platform/kernel/linux-starfive.git
4 years agonet: phy: mscc: split the driver into separate files
Antoine Tenart [Fri, 13 Mar 2020 09:48:01 +0000 (10:48 +0100)]
net: phy: mscc: split the driver into separate files

This patch splits the MSCC driver into separate files, per
functionality, to improve readability and maintenance as the codebase
grew a lot. The MACsec code is moved to a dedicated mscc_macsec.c file,
the mscc.c file is renamed to mscc_main.c to keep the driver binary to
be named mscc and common definition are put into a new mscc.h header.

Most of the code was just moved around, except for a few exceptions:
- Header inclusions were reworked to only keep what's needed.
- Three helpers were created in the MACsec code, to avoid #ifdef's in
  the main C file: vsc8584_macsec_init, vsc8584_handle_macsec_interrupt
  and vsc8584_config_macsec_intr.

The patch should not introduce any functional modification.

Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: phy: move the mscc driver to its own directory
Antoine Tenart [Fri, 13 Mar 2020 09:48:00 +0000 (10:48 +0100)]
net: phy: move the mscc driver to its own directory

The MSCC PHY driver is growing, with lots of space consuming features
(firmware support, full initialization, MACsec...). It's becoming hard
to read and navigate in its source code. This patch moves the MSCC
driver to its own directory, without modifying anything, as a
preparation for splitting up its features into dedicated files.

Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'RED-Introduce-an-ECN-tail-dropping-mode'
David S. Miller [Sun, 15 Mar 2020 04:03:47 +0000 (21:03 -0700)]
Merge branch 'RED-Introduce-an-ECN-tail-dropping-mode'

Petr Machata says:

====================
RED: Introduce an ECN tail-dropping mode

When the RED qdisc is currently configured to enable ECN, the RED algorithm
is used to decide whether a certain SKB should be marked. If that SKB is
not ECN-capable, it is early-dropped.

It is also possible to keep all traffic in the queue, and just mark the
ECN-capable subset of it, as appropriate under the RED algorithm. Some
switches support this mode, and some installations make use of it.
There is currently no way to put the RED qdiscs to this mode.

Therefore this patchset adds a new RED flag, TC_RED_TAILDROP. When the
qdisc is configured with this flag, non-ECT traffic is enqueued (and
tail-dropped when the queue size is exhausted) instead of being
early-dropped.

Unfortunately, adding a new RED flag is not as simple as it sounds. RED
flags are passed in tc_red_qopt.flags. However RED neglects to validate the
flag field, and just copies it over wholesale to its internal structure,
and later dumps it back.

A broken userspace can therefore configure a RED qdisc with arbitrary
unsupported flags, and later expect to see the flags on qdisc dump. The
current ABI thus allows storage of 5 bits of custom data along with the
qdisc instance.

GRED, SFQ and CHOKE qdiscs are in the same situation. (GRED validates VQ
flags, but not the flags for the main queue.) E.g. if SFQ ever needs to
support TC_RED_ADAPTATIVE, it needs another way of doing it, and at the
same time it needs to retain the possibility to store 6 bits of
uninterpreted data.

For RED, this problem is resolved in patch #2, which adds a new attribute,
and a way to separate flags from userbits that can be reused by other
qdiscs. The flag itself and related behavioral changes are added in patch

To test the new feature, patch #1 first introduces a TDC testsuite that
covers the existing RED flags. Patch #5 later extends it with taildrop
coverage. Patch #6 contains a forwarding selftest for the offloaded
datapath.

To test the SW datapath, I took the mlxsw selftest and adapted it in mostly
obvious ways. The test is stable enough to verify that RED, ECN and ECN
taildrop actually work. However, I have no confidence in its portability to
other people's machines or mildly different configurations. I therefore do
not find it suitable for upstreaming.

GRED and CHOKE can use the same method as RED if they ever need to support
extra flags. SFQ uses the length of TCA_OPTIONS to dispatch on binary
control structure version, and would therefore need a different approach.

v2:
- Patch #1
    - Require nsPlugin in each RED test
    - Match end-of-line to catch cases of more flags reported than
      requested
- Patch #2:
    - Replaced with another patch.
- Patch #3:
    - Fix red_use_taildrop() condition in red_enqueue switch for
      probabilistic case.
- Patch #5:
    - Require nsPlugin in each RED test
    - Match end-of-line to catch cases of more flags reported than
      requested
    - Add a test for creation of non-ECN taildrop, which should fail
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests: mlxsw: RED: Test RED ECN nodrop offload
Petr Machata [Thu, 12 Mar 2020 23:11:00 +0000 (01:11 +0200)]
selftests: mlxsw: RED: Test RED ECN nodrop offload

Extend RED testsuite to cover the new nodrop mode of RED-ECN. This test is
really similar to ECN test, diverging only in the last step, where UDP
traffic should go to backlog instead of being dropped. Thus extract a
common helper, ecn_test_common(), make do_ecn_test() into a relatively
simple wrapper, and add another one, do_ecn_nodrop_test().

Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests: qdiscs: RED: Add nodrop tests
Petr Machata [Thu, 12 Mar 2020 23:10:59 +0000 (01:10 +0200)]
selftests: qdiscs: RED: Add nodrop tests

Add tests for the new "nodrop" flag.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agomlxsw: spectrum_qdisc: Offload RED ECN nodrop mode
Petr Machata [Thu, 12 Mar 2020 23:10:58 +0000 (01:10 +0200)]
mlxsw: spectrum_qdisc: Offload RED ECN nodrop mode

RED ECN nodrop mode means that non-ECT traffic should not be early-dropped,
but enqueued normally instead. In Spectrum systems, this is achieved by
disabling CWTPM.ew (enable WRED) for a given traffic class.

So far CWTPM.ew was unconditionally enabled. Instead disable it when the
RED qdisc is in nodrop mode.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: RED: Introduce an ECN nodrop mode
Petr Machata [Thu, 12 Mar 2020 23:10:57 +0000 (01:10 +0200)]
net: sched: RED: Introduce an ECN nodrop mode

When the RED Qdisc is currently configured to enable ECN, the RED algorithm
is used to decide whether a certain SKB should be marked. If that SKB is
not ECN-capable, it is early-dropped.

It is also possible to keep all traffic in the queue, and just mark the
ECN-capable subset of it, as appropriate under the RED algorithm. Some
switches support this mode, and some installations make use of it.

To that end, add a new RED flag, TC_RED_NODROP. When the Qdisc is
configured with this flag, non-ECT traffic is enqueued instead of being
early-dropped.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: Allow extending set of supported RED flags
Petr Machata [Thu, 12 Mar 2020 23:10:56 +0000 (01:10 +0200)]
net: sched: Allow extending set of supported RED flags

The qdiscs RED, GRED, SFQ and CHOKE use different subsets of the same pool
of global RED flags. These are passed in tc_red_qopt.flags. However none of
these qdiscs validate the flag field, and just copy it over wholesale to
internal structures, and later dump it back. (An exception is GRED, which
does validate for VQs -- however not for the main setup.)

A broken userspace can therefore configure a qdisc with arbitrary
unsupported flags, and later expect to see the flags on qdisc dump. The
current ABI therefore allows storage of several bits of custom data to
qdisc instances of the types mentioned above. How many bits, depends on
which flags are meaningful for the qdisc in question. E.g. SFQ recognizes
flags ECN and HARDDROP, and the rest is not interpreted.

If SFQ ever needs to support ADAPTATIVE, it needs another way of doing it,
and at the same time it needs to retain the possibility to store 6 bits of
uninterpreted data. Likewise RED, which adds a new flag later in this
patchset.

To that end, this patch adds a new function, red_get_flags(), to split the
passed flags of RED-like qdiscs to flags and user bits, and
red_validate_flags() to validate the resulting configuration. It further
adds a new attribute, TCA_RED_FLAGS, to pass arbitrary flags.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests: qdiscs: Add TDC test for RED
Petr Machata [Thu, 12 Mar 2020 23:10:55 +0000 (01:10 +0200)]
selftests: qdiscs: Add TDC test for RED

Add a handful of tests for creating RED with different flags.

Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Roman Mashak <mrv@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agosfc: support configuring vf spoofchk on EF10 VFs
Edward Cree [Thu, 12 Mar 2020 19:21:39 +0000 (19:21 +0000)]
sfc: support configuring vf spoofchk on EF10 VFs

Corresponds to the MAC_SPOOFING_TX privilege in the hardware.
Some firmware versions on some cards don't support the feature, so check
 the TX_MAC_SECURITY capability and fail EOPNOTSUPP if trying to enable
 spoofchk on a NIC that doesn't support it.

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'net-phy-XLGMII-define-and-usage-in-PHYLINK'
David S. Miller [Sun, 15 Mar 2020 03:55:12 +0000 (20:55 -0700)]
Merge branch 'net-phy-XLGMII-define-and-usage-in-PHYLINK'

Jose Abreu says:

====================
net: phy: XLGMII define and usage in PHYLINK

Adds XLGMII defines and usage in PHYLINK.

Patch 1/2, adds the define for it, whilst 2/2 adds the usage of it in
PHYLINK.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: phylink: Add XLGMII support
Jose Abreu [Thu, 12 Mar 2020 17:10:10 +0000 (18:10 +0100)]
net: phylink: Add XLGMII support

Add XLGMII interface and the list of XLGMII speeds to PHYLINK.

Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: phy: Add XLGMII interface define
Jose Abreu [Thu, 12 Mar 2020 17:10:09 +0000 (18:10 +0100)]
net: phy: Add XLGMII interface define

Add a define for XLGMII interface.

Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: ena: ethtool: clean up minor indentation issue
Colin Ian King [Thu, 12 Mar 2020 14:05:22 +0000 (14:05 +0000)]
net: ena: ethtool: clean up minor indentation issue

There is a statement that is indented incorrectly, remove a space.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: move MAC configuration to .phylink_mac_link_up
Vladimir Oltean [Thu, 12 Mar 2020 12:19:51 +0000 (12:19 +0000)]
net: dsa: sja1105: move MAC configuration to .phylink_mac_link_up

The switches supported so far by the driver only have non-SerDes ports,
so they should be configured in the PHYLINK callback that provides the
resolved PHY link parameters.

Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agocxgb4: update T5/T6 adapter register ranges
Shahjada Abul Husain [Thu, 12 Mar 2020 11:42:40 +0000 (17:12 +0530)]
cxgb4: update T5/T6 adapter register ranges

Add more T5/T6 registers to be collected in register dump:

1. MPS register range 0x9810 to 0x9864 and 0xd000 to 0xd004.
2. NCSI register range 0x1a114 to 0x1a130 and 0x1a138 to 0x1a1c4.

Signed-off-by: Shahjada Abul Husain <shahjada@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'mlx5-updates-2020-03-13' of git://git.kernel.org/pub/scm/linux/kernel...
David S. Miller [Sat, 14 Mar 2020 04:04:03 +0000 (21:04 -0700)]
Merge tag 'mlx5-updates-2020-03-13' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2020-03-13

Misc update to mlx5 core and E-Switch driver:

1) Blue-Field, Update VF vports config when num of VFs changed

From Bodon, Various misc cleanups and refactoring
for vport enabling/disabling routines to allow them to be called
dynamically and not only on E-Switch load.

This will allow ECPF (ConnectX BlueField Smartnic) support for dynamic
num vf changes and dynamic vport creation and configuration as introduced
in "Update VF vports config when num of VFs changed" patch.

2) From Parav and Mark, trivial clean-ups.

3) Software steering support for flow table id as destination
and a clean-up patch to remove unnecessary function stubs, from Alex.
====================

Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
David S. Miller [Sat, 14 Mar 2020 03:52:03 +0000 (20:52 -0700)]
Merge git://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2020-03-13

The following pull-request contains BPF updates for your *net-next* tree.

We've added 86 non-merge commits during the last 12 day(s) which contain
a total of 107 files changed, 5771 insertions(+), 1700 deletions(-).

The main changes are:

1) Add modify_return attach type which allows to attach to a function via
   BPF trampoline and is run after the fentry and before the fexit programs
   and can pass a return code to the original caller, from KP Singh.

2) Generalize BPF's kallsyms handling and add BPF trampoline and dispatcher
   objects to be visible in /proc/kallsyms so they can be annotated in
   stack traces, from Jiri Olsa.

3) Extend BPF sockmap to allow for UDP next to existing TCP support in order
   in order to enable this for BPF based socket dispatch, from Lorenz Bauer.

4) Introduce a new bpftool 'prog profile' command which attaches to existing
   BPF programs via fentry and fexit hooks and reads out hardware counters
   during that period, from Song Liu. Example usage:

   bpftool prog profile id 337 duration 3 cycles instructions llc_misses

        4228 run_cnt
     3403698 cycles                                              (84.08%)
     3525294 instructions   #  1.04 insn per cycle               (84.05%)
          13 llc_misses     #  3.69 LLC misses per million isns  (83.50%)

5) Batch of improvements to libbpf, bpftool and BPF selftests. Also addition
   of a new bpf_link abstraction to keep in particular BPF tracing programs
   attached even when the applicaion owning them exits, from Andrii Nakryiko.

6) New bpf_get_current_pid_tgid() helper for tracing to perform PID filtering
   and which returns the PID as seen by the init namespace, from Carlos Neira.

7) Refactor of RISC-V JIT code to move out common pieces and addition of a
   new RV32G BPF JIT compiler, from Luke Nelson.

8) Add gso_size context member to __sk_buff in order to be able to know whether
   a given skb is GSO or not, from Willem de Bruijn.

9) Add a new bpf_xdp_output() helper which reuses XDP's existing perf RB output
   implementation but can be called from tracepoint programs, from Eelco Chaudron.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet/mlx5: DR, Remove unneeded functions deceleration
Alex Vesker [Sun, 8 Mar 2020 11:21:41 +0000 (13:21 +0200)]
net/mlx5: DR, Remove unneeded functions deceleration

Remove dummy functions declaration, the dummy functions are not needed
since fs_dr is the only one to call mlx5dr and both fs_dr and dr files
depend on the same config flag (MLX5_SW_STEERING).

Fixes: 70605ea545e8 ("net/mlx5: DR, Expose APIs for direct rule managing")
Signed-off-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: DR, Add support for flow table id destination action
Alex Vesker [Wed, 26 Feb 2020 09:39:45 +0000 (11:39 +0200)]
net/mlx5: DR, Add support for flow table id destination action

This action allows to go to a flow table based on the table id.
Goto flow table id is required for supporting user space SW.

Signed-off-by: Alex Vesker <valex@mellanox.com>
Reviewed-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: Avoid deriving mlx5_core_dev second time
Parav Pandit [Wed, 18 Dec 2019 05:16:11 +0000 (23:16 -0600)]
net/mlx5: Avoid deriving mlx5_core_dev second time

All callers needs to work on mlx5_core_dev and it is already derived
before calling mlx5_devlink_eswitch_check().
Hence, accept mlx5_core_dev in mlx5_devlink_eswitch_check().

Given that it works on mlx5_core_dev change helper function name to
drop devlink prefix.

Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-switch, Annotate esw state_lock mutex destroy
Parav Pandit [Wed, 18 Dec 2019 04:51:24 +0000 (22:51 -0600)]
net/mlx5: E-switch, Annotate esw state_lock mutex destroy

Invoke mutex_destroy() to catch any esw state_lock errors.

Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-switch, Annotate termtbl_mutex mutex destroy
Parav Pandit [Sat, 14 Dec 2019 09:24:25 +0000 (03:24 -0600)]
net/mlx5: E-switch, Annotate termtbl_mutex mutex destroy

Annotate mutex destroy to keep it symmetric to init sequence.
It should be destroyed after its users (representor netdevices) are
destroyed in below flow.

esw_offloads_disable()
  esw_offloads_unload_rep()

Hence, initialize the mutex before creating the representors which uses
it.

Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: Accept flow rules without match
Mark Bloch [Fri, 17 Jan 2020 18:30:32 +0000 (18:30 +0000)]
net/mlx5: Accept flow rules without match

Allow passing NULL spec when creating a flow rule. Such rules will act
as "catch all" flow rules.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Refactor unload all reps per rep type
Bodong Wang [Tue, 12 Nov 2019 17:56:12 +0000 (11:56 -0600)]
net/mlx5: E-Switch, Refactor unload all reps per rep type

Following introduction of per vport configuration of vport and rep,
unload all reps per rep type is still needed as IB reps can be
unloaded individually. However, a few internal functions exist purely
for this purpose, merge them to a single function.

This patch doesn't change any existing functionality.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Update VF vports config when num of VFs changed
Bodong Wang [Tue, 12 Nov 2019 17:30:10 +0000 (11:30 -0600)]
net/mlx5: E-Switch, Update VF vports config when num of VFs changed

Currently, ECPF eswitch manager does one-time only configuration for
VF vports when device switches to offloads mode. However, when num of
VFs changed from host side, driver doesn't update VF vports
configurations.

Hence, perform VFs vport configuration update whenever num_vfs change
event occurs.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Introduce per vport configuration for eswitch modes
Bodong Wang [Mon, 11 Nov 2019 22:40:35 +0000 (16:40 -0600)]
net/mlx5: E-Switch, Introduce per vport configuration for eswitch modes

Both legacy and offload modes require vport setup, only offload mode
requires rep setup. Before this patch, vport and rep operations are
separated applied to all relevant vports in different stages.

Change to use per vport configuration, so that vport and rep operations
are modularized per vport.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-switch, Make vport setup/cleanup sequence symmetric
Bodong Wang [Wed, 16 Oct 2019 16:19:34 +0000 (11:19 -0500)]
net/mlx5: E-switch, Make vport setup/cleanup sequence symmetric

Vport enable and disable sequence is incorrect. It should be:
  enable()
  esw_vport_setup_acl,
  esw_vport_setup,
  esw_vport_enable_qos.

  disable()
  esw_vport_disable_qos,
  esw_vport_cleanup,
  esw_vport_cleanup_acl.

Instead of having two setup functions for port and acl, merge
acl setup to port setup function.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Prepare for vport enable/disable refactor
Bodong Wang [Fri, 30 Aug 2019 20:41:09 +0000 (15:41 -0500)]
net/mlx5: E-Switch, Prepare for vport enable/disable refactor

Rename esw_apply_vport_config() to esw_vport_setup(), and add new
helper function esw_vport_cleanup() to make them symmetric.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Remove redundant warning when QoS enable failed
Bodong Wang [Thu, 17 Oct 2019 19:55:52 +0000 (14:55 -0500)]
net/mlx5: E-Switch, Remove redundant warning when QoS enable failed

esw_vport_enable_qos can return error in cases below:
1. QoS is already enabled. Warnning is useless in this case.
2. Create scheduling element cmd failed. There is already a warning.

Remove the redundant warnning if esw_vport_enable_qos returns err.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Hold mutex when querying drop counter in legacy mode
Bodong Wang [Fri, 13 Sep 2019 21:24:19 +0000 (16:24 -0500)]
net/mlx5: E-Switch, Hold mutex when querying drop counter in legacy mode

Consider scenario below, CPU 1 is at risk to query already destroyed
drop counters. Need to apply the same state mutex when disabling vport.

+-------------------------------+-------------------------------------+
| CPU 0                         | CPU 1                               |
+-------------------------------+-------------------------------------+
| mlx5_device_disable_sriov     | mlx5e_get_vf_stats                  |
| mlx5_eswitch_disable          | mlx5_eswitch_get_vport_stats        |
| esw_disable_vport             | mlx5_eswitch_query_vport_drop_stats |
| mlx5_fc_destroy(drop_counter) | mlx5_fc_query(drop_counter)         |
+-------------------------------+-------------------------------------+

Fixes: b8a0dbe3a90b ("net/mlx5e: E-switch, Add steering drop counters")
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agonet/mlx5: E-Switch, Remove redundant check of eswitch manager cap
Bodong Wang [Thu, 2 Jan 2020 21:30:52 +0000 (15:30 -0600)]
net/mlx5: E-Switch, Remove redundant check of eswitch manager cap

esw_vport_create_legacy_acl_tables bails out immediately for eswitch
manager, hence remove all the check of esw manager cap after.

Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
4 years agoMerge branch 'bpf-core-fixes'
Daniel Borkmann [Fri, 13 Mar 2020 22:30:53 +0000 (23:30 +0100)]
Merge branch 'bpf-core-fixes'

Andrii Nakryiko says:

====================
This patch set fixes bug in CO-RE relocation candidate finding logic, which
currently allows matching against forward declarations, functions, and other
named types, even though it makes no sense to even attempt. As part of
verifying the fix, add test using vmlinux.h with preserve_access_index
attribute and utilizing struct pt_regs heavily to trace nanosleep syscall
using 5 different types of tracing BPF programs.

This test also demonstrated problems using struct pt_regs in syscall
tracepoints and required a new set of macro, which were added in patch #3
into bpf_tracing.h.

Patch #1 fixes annoying issue with selftest failure messages being out of
sync.

v1->v2:
  - drop unused handle__probed() function (Martin).
====================

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
4 years agoselftests/bpf: Add vmlinux.h selftest exercising tracing of syscalls
Andrii Nakryiko [Fri, 13 Mar 2020 17:23:36 +0000 (10:23 -0700)]
selftests/bpf: Add vmlinux.h selftest exercising tracing of syscalls

Add vmlinux.h generation to selftest/bpf's Makefile. Use it from newly added
test_vmlinux to trace nanosleep syscall using 5 different types of programs:
  - tracepoint;
  - raw tracepoint;
  - raw tracepoint w/ direct memory reads (tp_btf);
  - kprobe;
  - fentry.

These programs are realistic variants of real-life tracing programs,
excercising vmlinux.h's usage with tracing applications.

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313172336.1879637-5-andriin@fb.com
4 years agolibbpf: Provide CO-RE variants of PT_REGS macros
Andrii Nakryiko [Fri, 13 Mar 2020 17:23:35 +0000 (10:23 -0700)]
libbpf: Provide CO-RE variants of PT_REGS macros

Syscall raw tracepoints have struct pt_regs pointer as tracepoint's first
argument. After that, reading any of pt_regs fields requires bpf_probe_read(),
even for tp_btf programs. Due to that, PT_REGS_PARMx macros are not usable as
is. This patch adds CO-RE variants of those macros that use BPF_CORE_READ() to
read necessary fields. This provides relocatable architecture-agnostic pt_regs
field accesses.

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313172336.1879637-4-andriin@fb.com
4 years agolibbpf: Ignore incompatible types with matching name during CO-RE relocation
Andrii Nakryiko [Fri, 13 Mar 2020 17:23:34 +0000 (10:23 -0700)]
libbpf: Ignore incompatible types with matching name during CO-RE relocation

When finding target type candidates, ignore forward declarations, functions,
and other named types of incompatible kind. Not doing this can cause false
errors.  See [0] for one such case (due to struct pt_regs forward
declaration).

  [0] https://github.com/iovisor/bcc/pull/2806#issuecomment-598543645

Fixes: ddc7c3042614 ("libbpf: implement BPF CO-RE offset relocation algorithm")
Reported-by: Wenbo Zhang <ethercflow@gmail.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313172336.1879637-3-andriin@fb.com
4 years agoselftests/bpf: Ensure consistent test failure output
Andrii Nakryiko [Fri, 13 Mar 2020 17:23:33 +0000 (10:23 -0700)]
selftests/bpf: Ensure consistent test failure output

printf() doesn't seem to honor using overwritten stdout/stderr (as part of
stdio hijacking), so ensure all "standard" invocations of printf() do
fprintf(stdout, ...) instead.

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313172336.1879637-2-andriin@fb.com
4 years agoselftests/bpf: Fix spurious failures in accept due to EAGAIN
Jakub Sitnicki [Fri, 13 Mar 2020 16:10:49 +0000 (17:10 +0100)]
selftests/bpf: Fix spurious failures in accept due to EAGAIN

Andrii Nakryiko reports that sockmap_listen test suite is frequently
failing due to accept() calls erroring out with EAGAIN:

  ./test_progs:connect_accept_thread:733: accept: Resource temporarily unavailable
  connect_accept_thread:FAIL:733

This is because we are using a non-blocking listening TCP socket to
accept() connections without polling on the socket.

While at first switching to blocking mode seems like the right thing to do,
this could lead to test process blocking indefinitely in face of a network
issue, like loopback interface being down, as Andrii pointed out.

Hence, stick to non-blocking mode for TCP listening sockets but with
polling for incoming connection for a limited time before giving up.

Apply this approach to all socket I/O calls in the test suite that we
expect to block indefinitely, that is accept() for TCP and recv() for UDP.

Fixes: 44d28be2b8d4 ("selftests/bpf: Tests for sockmap/sockhash holding listening sockets")
Reported-by: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200313161049.677700-1-jakub@cloudflare.com
4 years agotools/bpf: Move linux/types.h for selftests and bpftool
Tobias Klauser [Fri, 13 Mar 2020 11:31:05 +0000 (12:31 +0100)]
tools/bpf: Move linux/types.h for selftests and bpftool

Commit fe4eb069edb7 ("bpftool: Use linux/types.h from source tree for
profiler build") added a build dependency on tools/testing/selftests/bpf
to tools/bpf/bpftool. This is suboptimal with respect to a possible
stand-alone build of bpftool.

Fix this by moving tools/testing/selftests/bpf/include/uapi/linux/types.h
to tools/include/uapi/linux/types.h.

This requires an adjustment in the include search path order for the
tests in tools/testing/selftests/bpf so that tools/include/linux/types.h
is selected when building host binaries and
tools/include/uapi/linux/types.h is selected when building bpf binaries.

Verified by compiling bpftool and the bpf selftests on x86_64 with this
change.

Fixes: fe4eb069edb7 ("bpftool: Use linux/types.h from source tree for profiler build")
Suggested-by: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200313113105.6918-1-tklauser@distanz.ch
4 years agobpf: Add missing annotations for __bpf_prog_enter() and __bpf_prog_exit()
Jules Irenge [Wed, 11 Mar 2020 01:09:01 +0000 (01:09 +0000)]
bpf: Add missing annotations for __bpf_prog_enter() and __bpf_prog_exit()

Sparse reports a warning at __bpf_prog_enter() and __bpf_prog_exit()

warning: context imbalance in __bpf_prog_enter() - wrong count at exit
warning: context imbalance in __bpf_prog_exit() - unexpected unlock

The root cause is the missing annotation at __bpf_prog_enter()
and __bpf_prog_exit()

Add the missing __acquires(RCU) annotation
Add the missing __releases(RCU) annotation

Signed-off-by: Jules Irenge <jbi.octave@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200311010908.42366-2-jbi.octave@gmail.com
4 years agobpf_helpers_doc.py: Fix warning when compiling bpftool
Carlos Neira [Fri, 13 Mar 2020 15:46:50 +0000 (12:46 -0300)]
bpf_helpers_doc.py: Fix warning when compiling bpftool

When compiling bpftool the following warning is found: "declaration of
'struct bpf_pidns_info' will not be visible outside of this function."
This patch adds struct bpf_pidns_info to type_fwds array to fix this.

Fixes: b4490c5c4e02 ("bpf: Added new helper bpf_get_ns_current_pid_tgid")
Signed-off-by: Carlos Neira <cneirabustos@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313154650.13366-1-cneirabustos@gmail.com
4 years agoselftests/bpf: Fix usleep() implementation
Andrii Nakryiko [Fri, 13 Mar 2020 06:18:37 +0000 (23:18 -0700)]
selftests/bpf: Fix usleep() implementation

nanosleep syscall expects pointer to struct timespec, not nanoseconds
directly. Current implementation fulfills its purpose of invoking nanosleep
syscall, but doesn't really provide sleeping capabilities, which can cause
flakiness for tests relying on usleep() to wait for something.

Fixes: ec12a57b822c ("selftests/bpf: Guarantee that useep() calls nanosleep() syscall")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200313061837.3685572-1-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agoMerge branch 'generalize-bpf-ksym'
Alexei Starovoitov [Fri, 13 Mar 2020 02:23:12 +0000 (19:23 -0700)]
Merge branch 'generalize-bpf-ksym'

Jiri Olsa says:

====================
this patchset adds trampoline and dispatcher objects
to be visible in /proc/kallsyms.

  $ sudo cat /proc/kallsyms | tail -20
  ...
  ffffffffa050f000 t bpf_prog_5a2b06eab81b8f51    [bpf]
  ffffffffa0511000 t bpf_prog_6deef7357e7b4530    [bpf]
  ffffffffa0542000 t bpf_trampoline_13832 [bpf]
  ffffffffa0548000 t bpf_prog_96f1b5bf4e4cc6dc_mutex_lock [bpf]
  ffffffffa0572000 t bpf_prog_d1c63e29ad82c4ab_bpf_prog1  [bpf]
  ffffffffa0585000 t bpf_prog_e314084d332a5338__dissect   [bpf]
  ffffffffa0587000 t bpf_prog_59785a79eac7e5d2_mutex_unlock       [bpf]
  ffffffffa0589000 t bpf_prog_d0db6e0cac050163_mutex_lock [bpf]
  ffffffffa058d000 t bpf_prog_d8f047721e4d8321_bpf_prog2  [bpf]
  ffffffffa05df000 t bpf_trampoline_25637 [bpf]
  ffffffffa05e3000 t bpf_prog_d8f047721e4d8321_bpf_prog2  [bpf]
  ffffffffa05e5000 t bpf_prog_3b185187f1855c4c    [bpf]
  ffffffffa05e7000 t bpf_prog_d8f047721e4d8321_bpf_prog2  [bpf]
  ffffffffa05eb000 t bpf_prog_93cebb259dd5c4b2_do_sys_open        [bpf]
  ffffffffa0677000 t bpf_dispatcher_xdp   [bpf]

v5 changes:
  - keeping just 1 bpf_tree for all the objects and adding flag
    to recognize bpf_objects when searching for exception tables [Alexei]
  - no need for is_bpf_image_address call in kernel_text_address [Alexei]
  - removed the bpf_image tree, because it's no longer needed

v4 changes:
  - add trampoline and dispatcher to kallsyms once the it's allocated [Alexei]
  - omit the symbols sorting for kallsyms [Alexei]
  - small title change in one patch [Song]
  - some function renames:
     bpf_get_prog_name to bpf_prog_ksym_set_name
     bpf_get_prog_addr_region to bpf_prog_ksym_set_addr
  - added acks to changelogs
  - I checked and there'll be conflict on perftool side with
    upcoming changes from Adrian Hunter (text poke events),
    so I think it's better if Arnaldo takes the perf changes
    via perf tree and we will solve all conflicts there

v3 changes:
  - use container_of directly in bpf_get_ksym_start  [Daniel]
  - add more changelog explanations for ksym addresses [Daniel]

v2 changes:
  - omit extra condition in __bpf_ksym_add for sorting code (Andrii)
  - rename bpf_kallsyms_tree_ops to bpf_ksym_tree (Andrii)
  - expose only executable code in kallsyms (Andrii)
  - use full trampoline key as its kallsyms id (Andrii)
  - explained the BPF_TRAMP_REPLACE case (Andrii)
  - small format changes in bpf_trampoline_link_prog/bpf_trampoline_unlink_prog (Andrii)
  - propagate error value in bpf_dispatcher_update and update kallsym if it's successful (Andrii)
  - get rid of __always_inline for bpf_ksym_tree callbacks (Andrii)
  - added KSYMBOL notification for bpf_image add/removal
  - added perf tools changes to properly display trampoline/dispatcher
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Remove bpf_image tree
Jiri Olsa [Thu, 12 Mar 2020 19:56:07 +0000 (20:56 +0100)]
bpf: Remove bpf_image tree

Now that we have all the objects (bpf_prog, bpf_trampoline,
bpf_dispatcher) linked in bpf_tree, there's no need to have
separate bpf_image tree for images.

Reverting the bpf_image tree together with struct bpf_image,
because it's no longer needed.

Also removing bpf_image_alloc function and adding the original
bpf_jit_alloc_exec_page interface instead.

The kernel_text_address function can now rely only on is_bpf_text_address,
because it checks the bpf_tree that contains all the objects.

Keeping bpf_image_ksym_add and bpf_image_ksym_del because they are
useful wrappers with perf's ksymbol interface calls.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-13-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add dispatchers to kallsyms
Jiri Olsa [Thu, 12 Mar 2020 19:56:06 +0000 (20:56 +0100)]
bpf: Add dispatchers to kallsyms

Adding dispatchers to kallsyms. It's displayed as
  bpf_dispatcher_<NAME>

where NAME is the name of dispatcher.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-12-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add trampolines to kallsyms
Jiri Olsa [Thu, 12 Mar 2020 19:56:05 +0000 (20:56 +0100)]
bpf: Add trampolines to kallsyms

Adding trampolines to kallsyms. It's displayed as
  bpf_trampoline_<ID> [bpf]

where ID is the BTF id of the trampoline function.

Adding bpf_image_ksym_add/del functions that setup
the start/end values and call KSYMBOL perf events
handlers.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-11-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add bpf_ksym_add/del functions
Jiri Olsa [Thu, 12 Mar 2020 19:56:04 +0000 (20:56 +0100)]
bpf: Add bpf_ksym_add/del functions

Separating /proc/kallsyms add/del code and adding bpf_ksym_add/del
functions for that.

Moving bpf_prog_ksym_node_add/del functions to __bpf_ksym_add/del
and changing their argument to 'struct bpf_ksym' object. This way
we can call them for other bpf objects types like trampoline and
dispatcher.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-10-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add prog flag to struct bpf_ksym object
Jiri Olsa [Thu, 12 Mar 2020 19:56:03 +0000 (20:56 +0100)]
bpf: Add prog flag to struct bpf_ksym object

Adding 'prog' bool flag to 'struct bpf_ksym' to mark that
this object belongs to bpf_prog object.

This change allows having bpf_prog objects together with
other types (trampolines and dispatchers) in the single
bpf_tree. It's used when searching for bpf_prog exception
tables by the bpf_prog_ksym_find function, where we need
to get the bpf_prog pointer.

>From now we can safely add bpf_ksym support for trampoline
or dispatcher objects, because we can differentiate them
from bpf_prog objects.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-9-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Abstract away entire bpf_link clean up procedure
Andrii Nakryiko [Fri, 13 Mar 2020 00:21:28 +0000 (17:21 -0700)]
bpf: Abstract away entire bpf_link clean up procedure

Instead of requiring users to do three steps for cleaning up bpf_link, its
anon_inode file, and unused fd, abstract that away into bpf_link_cleanup()
helper. bpf_link_defunct() is removed, as it shouldn't be needed as an
individual operation anymore.

v1->v2:
- keep bpf_link_cleanup() static for now (Daniel).

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313002128.2028680-1-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add bpf_ksym_find function
Jiri Olsa [Thu, 12 Mar 2020 19:56:02 +0000 (20:56 +0100)]
bpf: Add bpf_ksym_find function

Adding bpf_ksym_find function that is used bpf bpf address
lookup functions:
  __bpf_address_lookup
  is_bpf_text_address

while keeping bpf_prog_kallsyms_find to be used only for lookup
of bpf_prog objects (will happen in following changes).

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-8-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agoselftests/bpf: Make tcp_rtt test more robust to failures
Andrii Nakryiko [Wed, 11 Mar 2020 22:27:49 +0000 (15:27 -0700)]
selftests/bpf: Make tcp_rtt test more robust to failures

Switch to non-blocking accept and wait for server thread to exit before
proceeding. I noticed that sometimes tcp_rtt server thread failure would
"spill over" into other tests (that would run after tcp_rtt), probably just
because server thread exits much later and tcp_rtt doesn't wait for it.

v1->v2:
  - add usleep() while waiting on initial non-blocking accept() (Stanislav);

Fixes: 8a03222f508b ("selftests/bpf: test_progs: fix client/server race in tcp_rtt")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/bpf/20200311222749.458015-1-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Move ksym_tnode to bpf_ksym
Jiri Olsa [Thu, 12 Mar 2020 19:56:01 +0000 (20:56 +0100)]
bpf: Move ksym_tnode to bpf_ksym

Moving ksym_tnode list node to 'struct bpf_ksym' object,
so the symbol itself can be chained and used in other
objects like bpf_trampoline and bpf_dispatcher.

We need bpf_ksym object to be linked both in bpf_kallsyms
via lnode for /proc/kallsyms and in bpf_tree via tnode for
bpf address lookup functions like __bpf_address_lookup or
bpf_prog_kallsyms_find.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200312195610.346362-7-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agoselftests/bpf: Guarantee that useep() calls nanosleep() syscall
Andrii Nakryiko [Wed, 11 Mar 2020 18:53:45 +0000 (11:53 -0700)]
selftests/bpf: Guarantee that useep() calls nanosleep() syscall

Some implementations of C runtime library won't call nanosleep() syscall from
usleep(). But a bunch of kprobe/tracepoint selftests rely on nanosleep being
called to trigger them. To make this more reliable, "override" usleep
implementation and call nanosleep explicitly.

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Julia Kartseva <hex@fb.com>
Link: https://lore.kernel.org/bpf/20200311185345.3874602-1-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Move lnode list node to struct bpf_ksym
Jiri Olsa [Thu, 12 Mar 2020 19:56:00 +0000 (20:56 +0100)]
bpf: Move lnode list node to struct bpf_ksym

Adding lnode list node to 'struct bpf_ksym' object,
so the struct bpf_ksym itself can be chained and used
in other objects like bpf_trampoline and bpf_dispatcher.

Changing iterator to bpf_ksym in bpf_get_kallsym function.

The ksym->start is holding the prog->bpf_func value,
so it's ok to use it as value in bpf_get_kallsym.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200312195610.346362-6-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agotools: bpftool: Restore message on failure to guess program type
Quentin Monnet [Wed, 11 Mar 2020 02:12:05 +0000 (02:12 +0000)]
tools: bpftool: Restore message on failure to guess program type

In commit 4a3d6c6a6e4d ("libbpf: Reduce log level for custom section
names"), log level for messages for libbpf_attach_type_by_name() and
libbpf_prog_type_by_name() was downgraded from "info" to "debug". The
latter function, in particular, is used by bpftool when attempting to
load programs, and this change caused bpftool to exit with no hint or
error message when it fails to detect the type of the program to load
(unless "-d" option was provided).

To help users understand why bpftool fails to load the program, let's do
a second run of the function with log level in "debug" mode in case of
failure.

Before:

    # bpftool prog load sample_ret0.o /sys/fs/bpf/sample_ret0
    # echo $?
    255

Or really verbose with -d flag:

    # bpftool -d prog load sample_ret0.o /sys/fs/bpf/sample_ret0
    libbpf: loading sample_ret0.o
    libbpf: section(1) .strtab, size 134, link 0, flags 0, type=3
    libbpf: skip section(1) .strtab
    libbpf: section(2) .text, size 16, link 0, flags 6, type=1
    libbpf: found program .text
    libbpf: section(3) .debug_abbrev, size 55, link 0, flags 0, type=1
    libbpf: skip section(3) .debug_abbrev
    libbpf: section(4) .debug_info, size 75, link 0, flags 0, type=1
    libbpf: skip section(4) .debug_info
    libbpf: section(5) .rel.debug_info, size 32, link 14, flags 0, type=9
    libbpf: skip relo .rel.debug_info(5) for section(4)
    libbpf: section(6) .debug_str, size 150, link 0, flags 30, type=1
    libbpf: skip section(6) .debug_str
    libbpf: section(7) .BTF, size 155, link 0, flags 0, type=1
    libbpf: section(8) .BTF.ext, size 80, link 0, flags 0, type=1
    libbpf: section(9) .rel.BTF.ext, size 32, link 14, flags 0, type=9
    libbpf: skip relo .rel.BTF.ext(9) for section(8)
    libbpf: section(10) .debug_frame, size 40, link 0, flags 0, type=1
    libbpf: skip section(10) .debug_frame
    libbpf: section(11) .rel.debug_frame, size 16, link 14, flags 0, type=9
    libbpf: skip relo .rel.debug_frame(11) for section(10)
    libbpf: section(12) .debug_line, size 74, link 0, flags 0, type=1
    libbpf: skip section(12) .debug_line
    libbpf: section(13) .rel.debug_line, size 16, link 14, flags 0, type=9
    libbpf: skip relo .rel.debug_line(13) for section(12)
    libbpf: section(14) .symtab, size 96, link 1, flags 0, type=2
    libbpf: looking for externs among 4 symbols...
    libbpf: collected 0 externs total
    libbpf: failed to guess program type from ELF section '.text'
    libbpf: supported section(type) names are: socket sk_reuseport kprobe/ [...]

After:

    # bpftool prog load sample_ret0.o /sys/fs/bpf/sample_ret0
    libbpf: failed to guess program type from ELF section '.text'
    libbpf: supported section(type) names are: socket sk_reuseport kprobe/ [...]

Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200311021205.9755-1-quentin@isovalent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add name to struct bpf_ksym
Jiri Olsa [Thu, 12 Mar 2020 19:55:59 +0000 (20:55 +0100)]
bpf: Add name to struct bpf_ksym

Adding name to 'struct bpf_ksym' object to carry the name
of the symbol for bpf_prog, bpf_trampoline, bpf_dispatcher
objects.

The current benefit is that name is now generated only when
the symbol is added to the list, so we don't need to generate
it every time it's accessed.

The future benefit is that we will have all the bpf objects
symbols represented by struct bpf_ksym.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200312195610.346362-5-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add struct bpf_ksym
Jiri Olsa [Thu, 12 Mar 2020 19:55:58 +0000 (20:55 +0100)]
bpf: Add struct bpf_ksym

Adding 'struct bpf_ksym' object that will carry the
kallsym information for bpf symbol. Adding the start
and end address to begin with. It will be used by
bpf_prog, bpf_trampoline, bpf_dispatcher objects.

The symbol_start/symbol_end values were originally used
to sort bpf_prog objects. For the address displayed in
/proc/kallsyms we are using prog->bpf_func value.

I'm using the bpf_func value for program symbol start
instead of the symbol_start, because it makes no difference
for sorting bpf_prog objects and we can use it directly as
an address to display it in /proc/kallsyms.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200312195610.346362-4-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agobpf: Add bpf_trampoline_ name prefix for DECLARE_BPF_DISPATCHER
Björn Töpel [Thu, 12 Mar 2020 19:55:57 +0000 (20:55 +0100)]
bpf: Add bpf_trampoline_ name prefix for DECLARE_BPF_DISPATCHER

Adding bpf_trampoline_ name prefix for DECLARE_BPF_DISPATCHER,
so all the dispatchers have the common name prefix.

And also a small '_' cleanup for bpf_dispatcher_nopfunc function
name.

Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200312195610.346362-3-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agox86/mm: Rename is_kernel_text to __is_kernel_text
Jiri Olsa [Thu, 12 Mar 2020 19:55:56 +0000 (20:55 +0100)]
x86/mm: Rename is_kernel_text to __is_kernel_text

The kbuild test robot reported compile issue on x86 in one of
the following patches that adds <linux/kallsyms.h> include into
<linux/bpf.h>, which is picked up by init_32.c object.

The problem is that <linux/kallsyms.h> defines global function
is_kernel_text which colides with the static function of the
same name defined in init_32.c:

  $ make ARCH=i386
  ...
  >> arch/x86/mm/init_32.c:241:19: error: redefinition of 'is_kernel_text'
    static inline int is_kernel_text(unsigned long addr)
                      ^~~~~~~~~~~~~~
   In file included from include/linux/bpf.h:21:0,
                    from include/linux/bpf-cgroup.h:5,
                    from include/linux/cgroup-defs.h:22,
                    from include/linux/cgroup.h:28,
                    from include/linux/hugetlb.h:9,
                    from arch/x86/mm/init_32.c:18:
   include/linux/kallsyms.h:31:19: note: previous definition of 'is_kernel_text' was here
    static inline int is_kernel_text(unsigned long addr)

Renaming the init_32.c is_kernel_text function to __is_kernel_text.

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200312195610.346362-2-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agoMerge tag 'ieee802154-for-davem-2020-03-13' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Fri, 13 Mar 2020 18:27:22 +0000 (11:27 -0700)]
Merge tag 'ieee802154-for-davem-2020-03-13' of git://git./linux/kernel/git/sschmidt/wpan-next

Stefan Schmidt says:

====================
pull-request: ieee802154-next 2020-03-13

An update from ieee802154 for *net-next*

Two small patches with updates targeting the whole tree.
Sergin does update SPI drivers to the new transfer delay handling
and Gustavo did one of his zero-length array replacement patches.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: fix spelling mistake "exceeeds" -> "exceeds"
Colin Ian King [Fri, 13 Mar 2020 11:25:34 +0000 (11:25 +0000)]
ethtool: fix spelling mistake "exceeeds" -> "exceeds"

There are a couple of spelling mistakes in NL_SET_ERR_MSG_ATTR messages.
Fix these.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
David S. Miller [Fri, 13 Mar 2020 04:29:30 +0000 (21:29 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Minor overlapping changes, nothing serious.

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'drm-fixes-2020-03-13' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 13 Mar 2020 01:05:19 +0000 (18:05 -0700)]
Merge tag 'drm-fixes-2020-03-13' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "It's a bit quieter, probably not as much as it could be.

  There is on large regression fix in here from Lyude for displayport
  bandwidth calculations, there've been reports of multi-monitor in
  docks not working since -rc1 and this has been tested to fix those.

  Otherwise it's a bunch of i915 (with some GVT fixes), a set of amdgpu
  watermark + bios fixes, and an exynos iommu cleanup fix.

  core:
   - DP MST bandwidth regression fix.

  i915:
   - hard lockup fix
   - GVT fixes
   - 32-bit alignment issue fix
   - timeline wait fixes
   - cacheline_retire and free

  amdgpu:
   - Update the display watermark bounding box for navi14
   - Fix fetching vbios directly from rom on vega20/arcturus
   - Navi and renoir watermark fixes

  exynos:
   - iommu object cleanup fix"

`

* tag 'drm-fixes-2020-03-13' of git://anongit.freedesktop.org/drm/drm:
  drm/dp_mst: Rewrite and fix bandwidth limit checks
  drm/dp_mst: Reprobe path resources in CSN handler
  drm/dp_mst: Use full_pbn instead of available_pbn for bandwidth checks
  drm/dp_mst: Rename drm_dp_mst_is_dp_mst_end_device() to be less redundant
  drm/i915: Defer semaphore priority bumping to a workqueue
  drm/i915/gt: Close race between cacheline_retire and free
  drm/i915/execlists: Enable timeslice on partial virtual engine dequeue
  drm/i915: be more solid in checking the alignment
  drm/i915/gvt: Fix dma-buf display blur issue on CFL
  drm/i915: Return early for await_start on same timeline
  drm/i915: Actually emit the await_start
  drm/amdgpu/powerplay: nv1x, renior copy dcn clock settings of watermark to smu during boot up
  drm/exynos: Fix cleanup of IOMMU related objects
  drm/amdgpu: correct ROM_INDEX/DATA offset for VEGA20
  drm/amd/display: update soc bb for nv14
  drm/i915/gvt: Fix emulated vbt size issue
  drm/i915/gvt: Fix unnecessary schedule timer when no vGPU exits

4 years agobpf: Add bpf_xdp_output() helper
Eelco Chaudron [Fri, 6 Mar 2020 08:59:23 +0000 (08:59 +0000)]
bpf: Add bpf_xdp_output() helper

Introduce new helper that reuses existing xdp perf_event output
implementation, but can be called from raw_tracepoint programs
that receive 'struct xdp_buff *' as a tracepoint argument.

Signed-off-by: Eelco Chaudron <echaudro@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/158348514556.2239.11050972434793741444.stgit@xdp-tutorial
4 years agoMerge branch 'bpf_get_ns_current_pid_tgid'
Alexei Starovoitov [Fri, 13 Mar 2020 00:33:12 +0000 (17:33 -0700)]
Merge branch 'bpf_get_ns_current_pid_tgid'

Carlos Neira says:

====================
Currently bpf_get_current_pid_tgid(), is used to do pid filtering in bcc's
scripts but this helper returns the pid as seen by the root namespace which is
fine when a bcc script is not executed inside a container.
When the process of interest is inside a container, pid filtering will not work
if bpf_get_current_pid_tgid() is used.
This helper addresses this limitation returning the pid as it's seen by the current
namespace where the script is executing.

In the future different pid_ns files may belong to different devices, according to the
discussion between Eric Biederman and Yonghong in 2017 Linux plumbers conference.
To address that situation the helper requires inum and dev_t from /proc/self/ns/pid.
This helper has the same use cases as bpf_get_current_pid_tgid() as it can be
used to do pid filtering even inside a container.
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
4 years agotools/testing/selftests/bpf: Add self-tests for new helper bpf_get_ns_current_pid_tgid.
Carlos Neira [Wed, 4 Mar 2020 20:41:57 +0000 (17:41 -0300)]
tools/testing/selftests/bpf: Add self-tests for new helper bpf_get_ns_current_pid_tgid.

Self tests added for new helper bpf_get_ns_current_pid_tgid

Signed-off-by: Carlos Neira <cneirabustos@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200304204157.58695-4-cneirabustos@gmail.com
4 years agoMerge tag 'topic/mst-bw-check-fixes-for-airlied-2020-03-12-2' of git://anongit.freede...
Dave Airlie [Fri, 13 Mar 2020 00:38:24 +0000 (10:38 +1000)]
Merge tag 'topic/mst-bw-check-fixes-for-airlied-2020-03-12-2' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

UAPI Changes: None

Cross-subsystem Changes: None

Core Changes: Fixed regressions introduced by commit cd82d82cbc04
("drm/dp_mst: Add branch bandwidth validation to MST atomic check"),
which would cause us to:

* Calculate the available bandwidth on an MST topology incorrectly, and
  as a result reject most display configurations that would try to enable
  more then one sink on a topology
* Occasionally expose MST connectors to userspace before finishing
  probing their PBN capabilities, resulting in us rejecting display
  configurations because we assumed briefly that no bandwidth was
  available

Driver Changes: None

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/bf16ee577567beed91c86b7d9cda3ec2e8c50a71.camel@redhat.com
4 years agobpf: Added new helper bpf_get_ns_current_pid_tgid
Carlos Neira [Wed, 4 Mar 2020 20:41:56 +0000 (17:41 -0300)]
bpf: Added new helper bpf_get_ns_current_pid_tgid

New bpf helper bpf_get_ns_current_pid_tgid,
This helper will return pid and tgid from current task
which namespace matches dev_t and inode number provided,
this will allows us to instrument a process inside a container.

Signed-off-by: Carlos Neira <cneirabustos@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200304204157.58695-3-cneirabustos@gmail.com
4 years agofs/nsfs.c: Added ns_match
Carlos Neira [Wed, 4 Mar 2020 20:41:55 +0000 (17:41 -0300)]
fs/nsfs.c: Added ns_match

ns_match returns true if the namespace inode and dev_t matches the ones
provided by the caller.

Signed-off-by: Carlos Neira <cneirabustos@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200304204157.58695-2-cneirabustos@gmail.com
4 years agoMerge tag 'drm-intel-fixes-2020-03-12' of git://anongit.freedesktop.org/drm/drm-intel...
Dave Airlie [Thu, 12 Mar 2020 23:41:15 +0000 (09:41 +1000)]
Merge tag 'drm-intel-fixes-2020-03-12' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

drm/i915 fixes for v5.6-rc6:
- hard lockup fix
- GVT fixes
- 32-bit alignment issue fix
- timeline wait fixes
- cacheline_retire and free

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87lfo6ksvw.fsf@intel.com
4 years agotools: bpftool: Fix minor bash completion mistakes
Quentin Monnet [Thu, 12 Mar 2020 18:46:08 +0000 (18:46 +0000)]
tools: bpftool: Fix minor bash completion mistakes

Minor fixes for bash completion: addition of program name completion for
two subcommands, and correction for program test-runs and map pinning.

The completion for the following commands is fixed or improved:

    # bpftool prog run [TAB]
    # bpftool prog pin [TAB]
    # bpftool map pin [TAB]
    # bpftool net attach xdp name [TAB]

Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200312184608.12050-3-quentin@isovalent.com
4 years agotools: bpftool: Allow all prog/map handles for pinning objects
Quentin Monnet [Thu, 12 Mar 2020 18:46:07 +0000 (18:46 +0000)]
tools: bpftool: Allow all prog/map handles for pinning objects

Documentation and interactive help for bpftool have always explained
that the regular handles for programs (id|name|tag|pinned) and maps
(id|name|pinned) can be passed to the utility when attempting to pin
objects (bpftool prog pin PROG / bpftool map pin MAP).

THIS IS A LIE!! The tool actually accepts only ids, as the parsing is
done in do_pin_any() in common.c instead of reusing the parsing
functions that have long been generic for program and map handles.

Instead of fixing the doc, fix the code. It is trivial to reuse the
generic parsing, and to simplify do_pin_any() in the process.

Do not accept to pin multiple objects at the same time with
prog_parse_fds() or map_parse_fds() (this would require a more complex
syntax for passing multiple sysfs paths and validating that they
correspond to the number of e.g. programs we find for a given name or
tag).

Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200312184608.12050-2-quentin@isovalent.com
4 years agoMerge tag 'amd-drm-fixes-5.6-2020-03-11' of git://people.freedesktop.org/~agd5f/linux...
Dave Airlie [Thu, 12 Mar 2020 23:22:25 +0000 (09:22 +1000)]
Merge tag 'amd-drm-fixes-5.6-2020-03-11' of git://people.freedesktop.org/~agd5f/linux into drm-fixes

amd-drm-fixes-5.6-2020-03-11:

amdgpu:
- Update the display watermark bounding box for navi14
- Fix fetching vbios directly from rom on vega20/arcturus
- Navi and renoir watermark fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200312020924.4161-1-alexander.deucher@amd.com
4 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 12 Mar 2020 23:19:19 +0000 (16:19 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Pull networking fixes from David Miller:
 "It looks like a decent sized set of fixes, but a lot of these are one
  liner off-by-one and similar type changes:

   1) Fix netlink header pointer to calcular bad attribute offset
      reported to user. From Pablo Neira Ayuso.

   2) Don't double clear PHY interrupts when ->did_interrupt is set,
      from Heiner Kallweit.

   3) Add missing validation of various (devlink, nl802154, fib, etc.)
      attributes, from Jakub Kicinski.

   4) Missing *pos increments in various netfilter seq_next ops, from
      Vasily Averin.

   5) Missing break in of_mdiobus_register() loop, from Dajun Jin.

   6) Don't double bump tx_dropped in veth driver, from Jiang Lidong.

   7) Work around FMAN erratum A050385, from Madalin Bucur.

   8) Make sure ARP header is pulled early enough in bonding driver,
      from Eric Dumazet.

   9) Do a cond_resched() during multicast processing of ipvlan and
      macvlan, from Mahesh Bandewar.

  10) Don't attach cgroups to unrelated sockets when in interrupt
      context, from Shakeel Butt.

  11) Fix tpacket ring state management when encountering unknown GSO
      types. From Willem de Bruijn.

  12) Fix MDIO bus PHY resume by checking mdio_bus_phy_may_suspend()
      only in the suspend context. From Heiner Kallweit"

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (112 commits)
  net: systemport: fix index check to avoid an array out of bounds access
  tc-testing: add ETS scheduler to tdc build configuration
  net: phy: fix MDIO bus PM PHY resuming
  net: hns3: clear port base VLAN when unload PF
  net: hns3: fix RMW issue for VLAN filter switch
  net: hns3: fix VF VLAN table entries inconsistent issue
  net: hns3: fix "tc qdisc del" failed issue
  taprio: Fix sending packets without dequeueing them
  net: mvmdio: avoid error message for optional IRQ
  net: dsa: mv88e6xxx: Add missing mask of ATU occupancy register
  net: memcg: fix lockdep splat in inet_csk_accept()
  s390/qeth: implement smarter resizing of the RX buffer pool
  s390/qeth: refactor buffer pool code
  s390/qeth: use page pointers to manage RX buffer pool
  seg6: fix SRv6 L2 tunnels to use IANA-assigned protocol number
  net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed
  net/packet: tpacket_rcv: do not increment ring index on drop
  sxgbe: Fix off by one in samsung driver strncpy size arg
  net: caif: Add lockdep expression to RCU traversal primitive
  MAINTAINERS: remove Sathya Perla as Emulex NIC maintainer
  ...

4 years agolibbpf: Split BTF presence checks into libbpf- and kernel-specific parts
Andrii Nakryiko [Thu, 12 Mar 2020 18:50:33 +0000 (11:50 -0700)]
libbpf: Split BTF presence checks into libbpf- and kernel-specific parts

Needs for application BTF being present differs between user-space libbpf needs and kernel
needs. Currently, BTF is mandatory only in kernel only when BPF application is
using STRUCT_OPS. While libbpf itself relies more heavily on presense of BTF:
  - for BTF-defined maps;
  - for Kconfig externs;
  - for STRUCT_OPS as well.

Thus, checks for presence and validness of bpf_object's BPF needs to be
performed separately, which is patch does.

Fixes: 5327644614a1 ("libbpf: Relax check whether BTF is mandatory")
Reported-by: Michal Rostecki <mrostecki@opensuse.org>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Cc: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200312185033.736911-1-andriin@fb.com
4 years agobpftool: Add _bpftool and profiler.skel.h to .gitignore
Song Liu [Thu, 12 Mar 2020 18:23:32 +0000 (11:23 -0700)]
bpftool: Add _bpftool and profiler.skel.h to .gitignore

These files are generated, so ignore them.

Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200312182332.3953408-4-songliubraving@fb.com
4 years agobpftool: Skeleton should depend on libbpf
Song Liu [Thu, 12 Mar 2020 18:23:31 +0000 (11:23 -0700)]
bpftool: Skeleton should depend on libbpf

Add the dependency to libbpf, to fix build errors like:

  In file included from skeleton/profiler.bpf.c:5:
  .../bpf_helpers.h:5:10: fatal error: 'bpf_helper_defs.h' file not found
  #include "bpf_helper_defs.h"
           ^~~~~~~~~~~~~~~~~~~
  1 error generated.
  make: *** [skeleton/profiler.bpf.o] Error 1
  make: *** Waiting for unfinished jobs....

Fixes: 47c09d6a9f67 ("bpftool: Introduce "prog profile" command")
Suggested-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200312182332.3953408-3-songliubraving@fb.com
4 years agobpftool: Only build bpftool-prog-profile if supported by clang
Song Liu [Thu, 12 Mar 2020 18:23:30 +0000 (11:23 -0700)]
bpftool: Only build bpftool-prog-profile if supported by clang

bpftool-prog-profile requires clang to generate BTF for global variables.
When compared with older clang, skip this command. This is achieved by
adding a new feature, clang-bpf-global-var, to tools/build/feature.

Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200312182332.3953408-2-songliubraving@fb.com
4 years agodrm/dp_mst: Rewrite and fix bandwidth limit checks
Lyude Paul [Mon, 9 Mar 2020 21:01:31 +0000 (17:01 -0400)]
drm/dp_mst: Rewrite and fix bandwidth limit checks

Sigh, this is mostly my fault for not giving commit cd82d82cbc04
("drm/dp_mst: Add branch bandwidth validation to MST atomic check")
enough scrutiny during review. The way we're checking bandwidth
limitations here is mostly wrong:

For starters, drm_dp_mst_atomic_check_bw_limit() determines the
pbn_limit of a branch by simply scanning each port on the current branch
device, then uses the last non-zero full_pbn value that it finds. It
then counts the sum of the PBN used on each branch device for that
level, and compares against the full_pbn value it found before.

This is wrong because ports can and will have different PBN limitations
on many hubs, especially since a number of DisplayPort hubs out there
will be clever and only use the smallest link rate required for each
downstream sink - potentially giving every port a different full_pbn
value depending on what link rate it's trained at. This means with our
current code, which max PBN value we end up with is not well defined.

Additionally, we also need to remember when checking bandwidth
limitations that the top-most device in any MST topology is a branch
device, not a port. This means that the first level of a topology
doesn't technically have a full_pbn value that needs to be checked.
Instead, we should assume that so long as our VCPI allocations fit we're
within the bandwidth limitations of the primary MSTB.

We do however, want to check full_pbn on every port including those of
the primary MSTB. However, it's important to keep in mind that this
value represents the minimum link rate /between a port's sink or mstb,
and the mstb itself/. A quick diagram to explain:

                                MSTB #1
                               /       \
                              /         \
                           Port #1    Port #2
       full_pbn for Port #1 → |          | ← full_pbn for Port #2
                           Sink #1    MSTB #2
                                         |
                                       etc...

Note that in the above diagram, the combined PBN from all VCPI
allocations on said hub should not exceed the full_pbn value of port #2,
and the display configuration on sink #1 should not exceed the full_pbn
value of port #1. However, port #1 and port #2 can otherwise consume as
much bandwidth as they want so long as their VCPI allocations still fit.

And finally - our current bandwidth checking code also makes the mistake
of not checking whether something is an end device or not before trying
to traverse down it.

So, let's fix it by rewriting our bandwidth checking helpers. We split
the function into one part for handling branches which simply adds up
the total PBN on each branch and returns it, and one for checking each
port to ensure we're not going over its PBN limit. Phew.

This should fix regressions seen, where we erroneously reject display
configurations due to thinking they're going over our bandwidth limits
when they're not.

Changes since v1:
* Took an even closer look at how PBN limitations are supposed to be
  handled, and did some experimenting with Sean Paul. Ended up rewriting
  these helpers again, but this time they should actually be correct!
Changes since v2:
* Small indenting fix
* Fix pbn_used check in drm_dp_mst_atomic_check_port_bw_limit()

Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: cd82d82cbc04 ("drm/dp_mst: Add branch bandwidth validation to MST atomic check")
Cc: Sean Paul <seanpaul@google.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Mikita Lipski <mikita.lipski@amd.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200309210131.1497545-1-lyude@redhat.com
4 years agodrm/dp_mst: Reprobe path resources in CSN handler
Lyude Paul [Fri, 6 Mar 2020 23:46:21 +0000 (18:46 -0500)]
drm/dp_mst: Reprobe path resources in CSN handler

We used to punt off reprobing path resources to the link address probe
work, but now that we handle CSNs asynchronously from the driver's HPD
handling we can do whatever the heck we want from the CSN!

So, reprobe the path resources from drm_dp_mst_handle_conn_stat(). Also,
get rid of the path resource reprobing code in
drm_dp_check_and_send_link_address() since it's needlessly complicated
when we already reprobe path resources from
drm_dp_handle_link_address_port(). And finally, teach
drm_dp_send_enum_path_resources() to return 1 on PBN changes so we know
if we need to send another hotplug or not.

This fixes issues where we've indicated to userspace that a port has
just been connected, before we actually probed it's available PBN -
something that results in unexpected atomic check failures.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: cd82d82cbc04 ("drm/dp_mst: Add branch bandwidth validation to MST atomic check")
Cc: Mikita Lipski <mikita.lipski@amd.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Sean Paul <sean@poorly.run>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306234623.547525-4-lyude@redhat.com
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
4 years agodrm/dp_mst: Use full_pbn instead of available_pbn for bandwidth checks
Lyude Paul [Fri, 6 Mar 2020 23:46:20 +0000 (18:46 -0500)]
drm/dp_mst: Use full_pbn instead of available_pbn for bandwidth checks

DisplayPort specifications are fun. For a while, it's been really
unclear to us what available_pbn actually does. There's a somewhat vague
explanation in the DisplayPort spec (starting from 1.2) that partially
explains it:

  The minimum payload bandwidth number supported by the path. Each node
  updates this number with its available payload bandwidth number if its
  payload bandwidth number is less than that in the Message Transaction
  reply.

So, it sounds like available_pbn represents the smallest link rate in
use between the source and the branch device. Cool, so full_pbn is just
the highest possible PBN that the branch device supports right?

Well, we assumed that for quite a while until Sean Paul noticed that on
some MST hubs, available_pbn will actually get set to 0 whenever there's
any active payloads on the respective branch device. This caused quite a
bit of confusion since clearing the payload ID table would end up fixing
the available_pbn value.

So, we just went with that until commit cd82d82cbc04 ("drm/dp_mst: Add
branch bandwidth validation to MST atomic check") started breaking
people's setups due to us getting erroneous available_pbn values. So, we
did some more digging and got confused until we finally looked at the
definition for full_pbn:

  The bandwidth of the link at the trained link rate and lane count
  between the DP Source device and the DP Sink device with no time slots
  allocated to VC Payloads, represented as a Payload Bandwidth Number. As
  with the Available_Payload_Bandwidth_Number, this number is determined
  by the link with the lowest lane count and link rate.

That's what we get for not reading specs closely enough, hehe. So, since
full_pbn is definitely what we want for doing bandwidth restriction
checks - let's start using that instead and ignore available_pbn
entirely.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: cd82d82cbc04 ("drm/dp_mst: Add branch bandwidth validation to MST atomic check")
Cc: Mikita Lipski <mikita.lipski@amd.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Sean Paul <sean@poorly.run>
Reviewed-by: Mikita Lipski <mikita.lipski@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306234623.547525-3-lyude@redhat.com
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
4 years agodrm/dp_mst: Rename drm_dp_mst_is_dp_mst_end_device() to be less redundant
Lyude Paul [Fri, 6 Mar 2020 23:46:19 +0000 (18:46 -0500)]
drm/dp_mst: Rename drm_dp_mst_is_dp_mst_end_device() to be less redundant

It's already prefixed by dp_mst, so we don't really need to repeat
ourselves here. One of the changes I should have picked up originally
when reviewing MST DSC support.

There should be no functional changes here

Cc: Mikita Lipski <mikita.lipski@amd.com>
Cc: Sean Paul <seanpaul@google.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306234623.547525-2-lyude@redhat.com
4 years agoinet: Use fallthrough;
Joe Perches [Thu, 12 Mar 2020 22:50:22 +0000 (15:50 -0700)]
inet: Use fallthrough;

Convert the various uses of fallthrough comments to fallthrough;

Done via script
Link: https://lore.kernel.org/lkml/b56602fcf79f849e733e7b521bb0e17895d390fa.1582230379.git.joe@perches.com/
And by hand:

net/ipv6/ip6_fib.c has a fallthrough comment outside of an #ifdef block
that causes gcc to emit a warning if converted in-place.

So move the new fallthrough; inside the containing #ifdef/#endif too.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoocteontx2-pf: unlock on error path in otx2_config_pause_frm()
Dan Carpenter [Thu, 12 Mar 2020 11:30:48 +0000 (14:30 +0300)]
octeontx2-pf: unlock on error path in otx2_config_pause_frm()

We need to unlock before returning if this allocation fails.

Fixes: 75f36270990c ("octeontx2-pf: Support to enable/disable pause frames via ethtool")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Thu, 12 Mar 2020 22:51:26 +0000 (15:51 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/viro/vfs

Pull vfs fixes from Al Viro:
 "A couple of fixes for old crap in ->atomic_open() instances"

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  cifs_atomic_open(): fix double-put on late allocation failure
  gfs2_atomic_open(): fix O_EXCL|O_CREAT handling on cold dcache

4 years agonet: systemport: fix index check to avoid an array out of bounds access
Colin Ian King [Thu, 12 Mar 2020 15:04:30 +0000 (15:04 +0000)]
net: systemport: fix index check to avoid an array out of bounds access

Currently the bounds check on index is off by one and can lead to
an out of bounds access on array priv->filters_loc when index is
RXCHK_BRCM_TAG_MAX.

Fixes: bb9051a2b230 ("net: systemport: Add support for WAKE_FILTER")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'ipa-fixes'
David S. Miller [Thu, 12 Mar 2020 22:48:37 +0000 (15:48 -0700)]
Merge branch 'ipa-fixes'

Alex Elder says:

====================
net: fix net-next

David: These patches resolve two issues caused by the IPA driver
being incorporated into net-next.  I hope you will merge
them as soon as you can.

The IPA driver was merged into net-next last week, but two problems
arise as a result, affecting net-next and linux-next:
  - The patch that defines field_max() was not incorporated into
    net-next, but is required by the IPA code
  - A patch that updates "sdm845.dtsi" *was* incorporated into
    net-next, but other changes to that file in the Qualcomm
    for-next branch lead to errors

Bjorn has agreed to incorporate the DTS file change into the
Qualcomm tree after it is reverted from net-next.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoRevert "arm64: dts: sdm845: add IPA information"
Alex Elder [Thu, 12 Mar 2020 16:44:28 +0000 (11:44 -0500)]
Revert "arm64: dts: sdm845: add IPA information"

This reverts commit 9cc5ae125f0eaee471bc87fb5cbf29385fd9272a.

This commit:
  b303f9f0050b arm64: dts: sdm845: Redefine interconnect provider DT nodes
found in the Qualcomm for-next tree removes/redefines the interconnect
provider node(s) used for IPA.  I'm not sure whether it technically
conflicts with the IPA change to "sdm845.dtsi" in for-next, but it renders
it broken.

Revert this commit in the for-next tree, with the plan to incorporate
it into the Qualcomm tree instead.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agobitfield.h: add FIELD_MAX() and field_max()
Alex Elder [Thu, 12 Mar 2020 16:44:27 +0000 (11:44 -0500)]
bitfield.h: add FIELD_MAX() and field_max()

Define FIELD_MAX(), which supplies the maximum value that can be
represented by a field value.  Define field_max() as well, to go
along with the lower-case forms of the field mask functions.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agotc-testing: add ETS scheduler to tdc build configuration
Davide Caratti [Thu, 12 Mar 2020 16:51:45 +0000 (17:51 +0100)]
tc-testing: add ETS scheduler to tdc build configuration

add CONFIG_NET_SCH_ETS to 'config', otherwise test suites using this file
to perform a full tdc run will encounter the following warning:

  ok 645 e90e - Add ETS qdisc using bands # skipped - "-----> teardown stage" did not complete successfully

Fixes: 82c664b69c8b ("selftests: qdiscs: Add test coverage for ETS Qdisc")
Reported-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: phy: fix MDIO bus PM PHY resuming
Heiner Kallweit [Thu, 12 Mar 2020 21:25:20 +0000 (22:25 +0100)]
net: phy: fix MDIO bus PM PHY resuming

So far we have the unfortunate situation that mdio_bus_phy_may_suspend()
is called in suspend AND resume path, assuming that function result is
the same. After the original change this is no longer the case,
resulting in broken resume as reported by Geert.

To fix this call mdio_bus_phy_may_suspend() in the suspend path only,
and let the phy_device store the info whether it was suspended by
MDIO bus PM.

Fixes: 503ba7c69610 ("net: phy: Avoid multiple suspends")
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'ethtool-netlink-interface-part-3'
David S. Miller [Thu, 12 Mar 2020 22:32:33 +0000 (15:32 -0700)]
Merge branch 'ethtool-netlink-interface-part-3'

Michal Kubecek says:

====================
ethtool netlink interface, part 3

Implementation of more netlink request types:

  - netdev features (ethtool -k/-K, patches 3-6)
  - private flags (--show-priv-flags / --set-priv-flags, patches 7-9)
  - ring sizes (ethtool -g/-G, patches 10-12)
  - channel counts (ethtool -l/-L, patches 13-15)

Patch 1 is a style cleanup suggested in part 2 review and patch 2 updates
the mapping between netdev features and legacy ioctl requests (which are
still used by ethtool for backward compatibility).

Changes in v2:
  - fix netdev reference leaks in error path of ethnl_set_rings() and
    ethnl_set_channels() (found by Jakub Kicinski)
  - use __set_bit() rather than set_bit() (suggested by David Miller)
  - in replies to RINGS_GET and CHANNELS_GET requests, omit ring and
    channel types not supported by driver/device (suggested by Jakub
    Kicinski)
  - more descriptive message size calculations in rings_reply_size() and
    channels_reply_size() (suggested by Jakub Kicinski)
  - coding style cleanup (suggested by Jakub Kicinski)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: add CHANNELS_NTF notification
Michal Kubecek [Thu, 12 Mar 2020 20:08:48 +0000 (21:08 +0100)]
ethtool: add CHANNELS_NTF notification

Send ETHTOOL_MSG_CHANNELS_NTF notification whenever channel counts of
a network device are modified using ETHTOOL_MSG_CHANNELS_SET netlink
message or ETHTOOL_SCHANNELS ioctl request.

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: set device channel counts with CHANNELS_SET request
Michal Kubecek [Thu, 12 Mar 2020 20:08:43 +0000 (21:08 +0100)]
ethtool: set device channel counts with CHANNELS_SET request

Implement CHANNELS_SET netlink request to set channel counts of a network
device. These are traditionally set with ETHTOOL_SCHANNELS ioctl request.

Like the ioctl implementation, the generic ethtool code checks if supplied
values do not exceed driver defined limits; if they do, first offending
attribute is reported using extack. Checks preventing removing channels
used for RX indirection table or zerocopy AF_XDP socket are also
implemented.

Move ethtool_get_max_rxfh_channel() helper into common.c so that it can be
used by both ioctl and netlink code.

v2:
  - fix netdev reference leak in error path (found by Jakub Kicinsky)

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: provide channel counts with CHANNELS_GET request
Michal Kubecek [Thu, 12 Mar 2020 20:08:38 +0000 (21:08 +0100)]
ethtool: provide channel counts with CHANNELS_GET request

Implement CHANNELS_GET request to get channel counts of a network device.
These are traditionally available via ETHTOOL_GCHANNELS ioctl request.

Omit attributes for channel types which are not supported by driver or
device (zero reported for maximum).

v2: (all suggested by Jakub Kicinski)
  - minor cleanup in channels_prepare_data()
  - more descriptive channels_reply_size()
  - omit attributes with zero max count

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: add RINGS_NTF notification
Michal Kubecek [Thu, 12 Mar 2020 20:08:33 +0000 (21:08 +0100)]
ethtool: add RINGS_NTF notification

Send ETHTOOL_MSG_RINGS_NTF notification whenever ring sizes of a network
device are modified using ETHTOOL_MSG_RINGS_SET netlink message or
ETHTOOL_SRINGPARAM ioctl request.

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: set device ring sizes with RINGS_SET request
Michal Kubecek [Thu, 12 Mar 2020 20:08:28 +0000 (21:08 +0100)]
ethtool: set device ring sizes with RINGS_SET request

Implement RINGS_SET netlink request to set ring sizes of a network device.
These are traditionally set with ETHTOOL_SRINGPARAM ioctl request.

Like the ioctl implementation, the generic ethtool code checks if supplied
values do not exceed driver defined limits; if they do, first offending
attribute is reported using extack.

v2:
  - fix netdev reference leak in error path (found by Jakub Kicinsky)

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: provide ring sizes with RINGS_GET request
Michal Kubecek [Thu, 12 Mar 2020 20:08:23 +0000 (21:08 +0100)]
ethtool: provide ring sizes with RINGS_GET request

Implement RINGS_GET request to get ring sizes of a network device. These
are traditionally available via ETHTOOL_GRINGPARAM ioctl request.

Omit attributes for ring types which are not supported by driver or device
(zero reported for maximum).

v2: (all suggested by Jakub Kicinski)
  - minor cleanup in rings_prepare_data()
  - more descriptive rings_reply_size()
  - omit attributes with zero max size

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: add PRIVFLAGS_NTF notification
Michal Kubecek [Thu, 12 Mar 2020 20:08:18 +0000 (21:08 +0100)]
ethtool: add PRIVFLAGS_NTF notification

Send ETHTOOL_MSG_PRIVFLAGS_NTF notification whenever private flags of
a network device are modified using ETHTOOL_MSG_PRIVFLAGS_SET netlink
message or ETHTOOL_SPFLAGS ioctl request.

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoethtool: set device private flags with PRIVFLAGS_SET request
Michal Kubecek [Thu, 12 Mar 2020 20:08:13 +0000 (21:08 +0100)]
ethtool: set device private flags with PRIVFLAGS_SET request

Implement PRIVFLAGS_SET netlink request to set private flags of a network
device. These are traditionally set with ETHTOOL_SPFLAGS ioctl request.

Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>