This patch converts iproute2 to use libbpf for loading and attaching
BPF programs when it is available, which is started by Toke's
implementation[1]. With libbpf iproute2 could correctly process BTF
information and support the new-style BTF-defined maps, while keeping
compatibility with the old internal map definition syntax.
The old iproute2 bpf code is kept and will be used if no suitable libbpf
is available. When using libbpf, wrapper code in bpf_legacy.c ensures that
iproute2 will still understand the old map definition format, including
populating map-in-map and tail call maps before load.
In bpf_libbpf.c, we init iproute2 ctx and elf info first to check the
legacy bytes. When handling the legacy maps, for map-in-maps, we create
them manually and re-use the fd as they are associated with id/inner_id.
For pin maps, we only set the pin path and let libbp load to handle it.
For tail calls, we find it first and update the element after prog load.
Other maps/progs will be loaded by libbpf directly.
[1] https://lore.kernel.org/bpf/20190820114706.18546-1-toke@redhat.com/
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Hangbin Liu <haliu@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
There are directly calls in libbpf for bpf program load/attach.
So we could just use two wrapper functions for ipvrf and convert
them with libbpf support.
Function bpf_prog_load() is removed as it's conflict with libbpf
function name.
bpf.c is moved to bpf_legacy.c for later main libbpf support in
iproute2.
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Hangbin Liu <haliu@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
This patch aim to add basic checking functions for later iproute2
libbpf support.
First we add check_libbpf() in configure to see if we have bpf library
support. By default the system libbpf will be used, but static linking
against a custom libbpf version can be achieved by passing libbpf DESTDIR
to variable LIBBPF_DIR for configure.
Another variable LIBBPF_FORCE is used to control whether to build iproute2
with libbpf. If set to on, then force to build with libbpf and exit if
not available. If set to off, then force to not build with libbpf.
When dynamically linking against libbpf, we can't be sure that the
version we discovered at compile time is actually the one we are
using at runtime. This can lead to hard-to-debug errors. So we add
a new file lib/bpf_glue.c and a helper function get_libbpf_version()
to get correct libbpf version at runtime.
Signed-off-by: Hangbin Liu <haliu@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
The DCB tool will have to provide an interface to a number of fixed-size
arrays. Unlike the egress- and ingress-qos-map, it makes good sense to have
an interface to set all members to the same value. For example to set
strict priority on all TCs besides select few, or to reset allocated
bandwidth to all zeroes, again besides several explicitly-given ones.
To support this usage, extend the parse_mapping() with a boolean that
determines whether this special use is supported. If "all" is given and
recognized, mapping_cb is called with the key of -1.
Have iplink_vlan pass false for allow_all.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
Currently argc and argv are not updated unless parsing of all of the
mapping was successful. However in that case, "ip link" will point at the
wrong argument when complaining:
# ip link add name eth0.100 link eth0 type vlan id 100 egress 1:1 2:foo
Error: argument "1" is wrong: invalid egress-qos-map
Update argc and argv even in the case of parsing error, so that the right
element is indicated.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
VLAN netdevices have two similar attributes: ingress-qos-map and
egress-qos-map. These attributes can be configured with a series of
802.1-priority-to-skb-priority (and vice versa) mappings. A reusable helper
along those lines will be handy for configuration of various
priority-to-tc, tc-to-algorithm, and other arrays in DCB.
Therefore extract the logic to a function parse_mapping(), move to utils.c,
and dispatch to utils.c from iplink_vlan.c. That necessitates extraction of
a VLAN-specific parse_qos_mapping(). Do that, and propagate addattr_l()
return value up, unlike the original.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
Receiving a message in libmnl is a somewhat involved operation. Devlink's
mnlg library has an implementation that is going to be handy for other
tools as well. Extract it into a new helper.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
Allocation of a new netlink message with the two usual headers is reusable
with other netlink netlink message types. Extract it into a helper,
mnlu_msg_prepare(). Take the second header as an argument, instead of
passing in parameters to initialize it, and copy it in.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
This little dance of mnl_socket_open(), option setting, and bind, is the
same regardless of tool. Extract into a new module that should hold helpers
for working with libmnl, mnl_util.c.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
The value of a number of booleans is shown as "on" and "off" in the plain
output, and as an actual boolean in JSON mode. Add a function that does
that.
RDMA tool already uses a function named print_on_off(). This function
always shows "on" and "off", even in JSON mode. Since there are probably
very few if any consumers of this interface at this point, migrate it to
the new central print_on_off() as well.
Signed-off-by: Petr Machata <me@pmachata.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Take from the macsec code parse_one_of() and adapt so that it passes the
primary result as the main return value, and error result through a
pointer. That is the simplest way to make the code reusable across data
types without introducing extra magic.
Also from macsec take the specialization of parse_one_of() for parsing
specifically the strings "off" and "on".
Convert the macsec code to the new helpers.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
The code for handling batches is largely the same across iproute2 tools.
Extract a helper to handle the batch, and adjust the tools to dispatch to
this helper. Sandwitch the invocation between prologue / epilogue code
specific for each tool.
Signed-off-by: Petr Machata <me@pmachata.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
On some systems (e.g. current Debian/stable) the inclusion
of utils.h pulled in some other things that may end up
defining __aligned, in a possibly different way than what
we had here.
Use our own definition only if there isn't one already.
Fixes: d5acae244f ("libnetlink: add nl_print_policy() helper")
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Add support for the new TCA_MPLS_ACT_MAC_PUSH action (kernel commit
a45294af9e96 ("net/sched: act_mpls: Add action to push MPLS LSE before
Ethernet header")). This action let TC push an MPLS header before the
MAC header of a frame.
Example (encapsulate all outgoing frames with label 20, then add an
outer Ethernet header):
# tc filter add dev ethX matchall \
action mpls mac_push label 20 ttl 64 \
action vlan push_eth dst_mac 0a:00:00:00:00:02 \
src_mac 0a:00:00:00:00:01
This patch also adds an alias for ETH_P_TEB, since it is useful when
decapsulating MPLS packets that contain an Ethernet frame.
With MAC_PUSH, there's no previous Ethertype to modify. However, the
"protocol" option is still needed, because the kernel uses it to set
skb->protocol. So rename can_modify_ethtype() to can_set_ethtype().
Also add a test suite for m_mpls, which covers the new action and the
pre-existing ones.
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
In case of bad entries in /proc/mounts just skip cgroup cache initialization.
Cgroups in output will be shown as "unreachable:cgroup_id".
Fixes: d5e6ee0dac ("ss: introduce cgroup2 cache and helper functions")
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
Reported-by: Donald Sharp <sharpd@nvidia.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This patch adds support for recently
added link IFLA_PROTO_DOWN_REASON attribute.
IFLA_PROTO_DOWN_REASON enumerates reasons
for the already existing IFLA_PROTO_DOWN link
attribute.
$ cat /etc/iproute2/protodown_reasons.d/r.conf
0 mlag
1 evpn
2 vrrp
3 psecurity
$ ip link set dev vx10 protodown on protodown_reason vrrp on
$ip link show dev vx10
14: vx10: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode
DEFAULT group default qlen 1000
link/ether f2:32:28:b8:35:ff brd ff:ff:ff:ff:ff:ff protodown on
protodown_reason <vrrp>
$ip -p -j link show dev vx10
[ {
<snip>
"proto_down": true,
"proto_down_reason": [ "vrrp" ]
} ]
$ip link set dev vx10 protodown_reason mlag on
$ip link show dev vx10
14: vx10: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode
DEFAULT group default qlen 1000
link/ether f2:32:28:b8:35:ff brd ff:ff:ff:ff:ff:ff protodown on
protodown_reason <mlag,vrrp>
$ip -p -j link show dev vx10
[ {
<snip>
"proto_down": true,
"protodown_reason": [ "mlag","vrrp" ]
} ]
$ip -p -j link show dev vx10
$ip link set dev vx10 protodown off protodown_reason vrrp off
Error: Cannot clear protodown, active reasons.
$ip link set dev vx10 protodown off protodown_reason mlag off
$
Note: for somereason the json and non-json key for protodown
are different (protodown and proto_down). I have kept the
same for protodown reason for consistency (protodown_reason and
proto_down_reason).
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
This prints out the data from the given nested attribute
to the given FILE pointer, interpreting the firmware that
the kernel has for showing netlink policies.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David Ahern <dsahern@gmail.com>
Before this patch check is happened only in case when we try to find
cgroup at cgroup2 mount point.
v2:
- add Fixes line before Signed-off-by (David Ahern)
Fixes: d5e6ee0dac ("ss: introduce cgroup2 cache and helper functions")
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Following inclusion in net-next, extend rtnl_rtprot_tab and rt_protos
to support Keepalived.
Signed-off-by: Alexandre Cassen <acassen@gmail.com>
Signed-off-by: David Ahern <dsahern@kernel.org>
gcc v9.3.1 reports:
bpf.c: In function ‘bpf_get_work_dir’:
bpf.c:784:49: warning: ‘snprintf’ output may be truncated before the last format character [-Wformat-truncation=]
784 | snprintf(bpf_wrk_dir, sizeof(bpf_wrk_dir), "%s/", mnt);
| ^
bpf.c:784:2: note: ‘snprintf’ output between 2 and 4097 bytes into a destination of size 4096
784 | snprintf(bpf_wrk_dir, sizeof(bpf_wrk_dir), "%s/", mnt);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fix this simply checking snprintf return code and properly handling the error.
Fixes: e42256699c ("bpf: make tc's bpf loader generic and move into lib")
Signed-off-by: Andrea Claudi <aclaudi@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This reverts commit c0325b0638.
It introduces a segfault in bpf_make_custom_path() when custom pinning is used.
This happens because asprintf allocates exactly the space needed to hold a
string in the buffer passed as its first argument, but if this buffer is later
used in strcat() or similar we have a buffer overrun.
As the aim of commit c0325b0638 is simply to fix a compiler warning, it
seems safe and reasonable to revert it.
Fixes: c0325b0638 ("bpf: replace snprintf with asprintf when dealing with long buffers")
Reported-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Andrea Claudi <aclaudi@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This patch prepares infrastructure for matching sockets by cgroups.
Two helper functions are added for transformation between cgroup v2 ID
and pathname. Cgroup v2 cache is implemented as hash table indexed by ID.
This cache is needed for faster lookups of socket cgroup.
v2:
- style fixes (David Ahern)
Signed-off-by: Dmitry Yakunin <zeil@yandex-team.ru>
Signed-off-by: David Ahern <dsahern@gmail.com>
Currently new json object opens (and delete_json_obj closes) the object as
an array, what adds prints for the matching bracket '[' ']' at the
start/end of the object. This patch adds new_json_obj_plain() and the
matching delete_json_obj_plain() to enable opening and closing json object,
not as array and leave it to the using function to decide which type of
object to open/close as the main object.
Signed-off-by: Ron Diskin <rondi@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Until now print_#type functions supported printing constant names and
unknown (variable) values only.
Add functions to allow printing when the name is also sent to the
function as a variable.
Signed-off-by: Ron Diskin <rondi@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
fread(3) returns size_t data type which is unsigned, thus check
`if (fread(...) < 0)' is always false. To check if fread(3) has
failed, user should check error indicator with ferror(3).
This commit also changes read logic a little bit by being less
forgiving for errors. Previous logic was checking if fread(3)
read *at least* required ammount of data, now code checks if
fread(3) read *exactly* expected ammount of data. This makes
sense because code parses very specific binary file, and reading
even 1 less/more byte than expected, will later corrupt data anyway.
Signed-off-by: Michał Łyszczek <michal.lyszczek@bofc.pl>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Extend ll_name_to_index() to get the index of a netdevice using
alternative interface name. Allow alternative long names to pass checks
in couple of ip link/addr commands.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Implement addition/deletion of lists of properties, currently
alternative ifnames. Also extent the ip link show command to list them.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Alternative names are related to the "parent name". That means,
whenever ll_remember_index() is called to add/delete/update and it founds
the "parent name" im object by ifindex, processes related
alternative name im objects too. Put them in a list which holds the
relationship with the parent.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
If two processes attempt to invoke bpf_map_attach() at the same time,
then they will both create maps, then the first will successfully pin
the map to the filesystem and the second will not pin the map, but will
continue operating with a reference to its own copy of the map. As a
result, the sharing of the same map will be broken from the two programs
that were concurrently loaded via loaders using this library.
Fix this by adding a retry in the case where the pinning fails because
the map already exists on the filesystem. In that case, re-attempt
opening a fd to the map on the filesystem as it shows that another
program already created and pinned a map at that location.
Signed-off-by: Joe Stringer <joe@wand.net.nz>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This reduces stack usage, as asprintf allocates memory on the heap.
This indirectly fixes a snprintf truncation warning (from gcc v9.2.1):
bpf.c: In function ‘bpf_get_work_dir’:
bpf.c:784:49: warning: ‘snprintf’ output may be truncated before the last format character [-Wformat-truncation=]
784 | snprintf(bpf_wrk_dir, sizeof(bpf_wrk_dir), "%s/", mnt);
| ^
bpf.c:784:2: note: ‘snprintf’ output between 2 and 4097 bytes into a destination of size 4096
784 | snprintf(bpf_wrk_dir, sizeof(bpf_wrk_dir), "%s/", mnt);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fixes: e42256699c ("bpf: make tc's bpf loader generic and move into lib")
Signed-off-by: Andrea Claudi <aclaudi@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
get_s64() uses internally strtoll() to parse the value out of a given
string. strtoll() returns a long long. However, the intermediate variable is
long only which might be 32 bit on some systems. So, fix it.
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
iproute has an utility function which checks if a string is a prefix for
another one, to allow use of abbreviated commands, e.g. 'addr' or 'a'
instead of 'address'.
This routine unfortunately considers an empty string as prefix
of any pattern, leading to undefined behaviour when an empty
argument is passed to ip:
# ip ''
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
# tc ''
qdisc noqueue 0: dev lo root refcnt 2
# ip address add 192.0.2.0/24 '' 198.51.100.1 dev dummy0
# ip addr show dev dummy0
6: dummy0: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN group default qlen 1000
link/ether 02:9d:5e:e9:3f:c0 brd ff:ff:ff:ff:ff:ff
inet 192.0.2.0/24 brd 198.51.100.1 scope global dummy0
valid_lft forever preferred_lft forever
Rewrite matches() so it takes care of an empty input, and doesn't
scan the input strings three times: the actual implementation
does 2 strlen and a memcpy to accomplish the same task.
Signed-off-by: Matteo Croce <mcroce@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Update the llproto_names array to allow users to reference the mpls
protocol ids with the names 'mpls_uc' for unicast MPLS and 'mpls_mc' for
multicast.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
The netns_{save,restore} functions are only used in ipnetns.c now, since
the restore is not needed anymore after the netns exec command.
Move them in ipnetns.c, and make them static.
Signed-off-by: Matteo Croce <mcroce@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
'ip netns exec' changes the current netns just before executing a child
process, and restores it after forking. This is needed if we're running
in batch or do_all mode.
Some cleanups must be done both in the parent and in the child: the
parent must restore the previous netns, while the child must reset any
VRF association.
Unfortunately, if do_all is set, the VRF are not reset in the child, and
the spawned processes are started with the wrong VRF context. This can
be triggered with this script:
# ip -b - <<-'EOF'
link add type vrf table 100
link set vrf0 up
link add type dummy
link set dummy0 vrf vrf0 up
netns add ns1
EOF
# ip -all -b - <<-'EOF'
vrf exec vrf0 true
netns exec setsid -f sleep 1h
EOF
# ip vrf pids vrf0
314 sleep
# ps 314
PID TTY STAT TIME COMMAND
314 ? Ss 0:00 sleep 1h
Refactor cmd_exec() and pass to it a function pointer which is called in
the child before the final exec. In the netns exec case the function just
resets the VRF and switches netns.
Doing it in the child is less error prone and safer, because the parent
environment is always kept unaltered.
After this refactor some utility functions became unused, so remove them.
Signed-off-by: Matteo Croce <mcroce@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
