David Ahern [Sat, 23 Apr 2022 15:22:59 +0000 (09:22 -0600)]
libbpf: Remove use of bpf_program__set_priv and bpf_program__priv
bpf_program__set_priv and bpf_program__priv are deprecated as of
libbpf v0.7+. Rather than store the map as priv on the program,
change find_legacy_tail_calls to take an argument to return a reference
to the map.
find_legacy_tail_calls is invoked twice from load_bpf_object - the
first time to check for programs that should be loaded. In this case
a reference to the map is not needed, but it does validate the map
exists. The second is invoked from update_legacy_tail_call_maps where
the map pointer is needed.
Signed-off-by: David Ahern <dsahern@kernel.org> Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Boris Sukholitko [Thu, 28 Apr 2022 08:32:33 +0000 (11:32 +0300)]
f_flower: add number of vlans man entry
The documentation was missing in the number of vlans commit.
Fixes: 5ba31bcf (f_flower: Add num of vlans parameter) Signed-off-by: Boris Sukholitko <boris.sukholitko@broadcom.com> Signed-off-by: David Ahern <dsahern@kernel.org>
David Ahern [Thu, 28 Apr 2022 02:18:25 +0000 (20:18 -0600)]
Merge branch 'flower-vlans' into next
Boris Sukholitko says:
====================
Our customers in the fiber telecom world have network configurations
where they would like to control their traffic according to the number
of tags appearing in the packet.
For example, TR247 GPON conformance test suite specification mostly
talks about untagged, single, double tagged packets and gives lax
guidelines on the vlan protocol vs. number of vlan tags.
This is different from the common IT networks where 802.1Q and 802.1ad
protocols are usually describe single and double tagged packet. GPON
configurations that we work with have arbitrary mix the above protocols
and number of vlan tags in the packet.
The following patch series implement number of vlans flower filter. They
add num_of_vlans flower filter as an alternative to vlan ethtype protocol
matching. The end result is that the following command becomes possible:
tc filter add dev eth1 ingress flower \
num_of_vlans 1 vlan_prio 5 action drop
Also, from our logs, we have redirect rules such that:
tc filter add dev $GPON ingress flower num_of_vlans $N \
action mirred egress redirect dev $DEV
where N can range from 0 to 3 and $DEV is the function of $N.
Also there are rules setting skb mark based on the number of vlans:
tc filter add dev $GPON ingress flower num_of_vlans $N vlan_prio \
$P action skbedit mark $M
Boris Sukholitko [Tue, 26 Apr 2022 09:14:16 +0000 (12:14 +0300)]
f_flower: Add num of vlans parameter
Our customers in the fiber telecom world have network configurations
where they would like to control their traffic according to the number
of tags appearing in the packet.
For example, TR247 GPON conformance test suite specification mostly
talks about untagged, single, double tagged packets and gives lax
guidelines on the vlan protocol vs. number of vlan tags.
This is different from the common IT networks where 802.1Q and 802.1ad
protocols are usually describe single and double tagged packet. GPON
configurations that we work with have arbitrary mix the above protocols
and number of vlan tags in the packet.
This patch adds num_of_vlans flower key and associated print and parse
routines. The following command becomes possible:
tc filter add dev eth1 ingress flower num_of_vlans 1 action drop
Signed-off-by: Boris Sukholitko <boris.sukholitko@broadcom.com> Signed-off-by: David Ahern <dsahern@kernel.org>
David Ahern [Thu, 28 Apr 2022 02:13:35 +0000 (20:13 -0600)]
Merge branch 'ip-stats' into next
Petr Machata says:
====================
A new rtnetlink message, RTM_SETSTATS, has been added recently in kernel
commit ca0a53dcec94 ("Merge branch 'net-hw-counters-for-soft-devices'").
At the same time, RTM_GETSTATS has been around for a while. The users of
this API are spread in a couple different places: "ip link xstats" reads
stats from the IFLA_STATS_LINK_XSTATS and _XSTATS_SLAVE subgroups, "ip
link afstats" then reads IFLA_STATS_AF_SPEC.
Finally, to read IFLA_STATS_LINK_OFFLOAD_XSTATS, one would use ifstats.
This does not seem to be a good fit for IFLA_OFFLOAD_XSTATS_HW_S_INFO in
particular.
The obvious place to expose all these offload stats suites would be
under a new link subcommand "ip link offload_xstats", or similar, which
would then have syntax for both showing stats and setting them.
However, this looks like a good opportunity to introduce a new top-level
command, "ip stats", that would be the go-to place to access anything
backed by RTM_GETSTATS and RTM_SETSTATS.
This patchset therefore does the following:
- It adds the new "stats" infrastructure
- It adds specifically the ability to toggle and show the suites that
were recently added to Linux, IFLA_OFFLOAD_XSTATS_HW_S_INFO and
IFLA_OFFLOAD_XSTATS_L3_STATS.
- It adds support to dump IFLA_OFFLOAD_XSTATS_CPU_HIT, which was not
available under "ip" at all.
- Does all this in a way that is easy to extend for new stats suites.
The patchset proceeds as follows:
- Patches #1 and #2 lay some groundwork and tweak existing code.
- Patch #3 adds the shell of the new "ip stats" command.
- Patch #4 adds "ip stats set" and the ability to toggle l3_stats in
particular.
- Patch #5 adds "ip stats show", but no actual stats suites.
- Patches #6-#9 add support for showing individual stats suites:
respectively, IFLA_STATS_LINK_64, IFLA_OFFLOAD_XSTATS_CPU_HIT,
IFLA_OFFLOAD_XSTATS_HW_S_INFO and IFLA_OFFLOAD_XSTATS_L3_STATS.
- Patch #10 adds support for monitoring stats events to "ip monitor".
- Patch #11 adds man page verbiage for the above.
The plan is to contribute support for afstats and xstats in a follow-up
patch set.
Petr Machata [Fri, 22 Apr 2022 08:30:58 +0000 (10:30 +0200)]
ipstats: Add offload subgroup "l3_stats"
Add into the group "offload" a subgroup "l3_stats" for showing
L3 statistics.
For example:
# ip stats show dev swp2.200 group offload subgroup l3_stats
4212: swp2.200: group offload subgroup l3_stats on used on
RX: bytes packets errors dropped mcast
1920 21 1 0 0
TX: bytes packets errors dropped
756 9 0 0
Petr Machata [Fri, 22 Apr 2022 08:30:57 +0000 (10:30 +0200)]
ipstats: Add offload subgroup "hw_stats_info"
Add into the group "offload" a subgroup "hw_stats_info" for showing
information about HW statistics counters.
For example:
# ip stats show dev swp1 group offload subgroup hw_stats_info
4178: swp1: group offload subgroup hw_stats_info
l3_stats on used off
# ip -j stats show dev swp1 group offload subgroup hw_stats_info | jq
[
{
"ifindex": 4178,
"ifname": "swp1",
"group": "offload",
"subgroup": "hw_stats_info",
"info": {
"l3_stats": {
"request": true,
"used": false
}
}
}
]
Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Petr Machata [Fri, 22 Apr 2022 08:30:56 +0000 (10:30 +0200)]
ipstats: Add a group "offload", subgroup "cpu_hit"
Add a new group, "offload", for showing counters from the
IFLA_STATS_LINK_OFFLOAD_XSTATS nest, and a subgroup "cpu_hit" for the
IFLA_OFFLOAD_XSTATS_CPU_HIT stats suite.
For example:
# ip stats show dev swp1 group offload subgroup cpu_hit
4178: swp1: group offload subgroup cpu_hit
RX: bytes packets errors dropped missed mcast
45522 353 0 0 0 0
TX: bytes packets errors dropped carrier collsns
46054 355 0 0 0 0
Petr Machata [Fri, 22 Apr 2022 08:30:52 +0000 (10:30 +0200)]
ip: Add a new family of commands, "stats"
Add a core of a new frontend tool for interfacing with the RTM_*STATS
family of messages. The following patches will add subcommands for showing
and setting individual statistics suites.
Note that in this patch, "ip stats" is made to be an invalid command line.
This will be changed in later patches to default to "show" when that is
introduced.
Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Petr Machata [Fri, 22 Apr 2022 08:30:51 +0000 (10:30 +0200)]
ip: Publish functions for stats formatting
Formatting struct rtnl_link_stats64 will be useful outside of iplink.c as
well. Extract from __print_link_stats() a new function, print_stats64(),
make it non-static and publish in the header file.
Additionally, publish the helper size_columns(), which will be useful for
formatting the new struct rtnl_hw_stats64.
Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Petr Machata [Fri, 22 Apr 2022 08:30:50 +0000 (10:30 +0200)]
libnetlink: Add filtering to rtnl_statsdump_req_filter()
A number of functions in the rtnl_*_req family accept a caller-provided
callback to set up arbitrary filtering. rtnl_statsdump_req_filter()
currently only allows setting a field in the IFSM header, not custom
attributes. So far these were not necessary, but with introduction of more
detailed filtering settings, the callback becomes necessary.
To that end, add a filter_fn and filter_data arguments to the function.
Unlike the other filters, this one is typed to expect an IFSM pointer, to
permit tweaking the header itself as well.
Pass NULLs in the existing callers.
Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
devlink: introduce -[he]x cmdline option to allow dumping numbers in hex format
For health reporter dumps it is quite convenient to have the numbers in
hexadecimal format. Introduce a command line option to allow user to
achieve that output.
Signed-off-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
David Ahern [Thu, 24 Mar 2022 03:27:39 +0000 (21:27 -0600)]
Merge branch 'ss-rpcinfo' into next
Andrea Claudi says:
====================
ss uses rpcinfo to get info about rpc service sockets. However, rpcinfo
is not part of iproute2 and it's an implicit dependency for ss.
This series uses libtirpc[1] API to implement the same feature of
rpcinfo for ss. This makes it possible to get info about rpc sockets,
provided ss is compiled with libtirpc support.
As a nice byproduct, this makes ss provide info about some ipv6 rpc
sockets that are not displayed using 'rpcinfo -p'.
- patch 1 adds a configure function to check for libtirpc;
- patch 2 actually rework ss to use libtirpc.
Andrea Claudi [Wed, 16 Mar 2022 18:52:14 +0000 (19:52 +0100)]
ss: remove an implicit dependency on rpcinfo
ss uses rpcinfo to get info about rpc services socket. This makes it
dependent on a tool not included in iproute2, and makes it impossible to
get info on rpc sockets if rpcinfo is not installed.
This reworks init_service_resolver() to use libtirpc, thus avoiding the
implicity dependency on rpcinfo. Moreover, this also makes it possible
to display info about ipv6 rpc socket that are not included in the
rpcinfo -p output.
For example, before this patch:
$ ss -rtap
LISTEN 0 5 localhost:ipp [::]:* users:(("cupsd",pid=1600,fd=9))
LISTEN 0 64 [::]:34265 [::]:*
LISTEN 0 64 [::]:rpc.nfs_acl [::]:*
LISTEN 0 128 [::]:42253 [::]:* users:(("rpc.statd",pid=146164,fd=12))
After this patch:
$ ss -rtap
LISTEN 0 5 localhost:ipp [::]:* users:(("cupsd",pid=1600,fd=9))
LISTEN 0 64 [::]:rpc.nlockmgr [::]:*
LISTEN 0 64 [::]:rpc.nfs_acl [::]:*
LISTEN 0 128 [::]:rpc.status [::]:* users:(("rpc.statd",pid=146164,fd=12))
Signed-off-by: Andrea Claudi <aclaudi@redhat.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Andrea Claudi [Wed, 16 Mar 2022 18:52:13 +0000 (19:52 +0100)]
configure: add check_libtirpc()
This patch adds a configure function to check if libtirpc is installed
on the build system. If this is the case, it makes iproute2 to compile
with libtirpc support.
Signed-off-by: Andrea Claudi <aclaudi@redhat.com> Signed-off-by: David Ahern <dsahern@kernel.org>
David Ahern [Sun, 20 Mar 2022 23:48:43 +0000 (17:48 -0600)]
Merge branch 'gtp-netdev' into next
Wojciech Drewek says:
====================
This patch series introduces GTP support to iproute2. Since this patch
series it is possible to create net devices of GTP type. Then, those
devices can be used in tc in order to offload GTP packets. New field
in tc flower (gtp_opts) can be used to match on QFI and PDU type.
Wojciech Drewek [Thu, 17 Mar 2022 16:27:55 +0000 (17:27 +0100)]
f_flower: Implement gtp options support
Add support for parsing TCA_FLOWER_KEY_ENC_OPTS_GTP.
Options are as follows: PDU_TYPE:QFI where each
option is represented as 8-bit hexadecimal value.
e.g.
# ip link add gtp_dev type gtp role sgsn
# tc qdisc add dev gtp_dev ingress
# tc filter add dev gtp_dev protocol ip parent ffff: \
flower \
enc_key_id 11 \
gtp_opts 1:8/ff:ff \
action mirred egress redirect dev eth0
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Wojciech Drewek [Thu, 17 Mar 2022 16:27:54 +0000 (17:27 +0100)]
ip: GTP support in ip link
Support for creating GTP devices through ip link. Two arguments
can be specified by the user when adding device of the GTP type.
- role (sgsn or ggsn) - indicates whether we are on the GGSN or SGSN
- hsize - indicates the size of the hash table where PDP sessions
are stored
IFLA_GTP_FD0 and IFLA_GTP_FD1 arguments would not be provided. Those
are file descriptores to the sockets created in the userspace. Since
we are not going to create sockets in ip link, we don't have to
provide them.
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com> Reviewed-by: Harald Welte <laforge@gnumonks.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 16 Mar 2022 09:02:56 +0000 (10:02 +0100)]
bridge: support for controlling mcast_router per port
The bridge vlan command supports setting mcast_router per-port and
per-vlan, what's however missing is the ability to set the per-port
mcast_router options, e.g. when VLAN filtering is disabled.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Antony Antony [Wed, 9 Mar 2022 20:52:58 +0000 (21:52 +0100)]
testsuite: link xfrm delete no if_id test
Since kernel commit 8dce43919566 ("xfrm: interface with if_id 0 should return error")
if_id should be non zero.
Delete the test without if_id, which defaulted if_id to zero.
Reported-by: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: Antony Antony <antony.antony@secunet.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Eli Cohen [Sun, 13 Mar 2022 17:12:19 +0000 (19:12 +0200)]
vdpa: Support reading device features
When showing the available management devices, check if
VDPA_ATTR_DEV_SUPPORTED_FEATURES feature is available and print the
supported features for a management device.
Examples:
$ vdpa mgmtdev show
auxiliary/mlx5_core.sf.1:
supported_classes net
max_supported_vqs 257
dev_features CSUM GUEST_CSUM MTU HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ MQ \
CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
Reviewed-by: Si-Wei Liu <si-wei.liu@oracle.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Eli Cohen <elic@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Eli Cohen [Sun, 13 Mar 2022 17:12:18 +0000 (19:12 +0200)]
vdpa: Support for configuring max VQ pairs for a device
Use VDPA_ATTR_DEV_MGMTDEV_MAX_VQS to specify max number of virtqueue
pairs to configure for a vdpa device when adding a device.
Examples:
1. Create a device with 3 virtqueue pairs:
$ vdpa dev add name vdpa-a mgmtdev auxiliary/mlx5_core.sf.1 max_vqp 3
2. Read the configuration of a vdpa device
$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 3 \
mtu 1500
negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
Reviewed-by: Si-Wei Liu <si-wei.liu@oracle.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Eli Cohen <elic@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Eli Cohen [Sun, 13 Mar 2022 17:12:17 +0000 (19:12 +0200)]
vdpa: Allow for printing negotiated features of a device
When reading the configuration of a vdpa device, check if the
VDPA_ATTR_DEV_NEGOTIATED_FEATURES is available. If it is, parse the
feature bits and print a string representation of each of the feature
bits.
We keep the strings in two different arrays. One for net device related
devices and one for generic feature bits.
In this patch we parse only net device specific features. Support for
other devices can be added later. If the device queried is not a net
device, we print its bit number only.
Examples:
1. Standard presentation
$ vdpa dev config show vdpa-a
vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 2 mtu 9000
negotiated_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS \
CTRL_VQ MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
2. json output
$ vdpa -j dev config show vdpa-a
{"config":{"vdpa-a":{"mac":"00:00:00:00:88:88","link":"up","link_announce":false,\
"max_vq_pairs":2,"mtu":9000,"negotiated_features":["CSUM","GUEST_CSUM",\
"MTU","MAC","HOST_TSO4","HOST_TSO6","STATUS","CTRL_VQ","MQ","CTRL_MAC_ADDR",\
"VERSION_1","ACCESS_PLATFORM"]}}}
Reviewed-by: Si-Wei Liu<si-wei.liu@oracle.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Eli Cohen <elic@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Eli Cohen [Sun, 13 Mar 2022 17:12:16 +0000 (19:12 +0200)]
vdpa: Remove unsupported command line option
"-v[erbose]" option is not supported.
Remove it.
Reviewed-by: Parav Pandit <parav@nvidia.com> Reviewed-by: Jianbo Liu <jianbol@mellanox.com> Reviewed-by: Si-Wei Liu <si-wei.liu@oracle.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Eli Cohen <elic@nvidia.com> Signed-off-by: David Ahern <dsahern@kernel.org>
David Ahern [Sat, 12 Mar 2022 15:59:03 +0000 (08:59 -0700)]
Merge branch 'bridge-broadcast-flooding' into next
Joachim Wiberg says:
====================
this patch set address a slight omission in controlling broadcast
flooding per bridge port, which the bridge has had support for a good
while now.
v3:
- Move bcast_flood option in manual files to before the mcast_flood
option, instead of breaking the two mcast options. Unfortunately
the other options are not alphabetically sorted, so this was the
least worst option. (Stephen)
- Add missing closing " for 'bridge mdb show' in bridge(8) SYNOPSIS
v2:
- Add bcast_flood also to ip/iplink_bridge_slave.c (Nik)
- Update man page for ip-link(8) with new bcast_flood flag
- Update mcast_flood in same man page slightly
- Fix minor weird whitespace issues causing sudden line breaks
v1:
- Add bcast_flood to bridge/link.c
- Update man page for bridge(8) with bcast_flood for brports
Joachim Wiberg [Wed, 9 Mar 2022 19:23:16 +0000 (20:23 +0100)]
man: ip-link: whitespace fixes to odd line breaks mid sentence
Some options, spread across the man page, were accidentally (?) space
indented (possible bullet list auto-indent in editors), causing odd line
breaks in presentation mode (emacs, nroff, etc.). This patch aligns the
multi-line descriptions to column zero, in line with other such option
descriptions.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:15 +0000 (20:23 +0100)]
man: ip-link: mention bridge port's default mcast_flood state
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:14 +0000 (20:23 +0100)]
man: ip-link: document new bcast_flood flag on bridge ports
The options are not alphabetically sorted, so placing bcast_flood right
before mcast_flood for now.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:13 +0000 (20:23 +0100)]
ip: iplink_bridge_slave: support for broadcast flooding
Add per-port support for controlling flooding of broadcast traffic.
Similar to unicast and multcast flooding that already exist.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:12 +0000 (20:23 +0100)]
man: bridge: add missing closing " in bridge show mdb
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:11 +0000 (20:23 +0100)]
man: bridge: document new bcast_flood flag for bridge ports
The bridge link options are not alphabetically sorted, so placing
bcast_flood right before mcast_flood for now.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Joachim Wiberg [Wed, 9 Mar 2022 19:23:10 +0000 (20:23 +0100)]
bridge: support for controlling flooding of broadcast per port
Add per-port support for controlling flooding of broadcast traffic.
Similar to unicast and multcast flooding that already exist.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by: David Ahern <dsahern@kernel.org>
Andrea Claudi [Tue, 8 Mar 2022 17:04:57 +0000 (18:04 +0100)]
rdma: make RES_PID and RES_KERN_NAME alternative to each other
RDMA_NLDEV_ATTR_RES_PID and RDMA_NLDEV_ATTR_RES_KERN_NAME cannot be set
together, as evident for the fill_res_name_pid() function in the kernel
infiniband driver. This commit makes this clear at first glance, using
an else branch for the RDMA_NLDEV_ATTR_RES_KERN_NAME case.
This also helps coverity to better understand this code and avoid
producing a bogus warning complaining about mnl_attr_get_str overwriting
comme, and thus leaking the storage that comm points to.
Signed-off-by: Andrea Claudi <aclaudi@redhat.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Andrea Claudi [Tue, 8 Mar 2022 17:04:56 +0000 (18:04 +0100)]
lib/fs: fix memory leak in get_task_name()
asprintf() allocates memory which is not freed on the error path of
get_task_name(), thus potentially leading to memory leaks.
%m specifier on fscanf allocates memory, too, which needs to be freed by
the caller.
This reworks get_task_name() to avoid memory allocation.
- Pass a buffer and its length to the function, similarly to what
get_command_name() does, thus avoiding to allocate memory for
the string to be returned;
- Use snprintf() instead of asprintf();
- Use fgets() instead of fscanf() to limit string length.
Fixes: 81bfd01a4c9e ("lib: move get_task_name() from rdma") Signed-off-by: Andrea Claudi <aclaudi@redhat.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Nicolas Escande [Fri, 4 Mar 2022 10:33:54 +0000 (11:33 +0100)]
ip/batadv: allow to specify RA when creating link
This patch adds the possibility to specify batadv specific options when
creating a new batman link. The only option available on link creation
is IFLA_BATADV_ALGO_NAME which specifies the routing algorithm.
Note there is no batadv specific attr to be handled on link dump.
Signed-off-by: Nicolas Escande <nico.escande@gmail.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Baowen Zheng [Fri, 4 Mar 2022 02:01:40 +0000 (10:01 +0800)]
tc: separate action print for filter and action dump
We need to separate action print for filter and action dump since
in action dump, we need to print hardware status and flags. But in
filter dump, we do not need to print action hardware status and
hardware related flags.
In filter dump, actions hardware status should be same with filter.
so we will not print action hardware status in this case.
David Ahern [Mon, 7 Feb 2022 01:00:32 +0000 (18:00 -0700)]
bpf: Remove use of bpf_create_map_xattr
bpf_create_map_xattr is deprecated in v0.7 in favor of bpf_map_create.
bpf_map_create and its bpf_map_create_opts are not available across the
range of v0.1 and up versions of libbpf, so change create_map to use
the bpf syscall directly.
David Ahern [Mon, 7 Feb 2022 00:32:43 +0000 (17:32 -0700)]
bpf_glue: Remove use of bpf_load_program from libbpf
bpf_load_program is deprecated starting in v0.7. The preferred
bpf_prog_load requires bpf_prog_load_opts from v0.6. This creates an
ugly scenario for iproute2 to work across libbpf versions from v0.1
and up.
Since bpf_program_load is only used to load the builtin vrf program,
just remove the libbpf call and use the legacy code.
David Ahern [Fri, 4 Mar 2022 16:18:24 +0000 (09:18 -0700)]
Merge branch '802.1X-locked-bridge-ports' into next
Hans Schultz says:
====================
This patch set is to complement the kernel locked port patches, such
that iproute2 can be used to lock/unlock a port and check if a port
is locked or not. To lock or unlock a port use the command:
bridge link set dev DEV locked {on | off}
To show the detailed setting of a port, including if the locked flag is
enabled for the port(s), use the command:
Geliang Tang [Wed, 23 Feb 2022 06:50:39 +0000 (14:50 +0800)]
mptcp: add port support for setting flags
This patch updated the port keyword check for the setting flags, allow
to use the port keyword with the non-signal flags. Don't allow to use
the port keyword with the id number.
With this patch, we can use setting flags in two forms, using the address
and port number directly or the id number of the address:
ip mptcp endpoint change id 1 fullmesh
ip mptcp endpoint change 10.0.2.1 fullmesh
ip mptcp endpoint change 10.0.2.1 port 10100 fullmesh
Acked-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: Geliang Tang <geliang.tang@suse.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Geliang Tang [Wed, 23 Feb 2022 06:50:38 +0000 (14:50 +0800)]
mptcp: add fullmesh support for setting flags
A pair of new flags, fullmesh and nofullmesh, had been added in the
setting flags of MPTCP PM netlink in kernel space recently by the commit 73c762c1f07d ("mptcp: set fullmesh flag in pm_netlink").
This patch added the corresponding logic to pass these two flags to the
netlink in user space.
These new flags can be used like this:
ip mptcp endpoint change id 1 fullmesh
ip mptcp endpoint change id 1 nofullmesh
ip mptcp endpoint change id 1 backup fullmesh
ip mptcp endpoint change id 1 nobackup nofullmesh
Here's an example of setting fullmesh flags:
> sudo ip mptcp endpoint add 10.0.2.1 subflow
> sudo ip mptcp endpoint show
10.0.2.1 id 1 subflow
> sudo ip mptcp endpoint change id 1 fullmesh
> sudo ip mptcp endpoint show
10.0.2.1 id 1 subflow fullmesh
> sudo ip mptcp endpoint change id 1 nofullmesh
> sudo ip mptcp endpoint show
10.0.2.1 id 1 subflow
It can be seen that 'ip mptcp endpoint show' already supports showing
the fullmesh flag.
Acked-by: Paolo Abeni <pabeni@redhat.com> Acked-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: Geliang Tang <geliang.tang@suse.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Wojciech Drewek [Thu, 10 Feb 2022 12:57:15 +0000 (13:57 +0100)]
f_flower: fix indentation for enc_key_id and u32
Commit b2450e46b7b2 ("flower: fix clang warnings") caused enc_key_id
and u32 to be printed without indentation. Fix this by printing two
spaces before calling print_uint_name_value.
Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com> Signed-off-by: David Ahern <dsahern@kernel.org>
Benjamin Poirier [Mon, 14 Feb 2022 01:44:46 +0000 (10:44 +0900)]
bridge: Remove vlan listing from `bridge link`
vlan listing from `bridge link -d` was broken by commit f32e4977dcb0 ("bridge:
add json support for link command"). print_vlan_info() expects to be passed a
IFLA_AF_SPEC attribute (as is done in print_vlan()) but that commit changed
the call in link.c to pass a IFLA_BRIDGE_VLAN_INFO attribute instead. As a
result, a struct bridge_vlan_info is mistakenly parsed as a struct rtattr and
print_vlan_info() usually exits early in this callpath.
The output style of print_vlan_info() (one line per vlan) is different from
the output style of `bridge link` (multiple attributes per line). The json
output is also unsuitable for `bridge link`. Since vlan listing is available
from `bridge vlan`, remove it from `bridge link` instead of trying to change
print_vlan_info().
Note that previously, bridge master devices would be included in the output
when specifying '-d' (and only in that case) but they are no longer
included because there is no detailed information to show for master
devices if we are not printing a vlan listing:
$ bridge link
4: vxlan0: <BROADCAST,MULTICAST> mtu 1500 master br0 state disabled priority 32 cost 100
$ bridge -d link
3: br0: <BROADCAST,MULTICAST> mtu 1500 master br0
4: vxlan0: <BROADCAST,MULTICAST> mtu 1500 master br0 state disabled priority 32 cost 100
hairpin off guard off root_block off fastleave off learning on flood on mcast_flood on mcast_to_unicast off neigh_suppress off vlan_tunnel on isolated off
$ ./bridge/bridge -d link
4: vxlan0: <BROADCAST,MULTICAST> mtu 1500 master br0 state disabled priority 32 cost 100
hairpin off guard off root_block off fastleave off learning on flood on mcast_flood on mcast_to_unicast off neigh_suppress off vlan_tunnel on isolated off
Fixes: f32e4977dcb0 ("bridge: add json support for link command") Signed-off-by: Benjamin Poirier <bpoirier@nvidia.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Benjamin Poirier [Mon, 14 Feb 2022 01:44:45 +0000 (10:44 +0900)]
bridge: Fix error string typo
Fixes: fab9a18a2e52 ("bridge: request vlans along with link information") Signed-off-by: Benjamin Poirier <bpoirier@nvidia.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Maxim Petrov [Tue, 15 Feb 2022 20:53:47 +0000 (23:53 +0300)]
lnstat: fix strdup leak in -w argument parsing
'tmp' string is used for safe tokenizing, but it is not required after
getting all the widths in -w option. As 'tmp' string is obtained by strdup
call, the caller has to deallocate it to avoid memory leak.
Signed-off-by: Maxim Petrov <mmrmaximuzz@gmail.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
The can help menu misses three spaces for the TDCV, TDCO and TDCF
parameters, making the closing curly bracket unaligned.
For reference, before this patch:
| $ ip link help can
| Usage: ip link set DEVICE type can
| [ bitrate BITRATE [ sample-point SAMPLE-POINT] ] |
| [ tq TQ prop-seg PROP_SEG phase-seg1 PHASE-SEG1
| phase-seg2 PHASE-SEG2 [ sjw SJW ] ]
|
| [ dbitrate BITRATE [ dsample-point SAMPLE-POINT] ] |
| [ dtq TQ dprop-seg PROP_SEG dphase-seg1 PHASE-SEG1
| dphase-seg2 PHASE-SEG2 [ dsjw SJW ] ]
| [ tdcv TDCV tdco TDCO tdcf TDCF ]
|
| [ loopback { on | off } ]
| [ listen-only { on | off } ]
| [ triple-sampling { on | off } ]
| [ one-shot { on | off } ]
| [ berr-reporting { on | off } ]
| [ fd { on | off } ]
| [ fd-non-iso { on | off } ]
| [ presume-ack { on | off } ]
| [ cc-len8-dlc { on | off } ]
| [ tdc-mode { auto | manual | off } ]
|
| [ restart-ms TIME-MS ]
| [ restart ]
|
| [ termination { 0..65535 } ]
|
| Where: BITRATE := { NUMBER in bps }
| SAMPLE-POINT := { 0.000..0.999 }
| TQ := { NUMBER in ns }
| PROP-SEG := { NUMBER in tq }
| PHASE-SEG1 := { NUMBER in tq }
| PHASE-SEG2 := { NUMBER in tq }
| SJW := { NUMBER in tq }
| TDCV := { NUMBER in tc}
| TDCO := { NUMBER in tc}
| TDCF := { NUMBER in tc}
| RESTART-MS := { 0 | NUMBER in ms }
... and after this patch:
| $ ip link help can
| Usage: ip link set DEVICE type can
| [ bitrate BITRATE [ sample-point SAMPLE-POINT] ] |
| [ tq TQ prop-seg PROP_SEG phase-seg1 PHASE-SEG1
| phase-seg2 PHASE-SEG2 [ sjw SJW ] ]
|
| [ dbitrate BITRATE [ dsample-point SAMPLE-POINT] ] |
| [ dtq TQ dprop-seg PROP_SEG dphase-seg1 PHASE-SEG1
| dphase-seg2 PHASE-SEG2 [ dsjw SJW ] ]
| [ tdcv TDCV tdco TDCO tdcf TDCF ]
|
| [ loopback { on | off } ]
| [ listen-only { on | off } ]
| [ triple-sampling { on | off } ]
| [ one-shot { on | off } ]
| [ berr-reporting { on | off } ]
| [ fd { on | off } ]
| [ fd-non-iso { on | off } ]
| [ presume-ack { on | off } ]
| [ cc-len8-dlc { on | off } ]
| [ tdc-mode { auto | manual | off } ]
|
| [ restart-ms TIME-MS ]
| [ restart ]
|
| [ termination { 0..65535 } ]
|
| Where: BITRATE := { NUMBER in bps }
| SAMPLE-POINT := { 0.000..0.999 }
| TQ := { NUMBER in ns }
| PROP-SEG := { NUMBER in tq }
| PHASE-SEG1 := { NUMBER in tq }
| PHASE-SEG2 := { NUMBER in tq }
| SJW := { NUMBER in tq }
| TDCV := { NUMBER in tc }
| TDCO := { NUMBER in tc }
| TDCF := { NUMBER in tc }
| RESTART-MS := { 0 | NUMBER in ms }
Fixes: 0c263d7c36ff ("iplink_can: add new CAN FD bittiming parameters:
Transmitter Delay Compensat ion (TDC)") Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr> Acked-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Petr Machata [Wed, 9 Feb 2022 14:41:40 +0000 (15:41 +0100)]
dcb: Fix error reporting when accessing "dcb app"
Currently dcb decodes the response from "dcb app add" and "del" by
interpreting the returned attribute as u8. But the value stored there is
actually a negative errno value.
Additionally, "dcb app" currently shows two sets of messages, one in
dcb_set_attribute_attr_cb() where the issue is detected, and another as a
result of error return from that function.
The current state is as follows:
# dcb app add dev swp36 dscp-prio 20:2
Error when attempting to set attribute: Unknown error 239
Attribute write: No such file or directory
Fix the "unknown error" issue by correctly decoding the attribute as i8 and
negating it. Furthermore, set errno to that value, and let the top-level
"attribute write" error message show the correct message.
Initialize errno to 0 before the dcb_talk() dispatch, and make the error
print conditional on errno != 0. This way the few error messages that are
worth describing in the place where they are detected will not cause the
second error message to be printed.
Maxim Petrov [Tue, 8 Feb 2022 17:20:45 +0000 (20:20 +0300)]
libnetlink: fix socket leak in rtnl_open_byproto()
rtnl_open_byproto() does not close the opened socket in case of
errors, and the socket is returned to the caller in the `fd` field of
the struct. However, none of the callers care about the socket, so
close it in the function immediately to avoid any potential resource
leaks.
Signed-off-by: Maxim Petrov <mmrmaximuzz@gmail.com> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
projects / thirdparty / iproute2.git / log
