* Re: [PATCH] [NET]: Remove PowerPC code from fec.c
From: Jochen Friedrich @ 2008-01-25 16:04 UTC (permalink / raw)
To: Frans Pop
Cc: galak, geert, gerg, jgarzik, linux-kernel, linux-m68k,
linuxppc-dev, netdev, scottwood, vitb
In-Reply-To: <E1JIQpQ-0006Xj-RR@faramir.fjphome.nl>
Hi Frans,
> Jochen Friedrich wrote:
>> +++ b/drivers/net/fec.c
>> @@ -23,6 +23,9 @@
>> *
>> * Bug fixes and cleanup by Philippe De Muyter (phdm@macqel.be)
>> * Copyright (c) 2004-2006 Macq Electronique SA.
>> + *
>> + * This driver is now only used on ColdFire processors. Remove conditional
>> + * Powerpc code.
>> */
>
> This comment makes sense for a changelog, but IMO it makes no sense at all
> to add it to the file.
I just added it to clarify this code is now only used on m68knommu (Coldfire).
The comments on top are mailny about MPC860T CPUs (PowerPC), however the driver is no
longer used for these CPUs.
Maybe the wording should be changed to:
This driver is now only used on ColdFire (m68knommu) processors. Conditional
PowerPC code has been removed.
Thanks,
Jochen
^ permalink raw reply
* ipcomp regression in 2.6.24?
From: Beschorner Daniel @ 2008-01-25 15:57 UTC (permalink / raw)
To: netdev
With 2.6.24 IPSEC/ESP tunnels to older kernels establish fine, data
flows in both directions, but no data comes out of the tunnel.
Needed to disable ipcomp.
Daniel
^ permalink raw reply
* Re: [PATCH] fib_trie: rescan if key is lost during dump
From: Stephen Hemminger @ 2008-01-25 16:13 UTC (permalink / raw)
To: Jarek Poplawski; +Cc: David Miller, kaber, netdev
In-Reply-To: <20080125082300.GA2257@ff.dom.local>
On Fri, 25 Jan 2008 09:23:00 +0100
Jarek Poplawski <jarkao2@gmail.com> wrote:
> On 24-01-2008 22:51, Stephen Hemminger wrote:
> > Normally during a dump the key of the last dumped entry is used for
> > continuation, but since lock is dropped it might be lost. In that case
> > fallback to the old counter based N^2 behaviour. This means the dump will end up
> > skipping some routes which matches what FIB_HASH does.
> >
> > Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
> ...
> > @@ -1918,35 +1931,37 @@ static int fn_trie_dump(struct fib_table
> > struct leaf *l;
> > struct trie *t = (struct trie *) tb->tb_data;
> > t_key key = cb->args[2];
> > + int count = cb->args[3];
> >
> > rcu_read_lock();
>
> Sorry, but I lost the point: is rtnl held or not held here at the moment?
> If held, how this rcu_read_lock can help? Maybe some additional comment
> in the code?
>
> Thanks,
> Jarek P.
There are two different issues, (therefore two different patches).
1. How to handle multipart resume when the key was deleted during the
period the lock was dropped.
That is what this patch addresses.
2. RCU is unnecessary here because of use of RTNL. I am going to defer
on this till after #1. That patch is much less important.
--
Stephen Hemminger <stephen.hemminger@vyatta.com>
^ permalink raw reply
* Re: [PATCH] TCP:Fix a bug in strategy_allowed_congestion_control
From: Stephen Hemminger @ 2008-01-25 16:25 UTC (permalink / raw)
To: shanwei; +Cc: davem, netdev, Sam Jansen
In-Reply-To: <47998B55.1030305@cn.fujitsu.com>
On Fri, 25 Jan 2008 15:10:13 +0800
shanwei <shanwei@cn.fujitsu.com> wrote:
> hi all:
>
> In strategy_allowed_congestion_control of the 2.6.24 kernel,
> when sysctl_string return 1 on success,it should call
> tcp_set_allowed_congestion_control to set the allowed congestion
> control.But, it don't.
> the sysctl_string return 1 on success, otherwise return negative,
> never return 0.The patch fix the problem.
>
> Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
>
> diff -Nuarp linux-2.6.24/net/ipv4/sysctl_net_ipv4.c linux-2.6.24-new/net/ipv4/sysctl_net_ipv4.c
> --- linux-2.6.24/net/ipv4/sysctl_net_ipv4.c 2008-01-25 06:58:37.000000000 +0800
> +++ linux-2.6.24-new/net/ipv4/sysctl_net_ipv4.c 2008-01-25 12:23:20.000000000 +0800
> @@ -248,7 +248,7 @@ static int strategy_allowed_congestion_c
>
> tcp_get_available_congestion_control(tbl.data, tbl.maxlen);
> ret = sysctl_string(&tbl, name, nlen, oldval, oldlenp, newval, newlen);
> - if (ret == 0 && newval && newlen)
> + if (ret == 1 && newval && newlen)
> ret = tcp_set_allowed_congestion_control(tbl.data);
> kfree(tbl.data);
>
>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
This parallels previous fix by Sam Jansen.
^ permalink raw reply
* Re: [patch net-2.6.25][IPV4][FIB] fix fib_proc compilation error
From: Stephen Hemminger @ 2008-01-25 16:30 UTC (permalink / raw)
To: Daniel Lezcano; +Cc: netdev
In-Reply-To: <4799E489.90008@fr.ibm.com>
EBOGUSWHITESPACE
> +static inline void fib_proc_exit(struct net *net)
> +{
> + return ;
> +}
> #endif
Whole patch is unnecessary, since the fib_proc_init is gone in the tree
going into 2.6.25
^ permalink raw reply
* Re: [PATCH] [NET]: Remove PowerPC code from fec.c
From: Frans Pop @ 2008-01-25 16:31 UTC (permalink / raw)
To: Jochen Friedrich
Cc: galak, geert, gerg, jgarzik, linux-kernel, linux-m68k,
linuxppc-dev, netdev, scottwood, vitb
In-Reply-To: <479A087F.5010305@scram.de>
On Friday 25 January 2008, Jochen Friedrich wrote:
> > Jochen Friedrich wrote:
> >> +++ b/drivers/net/fec.c
> >> @@ -23,6 +23,9 @@
> >> *
> >> * Bug fixes and cleanup by Philippe De Muyter (phdm@macqel.be)
> >> * Copyright (c) 2004-2006 Macq Electronique SA.
> >> + *
> >> + * This driver is now only used on ColdFire processors. Remove conditional
> >> + * Powerpc code.
> >> */
> >
> > This comment makes sense for a changelog, but IMO it makes no sense at
> > all to add it to the file.
>
> I just added it to clarify this code is now only used on m68knommu
> (Coldfire). The comments on top are mailny about MPC860T CPUs (PowerPC),
> however the driver is no longer used for these CPUs.
>
> Maybe the wording should be changed to:
>
> This driver is now only used on ColdFire (m68knommu) processors.
> Conditional PowerPC code has been removed.
Yes, that certainly makes more sense, although IMHO the second sentence is
still somewhat redundant. (My problem was mainly with the second sentence.
I should have made that more clear, sorry.)
^ permalink raw reply
* Re: [patch net-2.6.25][IPV4][FIB] fix fib_proc compilation error
From: Stephen Hemminger @ 2008-01-25 16:32 UTC (permalink / raw)
To: Daniel Lezcano; +Cc: netdev
In-Reply-To: <4799E489.90008@fr.ibm.com>
Nevermind, Patch looks fine for 2.6.25, it wasn't needed for 2.6.24
Sorry, need more coffee...
--
Stephen Hemminger <stephen.hemminger@vyatta.com>
^ permalink raw reply
* Re: [PATCH 3/7 net-2.6.25] [IPV4]: Prohibit assignment of 0.0.0.0 as interface address.
From: Stephen Hemminger @ 2008-01-25 16:34 UTC (permalink / raw)
To: Denis V. Lunev; +Cc: netdev
In-Reply-To: <1201269123-20378-3-git-send-email-den@openvz.org>
On Fri, 25 Jan 2008 16:51:59 +0300
"Denis V. Lunev" <den@openvz.org> wrote:
> I could hardly imagine why sombady needs to assign 0.0.0.0 as an interface
> address or interface destination address. The kernel will behave in a strage
> way in several places if this is possible, as ifa_local != 0 is considered
> as initialized/non-initialized state of the ifa.
>
> Signed-off-by: Denis V. Lunev <den@openvz.org>
>
This is used as a way to bring device up in lots of existing documentation.
So please don't change.
--
Stephen Hemminger <stephen.hemminger@vyatta.com>
^ permalink raw reply
* Re: [Bugme-new] [Bug 9811] New: Loopback address to eth0 interface changes scope permanently
From: Bjørn Mork @ 2008-01-25 16:34 UTC (permalink / raw)
To: netdev; +Cc: bugme-daemon
In-Reply-To: <20080125025016.f6b08754.akpm@linux-foundation.org>
>> Latest working kernel version: none
>> Earliest failing kernel version: 2.6.18 (verified, but most likely "any")
Looks like this was introduced in 2.1.68, which makes it a 10 year old
bug :-)
http://www.linuxhq.com/kernel/v2.1/68/net/ipv4/devinet.c
Bjørn
--
You sound like a real wimp
^ permalink raw reply
* Re: [PATCH 1/5] netns netfilter: per-netns ip6tables
From: Alexey Dobriyan @ 2008-01-25 16:39 UTC (permalink / raw)
To: Patrick McHardy; +Cc: netdev, netfilter-devel, devel
In-Reply-To: <4798CBF1.205@trash.net>
On Thu, Jan 24, 2008 at 06:33:37PM +0100, Patrick McHardy wrote:
> Alexey Dobriyan wrote:
> >* Propagate netns from userspace down to xt_find_table_lock()
> >* Register ip6 tables in netns (modules still use init_net)
> >
> >Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
> >---
> >
> > include/linux/netfilter_ipv6/ip6_tables.h | 3 +
> > net/ipv6/netfilter/ip6_tables.c | 50
> > +++++++++++++++---------------
> > net/ipv6/netfilter/ip6table_filter.c | 2 -
> > net/ipv6/netfilter/ip6table_mangle.c | 2 -
> > net/ipv6/netfilter/ip6table_raw.c | 2 -
> > 5 files changed, 31 insertions(+), 28 deletions(-)
>
> This adds checkpatch warnings:
>
> WARNING: line over 80 characters
> #96: FILE: net/ipv6/netfilter/ip6_tables.c:1361:
> +do_add_counters(struct net *net, void __user *user, unsigned int len,
> int compat)
>
> WARNING: line over 80 characters
> #229: FILE: net/ipv6/netfilter/ip6table_filter.c:135:
> + packet_filter = ip6t_register_table(&init_net, &__packet_filter,
> &initial_table.repl);
>
> WARNING: line over 80 characters
> #242: FILE: net/ipv6/netfilter/ip6table_mangle.c:167:
> + packet_mangler = ip6t_register_table(&init_net,
> &__packet_mangler, &initial_table.repl);
>
> WARNING: line over 80 characters
> #255: FILE: net/ipv6/netfilter/ip6table_raw.c:80:
> + packet_raw = ip6t_register_table(&init_net, &__packet_raw,
> &initial_table.repl);
>
> ERROR: Missing Signed-off-by: line(s)
Well, SOB line definitely was present.
> total: 1 errors, 4 warnings, 214 lines checked
>
> I'll fix them up, lets hope that it doesn't cause clashes
> with the following patches.
OK. It was quantum coding style violation. :-)
^ permalink raw reply
* Re: [PATCH] SCTP: Fix kernel panic while received AUTH chunk while enabled auth
From: Vlad Yasevich @ 2008-01-25 16:41 UTC (permalink / raw)
To: Wei Yongjun; +Cc: netdev, lksctp-developers
In-Reply-To: <4794C51B.8040904@cn.fujitsu.com>
Sorry for the delay. Was on vacation without net access.
Wei Yongjun wrote:
>
>
> This patch fix this probleam to treat AUTH chunk as unknow chunk if peer
> has initialized with no auth capable.
>
> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com>
>
> --- a/net/sctp/sm_statefuns.c 2008-01-21 00:03:25.000000000 -0500
> +++ b/net/sctp/sm_statefuns.c 2008-01-21 05:14:08.000000000 -0500
> @@ -3785,6 +3785,10 @@ sctp_disposition_t sctp_sf_eat_auth(cons
> struct sctp_chunk *err_chunk;
> sctp_ierror_t error;
>
> + /* Make sure that the peer has AUTH capable */
> + if (!asoc->peer.auth_capable)
> + return sctp_sf_unk_chunk(ep, asoc, type, arg, commands);
> +
> if (!sctp_vtag_verify(chunk, asoc)) {
> sctp_add_cmd_sf(commands, SCTP_CMD_REPORT_BAD_TAG,
> SCTP_NULL());
>
>
>
^ permalink raw reply
* [PATCH 1/3] netns netfilter: semi-rewrite of /proc/net/foo_tables_*
From: Alexey Dobriyan @ 2008-01-25 16:43 UTC (permalink / raw)
To: Patrick McHardy; +Cc: netdev, netfilter-devel, devel
Argh, there are many small but still wrong things with /proc/net/*_tables_*
so I decided to do overhaul simultaneously making it more suitable for
per-netns /proc/net/*_tables_* implementation.
Fix
a) xt_get_idx() duplicating now standard seq_list_start/seq_list_next
iterators
b) tables/matches/targets list was chosen again and again on every ->next
c) multiple useless "af >= NPROTO" checks -- we simple don't supply invalid
AFs there and registration function should BUG_ON instead.
Regardless, the one in ->next() is the most useless -- ->next doesn't
run at all if ->start fails.
d) Don't use mutex_lock_interruptible() -- it can fail and ->stop is
executed even if ->start failed, so unlock without lock is possible.
As side effect, streamline code by splitting xt_tgt_ops into xt_target_ops,
xt_matches_ops, xt_tables_ops.
xt_tables_ops hooks will be changed by per-netns code. Code of
xt_matches_ops, xt_target_ops is identical except the list chosen for
iterating, but I think consolidating code for two files not worth it
given "<< 16" hacks needed for it.
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---
net/netfilter/x_tables.c | 224 ++++++++++++++++++++++++++++++-----------------
1 file changed, 145 insertions(+), 79 deletions(-)
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -726,124 +726,190 @@ void *xt_unregister_table(struct xt_table *table)
EXPORT_SYMBOL_GPL(xt_unregister_table);
#ifdef CONFIG_PROC_FS
-static struct list_head *xt_get_idx(struct list_head *list, struct seq_file *seq, loff_t pos)
+static void *xt_table_seq_start(struct seq_file *seq, loff_t *pos)
{
- struct list_head *head = list->next;
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
- if (!head || list_empty(list))
- return NULL;
+ mutex_lock(&xt[af].mutex);
+ return seq_list_start(&init_net.xt.tables[af], *pos);
+}
- while (pos && (head = head->next)) {
- if (head == list)
- return NULL;
- pos--;
- }
- return pos ? NULL : head;
-}
-
-static struct list_head *type2list(u_int16_t af, u_int16_t type)
-{
- struct list_head *list;
-
- switch (type) {
- case TARGET:
- list = &xt[af].target;
- break;
- case MATCH:
- list = &xt[af].match;
- break;
- case TABLE:
- list = &init_net.xt.tables[af];
- break;
- default:
- list = NULL;
- break;
- }
+static void *xt_table_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
- return list;
+ return seq_list_next(v, &init_net.xt.tables[af], pos);
}
-static void *xt_tgt_seq_start(struct seq_file *seq, loff_t *pos)
+static void xt_table_seq_stop(struct seq_file *seq, void *v)
{
- struct proc_dir_entry *pde = (struct proc_dir_entry *) seq->private;
- u_int16_t af = (unsigned long)pde->data & 0xffff;
- u_int16_t type = (unsigned long)pde->data >> 16;
- struct list_head *list;
+ struct proc_dir_entry *pde = seq->private;
+ u_int16_t af = (unsigned long)pde->data;
- if (af >= NPROTO)
- return NULL;
+ mutex_unlock(&xt[af].mutex);
+}
- list = type2list(af, type);
- if (!list)
- return NULL;
+static int xt_table_seq_show(struct seq_file *seq, void *v)
+{
+ struct xt_table *table = list_entry(v, struct xt_table, list);
- if (mutex_lock_interruptible(&xt[af].mutex) != 0)
- return NULL;
+ if (strlen(table->name))
+ return seq_printf(seq, "%s\n", table->name);
+ else
+ return 0;
+}
+
+static const struct seq_operations xt_table_seq_ops = {
+ .start = xt_table_seq_start,
+ .next = xt_table_seq_next,
+ .stop = xt_table_seq_stop,
+ .show = xt_table_seq_show,
+};
+
+static int xt_table_open(struct inode *inode, struct file *file)
+{
+ int ret;
+
+ ret = seq_open(file, &xt_table_seq_ops);
+ if (!ret) {
+ struct seq_file *seq = file->private_data;
- return xt_get_idx(list, seq, *pos);
+ seq->private = PDE(inode);
+ }
+ return ret;
}
-static void *xt_tgt_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+static const struct file_operations xt_table_ops = {
+ .owner = THIS_MODULE,
+ .open = xt_table_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static void *xt_match_seq_start(struct seq_file *seq, loff_t *pos)
{
- struct proc_dir_entry *pde = seq->private;
- u_int16_t af = (unsigned long)pde->data & 0xffff;
- u_int16_t type = (unsigned long)pde->data >> 16;
- struct list_head *list;
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
- if (af >= NPROTO)
- return NULL;
+ mutex_lock(&xt[af].mutex);
+ return seq_list_start(&xt[af].match, *pos);
+}
- list = type2list(af, type);
- if (!list)
- return NULL;
+static void *xt_match_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
- (*pos)++;
- return xt_get_idx(list, seq, *pos);
+ return seq_list_next(v, &xt[af].match, pos);
}
-static void xt_tgt_seq_stop(struct seq_file *seq, void *v)
+static void xt_match_seq_stop(struct seq_file *seq, void *v)
{
struct proc_dir_entry *pde = seq->private;
- u_int16_t af = (unsigned long)pde->data & 0xffff;
+ u_int16_t af = (unsigned long)pde->data;
mutex_unlock(&xt[af].mutex);
}
-static int xt_name_seq_show(struct seq_file *seq, void *v)
+static int xt_match_seq_show(struct seq_file *seq, void *v)
{
- char *name = (char *)v + sizeof(struct list_head);
+ struct xt_match *match = list_entry(v, struct xt_match, list);
- if (strlen(name))
- return seq_printf(seq, "%s\n", name);
+ if (strlen(match->name))
+ return seq_printf(seq, "%s\n", match->name);
else
return 0;
}
-static const struct seq_operations xt_tgt_seq_ops = {
- .start = xt_tgt_seq_start,
- .next = xt_tgt_seq_next,
- .stop = xt_tgt_seq_stop,
- .show = xt_name_seq_show,
+static const struct seq_operations xt_match_seq_ops = {
+ .start = xt_match_seq_start,
+ .next = xt_match_seq_next,
+ .stop = xt_match_seq_stop,
+ .show = xt_match_seq_show,
};
-static int xt_tgt_open(struct inode *inode, struct file *file)
+static int xt_match_open(struct inode *inode, struct file *file)
{
int ret;
- ret = seq_open(file, &xt_tgt_seq_ops);
+ ret = seq_open(file, &xt_match_seq_ops);
if (!ret) {
struct seq_file *seq = file->private_data;
- struct proc_dir_entry *pde = PDE(inode);
- seq->private = pde;
+ seq->private = PDE(inode);
}
+ return ret;
+}
+
+static const struct file_operations xt_match_ops = {
+ .owner = THIS_MODULE,
+ .open = xt_match_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static void *xt_target_seq_start(struct seq_file *seq, loff_t *pos)
+{
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
+
+ mutex_lock(&xt[af].mutex);
+ return seq_list_start(&xt[af].target, *pos);
+}
+
+static void *xt_target_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
+ u_int16_t af = (unsigned long)pde->data;
+
+ return seq_list_next(v, &xt[af].target, pos);
+}
+
+static void xt_target_seq_stop(struct seq_file *seq, void *v)
+{
+ struct proc_dir_entry *pde = seq->private;
+ u_int16_t af = (unsigned long)pde->data;
+
+ mutex_unlock(&xt[af].mutex);
+}
+
+static int xt_target_seq_show(struct seq_file *seq, void *v)
+{
+ struct xt_target *target = list_entry(v, struct xt_target, list);
+
+ if (strlen(target->name))
+ return seq_printf(seq, "%s\n", target->name);
+ else
+ return 0;
+}
+
+static const struct seq_operations xt_target_seq_ops = {
+ .start = xt_target_seq_start,
+ .next = xt_target_seq_next,
+ .stop = xt_target_seq_stop,
+ .show = xt_target_seq_show,
+};
+
+static int xt_target_open(struct inode *inode, struct file *file)
+{
+ int ret;
+ ret = seq_open(file, &xt_target_seq_ops);
+ if (!ret) {
+ struct seq_file *seq = file->private_data;
+
+ seq->private = PDE(inode);
+ }
return ret;
}
-static const struct file_operations xt_file_ops = {
+static const struct file_operations xt_target_ops = {
.owner = THIS_MODULE,
- .open = xt_tgt_open,
+ .open = xt_target_open,
.read = seq_read,
.llseek = seq_lseek,
.release = seq_release,
@@ -869,25 +935,25 @@ int xt_proto_init(int af)
#ifdef CONFIG_PROC_FS
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TABLES, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_file_ops);
+ proc = proc_net_fops_create(&init_net, buf, 0440, &xt_table_ops);
if (!proc)
goto out;
- proc->data = (void *) ((unsigned long) af | (TABLE << 16));
+ proc->data = (void *)(unsigned long)af;
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_MATCHES, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_file_ops);
+ proc = proc_net_fops_create(&init_net, buf, 0440, &xt_match_ops);
if (!proc)
goto out_remove_tables;
- proc->data = (void *) ((unsigned long) af | (MATCH << 16));
+ proc->data = (void *)(unsigned long)af;
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TARGETS, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_file_ops);
+ proc = proc_net_fops_create(&init_net, buf, 0440, &xt_target_ops);
if (!proc)
goto out_remove_matches;
- proc->data = (void *) ((unsigned long) af | (TARGET << 16));
+ proc->data = (void *)(unsigned long)af;
#endif
return 0;
^ permalink raw reply
* [PATCH 2/3] netns netfilter: netns propagation for /proc/net/*_tables_names
From: Alexey Dobriyan @ 2008-01-25 16:44 UTC (permalink / raw)
To: Patrick McHardy; +Cc: netdev, netfilter-devel, devel
Propagate netns together with AF down to ->start/->next/->stop
iterators. Choose table based on netns and AF for showing.
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---
net/netfilter/x_tables.c | 31 +++++++++++++++++++------------
1 file changed, 19 insertions(+), 12 deletions(-)
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -726,27 +726,33 @@ void *xt_unregister_table(struct xt_table *table)
EXPORT_SYMBOL_GPL(xt_unregister_table);
#ifdef CONFIG_PROC_FS
+struct xt_names_priv {
+ struct seq_net_private p;
+ int af;
+};
static void *xt_table_seq_start(struct seq_file *seq, loff_t *pos)
{
- struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
- u_int16_t af = (unsigned long)pde->data;
+ struct xt_names_priv *priv = seq->private;
+ struct net *net = priv->p.net;
+ int af = priv->af;
mutex_lock(&xt[af].mutex);
- return seq_list_start(&init_net.xt.tables[af], *pos);
+ return seq_list_start(&net->xt.tables[af], *pos);
}
static void *xt_table_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct proc_dir_entry *pde = (struct proc_dir_entry *)seq->private;
- u_int16_t af = (unsigned long)pde->data;
+ struct xt_names_priv *priv = seq->private;
+ struct net *net = priv->p.net;
+ int af = priv->af;
- return seq_list_next(v, &init_net.xt.tables[af], pos);
+ return seq_list_next(v, &net->xt.tables[af], pos);
}
static void xt_table_seq_stop(struct seq_file *seq, void *v)
{
- struct proc_dir_entry *pde = seq->private;
- u_int16_t af = (unsigned long)pde->data;
+ struct xt_names_priv *priv = seq->private;
+ int af = priv->af;
mutex_unlock(&xt[af].mutex);
}
@@ -771,12 +777,13 @@ static const struct seq_operations xt_table_seq_ops = {
static int xt_table_open(struct inode *inode, struct file *file)
{
int ret;
+ struct xt_names_priv *priv;
- ret = seq_open(file, &xt_table_seq_ops);
+ ret = seq_open_net(inode, file, &xt_table_seq_ops,
+ sizeof(struct xt_names_priv));
if (!ret) {
- struct seq_file *seq = file->private_data;
-
- seq->private = PDE(inode);
+ priv = ((struct seq_file *)file->private_data)->private;
+ priv->af = (unsigned long)PDE(inode)->data;
}
return ret;
}
^ permalink raw reply
* [PATCH 3/3] netns netfilter: create per-netns /proc/net/*_tables_*
From: Alexey Dobriyan @ 2008-01-25 16:45 UTC (permalink / raw)
To: Patrick McHardy; +Cc: netdev, netfilter-devel, devel
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---
include/linux/netfilter/x_tables.h | 4 ++--
net/ipv4/netfilter/arp_tables.c | 21 ++++++++++++++++++---
net/ipv4/netfilter/ip_tables.c | 21 ++++++++++++++++++---
net/ipv6/netfilter/ip6_tables.c | 22 +++++++++++++++++++---
net/netfilter/x_tables.c | 20 ++++++++++----------
5 files changed, 67 insertions(+), 21 deletions(-)
--- a/include/linux/netfilter/x_tables.h
+++ b/include/linux/netfilter/x_tables.h
@@ -357,8 +357,8 @@ extern struct xt_table *xt_find_table_lock(struct net *net, int af,
const char *name);
extern void xt_table_unlock(struct xt_table *t);
-extern int xt_proto_init(int af);
-extern void xt_proto_fini(int af);
+extern int xt_proto_init(struct net *net, int af);
+extern void xt_proto_fini(struct net *net, int af);
extern struct xt_table_info *xt_alloc_table_info(unsigned int size);
extern void xt_free_table_info(struct xt_table_info *info);
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -1822,11 +1822,26 @@ static struct nf_sockopt_ops arpt_sockopts = {
.owner = THIS_MODULE,
};
+static int __net_init arp_tables_net_init(struct net *net)
+{
+ return xt_proto_init(net, NF_ARP);
+}
+
+static void __net_exit arp_tables_net_exit(struct net *net)
+{
+ xt_proto_fini(net, NF_ARP);
+}
+
+static struct pernet_operations arp_tables_net_ops = {
+ .init = arp_tables_net_init,
+ .exit = arp_tables_net_exit,
+};
+
static int __init arp_tables_init(void)
{
int ret;
- ret = xt_proto_init(NF_ARP);
+ ret = register_pernet_subsys(&arp_tables_net_ops);
if (ret < 0)
goto err1;
@@ -1851,7 +1866,7 @@ err4:
err3:
xt_unregister_target(&arpt_standard_target);
err2:
- xt_proto_fini(NF_ARP);
+ unregister_pernet_subsys(&arp_tables_net_ops);
err1:
return ret;
}
@@ -1861,7 +1876,7 @@ static void __exit arp_tables_fini(void)
nf_unregister_sockopt(&arpt_sockopts);
xt_unregister_target(&arpt_error_target);
xt_unregister_target(&arpt_standard_target);
- xt_proto_fini(NF_ARP);
+ unregister_pernet_subsys(&arp_tables_net_ops);
}
EXPORT_SYMBOL(arpt_register_table);
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -2211,11 +2211,26 @@ static struct xt_match icmp_matchstruct __read_mostly = {
.family = AF_INET,
};
+static int __net_init ip_tables_net_init(struct net *net)
+{
+ return xt_proto_init(net, AF_INET);
+}
+
+static void __net_exit ip_tables_net_exit(struct net *net)
+{
+ xt_proto_fini(net, AF_INET);
+}
+
+static struct pernet_operations ip_tables_net_ops = {
+ .init = ip_tables_net_init,
+ .exit = ip_tables_net_exit,
+};
+
static int __init ip_tables_init(void)
{
int ret;
- ret = xt_proto_init(AF_INET);
+ ret = register_pernet_subsys(&ip_tables_net_ops);
if (ret < 0)
goto err1;
@@ -2245,7 +2260,7 @@ err4:
err3:
xt_unregister_target(&ipt_standard_target);
err2:
- xt_proto_fini(AF_INET);
+ unregister_pernet_subsys(&ip_tables_net_ops);
err1:
return ret;
}
@@ -2258,7 +2273,7 @@ static void __exit ip_tables_fini(void)
xt_unregister_target(&ipt_error_target);
xt_unregister_target(&ipt_standard_target);
- xt_proto_fini(AF_INET);
+ unregister_pernet_subsys(&ip_tables_net_ops);
}
EXPORT_SYMBOL(ipt_register_table);
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -2235,11 +2235,26 @@ static struct xt_match icmp6_matchstruct __read_mostly = {
.family = AF_INET6,
};
+static int __net_init ip6_tables_net_init(struct net *net)
+{
+ return xt_proto_init(net, AF_INET6);
+}
+
+static void __net_exit ip6_tables_net_exit(struct net *net)
+{
+ xt_proto_fini(net, AF_INET6);
+}
+
+static struct pernet_operations ip6_tables_net_ops = {
+ .init = ip6_tables_net_init,
+ .exit = ip6_tables_net_exit,
+};
+
static int __init ip6_tables_init(void)
{
int ret;
- ret = xt_proto_init(AF_INET6);
+ ret = register_pernet_subsys(&ip6_tables_net_ops);
if (ret < 0)
goto err1;
@@ -2269,7 +2284,7 @@ err4:
err3:
xt_unregister_target(&ip6t_standard_target);
err2:
- xt_proto_fini(AF_INET6);
+ unregister_pernet_subsys(&ip6_tables_net_ops);
err1:
return ret;
}
@@ -2281,7 +2296,8 @@ static void __exit ip6_tables_fini(void)
xt_unregister_match(&icmp6_matchstruct);
xt_unregister_target(&ip6t_error_target);
xt_unregister_target(&ip6t_standard_target);
- xt_proto_fini(AF_INET6);
+
+ unregister_pernet_subsys(&ip6_tables_net_ops);
}
/*
--- a/net/netfilter/x_tables.c
+++ b/net/netfilter/x_tables.c
@@ -928,7 +928,7 @@ static const struct file_operations xt_target_ops = {
#endif /* CONFIG_PROC_FS */
-int xt_proto_init(int af)
+int xt_proto_init(struct net *net, int af)
{
#ifdef CONFIG_PROC_FS
char buf[XT_FUNCTION_MAXNAMELEN];
@@ -942,7 +942,7 @@ int xt_proto_init(int af)
#ifdef CONFIG_PROC_FS
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TABLES, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_table_ops);
+ proc = proc_net_fops_create(net, buf, 0440, &xt_table_ops);
if (!proc)
goto out;
proc->data = (void *)(unsigned long)af;
@@ -950,14 +950,14 @@ int xt_proto_init(int af)
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_MATCHES, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_match_ops);
+ proc = proc_net_fops_create(net, buf, 0440, &xt_match_ops);
if (!proc)
goto out_remove_tables;
proc->data = (void *)(unsigned long)af;
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TARGETS, sizeof(buf));
- proc = proc_net_fops_create(&init_net, buf, 0440, &xt_target_ops);
+ proc = proc_net_fops_create(net, buf, 0440, &xt_target_ops);
if (!proc)
goto out_remove_matches;
proc->data = (void *)(unsigned long)af;
@@ -969,34 +969,34 @@ int xt_proto_init(int af)
out_remove_matches:
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_MATCHES, sizeof(buf));
- proc_net_remove(&init_net, buf);
+ proc_net_remove(net, buf);
out_remove_tables:
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TABLES, sizeof(buf));
- proc_net_remove(&init_net, buf);
+ proc_net_remove(net, buf);
out:
return -1;
#endif
}
EXPORT_SYMBOL_GPL(xt_proto_init);
-void xt_proto_fini(int af)
+void xt_proto_fini(struct net *net, int af)
{
#ifdef CONFIG_PROC_FS
char buf[XT_FUNCTION_MAXNAMELEN];
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TABLES, sizeof(buf));
- proc_net_remove(&init_net, buf);
+ proc_net_remove(net, buf);
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_TARGETS, sizeof(buf));
- proc_net_remove(&init_net, buf);
+ proc_net_remove(net, buf);
strlcpy(buf, xt_prefix[af], sizeof(buf));
strlcat(buf, FORMAT_MATCHES, sizeof(buf));
- proc_net_remove(&init_net, buf);
+ proc_net_remove(net, buf);
#endif /*CONFIG_PROC_FS*/
}
EXPORT_SYMBOL_GPL(xt_proto_fini);
^ permalink raw reply
* Re: [Lksctp-developers] [PATCH] SCTP: Fix kernel panic while received AUTH chunk with BAD shared key identifier
From: Vlad Yasevich @ 2008-01-25 16:46 UTC (permalink / raw)
To: Neil Horman; +Cc: Wei Yongjun, netdev, lksctp-developers
In-Reply-To: <20080124121608.GA20633@hmsreliant.think-freely.org>
Neil Horman wrote:
> On Tue, Jan 22, 2008 at 05:29:20PM +0900, Wei Yongjun wrote:
>>
>> This patch fix this problem.
>>
>> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
>>
>> --- a/net/sctp/auth.c 2008-01-21 00:03:25.000000000 -0500
>> +++ b/net/sctp/auth.c 2008-01-21 21:31:47.000000000 -0500
>> @@ -420,15 +420,15 @@ struct sctp_shared_key *sctp_auth_get_sh
>> const struct sctp_association *asoc,
>> __u16 key_id)
>> {
>> - struct sctp_shared_key *key = NULL;
>> + struct sctp_shared_key *key;
>>
>> /* First search associations set of endpoint pair shared keys */
>> key_for_each(key, &asoc->endpoint_shared_keys) {
>> if (key->key_id == key_id)
>> - break;
>> + return key;
>> }
>>
>> - return key;
>> + return NULL;
>> }
>>
>> /*
>>
>
> FWIW, Ack from me. The assignment of NULL to key can safely be removed, since
> key_for_each (which is just list_for_each_entry under the covers does an initial
> assignment to key anyway).
>
> If the endpoint_shared_keys list is empty, or if the key_id being requested does
> not exist, the function as it currently stands returns the actuall list_head (in
> this case endpoint_shared_keys. Since that list_head isn't surrounded by an
> actuall data structure, the last iteration through list_for_each_entry will do a
> container_of on key, and we wind up returning a bogus pointer, instead of NULL,
> as we should. Wei's patch corrects that.
>
> Regards
> Neil
>
> Acked-by: Neil Horman <nhorman@tuxdriver.com>
>
Yep, the patch is correct.
Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com>
-vlad
^ permalink raw reply
* [patch 00/11][NETNS][IPV6] make a subset of the routing per namespace
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
This patchset makes a subset of the routing to be per
namespace. It takes into account fib tables, garbage
collecting timer, fib rules and route statistics.
The routing per namespace is not complete with this
patchset, so there are some interactions between the init
network namespace and other namespaces, but nothing
critical.
This patchset will be followed by another one which
will complete the routing per namespace. I just don't want
to flood netdev@ with a huge patchset.
--
^ permalink raw reply
* [patch 01/11][NETNS][IPV6] ip6_fib - dynamically allocate the tables
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-dynamic-allocation.patch --]
[-- Type: text/plain, Size: 4135 bytes --]
This patch changes the fib6 tables to be dynamically allocated.
That provides the ability to make several instances of them when
a new network namespace is created.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
net/ipv6/ip6_fib.c | 71 +++++++++++++++++++++++++++++++++++------------------
1 file changed, 48 insertions(+), 23 deletions(-)
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -166,20 +166,14 @@ static __inline__ void rt6_release(struc
dst_free(&rt->u.dst);
}
-static struct fib6_table fib6_main_tbl = {
- .tb6_id = RT6_TABLE_MAIN,
- .tb6_root = {
- .leaf = &ip6_null_entry,
- .fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO,
- },
-};
+static struct fib6_table *fib6_main_tbl;
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
#define FIB_TABLE_HASHSZ 256
#else
#define FIB_TABLE_HASHSZ 1
#endif
-static struct hlist_head fib_table_hash[FIB_TABLE_HASHSZ];
+static struct hlist_head *fib_table_hash;
static void fib6_link_table(struct fib6_table *tb)
{
@@ -201,13 +195,8 @@ static void fib6_link_table(struct fib6_
}
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
-static struct fib6_table fib6_local_tbl = {
- .tb6_id = RT6_TABLE_LOCAL,
- .tb6_root = {
- .leaf = &ip6_null_entry,
- .fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO,
- },
-};
+
+static struct fib6_table *fib6_local_tbl;
static struct fib6_table *fib6_alloc_table(u32 id)
{
@@ -263,8 +252,8 @@ struct fib6_table *fib6_get_table(u32 id
static void __init fib6_tables_init(void)
{
- fib6_link_table(&fib6_main_tbl);
- fib6_link_table(&fib6_local_tbl);
+ fib6_link_table(fib6_main_tbl);
+ fib6_link_table(fib6_local_tbl);
}
#else
@@ -276,18 +265,18 @@ struct fib6_table *fib6_new_table(u32 id
struct fib6_table *fib6_get_table(u32 id)
{
- return &fib6_main_tbl;
+ return fib6_main_tbl;
}
struct dst_entry *fib6_rule_lookup(struct flowi *fl, int flags,
pol_lookup_t lookup)
{
- return (struct dst_entry *) lookup(&fib6_main_tbl, fl, flags);
+ return (struct dst_entry *) lookup(fib6_main_tbl, fl, flags);
}
static void __init fib6_tables_init(void)
{
- fib6_link_table(&fib6_main_tbl);
+ fib6_link_table(fib6_main_tbl);
}
#endif
@@ -1479,22 +1468,53 @@ void fib6_run_gc(unsigned long dummy)
int __init fib6_init(void)
{
- int ret;
+ int ret = -ENOMEM;
fib6_node_kmem = kmem_cache_create("fib6_nodes",
sizeof(struct fib6_node),
0, SLAB_HWCACHE_ALIGN,
NULL);
if (!fib6_node_kmem)
- return -ENOMEM;
+ goto out;
+
+ fib_table_hash = kzalloc(sizeof(*fib_table_hash)*FIB_TABLE_HASHSZ,
+ GFP_KERNEL);
+ if (!fib_table_hash)
+ goto out_kmem_cache_create;
+
+ fib6_main_tbl = kzalloc(sizeof(*fib6_main_tbl), GFP_KERNEL);
+ if (!fib6_main_tbl)
+ goto out_fib_table_hash;
+
+ fib6_main_tbl->tb6_id = RT6_TABLE_MAIN;
+ fib6_main_tbl->tb6_root.leaf = &ip6_null_entry;
+ fib6_main_tbl->tb6_root.fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
+
+#ifdef CONFIG_IPV6_MULTIPLE_TABLES
+ fib6_local_tbl = kzalloc(sizeof(*fib6_local_tbl), GFP_KERNEL);
+ if (!fib6_local_tbl)
+ goto out_fib6_main_tbl;
+
+ fib6_local_tbl->tb6_id = RT6_TABLE_LOCAL;
+ fib6_local_tbl->tb6_root.leaf = &ip6_null_entry;
+ fib6_local_tbl->tb6_root.fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
+#endif
fib6_tables_init();
ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
if (ret)
- goto out_kmem_cache_create;
+ goto out_fib6_local_tbl;
out:
return ret;
+out_fib6_local_tbl:
+#ifdef CONFIG_IPV6_MULTIPLE_TABLES
+ kfree(fib6_local_tbl);
+out_fib6_main_tbl:
+#endif
+ kfree(fib6_main_tbl);
+out_fib_table_hash:
+ kfree(fib_table_hash);
out_kmem_cache_create:
kmem_cache_destroy(fib6_node_kmem);
goto out;
@@ -1503,5 +1523,10 @@ out_kmem_cache_create:
void fib6_gc_cleanup(void)
{
del_timer(&ip6_fib_timer);
+#ifdef CONFIG_IPV6_MULTIPLE_TABLES
+ kfree(fib6_local_tbl);
+#endif
+ kfree(fib6_main_tbl);
+ kfree(fib_table_hash);
kmem_cache_destroy(fib6_node_kmem);
}
--
^ permalink raw reply
* [patch 02/11][NETNS][IPV6] ip6_fib - make the tables per namespace
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-per-network-namespace.patch --]
[-- Type: text/plain, Size: 14465 bytes --]
The fib table for ipv6 are moved to the network namespace structure.
All references to them are made relatively to the network namespace.
All external calls to the ip6_fib functions taking the network namespace
parameter are made using the init_net variable, so the ip6_fib engine is
ready for the namespaces but the callers not yet.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
include/net/ip6_fib.h | 9 +-
include/net/netns/ipv6.h | 6 +
net/ipv6/fib6_rules.c | 8 +-
net/ipv6/ip6_fib.c | 161 +++++++++++++++++++++++++++--------------------
net/ipv6/route.c | 22 +++---
5 files changed, 120 insertions(+), 86 deletions(-)
Index: net-2.6.25/include/net/ip6_fib.h
===================================================================
--- net-2.6.25.orig/include/net/ip6_fib.h
+++ net-2.6.25/include/net/ip6_fib.h
@@ -194,10 +194,11 @@ typedef struct rt6_info *(*pol_lookup_t)
* exported functions
*/
-extern struct fib6_table * fib6_get_table(u32 id);
-extern struct fib6_table * fib6_new_table(u32 id);
-extern struct dst_entry * fib6_rule_lookup(struct flowi *fl, int flags,
- pol_lookup_t lookup);
+extern struct fib6_table *fib6_get_table(struct net *net, u32 id);
+extern struct fib6_table *fib6_new_table(struct net *net, u32 id);
+extern struct dst_entry *fib6_rule_lookup(struct net *net,
+ struct flowi *fl, int flags,
+ pol_lookup_t lookup);
extern struct fib6_node *fib6_lookup(struct fib6_node *root,
struct in6_addr *daddr,
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
--- net-2.6.25.orig/include/net/netns/ipv6.h
+++ net-2.6.25/include/net/netns/ipv6.h
@@ -31,5 +31,11 @@ struct netns_ipv6 {
struct ipv6_devconf *devconf_all;
struct ipv6_devconf *devconf_dflt;
struct netns_frags frags;
+
+ struct hlist_head *fib_table_hash;
+ struct fib6_table *fib6_main_tbl;
+#ifdef CONFIG_IPV6_MULTIPLE_TABLES
+ struct fib6_table *fib6_local_tbl;
+#endif
};
#endif
Index: net-2.6.25/net/ipv6/fib6_rules.c
===================================================================
--- net-2.6.25.orig/net/ipv6/fib6_rules.c
+++ net-2.6.25/net/ipv6/fib6_rules.c
@@ -31,8 +31,8 @@ struct fib6_rule
static struct fib_rules_ops fib6_rules_ops;
-struct dst_entry *fib6_rule_lookup(struct flowi *fl, int flags,
- pol_lookup_t lookup)
+struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi *fl,
+ int flags, pol_lookup_t lookup)
{
struct fib_lookup_arg arg = {
.lookup_ptr = lookup,
@@ -71,7 +71,7 @@ static int fib6_rule_action(struct fib_r
goto discard_pkt;
}
- table = fib6_get_table(rule->table);
+ table = fib6_get_table(&init_net, rule->table);
if (table)
rt = lookup(table, flp, flags);
@@ -151,7 +151,7 @@ static int fib6_rule_configure(struct fi
if (rule->table == RT6_TABLE_UNSPEC)
goto errout;
- if (fib6_new_table(rule->table) == NULL) {
+ if (fib6_new_table(&init_net, rule->table) == NULL) {
err = -ENOBUFS;
goto errout;
}
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -166,16 +166,13 @@ static __inline__ void rt6_release(struc
dst_free(&rt->u.dst);
}
-static struct fib6_table *fib6_main_tbl;
-
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
#define FIB_TABLE_HASHSZ 256
#else
#define FIB_TABLE_HASHSZ 1
#endif
-static struct hlist_head *fib_table_hash;
-static void fib6_link_table(struct fib6_table *tb)
+static void fib6_link_table(struct net *net, struct fib6_table *tb)
{
unsigned int h;
@@ -191,13 +188,11 @@ static void fib6_link_table(struct fib6_
* No protection necessary, this is the only list mutatation
* operation, tables never disappear once they exist.
*/
- hlist_add_head_rcu(&tb->tb6_hlist, &fib_table_hash[h]);
+ hlist_add_head_rcu(&tb->tb6_hlist, &net->ipv6.fib_table_hash[h]);
}
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
-static struct fib6_table *fib6_local_tbl;
-
static struct fib6_table *fib6_alloc_table(u32 id)
{
struct fib6_table *table;
@@ -212,26 +207,27 @@ static struct fib6_table *fib6_alloc_tab
return table;
}
-struct fib6_table *fib6_new_table(u32 id)
+struct fib6_table *fib6_new_table(struct net *net, u32 id)
{
struct fib6_table *tb;
if (id == 0)
id = RT6_TABLE_MAIN;
- tb = fib6_get_table(id);
+ tb = fib6_get_table(net, id);
if (tb)
return tb;
tb = fib6_alloc_table(id);
if (tb != NULL)
- fib6_link_table(tb);
+ fib6_link_table(net, tb);
return tb;
}
-struct fib6_table *fib6_get_table(u32 id)
+struct fib6_table *fib6_get_table(struct net *net, u32 id)
{
struct fib6_table *tb;
+ struct hlist_head *head;
struct hlist_node *node;
unsigned int h;
@@ -239,7 +235,8 @@ struct fib6_table *fib6_get_table(u32 id
id = RT6_TABLE_MAIN;
h = id & (FIB_TABLE_HASHSZ - 1);
rcu_read_lock();
- hlist_for_each_entry_rcu(tb, node, &fib_table_hash[h], tb6_hlist) {
+ head = &net->ipv6.fib_table_hash[h];
+ hlist_for_each_entry_rcu(tb, node, head, tb6_hlist) {
if (tb->tb6_id == id) {
rcu_read_unlock();
return tb;
@@ -250,33 +247,32 @@ struct fib6_table *fib6_get_table(u32 id
return NULL;
}
-static void __init fib6_tables_init(void)
+static void fib6_tables_init(struct net *net)
{
- fib6_link_table(fib6_main_tbl);
- fib6_link_table(fib6_local_tbl);
+ fib6_link_table(net, net->ipv6.fib6_main_tbl);
+ fib6_link_table(net, net->ipv6.fib6_local_tbl);
}
-
#else
-struct fib6_table *fib6_new_table(u32 id)
+struct fib6_table *fib6_new_table(struct net *net, u32 id)
{
- return fib6_get_table(id);
+ return fib6_get_table(net, id);
}
-struct fib6_table *fib6_get_table(u32 id)
+struct fib6_table *fib6_get_table(struct net *net, u32 id)
{
- return fib6_main_tbl;
+ return net->ipv6.fib6_main_tbl;
}
-struct dst_entry *fib6_rule_lookup(struct flowi *fl, int flags,
- pol_lookup_t lookup)
+struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi *fl,
+ int flags, pol_lookup_t lookup)
{
- return (struct dst_entry *) lookup(fib6_main_tbl, fl, flags);
+ return (struct dst_entry *) lookup(net->ipv6.fib6_main_tbl, fl, flags);
}
-static void __init fib6_tables_init(void)
+static void fib6_tables_init(struct net *net)
{
- fib6_link_table(fib6_main_tbl);
+ fib6_link_table(net, net->ipv6.fib6_main_tbl);
}
#endif
@@ -357,11 +353,9 @@ static int inet6_dump_fib(struct sk_buff
struct fib6_walker_t *w;
struct fib6_table *tb;
struct hlist_node *node;
+ struct hlist_head *head;
int res = 0;
- if (net != &init_net)
- return 0;
-
s_h = cb->args[0];
s_e = cb->args[1];
@@ -390,7 +384,8 @@ static int inet6_dump_fib(struct sk_buff
for (h = s_h; h < FIB_TABLE_HASHSZ; h++, s_e = 0) {
e = 0;
- hlist_for_each_entry(tb, node, &fib_table_hash[h], tb6_hlist) {
+ head = &net->ipv6.fib_table_hash[h];
+ hlist_for_each_entry(tb, node, head, tb6_hlist) {
if (e < s_e)
goto next;
res = fib6_dump_table(tb, skb, cb);
@@ -1360,12 +1355,13 @@ void fib6_clean_all(int (*func)(struct r
{
struct fib6_table *table;
struct hlist_node *node;
+ struct hlist_head *head;
unsigned int h;
rcu_read_lock();
for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
- hlist_for_each_entry_rcu(table, node, &fib_table_hash[h],
- tb6_hlist) {
+ head = &init_net.ipv6.fib_table_hash[h];
+ hlist_for_each_entry_rcu(table, node, head, tb6_hlist) {
write_lock_bh(&table->tb6_lock);
fib6_clean_tree(&table->tb6_root, func, prune, arg);
write_unlock_bh(&table->tb6_lock);
@@ -1466,55 +1462,88 @@ void fib6_run_gc(unsigned long dummy)
spin_unlock_bh(&fib6_gc_lock);
}
-int __init fib6_init(void)
+static int fib6_net_init(struct net *net)
{
- int ret = -ENOMEM;
- fib6_node_kmem = kmem_cache_create("fib6_nodes",
- sizeof(struct fib6_node),
- 0, SLAB_HWCACHE_ALIGN,
- NULL);
- if (!fib6_node_kmem)
- goto out;
+ int ret;
- fib_table_hash = kzalloc(sizeof(*fib_table_hash)*FIB_TABLE_HASHSZ,
- GFP_KERNEL);
- if (!fib_table_hash)
- goto out_kmem_cache_create;
+ ret = -ENOMEM;
+ net->ipv6.fib_table_hash =
+ kzalloc(sizeof(*net->ipv6.fib_table_hash)*FIB_TABLE_HASHSZ,
+ GFP_KERNEL);
+ if (!net->ipv6.fib_table_hash)
+ goto out;
- fib6_main_tbl = kzalloc(sizeof(*fib6_main_tbl), GFP_KERNEL);
- if (!fib6_main_tbl)
+ net->ipv6.fib6_main_tbl = kzalloc(sizeof(*net->ipv6.fib6_main_tbl),
+ GFP_KERNEL);
+ if (!net->ipv6.fib6_main_tbl)
goto out_fib_table_hash;
- fib6_main_tbl->tb6_id = RT6_TABLE_MAIN;
- fib6_main_tbl->tb6_root.leaf = &ip6_null_entry;
- fib6_main_tbl->tb6_root.fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
+ net->ipv6.fib6_main_tbl->tb6_id = RT6_TABLE_MAIN;
+ net->ipv6.fib6_main_tbl->tb6_root.leaf = &ip6_null_entry;
+ net->ipv6.fib6_main_tbl->tb6_root.fn_flags =
+ RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
- fib6_local_tbl = kzalloc(sizeof(*fib6_local_tbl), GFP_KERNEL);
- if (!fib6_local_tbl)
+ net->ipv6.fib6_local_tbl = kzalloc(sizeof(*net->ipv6.fib6_local_tbl),
+ GFP_KERNEL);
+ if (!net->ipv6.fib6_local_tbl)
goto out_fib6_main_tbl;
-
- fib6_local_tbl->tb6_id = RT6_TABLE_LOCAL;
- fib6_local_tbl->tb6_root.leaf = &ip6_null_entry;
- fib6_local_tbl->tb6_root.fn_flags = RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
+ net->ipv6.fib6_local_tbl->tb6_id = RT6_TABLE_LOCAL;
+ net->ipv6.fib6_local_tbl->tb6_root.leaf = &ip6_null_entry;
+ net->ipv6.fib6_local_tbl->tb6_root.fn_flags =
+ RTN_ROOT | RTN_TL_ROOT | RTN_RTINFO;
#endif
+ fib6_tables_init(net);
- fib6_tables_init();
-
- ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
- if (ret)
- goto out_fib6_local_tbl;
+ ret = 0;
out:
return ret;
-out_fib6_local_tbl:
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
- kfree(fib6_local_tbl);
out_fib6_main_tbl:
+ kfree(net->ipv6.fib6_main_tbl);
#endif
- kfree(fib6_main_tbl);
out_fib_table_hash:
- kfree(fib_table_hash);
+ kfree(net->ipv6.fib_table_hash);
+ goto out;
+ }
+
+static void fib6_net_exit(struct net *net)
+{
+#ifdef CONFIG_IPV6_MULTIPLE_TABLES
+ kfree(net->ipv6.fib6_local_tbl);
+#endif
+ kfree(net->ipv6.fib6_main_tbl);
+ kfree(net->ipv6.fib_table_hash);
+}
+
+static struct pernet_operations fib6_net_ops = {
+ .init = fib6_net_init,
+ .exit = fib6_net_exit,
+};
+
+int __init fib6_init(void)
+{
+ int ret = -ENOMEM;
+ fib6_node_kmem = kmem_cache_create("fib6_nodes",
+ sizeof(struct fib6_node),
+ 0, SLAB_HWCACHE_ALIGN,
+ NULL);
+ if (!fib6_node_kmem)
+ goto out;
+
+ ret = register_pernet_subsys(&fib6_net_ops);
+ if (ret)
+ goto out_kmem_cache_create;
+
+ ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
+ if (ret)
+ goto out_unregister_subsys;
+out:
+ return ret;
+
+out_unregister_subsys:
+ unregister_pernet_subsys(&fib6_net_ops);
out_kmem_cache_create:
kmem_cache_destroy(fib6_node_kmem);
goto out;
@@ -1523,10 +1552,6 @@ out_kmem_cache_create:
void fib6_gc_cleanup(void)
{
del_timer(&ip6_fib_timer);
-#ifdef CONFIG_IPV6_MULTIPLE_TABLES
- kfree(fib6_local_tbl);
-#endif
- kfree(fib6_main_tbl);
- kfree(fib_table_hash);
+ unregister_pernet_subsys(&fib6_net_ops);
kmem_cache_destroy(fib6_node_kmem);
}
Index: net-2.6.25/net/ipv6/route.c
===================================================================
--- net-2.6.25.orig/net/ipv6/route.c
+++ net-2.6.25/net/ipv6/route.c
@@ -569,7 +569,7 @@ struct rt6_info *rt6_lookup(struct in6_a
flags |= RT6_LOOKUP_F_HAS_SADDR;
}
- dst = fib6_rule_lookup(&fl, flags, ip6_pol_route_lookup);
+ dst = fib6_rule_lookup(&init_net, &fl, flags, ip6_pol_route_lookup);
if (dst->error == 0)
return (struct rt6_info *) dst;
@@ -756,7 +756,7 @@ void ip6_route_input(struct sk_buff *skb
if (rt6_need_strict(&iph->daddr))
flags |= RT6_LOOKUP_F_IFACE;
- skb->dst = fib6_rule_lookup(&fl, flags, ip6_pol_route_input);
+ skb->dst = fib6_rule_lookup(&init_net, &fl, flags, ip6_pol_route_input);
}
static struct rt6_info *ip6_pol_route_output(struct fib6_table *table,
@@ -775,7 +775,7 @@ struct dst_entry * ip6_route_output(stru
if (!ipv6_addr_any(&fl->fl6_src))
flags |= RT6_LOOKUP_F_HAS_SADDR;
- return fib6_rule_lookup(fl, flags, ip6_pol_route_output);
+ return fib6_rule_lookup(&init_net, fl, flags, ip6_pol_route_output);
}
EXPORT_SYMBOL(ip6_route_output);
@@ -1068,7 +1068,7 @@ int ip6_route_add(struct fib6_config *cf
if (cfg->fc_metric == 0)
cfg->fc_metric = IP6_RT_PRIO_USER;
- table = fib6_new_table(cfg->fc_table);
+ table = fib6_new_table(&init_net, cfg->fc_table);
if (table == NULL) {
err = -ENOBUFS;
goto out;
@@ -1274,7 +1274,7 @@ static int ip6_route_del(struct fib6_con
struct rt6_info *rt;
int err = -ESRCH;
- table = fib6_get_table(cfg->fc_table);
+ table = fib6_get_table(&init_net, cfg->fc_table);
if (table == NULL)
return err;
@@ -1389,7 +1389,9 @@ static struct rt6_info *ip6_route_redire
if (rt6_need_strict(dest))
flags |= RT6_LOOKUP_F_IFACE;
- return (struct rt6_info *)fib6_rule_lookup((struct flowi *)&rdfl, flags, __ip6_route_redirect);
+ return (struct rt6_info *)fib6_rule_lookup(&init_net,
+ (struct flowi *)&rdfl,
+ flags, __ip6_route_redirect);
}
void rt6_redirect(struct in6_addr *dest, struct in6_addr *src,
@@ -1588,7 +1590,7 @@ static struct rt6_info *rt6_get_route_in
struct rt6_info *rt = NULL;
struct fib6_table *table;
- table = fib6_get_table(RT6_TABLE_INFO);
+ table = fib6_get_table(&init_net, RT6_TABLE_INFO);
if (table == NULL)
return NULL;
@@ -1643,7 +1645,7 @@ struct rt6_info *rt6_get_dflt_router(str
struct rt6_info *rt;
struct fib6_table *table;
- table = fib6_get_table(RT6_TABLE_DFLT);
+ table = fib6_get_table(&init_net, RT6_TABLE_DFLT);
if (table == NULL)
return NULL;
@@ -1687,7 +1689,7 @@ void rt6_purge_dflt_routers(void)
struct fib6_table *table;
/* NOTE: Keep consistent with rt6_get_dflt_router */
- table = fib6_get_table(RT6_TABLE_DFLT);
+ table = fib6_get_table(&init_net, RT6_TABLE_DFLT);
if (table == NULL)
return;
@@ -1848,7 +1850,7 @@ struct rt6_info *addrconf_dst_alloc(stru
ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
rt->rt6i_dst.plen = 128;
- rt->rt6i_table = fib6_get_table(RT6_TABLE_LOCAL);
+ rt->rt6i_table = fib6_get_table(&init_net, RT6_TABLE_LOCAL);
atomic_set(&rt->u.dst.__refcnt, 1);
--
^ permalink raw reply
* [patch 03/11][NETNS][IPV6] ip6_fib - make fib6_clean_all per namespace
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-make-fib6_clean_all-per-namespace.patch --]
[-- Type: text/plain, Size: 5624 bytes --]
The function fib6_clean_all takes the network namespace
as parameter. That allows to flush the routes related to
a specific network namespace.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
include/net/ip6_fib.h | 3 ++-
include/net/ip6_route.h | 2 +-
net/ipv6/addrconf.c | 3 ++-
net/ipv6/ip6_fib.c | 7 ++++---
net/ipv6/route.c | 28 ++++++++++++++++++++--------
5 files changed, 29 insertions(+), 14 deletions(-)
Index: net-2.6.25/include/net/ip6_fib.h
===================================================================
--- net-2.6.25.orig/include/net/ip6_fib.h
+++ net-2.6.25/include/net/ip6_fib.h
@@ -208,7 +208,8 @@ struct fib6_node *fib6_locate(struct fi
struct in6_addr *daddr, int dst_len,
struct in6_addr *saddr, int src_len);
-extern void fib6_clean_all(int (*func)(struct rt6_info *, void *arg),
+extern void fib6_clean_all(struct net *net,
+ int (*func)(struct rt6_info *, void *arg),
int prune, void *arg);
extern int fib6_add(struct fib6_node *root,
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -1350,7 +1350,7 @@ static void fib6_clean_tree(struct fib6_
fib6_walk(&c.w);
}
-void fib6_clean_all(int (*func)(struct rt6_info *, void *arg),
+void fib6_clean_all(struct net *net, int (*func)(struct rt6_info *, void *arg),
int prune, void *arg)
{
struct fib6_table *table;
@@ -1360,7 +1360,7 @@ void fib6_clean_all(int (*func)(struct r
rcu_read_lock();
for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
- head = &init_net.ipv6.fib_table_hash[h];
+ head = &net->ipv6.fib_table_hash[h];
hlist_for_each_entry_rcu(table, node, head, tb6_hlist) {
write_lock_bh(&table->tb6_lock);
fib6_clean_tree(&table->tb6_root, func, prune, arg);
@@ -1450,7 +1450,8 @@ void fib6_run_gc(unsigned long dummy)
gc_args.more = 0;
ndisc_dst_gc(&gc_args.more);
- fib6_clean_all(fib6_age, 0, NULL);
+
+ fib6_clean_all(&init_net, fib6_age, 0, NULL);
if (gc_args.more)
mod_timer(&ip6_fib_timer, jiffies +
Index: net-2.6.25/net/ipv6/route.c
===================================================================
--- net-2.6.25.orig/net/ipv6/route.c
+++ net-2.6.25/net/ipv6/route.c
@@ -1867,9 +1867,9 @@ static int fib6_ifdown(struct rt6_info *
return 0;
}
-void rt6_ifdown(struct net_device *dev)
+void rt6_ifdown(struct net *net, struct net_device *dev)
{
- fib6_clean_all(fib6_ifdown, 0, dev);
+ fib6_clean_all(net, fib6_ifdown, 0, dev);
}
struct rt6_mtu_change_arg
@@ -1925,7 +1925,7 @@ void rt6_mtu_change(struct net_device *d
.mtu = mtu,
};
- fib6_clean_all(rt6_mtu_change_route, 0, &arg);
+ fib6_clean_all(dev->nd_net, rt6_mtu_change_route, 0, &arg);
}
static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
@@ -2313,13 +2313,25 @@ static int rt6_info_route(struct rt6_inf
static int ipv6_route_show(struct seq_file *m, void *v)
{
- fib6_clean_all(rt6_info_route, 0, m);
+ struct net *net = (struct net *)m->private;
+ fib6_clean_all(net, rt6_info_route, 0, m);
return 0;
}
static int ipv6_route_open(struct inode *inode, struct file *file)
{
- return single_open(file, ipv6_route_show, NULL);
+ struct net *net = get_proc_net(inode);
+ if (!net)
+ return -ENXIO;
+ return single_open(file, ipv6_route_show, net);
+}
+
+static int ipv6_route_release(struct inode *inode, struct file *file)
+{
+ struct seq_file *seq = file->private_data;
+ struct net *net = seq->private;
+ put_net(net);
+ return single_release(inode, file);
}
static const struct file_operations ipv6_route_proc_fops = {
@@ -2327,7 +2339,7 @@ static const struct file_operations ipv6
.open = ipv6_route_open,
.read = seq_read,
.llseek = seq_lseek,
- .release = single_release,
+ .release = ipv6_route_release,
};
static int rt6_stats_seq_show(struct seq_file *seq, void *v)
@@ -2551,7 +2563,7 @@ xfrm6_init:
out_proc_init:
ipv6_route_proc_fini(&init_net);
out_fib6_init:
- rt6_ifdown(NULL);
+ rt6_ifdown(&init_net, NULL);
fib6_gc_cleanup();
out_kmem_cache:
kmem_cache_destroy(ip6_dst_ops.kmem_cachep);
@@ -2563,7 +2575,7 @@ void ip6_route_cleanup(void)
fib6_rules_cleanup();
ipv6_route_proc_fini(&init_net);
xfrm6_fini();
- rt6_ifdown(NULL);
+ rt6_ifdown(&init_net, NULL);
fib6_gc_cleanup();
kmem_cache_destroy(ip6_dst_ops.kmem_cachep);
}
Index: net-2.6.25/include/net/ip6_route.h
===================================================================
--- net-2.6.25.orig/include/net/ip6_route.h
+++ net-2.6.25/include/net/ip6_route.h
@@ -123,7 +123,7 @@ struct rt6_rtnl_dump_arg
};
extern int rt6_dump_route(struct rt6_info *rt, void *p_arg);
-extern void rt6_ifdown(struct net_device *dev);
+extern void rt6_ifdown(struct net *net, struct net_device *dev);
extern void rt6_mtu_change(struct net_device *dev, unsigned mtu);
extern rwlock_t rt6_lock;
Index: net-2.6.25/net/ipv6/addrconf.c
===================================================================
--- net-2.6.25.orig/net/ipv6/addrconf.c
+++ net-2.6.25/net/ipv6/addrconf.c
@@ -2430,6 +2430,7 @@ static int addrconf_ifdown(struct net_de
{
struct inet6_dev *idev;
struct inet6_ifaddr *ifa, **bifa;
+ struct net *net = dev->nd_net;
int i;
ASSERT_RTNL();
@@ -2437,7 +2438,7 @@ static int addrconf_ifdown(struct net_de
if (dev == init_net.loopback_dev && how == 1)
how = 0;
- rt6_ifdown(dev);
+ rt6_ifdown(net, dev);
neigh_ifdown(&nd_tbl, dev);
idev = __in6_dev_get(dev);
--
^ permalink raw reply
* [patch 04/11][NETNS][IPV6] ip6_fib - pass the network namespace parameter to timer callback
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-add-net-to-gc-timer-parameter.patch --]
[-- Type: text/plain, Size: 5694 bytes --]
The fib tables are now relative to the network namespace. When the
garbage collector timer expires, we must have a network namespace
parameter in order to retrieve the tables. For now this is the
init_net, but we should be able to have a timer per namespace and
use the timer callback parameter to pass the network namespace from
the expired timer.
The timer callback, fib6_run_gc, is actually used to be called
synchronously by some functions and asynchronously when the timer
expires.
When the timer expires, the delay specified for fib6_run_gc parameter
is always zero. So, I changed fib6_run_gc to not be a timer callback
but a function called by the timer callback and I added a timer callback
where its work is just to retrieve from the data arg of the timer the
network namespace and call fib6_run_gc with zero expiring time and
the network namespace parameters. That makes the code cleaner for the
fib6_run_gc callers.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
include/net/ip6_fib.h | 3 ++-
net/ipv6/ip6_fib.c | 24 ++++++++++++++++--------
net/ipv6/ndisc.c | 5 +++--
net/ipv6/route.c | 8 +++++---
4 files changed, 26 insertions(+), 14 deletions(-)
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -93,7 +93,10 @@ static int fib6_walk_continue(struct fib
static __u32 rt_sernum;
-static DEFINE_TIMER(ip6_fib_timer, fib6_run_gc, 0, 0);
+static void fib6_gc_timer_cb(unsigned long arg);
+
+static DEFINE_TIMER(ip6_fib_timer, fib6_gc_timer_cb, 0,
+ (unsigned long)&init_net);
static struct fib6_walker_t fib6_walker_list = {
.prev = &fib6_walker_list,
@@ -1432,12 +1435,12 @@ static int fib6_age(struct rt6_info *rt,
static DEFINE_SPINLOCK(fib6_gc_lock);
-void fib6_run_gc(unsigned long dummy)
+void fib6_run_gc(unsigned long expires, struct net *net)
{
- if (dummy != ~0UL) {
+ if (expires != ~0UL) {
spin_lock_bh(&fib6_gc_lock);
- gc_args.timeout = dummy ? (int)dummy :
- init_net.ipv6.sysctl.ip6_rt_gc_interval;
+ gc_args.timeout = expires ? (int)expires :
+ net->ipv6.sysctl.ip6_rt_gc_interval;
} else {
local_bh_disable();
if (!spin_trylock(&fib6_gc_lock)) {
@@ -1445,17 +1448,17 @@ void fib6_run_gc(unsigned long dummy)
local_bh_enable();
return;
}
- gc_args.timeout = init_net.ipv6.sysctl.ip6_rt_gc_interval;
+ gc_args.timeout = net->ipv6.sysctl.ip6_rt_gc_interval;
}
gc_args.more = 0;
ndisc_dst_gc(&gc_args.more);
- fib6_clean_all(&init_net, fib6_age, 0, NULL);
+ fib6_clean_all(net, fib6_age, 0, NULL);
if (gc_args.more)
mod_timer(&ip6_fib_timer, jiffies +
- init_net.ipv6.sysctl.ip6_rt_gc_interval);
+ net->ipv6.sysctl.ip6_rt_gc_interval);
else {
del_timer(&ip6_fib_timer);
ip6_fib_timer.expires = 0;
@@ -1463,6 +1466,11 @@ void fib6_run_gc(unsigned long dummy)
spin_unlock_bh(&fib6_gc_lock);
}
+static void fib6_gc_timer_cb(unsigned long arg)
+{
+ fib6_run_gc(0, (struct net *)arg);
+}
+
static int fib6_net_init(struct net *net)
{
int ret;
Index: net-2.6.25/include/net/ip6_fib.h
===================================================================
--- net-2.6.25.orig/include/net/ip6_fib.h
+++ net-2.6.25/include/net/ip6_fib.h
@@ -222,7 +222,8 @@ extern int fib6_del(struct rt6_info *r
extern void inet6_rt_notify(int event, struct rt6_info *rt,
struct nl_info *info);
-extern void fib6_run_gc(unsigned long dummy);
+extern void fib6_run_gc(unsigned long expires,
+ struct net *net);
extern void fib6_gc_cleanup(void);
Index: net-2.6.25/net/ipv6/ndisc.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ndisc.c
+++ net-2.6.25/net/ipv6/ndisc.c
@@ -1613,6 +1613,7 @@ int ndisc_rcv(struct sk_buff *skb)
static int ndisc_netdev_event(struct notifier_block *this, unsigned long event, void *ptr)
{
struct net_device *dev = ptr;
+ struct net *net = dev->nd_net;
if (dev->nd_net != &init_net)
return NOTIFY_DONE;
@@ -1620,11 +1621,11 @@ static int ndisc_netdev_event(struct not
switch (event) {
case NETDEV_CHANGEADDR:
neigh_changeaddr(&nd_tbl, dev);
- fib6_run_gc(~0UL);
+ fib6_run_gc(~0UL, net);
break;
case NETDEV_DOWN:
neigh_ifdown(&nd_tbl, dev);
- fib6_run_gc(~0UL);
+ fib6_run_gc(~0UL, net);
break;
default:
break;
Index: net-2.6.25/net/ipv6/route.c
===================================================================
--- net-2.6.25.orig/net/ipv6/route.c
+++ net-2.6.25/net/ipv6/route.c
@@ -40,6 +40,7 @@
#include <linux/if_arp.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
+#include <linux/nsproxy.h>
#include <net/net_namespace.h>
#include <net/snmp.h>
#include <net/ipv6.h>
@@ -994,7 +995,7 @@ static int ip6_dst_gc(struct dst_ops *op
goto out;
expire++;
- fib6_run_gc(expire);
+ fib6_run_gc(expire, &init_net);
last_gc = now;
if (atomic_read(&ip6_dst_ops.entries) < ip6_dst_ops.gc_thresh)
expire = init_net.ipv6.sysctl.ip6_rt_gc_timeout>>1;
@@ -2408,10 +2409,11 @@ static
int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file * filp,
void __user *buffer, size_t *lenp, loff_t *ppos)
{
- int delay = init_net.ipv6.sysctl.flush_delay;
+ struct net *net = current->nsproxy->net_ns;
+ int delay = net->ipv6.sysctl.flush_delay;
if (write) {
proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
- fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay);
+ fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay, net);
return 0;
} else
return -EINVAL;
--
^ permalink raw reply
* [patch 07/11][NETNS][IPV6] make fib6_clean_node to use the network namespace
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-clean-node-use-namespace.patch --]
[-- Type: text/plain, Size: 3376 bytes --]
The fib6_clean_node function should have the network namespace
it is working on. The fib6_cleaner_t structure is extended with the
network namespace field to be passed to the fib6_clean_node function.
The different functions calling the fib6_clean_node function are extended
with the netns parameter when needed to propagate the netns pointer.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
---
net/ipv6/ip6_fib.c | 25 +++++++++++++++----------
1 file changed, 15 insertions(+), 10 deletions(-)
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -66,6 +66,7 @@ enum fib_walk_state_t
struct fib6_cleaner_t
{
struct fib6_walker_t w;
+ struct net *net;
int (*func)(struct rt6_info *, void *arg);
void *arg;
};
@@ -78,7 +79,8 @@ static DEFINE_RWLOCK(fib6_walker_lock);
#define FWS_INIT FWS_L
#endif
-static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt);
+static void fib6_prune_clones(struct net *net, struct fib6_node *fn,
+ struct rt6_info *rt);
static struct rt6_info * fib6_find_prefix(struct fib6_node *fn);
static struct fib6_node * fib6_repair_tree(struct fib6_node *fn);
static int fib6_walk(struct fib6_walker_t *w);
@@ -762,7 +764,7 @@ int fib6_add(struct fib6_node *root, str
if (err == 0) {
fib6_start_gc(info->nl_net, rt);
if (!(rt->rt6i_flags&RTF_CACHE))
- fib6_prune_clones(pn, rt);
+ fib6_prune_clones(info->nl_net, pn, rt);
}
out:
@@ -1168,7 +1170,7 @@ int fib6_del(struct rt6_info *rt, struct
pn = pn->parent;
}
#endif
- fib6_prune_clones(pn, rt);
+ fib6_prune_clones(info->nl_net, pn, rt);
}
/*
@@ -1298,12 +1300,12 @@ static int fib6_walk(struct fib6_walker_
static int fib6_clean_node(struct fib6_walker_t *w)
{
- struct nl_info info = {
- .nl_net = &init_net,
- };
int res;
struct rt6_info *rt;
struct fib6_cleaner_t *c = container_of(w, struct fib6_cleaner_t, w);
+ struct nl_info info = {
+ .nl_net = c->net,
+ };
for (rt = w->leaf; rt; rt = rt->u.dst.rt6_next) {
res = c->func(rt, c->arg);
@@ -1335,7 +1337,7 @@ static int fib6_clean_node(struct fib6_w
* ignoring pure split nodes) will be scanned.
*/
-static void fib6_clean_tree(struct fib6_node *root,
+static void fib6_clean_tree(struct net *net, struct fib6_node *root,
int (*func)(struct rt6_info *, void *arg),
int prune, void *arg)
{
@@ -1346,6 +1348,7 @@ static void fib6_clean_tree(struct fib6_
c.w.prune = prune;
c.func = func;
c.arg = arg;
+ c.net = net;
fib6_walk(&c.w);
}
@@ -1363,7 +1366,8 @@ void fib6_clean_all(struct net *net, int
head = &net->ipv6.fib_table_hash[h];
hlist_for_each_entry_rcu(table, node, head, tb6_hlist) {
write_lock_bh(&table->tb6_lock);
- fib6_clean_tree(&table->tb6_root, func, prune, arg);
+ fib6_clean_tree(net, &table->tb6_root,
+ func, prune, arg);
write_unlock_bh(&table->tb6_lock);
}
}
@@ -1380,9 +1384,10 @@ static int fib6_prune_clone(struct rt6_i
return 0;
}
-static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt)
+static void fib6_prune_clones(struct net *net, struct fib6_node *fn,
+ struct rt6_info *rt)
{
- fib6_clean_tree(fn, fib6_prune_clone, 1, rt);
+ fib6_clean_tree(net, fn, fib6_prune_clone, 1, rt);
}
/*
--
^ permalink raw reply
* [patch 08/11][NETNS][IPV6] fib6_rules - dynamically allocate the fib rules ops
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: fib6-rules-dynamic-allocation.patch --]
[-- Type: text/plain, Size: 2897 bytes --]
The fib6_rules_ops structure is dynamically allocated, so that allows
to make several instances of it per network namespace.
The global static fib6_rules_ops structure is renamed to
fib6_rules_ops_template in order to quickly memcopy it for the
structure initialization.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
---
net/ipv6/fib6_rules.c | 26 +++++++++++++++++---------
1 file changed, 17 insertions(+), 9 deletions(-)
Index: net-2.6.25/net/ipv6/fib6_rules.c
===================================================================
--- net-2.6.25.orig/net/ipv6/fib6_rules.c
+++ net-2.6.25/net/ipv6/fib6_rules.c
@@ -29,7 +29,7 @@ struct fib6_rule
u8 tclass;
};
-static struct fib_rules_ops fib6_rules_ops;
+static struct fib_rules_ops *fib6_rules_ops;
struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi *fl,
int flags, pol_lookup_t lookup)
@@ -38,7 +38,7 @@ struct dst_entry *fib6_rule_lookup(struc
.lookup_ptr = lookup,
};
- fib_rules_lookup(&fib6_rules_ops, fl, flags, &arg);
+ fib_rules_lookup(fib6_rules_ops, fl, flags, &arg);
if (arg.rule)
fib_rule_put(arg.rule);
@@ -234,7 +234,7 @@ static size_t fib6_rule_nlmsg_payload(st
+ nla_total_size(16); /* src */
}
-static struct fib_rules_ops fib6_rules_ops = {
+static struct fib_rules_ops fib6_rules_ops_template = {
.family = AF_INET6,
.rule_size = sizeof(struct fib6_rule),
.addr_size = sizeof(struct in6_addr),
@@ -247,7 +247,6 @@ static struct fib_rules_ops fib6_rules_o
.nlmsg_payload = fib6_rule_nlmsg_payload,
.nlgroup = RTNLGRP_IPV6_RULE,
.policy = fib6_rule_policy,
- .rules_list = LIST_HEAD_INIT(fib6_rules_ops.rules_list),
.owner = THIS_MODULE,
.fro_net = &init_net,
};
@@ -256,11 +255,18 @@ static int __init fib6_default_rules_ini
{
int err;
- err = fib_default_rule_add(&fib6_rules_ops, 0,
+ fib6_rules_ops = kmemdup(&fib6_rules_ops_template,
+ sizeof(*fib6_rules_ops), GFP_KERNEL);
+ if (!fib6_rules_ops)
+ return -ENOMEM;
+
+ INIT_LIST_HEAD(&fib6_rules_ops->rules_list);
+
+ err = fib_default_rule_add(fib6_rules_ops, 0,
RT6_TABLE_LOCAL, FIB_RULE_PERMANENT);
if (err < 0)
return err;
- err = fib_default_rule_add(&fib6_rules_ops, 0x7FFE, RT6_TABLE_MAIN, 0);
+ err = fib_default_rule_add(fib6_rules_ops, 0x7FFE, RT6_TABLE_MAIN, 0);
if (err < 0)
return err;
return 0;
@@ -274,18 +280,20 @@ int __init fib6_rules_init(void)
if (ret)
goto out;
- ret = fib_rules_register(&fib6_rules_ops);
+ ret = fib_rules_register(fib6_rules_ops);
if (ret)
goto out_default_rules_init;
out:
return ret;
out_default_rules_init:
- fib_rules_cleanup_ops(&fib6_rules_ops);
+ fib_rules_cleanup_ops(fib6_rules_ops);
+ kfree(fib6_rules_ops);
goto out;
}
void fib6_rules_cleanup(void)
{
- fib_rules_unregister(&fib6_rules_ops);
+ fib_rules_unregister(fib6_rules_ops);
+ kfree(fib6_rules_ops);
}
--
^ permalink raw reply
* [patch 06/11][NETNS][IPV6] ip6_fib - make the ip6 fib gc timer per network namespace
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-gc-timer-per-namespace.patch --]
[-- Type: text/plain, Size: 6564 bytes --]
Move the timer initialization at the network namespace creation
and store the network namespace in the timer argument.
That enables multiple timers (one per network namespace) to do garbage
collecting.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
include/net/ip6_route.h | 2 -
include/net/netns/ipv6.h | 1
net/ipv6/ip6_fib.c | 56 +++++++++++++++++++++++------------------------
net/ipv6/route.c | 5 +++-
4 files changed, 34 insertions(+), 30 deletions(-)
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -95,8 +95,6 @@ static __u32 rt_sernum;
static void fib6_gc_timer_cb(unsigned long arg);
-static struct timer_list *ip6_fib_timer;
-
static struct fib6_walker_t fib6_walker_list = {
.prev = &fib6_walker_list,
.next = &fib6_walker_list,
@@ -663,19 +661,19 @@ static int fib6_add_rt2node(struct fib6_
return 0;
}
-static __inline__ void fib6_start_gc(struct rt6_info *rt)
+static __inline__ void fib6_start_gc(struct net *net, struct rt6_info *rt)
{
- if (ip6_fib_timer->expires == 0 &&
+ if (net->ipv6.ip6_fib_timer->expires == 0 &&
(rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE)))
- mod_timer(ip6_fib_timer, jiffies +
- init_net.ipv6.sysctl.ip6_rt_gc_interval);
+ mod_timer(net->ipv6.ip6_fib_timer, jiffies +
+ net->ipv6.sysctl.ip6_rt_gc_interval);
}
-void fib6_force_start_gc(void)
+void fib6_force_start_gc(struct net *net)
{
- if (ip6_fib_timer->expires == 0)
- mod_timer(ip6_fib_timer, jiffies +
- init_net.ipv6.sysctl.ip6_rt_gc_interval);
+ if (net->ipv6.ip6_fib_timer->expires == 0)
+ mod_timer(net->ipv6.ip6_fib_timer, jiffies +
+ net->ipv6.sysctl.ip6_rt_gc_interval);
}
/*
@@ -762,7 +760,7 @@ int fib6_add(struct fib6_node *root, str
err = fib6_add_rt2node(fn, rt, info);
if (err == 0) {
- fib6_start_gc(rt);
+ fib6_start_gc(info->nl_net, rt);
if (!(rt->rt6i_flags&RTF_CACHE))
fib6_prune_clones(pn, rt);
}
@@ -1443,7 +1441,7 @@ void fib6_run_gc(unsigned long expires,
} else {
local_bh_disable();
if (!spin_trylock(&fib6_gc_lock)) {
- mod_timer(ip6_fib_timer, jiffies + HZ);
+ mod_timer(net->ipv6.ip6_fib_timer, jiffies + HZ);
local_bh_enable();
return;
}
@@ -1456,11 +1454,11 @@ void fib6_run_gc(unsigned long expires,
fib6_clean_all(net, fib6_age, 0, NULL);
if (gc_args.more)
- mod_timer(ip6_fib_timer, jiffies +
+ mod_timer(net->ipv6.ip6_fib_timer, jiffies +
net->ipv6.sysctl.ip6_rt_gc_interval);
else {
- del_timer(ip6_fib_timer);
- ip6_fib_timer->expires = 0;
+ del_timer(net->ipv6.ip6_fib_timer);
+ net->ipv6.ip6_fib_timer->expires = 0;
}
spin_unlock_bh(&fib6_gc_lock);
}
@@ -1473,13 +1471,21 @@ static void fib6_gc_timer_cb(unsigned lo
static int fib6_net_init(struct net *net)
{
int ret;
+ struct timer_list *timer;
ret = -ENOMEM;
+ timer = kzalloc(sizeof(*timer), GFP_KERNEL);
+ if (!timer)
+ goto out;
+
+ setup_timer(timer, fib6_gc_timer_cb, (unsigned long)net);
+ net->ipv6.ip6_fib_timer = timer;
+
net->ipv6.fib_table_hash =
kzalloc(sizeof(*net->ipv6.fib_table_hash)*FIB_TABLE_HASHSZ,
GFP_KERNEL);
if (!net->ipv6.fib_table_hash)
- goto out;
+ goto out_timer;
net->ipv6.fib6_main_tbl = kzalloc(sizeof(*net->ipv6.fib6_main_tbl),
GFP_KERNEL);
@@ -1513,11 +1519,15 @@ out_fib6_main_tbl:
#endif
out_fib_table_hash:
kfree(net->ipv6.fib_table_hash);
+out_timer:
+ kfree(timer);
goto out;
}
static void fib6_net_exit(struct net *net)
{
+ del_timer(net->ipv6.ip6_fib_timer);
+ kfree(net->ipv6.ip6_fib_timer);
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
kfree(net->ipv6.fib6_local_tbl);
#endif
@@ -1533,6 +1543,7 @@ static struct pernet_operations fib6_net
int __init fib6_init(void)
{
int ret = -ENOMEM;
+
fib6_node_kmem = kmem_cache_create("fib6_nodes",
sizeof(struct fib6_node),
0, SLAB_HWCACHE_ALIGN,
@@ -1540,16 +1551,9 @@ int __init fib6_init(void)
if (!fib6_node_kmem)
goto out;
- ret = -ENOMEM;
- ip6_fib_timer = kzalloc(sizeof(*ip6_fib_timer), GFP_KERNEL);
- if (!ip6_fib_timer)
- goto out_kmem_cache_create;
-
- setup_timer(ip6_fib_timer, fib6_gc_timer_cb, (unsigned long)&init_net);
-
ret = register_pernet_subsys(&fib6_net_ops);
if (ret)
- goto out_timer;
+ goto out_kmem_cache_create;
ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
if (ret)
@@ -1559,8 +1563,6 @@ out:
out_unregister_subsys:
unregister_pernet_subsys(&fib6_net_ops);
-out_timer:
- kfree(ip6_fib_timer);
out_kmem_cache_create:
kmem_cache_destroy(fib6_node_kmem);
goto out;
@@ -1568,8 +1570,6 @@ out_kmem_cache_create:
void fib6_gc_cleanup(void)
{
- del_timer(ip6_fib_timer);
- kfree(ip6_fib_timer);
unregister_pernet_subsys(&fib6_net_ops);
kmem_cache_destroy(fib6_node_kmem);
}
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
--- net-2.6.25.orig/include/net/netns/ipv6.h
+++ net-2.6.25/include/net/netns/ipv6.h
@@ -32,6 +32,7 @@ struct netns_ipv6 {
struct ipv6_devconf *devconf_dflt;
struct netns_frags frags;
+ struct timer_list *ip6_fib_timer;
struct hlist_head *fib_table_hash;
struct fib6_table *fib6_main_tbl;
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
Index: net-2.6.25/include/net/ip6_route.h
===================================================================
--- net-2.6.25.orig/include/net/ip6_route.h
+++ net-2.6.25/include/net/ip6_route.h
@@ -80,7 +80,7 @@ extern struct dst_entry *ndisc_dst_alloc
struct in6_addr *addr,
int (*output)(struct sk_buff *));
extern int ndisc_dst_gc(int *more);
-extern void fib6_force_start_gc(void);
+extern void fib6_force_start_gc(struct net *net);
extern struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
const struct in6_addr *addr,
Index: net-2.6.25/net/ipv6/route.c
===================================================================
--- net-2.6.25.orig/net/ipv6/route.c
+++ net-2.6.25/net/ipv6/route.c
@@ -951,7 +951,7 @@ struct dst_entry *ndisc_dst_alloc(struct
ndisc_dst_gc_list = &rt->u.dst;
spin_unlock_bh(&ndisc_lock);
- fib6_force_start_gc();
+ fib6_force_start_gc(dev->nd_net);
out:
return &rt->u.dst;
@@ -1229,6 +1229,9 @@ install_route:
rt->u.dst.dev = dev;
rt->rt6i_idev = idev;
rt->rt6i_table = table;
+
+ cfg->fc_nlinfo.nl_net = dev->nd_net;
+
return __ip6_ins_rt(rt, &cfg->fc_nlinfo);
out:
--
^ permalink raw reply
* [patch 10/11][NETNS][IPV6] rt6_stats - dynamically allocate the rt6_stats
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: rt6-stats-dynamic-allocation.patch --]
[-- Type: text/plain, Size: 3946 bytes --]
This patch allocates the rt6_stats struct dynamically when
the fib6 is initialized. That provides the ability to create
several instances of this structure for the network namespaces.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
include/net/ipv6.h | 2 +-
net/ipv6/ip6_fib.c | 22 +++++++++++++++-------
net/ipv6/route.c | 8 ++++----
3 files changed, 20 insertions(+), 12 deletions(-)
Index: net-2.6.25/include/net/ipv6.h
===================================================================
--- net-2.6.25.orig/include/net/ipv6.h
+++ net-2.6.25/include/net/ipv6.h
@@ -604,7 +604,7 @@ extern void ipv6_misc_proc_exit(void);
extern int snmp6_register_dev(struct inet6_dev *idev);
extern int snmp6_unregister_dev(struct inet6_dev *idev);
-extern struct rt6_statistics rt6_stats;
+extern struct rt6_statistics *rt6_stats;
#else
static inline int snmp6_register_dev(struct inet6_dev *idev)
{
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -48,7 +48,7 @@
#define RT6_TRACE(x...) do { ; } while (0)
#endif
-struct rt6_statistics rt6_stats;
+struct rt6_statistics *rt6_stats;
static struct kmem_cache * fib6_node_kmem __read_mostly;
@@ -653,10 +653,10 @@ static int fib6_add_rt2node(struct fib6_
rt->rt6i_node = fn;
atomic_inc(&rt->rt6i_ref);
inet6_rt_notify(RTM_NEWROUTE, rt, info);
- rt6_stats.fib_rt_entries++;
+ rt6_stats->fib_rt_entries++;
if ((fn->fn_flags & RTN_RTINFO) == 0) {
- rt6_stats.fib_route_nodes++;
+ rt6_stats->fib_route_nodes++;
fn->fn_flags |= RTN_RTINFO;
}
@@ -1093,8 +1093,8 @@ static void fib6_del_route(struct fib6_n
/* Unlink it */
*rtp = rt->u.dst.rt6_next;
rt->rt6i_node = NULL;
- rt6_stats.fib_rt_entries--;
- rt6_stats.fib_discarded_routes++;
+ rt6_stats->fib_rt_entries--;
+ rt6_stats->fib_discarded_routes++;
/* Reset round-robin state, if necessary */
if (fn->rr_ptr == rt)
@@ -1117,7 +1117,7 @@ static void fib6_del_route(struct fib6_n
/* If it was last route, expunge its radix tree node */
if (fn->leaf == NULL) {
fn->fn_flags &= ~RTN_RTINFO;
- rt6_stats.fib_route_nodes--;
+ rt6_stats->fib_route_nodes--;
fn = fib6_repair_tree(fn);
}
@@ -1556,9 +1556,14 @@ int __init fib6_init(void)
if (!fib6_node_kmem)
goto out;
+ ret = -ENOMEM;
+ rt6_stats = kzalloc(sizeof(*rt6_stats), GFP_KERNEL);
+ if (!rt6_stats)
+ goto out_kmem_cache_create;
+
ret = register_pernet_subsys(&fib6_net_ops);
if (ret)
- goto out_kmem_cache_create;
+ goto out_rt6_stats;
ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
if (ret)
@@ -1568,6 +1573,8 @@ out:
out_unregister_subsys:
unregister_pernet_subsys(&fib6_net_ops);
+out_rt6_stats:
+ kfree(rt6_stats);
out_kmem_cache_create:
kmem_cache_destroy(fib6_node_kmem);
goto out;
@@ -1576,5 +1583,6 @@ out_kmem_cache_create:
void fib6_gc_cleanup(void)
{
unregister_pernet_subsys(&fib6_net_ops);
+ kfree(rt6_stats);
kmem_cache_destroy(fib6_node_kmem);
}
Index: net-2.6.25/net/ipv6/route.c
===================================================================
--- net-2.6.25.orig/net/ipv6/route.c
+++ net-2.6.25/net/ipv6/route.c
@@ -2349,11 +2349,11 @@ static const struct file_operations ipv6
static int rt6_stats_seq_show(struct seq_file *seq, void *v)
{
seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n",
- rt6_stats.fib_nodes, rt6_stats.fib_route_nodes,
- rt6_stats.fib_rt_alloc, rt6_stats.fib_rt_entries,
- rt6_stats.fib_rt_cache,
+ rt6_stats->fib_nodes, rt6_stats->fib_route_nodes,
+ rt6_stats->fib_rt_alloc, rt6_stats->fib_rt_entries,
+ rt6_stats->fib_rt_cache,
atomic_read(&ip6_dst_ops.entries),
- rt6_stats.fib_discarded_routes);
+ rt6_stats->fib_discarded_routes);
return 0;
}
--
^ permalink raw reply
* [patch 05/11][NETNS][IPV6] ip6_fib - dynamically allocate the gc_timer
From: Daniel Lezcano @ 2008-01-25 16:50 UTC (permalink / raw)
To: davem; +Cc: benjamin.thery, netdev, den
In-Reply-To: <20080125165008.317745745@localhost.localdomain>
[-- Attachment #1: ip6-fib-dynamically-allocate-gc-timer.patch --]
[-- Type: text/plain, Size: 3216 bytes --]
The ip6_fib_timer gc timer is dynamically allocated and initialized in
the ip6 fib init function. There are no more references to a
static global variable. That will allow to make multiple instance
of the garbage collecting timer and make them per namespace.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
---
net/ipv6/ip6_fib.c | 33 +++++++++++++++++++++------------
1 file changed, 21 insertions(+), 12 deletions(-)
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
--- net-2.6.25.orig/net/ipv6/ip6_fib.c
+++ net-2.6.25/net/ipv6/ip6_fib.c
@@ -95,8 +95,7 @@ static __u32 rt_sernum;
static void fib6_gc_timer_cb(unsigned long arg);
-static DEFINE_TIMER(ip6_fib_timer, fib6_gc_timer_cb, 0,
- (unsigned long)&init_net);
+static struct timer_list *ip6_fib_timer;
static struct fib6_walker_t fib6_walker_list = {
.prev = &fib6_walker_list,
@@ -666,16 +665,16 @@ static int fib6_add_rt2node(struct fib6_
static __inline__ void fib6_start_gc(struct rt6_info *rt)
{
- if (ip6_fib_timer.expires == 0 &&
+ if (ip6_fib_timer->expires == 0 &&
(rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE)))
- mod_timer(&ip6_fib_timer, jiffies +
+ mod_timer(ip6_fib_timer, jiffies +
init_net.ipv6.sysctl.ip6_rt_gc_interval);
}
void fib6_force_start_gc(void)
{
- if (ip6_fib_timer.expires == 0)
- mod_timer(&ip6_fib_timer, jiffies +
+ if (ip6_fib_timer->expires == 0)
+ mod_timer(ip6_fib_timer, jiffies +
init_net.ipv6.sysctl.ip6_rt_gc_interval);
}
@@ -1444,7 +1443,7 @@ void fib6_run_gc(unsigned long expires,
} else {
local_bh_disable();
if (!spin_trylock(&fib6_gc_lock)) {
- mod_timer(&ip6_fib_timer, jiffies + HZ);
+ mod_timer(ip6_fib_timer, jiffies + HZ);
local_bh_enable();
return;
}
@@ -1457,11 +1456,11 @@ void fib6_run_gc(unsigned long expires,
fib6_clean_all(net, fib6_age, 0, NULL);
if (gc_args.more)
- mod_timer(&ip6_fib_timer, jiffies +
+ mod_timer(ip6_fib_timer, jiffies +
net->ipv6.sysctl.ip6_rt_gc_interval);
else {
- del_timer(&ip6_fib_timer);
- ip6_fib_timer.expires = 0;
+ del_timer(ip6_fib_timer);
+ ip6_fib_timer->expires = 0;
}
spin_unlock_bh(&fib6_gc_lock);
}
@@ -1541,9 +1540,16 @@ int __init fib6_init(void)
if (!fib6_node_kmem)
goto out;
+ ret = -ENOMEM;
+ ip6_fib_timer = kzalloc(sizeof(*ip6_fib_timer), GFP_KERNEL);
+ if (!ip6_fib_timer)
+ goto out_kmem_cache_create;
+
+ setup_timer(ip6_fib_timer, fib6_gc_timer_cb, (unsigned long)&init_net);
+
ret = register_pernet_subsys(&fib6_net_ops);
if (ret)
- goto out_kmem_cache_create;
+ goto out_timer;
ret = __rtnl_register(PF_INET6, RTM_GETROUTE, NULL, inet6_dump_fib);
if (ret)
@@ -1553,6 +1559,8 @@ out:
out_unregister_subsys:
unregister_pernet_subsys(&fib6_net_ops);
+out_timer:
+ kfree(ip6_fib_timer);
out_kmem_cache_create:
kmem_cache_destroy(fib6_node_kmem);
goto out;
@@ -1560,7 +1568,8 @@ out_kmem_cache_create:
void fib6_gc_cleanup(void)
{
- del_timer(&ip6_fib_timer);
+ del_timer(ip6_fib_timer);
+ kfree(ip6_fib_timer);
unregister_pernet_subsys(&fib6_net_ops);
kmem_cache_destroy(fib6_node_kmem);
}
--
^ 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