3] ptp: Added a clock that uses the eTSEC found on the MPC85xx.
From: Richard Cochran @ 2010-05-14 16:46 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss
In-Reply-To: <cover.1273855016.git.richard.cochran@omicron.at>
The eTSEC includes a PTP clock with quite a few features. This patch adds
support for the basic clock adjustment functions, plus two external time
stamps and one alarm.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
Documentation/powerpc/dts-bindings/fsl/tsec.txt | 56 +++
arch/powerpc/boot/dts/mpc8313erdb.dts | 14 +
arch/powerpc/boot/dts/p2020ds.dts | 14 +
arch/powerpc/boot/dts/p2020rdb.dts | 14 +
drivers/net/Makefile | 1 +
drivers/net/gianfar_ptp.c | 521 +++++++++++++++++++++++
drivers/net/gianfar_ptp_reg.h | 113 +++++
drivers/ptp/Kconfig | 13 +
8 files changed, 746 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/gianfar_ptp.c
create mode 100644 drivers/net/gianfar_ptp_reg.h
diff --git a/Documentation/powerpc/dts-bindings/fsl/tsec.txt b/Documentation/powerpc/dts-bindings/fsl/tsec.txt
index edb7ae1..b09ba66 100644
--- a/Documentation/powerpc/dts-bindings/fsl/tsec.txt
+++ b/Documentation/powerpc/dts-bindings/fsl/tsec.txt
@@ -74,3 +74,59 @@ Example:
interrupt-parent = <&mpic>;
phy-handle = <&phy0>
};
+
+* Gianfar PTP clock nodes
+
+General Properties:
+
+ - device_type Should be "ptp_clock"
+ - model Model of the device. Must be "eTSEC"
+ - reg Offset and length of the register set for the device
+ - interrupts There should be at least two and as many as four
+ PTP related interrupts
+
+Clock Properties:
+
+ - tclk_period Timer reference clock period in nanoseconds.
+ - tmr_prsc Prescaler, divides the output clock.
+ - tmr_add Frequency compensation value.
+ - cksel 0= external clock, 1= eTSEC system clock, 3= RTC clock input.
+ Currently the driver only supports choice "1".
+ - tmr_fiper1 Fixed interval period pulse generator.
+ - tmr_fiper2 Fixed interval period pulse generator.
+
+ These properties set the operational parameters for the PTP
+ clock. You must choose these carefully for the clock to work right.
+ Here is how to figure good values:
+
+ TimerOsc = system clock MHz
+ tclk_period = desired clock period nanoseconds
+ NominalFreq = 1000 / tclk_period MHz
+ FreqDivRatio = TimerOsc / NominalFreq (must be greater that 1.0)
+ tmr_add = ceil(2^32 / FreqDivRatio)
+ OutputClock = NominalFreq / tmr_prsc MHz
+ PulseWidth = 1 / OutputClock microseconds
+ FiperFreq1 = desired frequency in Hz
+ FiperDiv1 = 1000000 * OutputClock / FiperFreq1
+ tmr_fiper1 = tmr_prsc * tclk_period * FiperDiv1 - tclk_period
+
+ The calculation for tmr_fiper2 is the same as for tmr_fiper1. The
+ driver expects that tmr_fiper1 will be correctly set to produce a 1
+ Pulse Per Second (PPS) signal, since this will be offered to the PPS
+ subsystem to synchronize the Linux clock.
+
+Example:
+
+ ptp_clock@24E00 {
+ device_type = "ptp_clock";
+ model = "eTSEC";
+ reg = <0x24E00 0xB0>;
+ interrupts = <12 0x8 13 0x8>;
+ interrupt-parent = < &ipic >;
+ tclk_period = <10>;
+ tmr_prsc = <100>;
+ tmr_add = <0x999999A4>;
+ cksel = <0x1>;
+ tmr_fiper1 = <0x3B9AC9F6>;
+ tmr_fiper2 = <0x00018696>;
+ };
diff --git a/arch/powerpc/boot/dts/mpc8313erdb.dts b/arch/powerpc/boot/dts/mpc8313erdb.dts
index 183f2aa..6be6b16 100644
--- a/arch/powerpc/boot/dts/mpc8313erdb.dts
+++ b/arch/powerpc/boot/dts/mpc8313erdb.dts
@@ -208,6 +208,20 @@
sleep = <&pmc 0x00300000>;
};
+ ptp_clock@24E00 {
+ device_type = "ptp_clock";
+ model = "eTSEC";
+ reg = <0x24E00 0xB0>;
+ interrupts = <12 0x8 13 0x8>;
+ interrupt-parent = < &ipic >;
+ tclk_period = <10>;
+ tmr_prsc = <100>;
+ tmr_add = <0x999999A4>;
+ cksel = <0x1>;
+ tmr_fiper1 = <0x3B9AC9F6>;
+ tmr_fiper2 = <0x00018696>;
+ };
+
enet0: ethernet@24000 {
#address-cells = <1>;
#size-cells = <1>;
diff --git a/arch/powerpc/boot/dts/p2020ds.dts b/arch/powerpc/boot/dts/p2020ds.dts
index 1101914..f72353a 100644
--- a/arch/powerpc/boot/dts/p2020ds.dts
+++ b/arch/powerpc/boot/dts/p2020ds.dts
@@ -336,6 +336,20 @@
phy_type = "ulpi";
};
+ ptp_clock@24E00 {
+ device_type = "ptp_clock";
+ model = "eTSEC";
+ reg = <0x24E00 0xB0>;
+ interrupts = <68 2 69 2 70 2>;
+ interrupt-parent = < &mpic >;
+ tclk_period = <5>;
+ tmr_prsc = <200>;
+ tmr_add = <0xCCCCCCCD>;
+ cksel = <1>;
+ tmr_fiper1 = <0x3B9AC9FB>;
+ tmr_fiper2 = <0x0001869B>;
+ };
+
enet0: ethernet@24000 {
#address-cells = <1>;
#size-cells = <1>;
diff --git a/arch/powerpc/boot/dts/p2020rdb.dts b/arch/powerpc/boot/dts/p2020rdb.dts
index da4cb0d..3eb9e65 100644
--- a/arch/powerpc/boot/dts/p2020rdb.dts
+++ b/arch/powerpc/boot/dts/p2020rdb.dts
@@ -396,6 +396,20 @@
phy_type = "ulpi";
};
+ ptp_clock@24E00 {
+ device_type = "ptp_clock";
+ model = "eTSEC";
+ reg = <0x24E00 0xB0>;
+ interrupts = <68 2 69 2 70 2>;
+ interrupt-parent = < &mpic >;
+ tclk_period = <5>;
+ tmr_prsc = <200>;
+ tmr_add = <0xCCCCCCCD>;
+ cksel = <1>;
+ tmr_fiper1 = <0x3B9AC9FB>;
+ tmr_fiper2 = <0x0001869B>;
+ };
+
enet0: ethernet@24000 {
#address-cells = <1>;
#size-cells = <1>;
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 0a0512a..389c0d9 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -28,6 +28,7 @@ obj-$(CONFIG_ATL2) += atlx/
obj-$(CONFIG_ATL1E) += atl1e/
obj-$(CONFIG_ATL1C) += atl1c/
obj-$(CONFIG_GIANFAR) += gianfar_driver.o
+obj-$(CONFIG_PTP_1588_CLOCK_GIANFAR) += gianfar_ptp.o
obj-$(CONFIG_TEHUTI) += tehuti.o
obj-$(CONFIG_ENIC) += enic/
obj-$(CONFIG_JME) += jme.o
diff --git a/drivers/net/gianfar_ptp.c b/drivers/net/gianfar_ptp.c
new file mode 100644
index 0000000..46797aa
--- /dev/null
+++ b/drivers/net/gianfar_ptp.c
@@ -0,0 +1,521 @@
+/*
+ * PTP 1588 clock using the eTSEC
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/device.h>
+#include <linux/hrtimer.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/timex.h>
+#include <asm/io.h>
+
+#include <linux/ptp_clock_kernel.h>
+
+#include "gianfar_ptp_reg.h"
+#include "gianfar.h"
+
+#define DRIVER "gianfar_ptp"
+#define N_ALARM 1 /* first alarm is used internally to reset fipers */
+#define N_EXT_TS 2
+#define N_IRQS 2
+#define REG_SIZE sizeof(struct gianfar_ptp_registers)
+
+struct etsects {
+ struct gianfar_ptp_registers *regs;
+ int irq[N_IRQS];
+ u64 alarm_interval; /* for periodic alarm */
+ u64 alarm_value;
+ u32 tclk_period; /* nanoseconds */
+ u32 tmr_prsc;
+ u32 tmr_add;
+ u32 cksel;
+ u32 tmr_fiper1;
+ u32 tmr_fiper2;
+};
+
+/* Private globals */
+static struct ptp_clock *gianfar_clock;
+static struct etsects the_clock;
+DEFINE_SPINLOCK(register_lock);
+
+/*
+ * Register access functions
+ */
+
+static u64 tmr_cnt_read(struct etsects *etsects)
+{
+ u64 ns;
+ u32 lo, hi;
+
+ lo = gfar_read(&etsects->regs->tmr_cnt_l);
+ hi = gfar_read(&etsects->regs->tmr_cnt_h);
+ ns = ((u64) hi) << 32;
+ ns |= lo;
+ return ns;
+}
+
+static void tmr_cnt_write(struct etsects *etsects, u64 ns)
+{
+ u32 hi = ns >> 32;
+ u32 lo = ns & 0xffffffff;
+
+ gfar_write(&etsects->regs->tmr_cnt_l, lo);
+ gfar_write(&etsects->regs->tmr_cnt_h, hi);
+}
+
+static void set_alarm(struct etsects *etsects)
+{
+ u64 ns;
+ u32 lo, hi;
+
+ ns = tmr_cnt_read(etsects) + 1500000000ULL;
+ ns = div_u64(ns, 1000000000UL) * 1000000000ULL;
+ ns -= etsects->tclk_period;
+ hi = ns >> 32;
+ lo = ns & 0xffffffff;
+ gfar_write(&etsects->regs->tmr_alarm1_l, lo);
+ gfar_write(&etsects->regs->tmr_alarm1_h, hi);
+}
+
+static void set_fipers(struct etsects *etsects)
+{
+ u32 tmr_ctrl = gfar_read(&etsects->regs->tmr_ctrl);
+
+ gfar_write(&etsects->regs->tmr_ctrl, tmr_ctrl & (~TE));
+ gfar_write(&etsects->regs->tmr_prsc, etsects->tmr_prsc);
+ gfar_write(&etsects->regs->tmr_fiper1, etsects->tmr_fiper1);
+ gfar_write(&etsects->regs->tmr_fiper2, etsects->tmr_fiper2);
+ set_alarm(etsects);
+ gfar_write(&etsects->regs->tmr_ctrl, tmr_ctrl|TE);
+}
+
+/*
+ * Interrupt service routine
+ */
+
+static irqreturn_t isr(int irq, void *priv)
+{
+ struct etsects *etsects = priv;
+ struct ptp_clock_event event;
+ u64 ns;
+ u32 ack=0, lo, hi, mask, val;
+
+ val = gfar_read(&etsects->regs->tmr_tevent);
+
+ if (val & ETS1) {
+ ack |= ETS1;
+ hi = gfar_read(&etsects->regs->tmr_etts1_h);
+ lo = gfar_read(&etsects->regs->tmr_etts1_l);
+ event.type = PTP_CLOCK_EXTTS;
+ event.index = 0;
+ event.timestamp = ((u64) hi) << 32;
+ event.timestamp |= lo;
+ ptp_clock_event(gianfar_clock, &event);
+ }
+
+ if (val & ETS2) {
+ ack |= ETS2;
+ hi = gfar_read(&etsects->regs->tmr_etts2_h);
+ lo = gfar_read(&etsects->regs->tmr_etts2_l);
+ event.type = PTP_CLOCK_EXTTS;
+ event.index = 1;
+ event.timestamp = ((u64) hi) << 32;
+ event.timestamp |= lo;
+ ptp_clock_event(gianfar_clock, &event);
+ }
+
+ if (val & ALM2) {
+ ack |= ALM2;
+ if (etsects->alarm_value) {
+ event.type = PTP_CLOCK_ALARM;
+ event.index = 0;
+ event.timestamp = etsects->alarm_value;
+ ptp_clock_event(gianfar_clock, &event);
+ }
+ if (etsects->alarm_interval) {
+ ns = etsects->alarm_value + etsects->alarm_interval;
+ hi = ns >> 32;
+ lo = ns & 0xffffffff;
+ spin_lock(®ister_lock);
+ gfar_write(&etsects->regs->tmr_alarm2_l, lo);
+ gfar_write(&etsects->regs->tmr_alarm2_h, hi);
+ spin_unlock(®ister_lock);
+ etsects->alarm_value = ns;
+ } else {
+ gfar_write(&etsects->regs->tmr_tevent, ALM2);
+ spin_lock(®ister_lock);
+ mask = gfar_read(&etsects->regs->tmr_temask);
+ mask &= ~ALM2EN;
+ gfar_write(&etsects->regs->tmr_temask, mask);
+ spin_unlock(®ister_lock);
+ etsects->alarm_value = 0;
+ etsects->alarm_interval = 0;
+ }
+ }
+
+ gfar_write(&etsects->regs->tmr_tevent, ack);
+
+ return IRQ_HANDLED;
+}
+
+/*
+ * PTP clock operations
+ */
+
+static int ptp_gianfar_adjfreq(void *priv, s32 ppb)
+{
+ u64 adj;
+ u32 diff, tmr_add;
+ int neg_adj = 0;
+ struct etsects *etsects = priv;
+
+ if (!ppb)
+ return 0;
+
+ if (ppb < 0) {
+ neg_adj = 1;
+ ppb = -ppb;
+ }
+ tmr_add = etsects->tmr_add;
+ adj = tmr_add;
+ adj *= ppb;
+ diff = div_u64(adj, 1000000000ULL);
+
+ tmr_add = neg_adj ? tmr_add - diff : tmr_add + diff;
+
+ gfar_write(&etsects->regs->tmr_add, tmr_add);
+
+ return 0;
+}
+
+static int ptp_gianfar_adjtime(void *priv, struct timespec *ts)
+{
+ s64 delta, now;
+ unsigned long flags;
+ struct etsects *etsects = priv;
+
+ delta = 1000000000LL * ts->tv_sec;
+ delta += ts->tv_nsec;
+
+ spin_lock_irqsave(®ister_lock, flags);
+
+ now = tmr_cnt_read(etsects);
+ now += delta;
+ tmr_cnt_write(etsects, now);
+
+ spin_unlock_irqrestore(®ister_lock, flags);
+
+ set_fipers(etsects);
+
+ return 0;
+}
+
+static int ptp_gianfar_gettime(void *priv, struct timespec *ts)
+{
+ u64 ns;
+ u32 remainder;
+ unsigned long flags;
+ struct etsects *etsects = priv;
+
+ spin_lock_irqsave(®ister_lock, flags);
+
+ ns = tmr_cnt_read(etsects);
+
+ spin_unlock_irqrestore(®ister_lock, flags);
+
+ ts->tv_sec = div_u64_rem(ns, 1000000000, &remainder);
+ ts->tv_nsec = remainder;
+ return 0;
+}
+
+static int ptp_gianfar_settime(void *priv, struct timespec *ts)
+{
+ u64 ns;
+ unsigned long flags;
+ struct etsects *etsects = priv;
+
+ ns = ts->tv_sec * 1000000000ULL;
+ ns += ts->tv_nsec;
+
+ spin_lock_irqsave(®ister_lock, flags);
+
+ tmr_cnt_write(etsects, ns);
+ set_fipers(etsects);
+
+ spin_unlock_irqrestore(®ister_lock, flags);
+
+ return 0;
+}
+
+static int ptp_gianfar_gettimer(void *priv, int index, struct itimerspec *ts)
+{
+ u64 now, ns;
+ u32 remainder;
+ unsigned long flags;
+ struct etsects *etsects = priv;
+
+ ns = etsects->alarm_interval;
+
+ ts->it_interval.tv_sec = div_u64_rem(ns, 1000000000, &remainder);
+ ts->it_interval.tv_nsec = remainder;
+
+ spin_lock_irqsave(®ister_lock, flags);
+ now = tmr_cnt_read(etsects);
+ spin_unlock_irqrestore(®ister_lock, flags);
+
+ ns = etsects->alarm_value - now;
+
+ ts->it_value.tv_sec = div_u64_rem(ns, 1000000000, &remainder);
+ ts->it_value.tv_nsec = remainder;
+
+ return 0;
+}
+
+static int ptp_gianfar_settimer(void *p, int i, int abs, struct itimerspec *ts)
+{
+ u64 ns;
+ u32 lo, hi, mask;
+ unsigned long flags;
+ struct etsects *etsects = p;
+
+ ns = ts->it_interval.tv_sec * 1000000000ULL;
+ ns += ts->it_interval.tv_nsec;
+
+ etsects->alarm_interval = ns;
+
+ ns = ts->it_value.tv_sec * 1000000000ULL;
+ ns += ts->it_value.tv_nsec;
+
+ if (!ns) {
+ /* Cancel the timer. */
+ etsects->alarm_value = 0;
+ etsects->alarm_interval = 0;
+ return 0;
+ }
+
+ if (!abs) {
+ spin_lock_irqsave(®ister_lock, flags);
+ ns += tmr_cnt_read(etsects);
+ spin_unlock_irqrestore(®ister_lock, flags);
+ }
+
+ etsects->alarm_value = ns;
+
+ hi = ns >> 32;
+ lo = ns & 0xffffffff;
+
+ spin_lock_irqsave(®ister_lock, flags);
+
+ gfar_write(&etsects->regs->tmr_alarm2_l, lo);
+ gfar_write(&etsects->regs->tmr_alarm2_h, hi);
+
+ mask = gfar_read(&etsects->regs->tmr_temask);
+ mask |= ALM2EN;
+ gfar_write(&etsects->regs->tmr_temask, mask);
+
+ spin_unlock_irqrestore(®ister_lock, flags);
+
+ return 0;
+}
+
+static int ptp_gianfar_enable(void *priv, struct ptp_clock_request *rq, int on)
+{
+ struct etsects *etsects = priv;
+ unsigned long flags;
+ u32 bit, mask;
+
+ switch (rq->type) {
+ case PTP_REQUEST_EXTTS:
+ switch (rq->index) {
+ case 0:
+ bit = ETS1EN;
+ break;
+ case 1:
+ bit = ETS2EN;
+ break;
+ default:
+ return -EINVAL;
+ }
+ spin_lock_irqsave(®ister_lock, flags);
+ if (on) {
+ mask = gfar_read(&etsects->regs->tmr_temask);
+ mask |= bit;
+ gfar_write(&etsects->regs->tmr_temask, mask);
+ } else {
+ mask = gfar_read(&etsects->regs->tmr_temask);
+ mask &= ~bit;
+ gfar_write(&etsects->regs->tmr_temask, mask);
+ }
+ spin_unlock_irqrestore(®ister_lock, flags);
+ return 0;
+ }
+
+ return -EOPNOTSUPP;
+}
+
+static struct ptp_clock_info ptp_gianfar_caps = {
+ .owner = THIS_MODULE,
+ .name = "gianfar clock",
+ .max_adj = 512000,
+ .n_alarm = N_ALARM,
+ .n_ext_ts = N_EXT_TS,
+ .n_per_out = 0,
+ .pps = 0,
+ .priv = &the_clock,
+ .adjfreq = ptp_gianfar_adjfreq,
+ .adjtime = ptp_gianfar_adjtime,
+ .gettime = ptp_gianfar_gettime,
+ .settime = ptp_gianfar_settime,
+ .gettimer = ptp_gianfar_gettimer,
+ .settimer = ptp_gianfar_settimer,
+ .enable = ptp_gianfar_enable,
+};
+
+/* OF device tree */
+
+static int get_of_u32(struct device_node *node, char *str, u32 *val)
+{
+ int plen;
+ const u32 *prop = of_get_property(node, str, &plen);
+
+ if (!prop || plen != sizeof(*prop))
+ return -1;
+ *val = *prop;
+ return 0;
+}
+
+static int gianfar_ptp_probe(struct of_device* dev,
+ const struct of_device_id *match)
+{
+ u64 addr, size;
+ struct device_node *node = dev->node;
+ struct etsects *etsects = &the_clock;
+ struct timespec now;
+ phys_addr_t reg_addr;
+ unsigned long reg_size;
+ u32 tmr_ctrl;
+ int i;
+
+ if (get_of_u32(node, "tclk_period", &etsects->tclk_period) ||
+ get_of_u32(node, "tmr_prsc", &etsects->tmr_prsc) ||
+ get_of_u32(node, "tmr_add", &etsects->tmr_add) ||
+ get_of_u32(node, "cksel", &etsects->cksel) ||
+ get_of_u32(node, "tmr_fiper1", &etsects->tmr_fiper1) ||
+ get_of_u32(node, "tmr_fiper2", &etsects->tmr_fiper2))
+ return -ENODEV;
+
+ for (i = 0; i < N_IRQS; i++) {
+
+ etsects->irq[i] = irq_of_parse_and_map(node, i);
+
+ if (etsects->irq[i] == NO_IRQ) {
+ pr_err("irq[%d] not in device tree", i);
+ return -ENODEV;
+ }
+
+ if (request_irq(etsects->irq[i], isr, 0, DRIVER, etsects)) {
+ pr_err("request_irq failed irq %d", etsects->irq[i]);
+ return -ENODEV;
+ }
+ }
+
+ addr = of_translate_address(node, of_get_address(node, 0, &size, NULL));
+ reg_addr = addr;
+ reg_size = size;
+ if (reg_size < REG_SIZE) {
+ pr_warning("device tree reg range %lu too small\n", reg_size);
+ reg_size = REG_SIZE;
+ }
+ etsects->regs = ioremap(reg_addr, reg_size);
+ if (!etsects->regs) {
+ pr_err("ioremap ptp registers failed\n");
+ return -EINVAL;
+ }
+
+ tmr_ctrl =
+ (etsects->tclk_period & TCLK_PERIOD_MASK) << TCLK_PERIOD_SHIFT |
+ (etsects->cksel & CKSEL_MASK) << CKSEL_SHIFT;
+
+ getnstimeofday(&now);
+ ptp_gianfar_settime(etsects, &now);
+
+ gfar_write(&etsects->regs->tmr_ctrl, tmr_ctrl);
+ gfar_write(&etsects->regs->tmr_add, etsects->tmr_add);
+ gfar_write(&etsects->regs->tmr_prsc, etsects->tmr_prsc);
+ gfar_write(&etsects->regs->tmr_fiper1, etsects->tmr_fiper1);
+ gfar_write(&etsects->regs->tmr_fiper2, etsects->tmr_fiper2);
+ set_alarm(etsects);
+ gfar_write(&etsects->regs->tmr_ctrl, tmr_ctrl|FS|RTPE|TE);
+
+ gianfar_clock = ptp_clock_register(&ptp_gianfar_caps);
+
+ return IS_ERR(gianfar_clock) ? PTR_ERR(gianfar_clock) : 0;
+}
+
+static int gianfar_ptp_remove(struct of_device* dev)
+{
+ int i;
+
+ gfar_write(&the_clock.regs->tmr_temask, 0);
+ gfar_write(&the_clock.regs->tmr_ctrl, 0);
+
+ ptp_clock_unregister(gianfar_clock);
+
+ for (i = 0; i < N_IRQS; i++)
+ free_irq(the_clock.irq[i], &the_clock);
+
+ iounmap(the_clock.regs);
+
+ return 0;
+}
+
+static struct of_device_id match_table[] = {
+ { .type = "ptp_clock" },
+ {},
+};
+
+static struct of_platform_driver gianfar_ptp_driver = {
+ .name = "gianfar_ptp",
+ .match_table = match_table,
+ .owner = THIS_MODULE,
+ .probe = gianfar_ptp_probe,
+ .remove = gianfar_ptp_remove,
+};
+
+/* module operations */
+
+static void __exit ptp_gianfar_exit(void)
+{
+ of_unregister_platform_driver(&gianfar_ptp_driver);
+}
+
+static int __init ptp_gianfar_init(void)
+{
+ return of_register_platform_driver(&gianfar_ptp_driver);
+}
+
+module_init(ptp_gianfar_init);
+module_exit(ptp_gianfar_exit);
+
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_DESCRIPTION("PTP clock using the eTSEC");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/gianfar_ptp_reg.h b/drivers/net/gianfar_ptp_reg.h
new file mode 100644
index 0000000..6e2b82f
--- /dev/null
+++ b/drivers/net/gianfar_ptp_reg.h
@@ -0,0 +1,113 @@
+/* gianfar_ptp_reg.h
+ * Generated by regen.tcl on Thu May 13 01:38:57 PM CEST 2010
+ *
+ * PTP 1588 clock using the gianfar eTSEC
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * 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.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#ifndef _GIANFAR_PTP_REG_H_
+#define _GIANFAR_PTP_REG_H_
+
+struct gianfar_ptp_registers {
+ u32 tmr_ctrl; /* Timer control register */
+ u32 tmr_tevent; /* Timestamp event register */
+ u32 tmr_temask; /* Timer event mask register */
+ u32 tmr_pevent; /* Timestamp event register */
+ u32 tmr_pemask; /* Timer event mask register */
+ u32 tmr_stat; /* Timestamp status register */
+ u32 tmr_cnt_h; /* Timer counter high register */
+ u32 tmr_cnt_l; /* Timer counter low register */
+ u32 tmr_add; /* Timer drift compensation addend register */
+ u32 tmr_acc; /* Timer accumulator register */
+ u32 tmr_prsc; /* Timer prescale */
+ u8 res1[4];
+ u32 tmroff_h; /* Timer offset high */
+ u32 tmroff_l; /* Timer offset low */
+ u8 res2[8];
+ u32 tmr_alarm1_h; /* Timer alarm 1 high register */
+ u32 tmr_alarm1_l; /* Timer alarm 1 high register */
+ u32 tmr_alarm2_h; /* Timer alarm 2 high register */
+ u32 tmr_alarm2_l; /* Timer alarm 2 high register */
+ u8 res3[48];
+ u32 tmr_fiper1; /* Timer fixed period interval */
+ u32 tmr_fiper2; /* Timer fixed period interval */
+ u32 tmr_fiper3; /* Timer fixed period interval */
+ u8 res4[20];
+ u32 tmr_etts1_h; /* Timestamp of general purpose external trigger */
+ u32 tmr_etts1_l; /* Timestamp of general purpose external trigger */
+ u32 tmr_etts2_h; /* Timestamp of general purpose external trigger */
+ u32 tmr_etts2_l; /* Timestamp of general purpose external trigger */
+};
+
+/* Bit definitions for the TMR_CTRL register */
+#define ALM1P (1<<31) /* Alarm1 output polarity */
+#define ALM2P (1<<30) /* Alarm2 output polarity */
+#define FS (1<<28) /* FIPER start indication */
+#define PP1L (1<<27) /* Fiper1 pulse loopback mode enabled. */
+#define PP2L (1<<26) /* Fiper2 pulse loopback mode enabled. */
+#define TCLK_PERIOD_SHIFT (16) /* 1588 timer reference clock period. */
+#define TCLK_PERIOD_MASK (0x3ff)
+#define RTPE (1<<15) /* Record Tx Timestamp to PAL Enable. */
+#define FRD (1<<14) /* FIPER Realignment Disable */
+#define ESFDP (1<<11) /* External Tx/Rx SFD Polarity. */
+#define ESFDE (1<<10) /* External Tx/Rx SFD Enable. */
+#define ETEP2 (1<<9) /* External trigger 2 edge polarity */
+#define ETEP1 (1<<8) /* External trigger 1 edge polarity */
+#define COPH (1<<7) /* Generated clock (TSEC_1588_GCLK) output phase. */
+#define CIPH (1<<6) /* External oscillator input clock phase. */
+#define TMSR (1<<5) /* Timer soft reset. When enabled, it resets all the timer registers and state machines. */
+#define BYP (1<<3) /* Bypass drift compensated clock */
+#define TE (1<<2) /* 1588 timer enable. If not enabled, all the timer registers and state machines are disabled. */
+#define CKSEL_SHIFT (0) /* 1588 Timer reference clock source select. */
+#define CKSEL_MASK (0x3)
+
+/* Bit definitions for the TMR_TEVENT register */
+#define ETS2 (1<<25) /* External trigger 2 timestamp sampled */
+#define ETS1 (1<<24) /* External trigger 1 timestamp sampled */
+#define ALM2 (1<<17) /* Current time equaled alarm time register 2 */
+#define ALM1 (1<<16) /* Current time equaled alarm time register 1 */
+#define PP1 (1<<7) /* Indicates that a periodic pulse has been generated based on FIPER1 register */
+#define PP2 (1<<6) /* Indicates that a periodic pulse has been generated based on FIPER2 register */
+#define PP3 (1<<5) /* Indicates that a periodic pulse has been generated based on FIPER3 register */
+
+/* Bit definitions for the TMR_TEMASK register */
+#define ETS2EN (1<<25) /* External trigger 2 timestamp sample event enable */
+#define ETS1EN (1<<24) /* External trigger 1 timestamp sample event enable */
+#define ALM2EN (1<<17) /* Timer ALM2 event enable */
+#define ALM1EN (1<<16) /* Timer ALM1 event enable */
+#define PP1EN (1<<7) /* Periodic pulse event 1 enable */
+#define PP2EN (1<<6) /* Periodic pulse event 2 enable */
+
+/* Bit definitions for the TMR_PEVENT register */
+#define TXP2 (1<<9) /* Indicates that a PTP frame has been transmitted and its timestamp is stored in TXTS2 register */
+#define TXP1 (1<<8) /* Indicates that a PTP frame has been transmitted and its timestamp is stored in TXTS1 register */
+#define RXP (1<<0) /* Indicates that a PTP frame has been received */
+
+/* Bit definitions for the TMR_PEMASK register */
+#define TXP2EN (1<<9) /* Transmit PTP packet event 2 enable */
+#define TXP1EN (1<<8) /* Transmit PTP packet event 1 enable */
+#define RXPEN (1<<0) /* Receive PTP packet event enable */
+
+/* Bit definitions for the TMR_STAT register */
+#define STAT_VEC_SHIFT (0) /* Timer general purpose status vector */
+#define STAT_VEC_MASK (0x3f)
+
+/* Bit definitions for the TMR_PRSC register */
+#define PRSC_OCK_SHIFT (0) /* Output clock division/prescale factor. */
+#define PRSC_OCK_MASK (0xffff)
+
+#endif
diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
index 9390d44..3b7bd73 100644
--- a/drivers/ptp/Kconfig
+++ b/drivers/ptp/Kconfig
@@ -35,4 +35,17 @@ config PTP_1588_CLOCK_LINUX
To compile this driver as a module, choose M here: the module
will be called ptp_linux.
+config PTP_1588_CLOCK_GIANFAR
+ tristate "Freescale eTSEC as PTP clock"
+ depends on PTP_1588_CLOCK
+ depends on GIANFAR
+ help
+ This driver adds support for using the eTSEC as a PTP
+ clock. This clock is only useful if your PTP programs are
+ getting hardware time stamps on the PTP Ethernet packets
+ using the SO_TIMESTAMPING API.
+
+ To compile this driver as a module, choose M here: the module
+ will be called gianfar_ptp.
+
endmenu
--
1.6.3.3
^ permalink raw reply related
* RE: SR-IOV PCI quirk for 82599?
From: Ronciak, John @ 2010-05-14 16:57 UTC (permalink / raw)
To: Fischer, Anna, netdev@vger.kernel.org,
e1000-devel@lists.sourceforge.net
In-Reply-To: <0199E0D51A61344794750DC57738F58E70B078DCA0@GVW1118EXC.americas.hpqcorp.net>
> There is a PCI quirk for the 82576 controller that programs the PCI
> BARs to use Flash memory if the BIOS has not allocated resources for
> the SR-IOV VF BARs.
>
> Is there a similar quirk for the 82599, or can even the same one be
> used for that device?
>
> Thanks,
> Anna
The quirk is for system BIOS's that don't fully support SR-IOV. The quirk is for those systems with these bad BIOS's with our (and others) devices in them. So we think the same quirk could be used but we have never test with it in a system with a Niantic device in it. So we don't know if it would work or not. It should but without testing there is no way to know for sure.
Cheers,
John
^ permalink raw reply
* [PATCH -next] netfilter: xt_TEE depends on NF_CONNTRACK
From: Randy Dunlap @ 2010-05-14 17:12 UTC (permalink / raw)
To: Stephen Rothwell, netdev, Jan Engelhardt, netfilter-devel
Cc: linux-next, LKML, Patrick McHardy
In-Reply-To: <20100514161407.740da901.sfr@canb.auug.org.au>
From: Randy Dunlap <randy.dunlap@oracle.com>
Fix xt_TEE build for the case of NF_CONNTRACK=m and
NETFILTER_XT_TARGET_TEE=y:
xt_TEE.c:(.text+0x6df5c): undefined reference to `nf_conntrack_untracked'
4x
Built with all 4 m/y combinations.
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Jan Engelhardt <jengelh@medozas.de>
---
net/netfilter/Kconfig | 1 +
1 file changed, 1 insertion(+)
--- linux-next-20100514.orig/net/netfilter/Kconfig
+++ linux-next-20100514/net/netfilter/Kconfig
@@ -506,6 +506,7 @@ config NETFILTER_XT_TARGET_TEE
tristate '"TEE" - packet cloning to alternate destiantion'
depends on NETFILTER_ADVANCED
depends on (IPV6 || IPV6=n)
+ depends on !NF_CONNTRACK || NF_CONNTRACK
---help---
This option adds a "TEE" target with which a packet can be cloned and
this clone be rerouted to another nexthop.
^ permalink raw reply
* Re: [net-next-2.6 V7 PATCH 1/2] Add netlink support for virtual port management (was iovnl)
From: Scott Feldman @ 2010-05-14 17:19 UTC (permalink / raw)
To: Patrick McHardy, Arnd Bergmann; +Cc: davem, netdev, chrisw
In-Reply-To: <4BED7D64.3070500@trash.net>
On 5/14/10 9:42 AM, "Patrick McHardy" <kaber@trash.net> wrote:
> Arnd Bergmann wrote:
>> Maybe a better structure would be to separate the two cases, also allowing
>> a port profile to be associated with both the PF and with each of its VFs?
>>
>> Something like this:
>>
>> [IFLA_NUM_VF]
>> [IFLA_VF_PORTS]
>> [IFLA_VF_PORT]
>> [IFLA_VF_PORT_*], ...
>> [IFLA_VF_PORT]
>> [IFLA_VF_PORT_*], ...
>> [IFLA_PORT_SELF]
>> [IFLA_VF_PORT_*], ...
>
> That would also be fine.
I want to make sure I've got this right before starting on ver8 of patch:
- we'll use the layout listed above
- RTM_SETLINK msg includes the full nested layout
- contains IFLA_VF_PORTs for all VFs of a PF
- OR, contains IFLA_PORT_SELF if PF is it's own VF
- it's up to the receiver to compare for changes for each VF
- RTM_GETLINK msg includes the full nested layout
- same rules as RTM_SETLINK above
I think we should redo the other IFLA_VF_xxx msgs in the same style. I'm
not going to tackle that for IFLA_VF_PORTS patch, but it would be a good
followup patch.
Do we have a plan?
-scott
^ permalink raw reply
* Re: [PATCH -next] netfilter: xt_TEE depends on NF_CONNTRACK
From: Patrick McHardy @ 2010-05-14 17:20 UTC (permalink / raw)
To: Randy Dunlap
Cc: Stephen Rothwell, netdev, Jan Engelhardt, netfilter-devel,
linux-next, LKML, David S. Miller
In-Reply-To: <20100514101218.16cdb0ad.randy.dunlap@oracle.com>
Randy Dunlap wrote:
> From: Randy Dunlap <randy.dunlap@oracle.com>
>
> Fix xt_TEE build for the case of NF_CONNTRACK=m and
> NETFILTER_XT_TARGET_TEE=y:
>
> xt_TEE.c:(.text+0x6df5c): undefined reference to `nf_conntrack_untracked'
> 4x
>
> Built with all 4 m/y combinations.
>
> Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
> Cc: Patrick McHardy <kaber@trash.net>
> Cc: Jan Engelhardt <jengelh@medozas.de>
Acked-by: Patrick McHardy <kaber@trash.net>
Dave, please apply directly. Thanks!
> ---
> net/netfilter/Kconfig | 1 +
> 1 file changed, 1 insertion(+)
>
> --- linux-next-20100514.orig/net/netfilter/Kconfig
> +++ linux-next-20100514/net/netfilter/Kconfig
> @@ -506,6 +506,7 @@ config NETFILTER_XT_TARGET_TEE
> tristate '"TEE" - packet cloning to alternate destiantion'
> depends on NETFILTER_ADVANCED
> depends on (IPV6 || IPV6=n)
> + depends on !NF_CONNTRACK || NF_CONNTRACK
> ---help---
> This option adds a "TEE" target with which a packet can be cloned and
> this clone be rerouted to another nexthop.
>
^ permalink raw reply
* Re: loosing IPMI-card by loading netconsole
From: "Brandeburg, Jesse" @ 2010-05-14 17:20 UTC (permalink / raw)
To: Tejun Heo
Cc: Ronciak, John, Henning Fehrmann, Kirsher, Jeffrey T,
Allan, Bruce W, Waskiewicz Jr, Peter P, netdev@vger.kernel.org,
Matt Mackall, Carsten Aulbert, e1000-devel
In-Reply-To: <4BED79EB.1000204@kernel.org>
On Fri, 2010-05-14 at 09:27 -0700, Tejun Heo wrote:
> Hello, John.
>
> As Henning seems offline, I'll try to fill in.
>
> On 05/14/2010 04:51 PM, Ronciak, John wrote:
> > Sorry to hear about the problem you are having Henning. What do you
> > mean when you say "it disappears"?
>
> It stops responding to IPMI requests.
We've actually had quite a few problems like this over the years, so I'm
not quite so surprised to hear about something like this.
Its easy to break the reception of IPMI packets because there are a
couple of registers that if not correctly configured during all points
of driver lifetime (probe only, administratively down, up)
> > Can both eth0 and eth1 ping (or be pinged)? Do all the networking
> > devices still show up in the system when you do an 'lspci'?
>
> Yeah, everything other than IPMI works just fine.
>
> > What happens if you down and then up the interface you are having
> > problems with? Does 'rmmod' do the same thing as your removal
> > method?
>
> Haven't tried these but well I think rmmoding should achieve about the
> same thing.
>
> > Is there anything in the system logs saying anything about the
> > interfaces?
>
> Nope.
One thing that would really help us is to see the stats from ethtool -S
ethX when interface is up, and not receiving IPMI
The other "smoking gun" indicator is the output of the register dump
tool called ethregs that we have posted at sourceforge. Please gather
registers for the card in question before and after loading netconsole.
http://prdownloads.sf.net/e1000/ethregs-1.7.2.tar.gz
> > We have not had reports of this so this is a bit unusual. Please let us know.
> >
> > Does this happen on other systems as well or just one particular system?
>
> Yeah, it happens on at least several hundred machines, so not an
> isolated hardware issue at all.
>
> To sum up.
>
> On 2.6.27.39, netconsole + IPMI works fine. On 2.6.32.7, as soon as
> netconsole is loaded, IPMI stops working. Unloading netconsole
> doesn't revive IPMI but detaching the driver from the controller does.
> In both cases, usual networking works fine.
I think that "loading netconsole" means bringing the interface "UP", in
this case, is this correct? To ask another way: Is network traffic
active on the interface in question before netconsole is loaded?
Jesse
^ 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