mir/ovs
2
0
Fork 0
mirror of https://github.com/openvswitch/ovs synced 2025-08-24 02:47:14 +00:00
Code Issues Releases Wiki Activity
ovs/lib/odp-util.c

1055 lines
34 KiB
C
Raw Normal View History

Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
/*
odp-util: Improve formatting of bogus empty action attributes. Before this change, these were formatted as: ,***%u leftover bytes*** After this change, they are formatted as: <empty>,***%u leftover bytes*** Reviewed by Ethan Jackson <ethan@nicira.com>.
2011-01-11 15:35:56 -08:00
* Copyright (c) 2009, 2010, 2011 Nicira Networks.
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
*
Update primary code license to Apache 2.0.
2009-06-15 15:11:30 -07:00
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
*
Update primary code license to Apache 2.0.
2009-06-15 15:11:30 -07:00
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
*/
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
#include <arpa/inet.h>
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
#include <config.h>
#include "odp-util.h"
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
#include <errno.h>
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
#include <inttypes.h>
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
#include <netinet/icmp6.h>
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
#include <stdlib.h>
#include <string.h>
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
#include "byte-order.h"
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
#include "coverage.h"
#include "dynamic-string.h"
#include "flow.h"
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
#include "netlink.h"
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
#include "ofpbuf.h"
ovs-dpctl: Print extended information about vports. When "ovs-dpctl show" is run, return additional information about the port. For example, tunnel ports will print the remote_ip, local_ip, and in_key when defined.
2010-12-18 01:07:06 -08:00
#include "openvswitch/tunnel.h"
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
#include "packets.h"
#include "timeval.h"
#include "util.h"
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
/* The interface between userspace and kernel uses an "OVS_*" prefix.
* Since this is fairly non-specific for the OVS userspace components,
* "ODP_*" (Open vSwitch Datapath) is used as the prefix for
* interactions with the datapath.
*/
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
int
odp_action_len(uint16_t type)
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
if (type > OVS_ACTION_ATTR_MAX) {
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
return -1;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
switch ((enum ovs_action_type) type) {
case OVS_ACTION_ATTR_OUTPUT: return 4;
case OVS_ACTION_ATTR_USERSPACE: return 8;
case OVS_ACTION_ATTR_SET_DL_TCI: return 2;
case OVS_ACTION_ATTR_STRIP_VLAN: return 0;
case OVS_ACTION_ATTR_SET_DL_SRC: return ETH_ADDR_LEN;
case OVS_ACTION_ATTR_SET_DL_DST: return ETH_ADDR_LEN;
case OVS_ACTION_ATTR_SET_NW_SRC: return 4;
case OVS_ACTION_ATTR_SET_NW_DST: return 4;
case OVS_ACTION_ATTR_SET_NW_TOS: return 1;
case OVS_ACTION_ATTR_SET_TP_SRC: return 2;
case OVS_ACTION_ATTR_SET_TP_DST: return 2;
case OVS_ACTION_ATTR_SET_TUNNEL: return 8;
case OVS_ACTION_ATTR_SET_PRIORITY: return 4;
case OVS_ACTION_ATTR_POP_PRIORITY: return 0;
case OVS_ACTION_ATTR_UNSPEC:
case __OVS_ACTION_ATTR_MAX:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
return -1;
}
return -1;
}
static void
format_generic_odp_action(struct ds *ds, const struct nlattr *a)
{
odp-util: Improve formatting of bad ODP actions. nla_len is the size of a Netlink attribute including its header, but we only want to print the attribute payload, so use nl_attr_get_size() instead. Also, make it clear that what follows the "bad action" information is a dump of the action itself. Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-23 14:21:01 -08:00
size_t len = nl_attr_get_size(a);
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "action%"PRId16, nl_attr_type(a));
odp-util: Improve formatting of bad ODP actions. nla_len is the size of a Netlink attribute including its header, but we only want to print the attribute payload, so use nl_attr_get_size() instead. Also, make it clear that what follows the "bad action" information is a dump of the action itself. Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-23 14:21:01 -08:00
if (len) {
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
const uint8_t *unspec;
unsigned int i;
unspec = nl_attr_get(a);
odp-util: Improve formatting of bad ODP actions. nla_len is the size of a Netlink attribute including its header, but we only want to print the attribute payload, so use nl_attr_get_size() instead. Also, make it clear that what follows the "bad action" information is a dump of the action itself. Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-23 14:21:01 -08:00
for (i = 0; i < len; i++) {
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_char(ds, i ? ' ': '(');
ds_put_format(ds, "%02x", unspec[i]);
}
ds_put_char(ds, ')');
}
}
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
void
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
format_odp_action(struct ds *ds, const struct nlattr *a)
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
{
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
const uint8_t *eth;
ovs_be32 ip;
odp-util: Correct length check in format_odp_action(). When printing the action list we first check that the size of the action matches the expected length for that type. However, when doing the lookup we were passing in the length of the action, not the type, leading to bogus values.
2010-12-16 14:27:47 -08:00
if (nl_attr_get_size(a) != odp_action_len(nl_attr_type(a))) {
odp-util: Improve formatting of bad ODP actions. nla_len is the size of a Netlink attribute including its header, but we only want to print the attribute payload, so use nl_attr_get_size() instead. Also, make it clear that what follows the "bad action" information is a dump of the action itself. Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-23 14:21:01 -08:00
ds_put_format(ds, "bad length %zu, expected %d for: ",
odp-util: Correct length check in format_odp_action(). When printing the action list we first check that the size of the action matches the expected length for that type. However, when doing the lookup we were passing in the length of the action, not the type, leading to bogus values.
2010-12-16 14:27:47 -08:00
nl_attr_get_size(a), odp_action_len(nl_attr_type(a)));
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
format_generic_odp_action(ds, a);
return;
}
switch (nl_attr_type(a)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_OUTPUT:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "%"PRIu16, nl_attr_get_u32(a));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_USERSPACE:
Datapath action should not refer to controller ODP_ACTION_ATTR_CONTROLLER in the kernel actually sends packets to userspace, not the controller. To make it generic rename this action to ODP_ACTION_ATTR_USERSPACE. Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
2011-07-28 09:05:25 -07:00
ds_put_format(ds, "userspace(%"PRIu64")", nl_attr_get_u64(a));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_TUNNEL:
Expand tunnel IDs from 32 to 64 bits. We have a need to identify tunnels with keys longer than 32 bits. This commit adds basic datapath and OpenFlow support for such keys. It doesn't actually add any tunnel protocols that support 64-bit keys, so this is not very useful yet. The 'arg' member of struct odp_msg had to be expanded to 64-bits also, because it sometimes contains a tunnel ID. This member also contains the argument passed to ODPAT_CONTROLLER, so I expanded that action's argument to 64 bits also so that it can use the full width of the expanded 'arg'. Userspace doesn't take advantage of the new space though (it was only using 16 bits anyhow). This commit has been tested only to the extent that it doesn't disrupt basic Open vSwitch operation. I have not tested it with tunnel traffic. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com> Feature #3976.
2010-12-10 10:42:42 -08:00
ds_put_format(ds, "set_tunnel(%#"PRIx64")",
ntohll(nl_attr_get_be64(a)));
tunneling: Add support for tunnel ID. Add a tun_id field which contains the ID of the encapsulating tunnel on which a packet was received (0 if not received on a tunnel). Also add an action which allows the tunnel ID to be set for outgoing packets. At this point there aren't any tunnel implementations so these fields don't have any effect. The matching is exposed to OpenFlow by overloading the high 32 bits of the cookie as the tunnel ID. ovs-ofctl is capable of turning on this special behavior using a new "tun-cookie" command but this command is intentially undocumented to avoid it being used without a full understanding of the consequences.
2010-04-12 11:49:16 -04:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_DL_TCI:
datapath: Simplify ODPAT_SET_DL_TCI action. There's no need to have a mask in this action, because both parts of the TCI are part of the flow structure. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-10-18 11:18:10 -07:00
ds_put_format(ds, "set_tci(vid=%"PRIu16",pcp=%d)",
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
vlan_tci_to_vid(nl_attr_get_be16(a)),
vlan_tci_to_pcp(nl_attr_get_be16(a)));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_STRIP_VLAN:
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
ds_put_format(ds, "strip_vlan");
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_DL_SRC:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
eth = nl_attr_get_unspec(a, ETH_ADDR_LEN);
ds_put_format(ds, "set_dl_src("ETH_ADDR_FMT")", ETH_ADDR_ARGS(eth));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_DL_DST:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
eth = nl_attr_get_unspec(a, ETH_ADDR_LEN);
ds_put_format(ds, "set_dl_dst("ETH_ADDR_FMT")", ETH_ADDR_ARGS(eth));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_NW_SRC:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ip = nl_attr_get_be32(a);
ds_put_format(ds, "set_nw_src("IP_FMT")", IP_ARGS(&ip));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_NW_DST:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ip = nl_attr_get_be32(a);
ds_put_format(ds, "set_nw_dst("IP_FMT")", IP_ARGS(&ip));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_NW_TOS:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "set_nw_tos(%"PRIu8")", nl_attr_get_u8(a));
ofproto: Match VLAN PCP and rewrite ToS bits (OpenFlow 0.9) Starting in OpenFlow 0.9, it is possible to match on the VLAN PCP (priority) field and rewrite the IP ToS/DSCP bits. This check-in provides that support and bumps the wire protocol number to 0x98. NOTE: The wire changes come together over the set of OpenFlow 0.9 commits, so OVS will not be OpenFlow-compatible with any official release between this commit and the one that completes the set.
2009-11-11 14:59:49 -08:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_TP_SRC:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "set_tp_src(%"PRIu16")", ntohs(nl_attr_get_be16(a)));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_TP_DST:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "set_tp_dst(%"PRIu16")", ntohs(nl_attr_get_be16(a)));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_SET_PRIORITY:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
ds_put_format(ds, "set_priority(%#"PRIx32")", nl_attr_get_u32(a));
Add pretty-printers for QoS related OpenFlow and ODP actions.
2010-07-16 15:30:09 -07:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_ACTION_ATTR_POP_PRIORITY:
Add pretty-printers for QoS related OpenFlow and ODP actions.
2010-07-16 15:30:09 -07:00
ds_put_cstr(ds, "pop_priority");
Properly print drop_spoofed_arp actions when decoding OpenFlow and ODP. Also fix formatting of unknown Nicira actions in OpenFlow.
2010-09-15 13:26:08 -07:00
break;
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
default:
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
format_generic_odp_action(ds, a);
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
}
}
void
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
format_odp_actions(struct ds *ds, const struct nlattr *actions,
vswitchd: Consistently use size_t for action lengths. Currently the type of the datapath action length is mixture of size_t and unsigned int. However, size_t is really defined as an unsigned long, which causes the build to fail on 64-bit platforms. This consistently uses size_t.
2010-12-11 22:51:31 -08:00
size_t actions_len)
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
{
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
if (actions_len) {
const struct nlattr *a;
unsigned int left;
NL_ATTR_FOR_EACH (a, left, actions, actions_len) {
if (a != actions) {
ds_put_char(ds, ',');
}
format_odp_action(ds, a);
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
}
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
if (left) {
odp-util: Improve formatting of bogus empty action attributes. Before this change, these were formatted as: ,***%u leftover bytes*** After this change, they are formatted as: <empty>,***%u leftover bytes*** Reviewed by Ethan Jackson <ethan@nicira.com>.
2011-01-11 15:35:56 -08:00
if (left == actions_len) {
ds_put_cstr(ds, "<empty>");
}
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
ds_put_format(ds, ",***%u leftover bytes***", left);
datapath: Replace "struct odp_action" by Netlink attributes. In the medium term, we plan to migrate the datapath to use Netlink as its communication channel. In the short term, we need to be able to have actions with 64-bit arguments but "struct odp_action" only has room for 48 bits. So this patch shifts to variable-length arguments using Netlink attributes, which starts in on the Netlink transition and makes 64-bit arguments possible at the same time. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2010-12-10 10:40:58 -08:00
}
} else {
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
ds_put_cstr(ds, "drop");
}
}
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
/* Returns the correct length of the payload for a flow key attribute of the
* specified 'type', or -1 if 'type' is unknown. */
static int
odp_flow_key_attr_len(uint16_t type)
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
if (type > OVS_KEY_ATTR_MAX) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return -1;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
switch ((enum ovs_key_type) type) {
case OVS_KEY_ATTR_TUN_ID: return 8;
case OVS_KEY_ATTR_IN_PORT: return 4;
case OVS_KEY_ATTR_ETHERNET: return sizeof(struct ovs_key_ethernet);
case OVS_KEY_ATTR_8021Q: return sizeof(struct ovs_key_8021q);
case OVS_KEY_ATTR_ETHERTYPE: return 2;
case OVS_KEY_ATTR_IPV4: return sizeof(struct ovs_key_ipv4);
case OVS_KEY_ATTR_IPV6: return sizeof(struct ovs_key_ipv6);
case OVS_KEY_ATTR_TCP: return sizeof(struct ovs_key_tcp);
case OVS_KEY_ATTR_UDP: return sizeof(struct ovs_key_udp);
case OVS_KEY_ATTR_ICMP: return sizeof(struct ovs_key_icmp);
case OVS_KEY_ATTR_ICMPV6: return sizeof(struct ovs_key_icmpv6);
case OVS_KEY_ATTR_ARP: return sizeof(struct ovs_key_arp);
case OVS_KEY_ATTR_ND: return sizeof(struct ovs_key_nd);
case OVS_KEY_ATTR_UNSPEC:
case __OVS_KEY_ATTR_MAX:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return -1;
}
return -1;
}
static void
format_generic_odp_key(const struct nlattr *a, struct ds *ds)
{
size_t len = nl_attr_get_size(a);
ds_put_format(ds, "key%"PRId16, nl_attr_type(a));
if (len) {
const uint8_t *unspec;
unsigned int i;
unspec = nl_attr_get(a);
for (i = 0; i < len; i++) {
ds_put_char(ds, i ? ' ': '(');
ds_put_format(ds, "%02x", unspec[i]);
}
ds_put_char(ds, ')');
}
}
static void
format_odp_key_attr(const struct nlattr *a, struct ds *ds)
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
const struct ovs_key_ethernet *eth_key;
const struct ovs_key_8021q *q_key;
const struct ovs_key_ipv4 *ipv4_key;
const struct ovs_key_ipv6 *ipv6_key;
const struct ovs_key_tcp *tcp_key;
const struct ovs_key_udp *udp_key;
const struct ovs_key_icmp *icmp_key;
const struct ovs_key_icmpv6 *icmpv6_key;
const struct ovs_key_arp *arp_key;
const struct ovs_key_nd *nd_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (nl_attr_get_size(a) != odp_flow_key_attr_len(nl_attr_type(a))) {
ds_put_format(ds, "bad length %zu, expected %d for: ",
nl_attr_get_size(a),
odp_flow_key_attr_len(nl_attr_type(a)));
format_generic_odp_key(a, ds);
return;
}
switch (nl_attr_type(a)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_TUN_ID:
odp-util: Print tunnel ids in host-order. Suggested-by: Pankaj Thakkar <thakkar@nicira.com>
2011-03-22 16:45:03 -07:00
ds_put_format(ds, "tun_id(%#"PRIx64")", ntohll(nl_attr_get_be64(a)));
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_IN_PORT:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
ds_put_format(ds, "in_port(%"PRIu32")", nl_attr_get_u32(a));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ETHERNET:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
eth_key = nl_attr_get(a);
ds_put_format(ds, "eth(src="ETH_ADDR_FMT",dst="ETH_ADDR_FMT")",
ETH_ADDR_ARGS(eth_key->eth_src),
ETH_ADDR_ARGS(eth_key->eth_dst));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_8021Q:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
q_key = nl_attr_get(a);
ds_put_cstr(ds, "vlan(");
if (q_key->q_tpid != htons(ETH_TYPE_VLAN)) {
odp-util: Print EtherTypes consistently. The '#' format specifier doesn't respect the field width modifier, so EtherTypes are printed with variable length. Zero is not a valid EtherType so there isn't a need for the logic to dynamically insert the 0x prefix (if the EtherType isn't specified it won't be printed at all). This fixes the EtherType to have the intended format and also changes the vlan TPID to match.
2011-01-29 23:41:27 -08:00
ds_put_format(ds, "tpid=0x%04"PRIx16",", ntohs(q_key->q_tpid));
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
}
odp-util: Format VLAN headers more like other headers in ODP flow output. The rest of the headers all follow the form "header(value)" or "header(key1=value1,key2=value2,...)" but VLAN headers left out the "=" characters. This adds them in for consistency.
2011-08-04 13:32:19 -07:00
ds_put_format(ds, "vid=%"PRIu16",pcp=%d)",
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
vlan_tci_to_vid(q_key->q_tci),
vlan_tci_to_pcp(q_key->q_tci));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ETHERTYPE:
odp-util: Print EtherTypes consistently. The '#' format specifier doesn't respect the field width modifier, so EtherTypes are printed with variable length. Zero is not a valid EtherType so there isn't a need for the logic to dynamically insert the 0x prefix (if the EtherType isn't specified it won't be printed at all). This fixes the EtherType to have the intended format and also changes the vlan TPID to match.
2011-01-29 23:41:27 -08:00
ds_put_format(ds, "eth_type(0x%04"PRIx16")",
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
ntohs(nl_attr_get_be16(a)));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_IPV4:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
ipv4_key = nl_attr_get(a);
ds_put_format(ds, "ipv4(src="IP_FMT",dst="IP_FMT","
"proto=%"PRId8",tos=%"PRIu8")",
IP_ARGS(&ipv4_key->ipv4_src),
IP_ARGS(&ipv4_key->ipv4_dst),
ipv4_key->ipv4_proto, ipv4_key->ipv4_tos);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_IPV6: {
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
char src_str[INET6_ADDRSTRLEN];
char dst_str[INET6_ADDRSTRLEN];
ipv6_key = nl_attr_get(a);
inet_ntop(AF_INET6, ipv6_key->ipv6_src, src_str, sizeof src_str);
inet_ntop(AF_INET6, ipv6_key->ipv6_dst, dst_str, sizeof dst_str);
ds_put_format(ds, "ipv6(src=%s,dst=%s,proto=%"PRId8",tos=%"PRIu8")",
src_str, dst_str, ipv6_key->ipv6_proto,
ipv6_key->ipv6_tos);
break;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_TCP:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
tcp_key = nl_attr_get(a);
ds_put_format(ds, "tcp(src=%"PRIu16",dst=%"PRIu16")",
ntohs(tcp_key->tcp_src), ntohs(tcp_key->tcp_dst));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_UDP:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
udp_key = nl_attr_get(a);
ds_put_format(ds, "udp(src=%"PRIu16",dst=%"PRIu16")",
ntohs(udp_key->udp_src), ntohs(udp_key->udp_dst));
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ICMP:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
icmp_key = nl_attr_get(a);
ds_put_format(ds, "icmp(type=%"PRIu8",code=%"PRIu8")",
icmp_key->icmp_type, icmp_key->icmp_code);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ICMPV6:
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
icmpv6_key = nl_attr_get(a);
ds_put_format(ds, "icmpv6(type=%"PRIu8",code=%"PRIu8")",
icmpv6_key->icmpv6_type, icmpv6_key->icmpv6_code);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ARP:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
arp_key = nl_attr_get(a);
nicira-ext: Support matching ARP source and target hardware addresses. OpenFlow 1.0 doesn't allow matching on the ARP source and target hardware address. This has caused us to introduce hacks such as the Drop Spoofed ARP action. Now that we have extensible match, we can match on more fields within ARP: - Source Hardware Address (arp_sha) - Target Hardware Address (arp_tha) Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-07 14:02:17 -08:00
ds_put_format(ds, "arp(sip="IP_FMT",tip="IP_FMT",op=%"PRIu16","
"sha="ETH_ADDR_FMT",tha="ETH_ADDR_FMT")",
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
IP_ARGS(&arp_key->arp_sip), IP_ARGS(&arp_key->arp_tip),
nicira-ext: Support matching ARP source and target hardware addresses. OpenFlow 1.0 doesn't allow matching on the ARP source and target hardware address. This has caused us to introduce hacks such as the Drop Spoofed ARP action. Now that we have extensible match, we can match on more fields within ARP: - Source Hardware Address (arp_sha) - Target Hardware Address (arp_tha) Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-07 14:02:17 -08:00
ntohs(arp_key->arp_op), ETH_ADDR_ARGS(arp_key->arp_sha),
ETH_ADDR_ARGS(arp_key->arp_tha));
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ND: {
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
char target[INET6_ADDRSTRLEN];
nd_key = nl_attr_get(a);
inet_ntop(AF_INET6, nd_key->nd_target, target, sizeof target);
ds_put_format(ds, "nd(target=%s", target);
if (!eth_addr_is_zero(nd_key->nd_sll)) {
ds_put_format(ds, ",sll="ETH_ADDR_FMT,
ETH_ADDR_ARGS(nd_key->nd_sll));
}
if (!eth_addr_is_zero(nd_key->nd_tll)) {
ds_put_format(ds, ",tll="ETH_ADDR_FMT,
ETH_ADDR_ARGS(nd_key->nd_tll));
}
ds_put_char(ds, ')');
break;
}
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
default:
format_generic_odp_key(a, ds);
break;
}
}
/* Appends to 'ds' a string representation of the 'key_len' bytes of
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
* OVS_KEY_ATTR_* attributes in 'key'. */
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
void
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
odp_flow_key_format(const struct nlattr *key, size_t key_len, struct ds *ds)
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
{
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (key_len) {
const struct nlattr *a;
unsigned int left;
NL_ATTR_FOR_EACH (a, left, key, key_len) {
if (a != key) {
ds_put_char(ds, ',');
}
format_odp_key_attr(a, ds);
}
if (left) {
if (left == key_len) {
ds_put_cstr(ds, "<empty>");
}
ds_put_format(ds, ",***%u leftover bytes***", left);
}
} else {
ds_put_cstr(ds, "<empty>");
}
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
}
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
static int
put_nd_key(int n, const char *nd_target_s,
const uint8_t *nd_sll, const uint8_t *nd_tll, struct ofpbuf *key)
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_nd nd_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
memset(&nd_key, 0, sizeof nd_key);
if (inet_pton(AF_INET6, nd_target_s, nd_key.nd_target) != 1) {
return -EINVAL;
}
if (nd_sll) {
memcpy(nd_key.nd_sll, nd_sll, ETH_ADDR_LEN);
}
if (nd_tll) {
memcpy(nd_key.nd_tll, nd_tll, ETH_ADDR_LEN);
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_ND, &nd_key, sizeof nd_key);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
static int
parse_odp_key_attr(const char *s, struct ofpbuf *key)
{
/* Many of the sscanf calls in this function use oversized destination
* fields because some sscanf() implementations truncate the range of %i
* directives, so that e.g. "%"SCNi16 interprets input of "0xfedc" as a
* value of 0x7fff. The other alternatives are to allow only a single
* radix (e.g. decimal or hexadecimal) or to write more sophisticated
* parsers.
*
* The tun_id parser has to use an alternative approach because there is no
* type larger than 64 bits. */
{
char tun_id_s[32];
int n = -1;
if (sscanf(s, "tun_id(%31[x0123456789abcdefABCDEF])%n",
tun_id_s, &n) > 0 && n > 0) {
uint64_t tun_id = strtoull(tun_id_s, NULL, 0);
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_be64(key, OVS_KEY_ATTR_TUN_ID, htonll(tun_id));
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
unsigned long long int in_port;
int n = -1;
if (sscanf(s, "in_port(%lli)%n", &in_port, &n) > 0 && n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_u32(key, OVS_KEY_ATTR_IN_PORT, in_port);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ethernet eth_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
int n = -1;
if (sscanf(s,
"eth(src="ETH_ADDR_SCAN_FMT",dst="ETH_ADDR_SCAN_FMT")%n",
ETH_ADDR_SCAN_ARGS(eth_key.eth_src),
ETH_ADDR_SCAN_ARGS(eth_key.eth_dst), &n) > 0 && n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_ETHERNET,
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
&eth_key, sizeof eth_key);
return n;
}
}
{
uint16_t tpid = ETH_TYPE_VLAN;
uint16_t vid;
int pcp;
int n = -1;
if ((sscanf(s, "vlan(vid=%"SCNi16",pcp=%i)%n",
&vid, &pcp, &n) > 0 && n > 0) ||
(sscanf(s, "vlan(tpid=%"SCNi16",vid=%"SCNi16",pcp=%i)%n",
&tpid, &vid, &pcp, &n) > 0 && n > 0)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_8021q q_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
q_key.q_tpid = htons(tpid);
q_key.q_tci = htons((vid << VLAN_VID_SHIFT) |
(pcp << VLAN_PCP_SHIFT));
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_8021Q, &q_key, sizeof q_key);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
odp-util: Fix parsing of Ethertypes 0x8000 and above. An existing comment in the function being updated explains the problem: * Many of the sscanf calls in this function use oversized destination * fields because some sscanf() implementations truncate the range of %i * directives, so that e.g. "%"SCNi16 interprets input of "0xfedc" as a * value of 0x7fff. The other alternatives are to allow only a single * radix (e.g. decimal or hexadecimal) or to write more sophisticated * parsers.
2011-08-18 10:33:32 -07:00
int eth_type;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
int n = -1;
odp-util: Fix parsing of Ethertypes 0x8000 and above. An existing comment in the function being updated explains the problem: * Many of the sscanf calls in this function use oversized destination * fields because some sscanf() implementations truncate the range of %i * directives, so that e.g. "%"SCNi16 interprets input of "0xfedc" as a * value of 0x7fff. The other alternatives are to allow only a single * radix (e.g. decimal or hexadecimal) or to write more sophisticated * parsers.
2011-08-18 10:33:32 -07:00
if (sscanf(s, "eth_type(%i)%n", &eth_type, &n) > 0 && n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_be16(key, OVS_KEY_ATTR_ETHERTYPE, htons(eth_type));
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
ovs_be32 ipv4_src;
ovs_be32 ipv4_dst;
int ipv4_proto;
int ipv4_tos;
int n = -1;
if (sscanf(s, "ipv4(src="IP_SCAN_FMT",dst="IP_SCAN_FMT","
"proto=%i,tos=%i)%n",
IP_SCAN_ARGS(&ipv4_src),
IP_SCAN_ARGS(&ipv4_dst), &ipv4_proto, &ipv4_tos, &n) > 0
&& n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ipv4 ipv4_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
memset(&ipv4_key, 0, sizeof ipv4_key);
ipv4_key.ipv4_src = ipv4_src;
ipv4_key.ipv4_dst = ipv4_dst;
ipv4_key.ipv4_proto = ipv4_proto;
ipv4_key.ipv4_tos = ipv4_tos;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_IPV4,
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
&ipv4_key, sizeof ipv4_key);
return n;
}
}
{
char ipv6_src_s[IPV6_SCAN_LEN + 1];
char ipv6_dst_s[IPV6_SCAN_LEN + 1];
int ipv6_proto;
int ipv6_tos;
int n = -1;
if (sscanf(s, "ipv6(src="IPV6_SCAN_FMT",dst="IPV6_SCAN_FMT","
"proto=%i,tos=%i)%n",
ipv6_src_s, ipv6_dst_s,
&ipv6_proto, &ipv6_tos, &n) > 0 && n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ipv6 ipv6_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
memset(&ipv6_key, 0, sizeof ipv6_key);
if (inet_pton(AF_INET6, ipv6_src_s, &ipv6_key.ipv6_src) != 1 ||
inet_pton(AF_INET6, ipv6_dst_s, &ipv6_key.ipv6_dst) != 1) {
return -EINVAL;
}
ipv6_key.ipv6_proto = ipv6_proto;
ipv6_key.ipv6_tos = ipv6_tos;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_IPV6,
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
&ipv6_key, sizeof ipv6_key);
return n;
}
}
{
int tcp_src;
int tcp_dst;
int n = -1;
if (sscanf(s, "tcp(src=%i,dst=%i)%n",&tcp_src, &tcp_dst, &n) > 0
&& n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_tcp tcp_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
tcp_key.tcp_src = htons(tcp_src);
tcp_key.tcp_dst = htons(tcp_dst);
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_TCP, &tcp_key, sizeof tcp_key);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
int udp_src;
int udp_dst;
int n = -1;
if (sscanf(s, "udp(src=%i,dst=%i)%n", &udp_src, &udp_dst, &n) > 0
&& n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_udp udp_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
udp_key.udp_src = htons(udp_src);
udp_key.udp_dst = htons(udp_dst);
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_UDP, &udp_key, sizeof udp_key);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
int icmp_type;
int icmp_code;
int n = -1;
if (sscanf(s, "icmp(type=%i,code=%i)%n",
&icmp_type, &icmp_code, &n) > 0
&& n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_icmp icmp_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
icmp_key.icmp_type = icmp_type;
icmp_key.icmp_code = icmp_code;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_ICMP,
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
&icmp_key, sizeof icmp_key);
return n;
}
}
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_icmpv6 icmpv6_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
int n = -1;
if (sscanf(s, "icmpv6(type=%"SCNi8",code=%"SCNi8")%n",
&icmpv6_key.icmpv6_type, &icmpv6_key.icmpv6_code,&n) > 0
&& n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_ICMPV6,
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
&icmpv6_key, sizeof icmpv6_key);
return n;
}
}
{
ovs_be32 arp_sip;
ovs_be32 arp_tip;
int arp_op;
uint8_t arp_sha[ETH_ADDR_LEN];
uint8_t arp_tha[ETH_ADDR_LEN];
int n = -1;
if (sscanf(s, "arp(sip="IP_SCAN_FMT",tip="IP_SCAN_FMT","
"op=%i,sha="ETH_ADDR_SCAN_FMT",tha="ETH_ADDR_SCAN_FMT")%n",
IP_SCAN_ARGS(&arp_sip),
IP_SCAN_ARGS(&arp_tip),
&arp_op,
ETH_ADDR_SCAN_ARGS(arp_sha),
ETH_ADDR_SCAN_ARGS(arp_tha), &n) > 0 && n > 0) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_arp arp_key;
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
memset(&arp_key, 0, sizeof arp_key);
arp_key.arp_sip = arp_sip;
arp_key.arp_tip = arp_tip;
arp_key.arp_op = htons(arp_op);
memcpy(arp_key.arp_sha, arp_sha, ETH_ADDR_LEN);
memcpy(arp_key.arp_tha, arp_tha, ETH_ADDR_LEN);
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_unspec(key, OVS_KEY_ATTR_ARP, &arp_key, sizeof arp_key);
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
return n;
}
}
{
char nd_target_s[IPV6_SCAN_LEN + 1];
uint8_t nd_sll[ETH_ADDR_LEN];
uint8_t nd_tll[ETH_ADDR_LEN];
int n = -1;
if (sscanf(s, "nd(target="IPV6_SCAN_FMT")%n",
nd_target_s, &n) > 0 && n > 0) {
return put_nd_key(n, nd_target_s, NULL, NULL, key);
}
if (sscanf(s, "nd(target="IPV6_SCAN_FMT",sll="ETH_ADDR_SCAN_FMT")%n",
nd_target_s, ETH_ADDR_SCAN_ARGS(nd_sll), &n) > 0
&& n > 0) {
return put_nd_key(n, nd_target_s, nd_sll, NULL, key);
}
if (sscanf(s, "nd(target="IPV6_SCAN_FMT",tll="ETH_ADDR_SCAN_FMT")%n",
nd_target_s, ETH_ADDR_SCAN_ARGS(nd_tll), &n) > 0
&& n > 0) {
return put_nd_key(n, nd_target_s, NULL, nd_tll, key);
}
if (sscanf(s, "nd(target="IPV6_SCAN_FMT",sll="ETH_ADDR_SCAN_FMT","
"tll="ETH_ADDR_SCAN_FMT")%n",
nd_target_s, ETH_ADDR_SCAN_ARGS(nd_sll),
ETH_ADDR_SCAN_ARGS(nd_tll), &n) > 0
&& n > 0) {
return put_nd_key(n, nd_target_s, nd_sll, nd_tll, key);
}
}
return -EINVAL;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
/* Parses the string representation of a datapath flow key, in the
* format output by odp_flow_key_format(). Returns 0 if successful,
* otherwise a positive errno value. On success, the flow key is
* appended to 'key' as a series of Netlink attributes. On failure, no
* data is appended to 'key'. Either way, 'key''s data might be
* reallocated.
odp-util: New function odp_flow_key_from_string(). This will be used in upcoming commits.
2011-08-04 16:20:34 -07:00
*
* On success, the attributes appended to 'key' are individually syntactically
* valid, but they may not be valid as a sequence. 'key' might, for example,
* be missing an "in_port" key, have duplicated keys, or have keys in the wrong
* order. odp_flow_key_to_flow() will detect those errors. */
int
odp_flow_key_from_string(const char *s, struct ofpbuf *key)
{
const size_t old_size = key->size;
for (;;) {
int retval;
s += strspn(s, ", \t\r\n");
if (!*s) {
return 0;
}
retval = parse_odp_key_attr(s, key);
if (retval < 0) {
key->size = old_size;
return -retval;
}
s += retval;
}
return 0;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
/* Appends a representation of 'flow' as OVS_KEY_ATTR_* attributes to 'buf'. */
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
void
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
odp_flow_key_from_flow(struct ofpbuf *buf, const struct flow *flow)
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
{
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ethernet *eth_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (flow->tun_id != htonll(0)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_be64(buf, OVS_KEY_ATTR_TUN_ID, flow->tun_id);
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_u32(buf, OVS_KEY_ATTR_IN_PORT,
ofproto: Break apart into generic and hardware-specific parts. In addition to the changes to ofproto, this commit changes all of the instances of "struct flow" in the tree so that the "in_port" member is an OpenFlow port number. Previously, this member was an OpenFlow port number in some cases and an ODP port number in other cases.
2011-05-11 12:13:10 -07:00
ofp_port_to_odp_port(flow->in_port));
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
eth_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_ETHERNET,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *eth_key);
memcpy(eth_key->eth_src, flow->dl_src, ETH_ADDR_LEN);
memcpy(eth_key->eth_dst, flow->dl_dst, ETH_ADDR_LEN);
if (flow->vlan_tci != htons(0)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_8021q *q_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
q_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_8021Q,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *q_key);
q_key->q_tpid = htons(ETH_TYPE_VLAN);
q_key->q_tci = flow->vlan_tci & ~htons(VLAN_CFI);
}
if (ntohs(flow->dl_type) < ETH_TYPE_MIN) {
return;
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nl_msg_put_be16(buf, OVS_KEY_ATTR_ETHERTYPE, flow->dl_type);
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (flow->dl_type == htons(ETH_TYPE_IP)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ipv4 *ipv4_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
ipv4_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_IPV4,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *ipv4_key);
Zero padding bytes in odp_key_ipv4, odp_key_arp. This is a potential security issue for the kernel. In userspace it just provokes false-positive valgrind warnings (which is how I found it). Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-02-01 11:23:30 -08:00
memset(ipv4_key, 0, sizeof *ipv4_key);
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
ipv4_key->ipv4_src = flow->nw_src;
ipv4_key->ipv4_dst = flow->nw_dst;
ipv4_key->ipv4_proto = flow->nw_proto;
ipv4_key->ipv4_tos = flow->nw_tos;
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
} else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_ipv6 *ipv6_key;
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
ipv6_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_IPV6,
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
sizeof *ipv6_key);
Zero padding bytes in odp_key_ipv4, odp_key_arp. This is a potential security issue for the kernel. In userspace it just provokes false-positive valgrind warnings (which is how I found it). Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-02-01 11:23:30 -08:00
memset(ipv6_key, 0, sizeof *ipv6_key);
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
memcpy(ipv6_key->ipv6_src, &flow->ipv6_src, sizeof ipv6_key->ipv6_src);
memcpy(ipv6_key->ipv6_dst, &flow->ipv6_dst, sizeof ipv6_key->ipv6_dst);
ipv6_key->ipv6_proto = flow->nw_proto;
ipv6_key->ipv6_tos = flow->nw_tos;
} else if (flow->dl_type == htons(ETH_TYPE_ARP)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_arp *arp_key;
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
arp_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_ARP,
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
sizeof *arp_key);
Zero padding bytes in odp_key_ipv4, odp_key_arp. This is a potential security issue for the kernel. In userspace it just provokes false-positive valgrind warnings (which is how I found it). Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-02-01 11:23:30 -08:00
memset(arp_key, 0, sizeof *arp_key);
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
arp_key->arp_sip = flow->nw_src;
arp_key->arp_tip = flow->nw_dst;
arp_key->arp_op = htons(flow->nw_proto);
memcpy(arp_key->arp_sha, flow->arp_sha, ETH_ADDR_LEN);
memcpy(arp_key->arp_tha, flow->arp_tha, ETH_ADDR_LEN);
}
lib: Whitespace cleanup.
2011-08-04 16:50:25 -07:00
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
if (flow->dl_type == htons(ETH_TYPE_IP)
|| flow->dl_type == htons(ETH_TYPE_IPV6)) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
if (flow->nw_proto == IPPROTO_TCP) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_tcp *tcp_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
tcp_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_TCP,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *tcp_key);
tcp_key->tcp_src = flow->tp_src;
tcp_key->tcp_dst = flow->tp_dst;
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
} else if (flow->nw_proto == IPPROTO_UDP) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_udp *udp_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
udp_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_UDP,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *udp_key);
udp_key->udp_src = flow->tp_src;
udp_key->udp_dst = flow->tp_dst;
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
} else if (flow->dl_type == htons(ETH_TYPE_IP)
&& flow->nw_proto == IPPROTO_ICMP) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_icmp *icmp_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
icmp_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_ICMP,
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
sizeof *icmp_key);
icmp_key->icmp_type = ntohs(flow->tp_src);
icmp_key->icmp_code = ntohs(flow->tp_dst);
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
} else if (flow->dl_type == htons(ETH_TYPE_IPV6)
&& flow->nw_proto == IPPROTO_ICMPV6) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_icmpv6 *icmpv6_key;
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
icmpv6_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_ICMPV6,
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
sizeof *icmpv6_key);
icmpv6_key->icmpv6_type = ntohs(flow->tp_src);
icmpv6_key->icmpv6_code = ntohs(flow->tp_dst);
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
if (icmpv6_key->icmpv6_type == ND_NEIGHBOR_SOLICIT
|| icmpv6_key->icmpv6_type == ND_NEIGHBOR_ADVERT) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
struct ovs_key_nd *nd_key;
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
nd_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_ND,
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
sizeof *nd_key);
memcpy(nd_key->nd_target, &flow->nd_target,
sizeof nd_key->nd_target);
memcpy(nd_key->nd_sll, flow->arp_sha, ETH_ADDR_LEN);
memcpy(nd_key->nd_tll, flow->arp_tha, ETH_ADDR_LEN);
}
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
}
}
}
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
/* Converts the 'key_len' bytes of OVS_KEY_ATTR_* attributes in 'key' to a flow
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
* structure in 'flow'. Returns 0 if successful, otherwise EINVAL. */
int
odp_flow_key_to_flow(const struct nlattr *key, size_t key_len,
struct flow *flow)
{
const struct nlattr *nla;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
enum ovs_key_type prev_type;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
size_t left;
memset(flow, 0, sizeof *flow);
flow->dl_type = htons(FLOW_DL_TYPE_NONE);
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
prev_type = OVS_KEY_ATTR_UNSPEC;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
NL_ATTR_FOR_EACH (nla, left, key, key_len) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
const struct ovs_key_ethernet *eth_key;
const struct ovs_key_8021q *q_key;
const struct ovs_key_ipv4 *ipv4_key;
const struct ovs_key_ipv6 *ipv6_key;
const struct ovs_key_tcp *tcp_key;
const struct ovs_key_udp *udp_key;
const struct ovs_key_icmp *icmp_key;
const struct ovs_key_icmpv6 *icmpv6_key;
const struct ovs_key_arp *arp_key;
const struct ovs_key_nd *nd_key;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
uint16_t type = nl_attr_type(nla);
int len = odp_flow_key_attr_len(type);
if (nl_attr_get_size(nla) != len && len != -1) {
return EINVAL;
}
#define TRANSITION(PREV_TYPE, TYPE) (((PREV_TYPE) << 16) | (TYPE))
switch (TRANSITION(prev_type, type)) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_UNSPEC, OVS_KEY_ATTR_TUN_ID):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
flow->tun_id = nl_attr_get_be64(nla);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_UNSPEC, OVS_KEY_ATTR_IN_PORT):
case TRANSITION(OVS_KEY_ATTR_TUN_ID, OVS_KEY_ATTR_IN_PORT):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (nl_attr_get_u32(nla) >= UINT16_MAX) {
return EINVAL;
}
ofproto: Break apart into generic and hardware-specific parts. In addition to the changes to ofproto, this commit changes all of the instances of "struct flow" in the tree so that the "in_port" member is an OpenFlow port number. Previously, this member was an OpenFlow port number in some cases and an ODP port number in other cases.
2011-05-11 12:13:10 -07:00
flow->in_port = odp_port_to_ofp_port(nl_attr_get_u32(nla));
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_IN_PORT, OVS_KEY_ATTR_ETHERNET):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
eth_key = nl_attr_get(nla);
memcpy(flow->dl_src, eth_key->eth_src, ETH_ADDR_LEN);
memcpy(flow->dl_dst, eth_key->eth_dst, ETH_ADDR_LEN);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_ETHERNET, OVS_KEY_ATTR_8021Q):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
q_key = nl_attr_get(nla);
if (q_key->q_tpid != htons(ETH_TYPE_VLAN)) {
/* Only standard 0x8100 VLANs currently supported. */
return EINVAL;
}
if (q_key->q_tci & htons(VLAN_CFI)) {
return EINVAL;
}
flow->vlan_tci = q_key->q_tci | htons(VLAN_CFI);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_8021Q, OVS_KEY_ATTR_ETHERTYPE):
case TRANSITION(OVS_KEY_ATTR_ETHERNET, OVS_KEY_ATTR_ETHERTYPE):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
flow->dl_type = nl_attr_get_be16(nla);
if (ntohs(flow->dl_type) < 1536) {
return EINVAL;
}
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_ETHERTYPE, OVS_KEY_ATTR_IPV4):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (flow->dl_type != htons(ETH_TYPE_IP)) {
return EINVAL;
}
ipv4_key = nl_attr_get(nla);
flow->nw_src = ipv4_key->ipv4_src;
flow->nw_dst = ipv4_key->ipv4_dst;
flow->nw_proto = ipv4_key->ipv4_proto;
flow->nw_tos = ipv4_key->ipv4_tos;
if (flow->nw_tos & IP_ECN_MASK) {
return EINVAL;
}
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_ETHERTYPE, OVS_KEY_ATTR_IPV6):
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
if (flow->dl_type != htons(ETH_TYPE_IPV6)) {
return EINVAL;
}
ipv6_key = nl_attr_get(nla);
memcpy(&flow->ipv6_src, ipv6_key->ipv6_src, sizeof flow->ipv6_src);
memcpy(&flow->ipv6_dst, ipv6_key->ipv6_dst, sizeof flow->ipv6_dst);
flow->nw_proto = ipv6_key->ipv6_proto;
flow->nw_tos = ipv6_key->ipv6_tos;
if (flow->nw_tos & IP_ECN_MASK) {
return EINVAL;
}
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_IPV4, OVS_KEY_ATTR_TCP):
case TRANSITION(OVS_KEY_ATTR_IPV6, OVS_KEY_ATTR_TCP):
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
if (flow->nw_proto != IPPROTO_TCP) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
}
tcp_key = nl_attr_get(nla);
flow->tp_src = tcp_key->tcp_src;
flow->tp_dst = tcp_key->tcp_dst;
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_IPV4, OVS_KEY_ATTR_UDP):
case TRANSITION(OVS_KEY_ATTR_IPV6, OVS_KEY_ATTR_UDP):
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
if (flow->nw_proto != IPPROTO_UDP) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
}
udp_key = nl_attr_get(nla);
flow->tp_src = udp_key->udp_src;
flow->tp_dst = udp_key->udp_dst;
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_IPV4, OVS_KEY_ATTR_ICMP):
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
if (flow->nw_proto != IPPROTO_ICMP) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
}
icmp_key = nl_attr_get(nla);
flow->tp_src = htons(icmp_key->icmp_type);
flow->tp_dst = htons(icmp_key->icmp_code);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_IPV6, OVS_KEY_ATTR_ICMPV6):
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
if (flow->nw_proto != IPPROTO_ICMPV6) {
return EINVAL;
}
icmpv6_key = nl_attr_get(nla);
flow->tp_src = htons(icmpv6_key->icmpv6_type);
flow->tp_dst = htons(icmpv6_key->icmpv6_code);
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_ETHERTYPE, OVS_KEY_ATTR_ARP):
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (flow->dl_type != htons(ETH_TYPE_ARP)) {
return EINVAL;
}
arp_key = nl_attr_get(nla);
flow->nw_src = arp_key->arp_sip;
flow->nw_dst = arp_key->arp_tip;
if (arp_key->arp_op & htons(0xff00)) {
return EINVAL;
}
flow->nw_proto = ntohs(arp_key->arp_op);
nicira-ext: Support matching ARP source and target hardware addresses. OpenFlow 1.0 doesn't allow matching on the ARP source and target hardware address. This has caused us to introduce hacks such as the Drop Spoofed ARP action. Now that we have extensible match, we can match on more fields within ARP: - Source Hardware Address (arp_sha) - Target Hardware Address (arp_tha) Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-07 14:02:17 -08:00
memcpy(flow->arp_sha, arp_key->arp_sha, ETH_ADDR_LEN);
memcpy(flow->arp_tha, arp_key->arp_tha, ETH_ADDR_LEN);
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
break;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case TRANSITION(OVS_KEY_ATTR_ICMPV6, OVS_KEY_ATTR_ND):
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
if (flow->tp_src != htons(ND_NEIGHBOR_SOLICIT)
&& flow->tp_src != htons(ND_NEIGHBOR_ADVERT)) {
return EINVAL;
}
nd_key = nl_attr_get(nla);
memcpy(&flow->nd_target, nd_key->nd_target, sizeof flow->nd_target);
memcpy(flow->arp_sha, nd_key->nd_sll, ETH_ADDR_LEN);
memcpy(flow->arp_tha, nd_key->nd_tll, ETH_ADDR_LEN);
break;
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
default:
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
if (type == OVS_KEY_ATTR_UNSPEC
|| prev_type == OVS_KEY_ATTR_UNSPEC) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
}
return EINVAL;
}
prev_type = type;
}
if (left) {
return EINVAL;
}
switch (prev_type) {
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_UNSPEC:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_TUN_ID:
case OVS_KEY_ATTR_IN_PORT:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ETHERNET:
case OVS_KEY_ATTR_8021Q:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ETHERTYPE:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
if (flow->dl_type == htons(ETH_TYPE_IP)
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
|| flow->dl_type == htons(ETH_TYPE_IPV6)
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
|| flow->dl_type == htons(ETH_TYPE_ARP)) {
return EINVAL;
}
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_IPV4:
lib: Replace IP_TYPE_ references with IPPROTO_. A few common IP protocol types were defined in "lib/packets.h". However, we already assume the existence of <netinet/in.h> which contains a more exhaustive list and should be available on POSIX systems.
2011-02-02 11:33:20 -08:00
if (flow->nw_proto == IPPROTO_TCP
|| flow->nw_proto == IPPROTO_UDP
|| flow->nw_proto == IPPROTO_ICMP) {
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return EINVAL;
}
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_IPV6:
nicira-ext: Support matching IPv6 traffic. Provides ability to match over IPv6 traffic in the same manner as IPv4. Currently, the matching fields include: - IPv6 source and destination addresses (ipv6_src and ipv6_dst) - Traffic Class (nw_tos) - Next Header (nw_proto) - ICMPv6 Type and Code (icmp_type and icmp_code) - TCP and UDP Ports over IPv6 (tp_src and tp_dst) When defining IPv6 rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2010-12-29 19:03:46 -08:00
if (flow->nw_proto == IPPROTO_TCP
|| flow->nw_proto == IPPROTO_UDP
|| flow->nw_proto == IPPROTO_ICMPV6) {
return EINVAL;
}
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_ICMPV6:
nicira-ext: Support matching IPv6 Neighbor Discovery messages. IPv6 uses Neighbor Discovery messages in a similar manner to how IPv4 uses ARP. This commit adds support for matching deeper into the payloads of Neighbor Solicitation (NS) and Neighbor Advertisement (NA) messages. Currently, the matching fields include: - NS and NA Target (nd_target) - NS Source Link Layer Address (nd_sll) - NA Target Link Layer Address (nd_tll) When defining IPv6 Neighbor Discovery rules, the Nicira Extensible Match (NXM) extension to OVS must be used. Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
2011-02-01 22:54:11 -08:00
if (flow->icmp_type == htons(ND_NEIGHBOR_SOLICIT)
|| flow->icmp_type == htons(ND_NEIGHBOR_ADVERT)) {
return EINVAL;
}
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case OVS_KEY_ATTR_TCP:
case OVS_KEY_ATTR_UDP:
case OVS_KEY_ATTR_ICMP:
case OVS_KEY_ATTR_ARP:
case OVS_KEY_ATTR_ND:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
return 0;
datapath: Use "OVS_*" as opposed to "ODP_*" for user<->kernel interactions. The prefix "ODP_*" is not overly descriptive in the context of the larger Linux tree. This commit changes the prefix to "OVS_*" for the userpace to kernel interactions. The userspace libraries still use "ODP_" in many of their interfaces since it is more descriptive in the OVS oeuvre. Feature #6904 Signed-off-by: Justin Pettit <jpettit@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-08-18 10:35:40 -07:00
case __OVS_KEY_ATTR_MAX:
datapath: Convert odp_flow_key to use Netlink attributes instead. One of the goals for Open vSwitch is to decouple kernel and userspace software, so that either one can be upgraded or rolled back independent of the other. To do this in full generality, it must be possible to change the kernel's idea of the flow key separately from the userspace version. In turn, that means that flow keys must become variable-length. This commit makes that change using Netlink attribute sequences. This commit does not actually make userspace flexible enough to handle changes in the kernel flow key structure, because userspace doesn't yet have enough information to do that intelligently. Upcoming commits will fix that. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 18:44:44 -08:00
default:
NOT_REACHED();
}
flow: Separate "flow_t" from "struct odp_flow_key". The "struct odp_flow_key" used in the kernel datapath is conceptually separate from the "flow_t" used in userspace, but until now we have used the latter as a typedef for the former for convenience. This commit separates them. This makes it possible in upcoming commits to change them independently. This is cross-ported from the "wdp" branch, which has had it for months.
2010-10-11 13:31:35 -07:00
}
Reference in New Issue Copy Permalink