RFC8754 says:

ICMP error packets generated within the SR domain are sent to source
nodes within the SR domain.  The invoking packet in the ICMP error
message may contain an SRH.  Since the destination address of a packet
with an SRH changes as each segment is processed, it may not be the
destination used by the socket or application that generated the
invoking packet.

For the source of an invoking packet to process the ICMP error
message, the ultimate destination address of the IPv6 header may be
required.  The following logic is used to determine the destination
address for use by protocol-error handlers.

*  Walk all extension headers of the invoking IPv6 packet to the
   routing extension header preceding the upper-layer header.

   -  If routing header is type 4 Segment Routing Header (SRH)

      o  The SID at Segment List[0] may be used as the destination
         address of the invoking packet.

Mangle the skb so the network header points to the invoking packet
inside the ICMP packet. The seg6 helpers can then be used on the skb
to find any segment routing headers. If found, mark this fact in the
IPv6 control block of the skb, and store the offset into the packet of
the SRH. Then restore the skb back to its old state.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An ICMP error message can contain in its message body part of an IPv6
packet which invoked the error. Such a packet might contain a segment
router header. Export get_srh() so the ICMP code can make use of it.

Since his changes the scope of the function from local to global, add
the seg6_ prefix to keep the namespace clean. And move it into seg6.c
so it is always available, not just when IPV6_SEG6_LWTUNNEL is
enabled.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Simon Wunderlich says:

====================
Here is a batman-adv bugfix:

 - avoid sending link-local multicast to multicast routers,
   by Linus Lüssing

* tag 'batadv-net-pullrequest-20220103' of git://git.open-mesh.org/linux-merge:
  batman-adv: mcast: don't send link-local multicast to mcast routers
====================

Link: https://lore.kernel.org/r/20220103171203.1124980-1-sw@simonwunderlich.deSigned-off-by: Jakub Kicinski <kuba@kernel.org>

This reverts commit b45396af ("net: phy:
fixed_phy: Fix NULL vs IS_ERR() checking in __fixed_phy_register")
since it prevents any system that uses a fixed PHY without a GPIO
descriptor from properly working:

[    5.971952] brcm-systemport 9300000.ethernet: failed to register fixed PHY
[    5.978854] brcm-systemport: probe of 9300000.ethernet failed with error -22
[    5.986047] brcm-systemport 9400000.ethernet: failed to register fixed PHY
[    5.992947] brcm-systemport: probe of 9400000.ethernet failed with error -22

Fixes: b45396af ("net: phy: fixed_phy: Fix NULL vs IS_ERR() checking in __fixed_phy_register")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20220103193453.1214961-1-f.fainelli@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

As Nicolas noted, if gateway validation fails walking the multipath
attribute the code should jump to the cleanup to free previously
allocated memory.

Fixes: 1ff15a71 ("ipv6: Check attribute length for RTA_GATEWAY when deleting multipath route")
Signed-off-by: David Ahern <dsahern@kernel.org>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Link: https://lore.kernel.org/r/20220103170555.94638-1-dsahern@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

ip6_route_multipath_del loop continues processing the multipath
attribute even if delete of a nexthop path fails. For consistency,
do the same if the gateway attribute is invalid.

Fixes: 1ff15a71 ("ipv6: Check attribute length for RTA_GATEWAY when deleting multipath route")
Signed-off-by: David Ahern <dsahern@kernel.org>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Link: https://lore.kernel.org/r/20220103171911.94739-1-dsahern@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

commit 26eee021 ("net/fsl: fix a bug in xgmac_mdio") fixed a bug in
the QorIQ mdio driver but left the (now unused) incorrect bit definition
for MDIO_DATA_BSY in the code. This commit removes it.
Signed-off-by: Markus Koch <markus@notsyncing.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Hytera makes a range of digital (DMR) radios. These radios can be
programmed to a allow a computer to control them over Ethernet over USB,
either using NCM or RNDIS.

This commit adds support for RNDIS for Hytera radios. I tested with a
Hytera PD785 and a Hytera MD785G. When these radios are programmed to
set up a Radio to PC Network using RNDIS, an USB interface will be added
with class 2 (Communications), subclass 2 (Abstract Modem Control) and
an interface protocol of 255 ("vendor specific" - lsusb even hints "MSFT
RNDIS?").

This patch is similar to the solution of this StackOverflow user, but
that only works for the Hytera MD785:
https://stackoverflow.com/a/53550858

To use the "Radio to PC Network" functionality of Hytera DMR radios, the
radios need to be programmed correctly in CPS (Hytera's Customer
Programming Software). "Forward to PC" should be checked in "Network"
(under "General Setting" in "Conventional") and the "USB Network
Communication Protocol" should be set to RNDIS.
Signed-off-by: Thomas Toye <thomas@toye.io>
Signed-off-by: David S. Miller <davem@davemloft.net>

The same fix in commit 5ec7d18d ("sctp: use call_rcu to free endpoint")
is also needed for dumping one asoc and sock after the lookup.

Fixes: 86fdb344 ("sctp: ensure ep is not destroyed before doing the dump")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Arthur Kiyanovski says:

====================
ENA driver bug fixes

Patchset V2 chages:
-------------------
Updated SHA1 of Fixes tag in patch 3/3 to be 12 digits long

Original cover letter:
----------------------
ENA driver bug fixes
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The role of ena_calc_max_io_queue_num() is to return the number
of queues supported by the device, which means the return value
should be >=0.

The function that calls ena_calc_max_io_queue_num(), checks
the return value. If it is 0, it means the device reported
it supports 0 IO queues. This case is considered an error
and is handled by the calling function accordingly.

However the current implementation of ena_calc_max_io_queue_num()
is wrong, since when it detects the device supports 0 IO queues,
it returns -EFAULT.

In such a case the calling function doesn't detect the error,
and therefore doesn't handle it.

This commit changes ena_calc_max_io_queue_num() to return 0
in case the device reported it supports 0 queues, allowing the
calling function to properly handle the error case.

Fixes: 736ce3f4 ("net: ena: make ethtool -l show correct max number of queues")
Signed-off-by: Shay Agroskin <shayagr@amazon.com>
Signed-off-by: Arthur Kiyanovski <akiyano@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

A wrong request id received from the device is a sign that
something is wrong with it, therefore trigger a device reset.

Also add some debug info to the "Page is NULL" print to make
it easier to debug.

Fixes: 1738cd3e ("net: ena: Add a driver for Amazon Elastic Network Adapters (ENA)")
Signed-off-by: Arthur Kiyanovski <akiyano@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ena_com_tx_comp_req_id_get() checks the req_id of a received completion,
and if it is out of bounds returns -EINVAL. This is a sign that
something is wrong with the device and it needs to be reset.

The current code does not reset the device in this case, which leaves
the driver in an undefined state, where this completion is not properly
handled.

This commit adds a call to handle_invalid_req_id() in ena_clean_tx_irq()
and ena_clean_xdp_irq() which resets the device to fix the issue.

This commit also removes unnecessary request id checks from
validate_tx_req_id() and validate_xdp_req_id(). This check is unneeded
because it was already performed in ena_com_tx_comp_req_id_get(), which
is called right before these functions.

Fixes: 548c4940 ("net: ena: Implement XDP_TX action")
Signed-off-by: Shay Agroskin <shayagr@amazon.com>
Signed-off-by: Arthur Kiyanovski <akiyano@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add neighbour source flag in mctp_neigh_remove(...) to allow removal of
only static neighbours.

This should be a no-op change and might be useful later when mctp can
have MCTP_NEIGH_DISCOVER neighbours.
Signed-off-by: Gagan Kumar <gagan1kumar.cs@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The addition of routable multicast TX handling introduced a
bug/regression for packets with a link-local multicast destination:
These packets would be sent to all batman-adv nodes with a multicast
router and to all batman-adv nodes with an old version without multicast
router detection.

This even disregards the batman-adv multicast fanout setting, which can
potentially lead to an unwanted, high number of unicast transmissions or
even congestion.

Fixing this by avoiding to send link-local multicast packets to nodes in
the multicast router list.

Fixes: 11d458c1 ("batman-adv: mcast: apply optimizations for routable packets, too")
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>

struct tipc_socket_addr.ref has a 4-byte hole,and __tipc_getname() currently
copying it to user space,causing kernel-infoleak.

BUG: KMSAN: kernel-infoleak in instrument_copy_to_user include/linux/instrumented.h:121 [inline]
BUG: KMSAN: kernel-infoleak in instrument_copy_to_user include/linux/instrumented.h:121 [inline] lib/usercopy.c:33
BUG: KMSAN: kernel-infoleak in _copy_to_user+0x1c9/0x270 lib/usercopy.c:33 lib/usercopy.c:33
 instrument_copy_to_user include/linux/instrumented.h:121 [inline]
 instrument_copy_to_user include/linux/instrumented.h:121 [inline] lib/usercopy.c:33
 _copy_to_user+0x1c9/0x270 lib/usercopy.c:33 lib/usercopy.c:33
 copy_to_user include/linux/uaccess.h:209 [inline]
 copy_to_user include/linux/uaccess.h:209 [inline] net/socket.c:287
 move_addr_to_user+0x3f6/0x600 net/socket.c:287 net/socket.c:287
 __sys_getpeername+0x470/0x6b0 net/socket.c:1987 net/socket.c:1987
 __do_sys_getpeername net/socket.c:1997 [inline]
 __se_sys_getpeername net/socket.c:1994 [inline]
 __do_sys_getpeername net/socket.c:1997 [inline] net/socket.c:1994
 __se_sys_getpeername net/socket.c:1994 [inline] net/socket.c:1994
 __x64_sys_getpeername+0xda/0x120 net/socket.c:1994 net/socket.c:1994
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_x64 arch/x86/entry/common.c:51 [inline] arch/x86/entry/common.c:82
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Uninit was stored to memory at:
 tipc_getname+0x575/0x5e0 net/tipc/socket.c:757 net/tipc/socket.c:757
 __sys_getpeername+0x3b3/0x6b0 net/socket.c:1984 net/socket.c:1984
 __do_sys_getpeername net/socket.c:1997 [inline]
 __se_sys_getpeername net/socket.c:1994 [inline]
 __do_sys_getpeername net/socket.c:1997 [inline] net/socket.c:1994
 __se_sys_getpeername net/socket.c:1994 [inline] net/socket.c:1994
 __x64_sys_getpeername+0xda/0x120 net/socket.c:1994 net/socket.c:1994
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_x64 arch/x86/entry/common.c:51 [inline] arch/x86/entry/common.c:82
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Uninit was stored to memory at:
 msg_set_word net/tipc/msg.h:212 [inline]
 msg_set_destport net/tipc/msg.h:619 [inline]
 msg_set_word net/tipc/msg.h:212 [inline] net/tipc/socket.c:1486
 msg_set_destport net/tipc/msg.h:619 [inline] net/tipc/socket.c:1486
 __tipc_sendmsg+0x44fa/0x5890 net/tipc/socket.c:1486 net/tipc/socket.c:1486
 tipc_sendmsg+0xeb/0x140 net/tipc/socket.c:1402 net/tipc/socket.c:1402
 sock_sendmsg_nosec net/socket.c:704 [inline]
 sock_sendmsg net/socket.c:724 [inline]
 sock_sendmsg_nosec net/socket.c:704 [inline] net/socket.c:2409
 sock_sendmsg net/socket.c:724 [inline] net/socket.c:2409
 ____sys_sendmsg+0xe11/0x12c0 net/socket.c:2409 net/socket.c:2409
 ___sys_sendmsg net/socket.c:2463 [inline]
 ___sys_sendmsg net/socket.c:2463 [inline] net/socket.c:2492
 __sys_sendmsg+0x704/0x840 net/socket.c:2492 net/socket.c:2492
 __do_sys_sendmsg net/socket.c:2501 [inline]
 __se_sys_sendmsg net/socket.c:2499 [inline]
 __do_sys_sendmsg net/socket.c:2501 [inline] net/socket.c:2499
 __se_sys_sendmsg net/socket.c:2499 [inline] net/socket.c:2499
 __x64_sys_sendmsg+0xe2/0x120 net/socket.c:2499 net/socket.c:2499
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_x64 arch/x86/entry/common.c:51 [inline] arch/x86/entry/common.c:82
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Local variable skaddr created at:
 __tipc_sendmsg+0x2d0/0x5890 net/tipc/socket.c:1419 net/tipc/socket.c:1419
 tipc_sendmsg+0xeb/0x140 net/tipc/socket.c:1402 net/tipc/socket.c:1402

Bytes 4-7 of 16 are uninitialized
Memory access of size 16 starts at ffff888113753e00
Data copied to user address 0000000020000280

Reported-by: syzbot+cdbd40e0c3ca02cae3b7@syzkaller.appspotmail.com
Signed-off-by: Haimin Zhang <tcs_kernel@tencent.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
Link: https://lore.kernel.org/r/1640918123-14547-1-git-send-email-tcs.kernel@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

As Paolo pointed out, the result of ping IPv6 address depends on
the running distro. So explicitly checking the available ping feature,
as e.g. do the bareudp.sh self-tests.

Fixes: 8b3170e0 ("selftests: net: using ping6 for IPv6 in udpgro_fwd.sh")
Signed-off-by: Jianguo Wu <wujianguo@chinatelecom.cn>
Link: https://lore.kernel.org/r/825ee22b-4245-dbf7-d2f7-a230770d6e21@163.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Daniel Borkmann says:

====================
pull-request: bpf 2021-12-31

We've added 2 non-merge commits during the last 14 day(s) which contain
a total of 2 files changed, 3 insertions(+), 3 deletions(-).

The main changes are:

1) Revert of an earlier attempt to fix xsk's poll() behavior where it
   turned out that the fix for a rare problem made it much worse in
   general, from Magnus Karlsson. (Fyi, Magnus mentioned that a proper
   fix is coming early next year, so the revert is mainly to avoid
   slipping the behavior into 5.16.)

2) Minor misc spell fix in BPF selftests, from Colin Ian King.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpf, selftests: Fix spelling mistake "tained" -> "tainted"
  Revert "xsk: Do not sleep in poll() when need_wakeup set"
====================

Link: https://lore.kernel.org/r/20211231160050.16105-1-daniel@iogearbox.netSigned-off-by: Jakub Kicinski <kuba@kernel.org>

David Ahern says:

====================
net: Length checks for attributes within multipath routes

Add length checks for attributes within a multipath route (attributes
within RTA_MULTIPATH). Motivated by the syzbot report in patch 1 and
then expanded to other attributes as noted by Ido.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

lwtunnel_valid_encap_type_attr is used to validate encap attributes
within a multipath route. Add length validation checking to the type.

lwtunnel_valid_encap_type_attr is called converting attributes to
fib{6,}_config struct which means it is used before fib_get_nhs,
ip6_route_multipath_add, and ip6_route_multipath_del - other
locations that use rtnh_ok and then nla_get_u16 on RTA_ENCAP_TYPE
attribute.

Fixes: 9ed59592 ("lwtunnel: fix autoload of lwt modules")
Signed-off-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure RTA_GATEWAY for IPv6 multipath route has enough bytes to hold
an IPv6 address.

Fixes: 6b9ea5a6 ("ipv6: fix multipath route replace error recovery")
Signed-off-by: David Ahern <dsahern@kernel.org>
Cc: Roopa Prabhu <roopa@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit referenced in the Fixes tag used nla_memcpy for RTA_GATEWAY as
does the current nla_get_in6_addr. nla_memcpy protects against accessing
memory greater than what is in the attribute, but there is no check
requiring the attribute to have an IPv6 address. Add it.

Fixes: 51ebd318 ("ipv6: add support of equal cost multipath (ECMP)")
Signed-off-by: David Ahern <dsahern@kernel.org>
Cc: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make sure RTA_FLOW is at least 4B before using.

Fixes: 4e902c57 ("[IPv4]: FIB configuration using struct fib_config")
Signed-off-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

syzbot reported uninit-value:
============================================================
  BUG: KMSAN: uninit-value in fib_get_nhs+0xac4/0x1f80
  net/ipv4/fib_semantics.c:708
   fib_get_nhs+0xac4/0x1f80 net/ipv4/fib_semantics.c:708
   fib_create_info+0x2411/0x4870 net/ipv4/fib_semantics.c:1453
   fib_table_insert+0x45c/0x3a10 net/ipv4/fib_trie.c:1224
   inet_rtm_newroute+0x289/0x420 net/ipv4/fib_frontend.c:886

Add helper to validate RTA_GATEWAY length before using the attribute.

Fixes: 4e902c57 ("[IPv4]: FIB configuration using struct fib_config")
Reported-by: syzbot+d4b9a2851cc3ce998741@syzkaller.appspotmail.com
Signed-off-by: David Ahern <dsahern@kernel.org>
Cc: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull networking fixes from Jakub Kicinski:
 "Including fixes from.. Santa?

  No regressions on our radar at this point. The igc problem fixed here
  was the last one I was tracking but it was broken in previous
  releases, anyway. Mostly driver fixes and a couple of largish SMC
  fixes.

  Current release - regressions:

   - xsk: initialise xskb free_list_node, fixup for a -rc7 fix

  Current release - new code bugs:

   - mlx5: handful of minor fixes:

   - use first online CPU instead of hard coded CPU

   - fix some error handling paths in 'mlx5e_tc_add_fdb_flow()'

   - fix skb memory leak when TC classifier action offloads are disabled

   - fix memory leak with rules with internal OvS port

  Previous releases - regressions:

   - igc: do not enable crosstimestamping for i225-V models

  Previous releases - always broken:

   - udp: use datalen to cap ipv6 udp max gso segments

   - fix use-after-free in tw_timer_handler due to early free of stats

   - smc: fix kernel panic caused by race of smc_sock

   - smc: don't send CDC/LLC message if link not ready, avoid timeouts

   - sctp: use call_rcu to free endpoint, avoid UAF in sock diag

   - bridge: mcast: add and enforce query interval minimum

   - usb: pegasus: do not drop long Ethernet frames

   - mlx5e: fix ICOSQ recovery flow for XSK

   - nfc: uapi: use kernel size_t to fix user-space builds"

* tag 'net-5.16-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (47 commits)
  fsl/fman: Fix missing put_device() call in fman_port_probe
  selftests: net: using ping6 for IPv6 in udpgro_fwd.sh
  Documentation: fix outdated interpretation of ip_no_pmtu_disc
  net/ncsi: check for error return from call to nla_put_u32
  net: bridge: mcast: fix br_multicast_ctx_vlan_global_disabled helper
  net: fix use-after-free in tw_timer_handler
  selftests: net: Fix a typo in udpgro_fwd.sh
  selftests/net: udpgso_bench_tx: fix dst ip argument
  net: bridge: mcast: add and enforce startup query interval minimum
  net: bridge: mcast: add and enforce query interval minimum
  ipv6: raw: check passed optlen before reading
  xsk: Initialise xskb free_list_node
  net/mlx5e: Fix wrong features assignment in case of error
  net/mlx5e: TC, Fix memory leak with rules with internal port
  ionic: Initialize the 'lif->dbid_inuse' bitmap
  igc: Fix TX timestamp support for non-MSI-X platforms
  igc: Do not enable crosstimestamping for i225-V models
  net/smc: fix kernel panic caused by race of smc_sock
  net/smc: don't send CDC/LLC message if link not ready
  NFC: st21nfca: Fix memory leak in device probe and remove
  ...

Pull char/misc fixes from Greg KH:
 "Here are two misc driver fixes for 5.16-final:

   - binder accounting fix to resolve reported problem

   - nitro_enclaves fix for mmap assert warning output

  Both of these have been for over a week with no reported issues"

* tag 'char-misc-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  nitro_enclaves: Use get_user_pages_unlocked() call to handle mmap assert
  binder: fix async_free_space accounting for empty parcels

Pull USB fixes from Greg KH:
 "Here are some small USB driver fixes for 5.16 to resolve some reported
  problems:

   - mtu3 driver fixes

   - typec ucsi driver fix

   - xhci driver quirk added

   - usb gadget f_fs fix for reported crash

  All of these have been in linux-next for a while with no reported
  problems"

* tag 'usb-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  usb: typec: ucsi: Only check the contract if there is a connection
  xhci: Fresco FL1100 controller should not have BROKEN_MSI quirk set.
  usb: mtu3: set interval of FS intr and isoc endpoint
  usb: mtu3: fix list_head check warning
  usb: mtu3: add memory barrier before set GPD's HWO
  usb: mtu3: fix interval value for intr and isoc
  usb: gadget: f_fs: Clear ffs_eventfd in ffs_data_clear.

The reference taken by 'of_find_device_by_node()' must be released when
not needed anymore.
Add the corresponding 'put_device()' in the and error handling paths.

Fixes: 18a6c85f ("fsl/fman: Add FMan Port Support")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

udpgro_fwd.sh output following message:
  ping: 2001:db8:1::100: Address family for hostname not supported

Using ping6 when pinging IPv6 addresses.

Fixes: a062260a ("selftests: net: add UDP GRO forwarding self-tests")
Signed-off-by: Jianguo Wu <wujianguo@chinatelecom.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>

The updating way of pmtu has changed, but documentation is still in the
old way. So this patch updates the interpretation of ip_no_pmtu_disc and
min_pmtu.

See commit 28d35bcd ("net: ipv4: don't let PMTU updates increase
route MTU")
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: xu xin <xu.xin16@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>

Saeed Mahameed says:

====================
mlx5 fixes 2021-12-28

This series provides bug fixes to mlx5 driver.

* tag 'mlx5-fixes-2021-12-28' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Fix wrong features assignment in case of error
  net/mlx5e: TC, Fix memory leak with rules with internal port
====================

Link: https://lore.kernel.org/r/20211229065352.30178-1-saeed@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

As we can see from the comment of the nla_put() that it could return
-EMSGSIZE if the tailroom of the skb is insufficient.
Therefore, it should be better to check the return value of the
nla_put_u32 and return the error code if error accurs.
Also, there are many other functions have the same problem, and if this
patch is correct, I will commit a new version to fix all.

Fixes: 955dc68c ("net/ncsi: Add generic netlink family")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Link: https://lore.kernel.org/r/20211229032118.1706294-1-jiasheng@iscas.ac.cnSigned-off-by: Jakub Kicinski <kuba@kernel.org>

We need to first check if the context is a vlan one, then we need to
check the global bridge multicast vlan snooping flag, and finally the
vlan's multicast flag, otherwise we will unnecessarily enable vlan mcast
processing (e.g. querier timers).

Fixes: 7b54aaaf ("net: bridge: multicast: add vlan state initialization and control")
Signed-off-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Link: https://lore.kernel.org/r/20211228153142.536969-1-nikolay@nvidia.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

A real world panic issue was found as follow in Linux 5.4.

    BUG: unable to handle page fault for address: ffffde49a863de28
    PGD 7e6fe62067 P4D 7e6fe62067 PUD 7e6fe63067 PMD f51e064067 PTE 0
    RIP: 0010:tw_timer_handler+0x20/0x40
    Call Trace:
     <IRQ>
     call_timer_fn+0x2b/0x120
     run_timer_softirq+0x1ef/0x450
     __do_softirq+0x10d/0x2b8
     irq_exit+0xc7/0xd0
     smp_apic_timer_interrupt+0x68/0x120
     apic_timer_interrupt+0xf/0x20

This issue was also reported since 2017 in the thread [1],
unfortunately, the issue was still can be reproduced after fixing
DCCP.

The ipv4_mib_exit_net is called before tcp_sk_exit_batch when a net
namespace is destroyed since tcp_sk_ops is registered befrore
ipv4_mib_ops, which means tcp_sk_ops is in the front of ipv4_mib_ops
in the list of pernet_list. There will be a use-after-free on
net->mib.net_statistics in tw_timer_handler after ipv4_mib_exit_net
if there are some inflight time-wait timers.

This bug is not introduced by commit f2bf415c ("mib: add net to
NET_ADD_STATS_BH") since the net_statistics is a global variable
instead of dynamic allocation and freeing. Actually, commit
61a7e260 ("mib: put net statistics on struct net") introduces
the bug since it put net statistics on struct net and free it when
net namespace is destroyed.

Moving init_ipv4_mibs() to the front of tcp_init() to fix this bug
and replace pr_crit() with panic() since continuing is meaningless
when init_ipv4_mibs() fails.

[1] https://groups.google.com/g/syzkaller/c/p1tn-_Kc6l4/m/smuL_FMAAgAJ?pli=1

Fixes: 61a7e260 ("mib: put net statistics on struct net")
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Cc: Cong Wang <cong.wang@bytedance.com>
Cc: Fam Zheng <fam.zheng@bytedance.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20211228104145.9426-1-songmuchun@bytedance.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

$rvs -> $rcv

Fixes: a062260a ("selftests: net: add UDP GRO forwarding self-tests")
Signed-off-by: Jianguo Wu <wujianguo@chinatelecom.cn>
Link: https://lore.kernel.org/r/d247d7c8-a03a-0abf-3c71-4006a051d133@163.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

udpgso_bench_tx call setup_sockaddr() for dest address before
parsing all arguments, if we specify "-p ${dst_port}" after "-D ${dst_ip}",
then ${dst_port} will be ignored, and using default cfg_port 8000.

This will cause test case "multiple GRO socks" failed in udpgro.sh.

Setup sockaddr after parsing all arguments.

Fixes: 3a687bef ("selftests: udp gso benchmark")
Signed-off-by: Jianguo Wu <wujianguo@chinatelecom.cn>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/ff620d9f-5b52-06ab-5286-44b945453002@163.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

Nikolay Aleksandrov says:

====================
net: bridge: mcast: add and enforce query interval minimum

This set adds and enforces 1 second minimum value for bridge multicast
query and startup query intervals in order to avoid rearming the timers
too often which could lock and crash the host. I doubt anyone is using
such low values or anything lower than 1 second, so it seems like a good
minimum. In order to be compatible if the value is lower then it is
overwritten and a log message is emitted, since we can't return an error
at this point.

Eric, I looked for the syzbot reports in its dashboard but couldn't find
them so I've added you as the reporter.

I've prepared a global bridge igmp rate limiting patch but wasn't
sure if it's ok for -net. It adds a static limit of 32k packets per
second, I plan to send it for net-next with added drop counters for
each bridge so it can be easily debugged.

Original report can be seen at:
https://lore.kernel.org/netdev/e8b9ce41-57b9-b6e2-a46a-ff9c791cf0ba@gmail.com/
====================

Link: https://lore.kernel.org/r/20211227172116.320768-1-nikolay@nvidia.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

As reported[1] if startup query interval is set too low in combination with
large number of startup queries and we have multiple bridges or even a
single bridge with multiple querier vlans configured we can crash the
machine. Add a 1 second minimum which must be enforced by overwriting the
value if set lower (i.e. without returning an error) to avoid breaking
user-space. If that happens a log message is emitted to let the admin know
that the startup interval has been set to the minimum. It doesn't make
sense to make the startup interval lower than the normal query interval
so use the same value of 1 second. The issue has been present since these
intervals could be user-controlled.

[1] https://lore.kernel.org/netdev/e8b9ce41-57b9-b6e2-a46a-ff9c791cf0ba@gmail.com/

Fixes: d902eee4 ("bridge: Add multicast count/interval sysfs entries")
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

As reported[1] if query interval is set too low and we have multiple
bridges or even a single bridge with multiple querier vlans configured
we can crash the machine. Add a 1 second minimum which must be enforced
by overwriting the value if set lower (i.e. without returning an error) to
avoid breaking user-space. If that happens a log message is emitted to let
the administrator know that the interval has been set to the minimum.
The issue has been present since these intervals could be user-controlled.

[1] https://lore.kernel.org/netdev/e8b9ce41-57b9-b6e2-a46a-ff9c791cf0ba@gmail.com/

Fixes: d902eee4 ("bridge: Add multicast count/interval sysfs entries")
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Add a check that the user-provided option is at least as long as the
number of bytes we intend to read. Before this patch we would blindly
read sizeof(int) bytes even in cases where the user passed
optlen<sizeof(int), which would potentially read garbage or fault.

Discovered by new tests in https://github.com/google/gvisor/pull/6957 .

The original get_user call predates history in the git repo.
Signed-off-by: Tamir Duberstein <tamird@gmail.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20211229200947.2862255-1-willemdebruijn.kernel@gmail.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>