* Re: 8390 drivers (was: Re: Debian kernel 2.6.38-5)
From: Christian T. Steigies @ 2011-05-10 7:29 UTC (permalink / raw)
To: Geert Uytterhoeven
Cc: Finn Thain, Thorsten Glaser, linux-m68k, Stephen Hemminger,
David S. Miller, netdev
In-Reply-To: <BANLkTik27HjWOF0s5HRqgva5zYzQRis27Q@mail.gmail.com>
On Tue, May 10, 2011 at 08:52:48AM +0200, Geert Uytterhoeven wrote:
> >
> > zorro8390.mod.c
>
> That's a generated file, so you can ignore it.
ok
> Yeah, speed depends on CPU type and motherboard/CPU interface.
well, it was much faster for a while. Now, after downloading and during
installing packages (and switching to dependency based booting, etc), I get:
eth0: mismatched read page pointer: 1 vs ff.
and
eth0: mismatched read page pointer: 1 vs f0. (I think, maybe fe or 0)
A lot of them. So many that I can not stop the output or do anything on the
machine :-(
There is a different message as well, but it is not readable.
Christian
^ permalink raw reply
* Re: 8390 drivers (was: Re: Debian kernel 2.6.38-5)
From: Geert Uytterhoeven @ 2011-05-10 6:52 UTC (permalink / raw)
To: Geert Uytterhoeven, Finn Thain, Thorsten Glaser, linux-m68k,
Stephen Hemminger
In-Reply-To: <20110509213213.GA28675@chumley.earth.sol>
On Mon, May 9, 2011 at 23:32, Christian T. Steigies <cts@debian.org> wrote:
> On Mon, May 09, 2011 at 10:38:18PM +0200, Geert Uytterhoeven wrote:
>> On Mon, May 9, 2011 at 22:25, Christian T. Steigies <cts@debian.org> wrote:
>> > On Mon, May 09, 2011 at 09:16:16AM +0200, Geert Uytterhoeven wrote:
>> >> On Mon, May 9, 2011 at 01:14, Finn Thain <fthain@telegraphics.com.au> wrote:
>> >> > On Sun, 8 May 2011, Christian T. Steigies wrote:
>> >> >> PS 2.6.28 did not boot: kernel too old. When was TLS introduced? I'll try to
>> >> >> apply the patch you mentioned in your other message.
>> >> >
>> >> > I sometimes test network cards with busybox. It can be built without
>> >> > linking in glibc and doesn't need TLS.
>> >> >
>> >> >
>> >> >> [ 130.870000] eth0: trigger_send() called with the transmitter busy.
>> >> >> [ 132.240000] ------------[ cut here ]------------
>> >> >> [ 132.240000] WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x1ac/0x1cc()
>> >> >> [ 132.250000] NETDEV WATCHDOG: eth0 (): transmit queue 0 timed out
>> >> >
>> >> > Looks a lot like this problem:
>> >> >
>> >> > http://patchwork.ozlabs.org/patch/27774/
>> >>
>> >> Yeah, that's a very likely culprit. Thx!
>> >
>> > I applied the following patch and it works! apt-get update that is, there is
>> > no sshd on this freshly installed system yet...
>> >
>> > I also changed this line to have 4 underscores at the beginning:
>> >
>> > { 0xec940559, "____alloc_ei_netdev" },
>>
>> In which file is that?
>
> zorro8390.mod.c
That's a generated file, so you can ignore it.
>> > it does not show up in git diff, so I am not sure if this is a leftover from
>> > a previous checkout, or if it is also needed. And if it is needed, if it
>> > should be two underscores or four or none, and if so why or why not...
>> > Looking at the other drivers, there does not seem to be a lot of consistency?
>>
>> From this, I guess hydra and ne-h8300 are also affected, as they
>> include lib8390.c
>> theirselves and are linked with 8390.o?
>>
>> > diff --git a/drivers/net/zorro8390.c b/drivers/net/zorro8390.c
>> > index b78a38d9..8c7c522 100644
>> > --- a/drivers/net/zorro8390.c
>> > +++ b/drivers/net/zorro8390.c
>> > @@ -126,7 +126,7 @@ static int __devinit zorro8390_init_one(struct zorro_dev *z,
>> >
>> > board = z->resource.start;
>> > ioaddr = board+cards[i].offset;
>> > - dev = alloc_ei_netdev();
>> > + dev = ____alloc_ei_netdev(0);
>> > if (!dev)
>> > return -ENOMEM;
>> > if (!request_mem_region(ioaddr, NE_IO_EXTENT*2, DRV_NAME)) {
>> > @@ -146,15 +146,15 @@ static int __devinit zorro8390_init_one(struct zorro_dev *z,
>> > static const struct net_device_ops zorro8390_netdev_ops = {
>> > .ndo_open = zorro8390_open,
>> > .ndo_stop = zorro8390_close,
>> > - .ndo_start_xmit = ei_start_xmit,
>> > - .ndo_tx_timeout = ei_tx_timeout,
>> > - .ndo_get_stats = ei_get_stats,
>> > - .ndo_set_multicast_list = ei_set_multicast_list,
>> > + .ndo_start_xmit = __ei_start_xmit,
>> > + .ndo_tx_timeout = __ei_tx_timeout,
>> > + .ndo_get_stats = __ei_get_stats,
>> > + .ndo_set_multicast_list = __ei_set_multicast_list,
>> > .ndo_validate_addr = eth_validate_addr,
>> > .ndo_set_mac_address = eth_mac_addr,
>> > .ndo_change_mtu = eth_change_mtu,
>> > #ifdef CONFIG_NET_POLL_CONTROLLER
>> > - .ndo_poll_controller = ei_poll,
>> > + .ndo_poll_controller = __ei_poll,
>> > #endif
>> > };
>>
>> And 8390.o can be removed from drivers/net/Makefile?
>
> I did not change that and it still works, modprobe zorro8390 did not load
> 8390. The speed is rather slow, though, 10kbit/sec seems to be the max
> speed. I have not been able to log in via ssh yet, maybe after it has
> finished downloading packages.
Yeah, speed depends on CPU type and motherboard/CPU interface.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply
* [PATCH net-2.6] vlan: fix GVRP at dismantle time
From: Eric Dumazet @ 2011-05-10 6:40 UTC (permalink / raw)
To: David Miller; +Cc: mirqus, alex, netdev, jesse, greearb, Patrick McHardy
In-Reply-To: <1304972251.3050.11.camel@edumazet-laptop>
Le lundi 09 mai 2011 à 22:17 +0200, Eric Dumazet a écrit :
> Le lundi 09 mai 2011 à 21:05 +0200, Eric Dumazet a écrit :
>
> > BTW, bug must be present in net-2.6, if we unload vlan module (since in this
> > case we also had a non NULL head )
>
> Yes, I confirm we have the bug in linux-2.6
>
Here is a patch to address this problem.
Thanks !
[PATCH net-2.6] vlan: fix GVRP at dismantle time
ip link add link eth2 eth2.103 type vlan id 103 gvrp on loose_binding on
ip link set eth2.103 up
rmmod tg3 # driver providing eth2
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<ffffffffa0030c9e>] garp_request_leave+0x3e/0xc0 [garp]
PGD 11d251067 PUD 11b9e0067 PMD 0
Oops: 0000 [#1] SMP
last sysfs file: /sys/devices/virtual/net/eth2.104/ifindex
CPU 0
Modules linked in: tg3(-) 8021q garp nfsd lockd auth_rpcgss sunrpc libphy sg [last unloaded: x_tables]
Pid: 11494, comm: rmmod Tainted: G W 2.6.39-rc6-00261-gfd71257-dirty #580 HP ProLiant BL460c G6
RIP: 0010:[<ffffffffa0030c9e>] [<ffffffffa0030c9e>] garp_request_leave+0x3e/0xc0 [garp]
RSP: 0018:ffff88007a19bae8 EFLAGS: 00010286
RAX: 0000000000000000 RBX: ffff88011b5e2000 RCX: 0000000000000002
RDX: 0000000000000000 RSI: 0000000000000175 RDI: ffffffffa0030d5b
RBP: ffff88007a19bb18 R08: 0000000000000001 R09: ffff88011bd64a00
R10: ffff88011d34ec00 R11: 0000000000000000 R12: 0000000000000002
R13: ffff88007a19bc48 R14: ffff88007a19bb88 R15: 0000000000000001
FS: 0000000000000000(0000) GS:ffff88011fc00000(0063) knlGS:00000000f77d76c0
CS: 0010 DS: 002b ES: 002b CR0: 000000008005003b
CR2: 0000000000000000 CR3: 000000011a675000 CR4: 00000000000006f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process rmmod (pid: 11494, threadinfo ffff88007a19a000, task ffff8800798595c0)
Stack:
ffff88007a19bb36 ffff88011c84b800 ffff88011b5e2000 ffff88007a19bc48
ffff88007a19bb88 0000000000000006 ffff88007a19bb38 ffffffffa003a5f6
ffff88007a19bb38 670088007a19bba8 ffff88007a19bb58 ffffffffa00397e7
Call Trace:
[<ffffffffa003a5f6>] vlan_gvrp_request_leave+0x46/0x50 [8021q]
[<ffffffffa00397e7>] vlan_dev_stop+0xb7/0xc0 [8021q]
[<ffffffff8137e427>] __dev_close_many+0x87/0xe0
[<ffffffff8137e507>] dev_close_many+0x87/0x110
[<ffffffff8137e630>] rollback_registered_many+0xa0/0x240
[<ffffffff8137e7e9>] unregister_netdevice_many+0x19/0x60
[<ffffffffa00389eb>] vlan_device_event+0x53b/0x550 [8021q]
[<ffffffff8143f448>] ? ip6mr_device_event+0xa8/0xd0
[<ffffffff81479d03>] notifier_call_chain+0x53/0x80
[<ffffffff81062539>] __raw_notifier_call_chain+0x9/0x10
[<ffffffff81062551>] raw_notifier_call_chain+0x11/0x20
[<ffffffff8137df82>] call_netdevice_notifiers+0x32/0x60
[<ffffffff8137e69f>] rollback_registered_many+0x10f/0x240
[<ffffffff8137e85f>] rollback_registered+0x2f/0x40
[<ffffffff8137e8c8>] unregister_netdevice_queue+0x58/0x90
[<ffffffff8137e9eb>] unregister_netdev+0x1b/0x30
[<ffffffffa005d73f>] tg3_remove_one+0x6f/0x10b [tg3]
We should call vlan_gvrp_request_leave() from unregister_vlan_dev(),
not from vlan_dev_stop(), because vlan_gvrp_uninit_applicant()
is called right after unregister_netdevice_queue(). In batch mode,
unregister_netdevice_queue() doesn’t immediately call vlan_dev_stop().
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Ben Greear <greearb@candelatech.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Jesse Gross <jesse@nicira.com>
Cc: Michał Mirosław <mirq-linux@rere.qmqm.pl>
---
net/8021q/vlan.c | 3 +++
net/8021q/vlan_dev.c | 3 ---
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index 7850412..0eb1a88 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -124,6 +124,9 @@ void unregister_vlan_dev(struct net_device *dev, struct list_head *head)
grp->nr_vlans--;
+ if (vlan->flags & VLAN_FLAG_GVRP)
+ vlan_gvrp_request_leave(dev);
+
vlan_group_set_device(grp, vlan_id, NULL);
if (!grp->killall)
synchronize_net();
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index e34ea9e..b2ff6c8 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -487,9 +487,6 @@ static int vlan_dev_stop(struct net_device *dev)
struct vlan_dev_info *vlan = vlan_dev_info(dev);
struct net_device *real_dev = vlan->real_dev;
- if (vlan->flags & VLAN_FLAG_GVRP)
- vlan_gvrp_request_leave(dev);
-
dev_mc_unsync(real_dev, dev);
dev_uc_unsync(real_dev, dev);
if (dev->flags & IFF_ALLMULTI)
^ permalink raw reply related
* slcan: fix ldisc->open retval
From: Oliver Hartkopp @ 2011-05-10 6:09 UTC (permalink / raw)
To: David Miller; +Cc: netdev, matvejchikov
TTY layer expects 0 if the ldisc->open operation succeeded.
Reported-by: Matvejchikov Ilya <matvejchikov@gmail.com>
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
---
diff --git a/drivers/net/can/slcan.c b/drivers/net/can/slcan.c
index b423965..1b49df6 100644
--- a/drivers/net/can/slcan.c
+++ b/drivers/net/can/slcan.c
@@ -583,7 +583,9 @@ static int slcan_open(struct tty_struct *tty)
/* Done. We have linked the TTY line to a channel. */
rtnl_unlock();
tty->receive_room = 65536; /* We don't flow control */
- return sl->dev->base_addr;
+
+ /* TTY layer expects 0 on success */
+ return 0;
err_free_chan:
sl->tty = NULL;
^ permalink raw reply related
* [PATCH] xfrm: Don't allow esn with disabled anti replay detection
From: Steffen Klassert @ 2011-05-10 5:43 UTC (permalink / raw)
To: David Miller, Herbert Xu; +Cc: netdev
Unlike the standard case, disabled anti replay detection needs some
nontrivial extra treatment on ESN. RFC 4303 states:
Note: If a receiver chooses to not enable anti-replay for an SA, then
the receiver SHOULD NOT negotiate ESN in an SA management protocol.
Use of ESN creates a need for the receiver to manage the anti-replay
window (in order to determine the correct value for the high-order
bits of the ESN, which are employed in the ICV computation), which is
generally contrary to the notion of disabling anti-replay for an SA.
So return an error if an ESN state with disabled anti replay detection
is inserted for now and add the extra treatment later if we need it.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
---
net/xfrm/xfrm_replay.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/net/xfrm/xfrm_replay.c b/net/xfrm/xfrm_replay.c
index e8a7814..47f1b86 100644
--- a/net/xfrm/xfrm_replay.c
+++ b/net/xfrm/xfrm_replay.c
@@ -535,6 +535,9 @@ int xfrm_init_replay(struct xfrm_state *x)
replay_esn->bmp_len * sizeof(__u32) * 8)
return -EINVAL;
+ if ((x->props.flags & XFRM_STATE_ESN) && replay_esn->replay_window == 0)
+ return -EINVAL;
+
if ((x->props.flags & XFRM_STATE_ESN) && x->replay_esn)
x->repl = &xfrm_replay_esn;
else
--
1.7.0.4
^ permalink raw reply related
* [PATCH] xfrm: Assign the inner mode output function to the dst entry
From: Steffen Klassert @ 2011-05-10 5:36 UTC (permalink / raw)
To: David Miller, Herbert Xu; +Cc: netdev
As it is, we assign the outer modes output function to the dst entry
when we create the xfrm bundle. This leads to two problems on interfamily
scenarios. We might insert ipv4 packets into ip6_fragment when called
from xfrm6_output. The system crashes if we try to fragment an ipv4
packet with ip6_fragment. This issue was introduced with git commit
ad0081e4 (ipv6: Fragment locally generated tunnel-mode IPSec6 packets
as needed). The second issue is, that we might insert ipv4 packets in
netfilter6 and vice versa on interfamily scenarios.
With this patch we assign the inner mode output function to the dst entry
when we create the xfrm bundle. So xfrm4_output/xfrm6_output from the inner
mode is used and the right fragmentation and netfilter functions are called.
We switch then to outer mode with the output_finish functions.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
---
include/net/xfrm.h | 3 +++
net/ipv4/xfrm4_output.c | 8 ++++++--
net/ipv4/xfrm4_state.c | 1 +
net/ipv6/xfrm6_output.c | 6 +++---
net/ipv6/xfrm6_state.c | 1 +
net/xfrm/xfrm_policy.c | 14 +++++++++++++-
6 files changed, 27 insertions(+), 6 deletions(-)
diff --git a/include/net/xfrm.h b/include/net/xfrm.h
index 6ae4bc5..20afeaa 100644
--- a/include/net/xfrm.h
+++ b/include/net/xfrm.h
@@ -324,6 +324,7 @@ struct xfrm_state_afinfo {
int (*tmpl_sort)(struct xfrm_tmpl **dst, struct xfrm_tmpl **src, int n);
int (*state_sort)(struct xfrm_state **dst, struct xfrm_state **src, int n);
int (*output)(struct sk_buff *skb);
+ int (*output_finish)(struct sk_buff *skb);
int (*extract_input)(struct xfrm_state *x,
struct sk_buff *skb);
int (*extract_output)(struct xfrm_state *x,
@@ -1454,6 +1455,7 @@ static inline int xfrm4_rcv_spi(struct sk_buff *skb, int nexthdr, __be32 spi)
extern int xfrm4_extract_output(struct xfrm_state *x, struct sk_buff *skb);
extern int xfrm4_prepare_output(struct xfrm_state *x, struct sk_buff *skb);
extern int xfrm4_output(struct sk_buff *skb);
+extern int xfrm4_output_finish(struct sk_buff *skb);
extern int xfrm4_tunnel_register(struct xfrm_tunnel *handler, unsigned short family);
extern int xfrm4_tunnel_deregister(struct xfrm_tunnel *handler, unsigned short family);
extern int xfrm6_extract_header(struct sk_buff *skb);
@@ -1470,6 +1472,7 @@ extern __be32 xfrm6_tunnel_spi_lookup(struct net *net, xfrm_address_t *saddr);
extern int xfrm6_extract_output(struct xfrm_state *x, struct sk_buff *skb);
extern int xfrm6_prepare_output(struct xfrm_state *x, struct sk_buff *skb);
extern int xfrm6_output(struct sk_buff *skb);
+extern int xfrm6_output_finish(struct sk_buff *skb);
extern int xfrm6_find_1stfragopt(struct xfrm_state *x, struct sk_buff *skb,
u8 **prevhdr);
diff --git a/net/ipv4/xfrm4_output.c b/net/ipv4/xfrm4_output.c
index 571aa96..2d51840 100644
--- a/net/ipv4/xfrm4_output.c
+++ b/net/ipv4/xfrm4_output.c
@@ -69,7 +69,7 @@ int xfrm4_prepare_output(struct xfrm_state *x, struct sk_buff *skb)
}
EXPORT_SYMBOL(xfrm4_prepare_output);
-static int xfrm4_output_finish(struct sk_buff *skb)
+int xfrm4_output_finish(struct sk_buff *skb)
{
#ifdef CONFIG_NETFILTER
if (!skb_dst(skb)->xfrm) {
@@ -86,7 +86,11 @@ static int xfrm4_output_finish(struct sk_buff *skb)
int xfrm4_output(struct sk_buff *skb)
{
+ struct dst_entry *dst = skb_dst(skb);
+ struct xfrm_state *x = dst->xfrm;
+
return NF_HOOK_COND(NFPROTO_IPV4, NF_INET_POST_ROUTING, skb,
- NULL, skb_dst(skb)->dev, xfrm4_output_finish,
+ NULL, dst->dev,
+ x->outer_mode->afinfo->output_finish,
!(IPCB(skb)->flags & IPSKB_REROUTED));
}
diff --git a/net/ipv4/xfrm4_state.c b/net/ipv4/xfrm4_state.c
index 1717c64..805d63e 100644
--- a/net/ipv4/xfrm4_state.c
+++ b/net/ipv4/xfrm4_state.c
@@ -78,6 +78,7 @@ static struct xfrm_state_afinfo xfrm4_state_afinfo = {
.init_tempsel = __xfrm4_init_tempsel,
.init_temprop = xfrm4_init_temprop,
.output = xfrm4_output,
+ .output_finish = xfrm4_output_finish,
.extract_input = xfrm4_extract_input,
.extract_output = xfrm4_extract_output,
.transport_finish = xfrm4_transport_finish,
diff --git a/net/ipv6/xfrm6_output.c b/net/ipv6/xfrm6_output.c
index 8e688b3..49a91c5 100644
--- a/net/ipv6/xfrm6_output.c
+++ b/net/ipv6/xfrm6_output.c
@@ -79,7 +79,7 @@ int xfrm6_prepare_output(struct xfrm_state *x, struct sk_buff *skb)
}
EXPORT_SYMBOL(xfrm6_prepare_output);
-static int xfrm6_output_finish(struct sk_buff *skb)
+int xfrm6_output_finish(struct sk_buff *skb)
{
#ifdef CONFIG_NETFILTER
IP6CB(skb)->flags |= IP6SKB_XFRM_TRANSFORMED;
@@ -97,9 +97,9 @@ static int __xfrm6_output(struct sk_buff *skb)
if ((x && x->props.mode == XFRM_MODE_TUNNEL) &&
((skb->len > ip6_skb_dst_mtu(skb) && !skb_is_gso(skb)) ||
dst_allfrag(skb_dst(skb)))) {
- return ip6_fragment(skb, xfrm6_output_finish);
+ return ip6_fragment(skb, x->outer_mode->afinfo->output_finish);
}
- return xfrm6_output_finish(skb);
+ return x->outer_mode->afinfo->output_finish(skb);
}
int xfrm6_output(struct sk_buff *skb)
diff --git a/net/ipv6/xfrm6_state.c b/net/ipv6/xfrm6_state.c
index afe941e..248f0b2 100644
--- a/net/ipv6/xfrm6_state.c
+++ b/net/ipv6/xfrm6_state.c
@@ -178,6 +178,7 @@ static struct xfrm_state_afinfo xfrm6_state_afinfo = {
.tmpl_sort = __xfrm6_tmpl_sort,
.state_sort = __xfrm6_state_sort,
.output = xfrm6_output,
+ .output_finish = xfrm6_output_finish,
.extract_input = xfrm6_extract_input,
.extract_output = xfrm6_extract_output,
.transport_finish = xfrm6_transport_finish,
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 15792d8..b4d745e 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -1406,6 +1406,7 @@ static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy,
struct net *net = xp_net(policy);
unsigned long now = jiffies;
struct net_device *dev;
+ struct xfrm_mode *inner_mode;
struct dst_entry *dst_prev = NULL;
struct dst_entry *dst0 = NULL;
int i = 0;
@@ -1436,6 +1437,17 @@ static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy,
goto put_states;
}
+ if (xfrm[i]->sel.family == AF_UNSPEC) {
+ inner_mode = xfrm_ip2inner_mode(xfrm[i],
+ xfrm_af2proto(family));
+ if (!inner_mode) {
+ err = -EAFNOSUPPORT;
+ dst_release(dst);
+ goto put_states;
+ }
+ } else
+ inner_mode = xfrm[i]->inner_mode;
+
if (!dst_prev)
dst0 = dst1;
else {
@@ -1464,7 +1476,7 @@ static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy,
dst1->lastuse = now;
dst1->input = dst_discard;
- dst1->output = xfrm[i]->outer_mode->afinfo->output;
+ dst1->output = inner_mode->afinfo->output;
dst1->next = dst_prev;
dst_prev = dst1;
--
1.7.0.4
^ permalink raw reply related
* [PATCH 10/10] ipv4: xfrm: Eliminate ->rt_src reference in policy code.
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
Rearrange xfrm4_dst_lookup() so that it works by calling a helper
function __xfrm_dst_lookup() that takes an explicit flow key storage
area as an argument.
Use this new helper in xfrm4_get_saddr() so we can fetch the selected
source address from the flow instead of from rt->rt_src
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/xfrm4_policy.c | 34 +++++++++++++++++++++-------------
1 files changed, 21 insertions(+), 13 deletions(-)
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 7ff973b..981e43e 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -18,38 +18,46 @@
static struct xfrm_policy_afinfo xfrm4_policy_afinfo;
-static struct dst_entry *xfrm4_dst_lookup(struct net *net, int tos,
- const xfrm_address_t *saddr,
- const xfrm_address_t *daddr)
+static struct dst_entry *__xfrm4_dst_lookup(struct net *net, struct flowi4 *fl4,
+ int tos,
+ const xfrm_address_t *saddr,
+ const xfrm_address_t *daddr)
{
- struct flowi4 fl4 = {
- .daddr = daddr->a4,
- .flowi4_tos = tos,
- };
struct rtable *rt;
+ memset(fl4, 0, sizeof(*fl4));
+ fl4->daddr = daddr->a4;
+ fl4->flowi4_tos = tos;
if (saddr)
- fl4.saddr = saddr->a4;
+ fl4->saddr = saddr->a4;
- rt = __ip_route_output_key(net, &fl4);
+ rt = __ip_route_output_key(net, fl4);
if (!IS_ERR(rt))
return &rt->dst;
return ERR_CAST(rt);
}
+static struct dst_entry *xfrm4_dst_lookup(struct net *net, int tos,
+ const xfrm_address_t *saddr,
+ const xfrm_address_t *daddr)
+{
+ struct flowi4 fl4;
+
+ return __xfrm4_dst_lookup(net, &fl4, tos, saddr, daddr);
+}
+
static int xfrm4_get_saddr(struct net *net,
xfrm_address_t *saddr, xfrm_address_t *daddr)
{
struct dst_entry *dst;
- struct rtable *rt;
+ struct flowi4 fl4;
- dst = xfrm4_dst_lookup(net, 0, NULL, daddr);
+ dst = __xfrm4_dst_lookup(net, &fl4, 0, NULL, daddr);
if (IS_ERR(dst))
return -EHOSTUNREACH;
- rt = (struct rtable *)dst;
- saddr->a4 = rt->rt_src;
+ saddr->a4 = fl4.saddr;
dst_release(dst);
return 0;
}
--
1.7.5.1
^ permalink raw reply related
* [PATCH 9/10] infiniband: Remove rt->rt_src usage in addr4_resolve()
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
Use an explicit flow key and fetch it from there.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/infiniband/core/addr.c | 9 +++++++--
1 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c
index 4ffc224..8e21d45 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -185,15 +185,20 @@ static int addr4_resolve(struct sockaddr_in *src_in,
__be32 dst_ip = dst_in->sin_addr.s_addr;
struct rtable *rt;
struct neighbour *neigh;
+ struct flowi4 fl4;
int ret;
- rt = ip_route_output(&init_net, dst_ip, src_ip, 0, addr->bound_dev_if);
+ memset(&fl4, 0, sizeof(fl4));
+ fl4.daddr = dst_ip;
+ fl4.saddr = src_ip;
+ fl4.flowi4_oif = addr->bound_dev_if;
+ rt = ip_route_output_key(&init_net, &fl4);
if (IS_ERR(rt)) {
ret = PTR_ERR(rt);
goto out;
}
src_in->sin_family = AF_INET;
- src_in->sin_addr.s_addr = rt->rt_src;
+ src_in->sin_addr.s_addr = fl4.saddr;
if (rt->dst.dev->flags & IFF_LOOPBACK) {
ret = rdma_translate_ip((struct sockaddr *) dst_in, addr);
--
1.7.5.1
^ permalink raw reply related
* [PATCH 8/10] sctp: Remove rt->rt_src usage in sctp_v4_get_saddr()
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
Flow key is available, so fetch it from there.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/sctp/protocol.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index 4f270ac..4de77cb 100644
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -566,7 +566,7 @@ static void sctp_v4_get_saddr(struct sctp_sock *sk,
if (rt) {
saddr->v4.sin_family = AF_INET;
- saddr->v4.sin_addr.s_addr = rt->rt_src;
+ saddr->v4.sin_addr.s_addr = fl->u.ip4.saddr;
}
}
--
1.7.5.1
^ permalink raw reply related
* [PATCH 7/10] ipvs: Remove all remaining references to rt->rt_{src,dst}
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
These values are always obtainable via the ip_vs_conn flow key.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/netfilter/ipvs/ip_vs_core.c | 2 +-
net/netfilter/ipvs/ip_vs_xmit.c | 17 +++++++++++++----
2 files changed, 14 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 07accf6..fa8c1fd 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -1384,7 +1384,7 @@ ip_vs_in_icmp(struct sk_buff *skb, int *related, unsigned int hooknum)
skb_rtable(skb)->rt_flags & RTCF_LOCAL) {
IP_VS_DBG(1, "%s(): "
"local delivery to %pI4 but in FORWARD\n",
- __func__, &skb_rtable(skb)->rt_dst);
+ __func__, &cp->fl.u.ip4.daddr);
verdict = NF_DROP;
}
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 2a300fe..99e7644 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -510,6 +510,7 @@ ip_vs_nat_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
struct rtable *rt; /* Route to the other host */
int mtu;
struct iphdr *iph = ip_hdr(skb);
+ struct flowi4 *fl4;
int local;
EnterFunction(10);
@@ -549,8 +550,10 @@ ip_vs_nat_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
}
#endif
+ fl4 = &cp->fl.u.ip4;
+
/* From world but DNAT to loopback address? */
- if (local && ipv4_is_loopback(rt->rt_dst) &&
+ if (local && ipv4_is_loopback(fl4->daddr) &&
rt_is_input_route(skb_rtable(skb))) {
IP_VS_DBG_RL_PKT(1, AF_INET, pp, skb, 0, "ip_vs_nat_xmit(): "
"stopping DNAT to loopback address");
@@ -767,6 +770,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
__be16 df = old_iph->frag_off;
struct iphdr *iph; /* Our new IP header */
unsigned int max_headroom; /* The extra header space needed */
+ struct flowi4 *fl4;
int mtu;
int ret;
@@ -833,6 +837,8 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
skb_dst_drop(skb);
skb_dst_set(skb, &rt->dst);
+ fl4 = &cp->fl.u.ip4;
+
/*
* Push down and install the IPIP header.
*/
@@ -842,8 +848,8 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
iph->frag_off = df;
iph->protocol = IPPROTO_IPIP;
iph->tos = tos;
- iph->daddr = rt->rt_dst;
- iph->saddr = rt->rt_src;
+ iph->daddr = fl4->daddr;
+ iph->saddr = fl4->saddr;
iph->ttl = old_iph->ttl;
ip_select_ident(iph, &rt->dst, NULL);
@@ -1127,6 +1133,7 @@ ip_vs_icmp_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
struct ip_vs_protocol *pp, int offset)
{
struct rtable *rt; /* Route to the other host */
+ struct flowi4 *fl4;
int mtu;
int rc;
int local;
@@ -1176,8 +1183,10 @@ ip_vs_icmp_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
}
#endif
+ fl4 = &cp->fl.u.ip4;
+
/* From world but DNAT to loopback address? */
- if (local && ipv4_is_loopback(rt->rt_dst) &&
+ if (local && ipv4_is_loopback(fl4->daddr) &&
rt_is_input_route(skb_rtable(skb))) {
IP_VS_DBG(1, "%s(): "
"stopping DNAT to loopback %pI4\n",
--
1.7.5.1
^ permalink raw reply related
* [PATCH 6/10] ipvs: Store a flow key in ip_vs_conn and use it in route lookups.
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
This is a key step in being able to eliminate the remaining references
to rt->rt_{src,dst} in the IPVS code.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/ip_vs.h | 1 +
net/netfilter/ipvs/ip_vs_xmit.c | 25 +++++++++++++++++--------
2 files changed, 18 insertions(+), 8 deletions(-)
diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h
index e0b7f13..6122c71 100644
--- a/include/net/ip_vs.h
+++ b/include/net/ip_vs.h
@@ -481,6 +481,7 @@ struct ip_vs_conn {
struct net *net; /* Name space */
#endif
/* Protocol, addresses and port numbers */
+ struct flowi fl;
u16 af; /* address family */
__be16 cport;
__be16 vport;
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index e5ef75b..2a300fe 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -86,19 +86,25 @@ __ip_vs_dst_check(struct ip_vs_dest *dest, u32 rtos)
/* Get route to destination or remote server */
static struct rtable *
-__ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_dest *dest,
+__ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_conn *cp,
+ struct ip_vs_dest *dest,
__be32 daddr, u32 rtos, int rt_mode)
{
struct net *net = dev_net(skb_dst(skb)->dev);
+ struct flowi4 *fl4;
struct rtable *rt; /* Route to the other host */
struct rtable *ort; /* Original route */
int local;
+ fl4 = &cp->fl.u.ip4;
if (dest) {
spin_lock(&dest->dst_lock);
if (!(rt = (struct rtable *)
__ip_vs_dst_check(dest, rtos))) {
- rt = ip_route_output(net, dest->addr.ip, 0, rtos, 0);
+ memset(fl4, 0, sizeof(*fl4));
+ fl4->daddr = dest->addr.ip;
+ fl4->flowi4_tos = rtos;
+ rt = ip_route_output_key(net, fl4);
if (IS_ERR(rt)) {
spin_unlock(&dest->dst_lock);
IP_VS_DBG_RL("ip_route_output error, dest: %pI4\n",
@@ -113,7 +119,10 @@ __ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_dest *dest,
daddr = dest->addr.ip;
spin_unlock(&dest->dst_lock);
} else {
- rt = ip_route_output(net, daddr, 0, rtos, 0);
+ memset(fl4, 0, sizeof(*fl4));
+ fl4->daddr = daddr;
+ fl4->flowi4_tos = rtos;
+ rt = ip_route_output_key(net, fl4);
if (IS_ERR(rt)) {
IP_VS_DBG_RL("ip_route_output error, dest: %pI4\n",
&daddr);
@@ -386,7 +395,7 @@ ip_vs_bypass_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
- if (!(rt = __ip_vs_get_out_rt(skb, NULL, iph->daddr, RT_TOS(iph->tos),
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, NULL, iph->daddr, RT_TOS(iph->tos),
IP_VS_RT_MODE_NON_LOCAL)))
goto tx_error_icmp;
@@ -515,7 +524,7 @@ ip_vs_nat_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
IP_VS_DBG(10, "filled cport=%d\n", ntohs(*p));
}
- if (!(rt = __ip_vs_get_out_rt(skb, cp->dest, cp->daddr.ip,
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, cp->dest, cp->daddr.ip,
RT_TOS(iph->tos),
IP_VS_RT_MODE_LOCAL |
IP_VS_RT_MODE_NON_LOCAL |
@@ -763,7 +772,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
- if (!(rt = __ip_vs_get_out_rt(skb, cp->dest, cp->daddr.ip,
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, cp->dest, cp->daddr.ip,
RT_TOS(tos), IP_VS_RT_MODE_LOCAL |
IP_VS_RT_MODE_NON_LOCAL)))
goto tx_error_icmp;
@@ -994,7 +1003,7 @@ ip_vs_dr_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
- if (!(rt = __ip_vs_get_out_rt(skb, cp->dest, cp->daddr.ip,
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, cp->dest, cp->daddr.ip,
RT_TOS(iph->tos),
IP_VS_RT_MODE_LOCAL |
IP_VS_RT_MODE_NON_LOCAL)))
@@ -1141,7 +1150,7 @@ ip_vs_icmp_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
* mangle and send the packet here (only for VS/NAT)
*/
- if (!(rt = __ip_vs_get_out_rt(skb, cp->dest, cp->daddr.ip,
+ if (!(rt = __ip_vs_get_out_rt(skb, cp, cp->dest, cp->daddr.ip,
RT_TOS(ip_hdr(skb)->tos),
IP_VS_RT_MODE_LOCAL |
IP_VS_RT_MODE_NON_LOCAL |
--
1.7.5.1
^ permalink raw reply related
* [PATCH 5/10] ipvs: Eliminate rt->rt_dst usage in __ip_vs_get_out_rt().
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
We can simply track what destination address is used based upon which
code block is taken at the top of the function.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/netfilter/ipvs/ip_vs_xmit.c | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index c4a19f9..e5ef75b 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -110,6 +110,7 @@ __ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_dest *dest,
&dest->addr.ip,
atomic_read(&rt->dst.__refcnt), rtos);
}
+ daddr = dest->addr.ip;
spin_unlock(&dest->dst_lock);
} else {
rt = ip_route_output(net, daddr, 0, rtos, 0);
@@ -125,7 +126,7 @@ __ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_dest *dest,
rt_mode)) {
IP_VS_DBG_RL("Stopping traffic to %s address, dest: %pI4\n",
(rt->rt_flags & RTCF_LOCAL) ?
- "local":"non-local", &rt->rt_dst);
+ "local":"non-local", &daddr);
ip_rt_put(rt);
return NULL;
}
@@ -133,14 +134,14 @@ __ip_vs_get_out_rt(struct sk_buff *skb, struct ip_vs_dest *dest,
!((ort = skb_rtable(skb)) && ort->rt_flags & RTCF_LOCAL)) {
IP_VS_DBG_RL("Redirect from non-local address %pI4 to local "
"requires NAT method, dest: %pI4\n",
- &ip_hdr(skb)->daddr, &rt->rt_dst);
+ &ip_hdr(skb)->daddr, &daddr);
ip_rt_put(rt);
return NULL;
}
if (unlikely(!local && ipv4_is_loopback(ip_hdr(skb)->saddr))) {
IP_VS_DBG_RL("Stopping traffic from loopback address %pI4 "
"to non-local address, dest: %pI4\n",
- &ip_hdr(skb)->saddr, &rt->rt_dst);
+ &ip_hdr(skb)->saddr, &daddr);
ip_rt_put(rt);
return NULL;
}
--
1.7.5.1
^ permalink raw reply related
* [PATCH 4/10] ipvs: Use IP_VS_RT_MODE_* instead of magic constants.
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/netfilter/ipvs/ip_vs_xmit.c | 17 ++++++++++++-----
1 files changed, 12 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 6132b21..c4a19f9 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -440,7 +440,8 @@ ip_vs_bypass_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
- if (!(rt = __ip_vs_get_out_rt_v6(skb, NULL, &iph->daddr, NULL, 0, 2)))
+ if (!(rt = __ip_vs_get_out_rt_v6(skb, NULL, &iph->daddr, NULL, 0,
+ IP_VS_RT_MODE_NON_LOCAL)))
goto tx_error_icmp;
/* MTU checking */
@@ -632,7 +633,9 @@ ip_vs_nat_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
}
if (!(rt = __ip_vs_get_out_rt_v6(skb, cp->dest, &cp->daddr.in6, NULL,
- 0, 1|2|4)))
+ 0, (IP_VS_RT_MODE_LOCAL |
+ IP_VS_RT_MODE_NON_LOCAL |
+ IP_VS_RT_MODE_RDR))))
goto tx_error_icmp;
local = __ip_vs_is_local_route6(rt);
/*
@@ -875,7 +878,8 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
if (!(rt = __ip_vs_get_out_rt_v6(skb, cp->dest, &cp->daddr.in6,
- &saddr, 1, 1|2)))
+ &saddr, 1, (IP_VS_RT_MODE_LOCAL |
+ IP_VS_RT_MODE_NON_LOCAL))))
goto tx_error_icmp;
if (__ip_vs_is_local_route6(rt)) {
dst_release(&rt->dst);
@@ -1050,7 +1054,8 @@ ip_vs_dr_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
EnterFunction(10);
if (!(rt = __ip_vs_get_out_rt_v6(skb, cp->dest, &cp->daddr.in6, NULL,
- 0, 1|2)))
+ 0, (IP_VS_RT_MODE_LOCAL |
+ IP_VS_RT_MODE_NON_LOCAL))))
goto tx_error_icmp;
if (__ip_vs_is_local_route6(rt)) {
dst_release(&rt->dst);
@@ -1254,7 +1259,9 @@ ip_vs_icmp_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
*/
if (!(rt = __ip_vs_get_out_rt_v6(skb, cp->dest, &cp->daddr.in6, NULL,
- 0, 1|2|4)))
+ 0, (IP_VS_RT_MODE_LOCAL |
+ IP_VS_RT_MODE_NON_LOCAL |
+ IP_VS_RT_MODE_RDR))))
goto tx_error_icmp;
local = __ip_vs_is_local_route6(rt);
--
1.7.5.1
^ permalink raw reply related
* [PATCH 3/10] ipv4: udp: Eliminate remaining uses of rt->rt_src
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
We already track and pass around the correct flow key,
so simply use it in udp_send_skb().
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/udp.c | 13 ++++++-------
1 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 66341a3..599374f 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -706,12 +706,11 @@ static void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
}
}
-static int udp_send_skb(struct sk_buff *skb, __be32 daddr, __be32 dport)
+static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
{
struct sock *sk = skb->sk;
struct inet_sock *inet = inet_sk(sk);
struct udphdr *uh;
- struct rtable *rt = (struct rtable *)skb_dst(skb);
int err = 0;
int is_udplite = IS_UDPLITE(sk);
int offset = skb_transport_offset(skb);
@@ -723,7 +722,7 @@ static int udp_send_skb(struct sk_buff *skb, __be32 daddr, __be32 dport)
*/
uh = udp_hdr(skb);
uh->source = inet->inet_sport;
- uh->dest = dport;
+ uh->dest = fl4->fl4_dport;
uh->len = htons(len);
uh->check = 0;
@@ -737,14 +736,14 @@ static int udp_send_skb(struct sk_buff *skb, __be32 daddr, __be32 dport)
} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
- udp4_hwcsum(skb, rt->rt_src, daddr);
+ udp4_hwcsum(skb, fl4->saddr, fl4->daddr);
goto send;
} else
csum = udp_csum(skb);
/* add protocol-dependent pseudo-header */
- uh->check = csum_tcpudp_magic(rt->rt_src, daddr, len,
+ uh->check = csum_tcpudp_magic(fl4->saddr, fl4->daddr, len,
sk->sk_protocol, csum);
if (uh->check == 0)
uh->check = CSUM_MANGLED_0;
@@ -778,7 +777,7 @@ static int udp_push_pending_frames(struct sock *sk)
if (!skb)
goto out;
- err = udp_send_skb(skb, fl4->daddr, fl4->fl4_dport);
+ err = udp_send_skb(skb, fl4);
out:
up->len = 0;
@@ -963,7 +962,7 @@ back_from_confirm:
msg->msg_flags);
err = PTR_ERR(skb);
if (skb && !IS_ERR(skb))
- err = udp_send_skb(skb, daddr, dport);
+ err = udp_send_skb(skb, fl4);
goto out;
}
--
1.7.5.1
^ permalink raw reply related
* [PATCH 2/10] ipv4: icmp: Eliminate remaining uses of rt->rt_src
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
On input packets, rt->rt_src always equals ip_hdr(skb)->saddr
Anything that mangles or otherwise changes the IP header must
relookup the route found at skb_rtable(). Therefore this
invariant must always hold true.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/icmp.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 853a670..3314394 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -345,7 +345,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
icmp_param->data.icmph.checksum = 0;
inet->tos = ip_hdr(skb)->tos;
- daddr = ipc.addr = rt->rt_src;
+ daddr = ipc.addr = ip_hdr(skb)->saddr;
ipc.opt = NULL;
ipc.tx_flags = 0;
if (icmp_param->replyopts.opt.opt.optlen) {
@@ -930,12 +930,12 @@ static void icmp_address_reply(struct sk_buff *skb)
BUG_ON(mp == NULL);
for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
if (*mp == ifa->ifa_mask &&
- inet_ifa_match(rt->rt_src, ifa))
+ inet_ifa_match(ip_hdr(skb)->saddr, ifa))
break;
}
if (!ifa && net_ratelimit()) {
printk(KERN_INFO "Wrong address mask %pI4 from %s/%pI4\n",
- mp, dev->name, &rt->rt_src);
+ mp, dev->name, &ip_hdr(skb)->saddr);
}
}
}
--
1.7.5.1
^ permalink raw reply related
* [PATCH 1/10] ipv4: Pass explicit daddr arg to ip_send_reply().
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
This eliminates an access to rt->rt_src.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/ip.h | 4 ++--
net/ipv4/ip_output.c | 7 +++----
net/ipv4/tcp_ipv4.c | 4 ++--
3 files changed, 7 insertions(+), 8 deletions(-)
diff --git a/include/net/ip.h b/include/net/ip.h
index 0b30d3a..07118c7 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -174,8 +174,8 @@ static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
}
-void ip_send_reply(struct sock *sk, struct sk_buff *skb, struct ip_reply_arg *arg,
- unsigned int len);
+void ip_send_reply(struct sock *sk, struct sk_buff *skb, __be32 daddr,
+ struct ip_reply_arg *arg, unsigned int len);
struct ipv4_config {
int log_martians;
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index cd89d22..70778d4 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1459,20 +1459,19 @@ static int ip_reply_glue_bits(void *dptr, char *to, int offset,
* Should run single threaded per socket because it uses the sock
* structure to pass arguments.
*/
-void ip_send_reply(struct sock *sk, struct sk_buff *skb, struct ip_reply_arg *arg,
- unsigned int len)
+void ip_send_reply(struct sock *sk, struct sk_buff *skb, __be32 daddr,
+ struct ip_reply_arg *arg, unsigned int len)
{
struct inet_sock *inet = inet_sk(sk);
struct ip_options_data replyopts;
struct ipcm_cookie ipc;
struct flowi4 fl4;
- __be32 daddr;
struct rtable *rt = skb_rtable(skb);
if (ip_options_echo(&replyopts.opt.opt, skb))
return;
- daddr = ipc.addr = rt->rt_src;
+ ipc.addr = daddr;
ipc.opt = NULL;
ipc.tx_flags = 0;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 2b65503..f67fb34 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -651,7 +651,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
arg.flags = (sk && inet_sk(sk)->transparent) ? IP_REPLY_ARG_NOSRCCHECK : 0;
net = dev_net(skb_dst(skb)->dev);
- ip_send_reply(net->ipv4.tcp_sock, skb,
+ ip_send_reply(net->ipv4.tcp_sock, skb, ip_hdr(skb)->saddr,
&arg, arg.iov[0].iov_len);
TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
@@ -726,7 +726,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
if (oif)
arg.bound_dev_if = oif;
- ip_send_reply(net->ipv4.tcp_sock, skb,
+ ip_send_reply(net->ipv4.tcp_sock, skb, ip_hdr(skb)->saddr,
&arg, arg.iov[0].iov_len);
TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
--
1.7.5.1
^ permalink raw reply related
* [PATCH 0/10] Another batch of rt->rt_{src,dst} removals
From: David Miller @ 2011-05-10 5:31 UTC (permalink / raw)
To: netdev
After this series there are very nearly no references remaining
outside of net/ipv4/route.c
They all come in the usual pattern, making sure the flow key can be
materialized at the site where rt->rt_{src,dst} is needed.
Signed-off-by: David S. Miller <davem@davemloft.net>
^ permalink raw reply
* Re: Bug#625914: linux-image-2.6.38-2-amd64: bridging is not interacting well with multicast in 2.6.38-4
From: Noah Meyerhans @ 2011-05-10 4:38 UTC (permalink / raw)
To: Ben Hutchings; +Cc: 625914, netdev, bridge
In-Reply-To: <1304995124.4065.157.camel@localhost>
[-- Attachment #1: Type: text/plain, Size: 603 bytes --]
On Tue, May 10, 2011 at 03:38:44AM +0100, Ben Hutchings wrote:
> This is pretty weird. Debian version 2.6.38-3 has a few bridging
> changes from stable 2.6.38.3 and 2.6.38.4, but they don't look like they
> would cause this.
I have apparently filed the bug against the wrong version of Debian's
kernel. 2.6.38-3 is not affected, and works as expected. The change
was introduced in -4. That may have been clear from the report itself,
but the report was filed against -3. I've fixed that in the BTS.
I've also confirmed that -5 is affected, to no great surprise.
I'll investigate further.
noah
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply
* RE: [RFC V2 PATCH] rtnetlink: Add method to calculate dump info data size
From: Rose, Gregory V @ 2011-05-10 4:25 UTC (permalink / raw)
To: David Miller, eric.dumazet@gmail.com
Cc: shemminger@vyatta.com, netdev@vger.kernel.org,
bhutchings@solarflare.com
In-Reply-To: <20110509.204959.71127408.davem@davemloft.net>
> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Monday, May 09, 2011 8:50 PM
> To: eric.dumazet@gmail.com
> Cc: shemminger@vyatta.com; Rose, Gregory V; netdev@vger.kernel.org;
> bhutchings@solarflare.com
> Subject: Re: [RFC V2 PATCH] rtnetlink: Add method to calculate dump info
> data size
>
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Tue, 10 May 2011 05:45:27 +0200
>
> > We only need to maintain a global variable to hold min_dump_alloc
> >
> > [ and only increase this variable when necessary, not bother to decrease
> > it when removing a driver ]
>
> Agreed, this is probably the best approach.
[Greg Rose]
OK, let me go back and look at doing it this way. I was sort of scratching my head about doing that device lookup on every dump, so this is a good suggestion.
Thanks,
- Greg
^ permalink raw reply
* Re: [PATCH net-next] netdevice.h: Align struct netdevices members
From: Joe Perches @ 2011-05-10 4:22 UTC (permalink / raw)
To: Eric Dumazet; +Cc: David Miller, netdev
In-Reply-To: <1304999595.3050.45.camel@edumazet-laptop>
On Tue, 2011-05-10 at 05:53 +0200, Eric Dumazet wrote:
> Le lundi 09 mai 2011 à 20:42 -0700, Joe Perches a écrit :
> > Save a bit of space.
> Hmm... correct alignements are far more important for this structure.
I agree. I aligned these members on "natural" boundaries.
4 consecutive chars rather than 2 chars, unsigned long.
> Did you run benchmarks on 32bit and 64bit platforms ?
Nope.
> BTW we have ____cacheline_aligned_in_smp markers, I am not even sure
> this patch saves space.
That depends on whether or not CONFIG_SMP is defined and
all of the changes are before any ____cacheline_aligned markers.
^ permalink raw reply
* Re: [RFC V2 PATCH] rtnetlink: Add method to calculate dump info data size
From: Eric Dumazet @ 2011-05-10 4:03 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: Greg Rose, netdev, bhutchings, davem
In-Reply-To: <20110509205626.19dede92@nehalam>
Le lundi 09 mai 2011 à 20:56 -0700, Stephen Hemminger a écrit :
> What happens is dump iterates over the table and puts entries into
> skb. When space is exhausted in skb the iterator stops and records the
> key of the where to restart.
Exactly : We would increase min_dump_alloc at this point, and not fully
recompute it at each round ;)
Problem is we could return an empty answer if first device ever dumped
in "ip link" would be a 'large one' and min_dump_alloc not yet
increased, so "ip link" would consider table is empty ?
Thats why I suggested to increase min_dump_alloc() when bloated^W driver
is loaded.
^ permalink raw reply
* Re: [RFC V2 PATCH] rtnetlink: Add method to calculate dump info data size
From: David Miller @ 2011-05-10 3:58 UTC (permalink / raw)
To: shemminger; +Cc: eric.dumazet, gregory.v.rose, netdev, bhutchings
In-Reply-To: <20110509205626.19dede92@nehalam>
From: Stephen Hemminger <shemminger@vyatta.com>
Date: Mon, 9 May 2011 20:56:26 -0700
> I was hoping that the new interface dump would not need a pre-calculated
> size and could just incrementally add values. I was trying to use an
> analogy with route dumping. The current route dump does not precalculate
> size.
Stephen the problem here is that one netdevice being dumped is highly
variable because of the VF netlink attributes.
Routes have a fixed reasonably low upper bound in size, so we always are
able to fit one route in an SKB.
Netdevices, on the other hand, can have so many VFs that one netdevice
entry does not fit even in a single 8K allocated SKB.
That's the failure case this code is trying to fix.
^ permalink raw reply
* Re: 2.6.39-rc6-mmotm0506 - another lockdep splat (networking this time)
From: David Miller @ 2011-05-10 3:58 UTC (permalink / raw)
To: eric.dumazet; +Cc: Valdis.Kletnieks, akpm, linux-kernel, netdev
In-Reply-To: <1304999271.3050.41.camel@edumazet-laptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 10 May 2011 05:47:51 +0200
> [PATCH net-next-2.6] net: fix two lockdep splats
>
> Commit e67f88dd12f6 (net: dont hold rtnl mutex during netlink dump
> callbacks) switched rtnl protection to RCU, but we forgot to adjust two
> rcu_dereference() lockdep annotations :
>
> inet_get_link_af_size() or inet_fill_link_af() might be called with
> rcu_read_lock or rtnl held, so use rcu_dereference_rtnl()
> instead of rtnl_dereference()
>
> Reported-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Applied, thanks everyone.
^ permalink raw reply
* Re: [RFC V2 PATCH] rtnetlink: Add method to calculate dump info data size
From: Stephen Hemminger @ 2011-05-10 3:56 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Greg Rose, netdev, bhutchings, davem
In-Reply-To: <1304999127.3050.40.camel@edumazet-laptop>
On Tue, 10 May 2011 05:45:27 +0200
Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Le lundi 09 mai 2011 à 20:17 -0700, Stephen Hemminger a écrit :
> > On Tue, 10 May 2011 04:43:33 +0200
> > Eric Dumazet <eric.dumazet@gmail.com> wrote:
> >
> > > Le lundi 09 mai 2011 à 15:26 -0700, Greg Rose a écrit :
> > > > The message size allocated for rtnl info dumps was limited to a single
> > > > page. This is not enough for additional interface info available with
> > > > devices that support SR-IOV. Calculate the amount of data required so
> > > > the dump can allocate enough data to satisfy the request.
> > > >
> > > > V2 of this patch adds a new argument to the rtnl_register service that
> > > > allows for a new method to calculate the amount of data required to
> > > > complete the info dump request. So far the method is only implemented
> > > > for the RTM_GETLINK slot.
> > > >
> > > > Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
> > >
> > > >
> > > > +static u16 rtnl_calcit(struct sk_buff *skb)
> > > > +{
> > > > + struct net *net = sock_net(skb->sk);
> > > > + int h;
> > > > + int idx = 0, s_idx;
> > > > + struct net_device *dev;
> > > > + struct hlist_head *head;
> > > > + struct hlist_node *node;
> > > > + u16 alloc_size = 0;
> > > > +
> > > > + for (h = 0; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
> > > > + idx = 0;
> > > > + head = &net->dev_index_head[h];
> > > > + hlist_for_each_entry(dev, node, head, index_hlist) {
> > > > + if (idx < s_idx) {
> > > > + idx++;
> > > > + continue;
> > > > + }
> > > > + alloc_size = (u16)if_nlmsg_size(dev);
> > > > + break;
> > > > + }
> > > > + }
> > > > +
> > > > + return alloc_size;
> > > > +}
> > > > +
> > >
> > >
> > > Sorry this wont scale. Some machines have thousand of devices.
> > >
> > > Just make an upper approximation, you dont need an exact one ;)
> >
> > The route dump does scale, can't you use a similar logic?
> > The result doesn't come back as one huge allocation.
> > I regularly test 600K routes on small machines.
> >
>
> Not sure I understand you Stephen.
>
> In Greg patch, rtnl_calcit() would be called for every 4K/8K block "ip"
> gets from kernel.
>
> If you add a function to route dump that would scan the 600K routes to
> get the max route size, surely you notice O(N^2) complexity instead of
> O(N)
>
> We only need to maintain a global variable to hold min_dump_alloc
I was hoping that the new interface dump would not need a pre-calculated
size and could just incrementally add values. I was trying to use an
analogy with route dumping. The current route dump does not precalculate
size.
What happens is dump iterates over the table and puts entries into
skb. When space is exhausted in skb the iterator stops and records the
key of the where to restart. Then restarts with next skb from there.
This scales O(N) with number of routes and does not have to precompute
size.
--
^ permalink raw reply
* Re: [PATCH net-next] netdevice.h: Align struct netdevices members
From: Eric Dumazet @ 2011-05-10 3:53 UTC (permalink / raw)
To: Joe Perches; +Cc: David Miller, netdev
In-Reply-To: <1304998966.19586.104.camel@Joe-Laptop>
Le lundi 09 mai 2011 à 20:42 -0700, Joe Perches a écrit :
> Save a bit of space.
>
> Signed-off-by: Joe Perches <joe@perches.com>
> ---
> include/linux/netdevice.h | 8 ++++----
> 1 files changed, 4 insertions(+), 4 deletions(-)
>
Hmm... correct alignements are far more important for this structure.
Did you run benchmarks on 32bit and 64bit platforms ?
BTW we have ____cacheline_aligned_in_smp markers, I am not even sure
this patch saves space.
> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> index d5de66a..0fcb86b 100644
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -1020,9 +1020,6 @@ struct net_device {
> * part of the usual set specified in Space.c.
> */
>
> - unsigned char if_port; /* Selectable AUI, TP,..*/
> - unsigned char dma; /* DMA channel */
> -
> unsigned long state;
>
> struct list_head dev_list;
> @@ -1141,13 +1138,16 @@ struct net_device {
> const struct header_ops *header_ops;
>
> unsigned int flags; /* interface flags (a la BSD) */
> + unsigned int priv_flags; /* Like 'flags' but invisible to userspace. */
> unsigned short gflags;
> - unsigned int priv_flags; /* Like 'flags' but invisible to userspace. */
> unsigned short padded; /* How much padding added by alloc_netdev() */
>
> unsigned char operstate; /* RFC2863 operstate */
> unsigned char link_mode; /* mapping policy to operstate */
>
> + unsigned char if_port; /* Selectable AUI, TP,..*/
> + unsigned char dma; /* DMA channel */
> +
> unsigned int mtu; /* interface MTU value */
> unsigned short type; /* interface hardware type */
> unsigned short hard_header_len; /* hardware hdr length */
>
^ 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