]> git.proxmox.com Git - mirror_iproute2.git/log
mirror_iproute2.git
5 years agotc/pedit: make functions static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:34 +0000 (14:36 -0800)]
tc/pedit: make functions static

The parse and pack functions are only used by the pedit routines.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoss: make local variables static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:33 +0000 (14:36 -0800)]
ss: make local variables static

Several variables only used in this code.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc/police: make print_police static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:32 +0000 (14:36 -0800)]
tc/police: make print_police static

print_police function only used by m_police.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc/class: make filter variables static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:31 +0000 (14:36 -0800)]
tc/class: make filter variables static

Only used in this file.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotipc: make cmd_find static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:30 +0000 (14:36 -0800)]
tipc: make cmd_find static

Function only used in one file.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc: drop unused name_to_id function
Stephen Hemminger [Thu, 15 Nov 2018 22:36:29 +0000 (14:36 -0800)]
tc: drop unused name_to_id function

Not used in current code.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoipxfrm: make local functions static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:28 +0000 (14:36 -0800)]
ipxfrm: make local functions static

Make functions only used in ipxfrm.c static.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoipmonitor: make local variable static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:27 +0000 (14:36 -0800)]
ipmonitor: make local variable static

prefix_banner only used in one file.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip: make flag names const/static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:26 +0000 (14:36 -0800)]
ip: make flag names const/static

The table of filter flags is only used in ipaddress

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agobridge: make local variables static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:25 +0000 (14:36 -0800)]
bridge: make local variables static

enable_color and set_color_palette only used here.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agogenl: remove dead code
Stephen Hemminger [Thu, 15 Nov 2018 22:36:24 +0000 (14:36 -0800)]
genl: remove dead code

The function genl_ctrl_resolve_family is defined but never used
in current code.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolibnetlnk: unused and local functions cleanup
Stephen Hemminger [Thu, 15 Nov 2018 22:36:23 +0000 (14:36 -0800)]
libnetlnk: unused and local functions cleanup

rntl_talk_extack and parse_rtattr_index not used in current code.
rtnl_dump_filter_l is only used in this file.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolib/ll_map: make local function static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:22 +0000 (14:36 -0800)]
lib/ll_map: make local function static

ll_idx_a2n is only used in ll_map.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolib/color: make local functions static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:21 +0000 (14:36 -0800)]
lib/color: make local functions static

color_enable etc, only used here.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolib/utils: make local functions static
Stephen Hemminger [Thu, 15 Nov 2018 22:36:20 +0000 (14:36 -0800)]
lib/utils: make local functions static

Some of the print/parsing is only used internally.
Drop unused get_s8/get_s16.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolib/ll_addr: whitespace and indent cleanup
Stephen Hemminger [Thu, 15 Nov 2018 22:36:19 +0000 (14:36 -0800)]
lib/ll_addr: whitespace and indent cleanup

Run old ll_addr through kernel Lindent.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'master' of ra.kernel.org:/pub/scm/linux/kernel/git/shemminger/iproute2
Stephen Hemminger [Mon, 19 Nov 2018 19:40:37 +0000 (11:40 -0800)]
Merge branch 'master' of ra.kernel.org:/pub/scm/linux/kernel/git/shemminger/iproute2

5 years agoip-address: Fix filtering by negated address flags
Phil Sutter [Thu, 15 Nov 2018 13:28:59 +0000 (14:28 +0100)]
ip-address: Fix filtering by negated address flags

When disabling a flag, one needs to AND with the inverse not the flag
itself. Otherwise specifying for instance 'home -nodad' will effectively
clear the flags variable.

While being at it, simplify the code a bit by merging common parts of
negated and non-negated case branches. Also allow for the "special
cases" to be inverted, too.

Fixes: f73ac674d0abf ("ip: change flag names to an array")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: drop old kernel configs
Stephen Hemminger [Mon, 19 Nov 2018 19:37:08 +0000 (11:37 -0800)]
testsuite: drop old kernel configs

The testsuite directory had several ancient kernel configs.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip-route: Fix nexthop encap parsing
Phil Sutter [Tue, 13 Nov 2018 12:39:04 +0000 (13:39 +0100)]
ip-route: Fix nexthop encap parsing

When parsing nexthop parameters, a buffer of 4k bytes is provided. Yet,
in lwt_parse_encap() and some functions called by it, buffer size was
assumed to be 1k despite the actual size was provided. This led to
spurious buffer size errors if the buffer was filled by previous nexthop
parameters to exceed that 1k boundary.

Fixes: 1e5293056a02c ("lwtunnel: Add encapsulation support to ip route")
Fixes: 5866bddd9aa9e ("ila: Add support for ILA lwtunnels")
Fixes: ed67f83806538 ("ila: Support for checksum neutral translation")
Fixes: 86905c8f057c0 ("ila: support for configuring identifier and hook types")
Fixes: b15f440e78373 ("lwt: BPF support for LWT")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: colorize test result output
Stephen Hemminger [Mon, 12 Nov 2018 16:39:32 +0000 (08:39 -0800)]
testsuite: colorize test result output

When running testsuite it is easy as a human to miss failure.
Add symbol colorizing to SKIPED/PASS/FAIL output.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoman: ip-route.8: Document nexthop limit
Phil Sutter [Tue, 13 Nov 2018 15:55:13 +0000 (16:55 +0100)]
man: ip-route.8: Document nexthop limit

Add a note to 'nexthop' description stating the maximum number of
nexthops per command and pointing at 'append' command as a workaround.

Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: ss: Fix spacing in expected output for ssfilter.t
Stefano Brivio [Sat, 10 Nov 2018 09:21:59 +0000 (10:21 +0100)]
testsuite: ss: Fix spacing in expected output for ssfilter.t

Since commit 00240899ec0b ("ss: Actually print left delimiter for
columns") changes spacing in ss output, we also need to adjust for that in
the ss filter test.

Fixes: 00240899ec0b ("ss: Actually print left delimiter for columns")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: correctly use CC macros for generate_nlmsg
Luca Boccassi [Fri, 9 Nov 2018 16:31:36 +0000 (16:31 +0000)]
testsuite: correctly use CC macros for generate_nlmsg

It's $(QUIET_CC)$(CC) not $(QUIET_CC), copy-paste error. CI does
verbose build so it slipped through.

Fixes: 6e7d347aabbb ("testsuite: build generate_nlmsg with QUIET_CC")
Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agouapi: sctp header change
Stephen Hemminger [Fri, 9 Nov 2018 16:18:26 +0000 (08:18 -0800)]
uapi: sctp header change

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc: f_u32: allow skip_hw and skip_sw flags to be last
Jakub Kicinski [Tue, 6 Nov 2018 03:23:27 +0000 (19:23 -0800)]
tc: f_u32: allow skip_hw and skip_sw flags to be last

u32 uses NEXT_ARG() incorrectly when parsing skip_hw and skip_sw
flags.  NEXT_ARG() ensures there is another argument on the command
line, and is used in handling <keyword> <value> syntax to move past
<keyword> and ensure there is a <value> to read.

Commit 5e5b3008d1fb ("tc: f_u32: Add support for skip_hw and skip_sw
flags") seems to have copy pasted the handling from the previous
command - "police", which needs an extra parameter and is kind of
special due to the use of parse_police() helper.

The combination of NEXT_ARG() and continue worked fine as long as
skip_sw/skip_hw wasn't last, e.g.:

$ tc filter add dev dummy0 ingress prio 101 protocol ipv6 \
    u32 match ip6 priority 0xa0 0xe0 skip_hw action pass

But would fail if it was last:

$ tc filter add dev dummy0 ingress prio 101 protocol ipv6 \
    u32 match ip6 priority 0xa0 0xe0 flowid :1 skip_hw
Command line is not complete. Try option "help"

Remove the NEXT_ARG()s and the continues, and let the argc--; argv++;
at the end of the loop do its job.

Fixes: 5e5b3008d1fb ("tc: f_u32: Add support for skip_hw and skip_sw flags")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoPass CPPFLAGS to the compiler
Luca Boccassi [Fri, 2 Nov 2018 12:35:44 +0000 (12:35 +0000)]
Pass CPPFLAGS to the compiler

When building Debian packages pre-processor flags are passed via
CPPFLAGS, as the convention indicates. Specifically, the hardening
-D_FORTIFY_SOURCE=2 flag is used.
Pass CPPFLAGS to all calls of QUIET_CC together with CFLAGS.

Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: build generate_nlmsg with QUIET_CC
Luca Boccassi [Fri, 2 Nov 2018 12:35:43 +0000 (12:35 +0000)]
testsuite: build generate_nlmsg with QUIET_CC

Follow the standard pattern, and respect user's verbosity setting.

Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agodevlink: Add missing region option to devlink man page
Alex Vesker [Thu, 8 Nov 2018 09:14:13 +0000 (11:14 +0200)]
devlink: Add missing region option to devlink man page

The region field was not added to the devlink man page.

Fixes: 8b4fbf0bed8e6 ("devlink: Add support for devlink-region access")
Signed-off-by: Alex Vesker <valex@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoFix warning in tc-skbprio.8 manpage
Luca Boccassi [Fri, 2 Nov 2018 15:27:31 +0000 (15:27 +0000)]
Fix warning in tc-skbprio.8 manpage

". If" gets interpreted as a macro, so move the period to the previous
line:

  33: warning: macro `If' not defined

Fixes: 141b55f8544e ("Add SKB Priority qdisc support in tc(8)")
Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoman: ss.8: break and indent long line
Luca Boccassi [Fri, 2 Nov 2018 13:27:48 +0000 (13:27 +0000)]
man: ss.8: break and indent long line

Fixes groff warning:
  ss.8 92: warning [p 2, 2.8i]: can't break line

And makes the line also more readable.

Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agobridge: fdb: remove redundant dev string in show output
Roopa Prabhu [Wed, 7 Nov 2018 23:14:09 +0000 (15:14 -0800)]
bridge: fdb: remove redundant dev string in show output

After commit 4abb8c723a64 ("bridge: fdb: Fix for missing
keywords in non-JSON output"), I am seeing a double print for dev
in bridge fdb show. eg:
"44:38:39:00:6a:82 dev dev bridge vlan 1 master bridge permanent"

this patch removes the redundant print.

Fixes: 4abb8c723a64 ("bridge: fdb: Fix for missing keywords in non-JSON output")
CC: Phil Sutter <phil@nwl.cc>
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Acked-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoInclude bsd/string.h only in include/utils.h
Luca Boccassi [Thu, 1 Nov 2018 22:25:27 +0000 (22:25 +0000)]
Include bsd/string.h only in include/utils.h

This is simpler and cleaner, and avoids having to include the header
from every file where the functions are used. The prototypes of the
internal implementation are in this header, so utils.h will have to be
included anyway for those.

Fixes: 508f3c231efb ("Use libbsd for strlcpy if available")
Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agouapi: update headers to 4.20-rc1
Stephen Hemminger [Mon, 5 Nov 2018 16:37:41 +0000 (08:37 -0800)]
uapi: update headers to 4.20-rc1

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agordma: Refresh help section of resource information
Leon Romanovsky [Thu, 1 Nov 2018 08:35:05 +0000 (10:35 +0200)]
rdma: Refresh help section of resource information

After commit 4060e4c0d257 ("rdma: Add PD resource tracking
information"), the resource information shows PDs and MRs,
but help pages didn't fully reflect it.

Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip rule: Require at least one argument for add
David Ahern [Tue, 30 Oct 2018 20:59:05 +0000 (13:59 -0700)]
ip rule: Require at least one argument for add

'ip rule add' with no additional arguments just adds another rule
for the main table - which exists by default. Require at least
1 argument similar to delete.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip rule: Honor filter arguments on flush
David Ahern [Tue, 30 Oct 2018 20:58:56 +0000 (13:58 -0700)]
ip rule: Honor filter arguments on flush

'ip ru flush' currently removes all rules with priority > 0 regardless
of any other command line arguments passed in. Update flush_rule to
call filter_nlmsg to determine if the rule should be flushed or not.
This enables rule flushing such as 'ip ru flush table 1001' and
'ip ru flush pref 99'.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoUse libbsd for strlcpy if available
Luca Boccassi [Wed, 31 Oct 2018 18:00:11 +0000 (18:00 +0000)]
Use libbsd for strlcpy if available

If libc does not provide strlcpy check for libbsd with pkg-config to
avoid relying on inline version.

Signed-off-by: Luca Boccassi <bluca@debian.org>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agobpf: check map symbol type properly with newer llvm compiler
Yonghong Song [Mon, 29 Oct 2018 22:32:03 +0000 (15:32 -0700)]
bpf: check map symbol type properly with newer llvm compiler

With llvm 7.0 or earlier, the map symbol type is STT_NOTYPE.
  -bash-4.4$ cat t.c
  __attribute__((section("maps"))) int g;
  -bash-4.4$ clang -target bpf -O2 -c t.c
  -bash-4.4$ readelf -s t.o

  Symbol table '.symtab' contains 2 entries:
     Num:    Value          Size Type    Bind   Vis      Ndx Name
       0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
       1: 0000000000000000     0 NOTYPE  GLOBAL DEFAULT    3 g

The following llvm commit enables BPF target to generate
proper symbol type and size.
  commit bf6ec206615b9718869d48b4e5400d0c6e3638dd
  Author: Yonghong Song <yhs@fb.com>
  Date:   Wed Sep 19 16:04:13 2018 +0000

      [bpf] Symbol sizes and types in object file

      Clang-compiled object files currently don't include the symbol sizes and
      types.  Some tools however need that information.  For example, ctfconvert
      uses that information to generate FreeBSD's CTF representation from ELF
      files.
      With this patch, symbol sizes and types are included in object files.

Signed-off-by: Paul Chaignon <paul.chaignon@orange.com>
Reported-by: Yutaro Hayakawa <yhayakawa3720@gmail.com>
Hence, for llvm 8.0.0 (currently trunk), symbol type will be not NOTYPE, but OBJECT.
  -bash-4.4$ clang -target bpf -O2 -c t.c
  -bash-4.4$ readelf -s t.o

  Symbol table '.symtab' contains 3 entries:
     Num:    Value          Size Type    Bind   Vis      Ndx Name
       0: 0000000000000000     0 NOTYPE  LOCAL  DEFAULT  UND
       1: 0000000000000000     0 FILE    LOCAL  DEFAULT  ABS t.c
       2: 0000000000000000     4 OBJECT  GLOBAL DEFAULT    3 g

This patch makes sure bpf library accepts both NOTYPE and OBJECT types
of global map symbols.

Signed-off-by: Yonghong Song <yhs@fb.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoss: Actually print left delimiter for columns
Stefano Brivio [Mon, 29 Oct 2018 22:04:25 +0000 (23:04 +0100)]
ss: Actually print left delimiter for columns

While rendering columns, we use a local variable to keep track of the
field currently being printed, without touching current_field, which is
used for buffering.

Use the right pointer to access the left delimiter for the current column,
instead of always printing the left delimiter for the last buffered field,
which is usually an empty string.

This fixes an issue especially visible on narrow terminals, where some
columns might be displayed without separation.

Reported-by: YoyPa <yoann.p.public@gmail.com>
Fixes: 691bd854bf4a ("ss: Buffer raw fields first, then render them as a table")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Tested-by: YoyPa <yoann.p.public@gmail.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agobridge: fix vlan show stats formatting
Tobias Jungel [Fri, 26 Oct 2018 21:51:03 +0000 (23:51 +0200)]
bridge: fix vlan show stats formatting

The output of -statistics vlan show was broken previous change for json
output. This aligns the format to vlan show.

v2: fixed too greedy deletion that caused a -Wmaybe-uninitialized

Signed-off-by: Tobias Jungel <tobias.jungel@gmail.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoutils.h: provide fallback CLOCK_TAI definition
Peter Korsgaard [Sat, 27 Oct 2018 15:31:02 +0000 (17:31 +0200)]
utils.h: provide fallback CLOCK_TAI definition

q_{etf,taprio}.c uses CLOCK_TAI, which isn't exposed by glibc < 2.21 or
uClibc, breaking the build. Provide a fallback definition like it is done
for IPPROTO_MPLS and others.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'iproute2-master' into iproute2-next
David Ahern [Tue, 23 Oct 2018 17:55:09 +0000 (10:55 -0700)]
Merge branch 'iproute2-master' into iproute2-next

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoip/geneve: fix ttl inherit behavior
Hangbin Liu [Mon, 22 Oct 2018 07:46:18 +0000 (15:46 +0800)]
ip/geneve: fix ttl inherit behavior

Currently when we add geneve with "ttl inherit", we only set ttl to 0, which
is actually use whatever default value instead of inherit the inner protocol's
ttl value.

To make a difference with ttl inherit and ttl == 0, we add an attribute
IFLA_GENEVE_TTL_INHERIT in kernel commit 52d0d404d39dd ("geneve: add ttl
inherit support"). Now let's use "ttl inherit" to inherit the inner
protocol's ttl, and use "ttl auto" to means "use whatever default value",
the same behavior with ttl == 0.

v2:
1) remove IFLA_GENEVE_TTL_INHERIT defination in if_link.h as it's already
   updated.
2) Still use addattr8() so we can enable/disable ttl inherit, as Michal
   suggested.

v3: Update man page

Reported-by: Jianlin Shi <jishi@redhat.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agov4.19.0
Stephen Hemminger [Tue, 23 Oct 2018 17:14:57 +0000 (10:14 -0700)]
v4.19.0

5 years agotc: htb: Print default value in hex
Phil Sutter [Tue, 23 Oct 2018 10:36:24 +0000 (12:36 +0200)]
tc: htb: Print default value in hex

Value of 'default' is assumed to be hexadecimal when parsing, so
consequently it should be printed in hex as well. This is a regression
introduced when adding JSON output.

As requested, also change JSON output to print the value as hex string.

Fixes: f354fa6aa5ff0 ("tc: jsonify htb qdisc")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc: Remove pointless assignments in batch()
Phil Sutter [Thu, 18 Oct 2018 13:48:48 +0000 (15:48 +0200)]
tc: Remove pointless assignments in batch()

All these assignments are later overwritten without reading in between,
so just drop them.

Fixes: 485d0c6001c4a ("tc: Add batchsize feature for filter and actions")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotipc: Drop unused variable 'genl'
Phil Sutter [Thu, 18 Oct 2018 13:48:09 +0000 (15:48 +0200)]
tipc: Drop unused variable 'genl'

Although initialized by call to libmnl, the variable is used only in a
call to sizeof(). Drop it and call sizeof with its type instead.

Fixes: f043759dd4928 ("tipc: add new TIPC configuration tool")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip-route: Fix parse_encap_seg6() srh parsing
Phil Sutter [Thu, 18 Oct 2018 13:44:14 +0000 (15:44 +0200)]
ip-route: Fix parse_encap_seg6() srh parsing

In case caller did not specify 'segs' parameter, parse_srh() would read
garbage while iterating over 'segbuf'. Avoid this by initializing
'segbuf' to an empty string.

Fixes: e8493916a8ede ("iproute: add support for SR-IPv6 lwtunnel encapsulation")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agordma: Don't pass garbage to rd_check_is_filtered()
Phil Sutter [Thu, 18 Oct 2018 12:35:50 +0000 (14:35 +0200)]
rdma: Don't pass garbage to rd_check_is_filtered()

Variables 'src_port' and 'dst_port' are initialized only if attributes
RDMA_NLDEV_ATTR_RES_SRC_ADDR or RDMA_NLDEV_ATTR_RES_DST_ADDR are
present. Make sure to pass them over to rd_check_is_filtered() only if
that is the case.

Fixes: 9a362cc71a455 ("rdma: Add CM_ID resource tracking information")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip-route: Fix for memleak in error path
Phil Sutter [Thu, 18 Oct 2018 12:30:31 +0000 (14:30 +0200)]
ip-route: Fix for memleak in error path

If call to rta_addattr_l() failed, parse_encap_seg6() would leak memory.
Fix this by making sure calls to free() are not skipped.

Fixes: bd59e5b1517b0 ("ip-route: Fix segfault with many nexthops")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agordma: Fix for ineffective check in add_filter()
Phil Sutter [Thu, 18 Oct 2018 11:41:54 +0000 (13:41 +0200)]
rdma: Fix for ineffective check in add_filter()

With 'name' field defined as array in struct filters, it will always
contain a value irrespective of whether a name was assigned or not.

Fix this by turning the field into a const char pointer.

Fixes: 1174be72d1b4c ("rdma: Add filtering infrastructure")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agodevlink: Fix error reporting in cmd_resource_set()
Phil Sutter [Thu, 18 Oct 2018 11:28:23 +0000 (13:28 +0200)]
devlink: Fix error reporting in cmd_resource_set()

resource_path_parse() returns either zero or a negative error code,
hence the negated value must be passed to strerror().

Fixes: 8cd644095842a ("devlink: Add support for devlink resource abstraction")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotestsuite: Fix make check when need build generate_nlmsg
Petr Vorel [Tue, 25 Sep 2018 12:49:56 +0000 (14:49 +0200)]
testsuite: Fix make check when need build generate_nlmsg

make check from top level Makefile defines several flags which break
building generate_nlmsg:

$ make check
make -C tools
gcc  -Wall -Wstrict-prototypes  -Wmissing-prototypes -Wmissing-declarations -Wold-style-definition -Wformat=2 -O2 -I../include -I../include/uapi -DRESOLVE_HOSTNAMES -DLIBDIR=\"/usr/lib\" -DCONFDIR=\"/etc/iproute2\" -DNETNS_RUN_DIR=\"/var/run/netns\" -DNETNS_ETC_DIR=\"/etc/netns\" -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE  -DHAVE_SETNS -DHAVE_SELINUX -DHAVE_ELF -DHAVE_LIBMNL -I/usr/include/libmnl -DNEED_STRLCPY -DHAVE_LIBCAP ../lib/libutil.a ../lib/libnetlink.a -lselinux -lelf -lmnl -lcap  -I../../include -include../../include/uapi/linux/netlink.h -o generate_nlmsg generate_nlmsg.c ../../lib/libnetlink.c -lmnl
gcc: error: ../lib/libutil.a: No such file or directory
gcc: error: ../lib/libnetlink.a: No such file or directory
make[2]: *** [Makefile:5: generate_nlmsg] Error 1
make[1]: *** [Makefile:40: generate_nlmsg] Error 2

To fix it reset CFLAGS in sub Makefile and remove LDLIBS entirely (as
required -lmnl flag was specified in 5dc2204c ("testsuite: add libmnl").

Fixes: 8804a8c0 ("Makefile: Add check target")
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoiplink: Remove flags argument from iplink_get
David Ahern [Fri, 19 Oct 2018 19:55:44 +0000 (12:55 -0700)]
iplink: Remove flags argument from iplink_get

iplink_get has 1 caller and the flags arg is 0, so just remove it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoTree wide: Drop sockaddr_nl arg
David Ahern [Fri, 19 Oct 2018 20:42:36 +0000 (13:42 -0700)]
Tree wide: Drop sockaddr_nl arg

No function, filter, or print function uses the sockaddr_nl arg,
so just drop it.

Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'iproute2-master' into iproute2-next
David Ahern [Mon, 22 Oct 2018 16:43:33 +0000 (09:43 -0700)]
Merge branch 'iproute2-master' into iproute2-next

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoutil: spelling fix
Stephen Hemminger [Thu, 18 Oct 2018 20:23:38 +0000 (13:23 -0700)]
util: spelling fix

5 years agotipc: spelling fix
Stephen Hemminger [Thu, 18 Oct 2018 20:23:30 +0000 (13:23 -0700)]
tipc: spelling fix

5 years agoip: spelling fixes
Stephen Hemminger [Thu, 18 Oct 2018 20:23:11 +0000 (13:23 -0700)]
ip: spelling fixes

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc: spelling fixes
Stephen Hemminger [Thu, 18 Oct 2018 20:22:51 +0000 (13:22 -0700)]
tc: spelling fixes

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoconfig: spelling fixes
Stephen Hemminger [Thu, 18 Oct 2018 20:22:25 +0000 (13:22 -0700)]
config: spelling fixes

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoexamples: fix spelling errors
Stephen Hemminger [Thu, 18 Oct 2018 20:18:56 +0000 (13:18 -0700)]
examples: fix spelling errors

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agodoc/man: spelling fixes
Stephen Hemminger [Thu, 18 Oct 2018 20:15:45 +0000 (13:15 -0700)]
doc/man: spelling fixes

Use ispell and codespell to find/fix spelling errors in documentation
and man pages.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoip-addrlabel: Fix printing of label value
Phil Sutter [Mon, 15 Oct 2018 20:20:58 +0000 (22:20 +0200)]
ip-addrlabel: Fix printing of label value

Passing the return value of RTA_DATA() to rta_getattr_u32() is wrong
since that function will call RTA_DATA() by itself already.

Fixes: a7ad1c8a6845d ("ipaddrlabel: add json support")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoutils: fix get_rtnl_link_stats_rta stats parsing
Lorenzo Bianconi [Wed, 10 Oct 2018 15:00:58 +0000 (17:00 +0200)]
utils: fix get_rtnl_link_stats_rta stats parsing

iproute2 walks through the list of available tunnels using netlink
protocol in order to get device info instead of reading
them from proc filesystem. However the kernel reports device statistics
using IFLA_INET6_STATS/IFLA_INET6_ICMP6STATS attributes nested in
IFLA_PROTINFO one but iproutes expects these info in
IFLA_STATS64/IFLA_STATS attributes.
The issue can be triggered with the following reproducer:

$ip link add ip6d0 type ip6tnl mode ip6ip6 local 1111::1 remote 2222::1
$ip -6 -d -s tunnel show ip6d0
ip6d0: ipv6/ipv6 remote 2222::1 local 1111::1 encaplimit 4 hoplimit 64
tclass 0x00 flowlabel 0x00000 (flowinfo 0x00000000)
Dump terminated

Fix the issue introducing IFLA_INET6_STATS attribute parsing

Fixes: 3e953938717f ("iptunnel/ip6tunnel: Use netlink to walk through
tunnels list")

Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
5 years agouapi: add snmp header file
Lorenzo Bianconi [Wed, 10 Oct 2018 15:00:57 +0000 (17:00 +0200)]
uapi: add snmp header file

Introduce snmp header file. It will be used in subsequent patch in
order to parse device statistics reported in
IFLA_INET6_STATS/IFLA_INET6_ICMP6STATS netlink attributes

Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agomacsec: fix off-by-one when parsing attributes
Sabrina Dubroca [Fri, 12 Oct 2018 15:34:12 +0000 (17:34 +0200)]
macsec: fix off-by-one when parsing attributes

I seem to have had a massive brainfart with uses of
parse_rtattr_nested(). The rtattr* array must have MAX+1 elements, and
the call to parse_rtattr_nested must have MAX as its bound. Let's fix
those.

Fixes: b26fc590ce62 ("ip: add MACsec support")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agojson: make 0xhex handle u64
Sabrina Dubroca [Fri, 12 Oct 2018 15:34:32 +0000 (17:34 +0200)]
json: make 0xhex handle u64

Stephen converted macsec's sci to use 0xhex, but 0xhex handles
unsigned int's, not 64 bits ints. Thus, the output of the "ip macsec
show" command is mangled, with half of the SCI replaced with 0s:

# ip macsec show
11: macsec0: [...]
    cipher suite: GCM-AES-128, using ICV length 16
    TXSC: 0000000001560001 on SA 0

# ip -d link show macsec0
11: macsec0@ens3: [...]
    link/ether 52:54:00:12:01:56 brd ff:ff:ff:ff:ff:ff promiscuity 0
    macsec sci 5254001201560001 [...]

where TXSC and sci should match.

Fixes: c0b904de6211 ("macsec: support JSON")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agobridge: fdb: Fix for missing keywords in non-JSON output
Phil Sutter [Tue, 9 Oct 2018 12:44:08 +0000 (14:44 +0200)]
bridge: fdb: Fix for missing keywords in non-JSON output

While migrating to JSON print library, some keywords were dropped from
standard output by accident. Add them back to unbreak output parsers.

Fixes: c7c1a1ef51aea ("bridge: colorize output and use JSON print library")
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'master' into iproute2-next
David Ahern [Sun, 14 Oct 2018 02:31:37 +0000 (19:31 -0700)]
Merge branch 'master' into iproute2-next

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agobridge: add support for backup port
Nikolay Aleksandrov [Fri, 12 Oct 2018 11:42:55 +0000 (14:42 +0300)]
bridge: add support for backup port

This patch adds support for the new backup port option that can be set
on a bridge port. If the port's carrier goes down all of the traffic
gets redirected to the configured backup port. We add the following new
arguments:
$ ip link set dev brport type bridge_slave backup_port brport2
$ ip link set dev brport type bridge_slave nobackup_port

$ bridge link set dev brport backup_port brport2
$ bridge link set dev brport nobackup_port

The man pages are updated respectively.
Also 2 minor style adjustments:
- add missing space to bridge man page's state argument
- use lower starting case for vlan_tunnel in ip-link man page (to be
consistent with the rest)

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoipneigh: support for NTF_EXT_LEARNED flag on neigh entries
Roopa Prabhu [Thu, 11 Oct 2018 20:45:10 +0000 (13:45 -0700)]
ipneigh: support for NTF_EXT_LEARNED flag on neigh entries

Adds new option extern_learn to set NTF_EXT_LEARNED flag
on neigh entries.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: use local variable
Stephen Hemminger [Tue, 9 Oct 2018 16:46:11 +0000 (09:46 -0700)]
libnetlink: use local variable

Now that err->error is in local variable, use it consistently.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agolibnetlink: fix use-after-free of message buf
Vlad Buslov [Mon, 8 Oct 2018 20:52:26 +0000 (23:52 +0300)]
libnetlink: fix use-after-free of message buf

In __rtnl_talk_iov() main loop, err is a pointer to memory in dynamically
allocated 'buf' that is used to store netlink messages. If netlink message
is an error message, buf is deallocated before returning with error code.
However, on return err->error code is checked one more time to generate
return value, after memory which err points to has already been
freed. Save error code in temporary variable and use the variable to
generate return value.

Fixes: c60389e4f9ea ("libnetlink: fix leak and using unused memory on error")
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agotc: jsonify output of q_fifo
Jakub Kicinski [Fri, 5 Oct 2018 00:08:34 +0000 (17:08 -0700)]
tc: jsonify output of q_fifo

Print limits correctly in JSON context.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'taprio-scheduler' into iproute2-next
David Ahern [Sun, 7 Oct 2018 17:37:57 +0000 (10:37 -0700)]
Merge branch 'taprio-scheduler' into iproute2-next

Vinicius Costa Gomes  says:

====================

This is the iproute2 side of the taprio v1 series.

====================

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agotaprio: Add manpage for tc-taprio(8)
Vinicius Costa Gomes [Fri, 5 Oct 2018 23:25:22 +0000 (16:25 -0700)]
taprio: Add manpage for tc-taprio(8)

This documents the parameters and provides an example of usage.

Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agotc: Add support for configuring the taprio scheduler
Vinicius Costa Gomes [Fri, 5 Oct 2018 23:25:21 +0000 (16:25 -0700)]
tc: Add support for configuring the taprio scheduler

This traffic scheduler allows traffic classes states (transmission
allowed/not allowed, in the simplest case) to be scheduled, according
to a pre-generated time sequence. This is the basis of the IEEE
802.1Qbv specification.

Example configuration:

tc qdisc replace dev enp3s0 parent root handle 100 taprio \
          num_tc 3 \
  map 2 2 1 0 2 2 2 2 2 2 2 2 2 2 2 2 \
  queues 1@0 1@1 2@2 \
  base-time 1528743495910289987 \
  sched-entry S 01 300000 \
  sched-entry S 02 300000 \
  sched-entry S 04 300000 \
  clockid CLOCK_TAI

The configuration format is similar to mqprio. The main difference is
the presence of a schedule, built by multiple "sched-entry"
definitions, each entry has the following format:

     sched-entry <CMD> <GATE MASK> <INTERVAL>

The only supported <CMD> is "S", which means "SetGateStates",
following the IEEE 802.1Qbv-2015 definition (Table 8-6). <GATE MASK>
is a bitmask where each bit is a associated with a traffic class, so
bit 0 (the least significant bit) being "on" means that traffic class
0 is "active" for that schedule entry. <INTERVAL> is a time duration
in nanoseconds that specifies for how long that state defined by <CMD>
and <GATE MASK> should be held before moving to the next entry.

This schedule is circular, that is, after the last entry is executed
it starts from the first one, indefinitely.

The other parameters can be defined as follows:

 - base-time: specifies the instant when the schedule starts, if
  'base-time' is a time in the past, the schedule will start at

        base-time + (N * cycle-time)

   where N is the smallest integer so the resulting time is greater
   than "now", and "cycle-time" is the sum of all the intervals of the
   entries in the schedule;

 - clockid: specifies the reference clock to be used;

The parameters should be similar to what the IEEE 802.1Q family of
specification defines.

Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Jesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Add helper for getting a __s32 from netlink msgs
Jesus Sanchez-Palencia [Fri, 5 Oct 2018 23:25:19 +0000 (16:25 -0700)]
libnetlink: Add helper for getting a __s32 from netlink msgs

This function retrieves a signed 32-bit integer from a netlink message
and returns it.

Signed-off-by: Jesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoinclude: Add helper to retrieve a __s64 from a netlink msg
Vinicius Costa Gomes [Fri, 5 Oct 2018 23:25:18 +0000 (16:25 -0700)]
include: Add helper to retrieve a __s64 from a netlink msg

This allows signed 64-bit integers to be retrieved from a netlink
message.

Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoutils: Implement get_s64()
Vinicius Costa Gomes [Fri, 5 Oct 2018 23:25:17 +0000 (16:25 -0700)]
utils: Implement get_s64()

Add this helper to read signed 64-bit integers from a string.

Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoUpdate kernel headers
David Ahern [Sun, 7 Oct 2018 17:29:25 +0000 (10:29 -0700)]
Update kernel headers

update kernel headers to commit:
72438f8cef4e ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net")

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agotc: flower: expose hardware offload count
Vlad Buslov [Wed, 3 Oct 2018 20:44:48 +0000 (23:44 +0300)]
tc: flower: expose hardware offload count

Recently flower classifier was updated to expose count of devices that
filter is offloaded to. Add support to print this counter as 'in_hw_count'.

Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
5 years agovxlan: show correct ttl inherit info
Hangbin Liu [Thu, 27 Sep 2018 07:28:36 +0000 (15:28 +0800)]
vxlan: show correct ttl inherit info

We should only show ttl inherit when IFLA_VXLAN_TTL_INHERIT supplied.
Otherwise show the ttl number, or auto when it is 0.

Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
5 years agoMerge branch 'hdrs-for-dump-req' into iproute2-next
David Ahern [Wed, 3 Oct 2018 01:40:37 +0000 (18:40 -0700)]
Merge branch 'hdrs-for-dump-req' into iproute2-next

David Ahern says:

====================

iproute2 currently uses ifinfomsg as the header for all dumps using the
wilddump headers. This is wrong as each message type actually has its own
header type. While the kernel has traditionally let it go as it for the
most part only uses the family entry, the use of kernel side filters is
increasing to alter what is returned on a request. The kernel side filters
really need to use the proper header type.

To that end, fix iproute2 to use the proper header struct for the GET type.

====================

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Rename rtnl_wilddump_stats_req_filter to rtnl_statsdump_req_filter
David Ahern [Sat, 29 Sep 2018 17:12:59 +0000 (10:12 -0700)]
libnetlink: Rename rtnl_wilddump_stats_req_filter to rtnl_statsdump_req_filter

rtnl_wilddump_stats_req_filter only takes RTM_GETSTATS as the type argument
so rename to rtnl_statsdump_req_filter for consistency with other request
functions and hardcode the type argument.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Rename rtnl_wilddump_* to rtnl_linkdump_*
David Ahern [Sat, 29 Sep 2018 16:56:57 +0000 (09:56 -0700)]
libnetlink: Rename rtnl_wilddump_* to rtnl_linkdump_*

Rename rtnl_wilddump_req_filter to rtnl_linkdump_req_filter,
rtnl_wilddump_request to rtnl_linkdump_req and
rtnl_wilddump_req_filter_fn to rtnl_linkdump_req_filter_fn.

In all cases drop the type argument which at this point is only
RTM_GETLINK and hardcode in the functions.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETNSID dumps to use rtnl_nsiddump_req
David Ahern [Sat, 29 Sep 2018 16:51:08 +0000 (09:51 -0700)]
libnetlink: Convert GETNSID dumps to use rtnl_nsiddump_req

Add rtnl_nsiddump_req for namespace id dumps using the proper rtgenmsg
as the header. Convert existing RTM_GETNSID dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETNEIGHTBL dumps to use rtnl_neightbldump_req
David Ahern [Sat, 29 Sep 2018 16:47:41 +0000 (09:47 -0700)]
libnetlink: Convert GETNEIGHTBL dumps to use rtnl_neightbldump_req

Add rtnl_neightbldump_req for neighbor table dumps using the proper ndtmsg
as the header. Convert existing RTM_GETNEIGHTBL dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETNEIGH dumps to use rtnl_neighdump_req
David Ahern [Sat, 29 Sep 2018 16:42:40 +0000 (09:42 -0700)]
libnetlink: Convert GETNEIGH dumps to use rtnl_neighdump_req

Add rtnl_neighdump_req for neighbor dumps using the proper ndmsg
as the header. Convert existing rtnl_wilddump_request for RTM_GETNEIGH
to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETRULE dumps to use rtnl_ruledump_req
David Ahern [Sat, 29 Sep 2018 16:34:57 +0000 (09:34 -0700)]
libnetlink: Convert GETRULE dumps to use rtnl_ruledump_req

Add rtnl_ruledump_req for fib fule dumps using the proper fib_rule_hdr
as the header. Convert existing RTM_GETRULE dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETNETCONF dumps to use rtnl_netconfdump_req
David Ahern [Sat, 29 Sep 2018 16:29:29 +0000 (09:29 -0700)]
libnetlink: Convert GETNETCONF dumps to use rtnl_netconfdump_req

Add rtnl_netconfdump_req for netconf dumps using the proper netconfmsg
as the header. Convert existing RTM_GETNETCONF dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETMDB dumps to use rtnl_mdbdump_req
David Ahern [Sat, 29 Sep 2018 17:44:18 +0000 (10:44 -0700)]
libnetlink: Convert GETMDB dumps to use rtnl_mdbdump_req

Add rtnl_mdbdump_req for mdb dumps using the proper br_port_msg as
the header. Convert existing RTM_GETMDB dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETADDRLABEL dumps to use rtnl_addrlbldump_req
David Ahern [Sat, 29 Sep 2018 17:43:15 +0000 (10:43 -0700)]
libnetlink: Convert GETADDRLABEL dumps to use rtnl_addrlbldump_req

Add rtnl_addrlbldump_req for address label dumps using the proper
ifaddrlblmsg as the header. Convert existing RTM_GETADDRALBEL dumps
to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETROUTE dumps to use rtnl_routedump_req
David Ahern [Sat, 29 Sep 2018 15:46:31 +0000 (08:46 -0700)]
libnetlink: Convert GETROUTE dumps to use rtnl_routedump_req

Add rtnl_routedump_req for route dumps using the proper rtmsg
as the header. Convert existing RTM_GETROUTE dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agolibnetlink: Convert GETADDR dumps to use rtnl_addrdump_req
David Ahern [Sat, 29 Sep 2018 15:41:46 +0000 (08:41 -0700)]
libnetlink: Convert GETADDR dumps to use rtnl_addrdump_req

Add rtnl_addrdump_req for address dumps using the proper ifaddrmsg
as the header. Convert existing RTM_GETADDR dumps to use it.

Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agotc_util: Add support for showing TCA_STATS_BASIC_HW statistics
Eelco Chaudron [Tue, 2 Oct 2018 07:27:18 +0000 (03:27 -0400)]
tc_util: Add support for showing TCA_STATS_BASIC_HW statistics

Add support for showing hardware specific counters to easy
troubleshooting hardware offload.

$ tc -s filter show dev enp3s0np0 parent ffff:
filter protocol ip pref 1 flower chain 0
filter protocol ip pref 1 flower chain 0 handle 0x1
  eth_type ipv4
  dst_ip 2.0.0.0
  src_ip 1.0.0.0
  ip_flags nofrag
  in_hw
        action order 1: mirred (Egress Redirect to device eth1) stolen
        index 1 ref 1 bind 1 installed 0 sec used 0 sec
        Action statistics:
        Sent 534884742 bytes 8915697 pkt (dropped 0, overlimits 0 requeues 0)
        Sent software 187542 bytes 4077 pkt
        Sent hardware 534697200 bytes 8911620 pkt
        backlog 0b 0p requeues 0
        cookie 89173e6a44447001becfd486bda17e29

Signed-off-by: Eelco Chaudron <echaudro@redhat.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agotc: f_flower: add geneve option match support to flower
Pieter Jansen van Vuuren [Fri, 28 Sep 2018 14:03:39 +0000 (16:03 +0200)]
tc: f_flower: add geneve option match support to flower

Allow matching on options in Geneve tunnel headers.

The options can be described in the form
CLASS:TYPE:DATA/CLASS_MASK:TYPE_MASK:DATA_MASK, where CLASS is
represented as a 16bit hexadecimal value, TYPE as an 8bit
hexadecimal value and DATA as a variable length hexadecimal value.

e.g.
 # ip link add name geneve0 type geneve dstport 0 external
 # tc qdisc add dev geneve0 ingress
 # tc filter add dev geneve0 protocol ip parent ffff: \
     flower \
       enc_src_ip 10.0.99.192 \
       enc_dst_ip 10.0.99.193 \
       enc_key_id 11 \
       geneve_opts 0102:80:1122334421314151/ffff:ff:ffffffffffffffff \
       ip_proto udp \
       action mirred egress redirect dev eth1

Signed-off-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
5 years agoipneigh: update man page and help for router
Roopa Prabhu [Sun, 30 Sep 2018 02:48:11 +0000 (19:48 -0700)]
ipneigh: update man page and help for router

While at it also add missing text for proxy in the man page.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>