OSDN Git Service

uclinux-h8/linux.git
4 years agocxgb4: smt: Add lock for atomic_dec_and_test
Chuhong Yuan [Tue, 6 Aug 2019 02:58:46 +0000 (10:58 +0800)]
cxgb4: smt: Add lock for atomic_dec_and_test

The atomic_dec_and_test() is not safe because it is
outside of locks.
Move the locks of t4_smte_free() to its caller,
cxgb4_smt_release() to protect the atomic decrement.

Fixes: 3bdb376e6944 ("cxgb4: introduce SMT ops to prepare for SMAC rewrite support")
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests: Add l2tp tests
David Ahern [Mon, 5 Aug 2019 22:41:37 +0000 (15:41 -0700)]
selftests: Add l2tp tests

Add IPv4 and IPv6 l2tp tests. Current set is over IP and with
IPsec.

v2
- add l2tp.sh to TEST_PROGS in Makefile

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: delete "register" keyword
Alexey Dobriyan [Mon, 5 Aug 2019 20:34:41 +0000 (23:34 +0300)]
net: delete "register" keyword

Delete long obsoleted "register" keyword.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agomkiss: Use refcount_t for refcount
Chuhong Yuan [Fri, 2 Aug 2019 16:48:21 +0000 (00:48 +0800)]
mkiss: Use refcount_t for refcount

refcount_t is better for reference counters since its
implementation can prevent overflows.
So convert atomic_t ref counters to refcount_t.

Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodpaa_eth: Use refcount_t for refcount
Chuhong Yuan [Fri, 2 Aug 2019 16:47:59 +0000 (00:47 +0800)]
dpaa_eth: Use refcount_t for refcount

refcount_t is better for reference counters since its
implementation can prevent overflows.
So convert atomic_t ref counters to refcount_t.

Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'batadv-next-for-davem-20190808' of git://git.open-mesh.org/linux-merge
David S. Miller [Thu, 8 Aug 2019 18:28:40 +0000 (11:28 -0700)]
Merge tag 'batadv-next-for-davem-20190808' of git://git.open-mesh.org/linux-merge

Simon Wunderlich says:

====================
This feature/cleanup patchset includes the following patches:

 - bump version strings, by Simon Wunderlich

 - Replace usage of strlcpy with strscpy, by Sven Eckelmann

 - Add OGMv2 per-interface queue and aggregations, by Linus Luessing
   (2 patches)
====================

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 [Wed, 7 Aug 2019 01:44:57 +0000 (18:44 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Just minor overlapping changes in the conflicts here.

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Wed, 7 Aug 2019 00:11:59 +0000 (17:11 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Pull networking fixes from David Miller:
 "Yeah I should have sent a pull request last week, so there is a lot
  more here than usual:

   1) Fix memory leak in ebtables compat code, from Wenwen Wang.

   2) Several kTLS bug fixes from Jakub Kicinski (circular close on
      disconnect etc.)

   3) Force slave speed check on link state recovery in bonding 802.3ad
      mode, from Thomas Falcon.

   4) Clear RX descriptor bits before assigning buffers to them in
      stmmac, from Jose Abreu.

   5) Several missing of_node_put() calls, mostly wrt. for_each_*() OF
      loops, from Nishka Dasgupta.

   6) Double kfree_skb() in peak_usb can driver, from Stephane Grosjean.

   7) Need to hold sock across skb->destructor invocation, from Cong
      Wang.

   8) IP header length needs to be validated in ipip tunnel xmit, from
      Haishuang Yan.

   9) Use after free in ip6 tunnel driver, also from Haishuang Yan.

  10) Do not use MSI interrupts on r8169 chips before RTL8168d, from
      Heiner Kallweit.

  11) Upon bridge device init failure, we need to delete the local fdb.
      From Nikolay Aleksandrov.

  12) Handle erros from of_get_mac_address() properly in stmmac, from
      Martin Blumenstingl.

  13) Handle concurrent rename vs. dump in netfilter ipset, from Jozsef
      Kadlecsik.

  14) Setting NETIF_F_LLTX on mac80211 causes complete breakage with
      some devices, so revert. From Johannes Berg.

  15) Fix deadlock in rxrpc, from David Howells.

  16) Fix Kconfig deps of enetc driver, we must have PHYLIB. From Yue
      Haibing.

  17) Fix mvpp2 crash on module removal, from Matteo Croce.

  18) Fix race in genphy_update_link, from Heiner Kallweit.

  19) bpf_xdp_adjust_head() stopped working with generic XDP when we
      fixes generic XDP to support stacked devices properly, fix from
      Jesper Dangaard Brouer.

  20) Unbalanced RCU locking in rt6_update_exception_stamp_rt(), from
      David Ahern.

  21) Several memory leaks in new sja1105 driver, from Vladimir Oltean"

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (214 commits)
  net: dsa: sja1105: Fix memory leak on meta state machine error path
  net: dsa: sja1105: Fix memory leak on meta state machine normal path
  net: dsa: sja1105: Really fix panic on unregistering PTP clock
  net: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as well
  net: dsa: sja1105: Fix broken learning with vlan_filtering disabled
  net: dsa: qca8k: Add of_node_put() in qca8k_setup_mdio_bus()
  net: sched: sample: allow accessing psample_group with rtnl
  net: sched: police: allow accessing police->params with rtnl
  net: hisilicon: Fix dma_map_single failed on arm64
  net: hisilicon: fix hip04-xmit never return TX_BUSY
  net: hisilicon: make hip04_tx_reclaim non-reentrant
  tc-testing: updated vlan action tests with batch create/delete
  net sched: update vlan action for batched events operations
  net: stmmac: tc: Do not return a fragment entry
  net: stmmac: Fix issues when number of Queues >= 4
  net: stmmac: xgmac: Fix XGMAC selftests
  be2net: disable bh with spin_lock in be_process_mcc
  net: cxgb3_main: Fix a resource leak in a error path in 'init_one()'
  net: ethernet: sun4i-emac: Support phy-handle property for finding PHYs
  net: bridge: move default pvid init/deinit to NETDEV_REGISTER/UNREGISTER
  ...

4 years agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Tue, 6 Aug 2019 21:41:45 +0000 (14:41 -0700)]
Merge branch '40GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
40GbE Intel Wired LAN Driver Updates 2019-08-05

This series contains updates to i40e driver only.

Dmitrii adds missing statistic counters for VEB and VEB TC's.

Slawomir adds support for logging the "Disable Firmware LLDP" flag
option and its current status.

Jake fixes an issue where VF's being notified of their link status
before their queues are enabled which was causing issues.  So always
report link status down when the VF queues are not enabled.  Also adds
future proofing when statistics are added or removed by adding checks to
ensure the data pointer for the strings lines up with the expected
statistics count.

Czeslaw fixes the advertised mode reported in ethtool for FEC, where the
"None BaseR RS" was always being displayed no matter what the mode it
was in.  Also added logging information when the PF is entering or
leaving "allmulti" (or promiscuous) mode.  Fixed up the logging logic
for VF's when leaving multicast mode to not include unicast as well.

v2: drop Aleksandr's patch (previously patch #2 in the series) to
    display the VF MAC address that is set by the VF while community
    feedback is addressed.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoopenvswitch: Print error when ovs_execute_actions() fails
Yifeng Sun [Mon, 5 Aug 2019 02:56:11 +0000 (19:56 -0700)]
openvswitch: Print error when ovs_execute_actions() fails

Currently in function ovs_dp_process_packet(), return values of
ovs_execute_actions() are silently discarded. This patch prints out
an debug message when error happens so as to provide helpful hints
for debugging.
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'sja1105-fixes'
David S. Miller [Tue, 6 Aug 2019 21:37:02 +0000 (14:37 -0700)]
Merge branch 'sja1105-fixes'

Vladimir Oltean says:

====================
Fixes for SJA1105 DSA: FDBs, Learning and PTP

This is an assortment of functional fixes for the sja1105 switch driver
targeted for the "net" tree (although they apply on net-next just as
well).

Patch 1/5 ("net: dsa: sja1105: Fix broken learning with vlan_filtering
disabled") repairs a breakage introduced in the early development stages
of the driver: support for traffic from the CPU has broken "normal"
frame forwarding (based on DMAC) - there is connectivity through the
switch only because all frames are flooded.
I debated whether this patch qualifies as a fix, since it puts the
switch into a mode it has never operated in before (aka SVL). But
"normal" forwarding did use to work before the "Traffic support for
SJA1105 DSA driver" patchset, and arguably this patch should have been
part of that.
Also, it would be strange for this feature to be broken in the 5.2 LTS.

Patch 2/5 ("net: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as
well") is a simplification of a previous FDB-related patch that is
currently in the 5.3 rc's.

Patches 3/5 - 5/5 fix various crashes found while running linuxptp over the
switch ports for extended periods of time, or in conjunction with other
error conditions. The fixed-up commits were all introduced in 5.2.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: Fix memory leak on meta state machine error path
Vladimir Oltean [Sun, 4 Aug 2019 22:38:48 +0000 (01:38 +0300)]
net: dsa: sja1105: Fix memory leak on meta state machine error path

When RX timestamping is enabled and two link-local (non-meta) frames are
received in a row, this constitutes an error.

The tagger is always caching the last link-local frame, in an attempt to
merge it with the meta follow-up frame when that arrives. To recover
from the above error condition, the initial cached link-local frame is
dropped and the second frame in a row is cached (in expectance of the
second meta frame).

However, when dropping the initial link-local frame, its backing memory
was being leaked.

Fixes: f3097be21bf1 ("net: dsa: sja1105: Add a state machine for RX timestamping")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: Fix memory leak on meta state machine normal path
Vladimir Oltean [Sun, 4 Aug 2019 22:38:47 +0000 (01:38 +0300)]
net: dsa: sja1105: Fix memory leak on meta state machine normal path

After a meta frame is received, it is associated with the cached
sp->data->stampable_skb from the DSA tagger private structure.

Cached means its refcount is incremented with skb_get() in order for
dsa_switch_rcv() to not free it when the tagger .rcv returns NULL.

The mistake is that skb_unref() is not the correct function to use. It
will correctly decrement the refcount (which will go back to zero) but
the skb memory will not be freed.  That is the job of kfree_skb(), which
also calls skb_unref().

But it turns out that freeing the cached stampable_skb is in fact not
necessary.  It is still a perfectly valid skb, and now it is even
annotated with the partial RX timestamp.  So remove the skb_copy()
altogether and simply pass the stampable_skb with a refcount of 1
(incremented by us, decremented by dsa_switch_rcv) up the stack.

Fixes: f3097be21bf1 ("net: dsa: sja1105: Add a state machine for RX timestamping")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: Really fix panic on unregistering PTP clock
Vladimir Oltean [Sun, 4 Aug 2019 22:38:46 +0000 (01:38 +0300)]
net: dsa: sja1105: Really fix panic on unregistering PTP clock

The IS_ERR_OR_NULL(priv->clock) check inside
sja1105_ptp_clock_unregister() is preventing cancel_delayed_work_sync
from actually being run.

Additionally, sja1105_ptp_clock_unregister() does not actually get run,
when placed in sja1105_remove(). The DSA switch gets torn down, but the
sja1105 module does not get unregistered. So sja1105_ptp_clock_unregister
needs to be moved to sja1105_teardown, to be symmetrical with
sja1105_ptp_clock_register which is called from the DSA sja1105_setup.

It is strange to fix a "fixes" patch, but the probe failure can only be
seen when the attached PHY does not respond to MDIO (issue which I can't
pinpoint the reason to) and it goes away after I power-cycle the board.
This time the patch was validated on a failing board, and the kernel
panic from the fixed commit's message can no longer be seen.

Fixes: 29dd908d355f ("net: dsa: sja1105: Cancel PTP delayed work on unregister")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as well
Vladimir Oltean [Sun, 4 Aug 2019 22:38:45 +0000 (01:38 +0300)]
net: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as well

It looks like the FDB dump taken from first-generation switches also
contains information on whether entries are static or not. So use that
instead of searching through the driver's tables.

Fixes: d763778224ea ("net: dsa: sja1105: Implement is_static for FDB entries on E/T")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: sja1105: Fix broken learning with vlan_filtering disabled
Vladimir Oltean [Sun, 4 Aug 2019 22:38:44 +0000 (01:38 +0300)]
net: dsa: sja1105: Fix broken learning with vlan_filtering disabled

When put under a bridge with vlan_filtering 0, the SJA1105 ports will
flood all traffic as if learning was broken. This is because learning
interferes with the rx_vid's configured by dsa_8021q as unique pvid's.

So learning technically still *does* work, it's just that the learnt
entries never get matched due to their unique VLAN ID.

The setting that saves the day is Shared VLAN Learning, which on this
switch family works exactly as desired: VLAN tagging still works
(untagged traffic gets the correct pvid) and FDB entries are still
populated with the correct contents including VID. Also, a frame cannot
violate the forwarding domain restrictions enforced by its classified
VLAN. It is just that the VID is ignored when looking up the FDB for
taking a forwarding decision (selecting the egress port).

This patch activates SVL, and the result is that frames with a learnt
DMAC are no longer flooded in the scenario described above.

Now exactly *because* SVL works as desired, we have to revisit some
earlier patches:

- It is no longer necessary to manipulate the VID of the 'bridge fdb
  {add,del}' command when vlan_filtering is off. This is because now,
  SVL is enabled for that case, so the actual VID does not matter*.

- It is still desirable to hide dsa_8021q VID's in the FDB dump
  callback. But right now the dump callback should no longer hide
  duplicates (one per each front panel port's pvid, plus one for the
  VLAN that the CPU port is going to tag a TX frame with), because there
  shouldn't be any (the switch will match a single FDB entry no matter
  its VID anyway).

* Not really... It's no longer necessary to transform a 'bridge fdb add'
  into 5 fdb add operations, but the user might still add a fdb entry with
  any vid, and all of them would appear as duplicates in 'bridge fdb
  show'. So force a 'bridge fdb add' to insert the VID of 0**, so that we
  can prune the duplicates at insertion time.

** The VID of 0 is better than 1 because it is always guaranteed to be
   in the ports' hardware filter. DSA also avoids putting the VID inside
   the netlink response message towards the bridge driver when we return
   this particular VID, which makes it suitable for FDB entries learnt
   with vlan_filtering off.

Fixes: 227d07a07ef1 ("net: dsa: sja1105: Add support for traffic through standalone ports")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Georg Waibel <georg.waibel@sensor-technik.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: qca8k: Add of_node_put() in qca8k_setup_mdio_bus()
Nishka Dasgupta [Sun, 4 Aug 2019 15:30:18 +0000 (21:00 +0530)]
net: dsa: qca8k: Add of_node_put() in qca8k_setup_mdio_bus()

Each iteration of for_each_available_child_of_node() puts the previous
node, but in the case of a return from the middle of the loop, there
is no put, thus causing a memory leak. Hence add an of_node_put() before
the return.
Additionally, the local variable ports in the function
qca8k_setup_mdio_bus() takes the return value of of_get_child_by_name(),
which gets a node but does not put it. If the function returns without
putting ports, it may cause a memory leak. Hence put ports before the
mid-loop return statement, and also outside the loop after its last usage
in this function.
Issues found with Coccinelle.

Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'Support-tunnels-over-VLAN-in-NFP'
David S. Miller [Tue, 6 Aug 2019 21:24:22 +0000 (14:24 -0700)]
Merge branch 'Support-tunnels-over-VLAN-in-NFP'

John Hurley says:

====================
Support tunnels over VLAN in NFP

This patchset deals with tunnel encap and decap when the end-point IP
address is on an internal port (for example and OvS VLAN port). Tunnel
encap without VLAN is already supported in the NFP driver. This patchset
extends that to include a push VLAN along with tunnel header push.

Patches 1-4 extend the flow_offload IR API to include actions that use
skbedit to set the ptype of an SKB and that send a packet to port ingress
from the act_mirred module. Such actions are used in flower rules that
forward tunnel packets to internal ports where they can be decapsulated.
OvS and its TC API is an example of a user-space app that produces such
rules.

Patch 5 modifies the encap offload code to allow the pushing of a VLAN
header after a tunnel header push.

Patches 6-10 deal with tunnel decap when the end-point is on an internal
port. They detect 'pre-tunnel rules' which do not deal with tunnels
themselves but, rather, forward packets to internal ports where they
can be decapped if required. Such rules are offloaded to a table in HW
along with an indication of whether packets need to be passed to this
table of not (based on their destination MAC address). Matching against
this table prior to decapsulation in HW allows the correct parsing and
handling of outer VLANs on tunnelled packets and the correct updating of
stats for said 'pre-tunnel' rules.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: encode mac indexes with pre-tunnel rule check
John Hurley [Sun, 4 Aug 2019 15:10:49 +0000 (16:10 +0100)]
nfp: flower: encode mac indexes with pre-tunnel rule check

When a tunnel packet arrives on the NFP card, its destination MAC is
looked up and MAC index returned for it. This index can help verify the
tunnel by, for example, ensuring that the packet arrived on the expected
port. If the packet is destined for a known MAC that is not connected to a
given physical port then the mac index can have a global value (e.g. when
a series of bonded ports shared the same MAC).

If the packet is to be detunneled at a bridge device or internal port like
an Open vSwitch VLAN port, then it should first match a 'pre-tunnel' rule
to direct it to that internal port.

Use the MAC index to indicate if a packet should match a pre-tunnel rule
before decap is allowed. Do this by tracking the number of internal ports
associated with a MAC address and, if the number if >0, set a bit in the
mac_index to forward the packet to the pre-tunnel table before continuing
with decap.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: remove offloaded MACs when reprs are applied to OvS bridges
John Hurley [Sun, 4 Aug 2019 15:09:11 +0000 (16:09 +0100)]
nfp: flower: remove offloaded MACs when reprs are applied to OvS bridges

MAC addresses along with an identifying index are offloaded to firmware to
allow tunnel decapsulation. If a tunnel packet arrives with a matching
destination MAC address and a verified index, it can continue on the
decapsulation process. This replicates the MAC verifications carried out
in the kernel network stack.

When a netdev is added to a bridge (e.g. OvS) then packets arriving on
that dev are directed through the bridge datapath instead of passing
through the network stack. Therefore, tunnelled packets matching the MAC
of that dev will not be decapped here.

Replicate this behaviour on firmware by removing offloaded MAC addresses
when a MAC representer is added to an OvS bridge. This can prevent any
false positive tunnel decaps.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: offload pre-tunnel rules
John Hurley [Sun, 4 Aug 2019 15:09:10 +0000 (16:09 +0100)]
nfp: flower: offload pre-tunnel rules

Pre-tunnel rules are TC flower and OvS rules that forward a packet to the
tunnel end point where it can then pass through the network stack and be
decapsulated. These are required if the tunnel end point is, say, an OvS
internal port.

Currently, firmware determines that a packet is in a tunnel and decaps it
if it has a known destination IP and MAC address. However, this bypasses
the flower pre-tunnel rule and so does not update the stats. Further to
this it ignores VLANs that may exist outside of the tunnel header.

Offload pre-tunnel rules to the NFP. This embeds the pre-tunnel rule into
the tunnel decap process based on (firmware) mac index and VLAN. This
means that decap can be carried out correctly with VLANs and that stats
can be updated for all kernel rules correctly.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: verify pre-tunnel rules
John Hurley [Sun, 4 Aug 2019 15:09:09 +0000 (16:09 +0100)]
nfp: flower: verify pre-tunnel rules

Pre-tunnel rules must direct packets to an internal port based on L2
information. Rules that egress to an internal port are already indicated
by a non-NULL device in its nfp_fl_payload struct. Verfiy the rest of the
match fields indicate that the rule is a pre-tunnel rule. This requires a
full match on the destination MAC address, an option VLAN field, and no
specific matches on other lower layer fields (with the exception of L4
proto and flags).

If a rule is identified as a pre-tunnel rule then mark it for offload to
the pre-tunnel table. Similarly, remove it from the pre-tunnel table on
rule deletion. The actual offloading of these commands is left to a
following patch.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: detect potential pre-tunnel rules
John Hurley [Sun, 4 Aug 2019 15:09:08 +0000 (16:09 +0100)]
nfp: flower: detect potential pre-tunnel rules

Pre-tunnel rules are used when the tunnel end-point is on an 'internal
port'. These rules are used to direct the tunnelled packets (based on outer
header fields) to the internal port where they can be detunnelled. The
rule must send the packet to ingress the internal port at the TC layer.

Currently FW does not support an action to send to ingress so cannot
offload such rules. However, in preparation for populating the pre-tunnel
table to represent such rules, check for rules that send to the ingress of
an internal port and mark them as such. Further validation of such rules
is left to subsequent patches.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonfp: flower: push vlan after tunnel in merge
John Hurley [Sun, 4 Aug 2019 15:09:07 +0000 (16:09 +0100)]
nfp: flower: push vlan after tunnel in merge

NFP allows the merging of 2 flows together into a single offloaded flow.
In the kernel datapath the packet must match 1 flow, impliment its
actions, recirculate, match the 2nd flow and also impliment its actions.
Merging creates a single flow with all actions from the 2 original flows.

Firmware impliments a tunnel header push as the packet is about to egress
the card. Therefore, if the first merge rule candiate pushes a tunnel,
then the second rule can only have an egress action for a valid merge to
occur (or else the action ordering will be incorrect). This prevents the
pushing of a tunnel header followed by the pushing of a vlan header.

In order to support this behaviour, firmware allows VLAN information to
be encoded in the tunnel push action. If this is non zero then the fw will
push a VLAN after the tunnel header push meaning that 2 such flows with
these actions can be merged (with action order being maintained).

Support tunnel in VLAN pushes by encoding VLAN information in the tunnel
push action of any merge flow requiring this.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: add ingress mirred action to hardware IR
John Hurley [Sun, 4 Aug 2019 15:09:06 +0000 (16:09 +0100)]
net: sched: add ingress mirred action to hardware IR

TC mirred actions (redirect and mirred) can send to egress or ingress of a
device. Currently only egress is used for hw offload rules.

Modify the intermediate representation for hw offload to include mirred
actions that go to ingress. This gives drivers access to such rules and
can decide whether or not to offload them.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: tc_act: add helpers to detect ingress mirred actions
John Hurley [Sun, 4 Aug 2019 15:09:05 +0000 (16:09 +0100)]
net: tc_act: add helpers to detect ingress mirred actions

TC mirred actions can send to egress or ingress on a given netdev. Helpers
exist to detect actions that are mirred to egress. Extend the header file
to include helpers to detect ingress mirred actions.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: add skbedit of ptype action to hardware IR
John Hurley [Sun, 4 Aug 2019 15:09:04 +0000 (16:09 +0100)]
net: sched: add skbedit of ptype action to hardware IR

TC rules can impliment skbedit actions. Currently actions that modify the
skb mark are passed to offloading drivers via the hardware intermediate
representation in the flow_offload API.

Extend this to include skbedit actions that modify the packet type of the
skb. Such actions may be used to set the ptype to HOST when redirecting a
packet to ingress.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: tc_act: add skbedit_ptype helper functions
John Hurley [Sun, 4 Aug 2019 15:09:03 +0000 (16:09 +0100)]
net: tc_act: add skbedit_ptype helper functions

The tc_act header file contains an inline function that checks if an
action is changing the skb mark of a packet and a further function to
extract the mark.

Add similar functions to check for and get skbedit actions that modify
the packet type of the skb.

Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Tue, 6 Aug 2019 21:21:21 +0000 (14:21 -0700)]
Merge branch '100GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
100GbE Intel Wired LAN Driver Updates 2019-08-04

This series contains more updates to fm10k from Jake Keller.

Jake removes the unnecessary initialization of some variables to help
resolve static code checker warnings.  Explicitly return success during
resume, since the value of 'err' is always success.  Fixed a issue with
incrementing a void pointer, which can produce undefined behavior.  Used
the __always_unused macro for function templates that are passed as
parameters in functions, but are not used.  Simplified the code by
removing an unnecessary macro in determining the value of NON_Q_VECTORS.
Fixed an issue, using bitwise operations to prevent the low address
overwriting the high portion of the address.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agor8169: sync PCIe PHY init with vendor driver 8.047.01
Heiner Kallweit [Sun, 4 Aug 2019 07:52:33 +0000 (09:52 +0200)]
r8169: sync PCIe PHY init with vendor driver 8.047.01

Synchronize PCIe PHY initialization with vendor driver version 8.047.01.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agor8169: add helper r8168_mac_ocp_modify
Heiner Kallweit [Sun, 4 Aug 2019 07:47:51 +0000 (09:47 +0200)]
r8169: add helper r8168_mac_ocp_modify

Add a helper for MAC OCP read-modify-write operations.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agor8169: remove access to legacy register MultiIntr
Heiner Kallweit [Sun, 4 Aug 2019 07:42:57 +0000 (09:42 +0200)]
r8169: remove access to legacy register MultiIntr

This code piece was inherited from RTL8139 code, the register at
address 0x5c however has a different meaning on RTL8169 and is unused.
So we can remove this.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'fq_codel-small-optimizations'
David S. Miller [Tue, 6 Aug 2019 21:18:20 +0000 (14:18 -0700)]
Merge branch 'fq_codel-small-optimizations'

Dave Taht says:

====================
Two small fq_codel optimizations

These two patches improve fq_codel performance
under extreme network loads. The first patch
more rapidly escalates the codel count under
overload, the second just kills a totally useless
statistic.

(sent together because they'd otherwise conflict)
====================

Signed-off-by: Dave Taht <dave.taht@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agofq_codel: Kill useless per-flow dropped statistic
Dave Taht [Sat, 3 Aug 2019 23:37:29 +0000 (16:37 -0700)]
fq_codel: Kill useless per-flow dropped statistic

It is almost impossible to get anything other than a 0 out of
flow->dropped statistic with a tc class dump, as it resets to 0
on every round.

It also conflates ecn marks with drops.

It would have been useful had it kept a cumulative drop count, but
it doesn't. This patch doesn't change the API, it just stops
tracking a stat and state that is impossible to measure and nobody
uses.

Signed-off-by: Dave Taht <dave.taht@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoIncrease fq_codel count in the bulk dropper
Dave Taht [Sat, 3 Aug 2019 23:37:28 +0000 (16:37 -0700)]
Increase fq_codel count in the bulk dropper

In the field fq_codel is often used with a smaller memory or
packet limit than the default, and when the bulk dropper is hit,
the drop pattern bifircates into one that more slowly increases
the codel drop rate and hits the bulk dropper more than it should.

The scan through the 1024 queues happens more often than it needs to.

This patch increases the codel count in the bulk dropper, but
does not change the drop rate there, relying on the next codel round
to deliver the next packet at the original drop rate
(after that burst of loss), then escalate to a higher signaling rate.

Signed-off-by: Dave Taht <dave.taht@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'flow_offload-action-fixes'
David S. Miller [Tue, 6 Aug 2019 21:15:39 +0000 (14:15 -0700)]
Merge branch 'flow_offload-action-fixes'

Vlad Buslov says:

====================
action fixes for flow_offload infra compatibility

Fix rcu warnings due to usage of action helpers that expect rcu read lock
protection from rtnl-protected context of flow_offload infra.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: sample: allow accessing psample_group with rtnl
Vlad Buslov [Sat, 3 Aug 2019 13:36:19 +0000 (16:36 +0300)]
net: sched: sample: allow accessing psample_group with rtnl

Recently implemented support for sample action in flow_offload infra leads
to following rcu usage warning:

[ 1938.234856] =============================
[ 1938.234858] WARNING: suspicious RCU usage
[ 1938.234863] 5.3.0-rc1+ #574 Not tainted
[ 1938.234866] -----------------------------
[ 1938.234869] include/net/tc_act/tc_sample.h:47 suspicious rcu_dereference_check() usage!
[ 1938.234872]
               other info that might help us debug this:

[ 1938.234875]
               rcu_scheduler_active = 2, debug_locks = 1
[ 1938.234879] 1 lock held by tc/19540:
[ 1938.234881]  #0: 00000000b03cb918 (rtnl_mutex){+.+.}, at: tc_new_tfilter+0x47c/0x970
[ 1938.234900]
               stack backtrace:
[ 1938.234905] CPU: 2 PID: 19540 Comm: tc Not tainted 5.3.0-rc1+ #574
[ 1938.234908] Hardware name: Supermicro SYS-2028TP-DECR/X10DRT-P, BIOS 2.0b 03/30/2017
[ 1938.234911] Call Trace:
[ 1938.234922]  dump_stack+0x85/0xc0
[ 1938.234930]  tc_setup_flow_action+0xed5/0x2040
[ 1938.234944]  fl_hw_replace_filter+0x11f/0x2e0 [cls_flower]
[ 1938.234965]  fl_change+0xd24/0x1b30 [cls_flower]
[ 1938.234990]  tc_new_tfilter+0x3e0/0x970
[ 1938.235021]  ? tc_del_tfilter+0x720/0x720
[ 1938.235028]  rtnetlink_rcv_msg+0x389/0x4b0
[ 1938.235038]  ? netlink_deliver_tap+0x95/0x400
[ 1938.235044]  ? rtnl_dellink+0x2d0/0x2d0
[ 1938.235053]  netlink_rcv_skb+0x49/0x110
[ 1938.235063]  netlink_unicast+0x171/0x200
[ 1938.235073]  netlink_sendmsg+0x224/0x3f0
[ 1938.235091]  sock_sendmsg+0x5e/0x60
[ 1938.235097]  ___sys_sendmsg+0x2ae/0x330
[ 1938.235111]  ? __handle_mm_fault+0x12cd/0x19e0
[ 1938.235125]  ? __handle_mm_fault+0x12cd/0x19e0
[ 1938.235138]  ? find_held_lock+0x2b/0x80
[ 1938.235147]  ? do_user_addr_fault+0x22d/0x490
[ 1938.235160]  __sys_sendmsg+0x59/0xa0
[ 1938.235178]  do_syscall_64+0x5c/0xb0
[ 1938.235187]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 1938.235192] RIP: 0033:0x7ff9a4d597b8
[ 1938.235197] Code: 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 48 8d 05 65 8f 0c 00 8b 00 85 c0 75 17 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 58 c3 0f 1f 80 00 00 00 00 48 83
 ec 28 89 54
[ 1938.235200] RSP: 002b:00007ffcfe381c48 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[ 1938.235205] RAX: ffffffffffffffda RBX: 000000005d4497f9 RCX: 00007ff9a4d597b8
[ 1938.235208] RDX: 0000000000000000 RSI: 00007ffcfe381cb0 RDI: 0000000000000003
[ 1938.235211] RBP: 0000000000000000 R08: 0000000000000001 R09: 0000000000000006
[ 1938.235214] R10: 0000000000404ec2 R11: 0000000000000246 R12: 0000000000000001
[ 1938.235217] R13: 0000000000480640 R14: 0000000000000012 R15: 0000000000000001

Change tcf_sample_psample_group() helper to allow using it from both rtnl
and rcu protected contexts.

Fixes: a7a7be6087b0 ("net/sched: add sample action to the hardware intermediate representation")
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: sched: police: allow accessing police->params with rtnl
Vlad Buslov [Sat, 3 Aug 2019 13:36:18 +0000 (16:36 +0300)]
net: sched: police: allow accessing police->params with rtnl

Recently implemented support for police action in flow_offload infra leads
to following rcu usage warning:

[ 1925.881092] =============================
[ 1925.881094] WARNING: suspicious RCU usage
[ 1925.881098] 5.3.0-rc1+ #574 Not tainted
[ 1925.881100] -----------------------------
[ 1925.881104] include/net/tc_act/tc_police.h:57 suspicious rcu_dereference_check() usage!
[ 1925.881106]
               other info that might help us debug this:

[ 1925.881109]
               rcu_scheduler_active = 2, debug_locks = 1
[ 1925.881112] 1 lock held by tc/18591:
[ 1925.881115]  #0: 00000000b03cb918 (rtnl_mutex){+.+.}, at: tc_new_tfilter+0x47c/0x970
[ 1925.881124]
               stack backtrace:
[ 1925.881127] CPU: 2 PID: 18591 Comm: tc Not tainted 5.3.0-rc1+ #574
[ 1925.881130] Hardware name: Supermicro SYS-2028TP-DECR/X10DRT-P, BIOS 2.0b 03/30/2017
[ 1925.881132] Call Trace:
[ 1925.881138]  dump_stack+0x85/0xc0
[ 1925.881145]  tc_setup_flow_action+0x1771/0x2040
[ 1925.881155]  fl_hw_replace_filter+0x11f/0x2e0 [cls_flower]
[ 1925.881175]  fl_change+0xd24/0x1b30 [cls_flower]
[ 1925.881200]  tc_new_tfilter+0x3e0/0x970
[ 1925.881231]  ? tc_del_tfilter+0x720/0x720
[ 1925.881243]  rtnetlink_rcv_msg+0x389/0x4b0
[ 1925.881250]  ? netlink_deliver_tap+0x95/0x400
[ 1925.881257]  ? rtnl_dellink+0x2d0/0x2d0
[ 1925.881264]  netlink_rcv_skb+0x49/0x110
[ 1925.881275]  netlink_unicast+0x171/0x200
[ 1925.881284]  netlink_sendmsg+0x224/0x3f0
[ 1925.881299]  sock_sendmsg+0x5e/0x60
[ 1925.881305]  ___sys_sendmsg+0x2ae/0x330
[ 1925.881309]  ? task_work_add+0x43/0x50
[ 1925.881314]  ? fput_many+0x45/0x80
[ 1925.881329]  ? __lock_acquire+0x248/0x1930
[ 1925.881342]  ? find_held_lock+0x2b/0x80
[ 1925.881347]  ? task_work_run+0x7b/0xd0
[ 1925.881359]  __sys_sendmsg+0x59/0xa0
[ 1925.881375]  do_syscall_64+0x5c/0xb0
[ 1925.881381]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 1925.881384] RIP: 0033:0x7feb245047b8
[ 1925.881388] Code: 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 48 8d 05 65 8f 0c 00 8b 00 85 c0 75 17 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 58 c3 0f 1f 80 00 00 00 00 48 83
 ec 28 89 54
[ 1925.881391] RSP: 002b:00007ffc2d2a5788 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[ 1925.881395] RAX: ffffffffffffffda RBX: 000000005d4497ed RCX: 00007feb245047b8
[ 1925.881398] RDX: 0000000000000000 RSI: 00007ffc2d2a57f0 RDI: 0000000000000003
[ 1925.881400] RBP: 0000000000000000 R08: 0000000000000001 R09: 0000000000000006
[ 1925.881403] R10: 0000000000404ec2 R11: 0000000000000246 R12: 0000000000000001
[ 1925.881406] R13: 0000000000480640 R14: 0000000000000012 R15: 0000000000000001

Change tcf_police_rate_bytes_ps() and tcf_police_tcfp_burst() helpers to
allow using them from both rtnl and rcu protected contexts.

Fixes: 8c8cfc6ed274 ("net/sched: add police action to the hardware intermediate representation")
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'hisilicon-fixes'
David S. Miller [Tue, 6 Aug 2019 21:14:01 +0000 (14:14 -0700)]
Merge branch 'hisilicon-fixes'

Jiangfeng Xiao says:

====================
net: hisilicon: Fix a few problems with hip04_eth

During the use of the hip04_eth driver,
several problems were found,
which solved the hip04_tx_reclaim reentry problem,
fixed the problem that hip04_mac_start_xmit never
returns NETDEV_TX_BUSY
and the dma_map_single failed on the arm64 platform.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: hisilicon: Fix dma_map_single failed on arm64
Jiangfeng Xiao [Sat, 3 Aug 2019 12:31:41 +0000 (20:31 +0800)]
net: hisilicon: Fix dma_map_single failed on arm64

On the arm64 platform, executing "ifconfig eth0 up" will fail,
returning "ifconfig: SIOCSIFFLAGS: Input/output error."

ndev->dev is not initialized, dma_map_single->get_dma_ops->
dummy_dma_ops->__dummy_map_page will return DMA_ERROR_CODE
directly, so when we use dma_map_single, the first parameter
is to use the device of platform_device.

Signed-off-by: Jiangfeng Xiao <xiaojiangfeng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: hisilicon: fix hip04-xmit never return TX_BUSY
Jiangfeng Xiao [Sat, 3 Aug 2019 12:31:40 +0000 (20:31 +0800)]
net: hisilicon: fix hip04-xmit never return TX_BUSY

TX_DESC_NUM is 256, in tx_count, the maximum value of
mod(TX_DESC_NUM - 1) is 254, the variable "count" in
the hip04_mac_start_xmit function is never equal to
(TX_DESC_NUM - 1), so hip04_mac_start_xmit never
return NETDEV_TX_BUSY.

tx_count is modified to mod(TX_DESC_NUM) so that
the maximum value of tx_count can reach
(TX_DESC_NUM - 1), then hip04_mac_start_xmit can reurn
NETDEV_TX_BUSY.

Signed-off-by: Jiangfeng Xiao <xiaojiangfeng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: hisilicon: make hip04_tx_reclaim non-reentrant
Jiangfeng Xiao [Sat, 3 Aug 2019 12:31:39 +0000 (20:31 +0800)]
net: hisilicon: make hip04_tx_reclaim non-reentrant

If hip04_tx_reclaim is interrupted while it is running
and then __napi_schedule continues to execute
hip04_rx_poll->hip04_tx_reclaim, reentrancy occurs
and oops is generated. So you need to mask the interrupt
during the hip04_tx_reclaim run.

The kernel oops exception stack is as follows:

Unable to handle kernel NULL pointer dereference
at virtual address 00000050
pgd = c0003000
[00000050] *pgd=80000000a04003, *pmd=00000000
Internal error: Oops: 206 [#1] SMP ARM
Modules linked in: hip04_eth mtdblock mtd_blkdevs mtd
ohci_platform ehci_platform ohci_hcd ehci_hcd
vfat fat sd_mod usb_storage scsi_mod usbcore usb_common
CPU: 0 PID: 0 Comm: swapper/0 Tainted: G           O    4.4.185 #1
Hardware name: Hisilicon A15
task: c0a250e0 task.stack: c0a00000
PC is at hip04_tx_reclaim+0xe0/0x17c [hip04_eth]
LR is at hip04_tx_reclaim+0x30/0x17c [hip04_eth]
pc : [<bf30c3a4>]    lr : [<bf30c2f4>]    psr: 600e0313
sp : c0a01d88  ip : 00000000  fp : c0601f9c
r10: 00000000  r9 : c3482380  r8 : 00000001
r7 : 00000000  r6 : 000000e1  r5 : c3482000  r4 : 0000000c
r3 : f2209800  r2 : 00000000  r1 : 00000000  r0 : 00000000
Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
Control: 32c5387d  Table: 03d28c80  DAC: 55555555
Process swapper/0 (pid: 0, stack limit = 0xc0a00190)
Stack: (0xc0a01d88 to 0xc0a02000)
[<bf30c3a4>] (hip04_tx_reclaim [hip04_eth]) from [<bf30d2e0>]
                                                (hip04_rx_poll+0x88/0x368 [hip04_eth])
[<bf30d2e0>] (hip04_rx_poll [hip04_eth]) from [<c04c2d9c>] (net_rx_action+0x114/0x34c)
[<c04c2d9c>] (net_rx_action) from [<c021eed8>] (__do_softirq+0x218/0x318)
[<c021eed8>] (__do_softirq) from [<c021f284>] (irq_exit+0x88/0xac)
[<c021f284>] (irq_exit) from [<c0240090>] (msa_irq_exit+0x11c/0x1d4)
[<c0240090>] (msa_irq_exit) from [<c02677e0>] (__handle_domain_irq+0x110/0x148)
[<c02677e0>] (__handle_domain_irq) from [<c0201588>] (gic_handle_irq+0xd4/0x118)
[<c0201588>] (gic_handle_irq) from [<c0551700>] (__irq_svc+0x40/0x58)
Exception stack(0xc0a01f30 to 0xc0a01f78)
1f20:                                     c0ae8b40 00000000 00000000 00000000
1f40: 00000002 ffffe000 c0601f9c 00000000 ffffffff c0a2257c c0a22440 c0831a38
1f60: c0a01ec4 c0a01f80 c0203714 c0203718 600e0213 ffffffff
[<c0551700>] (__irq_svc) from [<c0203718>] (arch_cpu_idle+0x20/0x3c)
[<c0203718>] (arch_cpu_idle) from [<c025bfd8>] (cpu_startup_entry+0x244/0x29c)
[<c025bfd8>] (cpu_startup_entry) from [<c054b0d8>] (rest_init+0xc8/0x10c)
[<c054b0d8>] (rest_init) from [<c0800c58>] (start_kernel+0x468/0x514)
Code: a40599e5 016086e2 018088e2 7660efe6 (503090e5)
---[ end trace 1db21d6d09c49d74 ]---
Kernel panic - not syncing: Fatal exception in interrupt
CPU3: stopping
CPU: 3 PID: 0 Comm: swapper/3 Tainted: G      D    O    4.4.185 #1

Signed-off-by: Jiangfeng Xiao <xiaojiangfeng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: mdio-octeon: Fix Kconfig warnings and build errors
Nathan Chancellor [Sat, 3 Aug 2019 06:01:56 +0000 (23:01 -0700)]
net: mdio-octeon: Fix Kconfig warnings and build errors

After commit 171a9bae68c7 ("staging/octeon: Allow test build on
!MIPS"), the following combination of configs cause a few Kconfig
warnings and build errors (distilled from arm allyesconfig and Randy's
randconfig builds):

    CONFIG_NETDEVICES=y
    CONFIG_STAGING=y
    CONFIG_COMPILE_TEST=y

and CONFIG_OCTEON_ETHERNET as either a module or built-in.

WARNING: unmet direct dependencies detected for MDIO_OCTEON
  Depends on [n]: NETDEVICES [=y] && MDIO_DEVICE [=y] && MDIO_BUS [=y]
&& 64BIT [=n] && HAS_IOMEM [=y] && OF_MDIO [=n]
  Selected by [y]:
  - OCTEON_ETHERNET [=y] && STAGING [=y] && (CAVIUM_OCTEON_SOC ||
COMPILE_TEST [=y]) && NETDEVICES [=y]

In file included from ../drivers/net/phy/mdio-octeon.c:14:
../drivers/net/phy/mdio-cavium.h:111:36: error: implicit declaration of
function â€˜writeq’; did you mean â€˜writel’?
[-Werror=implicit-function-declaration]
  111 | #define oct_mdio_writeq(val, addr) writeq(val, (void *)addr)
      |                                    ^~~~~~

CONFIG_64BIT is not strictly necessary if the proper readq/writeq
definitions are included from io-64-nonatomic-lo-hi.h.

CONFIG_OF_MDIO is not needed when CONFIG_COMPILE_TEST is enabled because
of commit f9dc9ac51610 ("of/mdio: Add dummy functions in of_mdio.h.").

Fixes: 171a9bae68c7 ("staging/octeon: Allow test build on !MIPS")
Reported-by: kbuild test robot <lkp@intel.com>
Reported-by: Mark Brown <broonie@kernel.org>
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: dump CPU port regs through master
Vivien Didelot [Fri, 2 Aug 2019 19:34:55 +0000 (15:34 -0400)]
net: dsa: dump CPU port regs through master

Merge the CPU port registers dump into the master interface registers
dump through ethtool, by nesting the ethtool_drvinfo and ethtool_regs
structures of the CPU port into the dump.

drvinfo->regdump_len will contain the full data length, while regs->len
will contain only the master interface registers dump length.

This allows for example to dump the CPU port registers on a ZII Dev
C board like this:

    # ethtool -d eth1
    0x004:                                              0x00000000
    0x008:                                              0x0a8000aa
    0x010:                                              0x01000000
    0x014:                                              0x00000000
    0x024:                                              0xf0000102
    0x040:                                              0x6d82c800
    0x044:                                              0x00000020
    0x064:                                              0x40000000
    0x084: RCR (Receive Control Register)               0x47c00104
        MAX_FL (Maximum frame length)                   1984
        FCE (Flow control enable)                       0
        BC_REJ (Broadcast frame reject)                 0
        PROM (Promiscuous mode)                         0
        DRT (Disable receive on transmit)               0
        LOOP (Internal loopback)                        0
    0x0c4: TCR (Transmit Control Register)              0x00000004
        RFC_PAUSE (Receive frame control pause)         0
        TFC_PAUSE (Transmit frame control pause)        0
        FDEN (Full duplex enable)                       1
        HBC (Heartbeat control)                         0
        GTS (Graceful transmit stop)                    0
    0x0e4:                                              0x76735d6d
    0x0e8:                                              0x7e9e8808
    0x0ec:                                              0x00010000
    .
    .
    .
    88E6352  Switch Port Registers
    ------------------------------
    00: Port Status                            0x4d04
          Pause Enabled                        0
          My Pause                             1
          802.3 PHY Detected                   0
          Link Status                          Up
          Duplex                               Full
          Speed                                100 or 200 Mbps
          EEE Enabled                          0
          Transmitter Paused                   0
          Flow Control                         0
          Config Mode                          0x4
    01: Physical Control                       0x003d
          RGMII Receive Timing Control         Default
          RGMII Transmit Timing Control        Default
          200 BASE Mode                        100
          Flow Control's Forced value          0
          Force Flow Control                   0
          Link's Forced value                  Up
          Force Link                           1
          Duplex's Forced value                Full
          Force Duplex                         1
          Force Speed                          100 or 200 Mbps
    .
    .
    .

Signed-off-by: Vivien Didelot <vivien.didelot@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'Fix-batched-event-generation-for-vlan-action'
David S. Miller [Tue, 6 Aug 2019 21:05:40 +0000 (14:05 -0700)]
Merge branch 'Fix-batched-event-generation-for-vlan-action'

Roman Mashak says:

====================
Fix batched event generation for vlan action

When adding or deleting a batch of entries, the kernel sends up to
TCA_ACT_MAX_PRIO (defined to 32 in kernel) entries in an event to user
space. However it does not consider that the action sizes may vary and
require different skb sizes.

For example, consider the following script adding 32 entries with all
supported vlan parameters (in order to maximize netlink messages size):

% cat tc-batch.sh
TC="sudo /mnt/iproute2.git/tc/tc"

$TC actions flush action vlan
for i in `seq 1 $1`;
do
   cmd="action vlan push protocol 802.1q id 4094 priority 7 pipe \
               index $i cookie aabbccddeeff112233445566778800a1 "
   args=$args$cmd
done
$TC actions add $args
%
% ./tc-batch.sh 32
Error: Failed to fill netlink attributes while adding TC action.
We have an error talking to the kernel
%

patch 1 adds callback in tc_action_ops of vlan action, which calculates
the action size, and passes size to tcf_add_notify()/tcf_del_notify().

patch 2 updates the TDC test suite with relevant vlan test cases.
====================

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agotc-testing: updated vlan action tests with batch create/delete
Roman Mashak [Fri, 2 Aug 2019 19:16:47 +0000 (15:16 -0400)]
tc-testing: updated vlan action tests with batch create/delete

Update TDC tests with cases varifying ability of TC to install or delete
batches of vlan actions.

Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet sched: update vlan action for batched events operations
Roman Mashak [Fri, 2 Aug 2019 19:16:46 +0000 (15:16 -0400)]
net sched: update vlan action for batched events operations

Add get_fill_size() routine used to calculate the action size
when building a batch of events.

Fixes: c7e2b9689 ("sched: introduce vlan action")
Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'mips_fixes_5.3_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips...
Linus Torvalds [Tue, 6 Aug 2019 21:01:08 +0000 (14:01 -0700)]
Merge tag 'mips_fixes_5.3_1' of git://git./linux/kernel/git/mips/linux

Pull MIPS fixes from Paul Burton:
 "A few MIPS fixes for 5.3:

   - Various switch fall through annotations to fixup warnings & errors
     resulting from -Wimplicit-fallthrough.

   - A fix for systems (at least jazz) using an i8253 PIT as clocksource
     when it's not suitably configured.

   - Set struct cacheinfo's cpu_map_populated field to true, indicating
     that we filled in cache info detected from cop0 registers &
     avoiding complaints about that info being (intentionally) missing
     in devicetree"

* tag 'mips_fixes_5.3_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  MIPS: BCM63XX: Mark expected switch fall-through
  MIPS: OProfile: Mark expected switch fall-throughs
  MIPS: Annotate fall-through in Cavium Octeon code
  MIPS: Annotate fall-through in kvm/emulate.c
  mips: fix cacheinfo
  MIPS: kernel: only use i8253 clocksource with periodic clockevent

4 years agoMerge branch 'drop_monitor-Various-improvements-and-cleanups'
David S. Miller [Tue, 6 Aug 2019 19:37:56 +0000 (12:37 -0700)]
Merge branch 'drop_monitor-Various-improvements-and-cleanups'

Ido Schimmel says:

====================
drop_monitor: Various improvements and cleanups

This patchset performs various improvements and cleanups in drop monitor
with no functional changes intended. There are no changes in these
patches relative to the RFC I sent two weeks ago [1].

A followup patchset will extend drop monitor with a packet alert mode in
which the dropped packet is notified to user space instead of just a
summary of recent drops. Subsequent patchsets will add the ability to
monitor hardware originated drops via drop monitor.

[1] https://patchwork.ozlabs.org/cover/1135226/
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Use pre_doit / post_doit hooks
Ido Schimmel [Tue, 6 Aug 2019 13:19:56 +0000 (16:19 +0300)]
drop_monitor: Use pre_doit / post_doit hooks

Each operation from user space should be protected by the global drop
monitor mutex. Use the pre_doit / post_doit hooks to take / release the
lock instead of doing it explicitly in each function.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Add extack support
Ido Schimmel [Tue, 6 Aug 2019 13:19:55 +0000 (16:19 +0300)]
drop_monitor: Add extack support

Add various extack messages to make drop_monitor more user friendly.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Avoid multiple blank lines
Ido Schimmel [Tue, 6 Aug 2019 13:19:54 +0000 (16:19 +0300)]
drop_monitor: Avoid multiple blank lines

Remove multiple blank lines which are visually annoying and useless.

This suppresses the "Please don't use multiple blank lines" checkpatch
messages.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Document scope of spinlock
Ido Schimmel [Tue, 6 Aug 2019 13:19:53 +0000 (16:19 +0300)]
drop_monitor: Document scope of spinlock

While 'per_cpu_dm_data' is a per-CPU variable, its 'skb' and
'send_timer' fields can be accessed concurrently by the CPU sending the
netlink notification to user space from the workqueue and the CPU
tracing kfree_skb(). This spinlock is meant to protect against that.

Document its scope and suppress the checkpatch message "spinlock_t
definition without comment".

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Rename and document scope of mutex
Ido Schimmel [Tue, 6 Aug 2019 13:19:52 +0000 (16:19 +0300)]
drop_monitor: Rename and document scope of mutex

The 'trace_state_mutex' does not only protect the global 'trace_state'
variable, but also the global 'hw_stats_list'.

Subsequent patches are going add more operations from user space to
drop_monitor and these all need to be mutually exclusive.

Rename 'trace_state_mutex' to the more fitting 'net_dm_mutex' name and
document its scope.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrop_monitor: Use correct error code
Ido Schimmel [Tue, 6 Aug 2019 13:19:51 +0000 (16:19 +0300)]
drop_monitor: Use correct error code

The error code 'ENOTSUPP' is reserved for use with NFS. Use 'EOPNOTSUPP'
instead.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'stmmac-fixes'
David S. Miller [Tue, 6 Aug 2019 19:26:11 +0000 (12:26 -0700)]
Merge branch 'stmmac-fixes'

Jose Abreu says:

====================
net: stmmac: Fixes for -net

Couple of fixes for -net. More info in commit log.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: stmmac: tc: Do not return a fragment entry
Jose Abreu [Tue, 6 Aug 2019 13:16:18 +0000 (15:16 +0200)]
net: stmmac: tc: Do not return a fragment entry

Do not try to return a fragment entry from TC list. Otherwise we may not
clean properly allocated entries.

Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: stmmac: Fix issues when number of Queues >= 4
Jose Abreu [Tue, 6 Aug 2019 13:16:17 +0000 (15:16 +0200)]
net: stmmac: Fix issues when number of Queues >= 4

When queues >= 4 we use different registers but we were not subtracting
the offset of 4. Fix this.

Found out by Coverity.

Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: stmmac: xgmac: Fix XGMAC selftests
Jose Abreu [Tue, 6 Aug 2019 13:16:16 +0000 (15:16 +0200)]
net: stmmac: xgmac: Fix XGMAC selftests

Fixup the XGMAC selftests by correctly finishing the implementation of
set_filter callback.

Result:
$ ethtool -t enp4s0
The test result is PASS
The test extra info:
 1. MAC Loopback           0
 2. PHY Loopback           -95
 3. MMC Counters           -95
 4. EEE                    -95
 5. Hash Filter MC         0
 6. Perfect Filter UC      0
 7. MC Filter              0
 8. UC Filter              0
 9. Flow Control           0

Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: ksz: Drop NET_DSA_TAG_KSZ9477
Marek Vasut [Tue, 6 Aug 2019 13:06:09 +0000 (15:06 +0200)]
net: dsa: ksz: Drop NET_DSA_TAG_KSZ9477

This Kconfig option is unused, drop it.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: David S. Miller <davem@davemloft.net>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Tristram Ha <Tristram.Ha@microchip.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: ksz: Merge ksz_priv.h into ksz_common.h
Marek Vasut [Tue, 6 Aug 2019 13:06:08 +0000 (15:06 +0200)]
net: dsa: ksz: Merge ksz_priv.h into ksz_common.h

Merge the two headers into one, no functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: David S. Miller <davem@davemloft.net>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Tristram Ha <Tristram.Ha@microchip.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: dsa: ksz: Remove dead code and fix warnings
Marek Vasut [Tue, 6 Aug 2019 13:06:07 +0000 (15:06 +0200)]
net: dsa: ksz: Remove dead code and fix warnings

Remove ksz_port_cleanup(), which is unused. Add missing include
"ksz_common.h", which fixes the following warning when built with
make ... W=1

drivers/net/dsa/microchip/ksz_common.c:23:6: warning: no previous prototype for â€˜...’ [-Wmissing-prototypes]

Note that the order of the headers cannot be swapped, as that would
trigger missing forward declaration errors, which would indicate the
way forward is to merge the two headers into one.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: David S. Miller <davem@davemloft.net>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Tristram Ha <Tristram.Ha@microchip.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'wireless-drivers-for-davem-2019-08-06' of git://git.kernel.org/pub/scm...
David S. Miller [Tue, 6 Aug 2019 18:48:48 +0000 (11:48 -0700)]
Merge tag 'wireless-drivers-for-davem-2019-08-06' of git://git./linux/kernel/git/kvalo/wireless-drivers

Kalle Valo says:

====================
wireless-drivers fixes for 5.3

Second set of fixes for 5.3. Lots of iwlwifi fixes have accumulated
which consists most of patches in this pull request. Only most notable
iwlwifi fixes are listed below.

mwifiex

* fix a regression related to WPA1 networks since v5.3-rc1

iwlwifi

* fix use-after-free issues

* fix DMA mapping API usage errors

* fix frame drop occurring due to reorder buffer handling in
  RSS in certain conditions

* fix rate scale locking issues

* disable TX A-MSDU on older NICs as it causes problems and was
  never supposed to be supported

* new PCI IDs

* GEO_TX_POWER_LIMIT API issue that many people were hitting
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Linus Torvalds [Tue, 6 Aug 2019 18:47:23 +0000 (11:47 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/hid/hid

Pull HID fixes from Jiri Kosina:

 - functional regression fix for some of the Logitech unifying devices,
   from Hans de Goede

 - race condition fix in hid-sony for bug severely affecting
   Valve/Android deployments, from Roderick Colenbrander

 - several fixes for issues found by syzbot/kasan, from Oliver Neukum
   and Hillf Danton

 - functional regression fix for Wacom Cintiq device, from Aaron
   Armstrong Skomra

 - a few other assorted device-specific quirks

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
  HID: sony: Fix race condition between rumble and device remove.
  HID: hiddev: do cleanup in failure of opening a device
  HID: hiddev: avoid opening a disconnected device
  HID: input: fix a4tech horizontal wheel custom usage
  HID: Add quirk for HP X1200 PIXART OEM mouse
  HID: holtek: test for sanity of intfdata
  HID: wacom: fix bit shift for Cintiq Companion 2
  HID: quirks: Set the INCREMENT_USAGE_ON_DUPLICATE quirk on Saitek X52
  HID: logitech-dj: Really fix return value of logi_dj_recv_query_hidpp_devices
  HID: Add 044f:b320 ThrustMaster, Inc. 2 in 1 DT
  HID: logitech-dj: add the Powerplay receiver
  HID: logitech-hidpp: add USB PID for a few more supported mice
  HID: logitech-dj: rename "gaming" receiver to "lightspeed"

4 years agobe2net: disable bh with spin_lock in be_process_mcc
Denis Kirjanov [Tue, 6 Aug 2019 10:51:11 +0000 (12:51 +0200)]
be2net: disable bh with spin_lock in be_process_mcc

be_process_mcc() is invoked in 3 different places and
always with BHs disabled except the be_poll function
but since it's invoked from softirq with BHs
disabled it won't hurt.

v1->v2: added explanation to the patch
v2->v3: add a missing call from be_cmds.c

Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: cxgb3_main: Fix a resource leak in a error path in 'init_one()'
Christophe JAILLET [Tue, 6 Aug 2019 08:55:12 +0000 (10:55 +0200)]
net: cxgb3_main: Fix a resource leak in a error path in 'init_one()'

A call to 'kfree_skb()' is missing in the error handling path of
'init_one()'.
This is already present in 'remove_one()' but is missing here.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: ethernet: sun4i-emac: Support phy-handle property for finding PHYs
Chen-Yu Tsai [Tue, 6 Aug 2019 07:35:39 +0000 (15:35 +0800)]
net: ethernet: sun4i-emac: Support phy-handle property for finding PHYs

The sun4i-emac uses the "phy" property to find the PHY it's supposed to
use. This property was deprecated in favor of "phy-handle" in commit
8c5b09447625 ("dt-bindings: net: sun4i-emac: Convert the binding to a
schemas").

Add support for this new property name, and fall back to the old one in
case the device tree hasn't been updated.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge branch 'x86/grand-schemozzle' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 6 Aug 2019 18:22:22 +0000 (11:22 -0700)]
Merge branch 'x86/grand-schemozzle' of git://git./linux/kernel/git/tip/tip

Pull pti updates from Thomas Gleixner:
 "The performance deterioration departement is not proud at all to
  present yet another set of speculation fences to mitigate the next
  chapter in the 'what could possibly go wrong' story.

  The new vulnerability belongs to the Spectre class and affects GS
  based data accesses and has therefore been dubbed 'Grand Schemozzle'
  for secret communication purposes. It's officially listed as
  CVE-2019-1125.

  Conditional branches in the entry paths which contain a SWAPGS
  instruction (interrupts and exceptions) can be mis-speculated which
  results in speculative accesses with a wrong GS base.

  This can happen on entry from user mode through a mis-speculated
  branch which takes the entry from kernel mode path and therefore does
  not execute the SWAPGS instruction. The following speculative accesses
  are done with user GS base.

  On entry from kernel mode the mis-speculated branch executes the
  SWAPGS instruction in the entry from user mode path which has the same
  effect that the following GS based accesses are done with user GS
  base.

  If there is a disclosure gadget available in these code paths the
  mis-speculated data access can be leaked through the usual side
  channels.

  The entry from user mode issue affects all CPUs which have speculative
  execution. The entry from kernel mode issue affects only Intel CPUs
  which can speculate through SWAPGS. On CPUs from other vendors SWAPGS
  has semantics which prevent that.

  SMAP migitates both problems but only when the CPU is not affected by
  the Meltdown vulnerability.

  The mitigation is to issue LFENCE instructions in the entry from
  kernel mode path for all affected CPUs and on the affected Intel CPUs
  also in the entry from user mode path unless PTI is enabled because
  the CR3 write is serializing.

  The fences are as usual enabled conditionally and can be completely
  disabled on the kernel command line. The Spectre V1 documentation is
  updated accordingly.

  A big "Thank You!" goes to Josh for doing the heavy lifting for this
  round of hardware misfeature 'repair'. Of course also "Thank You!" to
  everybody else who contributed in one way or the other"

* 'x86/grand-schemozzle' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  Documentation: Add swapgs description to the Spectre v1 documentation
  x86/speculation/swapgs: Exclude ATOMs from speculation through SWAPGS
  x86/entry/64: Use JMP instead of JMPQ
  x86/speculation: Enable Spectre v1 swapgs mitigations
  x86/speculation: Prepare entry code for Spectre v1 swapgs mitigations

4 years agoHID: sony: Fix race condition between rumble and device remove.
Roderick Colenbrander [Fri, 2 Aug 2019 22:50:19 +0000 (15:50 -0700)]
HID: sony: Fix race condition between rumble and device remove.

Valve reported a kernel crash on Ubuntu 18.04 when disconnecting a DS4
gamepad while rumble is enabled. This issue is reproducible with a
frequency of 1 in 3 times in the game Borderlands 2 when using an
automatic weapon, which triggers many rumble operations.

We found the issue to be a race condition between sony_remove and the
final device destruction by the HID / input system. The problem was
that sony_remove didn't clean some of its work_item state in
"struct sony_sc". After sony_remove work, the corresponding evdev
node was around for sufficient time for applications to still queue
rumble work after "sony_remove".

On pre-4.19 kernels the race condition caused a kernel crash due to a
NULL-pointer dereference as "sc->output_report_dmabuf" got freed during
sony_remove. On newer kernels this crash doesn't happen due the buffer
now being allocated using devm_kzalloc. However we can still queue work,
while the driver is an undefined state.

This patch fixes the described problem, by guarding the work_item
"state_worker" with an initialized variable, which we are setting back
to 0 on cleanup.

Signed-off-by: Roderick Colenbrander <roderick.colenbrander@sony.com>
CC: stable@vger.kernel.org
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
4 years agoHID: hiddev: do cleanup in failure of opening a device
Hillf Danton [Tue, 6 Aug 2019 08:40:15 +0000 (16:40 +0800)]
HID: hiddev: do cleanup in failure of opening a device

Undo what we did for opening before releasing the memory slice.

Reported-by: syzbot <syzbot+62a1e04fd3ec2abf099e@syzkaller.appspotmail.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
4 years agoHID: hiddev: avoid opening a disconnected device
Hillf Danton [Tue, 6 Aug 2019 08:38:58 +0000 (16:38 +0800)]
HID: hiddev: avoid opening a disconnected device

syzbot found the following crash on:

HEAD commit:    e96407b4 usb-fuzzer: main usb gadget fuzzer driver
git tree:       https://github.com/google/kasan.git usb-fuzzer
console output: https://syzkaller.appspot.com/x/log.txt?x=147ac20c600000
kernel config:  https://syzkaller.appspot.com/x/.config?x=792eb47789f57810
link: https://syzkaller.appspot.com/bug?extid=62a1e04fd3ec2abf099e
compiler:       gcc (GCC) 9.0.0 20181231 (experimental)

==================================================================
BUG: KASAN: use-after-free in __lock_acquire+0x302a/0x3b50
kernel/locking/lockdep.c:3753
Read of size 8 at addr ffff8881cf591a08 by task syz-executor.1/26260

CPU: 1 PID: 26260 Comm: syz-executor.1 Not tainted 5.3.0-rc2+ #24
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
Google 01/01/2011
Call Trace:
  __dump_stack lib/dump_stack.c:77 [inline]
  dump_stack+0xca/0x13e lib/dump_stack.c:113
  print_address_description+0x6a/0x32c mm/kasan/report.c:351
  __kasan_report.cold+0x1a/0x33 mm/kasan/report.c:482
  kasan_report+0xe/0x12 mm/kasan/common.c:612
  __lock_acquire+0x302a/0x3b50 kernel/locking/lockdep.c:3753
  lock_acquire+0x127/0x320 kernel/locking/lockdep.c:4412
  __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
  _raw_spin_lock_irqsave+0x32/0x50 kernel/locking/spinlock.c:159
  hiddev_release+0x82/0x520 drivers/hid/usbhid/hiddev.c:221
  __fput+0x2d7/0x840 fs/file_table.c:280
  task_work_run+0x13f/0x1c0 kernel/task_work.c:113
  exit_task_work include/linux/task_work.h:22 [inline]
  do_exit+0x8ef/0x2c50 kernel/exit.c:878
  do_group_exit+0x125/0x340 kernel/exit.c:982
  get_signal+0x466/0x23d0 kernel/signal.c:2728
  do_signal+0x88/0x14e0 arch/x86/kernel/signal.c:815
  exit_to_usermode_loop+0x1a2/0x200 arch/x86/entry/common.c:159
  prepare_exit_to_usermode arch/x86/entry/common.c:194 [inline]
  syscall_return_slowpath arch/x86/entry/common.c:274 [inline]
  do_syscall_64+0x45f/0x580 arch/x86/entry/common.c:299
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x459829
Code: fd b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7
48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff
ff 0f 83 cb b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007f75b2a6ccf8 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
RAX: fffffffffffffe00 RBX: 000000000075c078 RCX: 0000000000459829
RDX: 0000000000000000 RSI: 0000000000000080 RDI: 000000000075c078
RBP: 000000000075c070 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 000000000075c07c
R13: 00007ffcdfe1023f R14: 00007f75b2a6d9c0 R15: 000000000075c07c

Allocated by task 104:
  save_stack+0x1b/0x80 mm/kasan/common.c:69
  set_track mm/kasan/common.c:77 [inline]
  __kasan_kmalloc mm/kasan/common.c:487 [inline]
  __kasan_kmalloc.constprop.0+0xbf/0xd0 mm/kasan/common.c:460
  kmalloc include/linux/slab.h:552 [inline]
  kzalloc include/linux/slab.h:748 [inline]
  hiddev_connect+0x242/0x5b0 drivers/hid/usbhid/hiddev.c:900
  hid_connect+0x239/0xbb0 drivers/hid/hid-core.c:1882
  hid_hw_start drivers/hid/hid-core.c:1981 [inline]
  hid_hw_start+0xa2/0x130 drivers/hid/hid-core.c:1972
  appleir_probe+0x13e/0x1a0 drivers/hid/hid-appleir.c:308
  hid_device_probe+0x2be/0x3f0 drivers/hid/hid-core.c:2209
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  hid_add_device+0x33c/0x990 drivers/hid/hid-core.c:2365
  usbhid_probe+0xa81/0xfa0 drivers/hid/usbhid/hid-core.c:1386
  usb_probe_interface+0x305/0x7a0 drivers/usb/core/driver.c:361
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  usb_set_configuration+0xdf6/0x1670 drivers/usb/core/message.c:2023
  generic_probe+0x9d/0xd5 drivers/usb/core/generic.c:210
  usb_probe_device+0x99/0x100 drivers/usb/core/driver.c:266
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  usb_new_device.cold+0x6a4/0xe79 drivers/usb/core/hub.c:2536
  hub_port_connect drivers/usb/core/hub.c:5098 [inline]
  hub_port_connect_change drivers/usb/core/hub.c:5213 [inline]
  port_event drivers/usb/core/hub.c:5359 [inline]
  hub_event+0x1b5c/0x3640 drivers/usb/core/hub.c:5441
  process_one_work+0x92b/0x1530 kernel/workqueue.c:2269
  worker_thread+0x96/0xe20 kernel/workqueue.c:2415
  kthread+0x318/0x420 kernel/kthread.c:255
  ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:352

Freed by task 104:
  save_stack+0x1b/0x80 mm/kasan/common.c:69
  set_track mm/kasan/common.c:77 [inline]
  __kasan_slab_free+0x130/0x180 mm/kasan/common.c:449
  slab_free_hook mm/slub.c:1423 [inline]
  slab_free_freelist_hook mm/slub.c:1470 [inline]
  slab_free mm/slub.c:3012 [inline]
  kfree+0xe4/0x2f0 mm/slub.c:3953
  hiddev_connect.cold+0x45/0x5c drivers/hid/usbhid/hiddev.c:914
  hid_connect+0x239/0xbb0 drivers/hid/hid-core.c:1882
  hid_hw_start drivers/hid/hid-core.c:1981 [inline]
  hid_hw_start+0xa2/0x130 drivers/hid/hid-core.c:1972
  appleir_probe+0x13e/0x1a0 drivers/hid/hid-appleir.c:308
  hid_device_probe+0x2be/0x3f0 drivers/hid/hid-core.c:2209
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  hid_add_device+0x33c/0x990 drivers/hid/hid-core.c:2365
  usbhid_probe+0xa81/0xfa0 drivers/hid/usbhid/hid-core.c:1386
  usb_probe_interface+0x305/0x7a0 drivers/usb/core/driver.c:361
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  usb_set_configuration+0xdf6/0x1670 drivers/usb/core/message.c:2023
  generic_probe+0x9d/0xd5 drivers/usb/core/generic.c:210
  usb_probe_device+0x99/0x100 drivers/usb/core/driver.c:266
  really_probe+0x281/0x650 drivers/base/dd.c:548
  driver_probe_device+0x101/0x1b0 drivers/base/dd.c:709
  __device_attach_driver+0x1c2/0x220 drivers/base/dd.c:816
  bus_for_each_drv+0x15c/0x1e0 drivers/base/bus.c:454
  __device_attach+0x217/0x360 drivers/base/dd.c:882
  bus_probe_device+0x1e4/0x290 drivers/base/bus.c:514
  device_add+0xae6/0x16f0 drivers/base/core.c:2114
  usb_new_device.cold+0x6a4/0xe79 drivers/usb/core/hub.c:2536
  hub_port_connect drivers/usb/core/hub.c:5098 [inline]
  hub_port_connect_change drivers/usb/core/hub.c:5213 [inline]
  port_event drivers/usb/core/hub.c:5359 [inline]
  hub_event+0x1b5c/0x3640 drivers/usb/core/hub.c:5441
  process_one_work+0x92b/0x1530 kernel/workqueue.c:2269
  worker_thread+0x96/0xe20 kernel/workqueue.c:2415
  kthread+0x318/0x420 kernel/kthread.c:255
  ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:352

The buggy address belongs to the object at ffff8881cf591900
  which belongs to the cache kmalloc-512 of size 512
The buggy address is located 264 bytes inside of
  512-byte region [ffff8881cf591900ffff8881cf591b00)
The buggy address belongs to the page:
page:ffffea00073d6400 refcount:1 mapcount:0 mapping:ffff8881da002500
index:0x0 compound_mapcount: 0
flags: 0x200000000010200(slab|head)
raw: 0200000000010200 0000000000000000 0000000100000001 ffff8881da002500
raw: 0000000000000000 00000000000c000c 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
  ffff8881cf591900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
  ffff8881cf591980: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff8881cf591a00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
                       ^
  ffff8881cf591a80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
  ffff8881cf591b00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================

In order to avoid opening a disconnected device, we need to check exist
again after acquiring the existance lock, and bail out if necessary.

Reported-by: syzbot <syzbot+62a1e04fd3ec2abf099e@syzkaller.appspotmail.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
4 years agoMIPS: BCM63XX: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 5 Aug 2019 18:55:33 +0000 (13:55 -0500)]
MIPS: BCM63XX: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning (Building: bcm63xx_defconfig mips):

arch/mips/pci/ops-bcm63xx.c: In function â€˜bcm63xx_pcie_can_access’:
arch/mips/pci/ops-bcm63xx.c:474:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (PCI_SLOT(devfn) == 0)
      ^
arch/mips/pci/ops-bcm63xx.c:477:2: note: here
  default:
  ^~~~~~~

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: James Hogan <jhogan@kernel.org>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: bcm-kernel-feedback-list@broadcom.com
Cc: linux-mips@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
4 years agoMerge tag 'linux-can-fixes-for-5.3-20190802' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Mon, 5 Aug 2019 21:44:53 +0000 (14:44 -0700)]
Merge tag 'linux-can-fixes-for-5.3-20190802' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2019-08-02

this is a pull request of 4 patches for net/master.

The first two patches are by Wang Xiayang, they force that the string buffer
during a dev_info() is properly NULL terminated.

The last two patches are by Tomas Bortoli and fix both a potential info leak of
kernel memory to USB devices.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: bridge: move default pvid init/deinit to NETDEV_REGISTER/UNREGISTER
Nikolay Aleksandrov [Fri, 2 Aug 2019 10:57:36 +0000 (13:57 +0300)]
net: bridge: move default pvid init/deinit to NETDEV_REGISTER/UNREGISTER

Most of the bridge device's vlan init bugs come from the fact that its
default pvid is created at the wrong time, way too early in ndo_init()
before the device is even assigned an ifindex. It introduces a bug when the
bridge's dev_addr is added as fdb during the initial default pvid creation
the notification has ifindex/NDA_MASTER both equal to 0 (see example below)
which really makes no sense for user-space[0] and is wrong.
Usually user-space software would ignore such entries, but they are
actually valid and will eventually have all necessary attributes.
It makes much more sense to send a notification *after* the device has
registered and has a proper ifindex allocated rather than before when
there's a chance that the registration might still fail or to receive
it with ifindex/NDA_MASTER == 0. Note that we can remove the fdb flush
from br_vlan_flush() since that case can no longer happen. At
NETDEV_REGISTER br->default_pvid is always == 1 as it's initialized by
br_vlan_init() before that and at NETDEV_UNREGISTER it can be anything
depending why it was called (if called due to NETDEV_REGISTER error
it'll still be == 1, otherwise it could be any value changed during the
device life time).

For the demonstration below a small change to iproute2 for printing all fdb
notifications is added, because it contained a workaround not to show
entries with ifindex == 0.
Command executed while monitoring: $ ip l add br0 type bridge
Before (both ifindex and master == 0):
$ bridge monitor fdb
36:7e:8a:b3:56:ba dev * vlan 1 master * permanent

After (proper br0 ifindex):
$ bridge monitor fdb
e6:2a:ae:7a:b7:48 dev br0 vlan 1 master br0 permanent

v4: move only the default pvid init/deinit to NETDEV_REGISTER/UNREGISTER
v3: send the correct v2 patch with all changes (stub should return 0)
v2: on error in br_vlan_init set br->vlgrp to NULL and return 0 in
    the br_vlan_bridge_event stub when bridge vlans are disabled

[0] https://bugzilla.kernel.org/show_bug.cgi?id=204389

Reported-by: michael-dev <michael-dev@fami-braun.de>
Fixes: 5be5a2df40f0 ("bridge: Add filtering support for default_pvid")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agodrivers/net/ethernet/marvell/mvmdio.c: Fix non OF case
Arnaud Patard (Rtp) [Fri, 2 Aug 2019 08:32:40 +0000 (10:32 +0200)]
drivers/net/ethernet/marvell/mvmdio.c: Fix non OF case

Orion5.x systems are still using machine files and not device-tree.
Commit 96cb4342382290c9 ("net: mvmdio: allow up to three clocks to be
specified for orion-mdio") has replaced devm_clk_get() with of_clk_get(),
leading to a oops at boot and not working network, as reported in
https://lists.debian.org/debian-arm/2019/07/msg00088.html and possibly in
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=908712.

Link: https://lists.debian.org/debian-arm/2019/07/msg00088.html
Fixes: 96cb4342382290c9 ("net: mvmdio: allow up to three clocks to be specified for orion-mdio")
Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet/smc: avoid fallback in case of non-blocking connect
Ursula Braun [Fri, 2 Aug 2019 08:47:50 +0000 (10:47 +0200)]
net/smc: avoid fallback in case of non-blocking connect

FASTOPEN is not possible with SMC. sendmsg() with msg_flag MSG_FASTOPEN
triggers a fallback to TCP if the socket is in state SMC_INIT.
But if a nonblocking connect is already started, fallback to TCP
is no longer possible, even though the socket may still be in state
SMC_INIT.
And if a nonblocking connect is already started, a listen() call
does not make sense.

Reported-by: syzbot+bd8cc73d665590a1fcad@syzkaller.appspotmail.com
Fixes: 50717a37db032 ("net/smc: nonblocking connect rework")
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet/smc: do not schedule tx_work in SMC_CLOSED state
Ursula Braun [Fri, 2 Aug 2019 08:16:38 +0000 (10:16 +0200)]
net/smc: do not schedule tx_work in SMC_CLOSED state

The setsockopts options TCP_NODELAY and TCP_CORK may schedule the
tx worker. Make sure the socket is not yet moved into SMC_CLOSED
state (for instance by a shutdown SHUT_RDWR call).

Reported-by: syzbot+92209502e7aab127c75f@syzkaller.appspotmail.com
Reported-by: syzbot+b972214bb803a343f4fe@syzkaller.appspotmail.com
Fixes: 01d2f7e2cdd31 ("net/smc: sockopts TCP_NODELAY and TCP_CORK")
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agocnic: Explicitly initialize all reference counts to 0.
Michael Chan [Fri, 2 Aug 2019 06:17:51 +0000 (02:17 -0400)]
cnic: Explicitly initialize all reference counts to 0.

The driver is relying on zero'ed allocated memory and does not
explicitly call atomic_set() to initialize the ref counts to 0.  Add
these atomic_set() calls so that it will be more straight forward
to convert atomic ref counts to refcount_t.

Reported-by: Chuhong Yuan <hslester96@gmail.com>
Cc: Rasesh Mody <rmody@marvell.com>
Cc: <GR-Linux-NIC-Dev@marvell.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agogve: Fix case where desc_cnt and data_cnt can get out of sync
Catherine Sullivan [Thu, 1 Aug 2019 23:07:31 +0000 (16:07 -0700)]
gve: Fix case where desc_cnt and data_cnt can get out of sync

desc_cnt and data_cnt should always be equal. In the case of a dropped
packet desc_cnt was still getting updated (correctly), data_cnt
was not. To eliminate this bug and prevent it from recurring this
patch combines them into one ring level cnt.

Signed-off-by: Catherine Sullivan <csully@google.com>
Reviewed-by: Sagi Shahar <sagis@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoipv6: have a single rcu unlock point in __ip6_rt_update_pmtu
David Ahern [Thu, 1 Aug 2019 22:18:08 +0000 (15:18 -0700)]
ipv6: have a single rcu unlock point in __ip6_rt_update_pmtu

Simplify the unlock path in __ip6_rt_update_pmtu by using a
single point where rcu_read_unlock is called.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoipv6: Fix unbalanced rcu locking in rt6_update_exception_stamp_rt
David Ahern [Thu, 1 Aug 2019 21:36:35 +0000 (14:36 -0700)]
ipv6: Fix unbalanced rcu locking in rt6_update_exception_stamp_rt

The nexthop path in rt6_update_exception_stamp_rt needs to call
rcu_read_unlock if it fails to find a fib6_nh match rather than
just returning.

Fixes: e659ba31d806 ("ipv6: Handle all fib6_nh in a nexthop in exception handling")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests/tls: add a litmus test for the socket reuse through shutdown
Jakub Kicinski [Thu, 1 Aug 2019 21:36:02 +0000 (14:36 -0700)]
selftests/tls: add a litmus test for the socket reuse through shutdown

Make sure that shutdown never works, and at the same time document how
I tested to came to the conclusion that currently reuse is not possible.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet/tls: partially revert fix transition through disconnect with close
Jakub Kicinski [Thu, 1 Aug 2019 21:36:01 +0000 (14:36 -0700)]
net/tls: partially revert fix transition through disconnect with close

Looks like we were slightly overzealous with the shutdown()
cleanup. Even though the sock->sk_state can reach CLOSED again,
socket->state will not got back to SS_UNCONNECTED once
connections is ESTABLISHED. Meaning we will see EISCONN if
we try to reconnect, and EINVAL if we try to listen.

Only listen sockets can be shutdown() and reused, but since
ESTABLISHED sockets can never be re-connected() or used for
listen() we don't need to try to clean up the ULP state early.

Fixes: 32857cf57f92 ("net/tls: fix transition through disconnect with close")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoMerge tag 'spi-fix-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Mon, 5 Aug 2019 18:49:02 +0000 (11:49 -0700)]
Merge tag 'spi-fix-v5.3-rc3' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A bunch of small, device specific things here plus a DT bindings fix
  for the new validatable YAML binding format.

  The most notable thing is the fix for GPIO chip selects which fixes a
  corner case in updates of that code to modern APIs, unfortunately due
  to a historical mess the code around GPIO support is obscure, fragile
  and an ABI which makes and attempt to improve the situation painful"

* tag 'spi-fix-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: pxa2xx: Add support for Intel Tiger Lake
  spi: bcm2835: Fix 3-wire mode if DMA is enabled
  spi: pxa2xx: Balance runtime PM enable/disable on error
  spi: gpio: Add SPI_MASTER_GPIO_SS flag
  spi: spi-fsl-qspi: change i.MX7D RX FIFO size
  spi: dt-bindings: spi-controller: remove unnecessary 'maxItems: 1' from reg

4 years agoMerge tag 'regulator-fix-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 5 Aug 2019 18:47:02 +0000 (11:47 -0700)]
Merge tag 'regulator-fix-v5.3-rc3' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "A few small driver specific fixes here plus one core fix for a
  refcounting problem with DT which will have little practical impact
  unless overlays are used"

* tag 'regulator-fix-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: of: Add of_node_put() before return in function
  regulator: lp87565: Fix probe failure for "ti,lp87565"
  regulator: axp20x: fix DCDC5 and DCDC6 for AXP803
  regulator: axp20x: fix DCDCA and DCDCD for AXP806

4 years agoMerge tag 'linux-kselftest-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Mon, 5 Aug 2019 18:43:16 +0000 (11:43 -0700)]
Merge tag 'linux-kselftest-5.3-rc4' of git://git./linux/kernel/git/shuah/linux-kselftest

Pull kselftest fixes from Shuah Khan:
 "A fix to the Kselftest framework to save and restore errno and a fix
  to livepatch to push and pop dynamic debug config"

* tag 'linux-kselftest-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
  selftests/livepatch: push and pop dynamic debug config
  kselftest: save-and-restore errno to allow for %m formatting

4 years agoi40e: Remove unicast log when VF is leaving multicast mode.
Czeslaw Zagorski [Tue, 2 Jul 2019 12:22:59 +0000 (08:22 -0400)]
i40e: Remove unicast log when VF is leaving multicast mode.

This patch removes unicast log when VF is leaving multicast mode.
Added check of vf->vf_states &
I40E_VF_STATE_MC_PROMISC/I40E_VF_STATE_UC_PROMISC.
Without this commit, leaving multicast mode logs "unset unicast"
in dmsg.

Signed-off-by: Czeslaw Zagorski <czeslawx.zagorski@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: verify string count matches even on early return
Jacob Keller [Tue, 2 Jul 2019 12:22:58 +0000 (08:22 -0400)]
i40e: verify string count matches even on early return

Similar to i40e_get_ethtool_stats, add a goto to verify that the data
pointer for the strings lines up with the expected stats count. This
helps ensure that bugs are not introduced when adding stats.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: Log info when PF is entering and leaving Allmulti mode.
Czeslaw Zagorski [Tue, 2 Jul 2019 12:22:57 +0000 (08:22 -0400)]
i40e: Log info when PF is entering and leaving Allmulti mode.

Add log when PF is entering and leaving allmulti mode. The
change of PF state is visible in dmesg now. Without this commit,
entering and leaving allmulti mode is not logged in dmesg.

Signed-off-by: Czeslaw Zagorski <czeslawx.zagorski@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: Update visual effect for advertised FEC mode.
Czeslaw Zagorski [Tue, 2 Jul 2019 12:22:56 +0000 (08:22 -0400)]
i40e: Update visual effect for advertised FEC mode.

Updates visual effect for advertised mode after setting desired mode.
The mode appears in advertised FEC mode correctly, when ethtool
interface command is called. Without this commit advertised FEC
is displayed regardless of the settings as "None BaseR RS".

Signed-off-by: Czeslaw Zagorski <czeslawx.zagorski@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: fix code comments
Jeff Kirsher [Tue, 2 Jul 2019 12:22:55 +0000 (08:22 -0400)]
i40e: fix code comments

Found a code comment that needed TLC to correct their formatting.

Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
4 years agoi40e: don't report link up for a VF who hasn't enabled queues
Jacob Keller [Tue, 2 Jul 2019 12:22:54 +0000 (08:22 -0400)]
i40e: don't report link up for a VF who hasn't enabled queues

Commit d3d657a90850 ("i40e: update VFs of link state after
GET_VF_RESOURCES") modified the PF driver to notify a VF of
its link status immediately after it requests resources.

This was intended to fix reporting on VF drivers, so that they would
properly report link status.

However, some older VF drivers do not respond well to receiving a link
up notification before queues are enabled. This can cause their state
machine to think that it is safe to send traffic. This results in a Tx
hang on the VF.

More recent versions of the old i40evf and all versions of iavf are
resilient to these early link status messages. However, if a VM happens
to run an older version of the VF driver, this can be problematic.

Record whether the PF has actually enabled queues for the VF. When
reporting link status, always report link down if the queues aren't
enabled. In this way, the VF driver will never receive a link up
notification until after its queues are enabled.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: Log disable-fw-lldp flag change by ethtool
Slawomir Laba [Tue, 2 Jul 2019 12:22:53 +0000 (08:22 -0400)]
i40e: Log disable-fw-lldp flag change by ethtool

Add logging for disable-fw-lldp flag by ethtool. Added check
for I40E_FLAG_DISABLE_FW_LLDP and logging state in dmesg.
Without this commit there was no clear statement in dmesg
about FW LLDP state in dmesg.

Signed-off-by: Slawomir Laba <slawomirx.laba@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoi40e: fix incorrect ethtool statistics veb and veb.tc_
Dmitrii Golovanov [Tue, 2 Jul 2019 12:22:51 +0000 (08:22 -0400)]
i40e: fix incorrect ethtool statistics veb and veb.tc_

This patch fixes missing call of i40e_update_veb_stats() in function
i40e_get_ethtool_stats() to update stats data of VEB and VEB TC
counters before they are written into ethtool buffer.
Before the patch ethtool counters may fell behind interface counters.

Signed-off-by: Dmitrii Golovanov <dmitrii.golovanov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
4 years agoMerge branch 'net-fix-regressions-for-generic-XDP'
David S. Miller [Mon, 5 Aug 2019 18:17:40 +0000 (11:17 -0700)]
Merge branch 'net-fix-regressions-for-generic-XDP'

Jesper Dangaard Brouer says:

====================
net: fix regressions for generic-XDP

Thanks to Brandon Cazander, who wrote a very detailed bug report that
even used perf probe's on xdp-newbies mailing list, we discovered that
generic-XDP contains some regressions when using bpf_xdp_adjust_head().

First issue were that my selftests script, that use bpf_xdp_adjust_head(),
by mistake didn't use generic-XDP any-longer. That selftest should have
caught the real regression introduced in commit 458bf2f224f0 ("net: core:
support XDP generic on stacked devices.").

To verify this patchset fix the regressions, you can invoked manually via:

  cd tools/testing/selftests/bpf/
  sudo ./test_xdp_vlan_mode_generic.sh
  sudo ./test_xdp_vlan_mode_native.sh
====================

Link: https://www.spinics.net/lists/xdp-newbies/msg01231.html
Fixes: 458bf2f224f0 ("net: core: support XDP generic on stacked devices.")
Reported by: Brandon Cazander <brandon.cazander@multapplied.net>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet: fix bpf_xdp_adjust_head regression for generic-XDP
Jesper Dangaard Brouer [Thu, 1 Aug 2019 18:00:31 +0000 (20:00 +0200)]
net: fix bpf_xdp_adjust_head regression for generic-XDP

When generic-XDP was moved to a later processing step by commit
458bf2f224f0 ("net: core: support XDP generic on stacked devices.")
a regression was introduced when using bpf_xdp_adjust_head.

The issue is that after this commit the skb->network_header is now
changed prior to calling generic XDP and not after. Thus, if the header
is changed by XDP (via bpf_xdp_adjust_head), then skb->network_header
also need to be updated again.  Fix by calling skb_reset_network_header().

Fixes: 458bf2f224f0 ("net: core: support XDP generic on stacked devices.")
Reported-by: Brandon Cazander <brandon.cazander@multapplied.net>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests/bpf: reduce time to execute test_xdp_vlan.sh
Jesper Dangaard Brouer [Thu, 1 Aug 2019 18:00:26 +0000 (20:00 +0200)]
selftests/bpf: reduce time to execute test_xdp_vlan.sh

Given the increasing number of BPF selftests, it makes sense to
reduce the time to execute these tests.  The ping parameters are
adjusted to reduce the time from measures 9 sec to approx 2.8 sec.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agoselftests/bpf: add wrapper scripts for test_xdp_vlan.sh
Jesper Dangaard Brouer [Thu, 1 Aug 2019 18:00:21 +0000 (20:00 +0200)]
selftests/bpf: add wrapper scripts for test_xdp_vlan.sh

In-order to test both native-XDP (xdpdrv) and generic-XDP (xdpgeneric)
create two wrapper test scripts, that start the test_xdp_vlan.sh script
with these modes.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agobpf: fix XDP vlan selftests test_xdp_vlan.sh
Jesper Dangaard Brouer [Thu, 1 Aug 2019 18:00:16 +0000 (20:00 +0200)]
bpf: fix XDP vlan selftests test_xdp_vlan.sh

Change BPF selftest test_xdp_vlan.sh to (default) use generic XDP.

This selftest was created together with a fix for generic XDP, in commit
297249569932 ("net: fix generic XDP to handle if eth header was
mangled"). And was suppose to catch if generic XDP was broken again.

The tests are using veth and assumed that veth driver didn't support
native driver XDP, thus it used the (ip link set) 'xdp' attach that fell
back to generic-XDP. But veth gained native-XDP support in 948d4f214fde
("veth: Add driver XDP"), which caused this test script to use
native-XDP.

Fixes: 948d4f214fde ("veth: Add driver XDP")
Fixes: 97396ff0bc2d ("selftests/bpf: add XDP selftests for modifying and popping VLAN headers")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
4 years agonet/mlx5e: always initialize frag->last_in_page
Qian Cai [Thu, 1 Aug 2019 13:52:54 +0000 (09:52 -0400)]
net/mlx5e: always initialize frag->last_in_page

The commit 069d11465a80 ("net/mlx5e: RX, Enhance legacy Receive Queue
memory scheme") introduced an undefined behaviour below due to
"frag->last_in_page" is only initialized in mlx5e_init_frags_partition()
when,

if (next_frag.offset + frag_info[f].frag_stride > PAGE_SIZE)

or after bailed out the loop,

for (i = 0; i < mlx5_wq_cyc_get_size(&rq->wqe.wq); i++)

As the result, there could be some "frag" have uninitialized
value of "last_in_page".

Later, get_frag() obtains those "frag" and check "frag->last_in_page" in
mlx5e_put_rx_frag() and triggers the error during boot. Fix it by always
initializing "frag->last_in_page" to "false" in
mlx5e_init_frags_partition().

UBSAN: Undefined behaviour in
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c:325:12
load of value 170 is not a valid value for type 'bool' (aka '_Bool')
Call trace:
 dump_backtrace+0x0/0x264
 show_stack+0x20/0x2c
 dump_stack+0xb0/0x104
 __ubsan_handle_load_invalid_value+0x104/0x128
 mlx5e_handle_rx_cqe+0x8e8/0x12cc [mlx5_core]
 mlx5e_poll_rx_cq+0xca8/0x1a94 [mlx5_core]
 mlx5e_napi_poll+0x17c/0xa30 [mlx5_core]
 net_rx_action+0x248/0x940
 __do_softirq+0x350/0x7b8
 irq_exit+0x200/0x26c
 __handle_domain_irq+0xc8/0x128
 gic_handle_irq+0x138/0x228
 el1_irq+0xb8/0x140
 arch_cpu_idle+0x1a4/0x348
 do_idle+0x114/0x1b0
 cpu_startup_entry+0x24/0x28
 rest_init+0x1ac/0x1dc
 arch_call_rest_init+0x10/0x18
 start_kernel+0x4d4/0x57c

Fixes: 069d11465a80 ("net/mlx5e: RX, Enhance legacy Receive Queue memory scheme")
Signed-off-by: Qian Cai <cai@lca.pw>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>