mir/ovs
2
0
Fork 0
mirror of https://github.com/openvswitch/ovs synced 2025-08-24 19:07:45 +00:00
Code Issues Releases Wiki Activity
ovs/lib/odp-util.c

604 lines
19 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
*/
#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>
#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"
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: 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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
if (type > ODP_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;
}
switch ((enum odp_action_type) type) {
datapath: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_ACTION_ATTR_OUTPUT: return 4;
case ODP_ACTION_ATTR_CONTROLLER: return 8;
case ODP_ACTION_ATTR_SET_DL_TCI: return 2;
case ODP_ACTION_ATTR_STRIP_VLAN: return 0;
case ODP_ACTION_ATTR_SET_DL_SRC: return ETH_ADDR_LEN;
case ODP_ACTION_ATTR_SET_DL_DST: return ETH_ADDR_LEN;
case ODP_ACTION_ATTR_SET_NW_SRC: return 4;
case ODP_ACTION_ATTR_SET_NW_DST: return 4;
case ODP_ACTION_ATTR_SET_NW_TOS: return 1;
case ODP_ACTION_ATTR_SET_TP_SRC: return 2;
case ODP_ACTION_ATTR_SET_TP_DST: return 2;
case ODP_ACTION_ATTR_SET_TUNNEL: return 8;
case ODP_ACTION_ATTR_SET_PRIORITY: return 4;
case ODP_ACTION_ATTR_POP_PRIORITY: return 0;
case ODP_ACTION_ATTR_DROP_SPOOFED_ARP: return 0;
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
datapath: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_ACTION_ATTR_UNSPEC:
case __ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_ACTION_ATTR_CONTROLLER:
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, "ctl(%"PRIu64")", nl_attr_get_u64(a));
Import from old repository commit 61ef2b42a9c4ba8e1600f15bb0236765edc2ad45.
2009-07-08 13:19:16 -07:00
break;
datapath: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_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");
break;
datapath: s/ODPAT_/ODP_ACTION_ATTR_/ to fit new naming scheme. Jesse suggested this naming scheme, so I'm adjusting existing names to fit it. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
2011-01-23 21:56:00 -08:00
case ODP_ACTION_ATTR_DROP_SPOOFED_ARP:
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
ds_put_cstr(ds, "drop_spoofed_arp");
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)
{
if (type > ODP_KEY_ATTR_MAX) {
return -1;
}
switch ((enum odp_key_type) type) {
case ODP_KEY_ATTR_TUN_ID: return 8;
case ODP_KEY_ATTR_IN_PORT: return 4;
case ODP_KEY_ATTR_ETHERNET: return sizeof(struct odp_key_ethernet);
case ODP_KEY_ATTR_8021Q: return sizeof(struct odp_key_8021q);
case ODP_KEY_ATTR_ETHERTYPE: return 2;
case ODP_KEY_ATTR_IPV4: return sizeof(struct odp_key_ipv4);
case ODP_KEY_ATTR_TCP: return sizeof(struct odp_key_tcp);
case ODP_KEY_ATTR_UDP: return sizeof(struct odp_key_udp);
case ODP_KEY_ATTR_ICMP: return sizeof(struct odp_key_icmp);
case ODP_KEY_ATTR_ARP: return sizeof(struct odp_key_arp);
case ODP_KEY_ATTR_UNSPEC:
case __ODP_KEY_ATTR_MAX:
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)
{
const struct odp_key_ethernet *eth_key;
const struct odp_key_8021q *q_key;
const struct odp_key_ipv4 *ipv4_key;
const struct odp_key_tcp *tcp_key;
const struct odp_key_udp *udp_key;
const struct odp_key_icmp *icmp_key;
const struct odp_key_arp *arp_key;
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)) {
case ODP_KEY_ATTR_TUN_ID:
ds_put_format(ds, "tun_id(%#"PRIx64")", nl_attr_get_be64(a));
break;
case ODP_KEY_ATTR_IN_PORT:
ds_put_format(ds, "in_port(%"PRIu32")", nl_attr_get_u32(a));
break;
case ODP_KEY_ATTR_ETHERNET:
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;
case ODP_KEY_ATTR_8021Q:
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
}
ds_put_format(ds, "vid%"PRIu16",pcp%d)",
vlan_tci_to_vid(q_key->q_tci),
vlan_tci_to_pcp(q_key->q_tci));
break;
case ODP_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;
case ODP_KEY_ATTR_IPV4:
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;
case ODP_KEY_ATTR_TCP:
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;
case ODP_KEY_ATTR_UDP:
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;
case ODP_KEY_ATTR_ICMP:
icmp_key = nl_attr_get(a);
ds_put_format(ds, "icmp(type=%"PRIu8",code=%"PRIu8")",
icmp_key->icmp_type, icmp_key->icmp_code);
break;
case ODP_KEY_ATTR_ARP:
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;
default:
format_generic_odp_key(a, ds);
break;
}
}
/* Appends to 'ds' a string representation of the 'key_len' bytes of
* ODP_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
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
/* Appends a representation of 'flow' as ODP_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: 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
struct odp_key_ethernet *eth_key;
if (flow->tun_id != htonll(0)) {
nl_msg_put_be64(buf, ODP_KEY_ATTR_TUN_ID, flow->tun_id);
}
nl_msg_put_u32(buf, ODP_KEY_ATTR_IN_PORT, flow->in_port);
eth_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_ETHERNET,
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)) {
struct odp_key_8021q *q_key;
q_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_8021Q,
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;
}
nl_msg_put_be16(buf, ODP_KEY_ATTR_ETHERTYPE, flow->dl_type);
if (flow->dl_type == htons(ETH_TYPE_IP)) {
struct odp_key_ipv4 *ipv4_key;
ipv4_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_IPV4,
sizeof *ipv4_key);
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;
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
struct odp_key_tcp *tcp_key;
tcp_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_TCP,
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: 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
struct odp_key_udp *udp_key;
udp_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_UDP,
sizeof *udp_key);
udp_key->udp_src = flow->tp_src;
udp_key->udp_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_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
struct odp_key_icmp *icmp_key;
icmp_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_ICMP,
sizeof *icmp_key);
icmp_key->icmp_type = ntohs(flow->tp_src);
icmp_key->icmp_code = ntohs(flow->tp_dst);
}
} else if (flow->dl_type == htons(ETH_TYPE_ARP)) {
struct odp_key_arp *arp_key;
arp_key = nl_msg_put_unspec_uninit(buf, ODP_KEY_ATTR_ARP,
sizeof *arp_key);
arp_key->arp_sip = flow->nw_src;
arp_key->arp_tip = flow->nw_dst;
arp_key->arp_op = htons(flow->nw_proto);
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(arp_key->arp_sha, flow->arp_sha, ETH_ADDR_LEN);
memcpy(arp_key->arp_tha, 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
}
}
/* Converts the 'key_len' bytes of ODP_KEY_ATTR_* attributes in 'key' to a flow
* 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;
enum odp_key_type prev_type;
size_t left;
memset(flow, 0, sizeof *flow);
flow->dl_type = htons(FLOW_DL_TYPE_NONE);
prev_type = ODP_KEY_ATTR_UNSPEC;
NL_ATTR_FOR_EACH (nla, left, key, key_len) {
const struct odp_key_ethernet *eth_key;
const struct odp_key_8021q *q_key;
const struct odp_key_ipv4 *ipv4_key;
const struct odp_key_tcp *tcp_key;
const struct odp_key_udp *udp_key;
const struct odp_key_icmp *icmp_key;
const struct odp_key_arp *arp_key;
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)) {
case TRANSITION(ODP_KEY_ATTR_UNSPEC, ODP_KEY_ATTR_TUN_ID):
flow->tun_id = nl_attr_get_be64(nla);
break;
case TRANSITION(ODP_KEY_ATTR_UNSPEC, ODP_KEY_ATTR_IN_PORT):
case TRANSITION(ODP_KEY_ATTR_TUN_ID, ODP_KEY_ATTR_IN_PORT):
if (nl_attr_get_u32(nla) >= UINT16_MAX) {
return EINVAL;
}
flow->in_port = nl_attr_get_u32(nla);
break;
case TRANSITION(ODP_KEY_ATTR_IN_PORT, ODP_KEY_ATTR_ETHERNET):
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;
case TRANSITION(ODP_KEY_ATTR_ETHERNET, ODP_KEY_ATTR_8021Q):
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;
case TRANSITION(ODP_KEY_ATTR_8021Q, ODP_KEY_ATTR_ETHERTYPE):
case TRANSITION(ODP_KEY_ATTR_ETHERNET, ODP_KEY_ATTR_ETHERTYPE):
flow->dl_type = nl_attr_get_be16(nla);
if (ntohs(flow->dl_type) < 1536) {
return EINVAL;
}
break;
case TRANSITION(ODP_KEY_ATTR_ETHERTYPE, ODP_KEY_ATTR_IPV4):
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;
case TRANSITION(ODP_KEY_ATTR_IPV4, ODP_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;
case TRANSITION(ODP_KEY_ATTR_IPV4, ODP_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;
case TRANSITION(ODP_KEY_ATTR_IPV4, ODP_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;
case TRANSITION(ODP_KEY_ATTR_ETHERTYPE, ODP_KEY_ATTR_ARP):
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;
default:
if (type == ODP_KEY_ATTR_UNSPEC
|| prev_type == ODP_KEY_ATTR_UNSPEC) {
return EINVAL;
}
return EINVAL;
}
prev_type = type;
}
if (left) {
return EINVAL;
}
switch (prev_type) {
case ODP_KEY_ATTR_UNSPEC:
return EINVAL;
case ODP_KEY_ATTR_TUN_ID:
case ODP_KEY_ATTR_IN_PORT:
return EINVAL;
case ODP_KEY_ATTR_ETHERNET:
case ODP_KEY_ATTR_8021Q:
return 0;
case ODP_KEY_ATTR_ETHERTYPE:
if (flow->dl_type == htons(ETH_TYPE_IP)
|| flow->dl_type == htons(ETH_TYPE_ARP)) {
return EINVAL;
}
return 0;
case ODP_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;
case ODP_KEY_ATTR_TCP:
case ODP_KEY_ATTR_UDP:
case ODP_KEY_ATTR_ICMP:
case ODP_KEY_ATTR_ARP:
return 0;
case __ODP_KEY_ATTR_MAX:
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