platform/kernel/linux-rpi.git
3 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
David S. Miller [Tue, 25 May 2021 22:59:24 +0000 (15:59 -0700)]
Merge git://git./pub/scm/linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2021-05-26

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

We've added 14 non-merge commits during the last 14 day(s) which contain
a total of 17 files changed, 513 insertions(+), 231 deletions(-).

The main changes are:

1) Fix bpf_skb_change_head() helper to reset mac_len, from Jussi Maki.

2) Fix masking direction swap upon off-reg sign change, from Daniel Borkmann.

3) Fix BPF offloads in verifier by reordering driver callback, from Yinjun Zhang.

4) BPF selftest for ringbuf mmap ro/rw restrictions, from Andrii Nakryiko.

5) Follow-up fixes to nested bprintf per-cpu buffers, from Florent Revest.

6) Fix bpftool sock_release attach point help info, from Liu Jian.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'mptcp-fixes'
David S. Miller [Tue, 25 May 2021 22:56:20 +0000 (15:56 -0700)]
Merge branch 'mptcp-fixes'

Mat Martineau says:

====================
MPTCP fixes

Here are a few fixes for the -net tree.

Patch 1 fixes an attempt to access a tcp-specific field that does not
exist in mptcp sockets.

Patches 2 and 3 remove warning/error log output that could be flooded.

Patch 4 performs more validation on address advertisement echo packets
to improve RFC 8684 compliance.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agomptcp: validate 'id' when stopping the ADD_ADDR retransmit timer
Davide Caratti [Tue, 25 May 2021 21:23:13 +0000 (14:23 -0700)]
mptcp: validate 'id' when stopping the ADD_ADDR retransmit timer

when Linux receives an echo-ed ADD_ADDR, it checks the IP address against
the list of "announced" addresses. In case of a positive match, the timer
that handles retransmissions is stopped regardless of the 'Address Id' in
the received packet: this behaviour does not comply with RFC8684 3.4.1.

Fix it by validating the 'Address Id' in received echo-ed ADD_ADDRs.
Tested using packetdrill, with the following captured output:

 unpatched kernel:

 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0
 In  <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 3013740213], length 0
 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0xfd2e62517888fe29,mptcp dss ack 3007449509], length 0
 In  <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 3013740213], length 0
        ^^^ retransmission is stopped here, but 'Address Id' is 90

 patched kernel:

 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0
 In  <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 1.2.3.4,mptcp dss ack 1672384568], length 0
 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0
 In  <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 90 198.51.100.2,mptcp dss ack 1672384568], length 0
 Out <...> Flags [.], ack 1, win 256, options [mptcp add-addr v1 id 1 198.51.100.2 hmac 0x1cf372d59e05f4b8,mptcp dss ack 3007449509], length 0
 In  <...> Flags [.], ack 1, win 257, options [mptcp add-addr v1-echo id 1 198.51.100.2,mptcp dss ack 1672384568], length 0
        ^^^ retransmission is stopped here, only when both 'Address Id' and 'IP Address' match

Fixes: 00cfd77b9063 ("mptcp: retransmit ADD_ADDR when timeout")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agomptcp: avoid error message on infinite mapping
Paolo Abeni [Tue, 25 May 2021 21:23:12 +0000 (14:23 -0700)]
mptcp: avoid error message on infinite mapping

Another left-over. Avoid flooding dmesg with useless text,
we already have a MIB for that event.

Fixes: 648ef4b88673 ("mptcp: Implement MPTCP receive path")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agomptcp: drop unconditional pr_warn on bad opt
Paolo Abeni [Tue, 25 May 2021 21:23:11 +0000 (14:23 -0700)]
mptcp: drop unconditional pr_warn on bad opt

This is a left-over of early day. A malicious peer can flood
the kernel logs with useless messages, just drop it.

Fixes: f296234c98a8 ("mptcp: Add handling of incoming MP_JOIN requests")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agomptcp: avoid OOB access in setsockopt()
Paolo Abeni [Tue, 25 May 2021 21:23:10 +0000 (14:23 -0700)]
mptcp: avoid OOB access in setsockopt()

We can't use tcp_set_congestion_control() on an mptcp socket, as
such function can end-up accessing a tcp-specific field -
prior_ssthresh - causing an OOB access.

To allow propagating the correct ca algo on subflow, cache the ca
name at initialization time.

Additionally avoid overriding the user-selected CA (if any) at
clone time.

Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/182
Fixes: aa1fbd94e5c7 ("mptcp: sockopt: add TCP_CONGESTION and TCP_INFO")
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonfp: update maintainer and mailing list addresses
Simon Horman [Tue, 25 May 2021 15:47:04 +0000 (17:47 +0200)]
nfp: update maintainer and mailing list addresses

Some of Netronome's activities and people have moved over to Corigine,
including NFP driver maintenance and myself.

Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: mvpp2: add buffer header handling in RX
Stefan Chulski [Tue, 25 May 2021 16:04:41 +0000 (19:04 +0300)]
net: mvpp2: add buffer header handling in RX

If Link Partner sends frames larger than RX buffer size, MAC mark it
as oversize but still would pass it to the Packet Processor.
In this scenario, Packet Processor scatter frame between multiple buffers,
but only a single buffer would be returned to the Buffer Manager pool and
it would not refill the poll.

Patch add handling of oversize error with buffer header handling, so all
buffers would be returned to the Buffer Manager pool.

Fixes: 3f518509dedc ("ethernet: Add new driver for Marvell Armada 375 network unit")
Reported-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Stefan Chulski <stefanc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agobnx2x: Fix missing error code in bnx2x_iov_init_one()
Jiapeng Chong [Tue, 25 May 2021 11:00:12 +0000 (19:00 +0800)]
bnx2x: Fix missing error code in bnx2x_iov_init_one()

Eliminate the follow smatch warning:

drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c:1227
bnx2x_iov_init_one() warn: missing error code 'err'.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: zero-initialize tc skb extension on allocation
Vlad Buslov [Tue, 25 May 2021 13:21:52 +0000 (16:21 +0300)]
net: zero-initialize tc skb extension on allocation

Function skb_ext_add() doesn't initialize created skb extension with any
value and leaves it up to the user. However, since extension of type
TC_SKB_EXT originally contained only single value tc_skb_ext->chain its
users used to just assign the chain value without setting whole extension
memory to zero first. This assumption changed when TC_SKB_EXT extension was
extended with additional fields but not all users were updated to
initialize the new fields which leads to use of uninitialized memory
afterwards. UBSAN log:

[  778.299821] UBSAN: invalid-load in net/openvswitch/flow.c:899:28
[  778.301495] load of value 107 is not a valid value for type '_Bool'
[  778.303215] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.12.0-rc7+ #2
[  778.304933] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
[  778.307901] Call Trace:
[  778.308680]  <IRQ>
[  778.309358]  dump_stack+0xbb/0x107
[  778.310307]  ubsan_epilogue+0x5/0x40
[  778.311167]  __ubsan_handle_load_invalid_value.cold+0x43/0x48
[  778.312454]  ? memset+0x20/0x40
[  778.313230]  ovs_flow_key_extract.cold+0xf/0x14 [openvswitch]
[  778.314532]  ovs_vport_receive+0x19e/0x2e0 [openvswitch]
[  778.315749]  ? ovs_vport_find_upcall_portid+0x330/0x330 [openvswitch]
[  778.317188]  ? create_prof_cpu_mask+0x20/0x20
[  778.318220]  ? arch_stack_walk+0x82/0xf0
[  778.319153]  ? secondary_startup_64_no_verify+0xb0/0xbb
[  778.320399]  ? stack_trace_save+0x91/0xc0
[  778.321362]  ? stack_trace_consume_entry+0x160/0x160
[  778.322517]  ? lock_release+0x52e/0x760
[  778.323444]  netdev_frame_hook+0x323/0x610 [openvswitch]
[  778.324668]  ? ovs_netdev_get_vport+0xe0/0xe0 [openvswitch]
[  778.325950]  __netif_receive_skb_core+0x771/0x2db0
[  778.327067]  ? lock_downgrade+0x6e0/0x6f0
[  778.328021]  ? lock_acquire+0x565/0x720
[  778.328940]  ? generic_xdp_tx+0x4f0/0x4f0
[  778.329902]  ? inet_gro_receive+0x2a7/0x10a0
[  778.330914]  ? lock_downgrade+0x6f0/0x6f0
[  778.331867]  ? udp4_gro_receive+0x4c4/0x13e0
[  778.332876]  ? lock_release+0x52e/0x760
[  778.333808]  ? dev_gro_receive+0xcc8/0x2380
[  778.334810]  ? lock_downgrade+0x6f0/0x6f0
[  778.335769]  __netif_receive_skb_list_core+0x295/0x820
[  778.336955]  ? process_backlog+0x780/0x780
[  778.337941]  ? mlx5e_rep_tc_netdevice_event_unregister+0x20/0x20 [mlx5_core]
[  778.339613]  ? seqcount_lockdep_reader_access.constprop.0+0xa7/0xc0
[  778.341033]  ? kvm_clock_get_cycles+0x14/0x20
[  778.342072]  netif_receive_skb_list_internal+0x5f5/0xcb0
[  778.343288]  ? __kasan_kmalloc+0x7a/0x90
[  778.344234]  ? mlx5e_handle_rx_cqe_mpwrq+0x9e0/0x9e0 [mlx5_core]
[  778.345676]  ? mlx5e_xmit_xdp_frame_mpwqe+0x14d0/0x14d0 [mlx5_core]
[  778.347140]  ? __netif_receive_skb_list_core+0x820/0x820
[  778.348351]  ? mlx5e_post_rx_mpwqes+0xa6/0x25d0 [mlx5_core]
[  778.349688]  ? napi_gro_flush+0x26c/0x3c0
[  778.350641]  napi_complete_done+0x188/0x6b0
[  778.351627]  mlx5e_napi_poll+0x373/0x1b80 [mlx5_core]
[  778.352853]  __napi_poll+0x9f/0x510
[  778.353704]  ? mlx5_flow_namespace_set_mode+0x260/0x260 [mlx5_core]
[  778.355158]  net_rx_action+0x34c/0xa40
[  778.356060]  ? napi_threaded_poll+0x3d0/0x3d0
[  778.357083]  ? sched_clock_cpu+0x18/0x190
[  778.358041]  ? __common_interrupt+0x8e/0x1a0
[  778.359045]  __do_softirq+0x1ce/0x984
[  778.359938]  __irq_exit_rcu+0x137/0x1d0
[  778.360865]  irq_exit_rcu+0xa/0x20
[  778.361708]  common_interrupt+0x80/0xa0
[  778.362640]  </IRQ>
[  778.363212]  asm_common_interrupt+0x1e/0x40
[  778.364204] RIP: 0010:native_safe_halt+0xe/0x10
[  778.365273] Code: 4f ff ff ff 4c 89 e7 e8 50 3f 40 fe e9 dc fe ff ff 48 89 df e8 43 3f 40 fe eb 90 cc e9 07 00 00 00 0f 00 2d 74 05 62 00 fb f4 <c3> 90 e9 07 00 00 00 0f 00 2d 64 05 62 00 f4 c3 cc cc 0f 1f 44 00
[  778.369355] RSP: 0018:ffffffff84407e48 EFLAGS: 00000246
[  778.370570] RAX: ffff88842de46a80 RBX: ffffffff84425840 RCX: ffffffff83418468
[  778.372143] RDX: 000000000026f1da RSI: 0000000000000004 RDI: ffffffff8343af5e
[  778.373722] RBP: fffffbfff0884b08 R08: 0000000000000000 R09: ffff88842de46bcb
[  778.375292] R10: ffffed1085bc8d79 R11: 0000000000000001 R12: 0000000000000000
[  778.376860] R13: ffffffff851124a0 R14: 0000000000000000 R15: dffffc0000000000
[  778.378491]  ? rcu_eqs_enter.constprop.0+0xb8/0xe0
[  778.379606]  ? default_idle_call+0x5e/0xe0
[  778.380578]  default_idle+0xa/0x10
[  778.381406]  default_idle_call+0x96/0xe0
[  778.382350]  do_idle+0x3d4/0x550
[  778.383153]  ? arch_cpu_idle_exit+0x40/0x40
[  778.384143]  cpu_startup_entry+0x19/0x20
[  778.385078]  start_kernel+0x3c7/0x3e5
[  778.385978]  secondary_startup_64_no_verify+0xb0/0xbb

Fix the issue by providing new function tc_skb_ext_alloc() that allocates
tc skb extension and initializes its memory to 0 before returning it to the
caller. Change all existing users to use new API instead of calling
skb_ext_add() directly.

Fixes: 038ebb1a713d ("net/sched: act_ct: fix miss set mru for ovs after defrag in act_ct")
Fixes: d29334c15d33 ("net/sched: act_api: fix miss set post_ct for ovs after do conntrack in act_ct")
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Acked-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns: Fix kernel-doc
Yang Li [Tue, 25 May 2021 10:52:47 +0000 (18:52 +0800)]
net: hns: Fix kernel-doc

Fix function name in hns_ethtool.c kernel-doc comment
to remove these warnings found by clang_w1.

drivers/net/ethernet/hisilicon/hns/hns_ethtool.c:202: warning: expecting
prototype for hns_nic_set_link_settings(). Prototype was for
hns_nic_set_link_ksettings() instead.
drivers/net/ethernet/hisilicon/hns/hns_ethtool.c:837: warning: expecting
prototype for get_ethtool_stats(). Prototype was for
hns_get_ethtool_stats() instead.
drivers/net/ethernet/hisilicon/hns/hns_ethtool.c:894: warning:
expecting prototype for get_strings(). Prototype was for
hns_get_strings() instead.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Fixes: 'commit 262b38cdb3e4 ("net: ethernet: hisilicon: hns: use phydev
from struct net_device")'
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agosctp: fix the proc_handler for sysctl encap_port
Xin Long [Tue, 25 May 2021 02:49:42 +0000 (22:49 -0400)]
sctp: fix the proc_handler for sysctl encap_port

proc_dointvec() cannot do min and max check for setting a value
when extra1/extra2 is set, so change it to proc_dointvec_minmax()
for sysctl encap_port.

Fixes: e8a3001c2120 ("sctp: add encap_port for netns sock asoc and transport")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agosctp: add the missing setting for asoc encap_port
Xin Long [Tue, 25 May 2021 02:49:24 +0000 (22:49 -0400)]
sctp: add the missing setting for asoc encap_port

This patch is to add the missing setting back for asoc encap_port.

Fixes: 8dba29603b5c ("sctp: add SCTP_REMOTE_UDP_ENCAPS_PORT sockopt")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agobpf, selftests: Adjust few selftest result_unpriv outcomes
Daniel Borkmann [Tue, 4 May 2021 08:58:25 +0000 (08:58 +0000)]
bpf, selftests: Adjust few selftest result_unpriv outcomes

Given we don't need to simulate the speculative domain for registers with
immediates anymore since the verifier uses direct imm-based rewrites instead
of having to mask, we can also lift a few cases that were previously rejected.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
3 years agobpf: No need to simulate speculative domain for immediates
Daniel Borkmann [Tue, 4 May 2021 08:58:25 +0000 (08:58 +0000)]
bpf: No need to simulate speculative domain for immediates

In 801c6058d14a ("bpf: Fix leakage of uninitialized bpf stack under
speculation") we replaced masking logic with direct loads of immediates
if the register is a known constant. Given in this case we do not apply
any masking, there is also no reason for the operation to be truncated
under the speculative domain.

Therefore, there is also zero reason for the verifier to branch-off and
simulate this case, it only needs to do it for unknown but bounded scalars.
As a side-effect, this also enables few test cases that were previously
rejected due to simulation under zero truncation.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Piotr Krysiuk <piotras@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
3 years agobpf: Fix mask direction swap upon off reg sign change
Daniel Borkmann [Fri, 21 May 2021 10:19:22 +0000 (10:19 +0000)]
bpf: Fix mask direction swap upon off reg sign change

Masking direction as indicated via mask_to_left is considered to be
calculated once and then used to derive pointer limits. Thus, this
needs to be placed into bpf_sanitize_info instead so we can pass it
to sanitize_ptr_alu() call after the pointer move. Piotr noticed a
corner case where the off reg causes masking direction change which
then results in an incorrect final aux->alu_limit.

Fixes: 7fedb63a8307 ("bpf: Tighten speculative pointer arithmetic mask")
Reported-by: Piotr Krysiuk <piotras@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Piotr Krysiuk <piotras@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
3 years agobpf: Wrap aux data inside bpf_sanitize_info container
Daniel Borkmann [Fri, 21 May 2021 10:17:36 +0000 (10:17 +0000)]
bpf: Wrap aux data inside bpf_sanitize_info container

Add a container structure struct bpf_sanitize_info which holds
the current aux info, and update call-sites to sanitize_ptr_alu()
to pass it in. This is needed for passing in additional state
later on.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Piotr Krysiuk <piotras@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
3 years agobpf: Fix BPF_LSM kconfig symbol dependency
Daniel Borkmann [Tue, 25 May 2021 18:35:29 +0000 (20:35 +0200)]
bpf: Fix BPF_LSM kconfig symbol dependency

Similarly as 6bdacdb48e94 ("bpf: Fix BPF_JIT kconfig symbol dependency") we
need to detangle the hard BPF_LSM dependency on NET. This was previously
implicit by its dependency on BPF_JIT which itself was dependent on NET (but
without any actual/real hard dependency code-wise). Given the latter was
lifted, so should be the former as BPF_LSMs could well exist on net-less
systems. This therefore also fixes a randconfig build error recently reported
by Randy:

  ld: kernel/bpf/bpf_lsm.o: in function `bpf_lsm_func_proto':
  bpf_lsm.c:(.text+0x1a0): undefined reference to `bpf_sk_storage_get_proto'
  ld: bpf_lsm.c:(.text+0x1b8): undefined reference to `bpf_sk_storage_delete_proto'
  [...]

Fixes: b24abcff918a ("bpf, kconfig: Add consolidated menu entry for bpf with core options")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
3 years agoselftests/bpf: Add test for l3 use of bpf_redirect_peer
Jussi Maki [Tue, 25 May 2021 10:29:55 +0000 (10:29 +0000)]
selftests/bpf: Add test for l3 use of bpf_redirect_peer

Add a test case for using bpf_skb_change_head() in combination with
bpf_redirect_peer() to redirect a packet from a L3 device to veth and back.

The test uses a BPF program that adds L2 headers to the packet coming
from a L3 device and then calls bpf_redirect_peer() to redirect the packet
to a veth device. The test fails as skb->mac_len is not set properly and
thus the ethernet headers are not properly skb_pull'd in cls_bpf_classify(),
causing tcp_v4_rcv() to point the TCP header into middle of the IP header.

Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210525102955.2811090-1-joamaki@gmail.com
3 years agobpftool: Add sock_release help info for cgroup attach/prog load command
Liu Jian [Tue, 25 May 2021 01:41:39 +0000 (09:41 +0800)]
bpftool: Add sock_release help info for cgroup attach/prog load command

The help information was not added at the time when the function got added.
Fix this and add the missing information to its cli, documentation and bash
completion.

Fixes: db94cc0b4805 ("bpftool: Add support for BPF_CGROUP_INET_SOCK_RELEASE")
Signed-off-by: Liu Jian <liujian56@huawei.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20210525014139.323859-1-liujian56@huawei.com
3 years agonet: dsa: microchip: enable phy errata workaround on 9567
George McCollister [Mon, 24 May 2021 20:29:53 +0000 (15:29 -0500)]
net: dsa: microchip: enable phy errata workaround on 9567

Also enable phy errata workaround on 9567 since has the same errata as
the 9477 according to the manufacture's documentation.

Signed-off-by: George McCollister <george.mccollister@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: usb: fix memory leak in smsc75xx_bind
Pavel Skripkin [Mon, 24 May 2021 20:02:08 +0000 (23:02 +0300)]
net: usb: fix memory leak in smsc75xx_bind

Syzbot reported memory leak in smsc75xx_bind().
The problem was is non-freed memory in case of
errors after memory allocation.

backtrace:
  [<ffffffff84245b62>] kmalloc include/linux/slab.h:556 [inline]
  [<ffffffff84245b62>] kzalloc include/linux/slab.h:686 [inline]
  [<ffffffff84245b62>] smsc75xx_bind+0x7a/0x334 drivers/net/usb/smsc75xx.c:1460
  [<ffffffff82b5b2e6>] usbnet_probe+0x3b6/0xc30 drivers/net/usb/usbnet.c:1728

Fixes: d0cad871703b ("smsc75xx: SMSC LAN75xx USB gigabit ethernet adapter driver")
Cc: stable@kernel.vger.org
Reported-and-tested-by: syzbot+b558506ba8165425fee2@syzkaller.appspotmail.com
Signed-off-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hsr: fix mac_len checks
George McCollister [Mon, 24 May 2021 18:50:54 +0000 (13:50 -0500)]
net: hsr: fix mac_len checks

Commit 2e9f60932a2c ("net: hsr: check skb can contain struct hsr_ethhdr
in fill_frame_info") added the following which resulted in -EINVAL
always being returned:
if (skb->mac_len < sizeof(struct hsr_ethhdr))
return -EINVAL;

mac_len was not being set correctly so this check completely broke
HSR/PRP since it was always 14, not 20.

Set mac_len correctly and modify the mac_len checks to test in the
correct places since sometimes it is legitimately 14.

Fixes: 2e9f60932a2c ("net: hsr: check skb can contain struct hsr_ethhdr in fill_frame_info")
Signed-off-by: George McCollister <george.mccollister@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: appletalk: cops: Fix data race in cops_probe1
Saubhik Mukherjee [Mon, 24 May 2021 13:37:12 +0000 (19:07 +0530)]
net: appletalk: cops: Fix data race in cops_probe1

In cops_probe1(), there is a write to dev->base_addr after requesting an
interrupt line and registering the interrupt handler cops_interrupt().
The handler might be called in parallel to handle an interrupt.
cops_interrupt() tries to read dev->base_addr leading to a potential
data race. So write to dev->base_addr before calling request_irq().

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Saubhik Mukherjee <saubhik.mukherjee@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'sja1105-fixes'
David S. Miller [Mon, 24 May 2021 20:20:24 +0000 (13:20 -0700)]
Merge branch 'sja1105-fixes'

Vladimir Oltean says:

====================
Fixes for SJA1105 DSA driver

This series contains some minor fixes in the sja1105 driver:
- improved error handling in the probe path
- rejecting an invalid phy-mode specified in the device tree
- register access fix for SJA1105P/Q/R/S for the virtual links through
  the dynamic reconfiguration interface
- handling 2 bridge VLANs where the second is supposed to overwrite the
  first
- making sure that the lack of a pvid results in the actual dropping of
  untagged traffic
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: update existing VLANs from the bridge VLAN list
Vladimir Oltean [Mon, 24 May 2021 09:25:27 +0000 (12:25 +0300)]
net: dsa: sja1105: update existing VLANs from the bridge VLAN list

When running this sequence of operations:

ip link add br0 type bridge vlan_filtering 1
ip link set swp4 master br0
bridge vlan add dev swp4 vid 1

We observe the traffic sent on swp4 is still untagged, even though the
bridge has overwritten the existing VLAN entry:

port    vlan ids
swp4     1 PVID

br0      1 PVID Egress Untagged

This happens because we didn't consider that the 'bridge vlan add'
command just overwrites VLANs like it's nothing. We treat the 'vid 1
pvid untagged' and the 'vid 1' as two separate VLANs, and the first
still has precedence when calling sja1105_build_vlan_table. Obviously
there is a disagreement regarding semantics, and we end up doing
something unexpected from the PoV of the bridge.

Let's actually consider an "existing VLAN" to be one which is on the
same port, and has the same VLAN ID, as one we already have, and update
it if it has different flags than we do.

The first blamed commit is the one introducing the bug, the second one
is the latest on top of which the bugfix still applies.

Fixes: ec5ae61076d0 ("net: dsa: sja1105: save/restore VLANs using a delta commit method")
Fixes: 5899ee367ab3 ("net: dsa: tag_8021q: add a context structure")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: use 4095 as the private VLAN for untagged traffic
Vladimir Oltean [Mon, 24 May 2021 09:25:26 +0000 (12:25 +0300)]
net: dsa: sja1105: use 4095 as the private VLAN for untagged traffic

One thing became visible when writing the blamed commit, and that was
that STP and PTP frames injected by net/dsa/tag_sja1105.c using the
deferred xmit mechanism are always classified to the pvid of the CPU
port, regardless of whatever VLAN there might be in these packets.

So a decision needed to be taken regarding the mechanism through which
we should ensure that delivery of STP and PTP traffic is possible when
we are in a VLAN awareness mode that involves tag_8021q. This is because
tag_8021q is not concerned with managing the pvid of the CPU port, since
as far as tag_8021q is concerned, no traffic should be sent as untagged
from the CPU port. So we end up not actually having a pvid on the CPU
port if we only listen to tag_8021q, and unless we do something about it.

The decision taken at the time was to keep VLAN 1 in the list of
priv->dsa_8021q_vlans, and make it a pvid of the CPU port. This ensures
that STP and PTP frames can always be sent to the outside world.

However there is a problem. If we do the following while we are in
the best_effort_vlan_filtering=true mode:

ip link add br0 type bridge vlan_filtering 1
ip link set swp2 master br0
bridge vlan del dev swp2 vid 1

Then untagged and pvid-tagged frames should be dropped. But we observe
that they aren't, and this is because of the precaution we took that VID
1 is always installed on all ports.

So clearly VLAN 1 is not good for this purpose. What about VLAN 0?
Well, VLAN 0 is managed by the 8021q module, and that module wants to
ensure that 802.1p tagged frames are always received by a port, and are
always transmitted as VLAN-tagged (with VLAN ID 0). Whereas we want our
STP and PTP frames to be untagged if the stack sent them as untagged -
we don't want the driver to just decide out of the blue that it adds
VID 0 to some packets.

So what to do?

Well, there is one other VLAN that is reserved, and that is 4095:
$ ip link add link swp2 name swp2.4095 type vlan id 4095
Error: 8021q: Invalid VLAN id.
$ bridge vlan add dev swp2 vid 4095
Error: bridge: Vlan id is invalid.

After we made this change, VLAN 1 is indeed forwarded and/or dropped
according to the bridge VLAN table, there are no further alterations
done by the sja1105 driver.

Fixes: ec5ae61076d0 ("net: dsa: sja1105: save/restore VLANs using a delta commit method")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: error out on unsupported PHY mode
Vladimir Oltean [Mon, 24 May 2021 09:25:25 +0000 (12:25 +0300)]
net: dsa: sja1105: error out on unsupported PHY mode

The driver continues probing when a port is configured for an
unsupported PHY interface type, instead it should stop.

Fixes: 8aa9ebccae87 ("net: dsa: Introduce driver for NXP SJA1105 5-port L2 switch")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: add error handling in sja1105_setup()
Vladimir Oltean [Mon, 24 May 2021 09:25:24 +0000 (12:25 +0300)]
net: dsa: sja1105: add error handling in sja1105_setup()

If any of sja1105_static_config_load(), sja1105_clocking_setup() or
sja1105_devlink_setup() fails, we can't just return in the middle of
sja1105_setup() or memory will leak. Add a cleanup path.

Fixes: 0a7bdbc23d8a ("net: dsa: sja1105: move devlink param code to sja1105_devlink.c")
Fixes: 8aa9ebccae87 ("net: dsa: Introduce driver for NXP SJA1105 5-port L2 switch")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: call dsa_unregister_switch when allocating memory fails
Vladimir Oltean [Mon, 24 May 2021 09:25:23 +0000 (12:25 +0300)]
net: dsa: sja1105: call dsa_unregister_switch when allocating memory fails

Unlike other drivers which pretty much end their .probe() execution with
dsa_register_switch(), the sja1105 does some extra stuff. When that
fails with -ENOMEM, the driver is quick to return that, forgetting to
call dsa_unregister_switch(). Not critical, but a bug nonetheless.

Fixes: 4d7525085a9b ("net: dsa: sja1105: offload the Credit-Based Shaper qdisc")
Fixes: a68578c20a96 ("net: dsa: Make deferred_xmit private to sja1105")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: sja1105: fix VL lookup command packing for P/Q/R/S
Vladimir Oltean [Mon, 24 May 2021 09:25:22 +0000 (12:25 +0300)]
net: dsa: sja1105: fix VL lookup command packing for P/Q/R/S

At the beginning of the sja1105_dynamic_config.c file there is a diagram
of the dynamic config interface layout:

 packed_buf

 |
 V
 +-----------------------------------------+------------------+
 |              ENTRY BUFFER               |  COMMAND BUFFER  |
 +-----------------------------------------+------------------+

 <----------------------- packed_size ------------------------>

So in order to pack/unpack the command bits into the buffer,
sja1105_vl_lookup_cmd_packing must first advance the buffer pointer by
the length of the entry. This is similar to what the other *cmd_packing
functions do.

This bug exists because the command packing function for P/Q/R/S was
copied from the E/T generation, and on E/T, the command was actually
embedded within the entry buffer itself.

Fixes: 94f94d4acfb2 ("net: dsa: sja1105: add static tables for virtual links")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hso: fix control-request directions
Johan Hovold [Mon, 24 May 2021 09:25:11 +0000 (11:25 +0200)]
net: hso: fix control-request directions

The direction of the pipe argument must match the request-type direction
bit or control requests may fail depending on the host-controller-driver
implementation.

Fix the tiocmset and rfkill requests which erroneously used
usb_rcvctrlpipe().

Fixes: 72dc1c096c70 ("HSO: add option hso driver")
Cc: stable@vger.kernel.org # 2.6.27
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agor8152: check the informaton of the device
Hayes Wang [Mon, 24 May 2021 06:49:42 +0000 (14:49 +0800)]
r8152: check the informaton of the device

Verify some fields of the USB descriptor to make sure the driver
could be used by the device.

Besides, remove the check of endpoint number in rtl8152_probe().
usb_find_common_endpoints() includes it.

BugLink: https://syzkaller.appspot.com/bug?id=912c9c373656996801b4de61f1e3cb326fe940aa
Reported-by: syzbot+95afd23673f5dd295c57@syzkaller.appspotmail.com
Fixes: c2198943e33b ("r8152: search the configuration of vendor mode")
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Reviewed-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agosch_dsmark: fix a NULL deref in qdisc_reset()
Taehee Yoo [Sun, 23 May 2021 14:38:53 +0000 (14:38 +0000)]
sch_dsmark: fix a NULL deref in qdisc_reset()

If Qdisc_ops->init() is failed, Qdisc_ops->reset() would be called.
When dsmark_init(Qdisc_ops->init()) is failed, it possibly doesn't
initialize dsmark_qdisc_data->q. But dsmark_reset(Qdisc_ops->reset())
uses dsmark_qdisc_data->q pointer wihtout any null checking.
So, panic would occur.

Test commands:
    sysctl net.core.default_qdisc=dsmark -w
    ip link add dummy0 type dummy
    ip link add vw0 link dummy0 type virt_wifi
    ip link set vw0 up

Splat looks like:
KASAN: null-ptr-deref in range [0x0000000000000018-0x000000000000001f]
CPU: 3 PID: 684 Comm: ip Not tainted 5.12.0+ #910
RIP: 0010:qdisc_reset+0x2b/0x680
Code: 1f 44 00 00 48 b8 00 00 00 00 00 fc ff df 41 57 41 56 41 55 41 54
55 48 89 fd 48 83 c7 18 53 48 89 fa 48 c1 ea 03 48 83 ec 20 <80> 3c 02
00 0f 85 09 06 00 00 4c 8b 65 18 0f 1f 44 00 00 65 8b 1d
RSP: 0018:ffff88800fda6bf8 EFLAGS: 00010282
RAX: dffffc0000000000 RBX: ffff8880050ed800 RCX: 0000000000000000
RDX: 0000000000000003 RSI: ffffffff99e34100 RDI: 0000000000000018
RBP: 0000000000000000 R08: fffffbfff346b553 R09: fffffbfff346b553
R10: 0000000000000001 R11: fffffbfff346b552 R12: ffffffffc0824940
R13: ffff888109e83800 R14: 00000000ffffffff R15: ffffffffc08249e0
FS:  00007f5042287680(0000) GS:ffff888119800000(0000)
knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000055ae1f4dbd90 CR3: 0000000006760002 CR4: 00000000003706e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 ? rcu_read_lock_bh_held+0xa0/0xa0
 dsmark_reset+0x3d/0xf0 [sch_dsmark]
 qdisc_reset+0xa9/0x680
 qdisc_destroy+0x84/0x370
 qdisc_create_dflt+0x1fe/0x380
 attach_one_default_qdisc.constprop.41+0xa4/0x180
 dev_activate+0x4d5/0x8c0
 ? __dev_open+0x268/0x390
 __dev_open+0x270/0x390

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoNFC: nfcmrvl: fix kernel-doc syntax in file headers
Aditya Srivastava [Sun, 23 May 2021 21:09:09 +0000 (02:39 +0530)]
NFC: nfcmrvl: fix kernel-doc syntax in file headers

The opening comment mark '/**' is used for highlighting the beginning of
kernel-doc comments.
The header for drivers/nfc/nfcmrvl follows this syntax, but the content
inside does not comply with kernel-doc.

This line was probably not meant for kernel-doc parsing, but is parsed
due to the presence of kernel-doc like comment syntax(i.e, '/**'), which
causes unexpected warnings from kernel-doc.
For e.g., running scripts/kernel-doc -none on drivers/nfc/nfcmrvl/spi.c
causes warning:
warning: expecting prototype for Marvell NFC(). Prototype was for SPI_WAIT_HANDSHAKE() instead

Provide a simple fix by replacing such occurrences with general comment
format, i.e. '/*', to prevent kernel-doc from parsing it.

Signed-off-by: Aditya Srivastava <yashsri421@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: mt7530: fix VLAN traffic leaks
DENG Qingfang [Sun, 23 May 2021 14:51:54 +0000 (22:51 +0800)]
net: dsa: mt7530: fix VLAN traffic leaks

PCR_MATRIX field was set to all 1's when VLAN filtering is enabled, but
was not reset when it is disabled, which may cause traffic leaks:

ip link add br0 type bridge vlan_filtering 1
ip link add br1 type bridge vlan_filtering 1
ip link set swp0 master br0
ip link set swp1 master br1
ip link set br0 type bridge vlan_filtering 0
ip link set br1 type bridge vlan_filtering 0
# traffic in br0 and br1 will start leaking to each other

As port_bridge_{add,del} have set up PCR_MATRIX properly, remove the
PCR_MATRIX write from mt7530_port_set_vlan_aware.

Fixes: 83163f7dca56 ("net: dsa: mediatek: add VLAN support for MT7530")
Signed-off-by: DENG Qingfang <dqfext@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'fq_pie-fixes'
David S. Miller [Mon, 24 May 2021 00:16:09 +0000 (17:16 -0700)]
Merge branch 'fq_pie-fixes'

Davide Caratti says:

====================
two fixes for the fq_pie scheduler

- patch 1/2 restores the possibility to use 65536 flows with fq_pie,
  preserving the fix for an endless loop in the control plane
- patch 2/2 fixes an OOB access that can be observed in the traffic
  path of fq_pie scheduler, when the classification selects a flow
  beyond the allocated space.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet/sched: fq_pie: fix OOB access in the traffic path
Davide Caratti [Sat, 22 May 2021 13:15:13 +0000 (15:15 +0200)]
net/sched: fq_pie: fix OOB access in the traffic path

the following script:

  # tc qdisc add dev eth0 handle 0x1 root fq_pie flows 2
  # tc qdisc add dev eth0 clsact
  # tc filter add dev eth0 egress matchall action skbedit priority 0x10002
  # ping 192.0.2.2 -I eth0 -c2 -w1 -q

produces the following splat:

 BUG: KASAN: slab-out-of-bounds in fq_pie_qdisc_enqueue+0x1314/0x19d0 [sch_fq_pie]
 Read of size 4 at addr ffff888171306924 by task ping/942

 CPU: 3 PID: 942 Comm: ping Not tainted 5.12.0+ #441
 Hardware name: Red Hat KVM, BIOS 1.11.1-4.module+el8.1.0+4066+0f1aadab 04/01/2014
 Call Trace:
  dump_stack+0x92/0xc1
  print_address_description.constprop.7+0x1a/0x150
  kasan_report.cold.13+0x7f/0x111
  fq_pie_qdisc_enqueue+0x1314/0x19d0 [sch_fq_pie]
  __dev_queue_xmit+0x1034/0x2b10
  ip_finish_output2+0xc62/0x2120
  __ip_finish_output+0x553/0xea0
  ip_output+0x1ca/0x4d0
  ip_send_skb+0x37/0xa0
  raw_sendmsg+0x1c4b/0x2d00
  sock_sendmsg+0xdb/0x110
  __sys_sendto+0x1d7/0x2b0
  __x64_sys_sendto+0xdd/0x1b0
  do_syscall_64+0x3c/0x80
  entry_SYSCALL_64_after_hwframe+0x44/0xae
 RIP: 0033:0x7fe69735c3eb
 Code: 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 f3 0f 1e fa 48 8d 05 75 42 2c 00 41 89 ca 8b 00 85 c0 75 14 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 75 c3 0f 1f 40 00 41 57 4d 89 c7 41 56 41 89
 RSP: 002b:00007fff06d7fb38 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
 RAX: ffffffffffffffda RBX: 000055e961413700 RCX: 00007fe69735c3eb
 RDX: 0000000000000040 RSI: 000055e961413700 RDI: 0000000000000003
 RBP: 0000000000000040 R08: 000055e961410500 R09: 0000000000000010
 R10: 0000000000000000 R11: 0000000000000246 R12: 00007fff06d81260
 R13: 00007fff06d7fb40 R14: 00007fff06d7fc30 R15: 000055e96140f0a0

 Allocated by task 917:
  kasan_save_stack+0x19/0x40
  __kasan_kmalloc+0x7f/0xa0
  __kmalloc_node+0x139/0x280
  fq_pie_init+0x555/0x8e8 [sch_fq_pie]
  qdisc_create+0x407/0x11b0
  tc_modify_qdisc+0x3c2/0x17e0
  rtnetlink_rcv_msg+0x346/0x8e0
  netlink_rcv_skb+0x120/0x380
  netlink_unicast+0x439/0x630
  netlink_sendmsg+0x719/0xbf0
  sock_sendmsg+0xe2/0x110
  ____sys_sendmsg+0x5ba/0x890
  ___sys_sendmsg+0xe9/0x160
  __sys_sendmsg+0xd3/0x170
  do_syscall_64+0x3c/0x80
  entry_SYSCALL_64_after_hwframe+0x44/0xae

 The buggy address belongs to the object at ffff888171306800
  which belongs to the cache kmalloc-256 of size 256
 The buggy address is located 36 bytes to the right of
  256-byte region [ffff888171306800ffff888171306900)
 The buggy address belongs to the page:
 page:00000000bcfb624e refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x171306
 head:00000000bcfb624e order:1 compound_mapcount:0
 flags: 0x17ffffc0010200(slab|head|node=0|zone=2|lastcpupid=0x1fffff)
 raw: 0017ffffc0010200 dead000000000100 dead000000000122 ffff888100042b40
 raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
 page dumped because: kasan: bad access detected

 Memory state around the buggy address:
  ffff888171306800: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  ffff888171306880: 00 00 00 00 00 00 00 00 00 00 00 00 fc fc fc fc
 >ffff888171306900: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
                                ^
  ffff888171306980: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
  ffff888171306a00: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb

fix fq_pie traffic path to avoid selecting 'q->flows + q->flows_cnt' as a
valid flow: it's an address beyond the allocated memory.

Fixes: ec97ecf1ebe4 ("net: sched: add Flow Queue PIE packet scheduler")
CC: stable@vger.kernel.org
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet/sched: fq_pie: re-factor fix for fq_pie endless loop
Davide Caratti [Sat, 22 May 2021 13:14:45 +0000 (15:14 +0200)]
net/sched: fq_pie: re-factor fix for fq_pie endless loop

the patch that fixed an endless loop in_fq_pie_init() was not considering
that 65535 is a valid class id. The correct bugfix for this infinite loop
is to change 'idx' to become an u32, like Colin proposed in the past [1].

Fix this as follows:
 - restore 65536 as maximum possible values of 'flows_cnt'
 - use u32 'idx' when iterating on 'q->flows'
 - fix the TDC selftest

This reverts commit bb2f930d6dd708469a587dc9ed1efe1ef969c0bf.

[1] https://lore.kernel.org/netdev/20210407163808.499027-1-colin.king@canonical.com/

CC: Colin Ian King <colin.king@canonical.com>
CC: stable@vger.kernel.org
Fixes: bb2f930d6dd7 ("net/sched: fix infinite loop in sch_fq_pie")
Fixes: ec97ecf1ebe4 ("net: sched: add Flow Queue PIE packet scheduler")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: macb: ensure the device is available before accessing GEMGXL control registers
Zong Li [Sat, 22 May 2021 09:16:11 +0000 (17:16 +0800)]
net: macb: ensure the device is available before accessing GEMGXL control registers

If runtime power menagement is enabled, the gigabit ethernet PLL would
be disabled after macb_probe(). During this period of time, the system
would hang up if we try to access GEMGXL control registers.

We can't put runtime_pm_get/runtime_pm_put/ there due to the issue of
sleep inside atomic section (7fa2955ff70ce453 ("sh_eth: Fix sleeping
function called from invalid context"). Add netif_running checking to
ensure the device is available before accessing GEMGXL device.

Changed in v2:
 - Use netif_running instead of its own flag

Signed-off-by: Zong Li <zong.li@sifive.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: ethernet: mtk_eth_soc: Fix packet statistics support for MT7628/88
Stefan Roese [Sat, 22 May 2021 07:56:30 +0000 (09:56 +0200)]
net: ethernet: mtk_eth_soc: Fix packet statistics support for MT7628/88

The MT7628/88 SoC(s) have other (limited) packet counter registers than
currently supported in the mtk_eth_soc driver. This patch adds support
for reading these registers, so that the packet statistics are correctly
updated.

Additionally the defines for the non-MT7628 variant packet counter
registers are added and used in this patch instead of using hard coded
values.

Signed-off-by: Stefan Roese <sr@denx.de>
Fixes: 296c9120752b ("net: ethernet: mediatek: Add MT7628/88 SoC support")
Cc: Felix Fietkau <nbd@nbd.name>
Cc: John Crispin <john@phrozen.org>
Cc: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Cc: Reto Schneider <code@reto-schneider.ch>
Cc: Reto Schneider <reto.schneider@husqvarnagroup.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMAINTAINERS: Add entries for CBS, ETF and taprio qdiscs
Vinicius Costa Gomes [Sat, 22 May 2021 00:46:54 +0000 (17:46 -0700)]
MAINTAINERS: Add entries for CBS, ETF and taprio qdiscs

Add Vinicius Costa Gomes as maintainer for these qdiscs.

These qdiscs are all TSN (Time Sensitive Networking) related.

Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Acked-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoipv6: record frag_max_size in atomic fragments in input path
Francesco Ruggeri [Fri, 21 May 2021 20:21:14 +0000 (13:21 -0700)]
ipv6: record frag_max_size in atomic fragments in input path

Commit dbd1759e6a9c ("ipv6: on reassembly, record frag_max_size")
filled the frag_max_size field in IP6CB in the input path.
The field should also be filled in case of atomic fragments.

Fixes: dbd1759e6a9c ('ipv6: on reassembly, record frag_max_size')
Signed-off-by: Francesco Ruggeri <fruggeri@arista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet:sfc: fix non-freed irq in legacy irq mode
Íñigo Huguet [Fri, 21 May 2021 14:38:35 +0000 (16:38 +0200)]
net:sfc: fix non-freed irq in legacy irq mode

SFC driver can be configured via modparam to work using MSI-X, MSI or
legacy IRQ interrupts. In the last one, the interrupt was not properly
released on module remove.

It was not freed because the flag irqs_hooked was not set during
initialization in the case of using legacy IRQ.

Example of (trimmed) trace during module remove without this fix:

remove_proc_entry: removing non-empty directory 'irq/125', leaking at least '0000:3b:00.1'
WARNING: CPU: 39 PID: 3658 at fs/proc/generic.c:715 remove_proc_entry+0x15c/0x170
...trimmed...
Call Trace:
 unregister_irq_proc+0xe3/0x100
 free_desc+0x29/0x70
 irq_free_descs+0x47/0x70
 mp_unmap_irq+0x58/0x60
 acpi_unregister_gsi_ioapic+0x2a/0x40
 acpi_pci_irq_disable+0x78/0xb0
 pci_disable_device+0xd1/0x100
 efx_pci_remove+0xa1/0x1e0 [sfc]
 pci_device_remove+0x38/0xa0
 __device_release_driver+0x177/0x230
 driver_detach+0xcb/0x110
 bus_remove_driver+0x58/0xd0
 pci_unregister_driver+0x2a/0xb0
 efx_exit_module+0x24/0xf40 [sfc]
 __do_sys_delete_module.constprop.0+0x171/0x280
 ? exit_to_user_mode_prepare+0x83/0x1d0
 do_syscall_64+0x3d/0x80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7f9f9385800b
...trimmed...

Signed-off-by: Íñigo Huguet <ihuguet@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoRDS tcp loopback connection can hang
Rao Shoaib [Fri, 21 May 2021 18:08:06 +0000 (11:08 -0700)]
RDS tcp loopback connection can hang

When TCP is used as transport and a program on the
system connects to RDS port 16385, connection is
accepted but denied per the rules of RDS. However,
RDS connections object is left in the list. Next
loopback connection will select that connection
object as it is at the head of list. The connection
attempt will hang as the connection object is set
to connect over TCP which is not allowed

The issue can be reproduced easily, use rds-ping
to ping a local IP address. After that use any
program like ncat to connect to the same IP
address and port 16385. This will hang so ctrl-c out.
Now try rds-ping, it will hang.

To fix the issue this patch adds checks to disallow
the connection object creation and destroys the
connection object.

Signed-off-by: Rao Shoaib <rao.shoaib@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMAINTAINERS: remove Ioana Radulescu from dpaa2-eth
Ioana Ciornei [Fri, 21 May 2021 21:01:00 +0000 (00:01 +0300)]
MAINTAINERS: remove Ioana Radulescu from dpaa2-eth

Remove Ioana Radulescu from dpaa2-eth since she is no longer working on
the DPAA2 set of drivers.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: lantiq: fix memory corruption in RX ring
Aleksander Jan Bajkowski [Fri, 21 May 2021 14:45:58 +0000 (16:45 +0200)]
net: lantiq: fix memory corruption in RX ring

In a situation where memory allocation or dma mapping fails, an
invalid address is programmed into the descriptor. This can lead
to memory corruption. If the memory allocation fails, DMA should
reuse the previous skb and mapping and drop the packet. This patch
also increments rx drop counter.

Fixes: fe1a56420cf2 ("net: lantiq: Add Lantiq / Intel VRX200 Ethernet driver ")
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: dsa: bcm_sf2: Fix bcm_sf2_reg_rgmii_cntrl() call for non-RGMII port
Florian Fainelli [Fri, 21 May 2021 17:46:14 +0000 (10:46 -0700)]
net: dsa: bcm_sf2: Fix bcm_sf2_reg_rgmii_cntrl() call for non-RGMII port

We cannot call bcm_sf2_reg_rgmii_cntrl() for a port that is not RGMII,
yet we do that in bcm_sf2_sw_mac_link_up() irrespective of the port's
interface. Move that read until we have properly qualified the PHY
interface mode. This avoids triggering a warning on 7278 platforms that
have GMII ports.

Fixes: 55cfeb396965 ("net: dsa: bcm_sf2: add function finding RGMII register")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMAINTAINERS: s390/net: add netdev list
Julian Wiedmann [Fri, 21 May 2021 13:28:56 +0000 (15:28 +0200)]
MAINTAINERS: s390/net: add netdev list

Discussions for network-related code should include the netdev list.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: ethernet: mtk_eth_soc: Fix DIM support for MT7628/88
Stefan Roese [Thu, 20 May 2021 08:43:18 +0000 (10:43 +0200)]
net: ethernet: mtk_eth_soc: Fix DIM support for MT7628/88

When updating to latest mainline for some testing on the GARDENA smart
gateway based on the MT7628, I noticed that ethernet does not work any
more. Commit e9229ffd550b ("net: ethernet: mtk_eth_soc: implement
dynamic interrupt moderation") introduced this problem, as it missed the
RX_DIM & TX_DIM configuration for this SoC variant. This patch fixes
this by calling mtk_dim_rx() & mtk_dim_tx() in this case as well.

Signed-off-by: Stefan Roese <sr@denx.de>
Fixes: e9229ffd550b ("net: ethernet: mtk_eth_soc: implement dynamic interrupt moderation")
Cc: Felix Fietkau <nbd@nbd.name>
Cc: John Crispin <john@phrozen.org>
Cc: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Cc: Reto Schneider <code@reto-schneider.ch>
Cc: Reto Schneider <reto.schneider@husqvarnagroup.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'stmmac-fixes'
David S. Miller [Thu, 20 May 2021 22:49:33 +0000 (15:49 -0700)]
Merge branch 'stmmac-fixes'

Joakim Zhang says:

====================
net: fixes for stmmac

Two clock fixes for stmmac driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: stmmac: fix system hang if change mac address after interface ifdown
Joakim Zhang [Thu, 20 May 2021 12:51:17 +0000 (20:51 +0800)]
net: stmmac: fix system hang if change mac address after interface ifdown

Fix system hang with below sequences:
~# ifconfig ethx down
~# ifconfig ethx hw ether xx:xx:xx:xx:xx:xx

After ethx down, stmmac all clocks gated off and then register access causes
system hang.

Fixes: 5ec55823438e ("net: stmmac: add clocks management for gmac driver")
Signed-off-by: Joakim Zhang <qiangqing.zhang@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: stmmac: correct clocks enabled in stmmac_vlan_rx_kill_vid()
Joakim Zhang [Thu, 20 May 2021 12:51:16 +0000 (20:51 +0800)]
net: stmmac: correct clocks enabled in stmmac_vlan_rx_kill_vid()

This should be a mistake to fix conflicts when removing RFC tag to
repost the patch.

Fixes: 5ec55823438e ("net: stmmac: add clocks management for gmac driver")
Signed-off-by: Joakim Zhang <qiangqing.zhang@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet/qla3xxx: fix schedule while atomic in ql_sem_spinlock
Zheyu Ma [Thu, 20 May 2021 12:32:36 +0000 (12:32 +0000)]
net/qla3xxx: fix schedule while atomic in ql_sem_spinlock

When calling the 'ql_sem_spinlock', the driver has already acquired the
spin lock, so the driver should not call 'ssleep' in atomic context.

This bug can be fixed by using 'mdelay' instead of 'ssleep'.

The KASAN's log reveals it:

[    3.238124 ] BUG: scheduling while atomic: swapper/0/1/0x00000002
[    3.238748 ] 2 locks held by swapper/0/1:
[    3.239151 ]  #0: ffff88810177b240 (&dev->mutex){....}-{3:3}, at:
__device_driver_lock+0x41/0x60
[    3.240026 ]  #1: ffff888107c60e28 (&qdev->hw_lock){....}-{2:2}, at:
ql3xxx_probe+0x2aa/0xea0
[    3.240873 ] Modules linked in:
[    3.241187 ] irq event stamp: 460854
[    3.241541 ] hardirqs last  enabled at (460853): [<ffffffff843051bf>]
_raw_spin_unlock_irqrestore+0x4f/0x70
[    3.242245 ] hardirqs last disabled at (460854): [<ffffffff843058ca>]
_raw_spin_lock_irqsave+0x2a/0x70
[    3.242245 ] softirqs last  enabled at (446076): [<ffffffff846002e4>]
__do_softirq+0x2e4/0x4b1
[    3.242245 ] softirqs last disabled at (446069): [<ffffffff811ba5e0>]
irq_exit_rcu+0x100/0x110
[    3.242245 ] Preemption disabled at:
[    3.242245 ] [<ffffffff828ca5ba>] ql3xxx_probe+0x2aa/0xea0
[    3.242245 ] Kernel panic - not syncing: scheduling while atomic
[    3.242245 ] CPU: 2 PID: 1 Comm: swapper/0 Not tainted
5.13.0-rc1-00145
-gee7dc339169-dirty #16
[    3.242245 ] Call Trace:
[    3.242245 ]  dump_stack+0xba/0xf5
[    3.242245 ]  ? ql3xxx_probe+0x1f0/0xea0
[    3.242245 ]  panic+0x15a/0x3f2
[    3.242245 ]  ? vprintk+0x76/0x150
[    3.242245 ]  ? ql3xxx_probe+0x2aa/0xea0
[    3.242245 ]  __schedule_bug+0xae/0xe0
[    3.242245 ]  __schedule+0x72e/0xa00
[    3.242245 ]  schedule+0x43/0xf0
[    3.242245 ]  schedule_timeout+0x28b/0x500
[    3.242245 ]  ? del_timer_sync+0xf0/0xf0
[    3.242245 ]  ? msleep+0x2f/0x70
[    3.242245 ]  msleep+0x59/0x70
[    3.242245 ]  ql3xxx_probe+0x307/0xea0
[    3.242245 ]  ? _raw_spin_unlock_irqrestore+0x3a/0x70
[    3.242245 ]  ? pci_device_remove+0x110/0x110
[    3.242245 ]  local_pci_probe+0x45/0xa0
[    3.242245 ]  pci_device_probe+0x12b/0x1d0
[    3.242245 ]  really_probe+0x2a9/0x610
[    3.242245 ]  driver_probe_device+0x90/0x1d0
[    3.242245 ]  ? mutex_lock_nested+0x1b/0x20
[    3.242245 ]  device_driver_attach+0x68/0x70
[    3.242245 ]  __driver_attach+0x124/0x1b0
[    3.242245 ]  ? device_driver_attach+0x70/0x70
[    3.242245 ]  bus_for_each_dev+0xbb/0x110
[    3.242245 ]  ? rdinit_setup+0x45/0x45
[    3.242245 ]  driver_attach+0x27/0x30
[    3.242245 ]  bus_add_driver+0x1eb/0x2a0
[    3.242245 ]  driver_register+0xa9/0x180
[    3.242245 ]  __pci_register_driver+0x82/0x90
[    3.242245 ]  ? yellowfin_init+0x25/0x25
[    3.242245 ]  ql3xxx_driver_init+0x23/0x25
[    3.242245 ]  do_one_initcall+0x7f/0x3d0
[    3.242245 ]  ? rdinit_setup+0x45/0x45
[    3.242245 ]  ? rcu_read_lock_sched_held+0x4f/0x80
[    3.242245 ]  kernel_init_freeable+0x2aa/0x301
[    3.242245 ]  ? rest_init+0x2c0/0x2c0
[    3.242245 ]  kernel_init+0x18/0x190
[    3.242245 ]  ? rest_init+0x2c0/0x2c0
[    3.242245 ]  ? rest_init+0x2c0/0x2c0
[    3.242245 ]  ret_from_fork+0x1f/0x30
[    3.242245 ] Dumping ftrace buffer:
[    3.242245 ]    (ftrace buffer empty)
[    3.242245 ] Kernel Offset: disabled
[    3.242245 ] Rebooting in 1 seconds.

Reported-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: encx24j600: fix kernel-doc syntax in file headers
Aditya Srivastava [Thu, 20 May 2021 18:49:15 +0000 (00:19 +0530)]
net: encx24j600: fix kernel-doc syntax in file headers

The opening comment mark '/**' is used for highlighting the beginning of
kernel-doc comments.
The header for drivers/net/ethernet/microchip/encx24j600 files follows
this syntax, but the content inside does not comply with kernel-doc.

This line was probably not meant for kernel-doc parsing, but is parsed
due to the presence of kernel-doc like comment syntax(i.e, '/**'), which
causes unexpected warning from kernel-doc.
For e.g., running scripts/kernel-doc -none
drivers/net/ethernet/microchip/encx24j600_hw.h emits:
warning: expecting prototype for h(). Prototype was for _ENCX24J600_HW_H() instead

Provide a simple fix by replacing such occurrences with general comment
format, i.e. '/*', to prevent kernel-doc from parsing it.

Signed-off-by: Aditya Srivastava <yashsri421@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoixgbe: fix large MTU request from VF
Jesse Brandeburg [Thu, 20 May 2021 18:18:35 +0000 (11:18 -0700)]
ixgbe: fix large MTU request from VF

Check that the MTU value requested by the VF is in the supported
range of MTUs before attempting to set the VF large packet enable,
otherwise reject the request. This also avoids unnecessary
register updates in the case of the 82599 controller.

Fixes: 872844ddb9e4 ("ixgbe: Enable jumbo frames support w/ SR-IOV")
Co-developed-by: Piotr Skajewski <piotrx.skajewski@intel.com>
Signed-off-by: Piotr Skajewski <piotrx.skajewski@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Co-developed-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Signed-off-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agobpf: Set mac_len in bpf_skb_change_head
Jussi Maki [Wed, 19 May 2021 15:47:42 +0000 (15:47 +0000)]
bpf: Set mac_len in bpf_skb_change_head

The skb_change_head() helper did not set "skb->mac_len", which is
problematic when it's used in combination with skb_redirect_peer().
Without it, redirecting a packet from a L3 device such as wireguard to
the veth peer device will cause skb->data to point to the middle of the
IP header on entry to tcp_v4_rcv() since the L2 header is not pulled
correctly due to mac_len=0.

Fixes: 3a0af8fd61f9 ("bpf: BPF for lightweight tunnel infrastructure")
Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210519154743.2554771-2-joamaki@gmail.com
3 years agobpf, offload: Reorder offload callback 'prepare' in verifier
Yinjun Zhang [Thu, 20 May 2021 08:58:34 +0000 (10:58 +0200)]
bpf, offload: Reorder offload callback 'prepare' in verifier

Commit 4976b718c355 ("bpf: Introduce pseudo_btf_id") switched the
order of resolve_pseudo_ldimm(), in which some pseudo instructions
are rewritten. Thus those rewritten instructions cannot be passed
to driver via 'prepare' offload callback.

Reorder the 'prepare' offload callback to fix it.

Fixes: 4976b718c355 ("bpf: Introduce pseudo_btf_id")
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20210520085834.15023-1-simon.horman@netronome.com
3 years agobpf: Avoid using ARRAY_SIZE on an uninitialized pointer
Florent Revest [Mon, 17 May 2021 09:28:30 +0000 (11:28 +0200)]
bpf: Avoid using ARRAY_SIZE on an uninitialized pointer

The cppcheck static code analysis reported the following error:

    if (WARN_ON_ONCE(nest_level > ARRAY_SIZE(bufs->tmp_bufs))) {
                                             ^
ARRAY_SIZE is a macro that expands to sizeofs, so bufs is not actually
dereferenced at runtime, and the code is actually safe. But to keep
things tidy, this patch removes the need for a call to ARRAY_SIZE by
extracting the size of the array into a macro. Cppcheck should no longer
be confused and the code ends up being a bit cleaner.

Fixes: e2d5b2bb769f ("bpf: Fix nested bpf_bprintf_prepare with more per-cpu buffers")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Florent Revest <revest@chromium.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/bpf/20210517092830.1026418-2-revest@chromium.org
3 years agobpf: Clarify a bpf_bprintf_prepare macro
Florent Revest [Mon, 17 May 2021 09:28:29 +0000 (11:28 +0200)]
bpf: Clarify a bpf_bprintf_prepare macro

The per-cpu buffers contain bprintf data rather than printf arguments.
The macro name and comment were a bit confusing, this rewords them in a
clearer way.

Signed-off-by: Florent Revest <revest@chromium.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/bpf/20210517092830.1026418-1-revest@chromium.org
3 years agoselftests/bpf: Test ringbuf mmap read-only and read-write restrictions
Andrii Nakryiko [Fri, 14 May 2021 18:07:26 +0000 (11:07 -0700)]
selftests/bpf: Test ringbuf mmap read-only and read-write restrictions

Extend ringbuf selftest to validate read/write and read-only restrictions on
memory mapping consumer/producer/data pages. Ensure no "escalations" from
PROT_READ to PROT_WRITE/PROT_EXEC is allowed. And test that mremap() fails to
expand mmap()'ed area.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210514180726.843157-1-andrii@kernel.org
3 years agoselftests/bpf: Convert static to global in tc_redirect progs
Stanislav Fomichev [Fri, 14 May 2021 17:05:28 +0000 (10:05 -0700)]
selftests/bpf: Convert static to global in tc_redirect progs

Both IFINDEX_SRC and IFINDEX_DST are set from the userspace
and it won't work once bpf merges with bpf-next.

Fixes: 096eccdef0b3 ("selftests/bpf: Rewrite test_tc_redirect.sh as prog_tests/tc_redirect.c")
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210514170528.3750250-1-sdf@google.com
3 years agobpf: Fix BPF_JIT kconfig symbol dependency
Daniel Borkmann [Wed, 12 May 2021 18:57:14 +0000 (20:57 +0200)]
bpf: Fix BPF_JIT kconfig symbol dependency

Randy reported a randconfig build error recently on i386:

  ld: arch/x86/net/bpf_jit_comp32.o: in function `do_jit':
  bpf_jit_comp32.c:(.text+0x28c9): undefined reference to `__bpf_call_base'
  ld: arch/x86/net/bpf_jit_comp32.o: in function `bpf_int_jit_compile':
  bpf_jit_comp32.c:(.text+0x3694): undefined reference to `bpf_jit_blind_constants'
  ld: bpf_jit_comp32.c:(.text+0x3719): undefined reference to `bpf_jit_binary_free'
  ld: bpf_jit_comp32.c:(.text+0x3745): undefined reference to `bpf_jit_binary_alloc'
  ld: bpf_jit_comp32.c:(.text+0x37d3): undefined reference to `bpf_jit_prog_release_other'
  [...]

The cause was that b24abcff918a ("bpf, kconfig: Add consolidated menu entry for
bpf with core options") moved BPF_JIT from net/Kconfig into kernel/bpf/Kconfig
and previously BPF_JIT was guarded by a 'if NET'. However, there is no actual
dependency on NET, it's just that menuconfig NET selects BPF. And the latter in
turn causes kernel/bpf/core.o to be built which contains above symbols. Randy's
randconfig didn't have NET set, and BPF wasn't either, but BPF_JIT otoh was.
Detangle this by making BPF_JIT depend on BPF instead. arm64 was the only arch
that pulled in its JIT in net/ via obj-$(CONFIG_NET), all others unconditionally
pull this dir in via obj-y. Do the same since CONFIG_NET guard there is really
useless as we compiled the JIT via obj-$(CONFIG_BPF_JIT) += bpf_jit_comp.o anyway.

Fixes: b24abcff918a ("bpf, kconfig: Add consolidated menu entry for bpf with core options")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
3 years agoselftests: Add .gitignore for nci test suite
David Matlack [Wed, 19 May 2021 21:33:33 +0000 (21:33 +0000)]
selftests: Add .gitignore for nci test suite

Building the nci test suite produces a binary, nci_dev, that git then
tries to track. Add a .gitignore file to tell git to ignore this binary.

Signed-off-by: David Matlack <dmatlack@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hso: bail out on interrupt URB allocation failure
Johan Hovold [Wed, 19 May 2021 12:47:17 +0000 (14:47 +0200)]
net: hso: bail out on interrupt URB allocation failure

Commit 31db0dbd7244 ("net: hso: check for allocation failure in
hso_create_bulk_serial_device()") recently started returning an error
when the driver fails to allocate resources for the interrupt endpoint
and tiocmget functionality.

For consistency let's bail out from probe also if the URB allocation
fails.

Signed-off-by: Johan Hovold <johan@kernel.org>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agodt-bindings: net: renesas,ether: Update Sergei's email address
Geert Uytterhoeven [Wed, 19 May 2021 13:02:53 +0000 (15:02 +0200)]
dt-bindings: net: renesas,ether: Update Sergei's email address

Update Sergei's email address, as per commit 534a8bf0ccdd7b3f
("MAINTAINERS: switch to my private email for Renesas Ethernet
drivers").

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Sergei Shtylyov <sergei.shtylyov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agocxgb4: avoid accessing registers when clearing filters
Raju Rangoju [Wed, 19 May 2021 11:18:31 +0000 (16:48 +0530)]
cxgb4: avoid accessing registers when clearing filters

Hardware register having the server TID base can contain
invalid values when adapter is in bad state (for example,
due to AER fatal error). Reading these invalid values in the
register can lead to out-of-bound memory access. So, fix
by using the saved server TID base when clearing filters.

Fixes: b1a79360ee86 ("cxgb4: Delete all hash and TCAM filters before resource cleanup")
Signed-off-by: Raju Rangoju <rajur@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge tag 'mlx5-fixes-2021-05-18' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Wed, 19 May 2021 19:19:30 +0000 (12:19 -0700)]
Merge tag 'mlx5-fixes-2021-05-18' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2021-05-18

This series introduces some fixes to mlx5 driver.

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

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoethtool: stats: Fix a copy-paste error
YueHaibing [Wed, 19 May 2021 02:10:38 +0000 (10:10 +0800)]
ethtool: stats: Fix a copy-paste error

data->ctrl_stats should be memset with correct size.

Fixes: bfad2b979ddc ("ethtool: add interface to read standard MAC Ctrl stats")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agomlx5e: add add missing BH locking around napi_schdule()
Jakub Kicinski [Wed, 5 May 2021 20:20:26 +0000 (13:20 -0700)]
mlx5e: add add missing BH locking around napi_schdule()

It's not correct to call napi_schedule() in pure process
context. Because we use __raise_softirq_irqoff() we require
callers to be in a context which will eventually lead to
softirq handling (hardirq, bh disabled, etc.).

With code as is users will see:

 NOHZ tick-stop error: Non-RCU local softirq work is pending, handler #08!!!

Fixes: a8dd7ac12fc3 ("net/mlx5e: Generalize RQ activation")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Set term table as an unmanaged flow table
Ariel Levkovich [Wed, 31 Mar 2021 07:09:02 +0000 (10:09 +0300)]
net/mlx5: Set term table as an unmanaged flow table

Termination tables are restricted to have the default miss action and
cannot be set to forward to another table in case of a miss.
If the fs prio of the termination table is not the last one in the
list, fs_core will attempt to attach it to another table.

Set the unmanaged ft flag when creating the termination table ft
and select the tc offload prio for it to prevent fs_core from selecting
the forwarding to next ft miss action and use the default one.

In addition, set the flow that forwards to the termination table to
ignore ft level restrictions since the ft level is not set by fs_core
for unamanged fts.

Fixes: 249ccc3c95bd ("net/mlx5e: Add support for offloading traffic from uplink to uplink")
Signed-off-by: Ariel Levkovich <lariel@nvidia.com>
3 years agonet/mlx5: Don't overwrite HCA capabilities when setting MSI-X count
Leon Romanovsky [Sun, 25 Apr 2021 10:28:10 +0000 (13:28 +0300)]
net/mlx5: Don't overwrite HCA capabilities when setting MSI-X count

During driver probe of device that has dynamic MSI-X feature enabled,
the following error is printed in some FW flavour (not released yet).

 mlx5_core 0000:06:00.0: firmware version: 4.7.4387
 mlx5_core 0000:06:00.0: 126.016 Gb/s available PCIe bandwidth (8.0 GT/s PCIe x16 link)
 mlx5_core 0000:06:00.0: mlx5_cmd_check:777:(pid 70599): SET_HCA_CAP(0x109) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x0)
 mlx5_core 0000:06:00.0: set_hca_cap:622:(pid 70599): handle_hca_cap failed
 mlx5_core 0000:06:00.0: mlx5_function_setup:1045:(pid 70599): set_hca_cap failed
 mlx5_core 0000:06:00.0: probe_one:1465:(pid 70599): mlx5_init_one failed with error code -22
 mlx5_core: probe of 0000:06:00.0 failed with error -22

In order to make the setting capability of MSI-X future proof, let's
query the current capabilities first.

Fixes: 604774add516 ("net/mlx5: Dynamically assign MSI-X vectors count")
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years ago{net,vdpa}/mlx5: Configure interface MAC into mpfs L2 table
Eli Cohen [Thu, 22 Apr 2021 12:48:10 +0000 (15:48 +0300)]
{net,vdpa}/mlx5: Configure interface MAC into mpfs L2 table

net/mlx5: Expose MPFS configuration API

MPFS is the multi physical function switch that bridges traffic between
the physical port and any physical functions associated with it. The
driver is required to add or remove MAC entries to properly forward
incoming traffic to the correct physical function.

We export the API to control MPFS so that other drivers, such as
mlx5_vdpa are able to add MAC addresses of their network interfaces.

The MAC address of the vdpa interface must be configured into the MPFS L2
address. Failing to do so could cause, in some NIC configurations, failure
to forward packets to the vdpa network device instance.

Fix this by adding calls to update the MPFS table.

CC: <mst@redhat.com>
CC: <jasowang@redhat.com>
CC: <virtualization@lists.linux-foundation.org>
Fixes: 1a86b377aa21 ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix error path of updating netdev queues
Aya Levin [Wed, 21 Apr 2021 11:26:31 +0000 (14:26 +0300)]
net/mlx5e: Fix error path of updating netdev queues

Avoid division by zero in the error flow. In the driver TC number can be
either 1 or 8. When TC count is set to 1, driver zero netdev->num_tc.
Hence, need to convert it back from 0 to 1 in the error flow.

Fixes: fa3748775b92 ("net/mlx5e: Handle errors from netif_set_real_num_{tx,rx}_queues")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Reject mirroring on source port change encap rules
Vlad Buslov [Tue, 20 Apr 2021 12:16:16 +0000 (15:16 +0300)]
net/mlx5e: Reject mirroring on source port change encap rules

Rules with MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE dest flag are
translated to destination FT in eswitch. Currently it is not possible to
mirror such rules because firmware doesn't support mixing FT and Vport
destinations in single rule when one of them adds encapsulation. Since the
only use case for MLX5_ESW_DEST_CHAIN_WITH_SRC_PORT_CHANGE destination is
support for tunnel endpoints on VF and trying to offload such rule with
mirror action causes either crash in fs_core or firmware error with
syndrome 0xff6a1d, reject all such rules in mlx5 TC layer.

Fixes: 10742efc20a4 ("net/mlx5e: VF tunnel TX traffic offloading")
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix multipath lag activation
Dima Chumak [Tue, 13 Apr 2021 19:43:08 +0000 (22:43 +0300)]
net/mlx5e: Fix multipath lag activation

When handling FIB_EVENT_ENTRY_REPLACE event for a new multipath route,
lag activation can be missed if a stale (struct lag_mp)->mfi pointer
exists, which was associated with an older multipath route that had been
removed.

Normally, when a route is removed, it triggers mlx5_lag_fib_event(),
which handles FIB_EVENT_ENTRY_DEL and clears mfi pointer. But, if
mlx5_lag_check_prereq() condition isn't met, for example when eswitch is
in legacy mode, the fib event is skipped and mfi pointer becomes stale.

Fix by resetting mfi pointer to NULL every time mlx5_lag_mp_init() is
called.

Fixes: 544fe7c2e654 ("net/mlx5e: Activate HW multipath and handle port affinity based on FIB events")
Signed-off-by: Dima Chumak <dchumak@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: reset XPS on error flow if netdev isn't registered yet
Saeed Mahameed [Thu, 25 Feb 2021 19:20:00 +0000 (11:20 -0800)]
net/mlx5e: reset XPS on error flow if netdev isn't registered yet

mlx5e_attach_netdev can be called prior to registering the netdevice:
Example stack:

ipoib_new_child_link ->
ipoib_intf_init->
rdma_init_netdev->
mlx5_rdma_setup_rn->

mlx5e_attach_netdev->
mlx5e_num_channels_changed ->
mlx5e_set_default_xps_cpumasks ->
netif_set_xps_queue ->
__netif_set_xps_queue -> kmalloc

If any later stage fails at any point after mlx5e_num_channels_changed()
returns, XPS allocated maps will never be freed as they
are only freed during netdev unregistration, which will never happen for
yet to be registered netdevs.

Fixes: 3909a12e7913 ("net/mlx5e: Fix configuration of XPS cpumasks and netdev queues in corner cases")
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
3 years agonet/mlx5e: Make sure fib dev exists in fib event
Roi Dayan [Sun, 2 May 2021 07:25:50 +0000 (10:25 +0300)]
net/mlx5e: Make sure fib dev exists in fib event

For unreachable route entry the fib dev does not exists.

Fixes: 8914add2c9e5 ("net/mlx5e: Handle FIB events to update tunnel endpoint device")
Reported-by: Dennis Afanasev <dennis.afanasev@stateless.net>
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix null deref accessing lag dev
Roi Dayan [Mon, 3 May 2021 15:01:02 +0000 (18:01 +0300)]
net/mlx5e: Fix null deref accessing lag dev

It could be the lag dev is null so stop processing the event.
In bond_enslave() the active/backup slave being set before setting the
upper dev so first event is without an upper dev.
After setting the upper dev with bond_master_upper_dev_link() there is
a second event and in that event we have an upper dev.

Fixes: 7e51891a237f ("net/mlx5e: Use netdev events to set/del egress acl forward-to-vport rule")
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix nullptr in mlx5e_tc_add_fdb_flow()
Dima Chumak [Mon, 26 Apr 2021 12:16:26 +0000 (15:16 +0300)]
net/mlx5e: Fix nullptr in mlx5e_tc_add_fdb_flow()

The result of __dev_get_by_index() is not checked for NULL, which then
passed to mlx5e_attach_encap() and gets dereferenced.

Also, in case of a successful lookup, the net_device reference count is
not incremented, which may result in net_device pointer becoming invalid
at any time during mlx5e_attach_encap() execution.

Fix by using dev_get_by_index(), which does proper reference counting on
the net_device pointer. Also, handle nullptr return value when mirred
device is not found.

It's safe to call dev_put() on the mirred net_device pointer, right
after mlx5e_attach_encap() call, because it's not being saved/copied
down the call chain.

Fixes: 3c37745ec614 ("net/mlx5e: Properly deal with encap flows add/del under neigh update")
Addresses-Coverity: ("Dereference null return value")
Signed-off-by: Dima Chumak <dchumak@nvidia.com>
Reviewed-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: SF, Fix show state inactive when its inactivated
Parav Pandit [Fri, 7 May 2021 07:08:47 +0000 (10:08 +0300)]
net/mlx5: SF, Fix show state inactive when its inactivated

When a SF is inactivated and when it is in a TEARDOWN_REQUEST
state, driver still returns its state as active. This is incorrect.
Fix it by treating TEARDOWN_REQEUST as inactive state. When a SF
is still attached to the driver, on user request to reactivate EINVAL
error is returned. Inform user about it with better code EBUSY and
informative error message.

Fixes: 6a3273217469 ("net/mlx5: SF, Port function state change support")
Signed-off-by: Parav Pandit <parav@nvidia.com>
Reviewed-by: Vu Pham <vuhuong@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Fix err prints and return when creating termination table
Roi Dayan [Thu, 13 May 2021 12:00:53 +0000 (15:00 +0300)]
net/mlx5: Fix err prints and return when creating termination table

Fix print to print correct error code and not using IS_ERR() which
will just result in always printing 1.
Also return real err instead of always -EOPNOTSUPP.

Fixes: 10caabdaad5a ("net/mlx5e: Use termination table for VLAN push actions")
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Set reformat action when needed for termination rules
Jianbo Liu [Fri, 30 Apr 2021 06:58:29 +0000 (06:58 +0000)]
net/mlx5: Set reformat action when needed for termination rules

For remote mirroring, after the tunnel packets are received, they are
decapsulated and sent to representor, then re-encapsulated and sent
out over another tunnel. So reformat action is set only when the
destination is required to do encapsulation.

Fixes: 249ccc3c95bd ("net/mlx5e: Add support for offloading traffic from uplink to uplink")
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix nullptr in add_vlan_push_action()
Dima Chumak [Mon, 26 Apr 2021 12:16:26 +0000 (15:16 +0300)]
net/mlx5e: Fix nullptr in add_vlan_push_action()

The result of dev_get_by_index_rcu() is not checked for NULL and then
gets dereferenced immediately.

Also, the RCU lock must be held by the caller of dev_get_by_index_rcu(),
which isn't satisfied by the call stack.

Fix by handling nullptr return value when iflink device is not found.
Add RCU locking around dev_get_by_index_rcu() to avoid possible adverse
effects while iterating over the net_device's hlist.

It is safe not to increment reference count of the net_device pointer in
case of a successful lookup, because it's already handled by VLAN code
during VLAN device registration (see register_vlan_dev and
netdev_upper_dev_link).

Fixes: 278748a95aa3 ("net/mlx5e: Offload TC e-switch rules with egress VLAN device")
Addresses-Coverity: ("Dereference null return value")
Signed-off-by: Dima Chumak <dchumak@nvidia.com>
Reviewed-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years ago{net, RDMA}/mlx5: Fix override of log_max_qp by other device
Maor Gottlieb [Wed, 12 May 2021 10:52:27 +0000 (13:52 +0300)]
{net, RDMA}/mlx5: Fix override of log_max_qp by other device

mlx5_core_dev holds pointer to static profile, hence when the
log_max_qp of the profile is override by some device, then it
effect all other mlx5 devices that share the same profile.
Fix it by having a profile instance for every mlx5 device.

Fixes: 883371c453b9 ("net/mlx5: Check FW limitations on log_max_qp before setting it")
Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agoMerge branch 'hns3-fixes'
David S. Miller [Tue, 18 May 2021 20:41:04 +0000 (13:41 -0700)]
Merge branch 'hns3-fixes'

Huazhong Tan says:

====================
net: hns3: fixes for -net

This series includes some bugfixes for the HNS3 ethernet driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: check the return of skb_checksum_help()
Yunsheng Lin [Tue, 18 May 2021 11:36:03 +0000 (19:36 +0800)]
net: hns3: check the return of skb_checksum_help()

Currently skb_checksum_help()'s return is ignored, but it may
return error when it fails to allocate memory when linearizing.

So adds checking for the return of skb_checksum_help().

Fixes: 76ad4f0ee747("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Fixes: 3db084d28dc0("net: hns3: Fix for vxlan tx checksum bug")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: fix user's coalesce configuration lost issue
Huazhong Tan [Tue, 18 May 2021 11:36:02 +0000 (19:36 +0800)]
net: hns3: fix user's coalesce configuration lost issue

Currently, when adaptive is on, the user's coalesce configuration
may be overwritten by the dynamic one. The reason is that user's
configurations are saved in struct hns3_enet_tqp_vector whose
value maybe changed by the dynamic algorithm. To fix it, use
struct hns3_nic_priv instead of struct hns3_enet_tqp_vector to
save and get the user's configuration.

BTW, operations of storing and restoring coalesce info in the reset
process are unnecessary now, so remove them as well.

Fixes: 434776a5fae2 ("net: hns3: add ethtool_ops.set_coalesce support to PF")
Fixes: 7e96adc46633 ("net: hns3: add ethtool_ops.get_coalesce support to PF")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: put off calling register_netdev() until client initialize complete
Jian Shen [Tue, 18 May 2021 11:36:01 +0000 (19:36 +0800)]
net: hns3: put off calling register_netdev() until client initialize complete

Currently, the netdevice is registered before client initializing
complete. So there is a timewindow between netdevice available
and usable. In this case, if user try to change the channel number
or ring param, it may cause the hns3_set_rx_cpu_rmap() being called
twice, and report bug.

[47199.416502] hns3 0000:35:00.0 eth1: set channels: tqp_num=1, rxfh=0
[47199.430340] hns3 0000:35:00.0 eth1: already uninitialized
[47199.438554] hns3 0000:35:00.0: rss changes from 4 to 1
[47199.511854] hns3 0000:35:00.0: Channels changed, rss_size from 4 to 1, tqps from 4 to 1
[47200.163524] ------------[ cut here ]------------
[47200.171674] kernel BUG at lib/cpu_rmap.c:142!
[47200.177847] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP
[47200.185259] Modules linked in: hclge(+) hns3(-) hns3_cae(O) hns_roce_hw_v2 hnae3 vfio_iommu_type1 vfio_pci vfio_virqfd vfio pv680_mii(O) [last unloaded: hclge]
[47200.205912] CPU: 1 PID: 8260 Comm: ethtool Tainted: G           O      5.11.0-rc3+ #1
[47200.215601] Hardware name:  , xxxxxx 02/04/2021
[47200.223052] pstate: 60400009 (nZCv daif +PAN -UAO -TCO BTYPE=--)
[47200.230188] pc : cpu_rmap_add+0x38/0x40
[47200.237472] lr : irq_cpu_rmap_add+0x84/0x140
[47200.243291] sp : ffff800010e93a30
[47200.247295] x29: ffff800010e93a30 x28: ffff082100584880
[47200.254155] x27: 0000000000000000 x26: 0000000000000000
[47200.260712] x25: 0000000000000000 x24: 0000000000000004
[47200.267241] x23: ffff08209ba03000 x22: ffff08209ba038c0
[47200.273789] x21: 000000000000003f x20: ffff0820e2bc1680
[47200.280400] x19: ffff0820c970ec80 x18: 00000000000000c0
[47200.286944] x17: 0000000000000000 x16: ffffb43debe4a0d0
[47200.293456] x15: fffffc2082990600 x14: dead000000000122
[47200.300059] x13: ffffffffffffffff x12: 000000000000003e
[47200.306606] x11: ffff0820815b8080 x10: ffff53e411988000
[47200.313171] x9 : 0000000000000000 x8 : ffff0820e2bc1700
[47200.319682] x7 : 0000000000000000 x6 : 000000000000003f
[47200.326170] x5 : 0000000000000040 x4 : ffff800010e93a20
[47200.332656] x3 : 0000000000000004 x2 : ffff0820c970ec80
[47200.339168] x1 : ffff0820e2bc1680 x0 : 0000000000000004
[47200.346058] Call trace:
[47200.349324]  cpu_rmap_add+0x38/0x40
[47200.354300]  hns3_set_rx_cpu_rmap+0x6c/0xe0 [hns3]
[47200.362294]  hns3_reset_notify_init_enet+0x1cc/0x340 [hns3]
[47200.370049]  hns3_change_channels+0x40/0xb0 [hns3]
[47200.376770]  hns3_set_channels+0x12c/0x2a0 [hns3]
[47200.383353]  ethtool_set_channels+0x140/0x250
[47200.389772]  dev_ethtool+0x714/0x23d0
[47200.394440]  dev_ioctl+0x4cc/0x640
[47200.399277]  sock_do_ioctl+0x100/0x2a0
[47200.404574]  sock_ioctl+0x28c/0x470
[47200.409079]  __arm64_sys_ioctl+0xb4/0x100
[47200.415217]  el0_svc_common.constprop.0+0x84/0x210
[47200.422088]  do_el0_svc+0x28/0x34
[47200.426387]  el0_svc+0x28/0x70
[47200.431308]  el0_sync_handler+0x1a4/0x1b0
[47200.436477]  el0_sync+0x174/0x180
[47200.441562] Code: 11000405 79000c45 f8247861 d65f03c0 (d4210000)
[47200.448869] ---[ end trace a01efe4ce42e5f34 ]---

The process is like below:
excuting hns3_client_init
|
register_netdev()
|                           hns3_set_channels()
|                           |
hns3_set_rx_cpu_rmap()      hns3_reset_notify_uninit_enet()
|                               |
|                            quit without calling function
|                            hns3_free_rx_cpu_rmap for flag
|                            HNS3_NIC_STATE_INITED is unset.
|                           |
|                           hns3_reset_notify_init_enet()
|                               |
set HNS3_NIC_STATE_INITED    call hns3_set_rx_cpu_rmap()-- crash

Fix it by calling register_netdev() at the end of function
hns3_client_init().

Fixes: 08a100689d4b ("net: hns3: re-organize vector handle")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: fix incorrect resp_msg issue
Jiaran Zhang [Tue, 18 May 2021 11:36:00 +0000 (19:36 +0800)]
net: hns3: fix incorrect resp_msg issue

In hclge_mbx_handler(), if there are two consecutive mailbox
messages that requires resp_msg, the resp_msg is not cleared
after processing the first message, which will cause the resp_msg
data of second message incorrect.

Fix it by clearing the resp_msg before processing every mailbox
message.

Fixes: bb5790b71bad ("net: hns3: refactor mailbox response scheme between PF and VF")
Signed-off-by: Jiaran Zhang <zhangjiaran@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: lan78xx: advertise tx software timestamping support
Markus Bloechl [Tue, 18 May 2021 09:54:11 +0000 (11:54 +0200)]
net: lan78xx: advertise tx software timestamping support

lan78xx already calls skb_tx_timestamp() in its lan78xx_start_xmit().
Override .get_ts_info to also advertise this capability
(SOF_TIMESTAMPING_TX_SOFTWARE) via ethtool.

Signed-off-by: Markus Blöchl <markus.bloechl@ipetronik.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agotipc: simplify the finalize work queue
Xin Long [Tue, 18 May 2021 02:09:08 +0000 (10:09 +0800)]
tipc: simplify the finalize work queue

This patch is to use "struct work_struct" for the finalize work queue
instead of "struct tipc_net_work", as it can get the "net" and "addr"
from tipc_net's other members and there is no need to add extra net
and addr in tipc_net by defining "struct tipc_net_work".

Note that it's safe to get net from tn->bcl as bcl is always released
after the finalize work queue is done.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: mdiobus: get rid of a BUG_ON()
Dan Carpenter [Mon, 17 May 2021 09:04:13 +0000 (12:04 +0300)]
net: mdiobus: get rid of a BUG_ON()

We spotted a bug recently during a review where a driver was
unregistering a bus that wasn't registered, which would trigger this
BUG_ON().  Let's handle that situation more gracefully, and just print
a warning and return.

Reported-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'gve-fixes'
David S. Miller [Mon, 17 May 2021 22:38:40 +0000 (15:38 -0700)]
Merge branch 'gve-fixes'

David Awogbemila says:

====================
GVE bug fixes

This patch series includes fixes to some bugs in the gve driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agogve: Correct SKB queue index validation.
David Awogbemila [Mon, 17 May 2021 21:08:15 +0000 (14:08 -0700)]
gve: Correct SKB queue index validation.

SKBs with skb_get_queue_mapping(skb) == tx_cfg.num_queues should also be
considered invalid.

Fixes: f5cedc84a30d ("gve: Add transmit and receive support")
Signed-off-by: David Awogbemila <awogbemila@google.com>
Acked-by: Willem de Brujin <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agogve: Upgrade memory barrier in poll routine
Catherine Sullivan [Mon, 17 May 2021 21:08:14 +0000 (14:08 -0700)]
gve: Upgrade memory barrier in poll routine

As currently written, if the driver checks for more work (via
gve_tx_poll or gve_rx_poll) before the device posts work and the
irq doorbell is not unmasked
(via iowrite32be(GVE_IRQ_ACK | GVE_IRQ_EVENT, ...)) before the device
attempts to raise an interrupt, an interrupt is lost and this could
potentially lead to the traffic being completely halted. For
example, if a tx queue has already been stopped, the driver won't get
the chance to complete work and egress will be halted.

We need a full memory barrier in the poll
routine to ensure that the irq doorbell is unmasked before the driver
checks for more work.

Fixes: f5cedc84a30d ("gve: Add transmit and receive support")
Signed-off-by: Catherine Sullivan <csully@google.com>
Signed-off-by: David Awogbemila <awogbemila@google.com>
Acked-by: Willem de Brujin <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agogve: Add NULL pointer checks when freeing irqs.
David Awogbemila [Mon, 17 May 2021 21:08:13 +0000 (14:08 -0700)]
gve: Add NULL pointer checks when freeing irqs.

When freeing notification blocks, we index priv->msix_vectors.
If we failed to allocate priv->msix_vectors (see abort_with_msix_vectors)
this could lead to a NULL pointer dereference if the driver is unloaded.

Fixes: 893ce44df565 ("gve: Add basic driver framework for Compute Engine Virtual NIC")
Signed-off-by: David Awogbemila <awogbemila@google.com>
Acked-by: Willem de Brujin <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agogve: Update mgmt_msix_idx if num_ntfy changes
David Awogbemila [Mon, 17 May 2021 21:08:12 +0000 (14:08 -0700)]
gve: Update mgmt_msix_idx if num_ntfy changes

If we do not get the expected number of vectors from
pci_enable_msix_range, we update priv->num_ntfy_blks but not
priv->mgmt_msix_idx. This patch fixes this so that priv->mgmt_msix_idx
is updated accordingly.

Fixes: f5cedc84a30d ("gve: Add transmit and receive support")
Signed-off-by: David Awogbemila <awogbemila@google.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agogve: Check TX QPL was actually assigned
Catherine Sullivan [Mon, 17 May 2021 21:08:11 +0000 (14:08 -0700)]
gve: Check TX QPL was actually assigned

Correctly check the TX QPL was assigned and unassigned if
other steps in the allocation fail.

Fixes: f5cedc84a30d (gve: Add transmit and receive support)
Signed-off-by: Catherine Sullivan <csully@google.com>
Signed-off-by: David Awogbemila <awogbemila@google.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonetlink: disable IRQs for netlink_lock_table()
Johannes Berg [Mon, 17 May 2021 14:38:09 +0000 (16:38 +0200)]
netlink: disable IRQs for netlink_lock_table()

Syzbot reports that in mac80211 we have a potential deadlock
between our "local->stop_queue_reasons_lock" (spinlock) and
netlink's nl_table_lock (rwlock). This is because there's at
least one situation in which we might try to send a netlink
message with this spinlock held while it is also possible to
take the spinlock from a hardirq context, resulting in the
following deadlock scenario reported by lockdep:

       CPU0                    CPU1
       ----                    ----
  lock(nl_table_lock);
                               local_irq_disable();
                               lock(&local->queue_stop_reason_lock);
                               lock(nl_table_lock);
  <Interrupt>
    lock(&local->queue_stop_reason_lock);

This seems valid, we can take the queue_stop_reason_lock in
any kind of context ("CPU0"), and call ieee80211_report_ack_skb()
with the spinlock held and IRQs disabled ("CPU1") in some
code path (ieee80211_do_stop() via ieee80211_free_txskb()).

Short of disallowing netlink use in scenarios like these
(which would be rather complex in mac80211's case due to
the deep callchain), it seems the only fix for this is to
disable IRQs while nl_table_lock is held to avoid hitting
this scenario, this disallows the "CPU0" portion of the
reported deadlock.

Note that the writer side (netlink_table_grab()) already
disables IRQs for this lock.

Unfortunately though, this seems like a huge hammer, and
maybe the whole netlink table locking should be reworked.

Reported-by: syzbot+69ff9dff50dcfe14ddd4@syzkaller.appspotmail.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>