openwrt/staging/blogic.git
5 years agonet: phylink: Fix flow control for fixed-link
René van Dorst [Sat, 27 Jul 2019 09:40:11 +0000 (11:40 +0200)]
net: phylink: Fix flow control for fixed-link

In phylink_parse_fixedlink() the pl->link_config.advertising bits are AND
with pl->supported, pl->supported is zeroed and only the speed/duplex
modes and MII bits are set.
So pl->link_config.advertising always loses the flow control/pause bits.

By setting Pause and Asym_Pause bits in pl->supported, the flow control
work again when devicetree "pause" is set in fixes-link node and the MAC
advertise that is supports pause.

Results with this patch.

Legend:
- DT = 'Pause' is set in the fixed-link in devicetree.
- validate() = ‘Yes’ means phylink_set(mask, Pause) is set in the
  validate().
- flow = results reported my link is Up line.

+-----+------------+-------+
| DT  | validate() | flow  |
+-----+------------+-------+
| Yes | Yes        | rx/tx |
| No  | Yes        | off   |
| Yes | No         | off   |
+-----+------------+-------+

Fixes: 9525ae83959b ("phylink: add phylink infrastructure")
Signed-off-by: René van Dorst <opensource@vdorst.com>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agogigaset: stop maintaining seperately
Paul Bolle [Fri, 26 Jul 2019 22:05:41 +0000 (00:05 +0200)]
gigaset: stop maintaining seperately

The Dutch consumer grade ISDN network will be shut down on September 1,
2019. This means I'll be converted to some sort of VOIP shortly. At that
point it would be unwise to try to maintain the gigaset driver, even for
odd fixes as I do. So I'll stop maintaining it as a seperate driver and
bump support to CAPI in staging. De facto this means the driver will be
unmaintained, since no-one seems to be working on CAPI.

I've lighty tested the hardware specific modules of this driver (bas-gigaset,
ser-gigaset, and usb-gigaset) for v5.3-rc1. The basic functionality appears to
be working. It's unclear whether anyone still cares. I'm aware of only one
person sort of using the driver a few years ago.

Thanks to Karsten Keil for the ISDN subsystems gigaset was using (I4L and
CAPI). And many thanks to Hansjoerg Lipp and Tilman Schmidt for writing and
upstreaming this driver.

Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: rds: Fix possible null-pointer dereferences in rds_rdma_cm_event_handler_cmn()
Jia-Ju Bai [Fri, 26 Jul 2019 14:17:05 +0000 (22:17 +0800)]
net: rds: Fix possible null-pointer dereferences in rds_rdma_cm_event_handler_cmn()

In rds_rdma_cm_event_handler_cmn(), there are some if statements to
check whether conn is NULL, such as on lines 65, 96 and 112.
But conn is not checked before being used on line 108:
    trans->cm_connect_complete(conn, event);
and on lines 140-143:
    rdsdebug("DISCONNECT event - dropping connection "
            "%pI6c->%pI6c\n", &conn->c_laddr,
            &conn->c_faddr);
    rds_conn_drop(conn);

Thus, possible null-pointer dereferences may occur.

To fix these bugs, conn is checked before being used.

These bugs are found by a static analysis tool STCheck written by us.

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoisdn: mISDN: hfcsusb: Fix possible null-pointer dereferences in start_isoc_chain()
Jia-Ju Bai [Fri, 26 Jul 2019 08:27:36 +0000 (16:27 +0800)]
isdn: mISDN: hfcsusb: Fix possible null-pointer dereferences in start_isoc_chain()

In start_isoc_chain(), usb_alloc_urb() on line 1392 may fail
and return NULL. At this time, fifo->iso[i].urb is assigned to NULL.

Then, fifo->iso[i].urb is used at some places, such as:
LINE 1405:    fill_isoc_urb(fifo->iso[i].urb, ...)
                  urb->number_of_packets = num_packets;
                  urb->transfer_flags = URB_ISO_ASAP;
                  urb->actual_length = 0;
                  urb->interval = interval;
LINE 1416:    fifo->iso[i].urb->...
LINE 1419:    fifo->iso[i].urb->...

Thus, possible null-pointer dereferences may occur.

To fix these bugs, "continue" is added to avoid using fifo->iso[i].urb
when it is NULL.

These bugs are found by a static analysis tool STCheck written by us.

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'mlx5-fixes-2019-07-25' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Fri, 26 Jul 2019 21:26:41 +0000 (14:26 -0700)]
Merge tag 'mlx5-fixes-2019-07-25' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
Mellanox, mlx5 fixes 2019-07-25

This series introduces some fixes to mlx5 driver.

1) Ariel is addressing an issue with enacp flow counter race condition
2) Aya fixes ethtool speed handling
3) Edward fixes modify_cq hw bits alignment
4) Maor fixes RDMA_RX capabilities handling
5) Mark reverses unregister devices order to address an issue with LAG
6) From Tariq,
  - wrong max num channels indication regression
  - TLS counters naming and documentation as suggested by Jakub
  - kTLS, Call WARN_ONCE on netdev mismatch

There is one patch in this series that touches nfp driver to align
TLS statistics names with latest documentation, Jakub is CC'ed.

Please pull and let me know if there is any problem.

For -stable v4.9:
  ('net/mlx5: Use reversed order when unregister devices')

For -stable v4.20
  ('net/mlx5e: Prevent encap flow counter update async to user query')
  ('net/mlx5: Fix modify_cq_in alignment')

For -stable v5.1
  ('net/mlx5e: Fix matching of speed to PRM link modes')

For -stable v5.2
  ('net/mlx5: Add missing RDMA_RX capabilities')
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: qualcomm: rmnet: Fix incorrect UL checksum offload logic
Subash Abhinov Kasiviswanathan [Thu, 25 Jul 2019 18:07:12 +0000 (12:07 -0600)]
net: qualcomm: rmnet: Fix incorrect UL checksum offload logic

The udp_ip4_ind bit is set only for IPv4 UDP non-fragmented packets
so that the hardware can flip the checksum to 0xFFFF if the computed
checksum is 0 per RFC768.

However, this bit had to be set for IPv6 UDP non fragmented packets
as well per hardware requirements. Otherwise, IPv6 UDP packets
with computed checksum as 0 were transmitted by hardware and were
dropped in the network.

In addition to setting this bit for IPv6 UDP, the field is also
appropriately renamed to udp_ind as part of this change.

Fixes: 5eb5f8608ef1 ("net: qualcomm: rmnet: Add support for TX checksum offload")
Cc: Sean Tranchetti <stranche@codeaurora.org>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoip6_tunnel: fix possible use-after-free on xmit
Haishuang Yan [Thu, 25 Jul 2019 16:40:17 +0000 (00:40 +0800)]
ip6_tunnel: fix possible use-after-free on xmit

ip4ip6/ip6ip6 tunnels run iptunnel_handle_offloads on xmit which
can cause a possible use-after-free accessing iph/ipv6h pointer
since the packet will be 'uncloned' running pskb_expand_head if
it is a cloned gso skb.

Fixes: 0e9a709560db ("ip6_tunnel, ip6_gre: fix setting of DSCP on encapsulated packets")
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoocelot: Cancel delayed work before wq destruction
Claudiu Manoil [Thu, 25 Jul 2019 13:33:18 +0000 (16:33 +0300)]
ocelot: Cancel delayed work before wq destruction

Make sure the delayed work for stats update is not pending before
wq destruction.
This fixes the module unload path.
The issue is there since day 1.

Fixes: a556c76adc05 ("net: mscc: Add initial Ocelot switch support")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoqed: RDMA - Fix the hw_ver returned in device attributes
Michal Kalderon [Thu, 25 Jul 2019 10:59:55 +0000 (13:59 +0300)]
qed: RDMA - Fix the hw_ver returned in device attributes

The hw_ver field was initialized to zero. Return the chip revision.
This is relevant for rdma driver.

Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: usb: qmi_wwan: Add the BroadMobi BM818 card
Bob Ham [Wed, 24 Jul 2019 14:52:27 +0000 (07:52 -0700)]
net: usb: qmi_wwan: Add the BroadMobi BM818 card

The BroadMobi BM818 M.2 card uses the QMI protocol

Signed-off-by: Bob Ham <bob.ham@puri.sm>
Signed-off-by: Angus Ainslie (Purism) <angus@akkea.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
David S. Miller [Fri, 26 Jul 2019 00:35:03 +0000 (17:35 -0700)]
Merge git://git./pub/scm/linux/kernel/git/bpf/bpf

Alexei Starovoitov says:

====================
pull-request: bpf 2019-07-25

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

The main changes are:

1) fix segfault in libbpf, from Andrii.

2) fix gso_segs access, from Eric.

3) tls/sockmap fixes, from Jakub and John.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipip: validate header length in ipip_tunnel_xmit
Haishuang Yan [Thu, 25 Jul 2019 03:07:56 +0000 (11:07 +0800)]
ipip: validate header length in ipip_tunnel_xmit

We need the same checks introduced by commit cb9f1b783850
("ip: validate header length on virtual device xmit") for
ipip tunnel.

Fixes: cb9f1b783850b ("ip: validate header length on virtual device xmit")
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoselftests/net: add missing gitignores (ipv6_flowlabel)
Jakub Kicinski [Thu, 25 Jul 2019 00:07:14 +0000 (17:07 -0700)]
selftests/net: add missing gitignores (ipv6_flowlabel)

ipv6_flowlabel and ipv6_flowlabel_mgr are missing from
gitignore.  Quentin points out that the original
commit 3fb321fde22d ("selftests/net: ipv6 flowlabel")
did add ignore entries, they are just missing the "ipv6_"
prefix.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agobnx2x: Disable multi-cos feature.
Sudarsana Reddy Kalluru [Wed, 24 Jul 2019 02:32:41 +0000 (19:32 -0700)]
bnx2x: Disable multi-cos feature.

Commit 3968d38917eb ("bnx2x: Fix Multi-Cos.") which enabled multi-cos
feature after prolonged time in driver added some regression causing
numerous issues (sudden reboots, tx timeout etc.) reported by customers.
We plan to backout this commit and submit proper fix once we have root
cause of issues reported with this feature enabled.

Fixes: 3968d38917eb ("bnx2x: Fix Multi-Cos.")
Signed-off-by: Sudarsana Reddy Kalluru <skalluru@marvell.com>
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoDocumentation: TLS: fix stat counters description
Tariq Toukan [Mon, 22 Jul 2019 10:43:03 +0000 (13:43 +0300)]
Documentation: TLS: fix stat counters description

Add missing description of counters.
Split tx_tls_encrypted counter into two, to give packets
and bytes indications.

Fixes: f42c104f2ec9 ("Documentation: add TLS offload documentation")
Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonfp: tls: rename tls packet counters
Tariq Toukan [Thu, 18 Jul 2019 13:34:05 +0000 (16:34 +0300)]
nfp: tls: rename tls packet counters

Align to the naming convention in TLS documentation.

Fixes: 51a5e563298d ("nfp: tls: add basic statistics")
Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: kTLS, Call WARN_ONCE on netdev mismatch
Tariq Toukan [Thu, 18 Jul 2019 13:32:31 +0000 (16:32 +0300)]
net/mlx5e: kTLS, Call WARN_ONCE on netdev mismatch

A netdev mismatch in the processed TLS SKB should not occur,
and indicates a kernel bug.
Add WARN_ONCE to spot such cases.

Fixes: d2ead1f360e8 ("net/mlx5e: Add kTLS TX HW offload support")
Suggested-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Prevent encap flow counter update async to user query
Ariel Levkovich [Sat, 6 Jul 2019 15:06:15 +0000 (18:06 +0300)]
net/mlx5e: Prevent encap flow counter update async to user query

This patch prevents a race between user invoked cached counters
query and a neighbor last usage updater.

The cached flow counter stats can be queried by calling
"mlx5_fc_query_cached" which provides the number of bytes and
packets that passed via this flow since the last time this counter
was queried.
It does so by reducting the last saved stats from the current, cached
stats and then updating the last saved stats with the cached stats.
It also provide the lastuse value for that flow.

Since "mlx5e_tc_update_neigh_used_value" needs to retrieve the
last usage time of encapsulation flows, it calls the flow counter
query method periodically and async to user queries of the flow counter
using cls_flower.
This call is causing the driver to update the last reported bytes and
packets from the cache and therefore, future user queries of the flow
stats will return lower than expected number for bytes and packets
since the last saved stats in the driver was updated async to the last
saved stats in cls_flower.

This causes wrong stats presentation of encapsulation flows to user.

Since the neighbor usage updater only needs the lastuse stats from the
cached counter, the fix is to use a dedicated lastuse query call that
returns the lastuse value without synching between the cached stats and
the last saved stats.

Fixes: f6dfb4c3f216 ("net/mlx5e: Update neighbour 'used' state using HW flow rules counters")
Signed-off-by: Ariel Levkovich <lariel@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Fix matching of speed to PRM link modes
Aya Levin [Sun, 16 Jun 2019 10:20:29 +0000 (13:20 +0300)]
net/mlx5e: Fix matching of speed to PRM link modes

Speed translation is performed based on legacy or extended PTYS
register. Translate speed with respect to:
1) Capability bit of extended PTYS table.
2) User request:
 a) When auto-negotiation is turned on, inspect advertisement whether it
 contains extended link modes.
 b) When auto-negotiation is turned off, speed > 100Gbps (maximal
 speed supported in legacy mode).
With both conditions fulfilled translation is done with extended PTYS
table otherwise use legacy PTYS table.
Without this patch 25/50/100 Gbps speed cannot be set, since try to
configure in extended mode but read from legacy mode.

Fixes: dd1b9e09c12b ("net/mlx5: ethtool, Allow legacy link-modes configuration via non-extended ptys")
Signed-off-by: Aya Levin <ayal@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Fix wrong max num channels indication
Tariq Toukan [Sun, 14 Jul 2019 08:43:43 +0000 (11:43 +0300)]
net/mlx5e: Fix wrong max num channels indication

No XSK support in the enhanced IPoIB driver and representors.
Add a profile property to specify this, and enhance the logic
that calculates the max number of channels to take it into
account.

Fixes: db05815b36cb ("net/mlx5e: Add XSK zero-copy support")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5: Fix modify_cq_in alignment
Edward Srouji [Tue, 23 Jul 2019 07:12:55 +0000 (10:12 +0300)]
net/mlx5: Fix modify_cq_in alignment

Fix modify_cq_in alignment to match the device specification.
After this fix the 'cq_umem_valid' field will be in the right offset.

Cc: <stable@vger.kernel.org> # 4.19
Fixes: bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits")
Signed-off-by: Edward Srouji <edwards@mellanox.com>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5: Add missing RDMA_RX capabilities
Maor Gottlieb [Sun, 14 Jul 2019 08:33:07 +0000 (11:33 +0300)]
net/mlx5: Add missing RDMA_RX capabilities

New flow table type RDMA_RX was added but the MLX5_CAP_FLOW_TABLE_TYPE
didn't handle this new flow table type.
This means that MLX5_CAP_FLOW_TABLE_TYPE returns an empty capability to
this flow table type.

Update both the macro and the maximum supported flow table type to
RDMA_RX.

Fixes: d83eb50e29de ("net/mlx5: Add support in RDMA RX steering")
Signed-off-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5: Use reversed order when unregister devices
Mark Zhang [Tue, 9 Jul 2019 02:37:12 +0000 (05:37 +0300)]
net/mlx5: Use reversed order when unregister devices

When lag is active, which is controlled by the bonded mlx5e netdev, mlx5
interface unregestering must happen in the reverse order where rdma is
unregistered (unloaded) first, to guarantee all references to the lag
context in hardware is removed, then remove mlx5e netdev interface which
will cleanup the lag context from hardware.

Without this fix during destroy of LAG interface, we observed following
errors:
 * mlx5_cmd_check:752:(pid 12556): DESTROY_LAG(0x843) op_mod(0x0) failed,
   status bad parameter(0x3), syndrome (0xe4ac33)
 * mlx5_cmd_check:752:(pid 12556): DESTROY_LAG(0x843) op_mod(0x0) failed,
   status bad parameter(0x3), syndrome (0xa5aee8).

Fixes: a31208b1e11d ("net/mlx5_core: New init and exit flow for mlx5_core")
Reviewed-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Mark Zhang <markz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agofsl/fman: Remove comment referring to non-existent function
Chris Packham [Tue, 23 Jul 2019 23:35:01 +0000 (11:35 +1200)]
fsl/fman: Remove comment referring to non-existent function

fm_set_max_frm() existed in the Freescale SDK as a callback for an
early_param. When this code was ported to the upstream kernel the
early_param was converted to a module_param making the reference to the
function incorrect. The rest of the comment already does a good job of
explaining the parameter so removing the reference to the non-existent
function seems like the best thing to do.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agost_nci_hci_connectivity_event_received: null check the allocation
Navid Emamdoost [Tue, 23 Jul 2019 22:11:51 +0000 (17:11 -0500)]
st_nci_hci_connectivity_event_received: null check the allocation

devm_kzalloc may fail and return NULL. So the null check is needed.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agost21nfca_connectivity_event_received: null check the allocation
Navid Emamdoost [Tue, 23 Jul 2019 22:04:30 +0000 (17:04 -0500)]
st21nfca_connectivity_event_received: null check the allocation

devm_kzalloc may fail and return null. So the null check is needed.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotun: mark small packets as owned by the tap sock
Alexis Bauvin [Tue, 23 Jul 2019 14:23:01 +0000 (16:23 +0200)]
tun: mark small packets as owned by the tap sock

- v1 -> v2: Move skb_set_owner_w to __tun_build_skb to reduce patch size

Small packets going out of a tap device go through an optimized code
path that uses build_skb() rather than sock_alloc_send_pskb(). The
latter calls skb_set_owner_w(), but the small packet code path does not.

The net effect is that small packets are not owned by the userland
application's socket (e.g. QEMU), while large packets are.
This can be seen with a TCP session, where packets are not owned when
the window size is small enough (around PAGE_SIZE), while they are once
the window grows (note that this requires the host to support virtio
tso for the guest to offload segmentation).
All this leads to inconsistent behaviour in the kernel, especially on
netfilter modules that uses sk->socket (e.g. xt_owner).

Fixes: 66ccbc9c87c2 ("tap: use build_skb() for small packet")
Signed-off-by: Alexis Bauvin <abauvin@scaleway.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'dim-fixes'
David S. Miller [Thu, 25 Jul 2019 18:34:39 +0000 (11:34 -0700)]
Merge branch 'dim-fixes'

Leon Romanovsky says:

====================
DIM fixes for 5.3

Those two fixes for recently merged DIM patches, both exposed through
RDMa DIM usage.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolib/dim: Fix -Wunused-const-variable warnings
Leon Romanovsky [Tue, 23 Jul 2019 07:22:48 +0000 (10:22 +0300)]
lib/dim: Fix -Wunused-const-variable warnings

DIM causes to the following warnings during kernel compilation
which indicates that tx_profile and rx_profile are supposed to
be declared in *.c and not in *.h files.

In file included from ./include/rdma/ib_verbs.h:64,
                 from ./include/linux/mlx5/device.h:37,
                 from ./include/linux/mlx5/driver.h:51,
                 from ./include/linux/mlx5/vport.h:36,
                 from drivers/infiniband/hw/mlx5/ib_virt.c:34:
./include/linux/dim.h:326:1: warning: _tx_profile_ defined but not used [-Wunused-const-variable=]
  326 | tx_profile[DIM_CQ_PERIOD_NUM_MODES][NET_DIM_PARAMS_NUM_PROFILES] = {
      | ^~~~~~~~~~
./include/linux/dim.h:320:1: warning: _rx_profile_ defined but not used [-Wunused-const-variable=]
  320 | rx_profile[DIM_CQ_PERIOD_NUM_MODES][NET_DIM_PARAMS_NUM_PROFILES] = {
      | ^~~~~~~~~~

Fixes: 4f75da3666c0 ("linux/dim: Move implementation to .c files")
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolinux/dim: Fix overflow in dim calculation
Yamin Friedman [Tue, 23 Jul 2019 07:22:47 +0000 (10:22 +0300)]
linux/dim: Fix overflow in dim calculation

While using net_dim, a dim_sample was used without ever initializing the
comps value. Added use of DIV_ROUND_DOWN_ULL() to prevent potential
overflow, it should not be a problem to save the final result in an int
because after the division by epms the value should not be larger than a
few thousand.

[ 1040.127124] UBSAN: Undefined behaviour in lib/dim/dim.c:78:23
[ 1040.130118] signed integer overflow:
[ 1040.131643] 134718714 * 100 cannot be represented in type 'int'

Fixes: 398c2b05bbee ("linux/dim: Add completions count to dim_sample")
Signed-off-by: Yamin Friedman <yaminf@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoife: error out when nla attributes are empty
Cong Wang [Tue, 23 Jul 2019 04:43:00 +0000 (21:43 -0700)]
ife: error out when nla attributes are empty

act_ife at least requires TCA_IFE_PARMS, so we have to bail out
when there is no attribute passed in.

Reported-by: syzbot+fbb5b288c9cb6a2eeac4@syzkaller.appspotmail.com
Fixes: ef6980b6becb ("introduce IFE action")
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolibbpf: silence GCC8 warning about string truncation
Andrii Nakryiko [Wed, 24 Jul 2019 21:47:53 +0000 (14:47 -0700)]
libbpf: silence GCC8 warning about string truncation

Despite a proper NULL-termination after strncpy(..., ..., IFNAMSIZ - 1),
GCC8 still complains about *expected* string truncation:

  xsk.c:330:2: error: 'strncpy' output may be truncated copying 15 bytes
  from a string of length 15 [-Werror=stringop-truncation]
    strncpy(ifr.ifr_name, xsk->ifname, IFNAMSIZ - 1);

This patch gets rid of the issue altogether by using memcpy instead.
There is no performance regression, as strncpy will still copy and fill
all of the bytes anyway.

v1->v2:
- rebase against bpf tree.

Cc: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Acked-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agonetrom: hold sock when setting skb->destructor
Cong Wang [Tue, 23 Jul 2019 03:41:22 +0000 (20:41 -0700)]
netrom: hold sock when setting skb->destructor

sock_efree() releases the sock refcnt, if we don't hold this refcnt
when setting skb->destructor to it, the refcnt would not be balanced.
This leads to several bug reports from syzbot.

I have checked other users of sock_efree(), all of them hold the
sock refcnt.

Fixes: c8c8218ec5af ("netrom: fix a memory leak in nr_rx_frame()")
Reported-and-tested-by: <syzbot+622bdabb128acc33427d@syzkaller.appspotmail.com>
Reported-and-tested-by: <syzbot+6eaef7158b19e3fec3a0@syzkaller.appspotmail.com>
Reported-and-tested-by: <syzbot+9399c158fcc09b21d0d2@syzkaller.appspotmail.com>
Reported-and-tested-by: <syzbot+a34e5f3d0300163f0c87@syzkaller.appspotmail.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoovs: datapath: hide clang frame-overflow warnings
Arnd Bergmann [Mon, 22 Jul 2019 15:00:01 +0000 (17:00 +0200)]
ovs: datapath: hide clang frame-overflow warnings

Some functions in the datapath code are factored out so that each
one has a stack frame smaller than 1024 bytes with gcc. However,
when compiling with clang, the functions are inlined more aggressively
and combined again so we get

net/openvswitch/datapath.c:1124:12: error: stack frame size of 1528 bytes in function 'ovs_flow_cmd_set' [-Werror,-Wframe-larger-than=]

Marking both get_flow_actions() and ovs_nla_init_match_and_action()
as 'noinline_for_stack' gives us the same behavior that we see with
gcc, and no warning. Note that this does not mean we actually use
less stack, as the functions call each other, and we still get
three copies of the large 'struct sw_flow_key' type on the stack.

The comment tells us that this was previously considered safe,
presumably since the netlink parsing functions are called with
a known backchain that does not also use a lot of stack space.

Fixes: 9cc9a5cb176c ("datapath: Avoid using stack larger than 1024.")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet/tls: add myself as a co-maintainer
Jakub Kicinski [Wed, 24 Jul 2019 18:02:48 +0000 (11:02 -0700)]
net/tls: add myself as a co-maintainer

I've been spending quite a bit of time fixing and
preventing bit rot in the core TLS code. TLS seems
to only be growing in importance, I'd like to help
ensuring the quality of our implementation.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: phy: mscc: initialize stats array
Andreas Schwab [Wed, 24 Jul 2019 15:32:57 +0000 (17:32 +0200)]
net: phy: mscc: initialize stats array

The memory allocated for the stats array may contain arbitrary data.

Fixes: e4f9ba642f0b ("net: phy: mscc: add support for VSC8514 PHY.")
Fixes: 00d70d8e0e78 ("net: phy: mscc: add support for VSC8574 PHY")
Fixes: a5afc1678044 ("net: phy: mscc: add support for VSC8584 PHY")
Fixes: f76178dc5218 ("net: phy: mscc: add ethtool statistics counters")
Signed-off-by: Andreas Schwab <schwab@suse.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: phylink: don't start and stop SGMII PHYs in SFP modules twice
Arseny Solokha [Wed, 24 Jul 2019 13:31:39 +0000 (20:31 +0700)]
net: phylink: don't start and stop SGMII PHYs in SFP modules twice

SFP modules connected using the SGMII interface have their own PHYs which
are handled by the struct phylink's phydev field. On the other hand, for
the modules connected using 1000Base-X interface that field is not set.

Since commit ce0aa27ff3f6 ("sfp: add sfp-bus to bridge between network
devices and sfp cages") phylink_start() ends up setting the phydev field
using the sfp-bus infrastructure, which eventually calls phy_start() on it,
and then calling phy_start() again on the same phydev from phylink_start()
itself. Similar call sequence holds for phylink_stop(), only in the reverse
order. This results in WARNs during network interface bringup and shutdown
when a copper SFP module is connected, as phy_start() and phy_stop() are
called twice in a row for the same phy_device:

  % ip link set up dev eth0
  ------------[ cut here ]------------
  called from state UP
  WARNING: CPU: 1 PID: 155 at drivers/net/phy/phy.c:895 phy_start+0x74/0xc0
  Modules linked in:
  CPU: 1 PID: 155 Comm: backend Not tainted 5.2.0+ #1
  NIP:  c0227bf0 LR: c0227bf0 CTR: c004d224
  REGS: df547720 TRAP: 0700   Not tainted  (5.2.0+)
  MSR:  00029000 <CE,EE,ME>  CR: 24002822  XER: 00000000

  GPR00: c0227bf0 df5477d8 df5d7080 00000014 df9d2370 df9d5ac4 1f4eb000 00000001
  GPR08: c061fe58 00000000 00000000 df5477d8 0000003c 100c8768 00000000 00000000
  GPR16: df486a00 c046f1c8 c046eea0 00000000 c046e904 c0239604 db68449c 00000000
  GPR24: e9083204 00000000 00000001 db684460 e9083404 00000000 db6dce00 db6dcc00
  NIP [c0227bf0] phy_start+0x74/0xc0
  LR [c0227bf0] phy_start+0x74/0xc0
  Call Trace:
  [df5477d8] [c0227bf0] phy_start+0x74/0xc0 (unreliable)
  [df5477e8] [c023cad0] startup_gfar+0x398/0x3f4
  [df547828] [c023cf08] gfar_enet_open+0x364/0x374
  [df547898] [c029d870] __dev_open+0xe4/0x140
  [df5478c8] [c029db70] __dev_change_flags+0xf0/0x188
  [df5478f8] [c029dc28] dev_change_flags+0x20/0x54
  [df547918] [c02ae304] do_setlink+0x310/0x818
  [df547a08] [c02b1eb8] __rtnl_newlink+0x384/0x6b0
  [df547c28] [c02b222c] rtnl_newlink+0x48/0x68
  [df547c48] [c02ad7c8] rtnetlink_rcv_msg+0x240/0x27c
  [df547c98] [c02cc068] netlink_rcv_skb+0x8c/0xf0
  [df547cd8] [c02cba3c] netlink_unicast+0x114/0x19c
  [df547d08] [c02cbd74] netlink_sendmsg+0x2b0/0x2c0
  [df547d58] [c027b668] sock_sendmsg_nosec+0x20/0x40
  [df547d68] [c027d080] ___sys_sendmsg+0x17c/0x1dc
  [df547e98] [c027df7c] __sys_sendmsg+0x68/0x84
  [df547ef8] [c027e430] sys_socketcall+0x1a0/0x204
  [df547f38] [c000d1d8] ret_from_syscall+0x0/0x38
  --- interrupt: c01 at 0xfd4e030
      LR = 0xfd4e010
  Instruction dump:
  813f0188 38800000 2b890005 419d0014 3d40c046 5529103a 394aa208 7c8a482e
  3c60c046 3863a1b8 4cc63182 4be009a1 <0fe0000048000030 3c60c046 3863a1d0
  ---[ end trace d4c095aeaf6ea998 ]---

and

  % ip link set down dev eth0
  ------------[ cut here ]------------
  called from state HALTED
  WARNING: CPU: 1 PID: 184 at drivers/net/phy/phy.c:858 phy_stop+0x3c/0x88

  <...>

  Call Trace:
  [df581788] [c0228450] phy_stop+0x3c/0x88 (unreliable)
  [df581798] [c022d548] sfp_sm_phy_detach+0x1c/0x44
  [df5817a8] [c022e8cc] sfp_sm_event+0x4b0/0x87c
  [df581848] [c022f04c] sfp_upstream_stop+0x34/0x44
  [df581858] [c0225608] phylink_stop+0x7c/0xe4
  [df581868] [c023c57c] stop_gfar+0x7c/0x94
  [df581888] [c023c5b8] gfar_close+0x24/0x94
  [df5818a8] [c0298688] __dev_close_many+0xdc/0xf8
  [df5818c8] [c029db58] __dev_change_flags+0xd8/0x188
  [df5818f8] [c029dc28] dev_change_flags+0x20/0x54
  [df581918] [c02ae304] do_setlink+0x310/0x818
  [df581a08] [c02b1eb8] __rtnl_newlink+0x384/0x6b0
  [df581c28] [c02b222c] rtnl_newlink+0x48/0x68
  [df581c48] [c02ad7c8] rtnetlink_rcv_msg+0x240/0x27c
  [df581c98] [c02cc068] netlink_rcv_skb+0x8c/0xf0
  [df581cd8] [c02cba3c] netlink_unicast+0x114/0x19c
  [df581d08] [c02cbd74] netlink_sendmsg+0x2b0/0x2c0
  [df581d58] [c027b668] sock_sendmsg_nosec+0x20/0x40
  [df581d68] [c027d080] ___sys_sendmsg+0x17c/0x1dc
  [df581e98] [c027df7c] __sys_sendmsg+0x68/0x84
  [df581ef8] [c027e430] sys_socketcall+0x1a0/0x204
  [df581f38] [c000d1d8] ret_from_syscall+0x0/0x38

  <...>

  ---[ end trace d4c095aeaf6ea999 ]---

SFP modules with the 1000Base-X interface are not affected.

Place explicit calls to phy_start() and phy_stop() before enabling or after
disabling an attached SFP module, where phydev is not yet set (or is
already unset), so they will be made only from the inside of sfp-bus, if
needed.

Fixes: 217962615662 ("net: phy: warn if phy_start is called from invalid state")
Signed-off-by: Arseny Solokha <asolokha@kb.kras.ru>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'linux-can-fixes-for-5.3-20190724' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Wed, 24 Jul 2019 21:14:50 +0000 (14:14 -0700)]
Merge tag 'linux-can-fixes-for-5.3-20190724' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2019-07-24

this is a pull reqeust of 7 patches for net/master.

The first patch is by Rasmus Villemoes add a missing netif_carrier_off() to
register_candev() so that generic netdev trigger based LEDs are initially off.

Nikita Yushchenko's patch for the rcar_canfd driver fixes a possible IRQ storm
on high load.

The patch by Weitao Hou for the mcp251x driver add missing error checking to
the work queue allocation.

Both Wen Yang's and Joakim Zhang's patch for the flexcan driver fix a problem
with the stop-mode.

Stephane Grosjean contributes a patch for the peak_usb driver to fix a
potential double kfree_skb().

The last patch is by YueHaibing and fixes the error path in can-gw's
cgw_module_init() function.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoip6_gre: reload ipv6h in prepare_ip6gre_xmit_ipv6
Haishuang Yan [Wed, 24 Jul 2019 12:00:42 +0000 (20:00 +0800)]
ip6_gre: reload ipv6h in prepare_ip6gre_xmit_ipv6

Since ip6_tnl_parse_tlv_enc_lim() can call pskb_may_pull()
which may change skb->data, so we need to re-load ipv6h at
the right place.

Fixes: 898b29798e36 ("ip6_gre: Refactor ip6gre xmit codes")
Cc: William Tu <u9012063@gmail.com>
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Acked-by: William Tu <u9012063@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet/ipv4: cleanup error condition testing
Pavel Machek [Wed, 24 Jul 2019 20:56:37 +0000 (13:56 -0700)]
net/ipv4: cleanup error condition testing

Cleanup testing for error condition.

Signed-off-by: Pavel Machek <pavel@denx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agocan: gw: Fix error path of cgw_module_init
YueHaibing [Sat, 18 May 2019 09:35:43 +0000 (17:35 +0800)]
can: gw: Fix error path of cgw_module_init

This patch add error path for cgw_module_init to avoid possible crash if
some error occurs.

Fixes: c1aabdf379bc ("can-gw: add netlink based CAN routing")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: peak_usb: fix potential double kfree_skb()
Stephane Grosjean [Fri, 5 Jul 2019 13:32:16 +0000 (15:32 +0200)]
can: peak_usb: fix potential double kfree_skb()

When closing the CAN device while tx skbs are inflight, echo skb could
be released twice. By calling close_candev() before unlinking all
pending tx urbs, then the internal echo_skb[] array is fully and
correctly cleared before the USB write callback and, therefore,
can_get_echo_skb() are called, for each aborted URB.

Fixes: bb4785551f64 ("can: usb: PEAK-System Technik USB adapters driver core")
Signed-off-by: Stephane Grosjean <s.grosjean@peak-system.com>
Cc: linux-stable <stable@vger.kernel.org>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: flexcan: fix stop mode acknowledgment
Joakim Zhang [Tue, 2 Jul 2019 01:45:41 +0000 (01:45 +0000)]
can: flexcan: fix stop mode acknowledgment

To enter stop mode, the CPU should manually assert a global Stop Mode
request and check the acknowledgment asserted by FlexCAN. The CPU must
only consider the FlexCAN in stop mode when both request and
acknowledgment conditions are satisfied.

Fixes: de3578c198c6 ("can: flexcan: add self wakeup support")
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Joakim Zhang <qiangqing.zhang@nxp.com>
Cc: linux-stable <stable@vger.kernel.org> # >= v5.0
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: flexcan: fix an use-after-free in flexcan_setup_stop_mode()
Wen Yang [Sat, 6 Jul 2019 03:37:20 +0000 (11:37 +0800)]
can: flexcan: fix an use-after-free in flexcan_setup_stop_mode()

The gpr_np variable is still being used in dev_dbg() after the
of_node_put() call, which may result in use-after-free.

Fixes: de3578c198c6 ("can: flexcan: add self wakeup support")
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Cc: linux-stable <stable@vger.kernel.org> # >= v5.0
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: mcp251x: add error check when wq alloc failed
Weitao Hou [Tue, 25 Jun 2019 12:50:48 +0000 (20:50 +0800)]
can: mcp251x: add error check when wq alloc failed

add error check when workqueue alloc failed, and remove redundant code
to make it clear.

Fixes: e0000163e30e ("can: Driver for the Microchip MCP251x SPI CAN controllers")
Signed-off-by: Weitao Hou <houweitaoo@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Tested-by: Sean Nyekjaer <sean@geanix.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: rcar_canfd: fix possible IRQ storm on high load
Nikita Yushchenko [Wed, 26 Jun 2019 13:08:48 +0000 (16:08 +0300)]
can: rcar_canfd: fix possible IRQ storm on high load

We have observed rcar_canfd driver entering IRQ storm under high load,
with following scenario:
- rcar_canfd_global_interrupt() in entered due to Rx available,
- napi_schedule_prep() is called, and sets NAPIF_STATE_SCHED in state
- Rx fifo interrupts are masked,
- rcar_canfd_global_interrupt() is entered again, this time due to
  error interrupt (e.g. due to overflow),
- since scheduled napi poller has not yet executed, condition for calling
  napi_schedule_prep() from rcar_canfd_global_interrupt() remains true,
  thus napi_schedule_prep() gets called and sets NAPIF_STATE_MISSED flag
  in state,
- later, napi poller function rcar_canfd_rx_poll() gets executed, and
  calls napi_complete_done(),
- due to NAPIF_STATE_MISSED flag in state, this call does not clear
  NAPIF_STATE_SCHED flag from state,
- on return from napi_complete_done(), rcar_canfd_rx_poll() unmasks Rx
  interrutps,
- Rx interrupt happens, rcar_canfd_global_interrupt() gets called
  and calls napi_schedule_prep(),
- since NAPIF_STATE_SCHED is set in state at this time, this call
  returns false,
- due to that false return, rcar_canfd_global_interrupt() returns
  without masking Rx interrupt
- and this results into IRQ storm: unmasked Rx interrupt happens again
  and again is misprocessed in the same way.

This patch fixes that scenario by unmasking Rx interrupts only when
napi_complete_done() returns true, which means it has cleared
NAPIF_STATE_SCHED in state.

Fixes: dd3bd23eb438 ("can: rcar_canfd: Add Renesas R-Car CAN FD driver")
Signed-off-by: Nikita Yushchenko <nikita.yoush@cogentembedded.com>
Cc: linux-stable <stable@vger.kernel.org>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agocan: dev: call netif_carrier_off() in register_candev()
Rasmus Villemoes [Mon, 24 Jun 2019 08:34:13 +0000 (08:34 +0000)]
can: dev: call netif_carrier_off() in register_candev()

CONFIG_CAN_LEDS is deprecated. When trying to use the generic netdev
trigger as suggested, there's a small inconsistency with the link
property: The LED is on initially, stays on when the device is brought
up, and then turns off (as expected) when the device is brought down.

Make sure the LED always reflects the state of the CAN device.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
5 years agolibbpf: fix using uninitialized ioctl results
Ilya Maximets [Tue, 23 Jul 2019 12:08:10 +0000 (15:08 +0300)]
libbpf: fix using uninitialized ioctl results

'channels.max_combined' initialized only on ioctl success and
errno is only valid on ioctl failure.

The code doesn't produce any runtime issues, but makes memory
sanitizers angry:

 Conditional jump or move depends on uninitialised value(s)
    at 0x55C056F: xsk_get_max_queues (xsk.c:336)
    by 0x55C05B2: xsk_create_bpf_maps (xsk.c:354)
    by 0x55C089F: xsk_setup_xdp_prog (xsk.c:447)
    by 0x55C0E57: xsk_socket__create (xsk.c:601)
  Uninitialised value was created by a stack allocation
    at 0x55C04CD: xsk_get_max_queues (xsk.c:318)

Additionally fixed warning on uninitialized bytes in ioctl arguments:

 Syscall param ioctl(SIOCETHTOOL) points to uninitialised byte(s)
    at 0x648D45B: ioctl (in /usr/lib64/libc-2.28.so)
    by 0x55C0546: xsk_get_max_queues (xsk.c:330)
    by 0x55C05B2: xsk_create_bpf_maps (xsk.c:354)
    by 0x55C089F: xsk_setup_xdp_prog (xsk.c:447)
    by 0x55C0E57: xsk_socket__create (xsk.c:601)
  Address 0x1ffefff378 is on thread 1's stack
  in frame #1, created by xsk_get_max_queues (xsk.c:318)
  Uninitialised value was created by a stack allocation
    at 0x55C04CD: xsk_get_max_queues (xsk.c:318)

CC: Magnus Karlsson <magnus.karlsson@intel.com>
Fixes: 1cad07884239 ("libbpf: add support for using AF_XDP sockets")
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agoMerge branch 'fix-gso_segs'
Alexei Starovoitov [Tue, 23 Jul 2019 21:12:38 +0000 (14:12 -0700)]
Merge branch 'fix-gso_segs'

Eric Dumazet says:

====================
First patch changes the kernel, second patch
adds a new test.

Note that other patches might be needed to take
care of similar issues in sock_ops_convert_ctx_access()
and SOCK_OPS_GET_FIELD()
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agoselftests/bpf: add another gso_segs access
Eric Dumazet [Tue, 23 Jul 2019 10:15:38 +0000 (03:15 -0700)]
selftests/bpf: add another gso_segs access

Use BPF_REG_1 for source and destination of gso_segs read,
to exercise "bpf: fix access to skb_shared_info->gso_segs" fix.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agobpf: fix access to skb_shared_info->gso_segs
Eric Dumazet [Tue, 23 Jul 2019 10:15:37 +0000 (03:15 -0700)]
bpf: fix access to skb_shared_info->gso_segs

It is possible we reach bpf_convert_ctx_access() with
si->dst_reg == si->src_reg

Therefore, we need to load BPF_REG_AX before eventually
mangling si->src_reg.

syzbot generated this x86 code :
   3:   55                      push   %rbp
   4:   48 89 e5                mov    %rsp,%rbp
   7:   48 81 ec 00 00 00 00    sub    $0x0,%rsp // Might be avoided ?
   e:   53                      push   %rbx
   f:   41 55                   push   %r13
  11:   41 56                   push   %r14
  13:   41 57                   push   %r15
  15:   6a 00                   pushq  $0x0
  17:   31 c0                   xor    %eax,%eax
  19:   48 8b bf c0 00 00 00    mov    0xc0(%rdi),%rdi
  20:   44 8b 97 bc 00 00 00    mov    0xbc(%rdi),%r10d
  27:   4c 01 d7                add    %r10,%rdi
  2a:   48 0f b7 7f 06          movzwq 0x6(%rdi),%rdi // Crash
  2f:   5b                      pop    %rbx
  30:   41 5f                   pop    %r15
  32:   41 5e                   pop    %r14
  34:   41 5d                   pop    %r13
  36:   5b                      pop    %rbx
  37:   c9                      leaveq
  38:   c3                      retq

Fixes: d9ff286a0f59 ("bpf: allow BPF programs access skb_shared_info->gso_segs field")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agonet: thunderx: Use fwnode_get_mac_address()
Andy Shevchenko [Tue, 23 Jul 2019 20:03:43 +0000 (23:03 +0300)]
net: thunderx: Use fwnode_get_mac_address()

Replace the custom implementation with fwnode_get_mac_address,
which works on both DT and ACPI platforms.

While here, replace memcpy() by ether_addr_copy().

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agobpf: fix narrower loads on s390
Ilya Leoshkevich [Fri, 19 Jul 2019 09:18:15 +0000 (11:18 +0200)]
bpf: fix narrower loads on s390

The very first check in test_pkt_md_access is failing on s390, which
happens because loading a part of a struct __sk_buff field produces
an incorrect result.

The preprocessed code of the check is:

{
__u8 tmp = *((volatile __u8 *)&skb->len +
((sizeof(skb->len) - sizeof(__u8)) / sizeof(__u8)));
if (tmp != ((*(volatile __u32 *)&skb->len) & 0xFF)) return 2;
};

clang generates the following code for it:

      0: 71 21 00 03 00 00 00 00 r2 = *(u8 *)(r1 + 3)
      1: 61 31 00 00 00 00 00 00 r3 = *(u32 *)(r1 + 0)
      2: 57 30 00 00 00 00 00 ff r3 &= 255
      3: 5d 23 00 1d 00 00 00 00 if r2 != r3 goto +29 <LBB0_10>

Finally, verifier transforms it to:

  0: (61) r2 = *(u32 *)(r1 +104)
  1: (bc) w2 = w2
  2: (74) w2 >>= 24
  3: (bc) w2 = w2
  4: (54) w2 &= 255
  5: (bc) w2 = w2

The problem is that when verifier emits the code to replace a partial
load of a struct __sk_buff field (*(u8 *)(r1 + 3)) with a full load of
struct sk_buff field (*(u32 *)(r1 + 104)), an optional shift and a
bitwise AND, it assumes that the machine is little endian and
incorrectly decides to use a shift.

Adjust shift count calculation to account for endianness.

Fixes: 31fd85816dbe ("bpf: permits narrower load from bpf program context fields")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
5 years agosky2: Disable MSI on ASUS P6T
Takashi Iwai [Tue, 23 Jul 2019 15:15:25 +0000 (17:15 +0200)]
sky2: Disable MSI on ASUS P6T

The onboard sky2 NIC on ASUS P6T WS PRO doesn't work after PM resume
due to the infamous IRQ problem.  Disabling MSI works around it, so
let's add it to the blacklist.

Unfortunately the BIOS on the machine doesn't fill the standard
DMI_SYS_* entry, so we pick up DMI_BOARD_* entries instead.

BugLink: https://bugzilla.suse.com/show_bug.cgi?id=1142496
Reported-and-tested-by: Marcus Seyfarth <m.seyfarth@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: dsa: sja1105: sja1105_main: Add of_node_put()
Nishka Dasgupta [Tue, 23 Jul 2019 10:44:48 +0000 (16:14 +0530)]
net: dsa: sja1105: sja1105_main: Add of_node_put()

Each iteration of for_each_child_of_node puts the previous node, but in
the case of a return from the middle of the loop, there is no put, thus
causing a memory leak. Hence add an of_node_put before the return.
Issue found with Coccinelle.

Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: dsa: mv88e6xxx: chip: Add of_node_put() before return
Nishka Dasgupta [Tue, 23 Jul 2019 10:43:07 +0000 (16:13 +0530)]
net: dsa: mv88e6xxx: chip: Add of_node_put() before return

Each iteration of for_each_available_child_of_node puts the previous
node, but in the case of a return from the middle of the loop, there is
no put, thus causing a memory leak. Hence add an of_node_put before the
return.
Issue found with Coccinelle.

Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'selftests-forwarding-GRE-multipath-fixes'
David S. Miller [Tue, 23 Jul 2019 20:06:49 +0000 (13:06 -0700)]
Merge branch 'selftests-forwarding-GRE-multipath-fixes'

Ido Schimmel says:

====================
selftests: forwarding: GRE multipath fixes

Patch #1 ensures IPv4 forwarding is enabled during the test.

Patch #2 fixes the flower filters used to measure the distribution of
the traffic between the two nexthops, so that the test will pass
regardless if traffic is offloaded or not.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoselftests: forwarding: gre_multipath: Fix flower filters
Ido Schimmel [Tue, 23 Jul 2019 08:19:26 +0000 (11:19 +0300)]
selftests: forwarding: gre_multipath: Fix flower filters

The TC filters used in the test do not work with veth devices because the
outer Ethertype is 802.1Q and not IPv4. The test passes with mlxsw
netdevs since the hardware always looks at "The first Ethertype that
does not point to either: VLAN, CNTAG or configurable Ethertype".

Fix this by matching on the VLAN ID instead, but on the ingress side.
The reason why this is not performed at egress is explained in the
commit cited below.

Fixes: 541ad323db3a ("selftests: forwarding: gre_multipath: Update next-hop statistics match criteria")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reported-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Tested-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoselftests: forwarding: gre_multipath: Enable IPv4 forwarding
Ido Schimmel [Tue, 23 Jul 2019 08:19:25 +0000 (11:19 +0300)]
selftests: forwarding: gre_multipath: Enable IPv4 forwarding

The test did not enable IPv4 forwarding during its setup phase, which
causes the test to fail on machines where IPv4 forwarding is disabled.

Fixes: 54818c4c4b93 ("selftests: forwarding: Test multipath tunneling")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reported-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Tested-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoRevert "net: hns: fix LED configuration for marvell phy"
David S. Miller [Tue, 23 Jul 2019 03:44:48 +0000 (20:44 -0700)]
Revert "net: hns: fix LED configuration for marvell phy"

This reverts commit f4e5f775db5a4631300dccd0de5eafb50a77c131.

Andrew Lunn says this should be handled another way.

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: stmmac: Do not cut down 1G modes
Jose Abreu [Mon, 22 Jul 2019 14:07:21 +0000 (16:07 +0200)]
net: stmmac: Do not cut down 1G modes

Some glue logic drivers support 1G without having GMAC/GMAC4/XGMAC.

Let's allow this speed by default.

Reported-by: Ondrej Jirman <megi@xff.cz>
Tested-by: Ondrej Jirman <megi@xff.cz>
Fixes: 5b0d7d7da64b ("net: stmmac: Add the missing speeds that XGMAC supports")
Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'stmmac-fixes'
David S. Miller [Tue, 23 Jul 2019 01:23:32 +0000 (18:23 -0700)]
Merge branch 'stmmac-fixes'

Jose Abreu says:

====================
net: stmmac: Two fixes

Two fixes targeting -net.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: stmmac: Use kcalloc() instead of kmalloc_array()
Jose Abreu [Mon, 22 Jul 2019 08:39:31 +0000 (10:39 +0200)]
net: stmmac: Use kcalloc() instead of kmalloc_array()

We need the memory to be zeroed upon allocation so use kcalloc()
instead.

Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: stmmac: RX Descriptors need to be clean before setting buffers
Jose Abreu [Mon, 22 Jul 2019 08:39:30 +0000 (10:39 +0200)]
net: stmmac: RX Descriptors need to be clean before setting buffers

RX Descriptors are being cleaned after setting the buffers which may
lead to buffer addresses being wiped out.

Fix this by clearing earlier the RX Descriptors.

Fixes: 2af6106ae949 ("net: stmmac: Introducing support for Page Pool")
Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns: fix LED configuration for marvell phy
Yonglong Liu [Mon, 22 Jul 2019 05:59:12 +0000 (13:59 +0800)]
net: hns: fix LED configuration for marvell phy

Since commit(net: phy: marvell: change default m88e1510 LED configuration),
the active LED of Hip07 devices is always off, because Hip07 just
use 2 LEDs.
This patch adds a phy_register_fixup_for_uid() for m88e1510 to
correct the LED configuration.

Fixes: 077772468ec1 ("net: phy: marvell: change default m88e1510 LED configuration")
Signed-off-by: Yonglong Liu <liuyonglong@huawei.com>
Reviewed-by: linyunsheng <linyunsheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: mvpp2: Don't check for 3 consecutive Idle frames for 10G links
Maxime Chevallier [Fri, 19 Jul 2019 14:38:48 +0000 (16:38 +0200)]
net: mvpp2: Don't check for 3 consecutive Idle frames for 10G links

PPv2's XLGMAC can wait for 3 idle frames before triggering a link up
event. This can cause the link to be stuck low when there's traffic on
the interface, so disable this feature.

Fixes: 4bb043262878 ("net: mvpp2: phylink support")
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agobonding: Force slave speed check after link state recovery for 802.3ad
Thomas Falcon [Tue, 16 Jul 2019 22:25:10 +0000 (17:25 -0500)]
bonding: Force slave speed check after link state recovery for 802.3ad

The following scenario was encountered during testing of logical
partition mobility on pseries partitions with bonded ibmvnic
adapters in LACP mode.

1. Driver receives a signal that the device has been
   swapped, and it needs to reset to initialize the new
   device.

2. Driver reports loss of carrier and begins initialization.

3. Bonding driver receives NETDEV_CHANGE notifier and checks
   the slave's current speed and duplex settings. Because these
   are unknown at the time, the bond sets its link state to
   BOND_LINK_FAIL and handles the speed update, clearing
   AD_PORT_LACP_ENABLE.

4. Driver finishes recovery and reports that the carrier is on.

5. Bond receives a new notification and checks the speed again.
   The speeds are valid but miimon has not altered the link
   state yet.  AD_PORT_LACP_ENABLE remains off.

Because the slave's link state is still BOND_LINK_FAIL,
no further port checks are made when it recovers. Though
the slave devices are operational and have valid speed
and duplex settings, the bond will not send LACPDU's. The
simplest fix I can see is to force another speed check
in bond_miimon_commit. This way the bond will update
AD_PORT_LACP_ENABLE if needed when transitioning from
BOND_LINK_FAIL to BOND_LINK_UP.

CC: Jarod Wilson <jarod@redhat.com>
CC: Jay Vosburgh <j.vosburgh@gmail.com>
CC: Veaceslav Falico <vfalico@gmail.com>
CC: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Thomas Falcon <tlfalcon@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Mon, 22 Jul 2019 16:30:34 +0000 (09:30 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull preemption Kconfig fix from Thomas Gleixner:
 "The PREEMPT_RT stub config renamed PREEMPT to PREEMPT_LL and defined
  PREEMPT outside of the menu and made it selectable by both PREEMPT_LL
  and PREEMPT_RT.

  Stupid me missed that 114 defconfigs select CONFIG_PREEMPT which
  obviously can't work anymore. oldconfig builds are affected as well,
  but it's more obvious as the user gets asked. [old]defconfig silently
  fixes it up and selects PREEMPT_NONE.

  Unbreak it by undoing the rename and adding a intermediate config
  symbol which is selected by both PREEMPT and PREEMPT_RT. That requires
  to chase down a few #ifdefs, but it's better than tweaking 114
  defconfigs and annoying users"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/rt, Kconfig: Unbreak def/oldconfig with CONFIG_PREEMPT=y

5 years agoMerge tag 'for-linus-20190722' of git://git.kernel.org/pub/scm/linux/kernel/git/braun...
Linus Torvalds [Mon, 22 Jul 2019 16:14:19 +0000 (09:14 -0700)]
Merge tag 'for-linus-20190722' of git://git./linux/kernel/git/brauner/linux

Pull pidfd polling fix from Christian Brauner:
 "A fix for pidfd polling. It ensures that the task's exit state is
  visible to all waiters"

* tag 'for-linus-20190722' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux:
  pidfd: fix a poll race when setting exit_state

5 years agoMerge tag 'for-5.3-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Mon, 22 Jul 2019 16:08:38 +0000 (09:08 -0700)]
Merge tag 'for-5.3-rc1-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:

 - fixes for leaks caused by recently merged patches

 - one build fix

 - a fix to prevent mixing of incompatible features

* tag 'for-5.3-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: don't leak extent_map in btrfs_get_io_geometry()
  btrfs: free checksum hash on in close_ctree
  btrfs: Fix build error while LIBCRC32C is module
  btrfs: inode: Don't compress if NODATASUM or NODATACOW set

5 years agosched/rt, Kconfig: Unbreak def/oldconfig with CONFIG_PREEMPT=y
Thomas Gleixner [Mon, 22 Jul 2019 15:59:19 +0000 (17:59 +0200)]
sched/rt, Kconfig: Unbreak def/oldconfig with CONFIG_PREEMPT=y

The merge of the CONFIG_PREEMPT_RT stub renamed CONFIG_PREEMPT to
CONFIG_PREEMPT_LL which causes all defconfigs which have CONFIG_PREEMPT=y
set to fall back to CONFIG_PREEMPT_NONE because CONFIG_PREEMPT depends on
the preemption mode choice wich defaults to NONE. This also affects
oldconfig builds.

So rather than changing 114 defconfig files and being an annoyance to
users, revert the rename and select a new config symbol PREEMPTION. That
keeps everything working smoothly and the revelant ifdef's are going to be
fixed up step by step.

Reported-by: Mark Rutland <mark.rutland@arm.com>
Fixes: a50a3f4b6a31 ("sched/rt, Kconfig: Introduce CONFIG_PREEMPT_RT")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
5 years agoMerge tag 'media/v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Mon, 22 Jul 2019 16:01:47 +0000 (09:01 -0700)]
Merge tag 'media/v5.3-2' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:
 "For two regressions in media core:

   - v4l2-subdev: fix regression in check_pad()

   - videodev2.h: change V4L2_PIX_FMT_BGRA444 define: fourcc was already
     in use"

* tag 'media/v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  media: videodev2.h: change V4L2_PIX_FMT_BGRA444 define: fourcc was already in use
  media: v4l2-subdev: fix regression in check_pad()

5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Mon, 22 Jul 2019 15:49:22 +0000 (08:49 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Several netfilter fixes including a nfnetlink deadlock fix from
    Florian Westphal and fix for dropping VRF packets from Miaohe Lin.

 2) Flow offload fixes from Pablo Neira Ayuso including a fix to restore
    proper block sharing.

 3) Fix r8169 PHY init from Thomas Voegtle.

 4) Fix memory leak in mac80211, from Lorenzo Bianconi.

 5) Missing NULL check on object allocation in cxgb4, from Navid
    Emamdoost.

 6) Fix scaling of RX power in sfp phy driver, from Andrew Lunn.

 7) Check that there is actually an ip header to access in skb->data in
    VRF, from Peter Kosyh.

 8) Remove spurious rcu unlock in hv_netvsc, from Haiyang Zhang.

 9) One more tweak the the TCP fragmentation memory limit changes, to be
    less harmful to applications setting small SO_SNDBUF values. From
    Eric Dumazet.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (40 commits)
  tcp: be more careful in tcp_fragment()
  hv_netvsc: Fix extra rcu_read_unlock in netvsc_recv_callback()
  vrf: make sure skb->data contains ip header to make routing
  connector: remove redundant input callback from cn_dev
  qed: Prefer pcie_capability_read_word()
  igc: Prefer pcie_capability_read_word()
  cxgb4: Prefer pcie_capability_read_word()
  be2net: Synchronize be_update_queues with dev_watchdog
  bnx2x: Prevent load reordering in tx completion processing
  net: phy: sfp: hwmon: Fix scaling of RX power
  net: sched: verify that q!=NULL before setting q->flags
  chelsio: Fix a typo in a function name
  allocate_flower_entry: should check for null deref
  net: hns3: typo in the name of a constant
  kbuild: add net/netfilter/nf_tables_offload.h to header-test blacklist.
  tipc: Fix a typo
  mac80211: don't warn about CW params when not using them
  mac80211: fix possible memory leak in ieee80211_assign_beacon
  nl80211: fix NL80211_HE_MAX_CAPABILITY_LEN
  nl80211: fix VENDOR_CMD_RAW_DATA
  ...

5 years agoselftests/bpf: fix sendmsg6_prog on s390
Ilya Leoshkevich [Fri, 19 Jul 2019 09:06:11 +0000 (11:06 +0200)]
selftests/bpf: fix sendmsg6_prog on s390

"sendmsg6: rewrite IP & port (C)" fails on s390, because the code in
sendmsg_v6_prog() assumes that (ctx->user_ip6[0] & 0xFFFF) refers to
leading IPv6 address digits, which is not the case on big-endian
machines.

Since checking bitwise operations doesn't seem to be the point of the
test, replace two short comparisons with a single int comparison.

Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agolibbpf: Avoid designated initializers for unnamed union members
Arnaldo Carvalho de Melo [Fri, 19 Jul 2019 14:34:07 +0000 (11:34 -0300)]
libbpf: Avoid designated initializers for unnamed union members

As it fails to build in some systems with:

  libbpf.c: In function 'perf_buffer__new':
  libbpf.c:4515: error: unknown field 'sample_period' specified in initializer
  libbpf.c:4516: error: unknown field 'wakeup_events' specified in initializer

Doing as:

    attr.sample_period = 1;

I.e. not as a designated initializer makes it build everywhere.

Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Fixes: fb84b8224655 ("libbpf: add perf buffer API")
Link: https://lkml.kernel.org/n/tip-hnlmch8qit1ieksfppmr32si@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agolibbpf: Fix endianness macro usage for some compilers
Arnaldo Carvalho de Melo [Fri, 19 Jul 2019 14:34:06 +0000 (11:34 -0300)]
libbpf: Fix endianness macro usage for some compilers

Using endian.h and its endianness macros makes this code build in a
wider range of compilers, as some don't have those macros
(__BYTE_ORDER__, __ORDER_LITTLE_ENDIAN__, __ORDER_BIG_ENDIAN__),
so use instead endian.h's macros (__BYTE_ORDER, __LITTLE_ENDIAN,
__BIG_ENDIAN) which makes this code even shorter :-)

Acked-by: Andrii Nakryiko <andriin@fb.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Fixes: 12ef5634a855 ("libbpf: simplify endianness check")
Fixes: e6c64855fd7a ("libbpf: add btf__parse_elf API to load .BTF and .BTF.ext")
Link: https://lkml.kernel.org/n/tip-eep5n8vgwcdphw3uc058k03u@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoMerge branch 'bpf-sockmap-tls-fixes'
Daniel Borkmann [Mon, 22 Jul 2019 14:04:17 +0000 (16:04 +0200)]
Merge branch 'bpf-sockmap-tls-fixes'

Jakub Kicinski says:

====================
John says:

Resolve a series of splats discovered by syzbot and an unhash
TLS issue noted by Eric Dumazet.

The main issues revolved around interaction between TLS and
sockmap tear down. TLS and sockmap could both reset sk->prot
ops creating a condition where a close or unhash op could be
called forever. A rare race condition resulting from a missing
rcu sync operation was causing a use after free. Then on the
TLS side dropping the sock lock and re-acquiring it during the
close op could hang. Finally, sockmap must be deployed before
tls for current stack assumptions to be met. This is enforced
now. A feature series can enable it.

To fix this first refactor TLS code so the lock is held for the
entire teardown operation. Then add an unhash callback to ensure
TLS can not transition from ESTABLISHED to LISTEN state. This
transition is a similar bug to the one found and fixed previously
in sockmap. Then apply three fixes to sockmap to fix up races
on tear down around map free and close. Finally, if sockmap
is destroyed before TLS we add a new ULP op update to inform
the TLS stack it should not call sockmap ops. This last one
appears to be the most commonly found issue from syzbot.

v4:
 - fix some use after frees;
 - disable disconnect work for offload (ctx lifetime is much
   more complex);
 - remove some of the dead code which made it hard to understand
   (for me) that things work correctly (e.g. the checks TLS is
   the top ULP);
 - add selftets.
====================

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoselftests/tls: add shutdown tests
Jakub Kicinski [Fri, 19 Jul 2019 17:29:27 +0000 (10:29 -0700)]
selftests/tls: add shutdown tests

Add test for killing the connection via shutdown.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoselftests/tls: close the socket with open record
Jakub Kicinski [Fri, 19 Jul 2019 17:29:26 +0000 (10:29 -0700)]
selftests/tls: close the socket with open record

Add test which sends some data with MSG_MORE and then
closes the socket (never calling send without MSG_MORE).
This should make sure we clean up open records correctly.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoselftests/tls: add a bidirectional test
Jakub Kicinski [Fri, 19 Jul 2019 17:29:25 +0000 (10:29 -0700)]
selftests/tls: add a bidirectional test

Add a simple test which installs the TLS state for both directions,
sends and receives data on both sockets.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoselftests/tls: test error codes around TLS ULP installation
Jakub Kicinski [Fri, 19 Jul 2019 17:29:24 +0000 (10:29 -0700)]
selftests/tls: test error codes around TLS ULP installation

Test the error codes returned when TCP connection is not
in ESTABLISHED state.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agoselftests/tls: add a test for ULP but no keys
Jakub Kicinski [Fri, 19 Jul 2019 17:29:23 +0000 (10:29 -0700)]
selftests/tls: add a test for ULP but no keys

Make sure we test the TLS_BASE/TLS_BASE case both with data
and the tear down/clean up path.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agobpf: sockmap/tls, close can race with map free
John Fastabend [Fri, 19 Jul 2019 17:29:22 +0000 (10:29 -0700)]
bpf: sockmap/tls, close can race with map free

When a map free is called and in parallel a socket is closed we
have two paths that can potentially reset the socket prot ops, the
bpf close() path and the map free path. This creates a problem
with which prot ops should be used from the socket closed side.

If the map_free side completes first then we want to call the
original lowest level ops. However, if the tls path runs first
we want to call the sockmap ops. Additionally there was no locking
around prot updates in TLS code paths so the prot ops could
be changed multiple times once from TLS path and again from sockmap
side potentially leaving ops pointed at either TLS or sockmap
when psock and/or tls context have already been destroyed.

To fix this race first only update ops inside callback lock
so that TLS, sockmap and lowest level all agree on prot state.
Second and a ULP callback update() so that lower layers can
inform the upper layer when they are being removed allowing the
upper layer to reset prot ops.

This gets us close to allowing sockmap and tls to be stacked
in arbitrary order but will save that patch for *next trees.

v4:
 - make sure we don't free things for device;
 - remove the checks which swap the callbacks back
   only if TLS is at the top.

Reported-by: syzbot+06537213db7ba2745c4a@syzkaller.appspotmail.com
Fixes: 02c558b2d5d6 ("bpf: sockmap, support for msg_peek in sk_msg with redirect ingress")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agobpf: sockmap, only create entry if ulp is not already enabled
John Fastabend [Fri, 19 Jul 2019 17:29:21 +0000 (10:29 -0700)]
bpf: sockmap, only create entry if ulp is not already enabled

Sockmap does not currently support adding sockets after TLS has been
enabled. There never was a real use case for this so it was never
added. But, we lost the test for ULP at some point so add it here
and fail the socket insert if TLS is enabled. Future work could
make sockmap support this use case but fixup the bug here.

Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agobpf: sockmap, synchronize_rcu before free'ing map
John Fastabend [Fri, 19 Jul 2019 17:29:20 +0000 (10:29 -0700)]
bpf: sockmap, synchronize_rcu before free'ing map

We need to have a synchronize_rcu before free'ing the sockmap because
any outstanding psock references will have a pointer to the map and
when they use this could trigger a use after free.

Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agobpf: sockmap, sock_map_delete needs to use xchg
John Fastabend [Fri, 19 Jul 2019 17:29:19 +0000 (10:29 -0700)]
bpf: sockmap, sock_map_delete needs to use xchg

__sock_map_delete() may be called from a tcp event such as unhash or
close from the following trace,

  tcp_bpf_close()
    tcp_bpf_remove()
      sk_psock_unlink()
        sock_map_delete_from_link()
          __sock_map_delete()

In this case the sock lock is held but this only protects against
duplicate removals on the TCP side. If the map is free'd then we have
this trace,

  sock_map_free
    xchg()                  <- replaces map entry
    sock_map_unref()
      sk_psock_put()
        sock_map_del_link()

The __sock_map_delete() call however uses a read, test, null over the
map entry which can result in both paths trying to free the map
entry.

To fix use xchg in TCP paths as well so we avoid having two references
to the same map entry.

Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agonet/tls: fix transition through disconnect with close
John Fastabend [Fri, 19 Jul 2019 17:29:18 +0000 (10:29 -0700)]
net/tls: fix transition through disconnect with close

It is possible (via shutdown()) for TCP socks to go through TCP_CLOSE
state via tcp_disconnect() without actually calling tcp_close which
would then call the tls close callback. Because of this a user could
disconnect a socket then put it in a LISTEN state which would break
our assumptions about sockets always being ESTABLISHED state.

More directly because close() can call unhash() and unhash is
implemented by sockmap if a sockmap socket has TLS enabled we can
incorrectly destroy the psock from unhash() and then call its close
handler again. But because the psock (sockmap socket representation)
is already destroyed we call close handler in sk->prot. However,
in some cases (TLS BASE/BASE case) this will still point at the
sockmap close handler resulting in a circular call and crash reported
by syzbot.

To fix both above issues implement the unhash() routine for TLS.

v4:
 - add note about tls offload still needing the fix;
 - move sk_proto to the cold cache line;
 - split TX context free into "release" and "free",
   otherwise the GC work itself is in already freed
   memory;
 - more TX before RX for consistency;
 - reuse tls_ctx_free();
 - schedule the GC work after we're done with context
   to avoid UAF;
 - don't set the unhash in all modes, all modes "inherit"
   TLS_BASE's callbacks anyway;
 - disable the unhash hook for TLS_HW.

Fixes: 3c4d7559159bf ("tls: kernel TLS support")
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agonet/tls: remove sock unlock/lock around strp_done()
John Fastabend [Fri, 19 Jul 2019 17:29:17 +0000 (10:29 -0700)]
net/tls: remove sock unlock/lock around strp_done()

The tls close() callback currently drops the sock lock to call
strp_done(). Split up the RX cleanup into stopping the strparser
and releasing most resources, syncing strparser and finally
freeing the context.

To avoid the need for a strp_done() call on the cleanup path
of device offload make sure we don't arm the strparser until
we are sure init will be successful.

Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agonet/tls: remove close callback sock unlock/lock around TX work flush
John Fastabend [Fri, 19 Jul 2019 17:29:16 +0000 (10:29 -0700)]
net/tls: remove close callback sock unlock/lock around TX work flush

The tls close() callback currently drops the sock lock, makes a
cancel_delayed_work_sync() call, and then relocks the sock.

By restructuring the code we can avoid droping lock and then
reclaiming it. To simplify this we do the following,

 tls_sk_proto_close
 set_bit(CLOSING)
 set_bit(SCHEDULE)
 cancel_delay_work_sync() <- cancel workqueue
 lock_sock(sk)
 ...
 release_sock(sk)
 strp_done()

Setting the CLOSING bit prevents the SCHEDULE bit from being
cleared by any workqueue items e.g. if one happens to be
scheduled and run between when we set SCHEDULE bit and cancel
work. Then because SCHEDULE bit is set now no new work will
be scheduled.

Tested with net selftests and bpf selftests.

Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agonet/tls: don't call tls_sk_proto_close for hw record offload
Jakub Kicinski [Fri, 19 Jul 2019 17:29:15 +0000 (10:29 -0700)]
net/tls: don't call tls_sk_proto_close for hw record offload

The deprecated TOE offload doesn't actually do anything in
tls_sk_proto_close() - all TLS code is skipped and context
not freed. Remove the callback to make it easier to refactor
tls_sk_proto_close().

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agonet/tls: don't arm strparser immediately in tls_set_sw_offload()
Jakub Kicinski [Fri, 19 Jul 2019 17:29:14 +0000 (10:29 -0700)]
net/tls: don't arm strparser immediately in tls_set_sw_offload()

In tls_set_device_offload_rx() we prepare the software context
for RX fallback and proceed to add the connection to the device.
Unfortunately, software context prep includes arming strparser
so in case of a later error we have to release the socket lock
to call strp_done().

In preparation for not releasing the socket lock half way through
callbacks move arming strparser into a separate function.
Following patches will make use of that.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
5 years agopidfd: fix a poll race when setting exit_state
Suren Baghdasaryan [Wed, 17 Jul 2019 17:21:00 +0000 (13:21 -0400)]
pidfd: fix a poll race when setting exit_state

There is a race between reading task->exit_state in pidfd_poll and
writing it after do_notify_parent calls do_notify_pidfd. Expected
sequence of events is:

CPU 0                            CPU 1
------------------------------------------------
exit_notify
  do_notify_parent
    do_notify_pidfd
  tsk->exit_state = EXIT_DEAD
                                  pidfd_poll
                                     if (tsk->exit_state)

However nothing prevents the following sequence:

CPU 0                            CPU 1
------------------------------------------------
exit_notify
  do_notify_parent
    do_notify_pidfd
                                   pidfd_poll
                                      if (tsk->exit_state)
  tsk->exit_state = EXIT_DEAD

This causes a polling task to wait forever, since poll blocks because
exit_state is 0 and the waiting task is not notified again. A stress
test continuously doing pidfd poll and process exits uncovered this bug.

To fix it, we make sure that the task's exit_state is always set before
calling do_notify_pidfd.

Fixes: b53b0b9d9a6 ("pidfd: add polling support")
Cc: kernel-team@android.com
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Link: https://lore.kernel.org/r/20190717172100.261204-1-joel@joelfernandes.org
[christian@brauner.io: adapt commit message and drop unneeded changes from wait_task_zombie]
Signed-off-by: Christian Brauner <christian@brauner.io>
5 years agotcp: be more careful in tcp_fragment()
Eric Dumazet [Fri, 19 Jul 2019 18:52:33 +0000 (11:52 -0700)]
tcp: be more careful in tcp_fragment()

Some applications set tiny SO_SNDBUF values and expect
TCP to just work. Recent patches to address CVE-2019-11478
broke them in case of losses, since retransmits might
be prevented.

We should allow these flows to make progress.

This patch allows the first and last skb in retransmit queue
to be split even if memory limits are hit.

It also adds the some room due to the fact that tcp_sendmsg()
and tcp_sendpage() might overshoot sk_wmem_queued by about one full
TSO skb (64KB size). Note this allowance was already present
in stable backports for kernels < 4.15

Note for < 4.15 backports :
 tcp_rtx_queue_tail() will probably look like :

static inline struct sk_buff *tcp_rtx_queue_tail(const struct sock *sk)
{
struct sk_buff *skb = tcp_send_head(sk);

return skb ? tcp_write_queue_prev(sk, skb) : tcp_write_queue_tail(sk);
}

Fixes: f070ef2ac667 ("tcp: tcp_fragment() should apply sane memory limits")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Andrew Prout <aprout@ll.mit.edu>
Tested-by: Andrew Prout <aprout@ll.mit.edu>
Tested-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Tested-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Christoph Paasch <cpaasch@apple.com>
Cc: Jonathan Looney <jtl@netflix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agohv_netvsc: Fix extra rcu_read_unlock in netvsc_recv_callback()
Haiyang Zhang [Fri, 19 Jul 2019 17:33:51 +0000 (17:33 +0000)]
hv_netvsc: Fix extra rcu_read_unlock in netvsc_recv_callback()

There is an extra rcu_read_unlock left in netvsc_recv_callback(),
after a previous patch that removes RCU from this function.
This patch removes the extra RCU unlock.

Fixes: 345ac08990b8 ("hv_netvsc: pass netvsc_device to receive callback")
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoLinus 5.3-rc1
Linus Torvalds [Sun, 21 Jul 2019 21:05:38 +0000 (14:05 -0700)]
Linus 5.3-rc1

5 years agovrf: make sure skb->data contains ip header to make routing
Peter Kosyh [Fri, 19 Jul 2019 08:11:47 +0000 (11:11 +0300)]
vrf: make sure skb->data contains ip header to make routing

vrf_process_v4_outbound() and vrf_process_v6_outbound() do routing
using ip/ipv6 addresses, but don't make sure the header is available
in skb->data[] (skb_headlen() is less then header size).

Case:

1) igb driver from intel.
2) Packet size is greater then 255.
3) MPLS forwards to VRF device.

So, patch adds pskb_may_pull() calls in vrf_process_v4/v6_outbound()
functions.

Signed-off-by: Peter Kosyh <p.kosyh@gmail.com>
Reviewed-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoconnector: remove redundant input callback from cn_dev
Vasily Averin [Thu, 18 Jul 2019 04:26:46 +0000 (07:26 +0300)]
connector: remove redundant input callback from cn_dev

A small cleanup: this callback is never used.
Originally fixed by Stanislav Kinsburskiy <skinsbursky@virtuozzo.com>
for OpenVZ7 bug OVZ-6877

cc: stanislav.kinsburskiy@gmail.com
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoqed: Prefer pcie_capability_read_word()
Frederick Lawler [Thu, 18 Jul 2019 02:07:42 +0000 (21:07 -0500)]
qed: Prefer pcie_capability_read_word()

Commit 8c0d3a02c130 ("PCI: Add accessors for PCI Express Capability")
added accessors for the PCI Express Capability so that drivers didn't
need to be aware of differences between v1 and v2 of the PCI
Express Capability.

Replace pci_read_config_word() and pci_write_config_word() calls with
pcie_capability_read_word() and pcie_capability_write_word().

Signed-off-by: Frederick Lawler <fred@fredlawl.com>
Acked-by: Michal Kalderon <michal.kalderon@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoigc: Prefer pcie_capability_read_word()
Frederick Lawler [Thu, 18 Jul 2019 02:07:39 +0000 (21:07 -0500)]
igc: Prefer pcie_capability_read_word()

Commit 8c0d3a02c130 ("PCI: Add accessors for PCI Express Capability")
added accessors for the PCI Express Capability so that drivers didn't
need to be aware of differences between v1 and v2 of the PCI
Express Capability.

Replace pci_read_config_word() and pci_write_config_word() calls with
pcie_capability_read_word() and pcie_capability_write_word().

Signed-off-by: Frederick Lawler <fred@fredlawl.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agocxgb4: Prefer pcie_capability_read_word()
Frederick Lawler [Thu, 18 Jul 2019 02:07:36 +0000 (21:07 -0500)]
cxgb4: Prefer pcie_capability_read_word()

Commit 8c0d3a02c130 ("PCI: Add accessors for PCI Express Capability")
added accessors for the PCI Express Capability so that drivers didn't
need to be aware of differences between v1 and v2 of the PCI
Express Capability.

Replace pci_read_config_word() and pci_write_config_word() calls with
pcie_capability_read_word() and pcie_capability_write_word().

Signed-off-by: Frederick Lawler <fred@fredlawl.com>
Reviewed-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>