* [PATCH] LRO ack aggregation
From: Andrew Gallatin @ 2007-10-23 15:11 UTC (permalink / raw)
To: netdev; +Cc: ossthema
[-- Attachment #1: Type: text/plain, Size: 3466 bytes --]
Hi,
We recently did some performance comparisons between the new inet_lro
LRO support in the kernel, and our Myri10GE in-driver LRO.
For receive, we found they were nearly identical. However, for
transmit, we found that Myri10GE's LRO shows much lower CPU
utilization. We traced the CPU utilization difference to our driver
LRO aggregating TCP acks, and the inet_lro module not doing so.
I've attached a patch which adds support to inet_lro for aggregating
pure acks. Aggregating pure acks (segments with TCP_PAYLOAD_LENGTH ==
0) entails freeing the skb (or putting the page in the frags case).
The patch also handles trimming (typical for 54-byte pure ack frames
which have been padded to the ethernet minimum 60 byte frame size).
In the frags case, I tried to keep things simple by only doing the
trim when the entire frame fits in the first frag. To be safe, I
ensure that the padding is all 0 (or, more exactly, was some pattern
whose checksum is -0) so that it doesn't impact hardware checksums.
This patch also fixes a small bug in the skb LRO path dealing with
vlans that I found when doing my own testing. Specifically, in the
desc->active case, the existing code always fails the
lro_tcp_ip_check() for NICs without LRO_F_EXTRACT_VLAN_ID, because it
fails to subtract the vlan_hdr_len from the skb->len.
Jan-Bernd Themann (ossthema@de.ibm.com) has tested the patch using
the eHEA driver (skb codepath), and I have tested it using Myri10GE
(both frags and skb codepath).
Using a pair of identical low-end 2.0GHz Athlon 64 x2 3800+ with
Myri10GE 10GbE NICs, I ran 10 iterations of netperf TCP_SENDFILE
tests, taking the median run for comparison purposes. The receiver
was running Myri10GE + patched inet_lro:
TCP SENDFILE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to rome-my
(192.168.1.16) port 0 AF_INET : cpu bind
Recv Send Send Utilization Service
Demand
Socket Socket Message Elapsed Send Recv Send Recv
Size Size Size Time Throughput local remote local
remote
bytes bytes bytes secs. 10^6bits/s % S % S us/KB us/KB
Myri10GE driver-specific LRO:
87380 65536 65536 60.02 9442.65 16.24 69.31 0.282 1.203
Myri10GE + unpatched inet_lro:
87380 65536 65536 60.02 9442.88 20.10 69.11 0.349 1.199
Myri10GE + patched inet_lro:
87380 65536 65536 60.02 9443.30 16.95 68.97 0.294 1.197
The important bit here is the sender's CPU utilization, and service
demand (cost per byte). As you can see, without aggregating ack's,
the overhead on the sender is roughly 20% higher, even when sending to
a receiver which uses LRO. The differences are even more dramatic
when sending to a receiver which does not use LRO (and hence sends
more frequent acks).
Below is the same benchmark, run between a pair of 4-way 3.0GHz Xeon
5160 machines (Dell 2950) with Myri10GE NICs. The receiver is running
Solaris 10U4, which does not do LRO, and is acking at approximately
8:1 (or ~100K acks/sec):
Myri10GE driver-specific LRO:
196712 65536 65536 60.01 9280.09 7.14 45.37 0.252 1.602
Myri10GE + unpatched inet_lro:
196712 65536 65536 60.01 8530.80 10.51 44.60 0.404 1.713
Myri10GE + patched inet_lro:
196712 65536 65536 60.00 9249.65 7.21 45.90 0.255 1.626
Signed off by: Andrew Gallatin <gallatin@myri.com>
Andrew Gallatin
Myricom Inc.
[-- Attachment #2: ack_aggr.diff --]
[-- Type: text/plain, Size: 3789 bytes --]
diff --git a/net/ipv4/inet_lro.c b/net/ipv4/inet_lro.c
index ac3b1d3..eba145b 100644
--- a/net/ipv4/inet_lro.c
+++ b/net/ipv4/inet_lro.c
@@ -58,9 +58,6 @@ static int lro_tcp_ip_check(struct iphdr
if (ntohs(iph->tot_len) != len)
return -1;
- if (TCP_PAYLOAD_LENGTH(iph, tcph) == 0)
- return -1;
-
if (iph->ihl != IPH_LEN_WO_OPTIONS)
return -1;
@@ -223,6 +220,11 @@ static void lro_add_packet(struct net_lr
lro_add_common(lro_desc, iph, tcph, tcp_data_len);
+ if (tcp_data_len == 0) {
+ dev_kfree_skb_any(skb);
+ return;
+ }
+
skb_pull(skb, (skb->len - tcp_data_len));
parent->truesize += skb->truesize;
@@ -244,6 +246,11 @@ static void lro_add_frags(struct net_lro
lro_add_common(lro_desc, iph, tcph, tcp_data_len);
+ if (tcp_data_len == 0) {
+ put_page(skb_frags[0].page);
+ return;
+ }
+
skb->truesize += truesize;
skb_frags[0].page_offset += hlen;
@@ -338,6 +345,8 @@ static int __lro_proc_skb(struct net_lro
struct tcphdr *tcph;
u64 flags;
int vlan_hdr_len = 0;
+ int pkt_len;
+ int trim;
if (!lro_mgr->get_skb_header
|| lro_mgr->get_skb_header(skb, (void *)&iph, (void *)&tcph,
@@ -355,6 +364,17 @@ static int __lro_proc_skb(struct net_lro
&& !test_bit(LRO_F_EXTRACT_VLAN_ID, &lro_mgr->features))
vlan_hdr_len = VLAN_HLEN;
+ /* strip padding from runts iff the padding is all zero */
+ pkt_len = vlan_hdr_len + ntohs(iph->tot_len);
+ trim = skb->len - pkt_len;
+ if (trim > 0) {
+ u8 *pad = skb_tail_pointer(skb) - trim;
+ if (unlikely(ip_compute_csum(pad, trim) != 0xffff)) {
+ goto out;
+ }
+ skb_trim(skb, skb->len - trim);
+ }
+
if (!lro_desc->active) { /* start new lro session */
if (lro_tcp_ip_check(iph, tcph, skb->len - vlan_hdr_len, NULL))
goto out;
@@ -368,7 +388,7 @@ static int __lro_proc_skb(struct net_lro
if (lro_desc->tcp_next_seq != ntohl(tcph->seq))
goto out2;
- if (lro_tcp_ip_check(iph, tcph, skb->len, lro_desc))
+ if (lro_tcp_ip_check(iph, tcph, skb->len - vlan_hdr_len, lro_desc))
goto out2;
lro_add_packet(lro_desc, skb, iph, tcph);
@@ -412,18 +432,20 @@ static struct sk_buff *lro_gen_skb(struc
memcpy(skb->data, mac_hdr, hdr_len);
- skb_frags = skb_shinfo(skb)->frags;
- while (data_len > 0) {
- *skb_frags = *frags;
- data_len -= frags->size;
- skb_frags++;
- frags++;
- skb_shinfo(skb)->nr_frags++;
+ if (skb->data_len == 0) {
+ put_page(frags[0].page);
+ } else {
+ skb_frags = skb_shinfo(skb)->frags;
+ while (data_len > 0) {
+ *skb_frags = *frags;
+ data_len -= frags->size;
+ skb_frags++;
+ frags++;
+ skb_shinfo(skb)->nr_frags++;
+ }
+ skb_shinfo(skb)->frags[0].page_offset += hdr_len;
+ skb_shinfo(skb)->frags[0].size -= hdr_len;
}
-
- skb_shinfo(skb)->frags[0].page_offset += hdr_len;
- skb_shinfo(skb)->frags[0].size -= hdr_len;
-
skb->ip_summed = ip_summed;
skb->csum = sum;
skb->protocol = eth_type_trans(skb, lro_mgr->dev);
@@ -445,6 +467,8 @@ static struct sk_buff *__lro_proc_segmen
int mac_hdr_len;
int hdr_len = LRO_MAX_PG_HLEN;
int vlan_hdr_len = 0;
+ int pkt_len;
+ int trim;
if (!lro_mgr->get_frag_header
|| lro_mgr->get_frag_header(frags, (void *)&mac_hdr, (void *)&iph,
@@ -463,6 +487,19 @@ static struct sk_buff *__lro_proc_segmen
if (!lro_desc)
goto out1;
+ /* strip padding from runts iff the padding is all zero */
+ pkt_len = mac_hdr_len + ntohs(iph->tot_len);
+ trim = len - pkt_len;
+ if (trim > 0 && pkt_len <= frags->size) {
+ u8 *pad = page_address(frags->page) + frags->page_offset +
+ pkt_len;
+ if (unlikely(ip_compute_csum(pad, trim) != 0xffff))
+ goto out1;
+ frags->size -= trim;
+ len -= trim;
+ true_size -= trim;
+ }
+
if (!lro_desc->active) { /* start new lro session */
if (lro_tcp_ip_check(iph, tcph, len - mac_hdr_len, NULL))
goto out1;
^ permalink raw reply related
* [PATCH] Consolidate sctp_ulpq_renege_xxx functions
From: Pavel Emelyanov @ 2007-10-23 15:17 UTC (permalink / raw)
To: David Miller; +Cc: Vlad Yasevich, lksctp-developers, Linux Netdev List, devel
Both are equal, except for the list to be traversed.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
---
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index b937095..4be92d0 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -908,8 +908,8 @@ void sctp_ulpq_skip(struct sctp_ulpq *ulpq, __u16 sid, __u16 ssn)
return;
}
-/* Renege 'needed' bytes from the ordering queue. */
-static __u16 sctp_ulpq_renege_order(struct sctp_ulpq *ulpq, __u16 needed)
+static __u16 sctp_ulpq_renege_list(struct sctp_ulpq *ulpq,
+ struct sk_buff_head *list, __u16 needed)
{
__u16 freed = 0;
__u32 tsn;
@@ -919,7 +919,7 @@ static __u16 sctp_ulpq_renege_order(struct sctp_ulpq *ulpq, __u16 needed)
tsnmap = &ulpq->asoc->peer.tsn_map;
- while ((skb = __skb_dequeue_tail(&ulpq->lobby)) != NULL) {
+ while ((skb = __skb_dequeue_tail(list)) != NULL) {
freed += skb_headlen(skb);
event = sctp_skb2event(skb);
tsn = event->tsn;
@@ -933,30 +933,16 @@ static __u16 sctp_ulpq_renege_order(struct sctp_ulpq *ulpq, __u16 needed)
return freed;
}
+/* Renege 'needed' bytes from the ordering queue. */
+static __u16 sctp_ulpq_renege_order(struct sctp_ulpq *ulpq, __u16 needed)
+{
+ return sctp_ulpq_renege_list(ulpq, &ulpq->lobby, needed);
+}
+
/* Renege 'needed' bytes from the reassembly queue. */
static __u16 sctp_ulpq_renege_frags(struct sctp_ulpq *ulpq, __u16 needed)
{
- __u16 freed = 0;
- __u32 tsn;
- struct sk_buff *skb;
- struct sctp_ulpevent *event;
- struct sctp_tsnmap *tsnmap;
-
- tsnmap = &ulpq->asoc->peer.tsn_map;
-
- /* Walk backwards through the list, reneges the newest tsns. */
- while ((skb = __skb_dequeue_tail(&ulpq->reasm)) != NULL) {
- freed += skb_headlen(skb);
- event = sctp_skb2event(skb);
- tsn = event->tsn;
-
- sctp_ulpevent_free(event);
- sctp_tsnmap_renege(tsnmap, tsn);
- if (freed >= needed)
- return freed;
- }
-
- return freed;
+ return sctp_ulpq_renege_list(ulpq, &ulpq->reasm, needed);
}
/* Partial deliver the first message as there is pressure on rwnd. */
^ permalink raw reply related
* Re: [PATCH] PowerPC: Add BCM5248 and Marvell 88E1111 PHY support to NEW EMAC.
From: Josh Boyer @ 2007-10-23 15:20 UTC (permalink / raw)
To: Jeff Garzik; +Cc: Valentine Barshak, Michael Ellerman, linuxppc-dev, netdev
In-Reply-To: <4713B10B.5000607@garzik.org>
On Mon, 15 Oct 2007 14:27:23 -0400
Jeff Garzik <jeff@garzik.org> wrote:
> Valentine Barshak wrote:
> > This patch adds BCM5248 and Marvell 88E1111 PHY support to NEW EMAC driver.
> > These PHY chips are used on PowerPC 440EPx boards.
> > The PHY code is based on the previous work by Stefan Roese <sr@denx.de>
> >
> > Signed-off-by: Stefan Roese <sr@denx.de>
> > Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
> > ---
> > drivers/net/ibm_newemac/phy.c | 39 +++++++++++++++++++++++++++++++++++++++
> > 1 files changed, 39 insertions(+)
> >
> > --- linux.orig/drivers/net/ibm_newemac/phy.c 2007-06-15 21:45:18.000000000 +0400
> > +++ linux/drivers/net/ibm_newemac/phy.c 2007-06-15 20:45:15.000000000 +0400
> > @@ -306,8 +306,47 @@
> > .ops = &cis8201_phy_ops
> > };
> >
> > +static struct mii_phy_def bcm5248_phy_def = {
> > +
> > + .phy_id = 0x0143bc00,
> > + .phy_id_mask = 0x0ffffff0,
> > + .name = "BCM5248 10/100 SMII Ethernet",
> > + .ops = &generic_phy_ops
> > +};
> > +
> > +static int m88e1111_init(struct mii_phy *phy)
> > +{
> > + printk("%s: Marvell 88E1111 Ethernet\n", __FUNCTION__);
> > + phy_write(phy, 0x14, 0x0ce3);
> > + phy_write(phy, 0x18, 0x4101);
> > + phy_write(phy, 0x09, 0x0e00);
> > + phy_write(phy, 0x04, 0x01e1);
> > + phy_write(phy, 0x00, 0x9140);
> > + phy_write(phy, 0x00, 0x1140);
> > +
> > + return 0;
> > +}
> > +
> > +static struct mii_phy_ops m88e1111_phy_ops = {
> > + .init = m88e1111_init,
> > + .setup_aneg = genmii_setup_aneg,
> > + .setup_forced = genmii_setup_forced,
> > + .poll_link = genmii_poll_link,
> > + .read_link = genmii_read_link
> > +};
> > +
> > +static struct mii_phy_def m88e1111_phy_def = {
> > +
> > + .phy_id = 0x01410CC0,
> > + .phy_id_mask = 0x0ffffff0,
> > + .name = "Marvell 88E1111 Ethernet",
> > + .ops = &m88e1111_phy_ops,
> > +};
> > +
> > static struct mii_phy_def *mii_phy_table[] = {
> > &cis8201_phy_def,
> > + &bcm5248_phy_def,
> > + &m88e1111_phy_def,
> > &genmii_phy_def,
>
> Seems sane to me -- ACK -- but we have multiple people sending me
> patches for a single driver. That's normal for janitorial cleanups
> across the whole tree, but discouraged when multiple people are actively
> working on the same driver.
>
> Please coordinate, and have ONE person send me patches...
Jeff, could you please pull in this patch for 2.6.24? We'll get the
coordination down for any further patches.
thx,
josh
>
^ permalink raw reply
* Re: iproute2: git pull request from debian repo.
From: Andreas Henriksson @ 2007-10-23 15:27 UTC (permalink / raw)
To: Patrick McHardy; +Cc: maximilian attems, shemminger, netdev
In-Reply-To: <471E0E5A.40306@trash.net>
On Tue, Oct 23, 2007 at 05:08:10PM +0200, Patrick McHardy wrote:
> Andreas Henriksson wrote:
> >On Tue, Oct 23, 2007 at 04:38:06PM +0200, Patrick McHardy wrote:
> >>Almost - the default behaviour shouldn't change IMO. Also please
> >>send a complete patch, not one on top of your old one.
> >
> >Which "default behaviour"? ip addr flush or ip neigh flush behaviour?
> >Both of those should behave the same IMHO.
>
> In both cases your changed behaviour should be optional and it should
> default to the old behaviour. Not actually flushing completely when the
> user specified "flush" is counter-intuitive.
>
I did not change the behaviour in both cases! I'm only sending in a change
for ip addr flush to make it match how ip neigh flush already works!
Two different behaviours -> one change -> same behaviour for both.
I've mentioned this multiple times already... I don't know how to express
myself in another/better way about this.
> >Defaulting to never bail out is causing problems in a bunch of firewalling
> >frontends which has been reported multiple times to both debian and ubuntu
> >(and possibly others I don't know about), and looping forever causes the
> >bootup to never finish.
> >Thats my reason for advocating not selecting the "loop forever" as the
> >default choice.
>
> Then those frontends should add "maxrounds N" if they want to abort
> before the flush is complete.
>
So you prefer that I rather change the default for ip neigh flush, then touch
ip addr flush? Making the problematic version the default for both, and
changing every application out there who uses any of these two functions.
Sorry, I'll pass on that one and let someone else do it if it's the preferred
solution.
--
Regards,
Andreas Henriksson
^ permalink raw reply
* Re: [PATCH] fix ACK processing after netlink_dump_start
From: Joseph Fannin @ 2007-10-23 15:36 UTC (permalink / raw)
To: Denis V. Lunev; +Cc: dcbw, tgraf, jfannin, davem, netdev, kuznet
In-Reply-To: <20071023144023.GA15035@iris.sw.ru>
On Tue, Oct 23, 2007 at 06:40:23PM +0400, Denis V. Lunev wrote:
> Revert to original netlink behavior. Do not reply with ACK if the
> netlink dump has bees successfully started.
>
> libnl has been broken by the cd40b7d3983c708aabe3d3008ec64ffce56d33b0
> The following command reproduce the problem:
> /nl-route-get 192.168.1.1
>
> Signed-off-by: Denis V. Lunev <den@openvz.org>
Network Manager works correctly here with this patch applied against
Linus's current -git.
(Sorry for the lack of response before -- I think I forgot to ask to
be CC'd on replies, and found the discussion of this in archives
only this morning.)
--
Joseph Fannin
jfannin@gmail.com
^ permalink raw reply
* Re: [ANNOUNCE] iproute2-2.6.23
From: Stephen Hemminger @ 2007-10-23 15:37 UTC (permalink / raw)
To: Herbert Xu; +Cc: netdev
In-Reply-To: <20071023061207.GA27560@gondor.apana.org.au>
On Tue, 23 Oct 2007 14:12:07 +0800
Herbert Xu <herbert@gondor.apana.org.au> wrote:
> On Wed, Oct 17, 2007 at 09:10:02AM +0800, Herbert Xu wrote:
> >
> > Any chance you could include the NAT patch for the next release?
> >
> > [TC]: Add NAT action
> >
> > This patch adds the parser for the nat action which is used
> > for stateless NAT.
> >
> > Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
>
> Hi Stephen:
>
> Is there a problem with merging this patch?
I was just holding off for a few days to see if there were any bug fixes
to 2.6.23 version that needed to go in first. It is first in queue for 2.6.24
related stuff.
Promise to get it in before 2.6.24-rc1
^ permalink raw reply
* [PATCH RESEND] ip_gre: sendto/recvfrom NBMA address
From: Timo Teräs @ 2007-10-23 15:08 UTC (permalink / raw)
To: Patrick McHardy; +Cc: davem, netdev, kuznet
In-Reply-To: <471E0638.9040405@trash.net>
When GRE tunnel is in NBMA mode, this patch allows an application to use
a PF_PACKET socket to:
- send a packet to specific NBMA address with sendto()
- use recvfrom() to receive packet and check which NBMA address it came from
This is required to implement properly NHRP over GRE tunnel.
Signed-off-by: Timo Teras <timo.teras@iki.fi>
---
> Your mailer mangled tabs, it won't apply like this. Try sending to
> yourself for testing, then repost when it works properly.
Sorry. Tabs were ok, but the initial white space on context lines was somehow
removed. Apparently a bug in my editor where I cut/pasted from.
net/ipv4/ip_gre.c | 12 +++++++++++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index f151900..b1e3816 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -1033,7 +1033,6 @@ static int ipgre_tunnel_change_mtu(struct net_device *dev, int new_mtu)
return 0;
}
-#ifdef CONFIG_NET_IPGRE_BROADCAST
/* Nice toy. Unfortunately, useless in real life :-)
It allows to construct virtual multiprotocol broadcast "LAN"
over the Internet, provided multicast routing is tuned.
@@ -1092,10 +1091,19 @@ static int ipgre_header(struct sk_buff *skb, struct net_device *dev,
return -t->hlen;
}
+static int ipgre_header_parse(const struct sk_buff *skb, unsigned char *haddr)
+{
+ struct iphdr *iph = (struct iphdr*) skb_mac_header(skb);
+ memcpy(haddr, &iph->saddr, 4);
+ return 4;
+}
+
static const struct header_ops ipgre_header_ops = {
.create = ipgre_header,
+ .parse = ipgre_header_parse,
};
+#ifdef CONFIG_NET_IPGRE_BROADCAST
static int ipgre_open(struct net_device *dev)
{
struct ip_tunnel *t = netdev_priv(dev);
@@ -1197,6 +1205,8 @@ static int ipgre_tunnel_init(struct net_device *dev)
dev->stop = ipgre_close;
}
#endif
+ } else {
+ dev->header_ops = &ipgre_header_ops;
}
if (!tdev && tunnel->parms.link)
^ permalink raw reply related
* [PATCH RESEND] ip_gre: sendto/recvfrom NBMA address
From: Timo Teräs @ 2007-10-23 15:57 UTC (permalink / raw)
To: Patrick McHardy; +Cc: davem, netdev, kuznet
In-Reply-To: <471E0638.9040405@trash.net>
When GRE tunnel is in NBMA mode, this patch allows an application to use
a PF_PACKET socket to:
- send a packet to specific NBMA address with sendto()
- use recvfrom() to receive packet and check which NBMA address it came from
This is required to implement properly NHRP over GRE tunnel.
Signed-off-by: Timo Teras <timo.teras@iki.fi>
---
Patrick McHardy wrote:
> Your mailer mangled tabs, it won't apply like this. Try sending to
> yourself for testing, then repost when it works properly.
Tabs were ok, I think. But somehow it seems that cut'n'paste from my
default editor loses the space from context lines. It should be good now.
net/ipv4/ip_gre.c | 12 +++++++++++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index f151900..b1e3816 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -1033,7 +1033,6 @@ static int ipgre_tunnel_change_mtu(struct net_device *dev, int new_mtu)
return 0;
}
-#ifdef CONFIG_NET_IPGRE_BROADCAST
/* Nice toy. Unfortunately, useless in real life :-)
It allows to construct virtual multiprotocol broadcast "LAN"
over the Internet, provided multicast routing is tuned.
@@ -1092,10 +1091,19 @@ static int ipgre_header(struct sk_buff *skb, struct net_device *dev,
return -t->hlen;
}
+static int ipgre_header_parse(const struct sk_buff *skb, unsigned char *haddr)
+{
+ struct iphdr *iph = (struct iphdr*) skb_mac_header(skb);
+ memcpy(haddr, &iph->saddr, 4);
+ return 4;
+}
+
static const struct header_ops ipgre_header_ops = {
.create = ipgre_header,
+ .parse = ipgre_header_parse,
};
+#ifdef CONFIG_NET_IPGRE_BROADCAST
static int ipgre_open(struct net_device *dev)
{
struct ip_tunnel *t = netdev_priv(dev);
@@ -1197,6 +1205,8 @@ static int ipgre_tunnel_init(struct net_device *dev)
dev->stop = ipgre_close;
}
#endif
+ } else {
+ dev->header_ops = &ipgre_header_ops;
}
if (!tdev && tunnel->parms.link)
^ permalink raw reply related
* Re: [ANNOUNCE] iproute2-2.6.23
From: Herbert Xu @ 2007-10-23 15:57 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: netdev
In-Reply-To: <20071023083702.2904c0a0@shemminger-laptop>
On Tue, Oct 23, 2007 at 08:37:02AM -0700, Stephen Hemminger wrote:
>
> I was just holding off for a few days to see if there were any bug fixes
> to 2.6.23 version that needed to go in first. It is first in queue for 2.6.24
> related stuff.
>
> Promise to get it in before 2.6.24-rc1
Thanks Stephen!
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH] Consolidate sctp_ulpq_renege_xxx functions
From: Vlad Yasevich @ 2007-10-23 15:57 UTC (permalink / raw)
To: Pavel Emelyanov; +Cc: David Miller, lksctp-developers, Linux Netdev List, devel
In-Reply-To: <471E1082.8010607@openvz.org>
Pavel Emelyanov wrote:
> Both are equal, except for the list to be traversed.
>
> Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
>
ACK. Good clean-up Pavel.
Thanks
-vlad
^ permalink raw reply
* [0/11] Merge bundle creation and other misc fixes/clean-ups
From: Herbert Xu @ 2007-10-23 16:02 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
Hi Dave:
This series of patches fixes the bug where the bundle creation
code treats IPv6 routes as IPv4 or vice versa.
I think this is suitable for 2.6.25.
They do fix bugs but I don't think any of them are serious enough
for inclusion in 2.6.24.
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* [PATCH 2/11] [IPSEC]: Use dst->header_len when resizing on output
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Use dst->header_len when resizing on output
Currently we use x->props.header_len when resizing on output. However,
if we're resizing at all we might as well go the whole hog and do it
for the whole dst.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
net/xfrm/xfrm_output.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c
index f4bfd6c..58d5a74 100644
--- a/net/xfrm/xfrm_output.c
+++ b/net/xfrm/xfrm_output.c
@@ -19,7 +19,8 @@
static int xfrm_state_check_space(struct xfrm_state *x, struct sk_buff *skb)
{
- int nhead = x->props.header_len + LL_RESERVED_SPACE(skb->dst->dev)
+ struct dst_entry *dst = skb->dst;
+ int nhead = dst->header_len + LL_RESERVED_SPACE(dst->dev)
- skb_headroom(skb);
if (nhead > 0)
^ permalink raw reply related
* [PATCH 1/11] [IPV6]: Only set nfheader_len for top xfrm dst
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPV6]: Only set nfheader_len for top xfrm dst
We only need to set nfheader_len in the top xfrm dst. This is because
we only ever read the nfheader_len from the top xfrm dst.
It is also easier to count nfheader_len as part of header_len which
then lets us remove the ugly wrapper functions for incrementing and
decrementing header lengths in xfrm6_policy.c.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
net/ipv6/ip6_output.c | 3 ++-
net/ipv6/xfrm6_policy.c | 26 ++++----------------------
2 files changed, 6 insertions(+), 23 deletions(-)
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 13565df..b1647f5 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1096,7 +1096,8 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
inet->cork.length = 0;
sk->sk_sndmsg_page = NULL;
sk->sk_sndmsg_off = 0;
- exthdrlen = rt->u.dst.header_len + (opt ? opt->opt_flen : 0);
+ exthdrlen = rt->u.dst.header_len + (opt ? opt->opt_flen : 0) -
+ rt->u.dst.nfheader_len;
length += exthdrlen;
transhdrlen += exthdrlen;
} else {
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 82e27b8..e5170bc 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -102,24 +102,6 @@ __xfrm6_bundle_addr_local(struct xfrm_state *x, struct in6_addr *addr)
(struct in6_addr*)&x->props.saddr;
}
-static inline void
-__xfrm6_bundle_len_inc(int *len, int *nflen, struct xfrm_state *x)
-{
- if (x->type->flags & XFRM_TYPE_NON_FRAGMENT)
- *nflen += x->props.header_len;
- else
- *len += x->props.header_len;
-}
-
-static inline void
-__xfrm6_bundle_len_dec(int *len, int *nflen, struct xfrm_state *x)
-{
- if (x->type->flags & XFRM_TYPE_NON_FRAGMENT)
- *nflen -= x->props.header_len;
- else
- *len -= x->props.header_len;
-}
-
/* Allocate chain of dst_entry's, attach known xfrm's, calculate
* all the metrics... Shortly, bundle a bundle.
*/
@@ -142,7 +124,6 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
int i;
int err = 0;
int header_len = 0;
- int nfheader_len = 0;
int trailer_len = 0;
dst = dst_prev = NULL;
@@ -175,7 +156,9 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst1->next = dst_prev;
dst_prev = dst1;
- __xfrm6_bundle_len_inc(&header_len, &nfheader_len, xfrm[i]);
+ if (xfrm[i]->type->flags & XFRM_TYPE_NON_FRAGMENT)
+ dst->nfheader_len += xfrm[i]->props.header_len;
+ header_len += xfrm[i]->props.header_len;
trailer_len += xfrm[i]->props.trailer_len;
if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
@@ -223,7 +206,6 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->flags |= DST_HOST;
dst_prev->lastuse = jiffies;
dst_prev->header_len = header_len;
- dst_prev->nfheader_len = nfheader_len;
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
@@ -242,7 +224,7 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
x->u.rt6.rt6i_src = rt0->rt6i_src;
x->u.rt6.rt6i_idev = rt0->rt6i_idev;
in6_dev_hold(rt0->rt6i_idev);
- __xfrm6_bundle_len_dec(&header_len, &nfheader_len, x->u.dst.xfrm);
+ header_len -= x->u.dst.xfrm->props.header_len;
trailer_len -= x->u.dst.xfrm->props.trailer_len;
}
^ permalink raw reply related
* [PATCH 3/11] [IPV6]: Move nfheader_len into rt6_info
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPV6]: Move nfheader_len into rt6_info
The dst member nfheader_len is only used by IPv6. It's also currently
creating a rather ugly alignment hole in struct dst. Therefore this patch
moves it from there into struct rt6_info.
It also reorders the fields in rt6_info to minimize holes.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/dst.h | 1 -
include/net/ip6_fib.h | 11 ++++++++---
net/ipv4/xfrm4_policy.c | 1 -
net/ipv6/ip6_output.c | 5 +++--
net/ipv6/xfrm6_policy.c | 3 ++-
5 files changed, 13 insertions(+), 8 deletions(-)
diff --git a/include/net/dst.h b/include/net/dst.h
index e9ff4a4..8f88c52 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -50,7 +50,6 @@ struct dst_entry
unsigned long expires;
unsigned short header_len; /* more space at head required */
- unsigned short nfheader_len; /* more non-fragment space at head required */
unsigned short trailer_len; /* space to reserve at tail */
u32 metrics[RTAX_MAX];
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index 8578213..4cefcff 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -99,16 +99,21 @@ struct rt6_info
u32 rt6i_flags;
u32 rt6i_metric;
atomic_t rt6i_ref;
- struct fib6_table *rt6i_table;
- struct rt6key rt6i_dst;
- struct rt6key rt6i_src;
+ /* more non-fragment space at head required */
+ unsigned short nfheader_len;
u8 rt6i_protocol;
+ struct fib6_table *rt6i_table;
+
+ struct rt6key rt6i_dst;
+
#ifdef CONFIG_XFRM
u32 rt6i_flow_cache_genid;
#endif
+
+ struct rt6key rt6i_src;
};
static inline struct inet6_dev *ip6_dst_idev(struct dst_entry *dst)
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index cc86fb1..5ee3a2f 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -161,7 +161,6 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->flags |= DST_HOST;
dst_prev->lastuse = jiffies;
dst_prev->header_len = header_len;
- dst_prev->nfheader_len = 0;
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index b1647f5..85c5eb8 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1097,7 +1097,7 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
sk->sk_sndmsg_page = NULL;
sk->sk_sndmsg_off = 0;
exthdrlen = rt->u.dst.header_len + (opt ? opt->opt_flen : 0) -
- rt->u.dst.nfheader_len;
+ rt->nfheader_len;
length += exthdrlen;
transhdrlen += exthdrlen;
} else {
@@ -1112,7 +1112,8 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
hh_len = LL_RESERVED_SPACE(rt->u.dst.dev);
- fragheaderlen = sizeof(struct ipv6hdr) + rt->u.dst.nfheader_len + (opt ? opt->opt_nflen : 0);
+ fragheaderlen = sizeof(struct ipv6hdr) + rt->nfheader_len +
+ (opt ? opt->opt_nflen : 0);
maxfraglen = ((mtu - fragheaderlen) & ~7) + fragheaderlen - sizeof(struct frag_hdr);
if (mtu <= sizeof(struct ipv6hdr) + IPV6_MAXPLEN) {
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index e5170bc..9095dfc 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -157,7 +157,8 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev = dst1;
if (xfrm[i]->type->flags & XFRM_TYPE_NON_FRAGMENT)
- dst->nfheader_len += xfrm[i]->props.header_len;
+ ((struct rt6_info *)dst)->nfheader_len +=
+ xfrm[i]->props.header_len;
header_len += xfrm[i]->props.header_len;
trailer_len += xfrm[i]->props.trailer_len;
^ permalink raw reply related
* [PATCH 4/11] [NET]: Eliminate duplicate copies of dst_discard
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[NET]: Eliminate duplicate copies of dst_discard
We have a number of copies of dst_discard scattered around the place which
all do the same thing, namely free a packet on the input or output paths.
This patch deletes all of them except dst_discard and points all the users
to it.
The only non-trivial bit is decnet where it returns an error. However,
conceptually this is identical to the blackhole functions used in IPv4
and IPv6 which do not return errors. So they should either all return
errors or all return zero. For now I've stuck with the majority and
picked zero as the return value.
It doesn't really matter in practice since few if any driver would react
differently depending on a zero return value or NET_RX_DROP.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/dst.h | 1 +
net/core/dst.c | 3 ++-
net/decnet/dn_route.c | 13 +------------
net/ipv4/route.c | 11 +++--------
net/ipv6/exthdrs.c | 13 ++-----------
net/ipv6/route.c | 21 ++++-----------------
6 files changed, 13 insertions(+), 49 deletions(-)
diff --git a/include/net/dst.h b/include/net/dst.h
index 8f88c52..4df103b 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -172,6 +172,7 @@ static inline struct dst_entry *dst_pop(struct dst_entry *dst)
return child;
}
+extern int dst_discard(struct sk_buff *skb);
extern void * dst_alloc(struct dst_ops * ops);
extern void __dst_free(struct dst_entry * dst);
extern struct dst_entry *dst_destroy(struct dst_entry * dst);
diff --git a/net/core/dst.c b/net/core/dst.c
index 16958e6..21b1d0f 100644
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -154,11 +154,12 @@ loop:
#endif
}
-static int dst_discard(struct sk_buff *skb)
+int dst_discard(struct sk_buff *skb)
{
kfree_skb(skb);
return 0;
}
+EXPORT_SYMBOL(dst_discard);
void * dst_alloc(struct dst_ops * ops)
{
diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c
index 97eee5e..e1f6669 100644
--- a/net/decnet/dn_route.c
+++ b/net/decnet/dn_route.c
@@ -769,17 +769,6 @@ drop:
}
/*
- * Drop packet. This is used for endnodes and for
- * when we should not be forwarding packets from
- * this dest.
- */
-static int dn_blackhole(struct sk_buff *skb)
-{
- kfree_skb(skb);
- return NET_RX_DROP;
-}
-
-/*
* Used to catch bugs. This should never normally get
* called.
*/
@@ -1402,7 +1391,7 @@ make_route:
default:
case RTN_UNREACHABLE:
case RTN_BLACKHOLE:
- rt->u.dst.input = dn_blackhole;
+ rt->u.dst.input = dst_discard;
}
rt->rt_flags = flags;
if (rt->u.dst.dev)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 21b12de..d5cbcee 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -92,6 +92,7 @@
#include <linux/jhash.h>
#include <linux/rcupdate.h>
#include <linux/times.h>
+#include <net/dst.h>
#include <net/net_namespace.h>
#include <net/protocol.h>
#include <net/ip.h>
@@ -2362,12 +2363,6 @@ static struct dst_ops ipv4_dst_blackhole_ops = {
};
-static int ipv4_blackhole_output(struct sk_buff *skb)
-{
- kfree_skb(skb);
- return 0;
-}
-
static int ipv4_dst_blackhole(struct rtable **rp, struct flowi *flp, struct sock *sk)
{
struct rtable *ort = *rp;
@@ -2379,8 +2374,8 @@ static int ipv4_dst_blackhole(struct rtable **rp, struct flowi *flp, struct sock
atomic_set(&new->__refcnt, 1);
new->__use = 1;
- new->input = ipv4_blackhole_output;
- new->output = ipv4_blackhole_output;
+ new->input = dst_discard;
+ new->output = dst_discard;
memcpy(new->metrics, ort->u.dst.metrics, RTAX_MAX*sizeof(u32));
new->dev = ort->u.dst.dev;
diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index 1e89efd..cee06b1 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -32,6 +32,7 @@
#include <linux/in6.h>
#include <linux/icmpv6.h>
+#include <net/dst.h>
#include <net/sock.h>
#include <net/snmp.h>
@@ -318,18 +319,8 @@ void __init ipv6_destopt_init(void)
printk(KERN_ERR "ipv6_destopt_init: Could not register protocol\n");
}
-/********************************
- NONE header. No data in packet.
- ********************************/
-
-static int ipv6_nodata_rcv(struct sk_buff *skb)
-{
- kfree_skb(skb);
- return 0;
-}
-
static struct inet6_protocol nodata_protocol = {
- .handler = ipv6_nodata_rcv,
+ .handler = dst_discard,
.flags = INET6_PROTO_NOPOLICY,
};
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 95f8e4a..7db3cdf 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -156,7 +156,6 @@ struct rt6_info ip6_null_entry = {
static int ip6_pkt_prohibit(struct sk_buff *skb);
static int ip6_pkt_prohibit_out(struct sk_buff *skb);
-static int ip6_pkt_blk_hole(struct sk_buff *skb);
struct rt6_info ip6_prohibit_entry = {
.u = {
@@ -185,8 +184,8 @@ struct rt6_info ip6_blk_hole_entry = {
.obsolete = -1,
.error = -EINVAL,
.metrics = { [RTAX_HOPLIMIT - 1] = 255, },
- .input = ip6_pkt_blk_hole,
- .output = ip6_pkt_blk_hole,
+ .input = dst_discard,
+ .output = dst_discard,
.ops = &ip6_dst_ops,
.path = (struct dst_entry*)&ip6_blk_hole_entry,
}
@@ -785,12 +784,6 @@ struct dst_entry * ip6_route_output(struct sock *sk, struct flowi *fl)
EXPORT_SYMBOL(ip6_route_output);
-static int ip6_blackhole_output(struct sk_buff *skb)
-{
- kfree_skb(skb);
- return 0;
-}
-
int ip6_dst_blackhole(struct sock *sk, struct dst_entry **dstp, struct flowi *fl)
{
struct rt6_info *ort = (struct rt6_info *) *dstp;
@@ -803,8 +796,8 @@ int ip6_dst_blackhole(struct sock *sk, struct dst_entry **dstp, struct flowi *fl
atomic_set(&new->__refcnt, 1);
new->__use = 1;
- new->input = ip6_blackhole_output;
- new->output = ip6_blackhole_output;
+ new->input = dst_discard;
+ new->output = dst_discard;
memcpy(new->metrics, ort->u.dst.metrics, RTAX_MAX*sizeof(u32));
new->dev = ort->u.dst.dev;
@@ -1814,12 +1807,6 @@ static int ip6_pkt_prohibit_out(struct sk_buff *skb)
return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_OUTNOROUTES);
}
-static int ip6_pkt_blk_hole(struct sk_buff *skb)
-{
- kfree_skb(skb);
- return 0;
-}
-
#endif
/*
^ permalink raw reply related
* [PATCH 5/11] [NET]: Remove unnecessary inclusion of dst.h
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[NET]: Remove unnecessary inclusion of dst.h
The file net/netevent.h only refers to struct dst_entry * so it doesn't
need to include dst.h. I've replaced it with a forward declaration.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/netevent.h | 2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/net/netevent.h b/include/net/netevent.h
index e5d2162..e82b7ba 100644
--- a/include/net/netevent.h
+++ b/include/net/netevent.h
@@ -12,7 +12,7 @@
*/
#ifdef __KERNEL__
-#include <net/dst.h>
+struct dst_entry;
struct netevent_redirect {
struct dst_entry *old;
^ permalink raw reply related
* [PATCH 7/11] [IPSEC]: Set dst->input to dst_discard
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Set dst->input to dst_discard
The input function should never be invoked on IPsec dst objects. This is
because we don't apply IPsec on input until after we've made the routing
decision.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
net/ipv4/xfrm4_policy.c | 3 ++-
net/ipv6/xfrm6_policy.c | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 7d250a1..c40a71b 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -10,6 +10,7 @@
#include <linux/compiler.h>
#include <linux/inetdevice.h>
+#include <net/dst.h>
#include <net/xfrm.h>
#include <net/ip.h>
@@ -167,7 +168,7 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
- dst_prev->input = rt->u.dst.input;
+ dst_prev->input = dst_discard;
dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
if (rt0->peer)
atomic_inc(&rt0->peer->refcnt);
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 15747f3..a1c6b7c 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -14,6 +14,7 @@
#include <linux/compiler.h>
#include <linux/netdevice.h>
#include <net/addrconf.h>
+#include <net/dst.h>
#include <net/xfrm.h>
#include <net/ip.h>
#include <net/ipv6.h>
@@ -214,7 +215,7 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
- dst_prev->input = rt->u.dst.input;
+ dst_prev->input = dst_discard;
dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
/* Sheit... I remember I did this right. Apparently,
* it was magically lost, so this code needs audit */
^ permalink raw reply related
* [PATCH 6/11] [IPSEC]: Only set neighbour on top xfrm dst
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Only set neighbour on top xfrm dst
The neighbour field is only used by dst_confirm which only ever happens on
the top-most xfrm dst. So it's a waste to duplicate for every other xfrm
dst. This patch moves its setting out of the loop so that only the top one
gets set.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
net/ipv4/xfrm4_policy.c | 5 +++--
net/ipv6/xfrm6_policy.c | 6 ++++--
2 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 5ee3a2f..7d250a1 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -144,6 +144,9 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->child = &rt->u.dst;
dst->path = &rt->u.dst;
+ /* Copy neighbout for reachability confirmation */
+ dst->neighbour = neigh_clone(rt->u.dst.neighbour);
+
*dst_p = dst;
dst = dst_prev;
@@ -164,8 +167,6 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
- /* Copy neighbout for reachability confirmation */
- dst_prev->neighbour = neigh_clone(rt->u.dst.neighbour);
dst_prev->input = rt->u.dst.input;
dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
if (rt0->peer)
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 9095dfc..15747f3 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -188,6 +188,10 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->child = &rt->u.dst;
dst->path = &rt->u.dst;
+
+ /* Copy neighbour for reachability confirmation */
+ dst->neighbour = neigh_clone(rt->u.dst.neighbour);
+
if (rt->rt6i_node)
((struct xfrm_dst *)dst)->path_cookie = rt->rt6i_node->fn_sernum;
@@ -210,8 +214,6 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev->trailer_len = trailer_len;
memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
- /* Copy neighbour for reachability confirmation */
- dst_prev->neighbour = neigh_clone(rt->u.dst.neighbour);
dst_prev->input = rt->u.dst.input;
dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
/* Sheit... I remember I did this right. Apparently,
^ permalink raw reply related
* [PATCH 8/11] [IPSEC]: Make sure idev is consistent with dev in xfrm_dst
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Make sure idev is consistent with dev in xfrm_dst
Previously we took the device from the bottom route and idev from the top
route. This is bad because idev may well point to a different device.
This patch changes it so that we get the idev from the device directly.
It also makes it an error if either dev or idev is NULL. This is consistent
with the rest of the routing code which also treats these cases as errors.
I've removed the err initialisation in xfrm6_policy.c because it achieves
no purpose and hid a bug when an initial version of this patch neglected
to set err to -ENODEV (fortunately the IPv4 version warned about it).
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
net/ipv4/xfrm4_policy.c | 13 +++++++++----
net/ipv6/xfrm6_policy.c | 15 ++++++++++-----
2 files changed, 19 insertions(+), 9 deletions(-)
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index c40a71b..d903c8b 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -153,14 +153,21 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev = *dst_p;
i = 0;
+ err = -ENODEV;
for (; dst_prev != &rt->u.dst; dst_prev = dst_prev->child) {
struct xfrm_dst *x = (struct xfrm_dst*)dst_prev;
x->u.rt.fl = *fl;
dst_prev->xfrm = xfrm[i++];
dst_prev->dev = rt->u.dst.dev;
- if (rt->u.dst.dev)
- dev_hold(rt->u.dst.dev);
+ if (!rt->u.dst.dev)
+ goto error;
+ dev_hold(rt->u.dst.dev);
+
+ x->u.rt.idev = in_dev_get(rt->u.dst.dev);
+ if (!x->u.rt.idev)
+ goto error;
+
dst_prev->obsolete = -1;
dst_prev->flags |= DST_HOST;
dst_prev->lastuse = jiffies;
@@ -181,8 +188,6 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
x->u.rt.rt_dst = rt0->rt_dst;
x->u.rt.rt_gateway = rt0->rt_gateway;
x->u.rt.rt_spec_dst = rt0->rt_spec_dst;
- x->u.rt.idev = rt0->idev;
- in_dev_hold(rt0->idev);
header_len -= x->u.dst.xfrm->props.header_len;
trailer_len -= x->u.dst.xfrm->props.trailer_len;
}
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index a1c6b7c..f04e718 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -123,7 +123,7 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
}
};
int i;
- int err = 0;
+ int err;
int header_len = 0;
int trailer_len = 0;
@@ -201,13 +201,20 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
dst_prev = *dst_p;
i = 0;
+ err = -ENODEV;
for (; dst_prev != &rt->u.dst; dst_prev = dst_prev->child) {
struct xfrm_dst *x = (struct xfrm_dst*)dst_prev;
dst_prev->xfrm = xfrm[i++];
dst_prev->dev = rt->u.dst.dev;
- if (rt->u.dst.dev)
- dev_hold(rt->u.dst.dev);
+ if (!rt->u.dst.dev)
+ goto error;
+ dev_hold(rt->u.dst.dev);
+
+ x->u.rt6.rt6i_idev = in6_dev_get(rt->u.dst.dev);
+ if (!x->u.rt6.rt6i_idev)
+ goto error;
+
dst_prev->obsolete = -1;
dst_prev->flags |= DST_HOST;
dst_prev->lastuse = jiffies;
@@ -226,8 +233,6 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
memcpy(&x->u.rt6.rt6i_gateway, &rt0->rt6i_gateway, sizeof(x->u.rt6.rt6i_gateway));
x->u.rt6.rt6i_dst = rt0->rt6i_dst;
x->u.rt6.rt6i_src = rt0->rt6i_src;
- x->u.rt6.rt6i_idev = rt0->rt6i_idev;
- in6_dev_hold(rt0->rt6i_idev);
header_len -= x->u.dst.xfrm->props.header_len;
trailer_len -= x->u.dst.xfrm->props.trailer_len;
}
^ permalink raw reply related
* [PATCH 9/11] [IPSEC]: Replace x->type->{local,remote}_addr with flags
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Replace x->type->{local,remote}_addr with flags
The functions local_addr and remote_addr are more than what they're needed
for. The same thing can be done easily with flags on the type object.
This patch does that and simplifies the wrapper functions in xfrm6_policy
accordingly.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/xfrm.h | 4 ++--
net/ipv6/mip6.c | 11 ++---------
net/ipv6/xfrm6_policy.c | 20 ++++++++------------
3 files changed, 12 insertions(+), 23 deletions(-)
diff --git a/include/net/xfrm.h b/include/net/xfrm.h
index 688f6f5..9ca4ae5 100644
--- a/include/net/xfrm.h
+++ b/include/net/xfrm.h
@@ -280,6 +280,8 @@ struct xfrm_type
__u8 flags;
#define XFRM_TYPE_NON_FRAGMENT 1
#define XFRM_TYPE_REPLAY_PROT 2
+#define XFRM_TYPE_LOCAL_COADDR 4
+#define XFRM_TYPE_REMOTE_COADDR 8
int (*init_state)(struct xfrm_state *x);
void (*destructor)(struct xfrm_state *);
@@ -287,8 +289,6 @@ struct xfrm_type
int (*output)(struct xfrm_state *, struct sk_buff *pskb);
int (*reject)(struct xfrm_state *, struct sk_buff *, struct flowi *);
int (*hdr_offset)(struct xfrm_state *, struct sk_buff *, u8 **);
- xfrm_address_t *(*local_addr)(struct xfrm_state *, xfrm_address_t *);
- xfrm_address_t *(*remote_addr)(struct xfrm_state *, xfrm_address_t *);
/* Estimate maximal size of result of transformation of a dgram */
u32 (*get_mtu)(struct xfrm_state *, int size);
};
diff --git a/net/ipv6/mip6.c b/net/ipv6/mip6.c
index 7fd841d..edfd9cd 100644
--- a/net/ipv6/mip6.c
+++ b/net/ipv6/mip6.c
@@ -34,11 +34,6 @@
#include <net/xfrm.h>
#include <net/mip6.h>
-static xfrm_address_t *mip6_xfrm_addr(struct xfrm_state *x, xfrm_address_t *addr)
-{
- return x->coaddr;
-}
-
static inline unsigned int calc_padlen(unsigned int len, unsigned int n)
{
return (n - len + 16) & 0x7;
@@ -337,14 +332,13 @@ static struct xfrm_type mip6_destopt_type =
.description = "MIP6DESTOPT",
.owner = THIS_MODULE,
.proto = IPPROTO_DSTOPTS,
- .flags = XFRM_TYPE_NON_FRAGMENT,
+ .flags = XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_LOCAL_COADDR,
.init_state = mip6_destopt_init_state,
.destructor = mip6_destopt_destroy,
.input = mip6_destopt_input,
.output = mip6_destopt_output,
.reject = mip6_destopt_reject,
.hdr_offset = mip6_destopt_offset,
- .local_addr = mip6_xfrm_addr,
};
static int mip6_rthdr_input(struct xfrm_state *x, struct sk_buff *skb)
@@ -467,13 +461,12 @@ static struct xfrm_type mip6_rthdr_type =
.description = "MIP6RT",
.owner = THIS_MODULE,
.proto = IPPROTO_ROUTING,
- .flags = XFRM_TYPE_NON_FRAGMENT,
+ .flags = XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_REMOTE_COADDR,
.init_state = mip6_rthdr_init_state,
.destructor = mip6_rthdr_destroy,
.input = mip6_rthdr_input,
.output = mip6_rthdr_output,
.hdr_offset = mip6_rthdr_offset,
- .remote_addr = mip6_xfrm_addr,
};
static int __init mip6_init(void)
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index f04e718..ed29bef 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -87,20 +87,16 @@ __xfrm6_find_bundle(struct flowi *fl, struct xfrm_policy *policy)
return dst;
}
-static inline struct in6_addr*
-__xfrm6_bundle_addr_remote(struct xfrm_state *x, struct in6_addr *addr)
+static inline xfrm_address_t *__xfrm6_bundle_addr_remote(struct xfrm_state *x)
{
- return (x->type->remote_addr) ?
- (struct in6_addr*)x->type->remote_addr(x, (xfrm_address_t *)addr) :
- (struct in6_addr*)&x->id.daddr;
+ return (x->type->flags & XFRM_TYPE_REMOTE_COADDR) ? x->coaddr :
+ &x->id.daddr;
}
-static inline struct in6_addr*
-__xfrm6_bundle_addr_local(struct xfrm_state *x, struct in6_addr *addr)
+static inline xfrm_address_t *__xfrm6_bundle_addr_local(struct xfrm_state *x)
{
- return (x->type->local_addr) ?
- (struct in6_addr*)x->type->local_addr(x, (xfrm_address_t *)addr) :
- (struct in6_addr*)&x->props.saddr;
+ return (x->type->flags & XFRM_TYPE_LOCAL_COADDR) ? x->coaddr :
+ &x->props.saddr;
}
/* Allocate chain of dst_entry's, attach known xfrm's, calculate
@@ -171,9 +167,9 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
fl_tunnel.fl4_src = xfrm[i]->props.saddr.a4;
break;
case AF_INET6:
- ipv6_addr_copy(&fl_tunnel.fl6_dst, __xfrm6_bundle_addr_remote(xfrm[i], &fl->fl6_dst));
+ ipv6_addr_copy(&fl_tunnel.fl6_dst, (struct in6_addr *)__xfrm6_bundle_addr_remote(xfrm[i]));
- ipv6_addr_copy(&fl_tunnel.fl6_src, __xfrm6_bundle_addr_local(xfrm[i], &fl->fl6_src));
+ ipv6_addr_copy(&fl_tunnel.fl6_src, (struct in6_addr *)__xfrm6_bundle_addr_local(xfrm[i]));
break;
default:
BUG_ON(1);
^ permalink raw reply related
* [PATCH 11/11] [IPSEC]: Merge common code into xfrm_bundle_create
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Merge common code into xfrm_bundle_create
Half of the code in xfrm4_bundle_create and xfrm6_bundle_create are common.
This patch extracts that logic and puts it into xfrm_bundle_create. The
rest of it are then accessed through afinfo.
As a result this fixes the problem with inter-family transforms where we
treat every xfrm dst in the bundle as if it belongs to the top family.
This patch also fixes a long-standing error-path bug where we may free the
xfrm states twice.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/xfrm.h | 11 +-
net/ipv4/xfrm4_policy.c | 134 ++++++-----------------------------
net/ipv6/xfrm6_policy.c | 143 ++++++-------------------------------
net/xfrm/xfrm_policy.c | 183 +++++++++++++++++++++++++++++++++++++++++-------
4 files changed, 215 insertions(+), 256 deletions(-)
diff --git a/include/net/xfrm.h b/include/net/xfrm.h
index 206c6f6..2da5ac5 100644
--- a/include/net/xfrm.h
+++ b/include/net/xfrm.h
@@ -225,6 +225,7 @@ struct km_event
u32 event;
};
+struct net_device;
struct xfrm_type;
struct xfrm_dst;
struct xfrm_policy_afinfo {
@@ -235,13 +236,11 @@ struct xfrm_policy_afinfo {
xfrm_address_t *daddr);
int (*get_saddr)(xfrm_address_t *saddr, xfrm_address_t *daddr);
struct dst_entry *(*find_bundle)(struct flowi *fl, struct xfrm_policy *policy);
- int (*bundle_create)(struct xfrm_policy *policy,
- struct xfrm_state **xfrm,
- int nx,
- struct flowi *fl,
- struct dst_entry **dst_p);
void (*decode_session)(struct sk_buff *skb,
struct flowi *fl);
+ int (*get_tos)(struct flowi *fl);
+ int (*fill_dst)(struct xfrm_dst *xdst,
+ struct net_device *dev);
};
extern int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo);
@@ -1092,7 +1091,6 @@ static inline int xfrm4_udp_encap_rcv(struct sock *sk, struct sk_buff *skb)
}
#endif
-extern struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos);
struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp);
extern int xfrm_policy_walk(u8 type, int (*func)(struct xfrm_policy *, int, int, void*), void *);
int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl);
@@ -1111,7 +1109,6 @@ extern int xfrm_policy_flush(u8 type, struct xfrm_audit *audit_info);
extern int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol);
extern int xfrm_bundle_ok(struct xfrm_policy *pol, struct xfrm_dst *xdst,
struct flowi *fl, int family, int strict);
-extern void xfrm_init_pmtu(struct dst_entry *dst);
#ifdef CONFIG_XFRM_MIGRATE
extern int km_migrate(struct xfrm_selector *sel, u8 dir, u8 type,
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index cebc847..1d75243 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -79,122 +79,39 @@ __xfrm4_find_bundle(struct flowi *fl, struct xfrm_policy *policy)
return dst;
}
-/* Allocate chain of dst_entry's, attach known xfrm's, calculate
- * all the metrics... Shortly, bundle a bundle.
- */
-
-static int
-__xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int nx,
- struct flowi *fl, struct dst_entry **dst_p)
+static int xfrm4_get_tos(struct flowi *fl)
{
- struct dst_entry *dst, *dst_prev;
- struct rtable *rt0 = (struct rtable*)(*dst_p);
- struct rtable *rt = rt0;
- int tos = fl->fl4_tos;
- int i;
- int err;
- int header_len = 0;
- int trailer_len = 0;
-
- dst = dst_prev = NULL;
- dst_hold(&rt->u.dst);
-
- for (i = 0; i < nx; i++) {
- struct dst_entry *dst1 = dst_alloc(&xfrm4_dst_ops);
- struct xfrm_dst *xdst;
-
- if (unlikely(dst1 == NULL)) {
- err = -ENOBUFS;
- dst_release(&rt->u.dst);
- goto error;
- }
+ return fl->fl4_tos;
+}
- if (!dst)
- dst = dst1;
- else {
- dst_prev->child = dst1;
- dst1->flags |= DST_NOHASH;
- dst_clone(dst1);
- }
+static int xfrm4_fill_dst(struct xfrm_dst *xdst, struct net_device *dev)
+{
+ struct rtable *rt = (struct rtable *)xdst->route;
- xdst = (struct xfrm_dst *)dst1;
- xdst->route = &rt->u.dst;
- xdst->genid = xfrm[i]->genid;
+ xdst->u.rt.fl = rt->fl;
- dst1->next = dst_prev;
- dst_prev = dst1;
+ xdst->u.dst.dev = dev;
+ dev_hold(dev);
- header_len += xfrm[i]->props.header_len;
- trailer_len += xfrm[i]->props.trailer_len;
+ xdst->u.rt.idev = in_dev_get(dev);
+ if (!xdst->u.rt.idev)
+ return -ENODEV;
- if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
- dst1 = xfrm_dst_lookup(xfrm[i], tos);
- err = PTR_ERR(dst1);
- if (IS_ERR(dst1))
- goto error;
+ xdst->u.rt.peer = rt->peer;
+ if (rt->peer)
+ atomic_inc(&rt->peer->refcnt);
- rt = (struct rtable *)dst1;
- } else
- dst_hold(&rt->u.dst);
- }
+ /* Sheit... I remember I did this right. Apparently,
+ * it was magically lost, so this code needs audit */
+ xdst->u.rt.rt_flags = rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST |
+ RTCF_LOCAL);
+ xdst->u.rt.rt_type = rt->rt_type;
+ xdst->u.rt.rt_src = rt->rt_src;
+ xdst->u.rt.rt_dst = rt->rt_dst;
+ xdst->u.rt.rt_gateway = rt->rt_gateway;
+ xdst->u.rt.rt_spec_dst = rt->rt_spec_dst;
- dst_prev->child = &rt->u.dst;
- dst->path = &rt->u.dst;
-
- /* Copy neighbout for reachability confirmation */
- dst->neighbour = neigh_clone(rt->u.dst.neighbour);
-
- *dst_p = dst;
- dst = dst_prev;
-
- dst_prev = *dst_p;
- i = 0;
- err = -ENODEV;
- for (; dst_prev != &rt->u.dst; dst_prev = dst_prev->child) {
- struct xfrm_dst *x = (struct xfrm_dst*)dst_prev;
- x->u.rt.fl = *fl;
-
- dst_prev->xfrm = xfrm[i++];
- dst_prev->dev = rt->u.dst.dev;
- if (!rt->u.dst.dev)
- goto error;
- dev_hold(rt->u.dst.dev);
-
- x->u.rt.idev = in_dev_get(rt->u.dst.dev);
- if (!x->u.rt.idev)
- goto error;
-
- dst_prev->obsolete = -1;
- dst_prev->flags |= DST_HOST;
- dst_prev->lastuse = jiffies;
- dst_prev->header_len = header_len;
- dst_prev->trailer_len = trailer_len;
- memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
-
- dst_prev->input = dst_discard;
- dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
- if (rt0->peer)
- atomic_inc(&rt0->peer->refcnt);
- x->u.rt.peer = rt0->peer;
- /* Sheit... I remember I did this right. Apparently,
- * it was magically lost, so this code needs audit */
- x->u.rt.rt_flags = rt0->rt_flags&(RTCF_BROADCAST|RTCF_MULTICAST|RTCF_LOCAL);
- x->u.rt.rt_type = rt0->rt_type;
- x->u.rt.rt_src = rt0->rt_src;
- x->u.rt.rt_dst = rt0->rt_dst;
- x->u.rt.rt_gateway = rt0->rt_gateway;
- x->u.rt.rt_spec_dst = rt0->rt_spec_dst;
- header_len -= x->u.dst.xfrm->props.header_len;
- trailer_len -= x->u.dst.xfrm->props.trailer_len;
- }
-
- xfrm_init_pmtu(dst);
return 0;
-
-error:
- if (dst)
- dst_free(dst);
- return err;
}
static void
@@ -330,8 +247,9 @@ static struct xfrm_policy_afinfo xfrm4_policy_afinfo = {
.dst_lookup = xfrm4_dst_lookup,
.get_saddr = xfrm4_get_saddr,
.find_bundle = __xfrm4_find_bundle,
- .bundle_create = __xfrm4_bundle_create,
.decode_session = _decode_session4,
+ .get_tos = xfrm4_get_tos,
+ .fill_dst = xfrm4_fill_dst,
};
static void __init xfrm4_policy_init(void)
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 864258f..50a3841 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -93,126 +93,34 @@ __xfrm6_find_bundle(struct flowi *fl, struct xfrm_policy *policy)
return dst;
}
-/* Allocate chain of dst_entry's, attach known xfrm's, calculate
- * all the metrics... Shortly, bundle a bundle.
- */
-
-static int
-__xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int nx,
- struct flowi *fl, struct dst_entry **dst_p)
+static int xfrm6_get_tos(struct flowi *fl)
{
- struct dst_entry *dst, *dst_prev;
- struct rt6_info *rt0 = (struct rt6_info*)(*dst_p);
- struct rt6_info *rt = rt0;
- int i;
- int err;
- int header_len = 0;
- int trailer_len = 0;
-
- dst = dst_prev = NULL;
- dst_hold(&rt->u.dst);
-
- for (i = 0; i < nx; i++) {
- struct dst_entry *dst1 = dst_alloc(&xfrm6_dst_ops);
- struct xfrm_dst *xdst;
-
- if (unlikely(dst1 == NULL)) {
- err = -ENOBUFS;
- dst_release(&rt->u.dst);
- goto error;
- }
-
- if (!dst)
- dst = dst1;
- else {
- dst_prev->child = dst1;
- dst1->flags |= DST_NOHASH;
- dst_clone(dst1);
- }
-
- xdst = (struct xfrm_dst *)dst1;
- xdst->route = &rt->u.dst;
- xdst->genid = xfrm[i]->genid;
- if (rt->rt6i_node)
- xdst->route_cookie = rt->rt6i_node->fn_sernum;
-
- dst1->next = dst_prev;
- dst_prev = dst1;
-
- if (xfrm[i]->type->flags & XFRM_TYPE_NON_FRAGMENT)
- ((struct rt6_info *)dst)->nfheader_len +=
- xfrm[i]->props.header_len;
- header_len += xfrm[i]->props.header_len;
- trailer_len += xfrm[i]->props.trailer_len;
-
- if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
- dst1 = xfrm_dst_lookup(xfrm[i], 0);
- err = PTR_ERR(dst1);
- if (IS_ERR(dst1))
- goto error;
-
- rt = (struct rt6_info *)dst1;
- } else
- dst_hold(&rt->u.dst);
- }
+ return 0;
+}
- dst_prev->child = &rt->u.dst;
- dst->path = &rt->u.dst;
-
- /* Copy neighbour for reachability confirmation */
- dst->neighbour = neigh_clone(rt->u.dst.neighbour);
-
- if (rt->rt6i_node)
- ((struct xfrm_dst *)dst)->path_cookie = rt->rt6i_node->fn_sernum;
-
- *dst_p = dst;
- dst = dst_prev;
-
- dst_prev = *dst_p;
- i = 0;
- err = -ENODEV;
- for (; dst_prev != &rt->u.dst; dst_prev = dst_prev->child) {
- struct xfrm_dst *x = (struct xfrm_dst*)dst_prev;
-
- dst_prev->xfrm = xfrm[i++];
- dst_prev->dev = rt->u.dst.dev;
- if (!rt->u.dst.dev)
- goto error;
- dev_hold(rt->u.dst.dev);
-
- x->u.rt6.rt6i_idev = in6_dev_get(rt->u.dst.dev);
- if (!x->u.rt6.rt6i_idev)
- goto error;
-
- dst_prev->obsolete = -1;
- dst_prev->flags |= DST_HOST;
- dst_prev->lastuse = jiffies;
- dst_prev->header_len = header_len;
- dst_prev->trailer_len = trailer_len;
- memcpy(&dst_prev->metrics, &x->route->metrics, sizeof(dst_prev->metrics));
-
- dst_prev->input = dst_discard;
- dst_prev->output = dst_prev->xfrm->outer_mode->afinfo->output;
- /* Sheit... I remember I did this right. Apparently,
- * it was magically lost, so this code needs audit */
- x->u.rt6.rt6i_flags = rt0->rt6i_flags&(RTCF_BROADCAST|RTCF_MULTICAST|RTCF_LOCAL);
- x->u.rt6.rt6i_metric = rt0->rt6i_metric;
- x->u.rt6.rt6i_node = rt0->rt6i_node;
- x->u.rt6.rt6i_gateway = rt0->rt6i_gateway;
- memcpy(&x->u.rt6.rt6i_gateway, &rt0->rt6i_gateway, sizeof(x->u.rt6.rt6i_gateway));
- x->u.rt6.rt6i_dst = rt0->rt6i_dst;
- x->u.rt6.rt6i_src = rt0->rt6i_src;
- header_len -= x->u.dst.xfrm->props.header_len;
- trailer_len -= x->u.dst.xfrm->props.trailer_len;
- }
+static int xfrm6_fill_dst(struct xfrm_dst *xdst, struct net_device *dev)
+{
+ struct rt6_info *rt = (struct rt6_info*)xdst->route;
+
+ xdst->u.dst.dev = dev;
+ dev_hold(dev);
+
+ xdst->u.rt6.rt6i_idev = in6_dev_get(rt->u.dst.dev);
+ if (!xdst->u.rt6.rt6i_idev)
+ return -ENODEV;
+
+ /* Sheit... I remember I did this right. Apparently,
+ * it was magically lost, so this code needs audit */
+ xdst->u.rt6.rt6i_flags = rt->rt6i_flags & (RTCF_BROADCAST |
+ RTCF_MULTICAST |
+ RTCF_LOCAL);
+ xdst->u.rt6.rt6i_metric = rt->rt6i_metric;
+ xdst->u.rt6.rt6i_node = rt->rt6i_node;
+ xdst->u.rt6.rt6i_gateway = rt->rt6i_gateway;
+ xdst->u.rt6.rt6i_dst = rt->rt6i_dst;
+ xdst->u.rt6.rt6i_src = rt->rt6i_src;
- xfrm_init_pmtu(dst);
return 0;
-
-error:
- if (dst)
- dst_free(dst);
- return err;
}
static inline void
@@ -355,8 +263,9 @@ static struct xfrm_policy_afinfo xfrm6_policy_afinfo = {
.dst_lookup = xfrm6_dst_lookup,
.get_saddr = xfrm6_get_saddr,
.find_bundle = __xfrm6_find_bundle,
- .bundle_create = __xfrm6_bundle_create,
.decode_session = _decode_session6,
+ .get_tos = xfrm6_get_tos,
+ .fill_dst = xfrm6_fill_dst,
};
static void __init xfrm6_policy_init(void)
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 6168341..5a03612 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -24,6 +24,7 @@
#include <linux/netfilter.h>
#include <linux/module.h>
#include <linux/cache.h>
+#include <net/dst.h>
#include <net/xfrm.h>
#include <net/ip.h>
@@ -50,6 +51,7 @@ static DEFINE_SPINLOCK(xfrm_policy_gc_lock);
static struct xfrm_policy_afinfo *xfrm_policy_get_afinfo(unsigned short family);
static void xfrm_policy_put_afinfo(struct xfrm_policy_afinfo *afinfo);
+static void xfrm_init_pmtu(struct dst_entry *dst);
static inline int
__xfrm4_selector_match(struct xfrm_selector *sel, struct flowi *fl)
@@ -85,7 +87,8 @@ int xfrm_selector_match(struct xfrm_selector *sel, struct flowi *fl,
return 0;
}
-struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos)
+static inline struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos,
+ int family)
{
xfrm_address_t *saddr = &x->props.saddr;
xfrm_address_t *daddr = &x->id.daddr;
@@ -97,7 +100,7 @@ struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos)
if (x->type->flags & XFRM_TYPE_REMOTE_COADDR)
daddr = x->coaddr;
- afinfo = xfrm_policy_get_afinfo(x->props.family);
+ afinfo = xfrm_policy_get_afinfo(family);
if (unlikely(afinfo == NULL))
return ERR_PTR(-EAFNOSUPPORT);
@@ -105,7 +108,6 @@ struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos)
xfrm_policy_put_afinfo(afinfo);
return dst;
}
-EXPORT_SYMBOL(xfrm_dst_lookup);
static inline unsigned long make_jiffies(long secs)
{
@@ -1235,24 +1237,164 @@ xfrm_find_bundle(struct flowi *fl, struct xfrm_policy *policy, unsigned short fa
return x;
}
-/* Allocate chain of dst_entry's, attach known xfrm's, calculate
- * all the metrics... Shortly, bundle a bundle.
- */
+static inline int xfrm_get_tos(struct flowi *fl, int family)
+{
+ struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
+ int tos;
-static int
-xfrm_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int nx,
- struct flowi *fl, struct dst_entry **dst_p,
- unsigned short family)
+ if (!afinfo)
+ return -EINVAL;
+
+ tos = afinfo->get_tos(fl);
+
+ xfrm_policy_put_afinfo(afinfo);
+
+ return tos;
+}
+
+static inline struct xfrm_dst *xfrm_alloc_dst(int family)
{
- int err;
struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
- if (unlikely(afinfo == NULL))
+ struct xfrm_dst *xdst;
+
+ if (!afinfo)
+ return ERR_PTR(-EINVAL);
+
+ xdst = dst_alloc(afinfo->dst_ops) ?: ERR_PTR(-ENOBUFS);
+
+ xfrm_policy_put_afinfo(afinfo);
+
+ return xdst;
+}
+
+static inline int xfrm_fill_dst(struct xfrm_dst *xdst, struct net_device *dev)
+{
+ struct xfrm_policy_afinfo *afinfo =
+ xfrm_policy_get_afinfo(xdst->u.dst.ops->family);
+ int err;
+
+ if (!afinfo)
return -EINVAL;
- err = afinfo->bundle_create(policy, xfrm, nx, fl, dst_p);
+
+ err = afinfo->fill_dst(xdst, dev);
+
xfrm_policy_put_afinfo(afinfo);
+
return err;
}
+/* Allocate chain of dst_entry's, attach known xfrm's, calculate
+ * all the metrics... Shortly, bundle a bundle.
+ */
+
+static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy,
+ struct xfrm_state **xfrm, int nx,
+ struct flowi *fl,
+ struct dst_entry *dst)
+{
+ unsigned long now = jiffies;
+ struct net_device *dev;
+ struct dst_entry *dst_prev = NULL;
+ struct dst_entry *dst0 = NULL;
+ int i = 0;
+ int err;
+ int header_len = 0;
+ int trailer_len = 0;
+ int tos;
+ int family = policy->selector.family;
+
+ tos = xfrm_get_tos(fl, family);
+ err = tos;
+ if (tos < 0)
+ goto put_states;
+
+ dst_hold(dst);
+
+ for (; i < nx; i++) {
+ struct xfrm_dst *xdst = xfrm_alloc_dst(family);
+ struct dst_entry *dst1 = &xdst->u.dst;
+
+ err = PTR_ERR(xdst);
+ if (IS_ERR(xdst)) {
+ dst_release(dst);
+ goto put_states;
+ }
+
+ if (!dst_prev)
+ dst0 = dst1;
+ else {
+ dst_prev->child = dst_clone(dst1);
+ dst1->flags |= DST_NOHASH;
+ }
+
+ xdst->route = dst;
+ memcpy(&dst1->metrics, &dst->metrics, sizeof(dst->metrics));
+
+ if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
+ family = xfrm[i]->props.family;
+ dst = xfrm_dst_lookup(xfrm[i], tos, family);
+ err = PTR_ERR(dst);
+ if (IS_ERR(dst))
+ goto put_states;
+ } else
+ dst_hold(dst);
+
+ dst1->xfrm = xfrm[i];
+ xdst->genid = xfrm[i]->genid;
+
+ dst1->obsolete = -1;
+ dst1->flags |= DST_HOST;
+ dst1->lastuse = now;
+
+ dst1->input = dst_discard;
+ dst1->output = xfrm[i]->outer_mode->afinfo->output;
+
+ dst1->next = dst_prev;
+ dst_prev = dst1;
+
+ header_len += xfrm[i]->props.header_len;
+ trailer_len += xfrm[i]->props.trailer_len;
+ }
+
+ dst_prev->child = dst;
+ dst0->path = dst;
+
+ err = -ENODEV;
+ dev = dst->dev;
+ if (!dev)
+ goto free_dst;
+
+ /* Copy neighbout for reachability confirmation */
+ dst0->neighbour = neigh_clone(dst->neighbour);
+
+ xfrm_init_pmtu(dst_prev);
+
+ for (dst_prev = dst0; dst_prev != dst; dst_prev = dst_prev->child) {
+ struct xfrm_dst *xdst = (struct xfrm_dst *)dst_prev;
+
+ err = xfrm_fill_dst(xdst, dev);
+ if (err)
+ goto free_dst;
+
+ dst_prev->header_len = header_len;
+ dst_prev->trailer_len = trailer_len;
+ header_len -= xdst->u.dst.xfrm->props.header_len;
+ trailer_len -= xdst->u.dst.xfrm->props.trailer_len;
+ }
+
+out:
+ return dst0;
+
+put_states:
+ for (; i < nx; i++)
+ xfrm_state_put(xfrm[i]);
+free_dst:
+ if (dst0)
+ dst_free(dst0);
+ dst0 = ERR_PTR(err);
+ goto out;
+}
+
static int inline
xfrm_dst_alloc_copy(void **target, void *src, int size)
{
@@ -1452,15 +1594,10 @@ restart:
return 0;
}
- dst = dst_orig;
- err = xfrm_bundle_create(policy, xfrm, nx, fl, &dst, family);
-
- if (unlikely(err)) {
- int i;
- for (i=0; i<nx; i++)
- xfrm_state_put(xfrm[i]);
+ dst = xfrm_bundle_create(policy, xfrm, nx, fl, dst_orig);
+ err = PTR_ERR(dst);
+ if (IS_ERR(dst))
goto error;
- }
for (pi = 0; pi < npols; pi++) {
read_lock_bh(&pols[pi]->lock);
@@ -1883,7 +2020,7 @@ static int xfrm_flush_bundles(void)
return 0;
}
-void xfrm_init_pmtu(struct dst_entry *dst)
+static void xfrm_init_pmtu(struct dst_entry *dst)
{
do {
struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
@@ -1904,8 +2041,6 @@ void xfrm_init_pmtu(struct dst_entry *dst)
} while ((dst = dst->next));
}
-EXPORT_SYMBOL(xfrm_init_pmtu);
-
/* Check that the bundle accepts the flow and its components are
* still valid.
*/
^ permalink raw reply related
* [PATCH 10/11] [IPSEC]: Move flow construction into xfrm_dst_lookup
From: Herbert Xu @ 2007-10-23 16:03 UTC (permalink / raw)
To: David S. Miller, YOSHIFUJI Hideaki, netdev
In-Reply-To: <20071023160217.GA32102@gondor.apana.org.au>
[IPSEC]: Move flow construction into xfrm_dst_lookup
This patch moves the flow construction from the callers of xfrm_dst_lookup
into that function. It also changes xfrm_dst_lookup so that it takes an
xfrm state as its argument instead of explicit addresses.
This removes any address-specific logic from the callers of xfrm_dst_lookup
which is needed to correctly support inter-family transforms.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
---
include/net/xfrm.h | 10 +---
net/ipv4/xfrm4_policy.c | 80 ++++++++++++++++++---------------------
net/ipv6/xfrm6_policy.c | 97 +++++++++++++++++-------------------------------
net/xfrm/xfrm_policy.c | 25 +++++++-----
4 files changed, 91 insertions(+), 121 deletions(-)
diff --git a/include/net/xfrm.h b/include/net/xfrm.h
index 9ca4ae5..206c6f6 100644
--- a/include/net/xfrm.h
+++ b/include/net/xfrm.h
@@ -231,7 +231,8 @@ struct xfrm_policy_afinfo {
unsigned short family;
struct dst_ops *dst_ops;
void (*garbage_collect)(void);
- int (*dst_lookup)(struct xfrm_dst **dst, struct flowi *fl);
+ struct dst_entry *(*dst_lookup)(int tos, xfrm_address_t *saddr,
+ xfrm_address_t *daddr);
int (*get_saddr)(xfrm_address_t *saddr, xfrm_address_t *daddr);
struct dst_entry *(*find_bundle)(struct flowi *fl, struct xfrm_policy *policy);
int (*bundle_create)(struct xfrm_policy *policy,
@@ -1077,7 +1078,6 @@ extern int xfrm6_find_1stfragopt(struct xfrm_state *x, struct sk_buff *skb,
#ifdef CONFIG_XFRM
extern int xfrm4_udp_encap_rcv(struct sock *sk, struct sk_buff *skb);
extern int xfrm_user_policy(struct sock *sk, int optname, u8 __user *optval, int optlen);
-extern int xfrm_dst_lookup(struct xfrm_dst **dst, struct flowi *fl, unsigned short family);
#else
static inline int xfrm_user_policy(struct sock *sk, int optname, u8 __user *optval, int optlen)
{
@@ -1090,13 +1090,9 @@ static inline int xfrm4_udp_encap_rcv(struct sock *sk, struct sk_buff *skb)
kfree_skb(skb);
return 0;
}
-
-static inline int xfrm_dst_lookup(struct xfrm_dst **dst, struct flowi *fl, unsigned short family)
-{
- return -EINVAL;
-}
#endif
+extern struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos);
struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp);
extern int xfrm_policy_walk(u8 type, int (*func)(struct xfrm_policy *, int, int, void*), void *);
int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl);
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index d903c8b..cebc847 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -8,7 +8,8 @@
*
*/
-#include <linux/compiler.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
#include <linux/inetdevice.h>
#include <net/dst.h>
#include <net/xfrm.h>
@@ -17,28 +18,44 @@
static struct dst_ops xfrm4_dst_ops;
static struct xfrm_policy_afinfo xfrm4_policy_afinfo;
-static int xfrm4_dst_lookup(struct xfrm_dst **dst, struct flowi *fl)
+static struct dst_entry *xfrm4_dst_lookup(int tos, xfrm_address_t *saddr,
+ xfrm_address_t *daddr)
{
- return __ip_route_output_key((struct rtable**)dst, fl);
-}
-
-static int xfrm4_get_saddr(xfrm_address_t *saddr, xfrm_address_t *daddr)
-{
- struct rtable *rt;
- struct flowi fl_tunnel = {
+ struct flowi fl = {
.nl_u = {
.ip4_u = {
+ .tos = tos,
.daddr = daddr->a4,
},
},
};
+ struct dst_entry *dst;
+ struct rtable *rt;
+ int err;
- if (!xfrm4_dst_lookup((struct xfrm_dst **)&rt, &fl_tunnel)) {
- saddr->a4 = rt->rt_src;
- dst_release(&rt->u.dst);
- return 0;
- }
- return -EHOSTUNREACH;
+ if (saddr)
+ fl.fl4_src = saddr->a4;
+
+ err = __ip_route_output_key(&rt, &fl);
+ dst = &rt->u.dst;
+ if (err)
+ dst = ERR_PTR(err);
+ return dst;
+}
+
+static int xfrm4_get_saddr(xfrm_address_t *saddr, xfrm_address_t *daddr)
+{
+ struct dst_entry *dst;
+ struct rtable *rt;
+
+ dst = xfrm4_dst_lookup(0, NULL, daddr);
+ if (IS_ERR(dst))
+ return -EHOSTUNREACH;
+
+ rt = (struct rtable *)dst;
+ saddr->a4 = rt->rt_src;
+ dst_release(dst);
+ return 0;
}
static struct dst_entry *
@@ -73,15 +90,7 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
struct dst_entry *dst, *dst_prev;
struct rtable *rt0 = (struct rtable*)(*dst_p);
struct rtable *rt = rt0;
- struct flowi fl_tunnel = {
- .nl_u = {
- .ip4_u = {
- .saddr = fl->fl4_src,
- .daddr = fl->fl4_dst,
- .tos = fl->fl4_tos
- }
- }
- };
+ int tos = fl->fl4_tos;
int i;
int err;
int header_len = 0;
@@ -119,25 +128,12 @@ __xfrm4_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
trailer_len += xfrm[i]->props.trailer_len;
if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
- unsigned short encap_family = xfrm[i]->props.family;
- switch (encap_family) {
- case AF_INET:
- fl_tunnel.fl4_dst = xfrm[i]->id.daddr.a4;
- fl_tunnel.fl4_src = xfrm[i]->props.saddr.a4;
- break;
-#if defined(CONFIG_IPV6) || defined (CONFIG_IPV6_MODULE)
- case AF_INET6:
- ipv6_addr_copy(&fl_tunnel.fl6_dst, (struct in6_addr*)&xfrm[i]->id.daddr.a6);
- ipv6_addr_copy(&fl_tunnel.fl6_src, (struct in6_addr*)&xfrm[i]->props.saddr.a6);
- break;
-#endif
- default:
- BUG_ON(1);
- }
- err = xfrm_dst_lookup((struct xfrm_dst **)&rt,
- &fl_tunnel, encap_family);
- if (err)
+ dst1 = xfrm_dst_lookup(xfrm[i], tos);
+ err = PTR_ERR(dst1);
+ if (IS_ERR(dst1))
goto error;
+
+ rt = (struct rtable *)dst1;
} else
dst_hold(&rt->u.dst);
}
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index ed29bef..864258f 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -11,7 +11,8 @@
*
*/
-#include <linux/compiler.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <net/addrconf.h>
#include <net/dst.h>
@@ -26,35 +27,40 @@
static struct dst_ops xfrm6_dst_ops;
static struct xfrm_policy_afinfo xfrm6_policy_afinfo;
-static int xfrm6_dst_lookup(struct xfrm_dst **xdst, struct flowi *fl)
+static struct dst_entry *xfrm6_dst_lookup(int tos, xfrm_address_t *saddr,
+ xfrm_address_t *daddr)
{
- struct dst_entry *dst = ip6_route_output(NULL, fl);
- int err = dst->error;
- if (!err)
- *xdst = (struct xfrm_dst *) dst;
- else
+ struct flowi fl = {};
+ struct dst_entry *dst;
+ int err;
+
+ memcpy(&fl.fl6_dst, daddr, sizeof(fl.fl6_dst));
+ if (saddr)
+ memcpy(&fl.fl6_src, saddr, sizeof(fl.fl6_src));
+
+ dst = ip6_route_output(NULL, &fl);
+
+ err = dst->error;
+ if (dst->error) {
dst_release(dst);
- return err;
+ dst = ERR_PTR(err);
+ }
+
+ return dst;
}
static int xfrm6_get_saddr(xfrm_address_t *saddr, xfrm_address_t *daddr)
{
- struct rt6_info *rt;
- struct flowi fl_tunnel = {
- .nl_u = {
- .ip6_u = {
- .daddr = *(struct in6_addr *)&daddr->a6,
- },
- },
- };
-
- if (!xfrm6_dst_lookup((struct xfrm_dst **)&rt, &fl_tunnel)) {
- ipv6_get_saddr(&rt->u.dst, (struct in6_addr *)&daddr->a6,
- (struct in6_addr *)&saddr->a6);
- dst_release(&rt->u.dst);
- return 0;
- }
- return -EHOSTUNREACH;
+ struct dst_entry *dst;
+
+ dst = xfrm6_dst_lookup(0, NULL, daddr);
+ if (IS_ERR(dst))
+ return -EHOSTUNREACH;
+
+ ipv6_get_saddr(dst, (struct in6_addr *)&daddr->a6,
+ (struct in6_addr *)&saddr->a6);
+ dst_release(dst);
+ return 0;
}
static struct dst_entry *
@@ -87,18 +93,6 @@ __xfrm6_find_bundle(struct flowi *fl, struct xfrm_policy *policy)
return dst;
}
-static inline xfrm_address_t *__xfrm6_bundle_addr_remote(struct xfrm_state *x)
-{
- return (x->type->flags & XFRM_TYPE_REMOTE_COADDR) ? x->coaddr :
- &x->id.daddr;
-}
-
-static inline xfrm_address_t *__xfrm6_bundle_addr_local(struct xfrm_state *x)
-{
- return (x->type->flags & XFRM_TYPE_LOCAL_COADDR) ? x->coaddr :
- &x->props.saddr;
-}
-
/* Allocate chain of dst_entry's, attach known xfrm's, calculate
* all the metrics... Shortly, bundle a bundle.
*/
@@ -110,14 +104,6 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
struct dst_entry *dst, *dst_prev;
struct rt6_info *rt0 = (struct rt6_info*)(*dst_p);
struct rt6_info *rt = rt0;
- struct flowi fl_tunnel = {
- .nl_u = {
- .ip6_u = {
- .saddr = fl->fl6_src,
- .daddr = fl->fl6_dst,
- }
- }
- };
int i;
int err;
int header_len = 0;
@@ -160,25 +146,12 @@ __xfrm6_bundle_create(struct xfrm_policy *policy, struct xfrm_state **xfrm, int
trailer_len += xfrm[i]->props.trailer_len;
if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) {
- unsigned short encap_family = xfrm[i]->props.family;
- switch(encap_family) {
- case AF_INET:
- fl_tunnel.fl4_dst = xfrm[i]->id.daddr.a4;
- fl_tunnel.fl4_src = xfrm[i]->props.saddr.a4;
- break;
- case AF_INET6:
- ipv6_addr_copy(&fl_tunnel.fl6_dst, (struct in6_addr *)__xfrm6_bundle_addr_remote(xfrm[i]));
-
- ipv6_addr_copy(&fl_tunnel.fl6_src, (struct in6_addr *)__xfrm6_bundle_addr_local(xfrm[i]));
- break;
- default:
- BUG_ON(1);
- }
-
- err = xfrm_dst_lookup((struct xfrm_dst **) &rt,
- &fl_tunnel, encap_family);
- if (err)
+ dst1 = xfrm_dst_lookup(xfrm[i], 0);
+ err = PTR_ERR(dst1);
+ if (IS_ERR(dst1))
goto error;
+
+ rt = (struct rt6_info *)dst1;
} else
dst_hold(&rt->u.dst);
}
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index b702bd8..6168341 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -13,6 +13,7 @@
*
*/
+#include <linux/err.h>
#include <linux/slab.h>
#include <linux/kmod.h>
#include <linux/list.h>
@@ -84,21 +85,25 @@ int xfrm_selector_match(struct xfrm_selector *sel, struct flowi *fl,
return 0;
}
-int xfrm_dst_lookup(struct xfrm_dst **dst, struct flowi *fl,
- unsigned short family)
+struct dst_entry *xfrm_dst_lookup(struct xfrm_state *x, int tos)
{
- struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
- int err = 0;
+ xfrm_address_t *saddr = &x->props.saddr;
+ xfrm_address_t *daddr = &x->id.daddr;
+ struct xfrm_policy_afinfo *afinfo;
+ struct dst_entry *dst;
+ if (x->type->flags & XFRM_TYPE_LOCAL_COADDR)
+ saddr = x->coaddr;
+ if (x->type->flags & XFRM_TYPE_REMOTE_COADDR)
+ daddr = x->coaddr;
+
+ afinfo = xfrm_policy_get_afinfo(x->props.family);
if (unlikely(afinfo == NULL))
- return -EAFNOSUPPORT;
+ return ERR_PTR(-EAFNOSUPPORT);
- if (likely(afinfo->dst_lookup != NULL))
- err = afinfo->dst_lookup(dst, fl);
- else
- err = -EINVAL;
+ dst = afinfo->dst_lookup(tos, saddr, daddr);
xfrm_policy_put_afinfo(afinfo);
- return err;
+ return dst;
}
EXPORT_SYMBOL(xfrm_dst_lookup);
^ permalink raw reply related
* Re: [PATCH RESEND] ip_gre: sendto/recvfrom NBMA address
From: Patrick McHardy @ 2007-10-23 16:10 UTC (permalink / raw)
To: Timo Teräs; +Cc: davem, netdev, kuznet
In-Reply-To: <471E19E5.5040509@iki.fi>
Timo Teräs wrote:
> When GRE tunnel is in NBMA mode, this patch allows an application to use
> a PF_PACKET socket to:
> - send a packet to specific NBMA address with sendto()
> - use recvfrom() to receive packet and check which NBMA address it
> came from
>
> This is required to implement properly NHRP over GRE tunnel.
>
> Signed-off-by: Timo Teras <timo.teras@iki.fi>
>
> ---
> Patrick McHardy wrote:
>> Your mailer mangled tabs, it won't apply like this. Try sending to
>> yourself for testing, then repost when it works properly.
>
> Tabs were ok, I think. But somehow it seems that cut'n'paste from my
> default editor loses the space from context lines. It should be good now.
Might have been a problem on my side, I just updated my mailer
and it still displays tabs broken, but the actual patch seems
fine. Sorry for the noise ...
^ permalink raw reply
* Re: [PATCH] PowerPC: Add BCM5248 and Marvell 88E1111 PHY support to NEW EMAC.
From: Kumar Gala @ 2007-10-23 16:13 UTC (permalink / raw)
To: Josh Boyer
Cc: Jeff Garzik, Valentine Barshak, Michael Ellerman, linuxppc-dev,
netdev
In-Reply-To: <20071023102035.6d35d3fe@weaponx.rchland.ibm.com>
On Oct 23, 2007, at 10:20 AM, Josh Boyer wrote:
> On Mon, 15 Oct 2007 14:27:23 -0400
> Jeff Garzik <jeff@garzik.org> wrote:
>
>> Valentine Barshak wrote:
>>> This patch adds BCM5248 and Marvell 88E1111 PHY support to NEW
>>> EMAC driver.
>>> These PHY chips are used on PowerPC 440EPx boards.
>>> The PHY code is based on the previous work by Stefan Roese
>>> <sr@denx.de>
>>>
>>> Signed-off-by: Stefan Roese <sr@denx.de>
>>> Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
>>> ---
You guys should really look at moving emac over to the phylib so we
don't have to duplicate drivers for the same phys all over the place :)
- k
^ permalink raw reply
* Re: [PATCH] PowerPC: Add BCM5248 and Marvell 88E1111 PHY support to NEW EMAC.
From: Josh Boyer @ 2007-10-23 16:17 UTC (permalink / raw)
To: Kumar Gala
Cc: Jeff Garzik, Valentine Barshak, Michael Ellerman, linuxppc-dev,
netdev
In-Reply-To: <38505C5B-4252-4AA7-B4F5-6B5267A6DBAA@kernel.crashing.org>
On Tue, 23 Oct 2007 11:13:48 -0500
Kumar Gala <galak@kernel.crashing.org> wrote:
>
> On Oct 23, 2007, at 10:20 AM, Josh Boyer wrote:
>
> > On Mon, 15 Oct 2007 14:27:23 -0400
> > Jeff Garzik <jeff@garzik.org> wrote:
> >
> >> Valentine Barshak wrote:
> >>> This patch adds BCM5248 and Marvell 88E1111 PHY support to NEW
> >>> EMAC driver.
> >>> These PHY chips are used on PowerPC 440EPx boards.
> >>> The PHY code is based on the previous work by Stefan Roese
> >>> <sr@denx.de>
> >>>
> >>> Signed-off-by: Stefan Roese <sr@denx.de>
> >>> Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
> >>> ---
>
> You guys should really look at moving emac over to the phylib so we
> don't have to duplicate drivers for the same phys all over the place :)
Yes, we should. It's on the list. Just not for 2.6.24 since it's way
too late.
josh
^ 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