This reverts commit 1b479fb8
("drivers/net/wan/hdlc_fr: Fix a double free in pvc_xmit").

1. This commit is incorrect. "__skb_pad" will NOT free the skb on
failure when its "free_on_error" parameter is "false".

2. This commit claims to fix my commit. But it didn't CC me??

Fixes: 1b479fb8

 ("drivers/net/wan/hdlc_fr: Fix a double free in pvc_xmit")
Cc: Lv Yunlong <lyl2019@mail.ustc.edu.cn>
Signed-off-by: Xie He <xie.he.0141@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Xin Long says:

====================
sctp: fix the race condition in sctp_destroy_sock in a proper way

The original fix introduced a dead lock, and has to be removed in
Patch 1/2, and we will get a proper way to fix it in Patch 2/2.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

As Or Cohen described:

  If sctp_destroy_sock is called without sock_net(sk)->sctp.addr_wq_lock
  held and sp->do_auto_asconf is true, then an element is removed
  from the auto_asconf_splist without any proper locking.

  This can happen in the following functions:
  1. In sctp_accept, if sctp_sock_migrate fails.
  2. In inet_create or inet6_create, if there is a bpf program
     attached to BPF_CGROUP_INET_SOCK_CREATE which denies
     creation of the sctp socket.

This patch is to fix it by moving the auto_asconf init out of
sctp_init_sock(), by which inet_create()/inet6_create() won't
need to operate it in sctp_destroy_sock() when calling
sk_common_release().

It also makes more sense to do auto_asconf init while binding the
first addr, as auto_asconf actually requires an ANY addr bind,
see it in sctp_addr_wq_timeout_handler().

This addresses CVE-2021-23133.

Fixes: 61023658

 ("bpf: Add new cgroup attach type to enable sock modifications")
Reported-by: Or Cohen <orcohen@paloaltonetworks.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit b166a20b

.

This one has to be reverted as it introduced a dead lock, as
syzbot reported:

       CPU0                    CPU1
       ----                    ----
  lock(&net->sctp.addr_wq_lock);
                               lock(slock-AF_INET6);
                               lock(&net->sctp.addr_wq_lock);
  lock(slock-AF_INET6);

CPU0 is the thread of sctp_addr_wq_timeout_handler(), and CPU1
is that of sctp_close().

The original issue this commit fixed will be fixed in the next
patch.

Reported-by:  <syzbot+959223586843e69a2674@syzkaller.appspotmail.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Check at start of fill_frame_info that the MAC header in the supplied
skb is large enough to fit a struct hsr_ethhdr, as otherwise this is
not a valid HSR frame. If it is too small, return an error which will
then cause the callers to clean up the skb. Fixes a KMSAN-found
uninit-value bug reported by syzbot at:
https://syzkaller.appspot.com/bug?id=f7e9b601f1414f814f7602a82b6619a8d80bce3f



Reported-by:  <syzbot+e267bed19bfc5478fb33@syzkaller.appspotmail.com>
Signed-off-by: Phillip Potter <phil@philpotter.co.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

Normally SCTP_MIB_CURRESTAB is always incremented once asoc enter into
ESTABLISHED from the state < ESTABLISHED and decremented when the asoc
is being deleted.

However, in sctp_sf_do_dupcook_b(), the asoc's state can be changed to
ESTABLISHED from the state >= ESTABLISHED where it shouldn't increment
SCTP_MIB_CURRESTAB. Otherwise, one asoc may increment MIB_CURRESTAB
multiple times but only decrement once at the end.

I was able to reproduce it by using scapy to do the 4-way shakehands,
after that I replayed the COOKIE-ECHO chunk with 'peer_vtag' field
changed to different values, and SCTP_MIB_CURRESTAB was incremented
multiple times and never went back to 0 even when the asoc was freed.

This patch is to fix it by only incrementing SCTP_MIB_CURRESTAB when
the state < ESTABLISHED in sctp_sf_do_dupcook_b().

Fixes: 1da177e4

 ("Linux-2.6.12-rc2")
Reported-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Xin Long says:

====================
sctp: fix the incorrect revert

commit 35b4f244 ("sctp: do asoc update earlier in
sctp_sf_do_dupcook_a") only keeps the SHUTDOWN and
COOKIE-ACK with the same asoc, not transport.

So instead of revert commit 145cb2f7 ("sctp: Fix bundling
of SHUTDOWN with COOKIE-ACK"), we should revert 12dfd78e


("sctp: Fix SHUTDOWN CTSN Ack in the peer restart case").
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit 12dfd78e.

This can be reverted as shutdown and cookie_ack chunk are using the
same asoc since commit 35b4f244

 ("sctp: do asoc update earlier
in sctp_sf_do_dupcook_a").

Reported-by: Jere Leppänen <jere.leppanen@nokia.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit 7e9269a5.

As Jere notice, commit 35b4f244

 ("sctp: do asoc update earlier
in sctp_sf_do_dupcook_a") only keeps the SHUTDOWN and COOKIE-ACK
with the same asoc, not transport. So we have to bring this patch
back.

Reported-by: Jere Leppänen <jere.leppanen@nokia.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In enic_hard_start_xmit, it calls enic_queue_wq_skb(). Inside
enic_queue_wq_skb, if some error happens, the skb will be freed
by dev_kfree_skb(skb). But the freed skb is still used in
skb_tx_timestamp(skb).

My patch makes enic_queue_wq_skb() return error and goto spin_unlock()
incase of error. The solution is provided by Govind.
See https://lkml.org/lkml/2021/4/30/961.

Fixes: fb7516d4

 ("enic: add sw timestamp support")
Signed-off-by: Lv Yunlong <lyl2019@mail.ustc.edu.cn>
Acked-by: Govindarajulu Varadarajan <gvaradar@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In commit f4da5652

 ("net: stmmac: Add support for external
trigger timestamping"), struct stmmac_priv was declared at line 507
which caused duplicate struct declarations.
Remove later duplicate declaration here.

Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Reviewed-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

A number of PHYs support the PHY tunable to set and get
downshift. However, only 88E1116R enables downshift by default. Extend
this default enabled to all the PHYs that support the downshift
tunable.

Signed-off-by: Maxim Kochetkov <fido_max@inbox.ru>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Xin Long says:

====================
sctp: always send a chunk with the asoc that it belongs to

Currently when processing a duplicate COOKIE-ECHO chunk, a new temp
asoc would be created, then it creates the chunks with the new asoc.
However, later on it uses the old asoc to send these chunks, which
has caused quite a few issues.

This patchset is to fix this and make sure that the COOKIE-ACK and
SHUTDOWN chunks are created with the same asoc that will be used to
send them out.

v1->v2:
  - see Patch 3/3.
====================

Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The same thing should be done for sctp_sf_do_dupcook_b().
Meanwhile, SCTP_CMD_UPDATE_ASSOC cmd can be removed.

v1->v2:
  - Fix the return value in sctp_sf_do_assoc_update().

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This can be reverted as shutdown and cookie_ack chunk are using the
same asoc since the last patch.

This reverts commit 145cb2f7

.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There's a panic that occurs in a few of envs, the call trace is as below:

  [] general protection fault, ... 0x29acd70f1000a: 0000 [#1] SMP PTI
  [] RIP: 0010:sctp_ulpevent_notify_peer_addr_change+0x4b/0x1fa [sctp]
  []  sctp_assoc_control_transport+0x1b9/0x210 [sctp]
  []  sctp_do_8_2_transport_strike.isra.16+0x15c/0x220 [sctp]
  []  sctp_cmd_interpreter.isra.21+0x1231/0x1a10 [sctp]
  []  sctp_do_sm+0xc3/0x2a0 [sctp]
  []  sctp_generate_timeout_event+0x81/0xf0 [sctp]

This is caused by a transport use-after-free issue. When processing a
duplicate COOKIE-ECHO chunk in sctp_sf_do_dupcook_a(), both COOKIE-ACK
and SHUTDOWN chunks are allocated with the transort from the new asoc.
However, later in the sideeffect machine, the old asoc is used to send
them out and old asoc's shutdown_last_sent_to is set to the transport
that SHUTDOWN chunk attached to in sctp_cmd_setup_t2(), which actually
belongs to the new asoc. After the new_asoc is freed and the old asoc
T2 timeout, the old asoc's shutdown_last_sent_to that is already freed
would be accessed in sctp_sf_t2_timer_expire().

Thanks Alexander and Jere for helping dig into this issue.

To fix it, this patch is to do the asoc update first, then allocate
the COOKIE-ACK and SHUTDOWN chunks with the 'updated' old asoc. This
would make more sense, as a chunk from an asoc shouldn't be sent out
with another asoc. We had fixed quite a few issues caused by this.

Fixes: 145cb2f7

 ("sctp: Fix bundling of SHUTDOWN with COOKIE-ACK")
Reported-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Reported-by:  <syzbot+bbe538efd1046586f587@syzkaller.appspotmail.com>
Reported-by: Michal Tesar <mtesar@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add Marc Dionne as a co-maintainer for kafs and rxrpc.

Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove it from the MODULE_AUTHOR statements referencing it.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Variable 'err' is set to zero but this value is never read as it is
overwritten with a new value later on, hence it is a redundant
assignment and can be removed.

Clean up the following clang-analyzer warning:

net/vmw_vsock/vmci_transport.c:948:2: warning: Value stored to 'err' is
never read [clang-analyzer-deadcode.DeadStores]

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Huazhong Tan says:

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

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

Signed-off-by: David S. Miller <davem@davemloft.net>

If selftest and reset are performed at the same time, the phy
loopback setting may be still in enable state after the reset,
and device cannot link up. So fix this issue by disabling phy
loopback before phy_start().

Fixes: 256727da

 ("net: hns3: Add MDIO support to HNS3 Ethernet driver for hip08 SoC")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

HW error and global reset are reported through MSIX interrupts.
The same error may be reported to different functions at the
same time. When global reset begins, the pending reset request
set by this error is unnecessary. So clear the pending reset
request after the reset is complete to avoid the repeated reset.

Fixes: f6162d44

 ("net: hns3: add handling of hw errors reported through MSIX")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, netif_tx_stop_all_queues() is used to ensure that
the xmit is not running, but for the concurrent case it will
not take effect, since netif_tx_stop_all_queues() just sets
a flag without locking to indicate that the xmit queue(s)
should not be run.

So use netif_tx_disable() to replace netif_tx_stop_all_queues(),
it takes the xmit queue lock while marking the queue stopped.

Fixes: 76ad4f0e

 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When skb->ip_summed is CHECKSUM_PARTIAL, for non-tunnel udp packet,
which has a dest port as the IANA assigned, the hardware is expected
to do the checksum offload, but the hardware whose version is below
V3 will not do the checksum offload when udp dest port is 4790.

So fixes it by doing the checksum in software for this case.

Fixes: 76ad4f0e

 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An issue found when network interface is down and up again, FPE handshake
fails to trigger. This is due to __FPE_REMOVING bit remains being set in
stmmac_fpe_stop_wq() but not cleared in stmmac_fpe_start_wq(). This
cause FPE workqueue task, stmmac_fpe_lp_task() not able to be executed.

To fix this, add clearing __FPE_REMOVING bit in stmmac_fpe_start_wq().

Fixes: 5a558611

 ("net: stmmac: support FPE link partner hand-shaking procedure")
Signed-off-by: Mohammad Athari Bin Ismail <mohammad.athari.ismail@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ksz_switch_alloc() will return NULL only if allocation is failed. So,
the proper return value is -ENOMEM.

Fixes: 60a36476

 ("net: dsa: microchip: Add Microchip KSZ8863 SMI based driver support")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix possible NULL pointer dereference in case devm_kzalloc() failed to
allocate memory

Fixes: cc13e52c

 ("net: dsa: microchip: Add Microchip KSZ8863 SPI based driver support")
Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix possible NULL pointer dereference in case devm_kzalloc() failed to
allocate memory.

Fixes: 60a36476

 ("net: dsa: microchip: Add Microchip KSZ8863 SMI based driver support")
Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Variable 'err' is set to -EIO but this value is never read as it is
overwritten with a new value later on, hence it is a redundant
assignment and can be removed.

Clean up the following clang-analyzer warning:
drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c:1195:2: warning: Value
stored to 'err' is never read [clang-analyzer-deadcode.DeadStores]

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Variable queue is set to bp->queues but these values is not used as it
is overwritten later on, hence redundant assignment  can be removed.

Cleans up the following clang-analyzer warning:

drivers/net/ethernet/cadence/macb_main.c:4919:21: warning: Value stored
to 'queue' during its initialization is never read
[clang-analyzer-deadcode.DeadStores].

drivers/net/ethernet/cadence/macb_main.c:4832:21: warning: Value stored
to 'queue' during its initialization is never read
[clang-analyzer-deadcode.DeadStores].

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

His email bounces with permanent error "550 Invalid recipient". His last
email was from 2020-09-09 on the LKML and he seems to have left TI.

Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>

His email bounces with permanent error "550 Invalid recipient". His last
email on the LKML was from 2015-10-22 on the LKML.

Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>

Huazhong Tan says:

====================
net: hns3: add some fixes for -net

This series adds some fixes for the HNS3 ethernet driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

In some cases, the device is not initialized because reset failed.
If another task calls hns3_reset_notify_up_enet() before reset
retry, it will cause an error since uninitialized pointer access.
So add check for HNS3_NIC_STATE_INITED before calling
hns3_nic_net_open() in hns3_reset_notify_up_enet().

Fixes: bb6b94a8

 ("net: hns3: Add reset interface implementation in client")
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>

The message sent to VF should be initialized, otherwise random
value of some contents may cause improper processing by the target.
So add a initialization to message in hclge_get_link_mode().

Fixes: 9194d18b

 ("net: hns3: fix the problem that the supported port is empty")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

According to the UM, the type and enable status of igu_egu_hw_err
should be configured separately. Currently, the type field is
incorrect when disable this error. So fix it by configuring these
two fields separately.

Fixes: bf1faf94

 ("net: hns3: Add enable and process hw errors from IGU, EGU and NCSI")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Variable 'err' is set to -ENOMEM but this value is never read as it is
overwritten with a new value later on, hence the 'If statements' and
assignments are redundantand and can be removed.

Cleans up the following clang-analyzer warning:

net/ipv6/seg6.c:126:4: warning: Value stored to 'err' is never read
[clang-analyzer-deadcode.DeadStores]

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is a crash in the function br_get_link_af_size_filtered,
as the port_exists(dev) is true and the rx_handler_data of dev is NULL.
But the rx_handler_data of dev is correct saved in vmcore.

The oops looks something like:
 ...
 pc : br_get_link_af_size_filtered+0x28/0x1c8 [bridge]
 ...
 Call trace:
  br_get_link_af_size_filtered+0x28/0x1c8 [bridge]
  if_nlmsg_size+0x180/0x1b0
  rtnl_calcit.isra.12+0xf8/0x148
  rtnetlink_rcv_msg+0x334/0x370
  netlink_rcv_skb+0x64/0x130
  rtnetlink_rcv+0x28/0x38
  netlink_unicast+0x1f0/0x250
  netlink_sendmsg+0x310/0x378
  sock_sendmsg+0x4c/0x70
  __sys_sendto+0x120/0x150
  __arm64_sys_sendto+0x30/0x40
  el0_svc_common+0x78/0x130
  el0_svc_handler+0x38/0x78
  el0_svc+0x8/0xc

In br_add_if(), we found there is no guarantee that
assigning rx_handler_data to dev->rx_handler_data
will before setting the IFF_BRIDGE_PORT bit of priv_flags.
So there is a possible data competition:

CPU 0:                                                        CPU 1:
(RCU read lock)                                               (RTNL lock)
rtnl_calcit()                                                 br_add_slave()
  if_nlmsg_size()                                               br_add_if()
    br_get_link_af_size_filtered()                              -> netdev_rx_handler_register
                                                                    ...
                                                                    // The order is not guaranteed
      ...                                                           -> dev->priv_flags |= IFF_BRIDGE_PORT;
      // The IFF_BRIDGE_PORT bit of priv_flags has been set
      -> if (br_port_exists(dev)) {
        // The dev->rx_handler_data has NOT been assigned
        -> p = br_port_get_rcu(dev);
        ....
                                                                    -> rcu_assign_pointer(dev->rx_handler_data, rx_handler_data);
                                                                     ...

Fix it in br_get_link_af_size_filtered, using br_port_get_check_rcu() and checking the return value.

Signed-off-by: Zhang Zhengming <zhangzhengming@huawei.com>
Reviewed-by: Zhao Lei <zhaolei69@huawei.com>
Reviewed-by: Wang Xiaogang <wangxiaogang3@huawei.com>
Suggested-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Davide Caratti says:

====================
fix stack OOB read while fragmenting IPv4 packets

- patch 1/2 fixes openvswitch IPv4 fragmentation, that does a stack OOB
read after commit d52e5a7e

 ("ipv4: lock mtu in fnhe when received
PMTU < net.ipv4.route.min_pmt")
- patch 2/2 fixes the same issue in TC 'sch_frag' code
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

when 'act_mirred' tries to fragment IPv4 packets that had been previously
re-assembled using 'act_ct', splats like the following can be observed on
kernels built with KASAN:

 BUG: KASAN: stack-out-of-bounds in ip_do_fragment+0x1b03/0x1f60
 Read of size 1 at addr ffff888147009574 by task ping/947

 CPU: 0 PID: 947 Comm: ping Not tainted 5.12.0-rc6+ #418
 Hardware name: Red Hat KVM, BIOS 1.11.1-4.module+el8.1.0+4066+0f1aadab 04/01/2014
 Call Trace:
  <IRQ>
  dump_stack+0x92/0xc1
  print_address_description.constprop.7+0x1a/0x150
  kasan_report.cold.13+0x7f/0x111
  ip_do_fragment+0x1b03/0x1f60
  sch_fragment+0x4bf/0xe40
  tcf_mirred_act+0xc3d/0x11a0 [act_mirred]
  tcf_action_exec+0x104/0x3e0
  fl_classify+0x49a/0x5e0 [cls_flower]
  tcf_classify_ingress+0x18a/0x820
  __netif_receive_skb_core+0xae7/0x3340
  __netif_receive_skb_one_core+0xb6/0x1b0
  process_backlog+0x1ef/0x6c0
  __napi_poll+0xaa/0x500
  net_rx_action+0x702/0xac0
  __do_softirq+0x1e4/0x97f
  do_softirq+0x71/0x90
  </IRQ>
  __local_bh_enable_ip+0xdb/0xf0
  ip_finish_output2+0x760/0x2120
  ip_do_fragment+0x15a5/0x1f60
  __ip_finish_output+0x4c2/0xea0
  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+0x33/0x40
  entry_SYSCALL_64_after_hwframe+0x44/0xae
 RIP: 0033:0x7f82e13853eb
 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:00007ffe01fad888 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
 RAX: ffffffffffffffda RBX: 00005571aac13700 RCX: 00007f82e13853eb
 RDX: 0000000000002330 RSI: 00005571aac13700 RDI: 0000000000000003
 RBP: 0000000000002330 R08: 00005571aac10500 R09: 0000000000000010
 R10: 0000000000000000 R11: 0000000000000246 R12: 00007ffe01faefb0
 R13: 00007ffe01fad890 R14: 00007ffe01fad980 R15: 00005571aac0f0a0

 The buggy address belongs to the page:
 page:000000001dff2e03 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x147009
 flags: 0x17ffffc0001000(reserved)
 raw: 0017ffffc0001000 ffffea00051c0248 ffffea00051c0248 0000000000000000
 raw: 0000000000000000 0000000000000000 00000001ffffffff 0000000000000000
 page dumped because: kasan: bad access detected

 Memory state around the buggy address:
  ffff888147009400: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  ffff888147009480: f1 f1 f1 f1 04 f2 f2 f2 f2 f2 f2 f2 00 00 00 00
 >ffff888147009500: 00 00 00 00 00 00 00 00 00 00 f2 f2 f2 f2 f2 f2
                                                              ^
  ffff888147009580: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  ffff888147009600: 00 00 00 00 00 00 00 00 00 00 00 00 00 f2 f2 f2

for IPv4 packets, sch_fragment() uses a temporary struct dst_entry. Then,
in the following call graph:

  ip_do_fragment()
    ip_skb_dst_mtu()
      ip_dst_mtu_maybe_forward()
        ip_mtu_locked()

the pointer to struct dst_entry is used as pointer to struct rtable: this
turns the access to struct members like rt_mtu_locked into an OOB read in
the stack. Fix this changing the temporary variable used for IPv4 packets
in sch_fragment(), similarly to what is done for IPv6 few lines below.

Fixes: c129412f

 ("net/sched: sch_frag: add generic packet fragment support.")
Cc: <stable@vger.kernel.org> # 5.11
Reported-by: Shuang Li <shuali@redhat.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>