openwrt/staging/blogic.git
5 years agosfc: change ARFS expiry mechanism
Edward Cree [Fri, 22 Nov 2019 17:57:03 +0000 (17:57 +0000)]
sfc: change ARFS expiry mechanism

The old rfs_filters_added method for determining the quota could potentially
 allow the NIC to become filled with old filters, which never get tested for
 expiry.  Instead, explicitly make expiry check work depend on the number of
 filters installed, and don't count checking slots without filters in as
 doing work.  This guarantees that each filter will be checked for expiry at
 least once every thirty seconds (assuming the channel to which it belongs is
 NAPI polling actively) regardless of fill level.

Signed-off-by: Edward Cree <ecree@solarflare.com>
Tested-by: David Ahern <dahern@digitalocean.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
Jakub Kicinski [Sat, 23 Nov 2019 00:59:00 +0000 (16:59 -0800)]
Merge branch '100GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
This series contains updates to the ice driver only.

Bruce updates the driver to store the number of functions the device has
so that it won't have to compute it when setting safe mode capabilities.
Adds a check to adjust the reporting of capabilities for devices with
more than 4 ports, which differ for devices with less than 4 ports.

Brett adds a helper function to determine if the VF is allowed to do
VLAN operations based on the host's VF configuration.  Also adds a new
function that initializes VLAN stripping (enabled/disabled) for the VF
based on the device supported capabilities.  Adds a check if the vector
index is valid with the respect to the number of transmit and receive
queues configured when we set coalesce settings for DCB.  Adds a check
if the promisc_mask contains ICE_PROMISC_VLAN_RX or ICE_PROMISC_VLAN_TX
so that VLAN 0 promiscuous rules to be removed.  Add a helper macro for
a commonly used de-reference of a pointer to &pf->dev->pdev.

Jesse fixes an issue where if an invalid virtchnl request from the VF,
the driver would return uninitialized data to the VF from the PF stack,
so ensure the stack variable is initialized earlier.  Add helpers to the
virtchnl interface make the reporting of strings consistent and help
reduce stack space.  Implements VF statistics gathering via the kernel
ndo_get_vf_stats().

Akeem ensures we disable the state flag for each VF when its resources
are returned to the device.

Tony does additional cleanup in the driver to ensure the when we
allocate and free memory within the same function, we should not be
using devm_* variants; use regular alloc and free functions.

Henry implements code to query and set the number of channels on the
primary VSI for a PF via ethtool.

Jake cleans up needless NULL checks in ice_sched_cleanup_all().

Kevin updates the firmware API version to align with current NVM images.

v2: Added "Fixes:" tag to patch 5 commit description and added the use
    of netif_is_rxfh_configured() in patch 13 to see if RSS has been
    configured by the user, if so do not overwrite that configuration.
====================

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agonet: inet_is_local_reserved_port() should return bool not int
Maciej Żenczykowski [Fri, 22 Nov 2019 21:50:52 +0000 (13:50 -0800)]
net: inet_is_local_reserved_port() should return bool not int

Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agoMerge branch 'cxgb4-add-udp-segmentation-offload-support'
Jakub Kicinski [Sat, 23 Nov 2019 00:46:33 +0000 (16:46 -0800)]
Merge branch 'cxgb4-add-udp-segmentation-offload-support'

Rahul Lakkireddy says:

====================
This series of patches add UDP Segmentation Offload (USO) supported
by Chelsio T5/T6 NICs.

Patch 1 updates the current Scatter Gather List (SGL) DMA unmap logic
for USO requests.

Patch 2 adds USO support for NIC and MQPRIO QoS offload Tx path.

Patch 3 adds missing stats for MQPRIO QoS offload Tx path.
====================

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agocxgb4: add stats for MQPRIO QoS offload Tx path
Rahul Lakkireddy [Fri, 22 Nov 2019 01:00:03 +0000 (06:30 +0530)]
cxgb4: add stats for MQPRIO QoS offload Tx path

Export necessary stats for traffic flowing through MQPRIO QoS offload
Tx path.

v2:
- No change.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agocxgb4: add UDP segmentation offload support
Rahul Lakkireddy [Fri, 22 Nov 2019 01:00:02 +0000 (06:30 +0530)]
cxgb4: add UDP segmentation offload support

Implement and export UDP segmentation offload (USO) support for both
NIC and MQPRIO QoS offload Tx path. Update appropriate logic in Tx to
parse GSO info in skb and configure FW_ETH_TX_EO_WR request needed to
perform USO.

v2:
- Remove inline keyword from write_eo_udp_wr() in sge.c. Let the
  compiler decide.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agocxgb4/chcr: update SGL DMA unmap for USO
Rahul Lakkireddy [Fri, 22 Nov 2019 01:00:01 +0000 (06:30 +0530)]
cxgb4/chcr: update SGL DMA unmap for USO

The FW_ETH_TX_EO_WR used for sending UDP Segmentation Offload (USO)
requests expects the headers to be part of the descriptor and the
payload to be part of the SGL containing the DMA mapped addresses.
Hence, the DMA address in the first entry of the SGL can start after
the packet headers. Currently, unmap_sgl() tries to unmap from this
wrong offset, instead of the originally mapped DMA address.

So, use existing unmap_skb() instead, which takes originally saved DMA
addresses as input. Update all necessary Tx paths to save the original
DMA addresses, so that unmap_skb() can unmap them properly.

v2:
- No change.

Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Sat, 23 Nov 2019 00:27:24 +0000 (16:27 -0800)]
Merge git://git./linux/kernel/git/netdev/net

Minor conflict in drivers/s390/net/qeth_l2_main.c, kept the lock
from commit c8183f548902 ("s390/qeth: fix potential deadlock on
workqueue flush"), removed the code which was removed by commit
9897d583b015 ("s390/qeth: consolidate some duplicated HW cmd code").

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Fri, 22 Nov 2019 22:28:14 +0000 (14:28 -0800)]
Merge git://git./linux/kernel/git/netdev/net

Pull networking fixes from David Miller:

 1) Validate tunnel options length in act_tunnel_key, from Xin Long.

 2) Fix DMA sync bug in gve driver, from Adi Suresh.

 3) TSO kills performance on some r8169 chips due to HW issues, disable
    by default in that case, from Corinna Vinschen.

 4) Fix clock disable mismatch in fec driver, from Chubong Yuan.

 5) Fix interrupt status bits define in hns3 driver, from Huazhong Tan.

 6) Fix workqueue deadlocks in qeth driver, from Julian Wiedmann.

 7) Don't napi_disable() twice in r8152 driver, from Hayes Wang.

 8) Fix SKB extension memory leak, from Florian Westphal.

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (54 commits)
  r8152: avoid to call napi_disable twice
  MAINTAINERS: Add myself as maintainer of virtio-vsock
  udp: drop skb extensions before marking skb stateless
  net: rtnetlink: prevent underflows in do_setvfinfo()
  can: m_can_platform: remove unnecessary m_can_class_resume() call
  can: m_can_platform: set net_device structure as driver data
  hv_netvsc: Fix send_table offset in case of a host bug
  hv_netvsc: Fix offset usage in netvsc_send_table()
  net-ipv6: IPV6_TRANSPARENT - check NET_RAW prior to NET_ADMIN
  sfc: Only cancel the PPS workqueue if it exists
  nfc: port100: handle command failure cleanly
  net-sysfs: fix netdev_queue_add_kobject() breakage
  r8152: Re-order napi_disable in rtl8152_close
  net: qca_spi: Move reset_count to struct qcaspi
  net: qca_spi: fix receive buffer size check
  net/ibmvnic: Ignore H_FUNCTION return from H_EOI to tolerate XIVE mode
  Revert "net/ibmvnic: Fix EOI when running in XIVE mode"
  net/mlxfw: Verify FSM error code translation doesn't exceed array size
  net/mlx5: Update the list of the PCI supported devices
  net/mlx5: Fix auto group size calculation
  ...

5 years agoafs: Fix large file support
Marc Dionne [Thu, 21 Nov 2019 15:37:26 +0000 (15:37 +0000)]
afs: Fix large file support

By default s_maxbytes is set to MAX_NON_LFS, which limits the usable
file size to 2GB, enforced by the vfs.

Commit b9b1f8d5930a ("AFS: write support fixes") added support for the
64-bit fetch and store server operations, but did not change this value.
As a result, attempts to write past the 2G mark result in EFBIG errors:

 $ dd if=/dev/zero of=foo bs=1M count=1 seek=2048
 dd: error writing 'foo': File too large

Set s_maxbytes to MAX_LFS_FILESIZE.

Fixes: b9b1f8d5930a ("AFS: write support fixes")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoafs: Fix possible assert with callbacks from yfs servers
Marc Dionne [Thu, 21 Nov 2019 15:26:15 +0000 (15:26 +0000)]
afs: Fix possible assert with callbacks from yfs servers

Servers sending callback breaks to the YFS_CM_SERVICE service may
send up to YFSCBMAX (1024) fids in a single RPC.  Anything over
AFSCBMAX (50) will cause the assert in afs_break_callbacks to trigger.

Remove the assert, as the count has already been checked against
the appropriate max values in afs_deliver_cb_callback and
afs_deliver_yfs_cb_callback.

Fixes: 35dbfba3111a ("afs: Implement the YFS cache manager service")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoice: Update FW API minor version
Kevin Scott [Fri, 8 Nov 2019 14:23:31 +0000 (06:23 -0800)]
ice: Update FW API minor version

Update FW API minor version to align to current value advertised
by FW in new NVM images.

Signed-off-by: Kevin Scott <kevin.c.scott@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: remove pointless NULL check of port_info
Jacob Keller [Fri, 8 Nov 2019 14:23:30 +0000 (06:23 -0800)]
ice: remove pointless NULL check of port_info

The code in ice_sched_cleanup_all checks whether the port info is NULL
prior to calling ice_sched_clear_port. However, ice_sched_clear_port
already checks whether port info is non-NULL.

More importantly, it also checks whether the port structure has been
initialized by checking its port_state field as well.

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>
5 years agoice: Implement ethtool ops for channels
Henry Tieman [Fri, 8 Nov 2019 14:23:29 +0000 (06:23 -0800)]
ice: Implement ethtool ops for channels

Add code to query and set the number of channels on the primary VSI for a
PF. This is accessed from the 'ethtool -l' and 'ethtool -L' commands,
respectively.  Though the ice driver supports asymmetric queues report an
IRQ vector that has both Rx and Tx queues attached and is counted as a
'combined' channel.

Signed-off-by: Henry Tieman <henry.w.tieman@intel.com>
Co-developed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: implement VF stats NDO
Jesse Brandeburg [Fri, 8 Nov 2019 14:23:28 +0000 (06:23 -0800)]
ice: implement VF stats NDO

Implement the VF stats gathering via the kernel via ndo_get_vf_stats().
The driver will show per-VF stats in the output of the
ip -s link show dev <PF> command.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: add helpers for virtchnl
Jesse Brandeburg [Fri, 8 Nov 2019 14:23:27 +0000 (06:23 -0800)]
ice: add helpers for virtchnl

The virtchannel interface was repeating a lot of strings
and wasting storage space in the kernel.  There was also
inconsistent messages for the same thing.  Consolidate all
those messages and bit checks into a couple of helper functions.

Also, reduce stack space usage by simplifying getting the pointer
to the pf using a helper.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Co-developed-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Add ice_pf_to_dev(pf) macro
Brett Creeley [Fri, 8 Nov 2019 14:23:26 +0000 (06:23 -0800)]
ice: Add ice_pf_to_dev(pf) macro

We use &pf->dev->pdev all over the code. Add a simple
macro to do this for us. When multiple de-references
like this are being done add a local struct device
variable.

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Do not use devm* functions for local uses
Tony Nguyen [Fri, 8 Nov 2019 14:23:25 +0000 (06:23 -0800)]
ice: Do not use devm* functions for local uses

In situations where we alloc and free memory within the same function do
not use the devm_* variants; use regular alloc and free functions. Remove
any unused vars if there are no usages after these changes.

Also, replace an allocate and copy with kmemdup() and remove an
unnecessary memset() to 0 after a kzalloc().

Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Refactor removal of VLAN promiscuous rules
Brett Creeley [Fri, 8 Nov 2019 14:23:24 +0000 (06:23 -0800)]
ice: Refactor removal of VLAN promiscuous rules

Currently ice_clear_vsi_promisc() detects if the VLAN ID sent is not 0
and sets the recipe_id to ICE_SW_LKUP_PROMISC_VLAN in that case and
ICE_SW_LKUP_PROMISC if the VLAN_ID is 0. However this doesn't allow VLAN
0 promiscuous rules to be removed, but they can be added. Fix this by
checking if the promisc_mask contains ICE_PROMISC_VLAN_RX or
ICE_PROMISC_VLAN_TX. This change was made to match what is being done
for ice_set_vsi_promisc().

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Fix setting coalesce to handle DCB configuration
Brett Creeley [Fri, 8 Nov 2019 14:23:23 +0000 (06:23 -0800)]
ice: Fix setting coalesce to handle DCB configuration

Currently there can be a case where a DCB map is applied and there are
more interrupt vectors (vsi->num_q_vectors) than Rx queues (vsi->num_rxq)
and Tx queues (vsi->num_txq). If we try to set coalesce settings in this
case it will report a false failure. Fix this by checking if vector index
is valid with respect to the number of Tx and Rx queues configured.

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Only disable VF state when freeing each VF resources
Akeem G Abodunrin [Fri, 8 Nov 2019 14:23:22 +0000 (06:23 -0800)]
ice: Only disable VF state when freeing each VF resources

It is wrong to set PF disable state flag for all VFs when freeing VF
resources - Instead, we should set VF disable state flag for each VF with
its resources being returned to the device. Right now, all VF opcodes,
mailbox communication to clear its resources as well fails - since we
already indicate that PF is in disable state, with all VFs not active. In
addition, we don't need to notify VF that PF is intending to reset it, if
it is already in disabled state.

Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: fix stack leakage
Jesse Brandeburg [Fri, 8 Nov 2019 14:23:21 +0000 (06:23 -0800)]
ice: fix stack leakage

In the case of an invalid virtchannel request the driver
would return uninitialized data to the VF from the PF stack
which is a bug.  Fix by initializing the stack variable
earlier in the function before any return paths can be taken.

Fixes: 1071a8358a28 ("ice: Implement virtchnl commands for AVF support")
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Don't modify stripping for add/del VLANs on VF
Brett Creeley [Fri, 8 Nov 2019 14:23:20 +0000 (06:23 -0800)]
ice: Don't modify stripping for add/del VLANs on VF

Currently when adding/deleting vlans in ice_vc_process_vlan_msg()
we are calling ice_vsi_manage_vlan_stripping() to enable/disable
when adding and deleting a VLAN respectively. This is wrong
because adding/deleting VLANs has nothing to do with configuring
VLAN stripping. VLAN stripping is configured through the
following VIRTCHNL operations:
VIRTCHNL_OP_ENABLE_VLAN_STRIPPING
VIRTCHNL_OP_DISABLE_VLAN_STRIPPING

Unfortunately we can't just remove this because then stripping
will never be configured on VF initialization. Fix this by
adding a new function that initializes (disables/enables) VLAN
stripping for the VF based on the device supported capabilities.
This allows us to remove the call to
ice_vsi_manage_vlan_stripping() in ice_vc_process_vlan_msg().

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Disallow VF VLAN opcodes if VLAN offloads disabled
Brett Creeley [Fri, 8 Nov 2019 14:23:19 +0000 (06:23 -0800)]
ice: Disallow VF VLAN opcodes if VLAN offloads disabled

Currently if the host disables VLAN offloads on the VF by
not setting the VIRTCHNL_VF_OFFLOAD_VLAN capability bit
we will still honor VF VLAN configuration messages over
VIRTCHNL. These messages (i.e. enable/disable VLAN stripping
and VLAN filtering) should be blocked when the feature
is not supported. Fix that by adding a helper function to
determine if the VF is allowed to do VLAN operations based
on the host's VF configuration.

Also, mirror the VF communicated capabilities in the host's
VF configuration.

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Correct capabilities reporting of max TCs
Bruce Allan [Fri, 8 Nov 2019 14:23:18 +0000 (06:23 -0800)]
ice: Correct capabilities reporting of max TCs

Firmware always returns 8 as the max number of supported TCs. However on
devices with more than 4 ports, the maximum number of TCs per port is
limited to 4. Check and, if necessary, correct the reporting of
capabilities for devices with more than 4 ports.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agoice: Store number of functions for the device
Bruce Allan [Fri, 8 Nov 2019 14:23:17 +0000 (06:23 -0800)]
ice: Store number of functions for the device

Store the number of functions the device has and use this number when
setting safe mode capabilities instead of calculating it.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Co-developed-by: Kevin Scott <kevin.c.scott@intel.com>
Signed-off-by: Kevin Scott <kevin.c.scott@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
5 years agonet: dsa: ocelot: fix "should it be static?" warnings
Chen Wandun [Fri, 22 Nov 2019 12:32:45 +0000 (20:32 +0800)]
net: dsa: ocelot: fix "should it be static?" warnings

Fix following sparse warnings:
drivers/net/dsa/ocelot/felix.c:351:6: warning: symbol 'felix_txtstamp' was not declared. Should it be static?

Signed-off-by: Chen Wandun <chenwandun@huawei.com>
Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agor8152: avoid to call napi_disable twice
Hayes Wang [Fri, 22 Nov 2019 08:21:09 +0000 (16:21 +0800)]
r8152: avoid to call napi_disable twice

Call napi_disable() twice would cause dead lock. There are three situations
may result in the issue.

1. rtl8152_pre_reset() and set_carrier() are run at the same time.
2. Call rtl8152_set_tunable() after rtl8152_close().
3. Call rtl8152_set_ringparam() after rtl8152_close().

For #1, use the same solution as commit 84811412464d ("r8152: Re-order
napi_disable in rtl8152_close"). For #2 and #3, add checking the flag
of IFF_UP and using napi_disable/napi_enable during mutex.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Fri, 22 Nov 2019 17:49:08 +0000 (09:49 -0800)]
Merge branch 'akpm' (patches from Andrew)

Merge misc fixes from Andrew Morton:
 "Three fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  mm/ksm.c: don't WARN if page is still mapped in remove_stable_node()
  mm/memory_hotplug: don't access uninitialized memmaps in shrink_zone_span()
  Revert "fs: ocfs2: fix possible null-pointer dereferences in ocfs2_xa_prepare_entry()"

5 years agoseg6: allow local packet processing for SRv6 End.DT6 behavior
Andrea Mayer [Fri, 22 Nov 2019 16:22:42 +0000 (17:22 +0100)]
seg6: allow local packet processing for SRv6 End.DT6 behavior

End.DT6 behavior makes use of seg6_lookup_nexthop() function which drops
all packets that are destined to be locally processed. However, DT* should
be able to deliver decapsulated packets that are destined to local
addresses. Function seg6_lookup_nexthop() is also used by DX6, so in order
to maintain compatibility I created another routing helper function which
is called seg6_lookup_any_nexthop(). This function is able to take into
account both packets that have to be processed locally and the ones that
are destined to be forwarded directly to another machine. Hence,
seg6_lookup_any_nexthop() is used in DT6 rather than seg6_lookup_nexthop()
to allow local delivery.

Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: flow_dissector: Wrap unionized VLAN fields in a struct
Petr Machata [Fri, 22 Nov 2019 15:47:21 +0000 (15:47 +0000)]
net: flow_dissector: Wrap unionized VLAN fields in a struct

In commit a82055af5959 ("netfilter: nft_payload: add VLAN offload
support"), VLAN fields in struct flow_dissector_key_vlan were unionized
with the intention of introducing another field that covered the whole TCI
header. However without a wrapping struct the subfields end up sharing the
same bits. As a result, "tc filter add ... flower vlan_id 14" specifies not
only vlan_id, but also vlan_priority.

Fix by wrapping the individual VLAN fields in a struct.

Fixes: a82055af5959 ("netfilter: nft_payload: add VLAN offload support")
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'linux-can-fixes-for-5.4-20191122' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Fri, 22 Nov 2019 17:42:11 +0000 (09:42 -0800)]
Merge tag 'linux-can-fixes-for-5.4-20191122' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2019-11-22

this is a pull request of 2 patches for net/master, if possible for the
current release cycle. Otherwise these patches should hit v5.4 via the
stable tree.

Both patches of this pull request target the m_can driver. Pankaj Sharma
fixes the fallout in the m_can_platform part, which appeared with the
introduction of the m_can platform framework.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'mac80211-next-for-net-next-2019-11-22' of git://git.kernel.org/pub/scm...
David S. Miller [Fri, 22 Nov 2019 17:40:52 +0000 (09:40 -0800)]
Merge tag 'mac80211-next-for-net-next-2019-11-22' of git://git./linux/kernel/git/jberg/mac80211-next

Johannes Berg says:

====================
The interesting new thing here is AQL, the Airtime Queue Limit
patchset from Kan Yan (Google) and Toke Høiland-Jørgensen (Redhat).
The effect is intended to eventually be similar to BQL, but byte
queue limits are not useful in wifi where the actual throughput can
vary by around 4 orders of magnitude. There are more details in the
patches themselves.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMAINTAINERS: Add myself as maintainer of virtio-vsock
Stefano Garzarella [Fri, 22 Nov 2019 10:20:10 +0000 (11:20 +0100)]
MAINTAINERS: Add myself as maintainer of virtio-vsock

Since I'm actively working on vsock and virtio/vhost transports,
Stefan suggested to help him to maintain it.

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotipc: support in-order name publication events
Tuong Lien [Thu, 21 Nov 2019 08:34:58 +0000 (15:34 +0700)]
tipc: support in-order name publication events

It is observed that TIPC service binding order will not be kept in the
publication event report to user if the service is subscribed after the
bindings.

For example, services are bound by application in the following order:

Server: bound port A to {18888,66,66} scope 2
Server: bound port A to {18888,33,33} scope 2

Now, if a client subscribes to the service range (e.g. {18888, 0-100}),
it will get the 'TIPC_PUBLISHED' events in that binding order only when
the subscription is started before the bindings.
Otherwise, if started after the bindings, the events will arrive in the
opposite order:

Client: received event for published {18888,33,33}
Client: received event for published {18888,66,66}

For the latter case, it is clear that the bindings have existed in the
name table already, so when reported, the events' order will follow the
order of the rbtree binding nodes (- a node with lesser 'lower'/'upper'
range value will be first).

This is correct as we provide the tracking on a specific service status
(available or not), not the relationship between multiple services.
However, some users expect to see the same order of arriving events
irrespective of when the subscription is issued. This turns out to be
easy to fix. We now add functionality to ensure that publication events
always are issued in the same temporal order as the corresponding
bindings were performed.

v2: replace the unnecessary macro - 'publication_after()' with inline
function.
v3: reuse 'time_after32()' instead of reinventing the same exact code.

Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotipc: update replicast capability for broadcast send link
Hoang Le [Thu, 21 Nov 2019 03:01:09 +0000 (10:01 +0700)]
tipc: update replicast capability for broadcast send link

When setting up a cluster with non-replicast/replicast capability
supported. This capability will be disabled for broadcast send link
in order to be backwards compatible.

However, when these non-support nodes left and be removed out the cluster.
We don't update this capability on broadcast send link. Then, some of
features that based on this capability will also disabling as unexpected.

In this commit, we make sure the broadcast send link capabilities will
be re-calculated as soon as a node removed/rejoined a cluster.

Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoudp: drop skb extensions before marking skb stateless
Florian Westphal [Thu, 21 Nov 2019 05:56:23 +0000 (06:56 +0100)]
udp: drop skb extensions before marking skb stateless

Once udp stack has set the UDP_SKB_IS_STATELESS flag, later skb free
assumes all skb head state has been dropped already.

This will leak the extension memory in case the skb has extensions other
than the ipsec secpath, e.g. bridge nf data.

To fix this, set the UDP_SKB_IS_STATELESS flag only if we don't have
extensions or if the extension space can be free'd.

Fixes: 895b5c9f206eb7d25dc1360a ("netfilter: drop bridge nf reset from nf_reset")
Cc: Paolo Abeni <pabeni@redhat.com>
Reported-by: Byron Stanoszek <gandalf@winds.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: rtnetlink: prevent underflows in do_setvfinfo()
Dan Carpenter [Wed, 20 Nov 2019 12:34:38 +0000 (15:34 +0300)]
net: rtnetlink: prevent underflows in do_setvfinfo()

The "ivm->vf" variable is a u32, but the problem is that a number of
drivers cast it to an int and then forget to check for negatives.  An
example of this is in the cxgb4 driver.

drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
  2890  static int cxgb4_mgmt_get_vf_config(struct net_device *dev,
  2891                                      int vf, struct ifla_vf_info *ivi)
                                            ^^^^^^
  2892  {
  2893          struct port_info *pi = netdev_priv(dev);
  2894          struct adapter *adap = pi->adapter;
  2895          struct vf_info *vfinfo;
  2896
  2897          if (vf >= adap->num_vfs)
                    ^^^^^^^^^^^^^^^^^^^
  2898                  return -EINVAL;
  2899          vfinfo = &adap->vfinfo[vf];
                ^^^^^^^^^^^^^^^^^^^^^^^^^^

There are 48 functions affected.

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c:8435 hclge_set_vf_vlan_filter() warn: can 'vfid' underflow 's32min-2147483646'
drivers/net/ethernet/freescale/enetc/enetc_pf.c:377 enetc_pf_set_vf_mac() warn: can 'vf' underflow 's32min-2147483646'
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:2899 cxgb4_mgmt_get_vf_config() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:2960 cxgb4_mgmt_set_vf_rate() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:3019 cxgb4_mgmt_set_vf_rate() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:3038 cxgb4_mgmt_set_vf_vlan() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c:3086 cxgb4_mgmt_set_vf_link_state() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/chelsio/cxgb/cxgb2.c:791 get_eeprom() warn: can 'i' underflow 's32min-(-4),0,4-s32max'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:82 bnxt_set_vf_spoofchk() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:164 bnxt_set_vf_trust() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:186 bnxt_get_vf_config() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:228 bnxt_set_vf_mac() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:264 bnxt_set_vf_vlan() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:293 bnxt_set_vf_bw() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c:333 bnxt_set_vf_link_state() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c:2595 bnx2x_vf_op_prep() warn: can 'vfidx' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c:2595 bnx2x_vf_op_prep() warn: can 'vfidx' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c:2281 bnx2x_post_vf_bulletin() warn: can 'vf' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c:2285 bnx2x_post_vf_bulletin() warn: can 'vf' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c:2286 bnx2x_post_vf_bulletin() warn: can 'vf' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c:2292 bnx2x_post_vf_bulletin() warn: can 'vf' underflow 's32min-63'
drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c:2297 bnx2x_post_vf_bulletin() warn: can 'vf' underflow 's32min-63'
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c:1832 qlcnic_sriov_set_vf_mac() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c:1864 qlcnic_sriov_set_vf_tx_rate() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c:1937 qlcnic_sriov_set_vf_vlan() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c:2005 qlcnic_sriov_get_vf_config() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c:2036 qlcnic_sriov_set_vf_spoofchk() warn: can 'vf' underflow 's32min-254'
drivers/net/ethernet/emulex/benet/be_main.c:1914 be_get_vf_config() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/emulex/benet/be_main.c:1915 be_get_vf_config() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/emulex/benet/be_main.c:1922 be_set_vf_tvt() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/emulex/benet/be_main.c:1951 be_clear_vf_tvt() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/emulex/benet/be_main.c:2063 be_set_vf_tx_rate() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/emulex/benet/be_main.c:2091 be_set_vf_link_state() warn: can 'vf' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:2609 ice_set_vf_port_vlan() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:3050 ice_get_vf_cfg() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:3103 ice_set_vf_spoofchk() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:3181 ice_set_vf_mac() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:3237 ice_set_vf_trust() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c:3286 ice_set_vf_link_state() warn: can 'vf_id' underflow 's32min-65534'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:3919 i40e_validate_vf() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:3957 i40e_ndo_set_vf_mac() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4104 i40e_ndo_set_vf_port_vlan() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4263 i40e_ndo_set_vf_bw() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4309 i40e_ndo_get_vf_config() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4371 i40e_ndo_set_vf_link_state() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4441 i40e_ndo_set_vf_spoofchk() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4441 i40e_ndo_set_vf_spoofchk() warn: can 'vf_id' underflow 's32min-2147483646'
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c:4504 i40e_ndo_set_vf_trust() warn: can 'vf_id' underflow 's32min-2147483646'

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'pm-5.4-final' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 22 Nov 2019 17:18:16 +0000 (09:18 -0800)]
Merge tag 'pm-5.4-final' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management regression fix from Rafael Wysocki:
 "Fix problems with switching cpufreq drivers on some x86 systems with
  ACPI (and with changing the operation modes of the intel_pstate driver
  on those systems) introduced by recent changes related to the
  management of frequency limits in cpufreq"

* tag 'pm-5.4-final' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM: QoS: Invalidate frequency QoS requests after removal

5 years agoMerge tag 'drm-fixes-2019-11-22' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 22 Nov 2019 17:14:30 +0000 (09:14 -0800)]
Merge tag 'drm-fixes-2019-11-22' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Two sets of fixes in here, one for amdgpu, and one for i915.

  The amdgpu ones are pretty small, i915's CI system seems to have a few
  problems in the last week or so, there is one major regression fix for
  fb_mmap, but there are a bunch of other issues fixed in there as well,
  oops, screen flashes and rcu related.

  amdgpu:
   - Remove experimental flag for navi14
   - Fix confusing power message failures on older VI parts
   - Hang fix for gfxoff when using the read register interface
   - Two stability regression fixes for Raven

  i915:
   - Fix kernel oops on dumb_create ioctl on no crtc situation
   - Fix bad ugly colored flash on VLV/CHV related to gamma LUT update
   - Fix unity of the frequencies reported on PMU
   - Fix kernel oops on set_page_dirty using better locks around it
   - Protect the request pointer with RCU to prevent it being freed
     while we might need still
   - Make pool objects read-only
   - Restore physical addresses for fb_map to avoid corrupted page
     table"

* tag 'drm-fixes-2019-11-22' of git://anongit.freedesktop.org/drm/drm:
  drm/i915/fbdev: Restore physical addresses for fb_mmap()
  Revert "drm/amd/display: enable S/G for RAVEN chip"
  drm/amdgpu: disable gfxoff on original raven
  drm/amdgpu: disable gfxoff when using register read interface
  drm/amd/powerplay: correct fine grained dpm force level setting
  drm/amd/powerplay: issue no PPSMC_MSG_GetCurrPkgPwr on unsupported ASICs
  drm/amdgpu: remove experimental flag for Navi14
  drm/i915: make pool objects read-only
  drm/i915: Protect request peeking with RCU
  drm/i915/userptr: Try to acquire the page lock around set_page_dirty()
  drm/i915/pmu: "Frequency" is reported as accumulated cycles
  drm/i915: Preload LUTs if the hw isn't currently using them
  drm/i915: Don't oops in dumb_create ioctl if we have no crtcs

5 years agomm/ksm.c: don't WARN if page is still mapped in remove_stable_node()
Andrey Ryabinin [Fri, 22 Nov 2019 01:54:01 +0000 (17:54 -0800)]
mm/ksm.c: don't WARN if page is still mapped in remove_stable_node()

It's possible to hit the WARN_ON_ONCE(page_mapped(page)) in
remove_stable_node() when it races with __mmput() and squeezes in
between ksm_exit() and exit_mmap().

  WARNING: CPU: 0 PID: 3295 at mm/ksm.c:888 remove_stable_node+0x10c/0x150

  Call Trace:
   remove_all_stable_nodes+0x12b/0x330
   run_store+0x4ef/0x7b0
   kernfs_fop_write+0x200/0x420
   vfs_write+0x154/0x450
   ksys_write+0xf9/0x1d0
   do_syscall_64+0x99/0x510
   entry_SYSCALL_64_after_hwframe+0x49/0xbe

Remove the warning as there is nothing scary going on.

Link: http://lkml.kernel.org/r/20191119131850.5675-1-aryabinin@virtuozzo.com
Fixes: cbf86cfe04a6 ("ksm: remove old stable nodes more thoroughly")
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agomm/memory_hotplug: don't access uninitialized memmaps in shrink_zone_span()
David Hildenbrand [Fri, 22 Nov 2019 01:53:56 +0000 (17:53 -0800)]
mm/memory_hotplug: don't access uninitialized memmaps in shrink_zone_span()

Let's limit shrinking to !ZONE_DEVICE so we can fix the current code.
We should never try to touch the memmap of offline sections where we
could have uninitialized memmaps and could trigger BUGs when calling
page_to_nid() on poisoned pages.

There is no reliable way to distinguish an uninitialized memmap from an
initialized memmap that belongs to ZONE_DEVICE, as we don't have
anything like SECTION_IS_ONLINE we can use similar to
pfn_to_online_section() for !ZONE_DEVICE memory.

E.g., set_zone_contiguous() similarly relies on pfn_to_online_section()
and will therefore never set a ZONE_DEVICE zone consecutive.  Stopping
to shrink the ZONE_DEVICE therefore results in no observable changes,
besides /proc/zoneinfo indicating different boundaries - something we
can totally live with.

Before commit d0dc12e86b31 ("mm/memory_hotplug: optimize memory
hotplug"), the memmap was initialized with 0 and the node with the right
value.  So the zone might be wrong but not garbage.  After that commit,
both the zone and the node will be garbage when touching uninitialized
memmaps.

Toshiki reported a BUG (race between delayed initialization of
ZONE_DEVICE memmaps without holding the memory hotplug lock and
concurrent zone shrinking).

  https://lkml.org/lkml/2019/11/14/1040

"Iteration of create and destroy namespace causes the panic as below:

      kernel BUG at mm/page_alloc.c:535!
      CPU: 7 PID: 2766 Comm: ndctl Not tainted 5.4.0-rc4 #6
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.11.0-0-g63451fca13-prebuilt.qemu-project.org 04/01/2014
      RIP: 0010:set_pfnblock_flags_mask+0x95/0xf0
      Call Trace:
       memmap_init_zone_device+0x165/0x17c
       memremap_pages+0x4c1/0x540
       devm_memremap_pages+0x1d/0x60
       pmem_attach_disk+0x16b/0x600 [nd_pmem]
       nvdimm_bus_probe+0x69/0x1c0
       really_probe+0x1c2/0x3e0
       driver_probe_device+0xb4/0x100
       device_driver_attach+0x4f/0x60
       bind_store+0xc9/0x110
       kernfs_fop_write+0x116/0x190
       vfs_write+0xa5/0x1a0
       ksys_write+0x59/0xd0
       do_syscall_64+0x5b/0x180
       entry_SYSCALL_64_after_hwframe+0x44/0xa9

  While creating a namespace and initializing memmap, if you destroy the
  namespace and shrink the zone, it will initialize the memmap outside
  the zone and trigger VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page),
  pfn), page) in set_pfnblock_flags_mask()."

This BUG is also mitigated by this commit, where we for now stop to
shrink the ZONE_DEVICE zone until we can do it in a safe and clean way.

Link: http://lkml.kernel.org/r/20191006085646.5768-5-david@redhat.com
Fixes: f1dd2cd13c4b ("mm, memory_hotplug: do not associate hotadded memory to zones until online") [visible after d0dc12e86b319]
Signed-off-by: David Hildenbrand <david@redhat.com>
Reported-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Reported-by: Toshiki Fukasawa <t-fukasawa@vx.jp.nec.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Damian Tometzki <damian.tometzki@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Halil Pasic <pasic@linux.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Jun Yao <yaojun8558363@gmail.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Pankaj Gupta <pagupta@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Pavel Tatashin <pavel.tatashin@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Qian Cai <cai@lca.pw>
Cc: Rich Felker <dalias@libc.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Steve Capper <steve.capper@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Will Deacon <will@kernel.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Yu Zhao <yuzhao@google.com>
Cc: <stable@vger.kernel.org> [4.13+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoRevert "fs: ocfs2: fix possible null-pointer dereferences in ocfs2_xa_prepare_entry()"
Joseph Qi [Fri, 22 Nov 2019 01:53:52 +0000 (17:53 -0800)]
Revert "fs: ocfs2: fix possible null-pointer dereferences in ocfs2_xa_prepare_entry()"

This reverts commit 56e94ea132bb5c2c1d0b60a6aeb34dcb7d71a53d.

Commit 56e94ea132bb ("fs: ocfs2: fix possible null-pointer dereferences
in ocfs2_xa_prepare_entry()") introduces a regression that fail to
create directory with mount option user_xattr and acl.  Actually the
reported NULL pointer dereference case can be correctly handled by
loc->xl_ops->xlo_add_entry(), so revert it.

Link: http://lkml.kernel.org/r/1573624916-83825-1-git-send-email-joseph.qi@linux.alibaba.com
Fixes: 56e94ea132bb ("fs: ocfs2: fix possible null-pointer dereferences in ocfs2_xa_prepare_entry()")
Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Reported-by: Thomas Voegtle <tv@lio96.de>
Acked-by: Changwei Ge <gechangwei@live.cn>
Cc: Jia-Ju Bai <baijiaju1990@gmail.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agocan: m_can_platform: remove unnecessary m_can_class_resume() call
Pankaj Sharma [Tue, 19 Nov 2019 10:20:38 +0000 (15:50 +0530)]
can: m_can_platform: remove unnecessary m_can_class_resume() call

The function m_can_runtime_resume() is getting recursively called from
m_can_class_resume(). This results in a lock up.

We need not call m_can_class_resume() during m_can_runtime_resume().

Fixes: f524f829b75a ("can: m_can: Create a m_can platform framework")
Signed-off-by: Pankaj Sharma <pankj.sharma@samsung.com>
Signed-off-by: Sriram Dash <sriram.dash@samsung.com>
Acked-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: m_can_platform: set net_device structure as driver data
Pankaj Sharma [Tue, 19 Nov 2019 10:20:37 +0000 (15:50 +0530)]
can: m_can_platform: set net_device structure as driver data

The current code is failing during clock prepare enable because of not
getting proper clock from platform device.

[    0.852089] Call trace:
[    0.854516]  0xffff0000fa22a668
[    0.857638]  clk_prepare+0x20/0x34
[    0.861019]  m_can_runtime_resume+0x2c/0xe4
[    0.865180]  pm_generic_runtime_resume+0x28/0x38
[    0.869770]  __rpm_callback+0x16c/0x1bc
[    0.873583]  rpm_callback+0x24/0x78
[    0.877050]  rpm_resume+0x428/0x560
[    0.880517]  __pm_runtime_resume+0x7c/0xa8
[    0.884593]  m_can_clk_start.isra.9.part.10+0x1c/0xa8
[    0.889618]  m_can_class_register+0x138/0x370
[    0.893950]  m_can_plat_probe+0x120/0x170
[    0.897939]  platform_drv_probe+0x4c/0xa0
[    0.901924]  really_probe+0xd8/0x31c
[    0.905477]  driver_probe_device+0x58/0xe8
[    0.909551]  device_driver_attach+0x68/0x70
[    0.913711]  __driver_attach+0x9c/0xf8
[    0.917437]  bus_for_each_dev+0x50/0xa0
[    0.921251]  driver_attach+0x20/0x28
[    0.924804]  bus_add_driver+0x148/0x1fc
[    0.928617]  driver_register+0x6c/0x124
[    0.932431]  __platform_driver_register+0x48/0x50
[    0.937113]  m_can_plat_driver_init+0x18/0x20
[    0.941446]  do_one_initcall+0x4c/0x19c
[    0.945259]  kernel_init_freeable+0x1d0/0x280
[    0.949591]  kernel_init+0x10/0x100
[    0.953057]  ret_from_fork+0x10/0x18
[    0.956614] Code: 00000000 00000000 00000000 00000000 (fa22a668)
[    0.962681] ---[ end trace 881f71bd609de763 ]---
[    0.967301] Kernel panic - not syncing: Attempted to kill init!

A device driver for CAN controller hardware registers itself with the
Linux network layer as a network device. So, the driver data for m_can
should ideally be of type net_device.

Fixes: f524f829b75a ("can: m_can: Create a m_can platform framework")
Signed-off-by: Pankaj Sharma <pankj.sharma@samsung.com>
Signed-off-by: Sriram Dash <sriram.dash@samsung.com>
Acked-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agomac80211: Use Airtime-based Queue Limits (AQL) on packet dequeue
Toke Høiland-Jørgensen [Tue, 19 Nov 2019 06:06:10 +0000 (22:06 -0800)]
mac80211: Use Airtime-based Queue Limits (AQL) on packet dequeue

The previous commit added the ability to throttle stations when they queue
too much airtime in the hardware. This commit enables the functionality by
calculating the expected airtime usage of each packet that is dequeued from
the TXQs in mac80211, and accounting that as pending airtime.

The estimated airtime for each skb is stored in the tx_info, so we can
subtract the same amount from the running total when the skb is freed or
recycled. The throttling mechanism relies on this accounting to be
accurate (i.e., that we are not freeing skbs without subtracting any
airtime they were accounted for), so we put the subtraction into
ieee80211_report_used_skb(). As an optimisation, we also subtract the
airtime on regular TX completion, zeroing out the value stored in the
packet afterwards, to avoid having to do an expensive lookup of the station
from the packet data on every packet.

This patch does *not* include any mechanism to wake a throttled TXQ again,
on the assumption that this will happen anyway as a side effect of whatever
freed the skb (most commonly a TX completion).

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20191119060610.76681-5-kyan@google.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: Implement Airtime-based Queue Limit (AQL)
Kan Yan [Tue, 19 Nov 2019 06:06:09 +0000 (22:06 -0800)]
mac80211: Implement Airtime-based Queue Limit (AQL)

In order for the Fq_CoDel algorithm integrated in mac80211 layer to operate
effectively to control excessive queueing latency, the CoDel algorithm
requires an accurate measure of how long packets stays in the queue, AKA
sojourn time. The sojourn time measured at the mac80211 layer doesn't
include queueing latency in the lower layer (firmware/hardware) and CoDel
expects lower layer to have a short queue. However, most 802.11ac chipsets
offload tasks such TX aggregation to firmware or hardware, thus have a deep
lower layer queue.

Without a mechanism to control the lower layer queue size, packets only
stay in mac80211 layer transiently before being sent to firmware queue.
As a result, the sojourn time measured by CoDel in the mac80211 layer is
almost always lower than the CoDel latency target, hence CoDel does little
to control the latency, even when the lower layer queue causes excessive
latency.

The Byte Queue Limits (BQL) mechanism is commonly used to address the
similar issue with wired network interface. However, this method cannot be
applied directly to the wireless network interface. "Bytes" is not a
suitable measure of queue depth in the wireless network, as the data rate
can vary dramatically from station to station in the same network, from a
few Mbps to over Gbps.

This patch implements an Airtime-based Queue Limit (AQL) to make CoDel work
effectively with wireless drivers that utilized firmware/hardware
offloading. AQL allows each txq to release just enough packets to the lower
layer to form 1-2 large aggregations to keep hardware fully utilized and
retains the rest of the frames in mac80211 layer to be controlled by the
CoDel algorithm.

Signed-off-by: Kan Yan <kyan@google.com>
[ Toke: Keep API to set pending airtime internal, fix nits in commit msg ]
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20191119060610.76681-4-kyan@google.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: Import airtime calculation code from mt76
Toke Høiland-Jørgensen [Tue, 19 Nov 2019 06:06:08 +0000 (22:06 -0800)]
mac80211: Import airtime calculation code from mt76

Felix recently added code to calculate airtime of packets to the mt76
driver. Import this into mac80211 so we can use it for airtime queue limit
calculations.

The airtime.c file is copied verbatim from the mt76 driver, and adjusted to
be usable in mac80211. This involves:

- Switching to mac80211 data structures.
- Adding support for 160 MHz channels and HE mode.
- Moving the symbol and duration calculations around a bit to avoid
  rounding with the higher rates and longer symbol times used for HE rates.

The per-rate TX rate calculation is also split out to its own function so
it can be used directly for the AQL calculations later.

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20191119060610.76681-3-kyan@google.com
[fix HE_GROUP_IDX() to use 3 * bw, since there are 3 _gi values]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agovirt_wifi: fix use-after-free in virt_wifi_newlink()
Taehee Yoo [Thu, 21 Nov 2019 12:26:45 +0000 (12:26 +0000)]
virt_wifi: fix use-after-free in virt_wifi_newlink()

When virt_wifi interface is created, virt_wifi_newlink() is called and
it calls register_netdevice().
if register_netdevice() fails, it internally would call
->priv_destructor(), which is virt_wifi_net_device_destructor() and
it frees netdev. but virt_wifi_newlink() still use netdev.
So, use-after-free would occur in virt_wifi_newlink().

Test commands:
    ip link add dummy0 type dummy
    modprobe bonding
    ip link add bonding_masters link dummy0 type virt_wifi

Splat looks like:
[  202.220554] BUG: KASAN: use-after-free in virt_wifi_newlink+0x88b/0x9a0 [virt_wifi]
[  202.221659] Read of size 8 at addr ffff888061629cb8 by task ip/852

[  202.222896] CPU: 1 PID: 852 Comm: ip Not tainted 5.4.0-rc5 #3
[  202.223765] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[  202.225073] Call Trace:
[  202.225532]  dump_stack+0x7c/0xbb
[  202.226869]  print_address_description.constprop.5+0x1be/0x360
[  202.229362]  __kasan_report+0x12a/0x16f
[  202.230714]  kasan_report+0xe/0x20
[  202.232595]  virt_wifi_newlink+0x88b/0x9a0 [virt_wifi]
[  202.233370]  __rtnl_newlink+0xb9f/0x11b0
[  202.244909]  rtnl_newlink+0x65/0x90
[ ... ]

Cc: stable@vger.kernel.org
Fixes: c7cdba31ed8b ("mac80211-next: rtnetlink wifi simulation device")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Link: https://lore.kernel.org/r/20191121122645.9355-1-ap420073@gmail.com
[trim stack dump a bit]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: consider QoS Null frames for STA_NULLFUNC_ACKED
Thomas Pedersen [Tue, 19 Nov 2019 05:35:38 +0000 (21:35 -0800)]
mac80211: consider QoS Null frames for STA_NULLFUNC_ACKED

Commit 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing")
let STAs send QoS Null frames as PS triggers if the AP was
a QoS STA.  However, the mac80211 PS stack relies on an
interface flag IEEE80211_STA_NULLFUNC_ACKED for
determining trigger frame ACK, which was not being set for
acked non-QoS Null frames. The effect is an inability to
trigger hardware sleep via IEEE80211_CONF_PS since the QoS
Null frame was seemingly never acked.

This bug only applies to drivers which set both
IEEE80211_HW_REPORTS_TX_ACK_STATUS and
IEEE80211_HW_PS_NULLFUNC_STACK.

Detect the acked QoS Null frame to restore STA power save.

Fixes: 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing")
Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20191119053538.25979-4-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: expose HW conf flags through debugfs
Thomas Pedersen [Tue, 19 Nov 2019 05:35:37 +0000 (21:35 -0800)]
mac80211: expose HW conf flags through debugfs

This is useful during testing to eg. check the currently
configured HW power save state.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20191119053538.25979-3-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: Add new sta_info getter by sta/vif addrs
Toke Høiland-Jørgensen [Tue, 12 Nov 2019 13:08:35 +0000 (14:08 +0100)]
mac80211: Add new sta_info getter by sta/vif addrs

In ieee80211_tx_status() we don't have an sdata struct when looking up the
destination sta. Instead, we just do a lookup by the vif addr that is the
source of the packet being completed. Factor this out into a new sta_info
getter helper, since we need to use it for accounting AQL as well.

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20191112130835.382062-1-toke@redhat.com
[remove internal rcu_read_lock(), document instead]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agomac80211: add a comment about monitor-to-dev injection
Johannes Berg [Fri, 22 Nov 2019 11:42:42 +0000 (12:42 +0100)]
mac80211: add a comment about monitor-to-dev injection

Add a note with a use-case for the monitor-to-dev injection
mechanism in mac80211, reported by Ben Greear.

Change-Id: I6456997ef9bc40b24ede860b6ef2fed5af49cf44
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
5 years agoMerge branch 'hv_netvsc-Fix-send-indirection-table-offset'
David S. Miller [Fri, 22 Nov 2019 03:32:23 +0000 (19:32 -0800)]
Merge branch 'hv_netvsc-Fix-send-indirection-table-offset'

Haiyang Zhang says:

====================
hv_netvsc: Fix send indirection table offset

Fix send indirection table offset issues related to guest and
host bugs.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agohv_netvsc: Fix send_table offset in case of a host bug
Haiyang Zhang [Thu, 21 Nov 2019 21:33:41 +0000 (13:33 -0800)]
hv_netvsc: Fix send_table offset in case of a host bug

If negotiated NVSP version <= NVSP_PROTOCOL_VERSION_6, the offset may
be wrong (too small) due to a host bug. This can cause missing the
end of the send indirection table, and add multiple zero entries from
leading zeros before the data region. This bug adds extra burden on
channel 0.

So fix the offset by computing it from the data structure sizes. This
will ensure netvsc driver runs normally on unfixed hosts, and future
fixed hosts.

Fixes: 5b54dac856cb ("hyperv: Add support for virtual Receive Side Scaling (vRSS)")
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agohv_netvsc: Fix offset usage in netvsc_send_table()
Haiyang Zhang [Thu, 21 Nov 2019 21:33:40 +0000 (13:33 -0800)]
hv_netvsc: Fix offset usage in netvsc_send_table()

To reach the data region, the existing code adds offset in struct
nvsp_5_send_indirect_table on the beginning of this struct. But the
offset should be based on the beginning of its container,
struct nvsp_message. This bug causes the first table entry missing,
and adds an extra zero from the zero pad after the data region.
This can put extra burden on the channel 0.

So, correct the offset usage. Also add a boundary check to ensure
not reading beyond data region.

Fixes: 5b54dac856cb ("hyperv: Add support for virtual Receive Side Scaling (vRSS)")
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoenetc: make enetc_setup_tc_mqprio static
Mao Wenan [Fri, 22 Nov 2019 02:52:40 +0000 (10:52 +0800)]
enetc: make enetc_setup_tc_mqprio static

While using ARCH=mips CROSS_COMPILE=mips-linux-gnu- command to compile,
make C=2 drivers/net/ethernet/freescale/enetc/enetc.o

one warning can be found:
drivers/net/ethernet/freescale/enetc/enetc.c:1439:5:
warning: symbol 'enetc_setup_tc_mqprio' was not declared.
Should it be static?

This patch make symbol enetc_setup_tc_mqprio static.
Fixes: 34c6adf1977b ("enetc: Configure the Time-Aware Scheduler via tc-taprio offload")
Signed-off-by: Mao Wenan <maowenan@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet-ipv6: IPV6_TRANSPARENT - check NET_RAW prior to NET_ADMIN
Maciej Żenczykowski [Thu, 21 Nov 2019 21:19:08 +0000 (13:19 -0800)]
net-ipv6: IPV6_TRANSPARENT - check NET_RAW prior to NET_ADMIN

NET_RAW is less dangerous, so more likely to be available to a process,
so check it first to prevent some spurious logging.

This matches IP_TRANSPARENT which checks NET_RAW first.

Signed-off-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'drm-intel-fixes-2019-11-21' of git://anongit.freedesktop.org/drm/drm-intel...
Dave Airlie [Fri, 22 Nov 2019 00:23:22 +0000 (10:23 +1000)]
Merge tag 'drm-intel-fixes-2019-11-21' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

- Fix kernel oops on dumb_create ioctl on no crtc situation
- Fix bad ugly colored flash on VLV/CHV related to gamma LUT update
- Fix unity of the frequencies reported on PMU
- Fix kernel oops on set_page_dirty using better locks around it
- Protect the request pointer with RCU to prevent it being freed while we might need still
- Make pool objects read-only
- Restore physical addresses for fb_map to avoid corrupted page table

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191121165339.GA23920@intel.com
5 years agoMerge branch 'net-introduce-and-use-route-hint'
David S. Miller [Thu, 21 Nov 2019 22:45:55 +0000 (14:45 -0800)]
Merge branch 'net-introduce-and-use-route-hint'

Paolo Abeni says:

====================
net: introduce and use route hint

This series leverages the listification infrastructure to avoid
unnecessary route lookup on ingress packets. In absence of custom rules,
packets with equal daddr will usually land on the same dst.

When processing packet bursts (lists) we can easily reference the previous
dst entry. When we hit the 'same destination' condition we can avoid the
route lookup, coping the already available dst.

Detailed performance numbers are available in the individual commit
messages.

v3 -> v4:
 - move helpers to their own patches (Eric D.)
 - enable hints for SUBTREE builds (David A.)
 - re-enable hints for ipv4 forward (David A.)

v2 -> v3:
 - use fib*_has_custom_rules() helpers (David A.)
 - add ip*_extract_route_hint() helper (Edward C.)
 - use prev skb as hint instead of copying data (Willem )

v1 -> v2:
 - fix build issue with !CONFIG_IP*_MULTIPLE_TABLES
 - fix potential race in ip6_list_rcv_finish()
====================

Acked-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv4: use dst hint for ipv4 list receive
Paolo Abeni [Wed, 20 Nov 2019 12:47:37 +0000 (13:47 +0100)]
ipv4: use dst hint for ipv4 list receive

This is alike the previous change, with some additional ipv4 specific
quirk. Even when using the route hint we still have to do perform
additional per packet checks about source address validity: a new
helper is added to wrap them.

Hints are explicitly disabled if the destination is a local broadcast,
that keeps the code simple and local broadcast are a slower path anyway.

UDP flood performances vs recvmmsg() receiver:

vanilla patched delta
Kpps Kpps %
1683 1871 +11

In the worst case scenario - each packet has a different
destination address - the performance delta is within noise
range.

v3 -> v4:
 - re-enable hints for forward

v2 -> v3:
 - really fix build (sic) and hint usage check
 - use fib4_has_custom_rules() helpers (David A.)
 - add ip_extract_route_hint() helper (Edward C.)
 - use prev skb as hint instead of copying data (Willem)

v1 -> v2:
 - fix build issue with !CONFIG_IP_MULTIPLE_TABLES

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv4: move fib4_has_custom_rules() helper to public header
Paolo Abeni [Wed, 20 Nov 2019 12:47:36 +0000 (13:47 +0100)]
ipv4: move fib4_has_custom_rules() helper to public header

So that we can use it in the next patch.
Additionally constify the helper argument.

Suggested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: introduce and uses route look hints for list input.
Paolo Abeni [Wed, 20 Nov 2019 12:47:35 +0000 (13:47 +0100)]
ipv6: introduce and uses route look hints for list input.

When doing RX batch packet processing, we currently always repeat
the route lookup for each ingress packet. When no custom rules are
in place, and there aren't routes depending on source addresses,
we know that packets with the same destination address will use
the same dst.

This change tries to avoid per packet route lookup caching
the destination address of the latest successful lookup, and
reusing it for the next packet when the above conditions are
in place. Ingress traffic for most servers should fit.

The measured performance delta under UDP flood vs a recvmmsg
receiver is as follow:

vanilla patched delta
Kpps Kpps %
1431 1674 +17

In the worst-case scenario - each packet has a different
destination address - the performance delta is within noise
range.

v3 -> v4:
 - support hints for SUBFLOW build, too (David A.)
 - several style fixes (Eric)

v2 -> v3:
 - add fib6_has_custom_rules() helpers (David A.)
 - add ip6_extract_route_hint() helper (Edward C.)
 - use hint directly in ip6_list_rcv_finish() (Willem)

v1 -> v2:
 - fix build issue with !CONFIG_IPV6_MULTIPLE_TABLES
 - fix potential race when fib6_has_custom_rules is set
   while processing a packet batch

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: keep track of routes using src
Paolo Abeni [Wed, 20 Nov 2019 12:47:34 +0000 (13:47 +0100)]
ipv6: keep track of routes using src

Use a per namespace counter, increment it on successful creation
of any route using the source address, decrement it on deletion
of such routes.

This allows us to check easily if the routing decision in the
current namespace depends on the packet source. Will be used
by the next patch.

Suggested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: add fib6_has_custom_rules() helper
Paolo Abeni [Wed, 20 Nov 2019 12:47:33 +0000 (13:47 +0100)]
ipv6: add fib6_has_custom_rules() helper

It wraps the namespace field with the same name, to easily
access it regardless of build options.

Suggested-by: David Ahern <dsahern@gmail.com>
Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'DSA-Felix-PTP'
David S. Miller [Thu, 21 Nov 2019 22:39:03 +0000 (14:39 -0800)]
Merge branch 'DSA-Felix-PTP'

Yangbo Lu says:

====================
Support PTP clock and hardware timestamping for DSA Felix driver

This patch-set is to support PTP clock and hardware timestamping
for DSA Felix driver. Some functions in ocelot.c/ocelot_board.c
driver were reworked/exported, so that DSA Felix driver was able
to reuse them as much as possible.

On TX path, timestamping works on packet which requires timestamp.
The injection header will be configured accordingly, and skb clone
requires timestamp will be added into a list. The TX timestamp
is final handled in threaded interrupt handler when PTP timestamp
FIFO is ready.
On RX path, timestamping is always working. The RX timestamp could
be got from extraction header.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: dsa: ocelot: add hardware timestamping support for Felix
Yangbo Lu [Wed, 20 Nov 2019 08:23:18 +0000 (16:23 +0800)]
net: dsa: ocelot: add hardware timestamping support for Felix

This patch is to reuse ocelot functions as possible to enable PTP
clock and to support hardware timestamping on Felix.
On TX path, timestamping works on packet which requires timestamp.
The injection header will be configured accordingly, and skb clone
requires timestamp will be added into a list. The TX timestamp
is final handled in threaded interrupt handler when PTP timestamp
FIFO is ready.
On RX path, timestamping is always working. The RX timestamp could
be got from extraction header.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: dsa: ocelot: define PTP registers for felix_vsc9959
Yangbo Lu [Wed, 20 Nov 2019 08:23:17 +0000 (16:23 +0800)]
net: dsa: ocelot: define PTP registers for felix_vsc9959

This patch is to define PTP registers for felix_vsc9959.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: mscc: ocelot: convert to use ocelot_port_add_txtstamp_skb()
Yangbo Lu [Wed, 20 Nov 2019 08:23:16 +0000 (16:23 +0800)]
net: mscc: ocelot: convert to use ocelot_port_add_txtstamp_skb()

Convert to use ocelot_port_add_txtstamp_skb() for adding skbs which
require TX timestamp into list. Export it so that DSA Felix driver
could reuse it too.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: mscc: ocelot: convert to use ocelot_get_txtstamp()
Yangbo Lu [Wed, 20 Nov 2019 08:23:15 +0000 (16:23 +0800)]
net: mscc: ocelot: convert to use ocelot_get_txtstamp()

The method getting TX timestamp by reading timestamp FIFO and
matching skbs list is common for DSA Felix driver too.
So move code out of ocelot_board.c, convert to use
ocelot_get_txtstamp() function and export it.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: mscc: ocelot: export ocelot_hwstamp_get/set functions
Yangbo Lu [Wed, 20 Nov 2019 08:23:14 +0000 (16:23 +0800)]
net: mscc: ocelot: export ocelot_hwstamp_get/set functions

Export ocelot_hwstamp_get/set functions so that DSA driver
is able to reuse them.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agobpf: skmsg, fix potential psock NULL pointer dereference
John Fastabend [Thu, 21 Nov 2019 16:25:09 +0000 (08:25 -0800)]
bpf: skmsg, fix potential psock NULL pointer dereference

Report from Dan Carpenter,

 net/core/skmsg.c:792 sk_psock_write_space()
 error: we previously assumed 'psock' could be null (see line 790)

 net/core/skmsg.c
   789 psock = sk_psock(sk);
   790 if (likely(psock && sk_psock_test_state(psock, SK_PSOCK_TX_ENABLED)))
 Check for NULL
   791 schedule_work(&psock->work);
   792 write_space = psock->saved_write_space;
                     ^^^^^^^^^^^^^^^^^^^^^^^^
   793          rcu_read_unlock();
   794          write_space(sk);

Ensure psock dereference on line 792 only occurs if psock is not null.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Thu, 21 Nov 2019 20:15:24 +0000 (12:15 -0800)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fix from Will Deacon:
 "Ensure PAN is re-enabled following user fault in uaccess routines.

  After I thought we were done for 5.4, we had a report this week of a
  nasty issue that has been shown to leak data between different user
  address spaces thanks to corruption of entries in the TLB. In
  hindsight, we should have spotted this in review when the PAN code was
  merged back in v4.3, but hindsight is 20/20 and I'm trying not to beat
  myself up too much about it despite being fairly miserable.

  Anyway, the fix is "obvious" but the actual failure is more more
  subtle, and is described in the commit message. I've included a fairly
  mechanical follow-up patch here as well, which moves this checking out
  into the C wrappers which is what we do for {get,put}_user() already
  and allows us to remove these bloody assembly macros entirely. The
  patches have passed kernelci [1] [2] [3] and CKI [4] tests over night,
  as well as some targetted testing [5] for this particular issue.

  The first patch is tagged for stable and should be applied to 4.14,
  4.19 and 5.3. I have separate backports for 4.4 and 4.9, which I'll
  send out once this has landed in your tree (although the original
  patch applies cleanly, it won't build for those two trees).

  Thanks to Pavel Tatashin for reporting this and Mark Rutland for
  helping to diagnose the issue and review/test the solution"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: uaccess: Remove uaccess_*_not_uao asm macros
  arm64: uaccess: Ensure PAN is re-enabled after unhandled uaccess fault

5 years agoaudit: Move audit_log_task declaration under CONFIG_AUDITSYSCALL
Jiri Olsa [Thu, 21 Nov 2019 15:58:53 +0000 (16:58 +0100)]
audit: Move audit_log_task declaration under CONFIG_AUDITSYSCALL

The 0-DAY found that audit_log_task is not declared under
CONFIG_AUDITSYSCALL which causes compilation error when
it is not defined:

    kernel/bpf/syscall.o: In function `bpf_audit_prog.isra.30':
 >> syscall.c:(.text+0x860): undefined reference to `audit_log_task'

Adding the audit_log_task declaration and stub within
CONFIG_AUDITSYSCALL ifdef.

Fixes: 91e6015b082b ("bpf: Emit audit messages upon successful prog load and unload")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agosfc: Only cancel the PPS workqueue if it exists
Martin Habets [Thu, 21 Nov 2019 17:52:15 +0000 (17:52 +0000)]
sfc: Only cancel the PPS workqueue if it exists

The workqueue only exists for the primary PF. For other functions
we hit a WARN_ON in kernel/workqueue.c.

Fixes: 7c236c43b838 ("sfc: Add support for IEEE-1588 PTP")
Signed-off-by: Martin Habets <mhabets@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'for-linus-20191121' of git://git.kernel.dk/linux-block
Linus Torvalds [Thu, 21 Nov 2019 20:04:50 +0000 (12:04 -0800)]
Merge tag 'for-linus-20191121' of git://git.kernel.dk/linux-block

Pull block fix from Jens Axboe:
 "Just a single fix for an issue in nbd introduced in this cycle"

* tag 'for-linus-20191121' of git://git.kernel.dk/linux-block:
  nbd:fix memory leak in nbd_get_socket()

5 years agoMerge tag 'gpio-v5.4-5' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux...
Linus Torvalds [Thu, 21 Nov 2019 20:01:30 +0000 (12:01 -0800)]
Merge tag 'gpio-v5.4-5' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fixes from Linus Walleij:
 "A last set of small fixes for GPIO, this cycle was quite busy.

   - Fix debounce delays on the MAX77620 GPIO expander

   - Use the correct unit for debounce times on the BD70528 GPIO expander

   - Get proper deps for parallel builds of the GPIO tools

   - Add a specific ACPI quirk for the Terra Pad 1061"

* tag 'gpio-v5.4-5' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpiolib: acpi: Add Terra Pad 1061 to the run_edge_events_on_boot_blacklist
  tools: gpio: Correctly add make dependencies for gpio_utils
  gpio: bd70528: Use correct unit for debounce times
  gpio: max77620: Fixup debounce delays

5 years agonet: Fix Kconfig indentation, continued
Krzysztof Kozlowski [Thu, 21 Nov 2019 13:28:35 +0000 (21:28 +0800)]
net: Fix Kconfig indentation, continued

Adjust indentation from spaces to tab (+optional two spaces) as in
coding style.  This fixes various indentation mixups (seven spaces,
tab+one space, etc).

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agodrivers: net: Fix Kconfig indentation, continued
Krzysztof Kozlowski [Thu, 21 Nov 2019 13:28:28 +0000 (21:28 +0800)]
drivers: net: Fix Kconfig indentation, continued

Adjust indentation from spaces to tab (+optional two spaces) as in
coding style.  This fixes various indentation mixups (seven spaces,
tab+one space, etc).

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'for-linus-2019-11-21' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 21 Nov 2019 19:51:49 +0000 (11:51 -0800)]
Merge tag 'for-linus-2019-11-21' of git://git./linux/kernel/git/brauner/linux

Pull pidfd fixlet from Christian Brauner:
 "This contains a simple fix for the pidfd poll method. In the original
  patchset pidfd_poll() was made to return an unsigned int. However, the
  poll method is defined to return a __poll_t. While the unsigned int is
  not a huge deal it's just nicer to return a __poll_t.

  I've decided to send it right before the 5.4 release mainly so that
  stable doesn't need to backport it to both 5.4 and 5.3"

* tag 'for-linus-2019-11-21' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux:
  fork: fix pidfd_poll()'s return type

5 years agonfc: port100: handle command failure cleanly
Oliver Neukum [Thu, 21 Nov 2019 10:37:10 +0000 (11:37 +0100)]
nfc: port100: handle command failure cleanly

If starting the transfer of a command suceeds but the transfer for the reply
fails, it is not enough to initiate killing the transfer for the
command may still be running. You need to wait for the killing to finish
before you can reuse URB and buffer.

Reported-and-tested-by: syzbot+711468aa5c3a1eabf863@syzkaller.appspotmail.com
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolwtunnel: check erspan options before allocating tun_info
Xin Long [Thu, 21 Nov 2019 10:14:50 +0000 (18:14 +0800)]
lwtunnel: check erspan options before allocating tun_info

As Jakub suggested on another patch, it's better to do the check
on erspan options before allocating memory.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolwtunnel: be STRICT to validate the new LWTUNNEL_IP(6)_OPTS
Xin Long [Thu, 21 Nov 2019 10:11:27 +0000 (18:11 +0800)]
lwtunnel: be STRICT to validate the new LWTUNNEL_IP(6)_OPTS

LWTUNNEL_IP(6)_OPTS are the new items in ip(6)_tun_policy, which
are parsed by nla_parse_nested_deprecated(). We should check it
strictly by setting .strict_start_type = LWTUNNEL_IP(6)_OPTS.

This patch also adds missing LWTUNNEL_IP6_OPTS in ip6_tun_policy.

Fixes: 4ece47787077 ("lwtunnel: add options setting and dumping for geneve")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: remove the unnecessary strict_start_type in some policies
Xin Long [Thu, 21 Nov 2019 10:08:38 +0000 (18:08 +0800)]
net: remove the unnecessary strict_start_type in some policies

ct_policy and mpls_policy are parsed with nla_parse_nested(), which
does NL_VALIDATE_STRICT validation, strict_start_type is not needed
to set as it is actually trying to make some attributes parsed with
NL_VALIDATE_STRICT.

This patch is to remove it, and do the same on rtm_nh_policy which
is parsed by nlmsg_parse().

Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'net-sched-support-vxlan-and-erspan-options'
David S. Miller [Thu, 21 Nov 2019 19:44:06 +0000 (11:44 -0800)]
Merge branch 'net-sched-support-vxlan-and-erspan-options'

Xin Long says:

====================
net: sched: support vxlan and erspan options

This patchset is to add vxlan and erspan options support in
cls_flower and act_tunnel_key. The form is pretty much like
geneve_opts in:

  https://patchwork.ozlabs.org/patch/935272/
  https://patchwork.ozlabs.org/patch/954564/

but only one option is allowed for vxlan and erspan.

v1->v2:
  - see each patch changelog.
====================

Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: allow flower to match erspan options
Xin Long [Thu, 21 Nov 2019 10:03:29 +0000 (18:03 +0800)]
net: sched: allow flower to match erspan options

This patch is to allow matching options in erspan.

The options can be described in the form:
VER:INDEX:DIR:HWID/VER:INDEX_MASK:DIR_MASK:HWID_MASK.
When ver is set to 1, index will be applied while dir
and hwid will be ignored, and when ver is set to 2,
dir and hwid will be used while index will be ignored.

Different from geneve, only one option can be set. And
also, geneve options, vxlan options or erspan options
can't be set at the same time.

  # ip link add name erspan1 type erspan external
  # tc qdisc add dev erspan1 ingress
  # tc filter add dev erspan1 protocol ip parent ffff: \
      flower \
        enc_src_ip 10.0.99.192 \
        enc_dst_ip 10.0.99.193 \
        enc_key_id 11 \
        erspan_opts 1:12:0:0/1:ffff:0:0 \
        ip_proto udp \
        action mirred egress redirect dev eth0

v1->v2:
  - improve some err msgs of extack.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: allow flower to match vxlan options
Xin Long [Thu, 21 Nov 2019 10:03:28 +0000 (18:03 +0800)]
net: sched: allow flower to match vxlan options

This patch is to allow matching gbp option in vxlan.

The options can be described in the form GBP/GBP_MASK,
where GBP is represented as a 32bit hexadecimal value.
Different from geneve, only one option can be set. And
also, geneve options and vxlan options can't be set at
the same time.

  # ip link add name vxlan0 type vxlan dstport 0 external
  # tc qdisc add dev vxlan0 ingress
  # tc filter add dev vxlan0 protocol ip parent ffff: \
      flower \
        enc_src_ip 10.0.99.192 \
        enc_dst_ip 10.0.99.193 \
        enc_key_id 11 \
        vxlan_opts 01020304/ffffffff \
        ip_proto udp \
        action mirred egress redirect dev eth0

v1->v2:
  - add .strict_start_type for enc_opts_policy as Jakub noticed.
  - use Duplicate instead of Wrong in err msg for extack as Jakub
    suggested.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: add erspan option support to act_tunnel_key
Xin Long [Thu, 21 Nov 2019 10:03:27 +0000 (18:03 +0800)]
net: sched: add erspan option support to act_tunnel_key

This patch is to allow setting erspan options using the
act_tunnel_key action. Different from geneve options,
only one option can be set. And also, geneve options,
vxlan options or erspan options can't be set at the
same time.

Options are expressed as ver:index:dir:hwid, when ver
is set to 1, index will be applied while dir and hwid
will be ignored, and when ver is set to 2, dir and
hwid will be used while index will be ignored.

  # ip link add name erspan1 type erspan external
  # tc qdisc add dev eth0 ingress
  # tc filter add dev eth0 protocol ip parent ffff: \
           flower indev eth0 \
              ip_proto udp \
              action tunnel_key \
                  set src_ip 10.0.99.192 \
                  dst_ip 10.0.99.193 \
                  dst_port 6081 \
                  id 11 \
   erspan_opts 1:2:0:0 \
          action mirred egress redirect dev erspan1

v1->v2:
  - do the validation when dst is not yet allocated as Jakub suggested.
  - use Duplicate instead of Wrong in err msg for extack.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: add vxlan option support to act_tunnel_key
Xin Long [Thu, 21 Nov 2019 10:03:26 +0000 (18:03 +0800)]
net: sched: add vxlan option support to act_tunnel_key

This patch is to allow setting vxlan options using the
act_tunnel_key action. Different from geneve options,
only one option can be set. And also, geneve options
and vxlan options can't be set at the same time.

gbp is the only param for vxlan options:

  # ip link add name vxlan0 type vxlan dstport 0 external
  # tc qdisc add dev eth0 ingress
  # tc filter add dev eth0 protocol ip parent ffff: \
           flower indev eth0 \
              ip_proto udp \
              action tunnel_key \
                  set src_ip 10.0.99.192 \
                  dst_ip 10.0.99.193 \
                  dst_port 6081 \
                  id 11 \
     vxlan_opts 01020304 \
          action mirred egress redirect dev vxlan0

v1->v2:
  - add .strict_start_type for enc_opts_policy as Jakub noticed.
  - use Duplicate instead of Wrong in err msg for extack as Jakub
    suggested.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoocteontx2-af: Fix uninitialized variable in debugfs
Dan Carpenter [Thu, 21 Nov 2019 09:21:46 +0000 (12:21 +0300)]
octeontx2-af: Fix uninitialized variable in debugfs

If rvu_get_blkaddr() fails, then this rvu_cgx_nix_cuml_stats() returns
zero and we write some uninitialized data into the debugfs output.

On the error paths, the use of the uninitialized "*stat" is harmless,
but it will lead to a Smatch warning (static analysis) and a UBSan
warning (runtime analysis) so we should prevent that as well.

Fixes: f967488d095e ("octeontx2-af: Add per CGX port level NIX Rx/Tx counters")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agovsock: avoid to assign transport if its initialization fails
Stefano Garzarella [Thu, 21 Nov 2019 09:06:09 +0000 (10:06 +0100)]
vsock: avoid to assign transport if its initialization fails

If transport->init() fails, we can't assign the transport to the
socket, because it's not initialized correctly, and any future
calls to the transport callbacks would have an unexpected behavior.

Fixes: c0cfa2d8a788 ("vsock: add multi-transports support")
Reported-and-tested-by: syzbot+e2e5c07bf353b2f79daa@syzkaller.appspotmail.com
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Jorgen Hansen <jhansen@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agodrm/i915/fbdev: Restore physical addresses for fb_mmap()
Chris Wilson [Wed, 13 Nov 2019 18:06:33 +0000 (18:06 +0000)]
drm/i915/fbdev: Restore physical addresses for fb_mmap()

fbdev uses the physical address of our framebuffer for its fb_mmap()
routine. While we need to adapt this address for the new io BAR, we have
to fix v5.4 first! The simplest fix is to restore the smem back to v5.3
and we will then probably have to implement our fbops->fb_mmap() callback
to handle local memory.

Reported-by: Neil MacLeod <freedesktop@nmacleod.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112256
Fixes: 5f889b9a61dd ("drm/i915: Disregard drm_mode_config.fb_base")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Tested-by: Neil MacLeod <freedesktop@nmacleod.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191113180633.3947-1-chris@chris-wilson.co.uk
(cherry picked from commit abc5520704ab438099fe352636b30b05c1253bea)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
(cherry picked from commit 9faf5fa4d3dad3b0c0fa6e67689c144981a11c27)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
5 years agonet: sfp: soft status and control support
Russell King [Wed, 20 Nov 2019 12:29:59 +0000 (12:29 +0000)]
net: sfp: soft status and control support

Add support for the soft status and control register, which allows
TX_FAULT and RX_LOS to be monitored and TX_DISABLE to be set.  We
make use of this when the board does not support GPIOs for these
signals.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'sfp-quirks'
David S. Miller [Thu, 21 Nov 2019 06:28:53 +0000 (22:28 -0800)]
Merge branch 'sfp-quirks'

Russell King says:

====================
Add rudimentary SFP module quirk support

The SFP module EEPROM describes the capabilities of the module, but
doesn't describe the host interface.  We have a certain amount of
guess-work to work out how to configure the host - which works most
of the time.

However, there are some (such as GPON) modules which are able to
support different host interfaces, such as 1000BASE-X and 2500BASE-X.
The module will switch between each mode until it achieves link with
the host.

There is no defined way to describe this in the SFP EEPROM, so we can
only recognise the module and handle it appropriately.  This series
adds the necessary recognition of the modules using a quirk system,
and tweaks the support mask to allow them to link with the host at
2500BASE-X, thereby allowing the user to achieve full line rate.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sfp: add some quirks for GPON modules
Russell King [Wed, 20 Nov 2019 11:42:47 +0000 (11:42 +0000)]
net: sfp: add some quirks for GPON modules

Marc Micalizzi reports that Huawei MA5671A and Alcatel/Lucent G-010S-P
modules are capable of 2500base-X, but incorrectly report their
capabilities in the EEPROM.  It seems rather common that GPON modules
mis-report.

Let's fix these modules by adding some quirks.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sfp: add support for module quirks
Russell King [Wed, 20 Nov 2019 11:42:42 +0000 (11:42 +0000)]
net: sfp: add support for module quirks

Add support for applying module quirks to the list of supported
ethtool link modes.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotcp: warn if offset reach the maxlen limit when using snprintf
Hangbin Liu [Wed, 20 Nov 2019 08:38:08 +0000 (16:38 +0800)]
tcp: warn if offset reach the maxlen limit when using snprintf

snprintf returns the number of chars that would be written, not number
of chars that were actually written. As such, 'offs' may get larger than
'tbl.maxlen', causing the 'tbl.maxlen - offs' being < 0, and since the
parameter is size_t, it would overflow.

Since using scnprintf may hide the limit error, while the buffer is still
enough now, let's just add a WARN_ON_ONCE in case it reach the limit
in future.

v2: Use WARN_ON_ONCE as Jiri and Eric suggested.

Suggested-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoip_gre: Make none-tun-dst gre tunnel store tunnel info as metadat_dst in recv
wenxu [Wed, 20 Nov 2019 02:59:39 +0000 (10:59 +0800)]
ip_gre: Make none-tun-dst gre tunnel store tunnel info as metadat_dst in recv

Currently collect_md gre tunnel will store the tunnel info(metadata_dst)
to skb_dst.
And now the non-tun-dst gre tunnel already can add tunnel header through
lwtunnel.

When received a arp_request on the non-tun-dst gre tunnel. The packet of
arp response will send through the non-tun-dst tunnel without tunnel info
which will lead the arp response packet to be dropped.

If the non-tun-dst gre tunnel also store the tunnel info as metadata_dst,
The arp response packet will set the releted tunnel info in the
iptunnel_metadata_reply.

The following is the test script:

ip netns add cl
ip l add dev vethc type veth peer name eth0 netns cl

ifconfig vethc 172.168.0.7/24 up
ip l add dev tun1000 type gretap key 1000

ip link add user1000 type vrf table 1
ip l set user1000 up
ip l set dev tun1000 master user1000
ifconfig tun1000 10.0.1.1/24 up

ip netns exec cl ifconfig eth0 172.168.0.17/24 up
ip netns exec cl ip l add dev tun type gretap local 172.168.0.17 remote 172.168.0.7 key 1000
ip netns exec cl ifconfig tun 10.0.1.7/24 up
ip r r 10.0.1.7 encap ip id 1000 dst 172.168.0.17 key dev tun1000 table 1

With this patch
ip netns exec cl ping 10.0.1.1 can success

Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet-sysfs: fix netdev_queue_add_kobject() breakage
Eric Dumazet [Thu, 21 Nov 2019 03:19:07 +0000 (19:19 -0800)]
net-sysfs: fix netdev_queue_add_kobject() breakage

kobject_put() should only be called in error path.

Fixes: b8eb718348b8 ("net-sysfs: Fix reference count leak in rx|netdev_queue_add_kobject")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jouni Hogander <jouni.hogander@unikie.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'drm-fixes-5.4-2019-11-20' of git://people.freedesktop.org/~agd5f/linux...
Dave Airlie [Thu, 21 Nov 2019 05:07:35 +0000 (15:07 +1000)]
Merge tag 'drm-fixes-5.4-2019-11-20' of git://people.freedesktop.org/~agd5f/linux into drm-fixes

drm-fixes-5.4-2019-11-20:

amdgpu:
- Remove experimental flag for navi14
- Fix confusing power message failures on older VI parts
- Hang fix for gfxoff when using the read register interface
- Two stability regression fixes for Raven

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191120235130.23755-1-alexander.deucher@amd.com