Merge branch 'rswitch-jumbo-frames'

Yoshihiro Shimoda says:

====================
net: rswitch: Add jumbo frames support

This patch series is based on the latest net-next.git / main branch.

Changes from v3:
https://lore.kernel.org/all/20231204012058.3876078-1-yoshihiro.shimoda.uh@renesas.com/
 - Based on the latest net-next.git / main branch.
 - Modify for code consistancy in the patch 3/9.
 - Add a condition in the patch 3/9.
 - Fix usage of dma_addr in the patch 8/9.

Changes from v2:
https://lore.kernel.org/all/20231201054655.3731772-1-yoshihiro.shimoda.uh@renesas.com/
 - Based on the latest net-next.git / main branch.
 - Fix using a variable in the patch 8/9.
 - Add Reviewed-by tag in the patch 1/9.

Changes from v1:
https://lore.kernel.org/all/20231127115334.3670790-1-yoshihiro.shimoda.uh@renesas.com/
 - Based on the latest net-next.git / main branch.
 - Fix commit descriptions (s/near the future/the near future/).
====================

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

net: rswitch: Allow jumbo frames

Allow jumbo frames by changing maximum MTU size and number of RX queues.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Add jumbo frames handling for TX

If the driver would like to transmit a jumbo frame like 2KiB or more,
it should be split into multiple queues. In the near future, to support
this, add handling specific descriptor types F{START,MID,END}. However,
such jumbo frames will not happen yet because the maximum MTU size is
still default for now.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Add jumbo frames handling for RX

If this hardware receives a jumbo frame like 2KiB or more, it will be
split into multiple queues. In the near future, to support this,
add handling specific descriptor types F{START,MID,END}. However, such
jumbo frames will not happen yet because the maximum MTU size is still
default for now.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Set GWMDNC register

To support jumbo frames, set GWMDNC register with acceptable maximum
values for TX and RX.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Add a setting ext descriptor function

If the driver would like to transmit a jumbo frame like 2KiB or more,
it should be split into multiple queues. In the near future, to support
this, add a setting ext descriptor function to improve code readability.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Add unmap_addrs instead of dma address in each desc

If the driver would like to transmit a jumbo frame like 2KiB or more,
it should be split into multiple queues. In the near future, to support
this, add unmap_addrs array to unmap dma mapping address instead of dma
address in each TX descriptor because the descriptors may not have
the top dma address.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Use build_skb() for RX

If this hardware receives a jumbo frame like 2KiB or more, it will be
split into multiple queues. In the near future, to support this, use
build_skb() instead of netdev_alloc_skb_ip_align().

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Use unsigned int for desc related array index

Array index should not be negative, so use unsigned int for
descriptors related array index.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: rswitch: Drop unused argument/return value

Drop unused argument and return value of rswitch_tx_free() to
simplify the code.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: mdio_bus: replace deprecated strncpy with strscpy

strncpy() is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous string
interfaces.

We expect mdiodev->modalias to be NUL-terminated based on its usage with
strcmp():
|       return strcmp(mdiodev->modalias, drv->name) == 0;

Moreover, mdiodev->modalias is already zero-allocated:
|       mdiodev = kzalloc(sizeof(*mdiodev), GFP_KERNEL);
... which means the NUL-padding strncpy provides is not necessary.

Considering the above, a suitable replacement is `strscpy` [2] due to
the fact that it guarantees NUL-termination on the destination buffer
without unnecessarily NUL-padding.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings
Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html
Link: https://github.com/KSPP/linux/issues/90
Cc: linux-hardening@vger.kernel.org
Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

qlcnic: replace deprecated strncpy with strscpy

strncpy() is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous string
interfaces.

We expect fw_info->fw_file_name to be NUL-terminated based on its use
within _request_firmware_prepare() wherein `name` refers to it:
|       if (firmware_request_builtin_buf(firmware, name, dbuf, size)) {
|               dev_dbg(device, "using built-in %s\n", name);
|               return 0; /* assigned */
|       }
... and with firmware_request_builtin() also via `name`:
|       if (strcmp(name, b_fw->name) == 0) {

There is no evidence that NUL-padding is required.

Additionally replace size macro (QLC_FW_FILE_NAME_LEN) with
sizeof(fw_info->fw_file_name) to more directly tie the maximum buffer
size to the destination buffer:

Considering the above, a suitable replacement is `strscpy` [2] due to
the fact that it guarantees NUL-termination on the destination buffer
without unnecessarily NUL-padding.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings
Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html
Link: https://github.com/KSPP/linux/issues/90
Cc: linux-hardening@vger.kernel.org
Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ena: replace deprecated strncpy with strscpy

`strncpy` is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous string
interfaces.

A suitable replacement is `strscpy` [2] due to the fact that it
guarantees NUL-termination on the destination buffer without
unnecessarily NUL-padding.

host_info allocation is done in ena_com_allocate_host_info() via
dma_alloc_coherent() and is not zero initialized by alloc_etherdev_mq().

However zero initialization of the destination doesn't matter in this case,
because strscpy() guarantees a NULL termination.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings
Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html
Link: https://github.com/KSPP/linux/issues/90
Cc: linux-hardening@vger.kernel.org
Signed-off-by: Justin Stitt <justinstitt@google.com>
Acked-by: Arthur Kiyanovski <akiyano@amazon.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: do not check fib6_has_expires() in fib6_info_release()

My prior patch went a bit too far, because apparently fib6_has_expires()
could be true while f6i->gc_link is not hashed yet.

fib6_set_expires_locked() can indeed set RTF_EXPIRES
while f6i->fib6_table is NULL.

Original syzbot reports were about corruptions caused
by dangling f6i->gc_link.

Fixes: 5a08d0065a91 ("ipv6: add debug checks in fib6_info_release()")
Reported-by: syzbot+c15aa445274af8674f41@syzkaller.appspotmail.com
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Kui-Feng Lee <thinker.li@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20231207201322.549000-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: sysfs: fix locking in carrier read

My previous patch added a call to linkwatch_sync_dev(),
but that of course needs to be called under RTNL, which
I missed earlier, but now saw RCU warnings from.

Fix that by acquiring the RTNL in a similar fashion to
how other files do it here.

Fixes: facd15dfd691 ("net: core: synchronize link-watch when carrier is queried")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20231206172122.859df6ba937f.I9c80608bcfbab171943ff4942b52dbd5e97fe06e@changeid
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

netlink: Return unsigned value for nla_len()

The return value from nla_len() is never expected to be negative, and can
never be more than struct nlattr::nla_len (a u16). Adjust the prototype
on the function. This will let GCC's value range optimization passes
know that the return can never be negative, and can never be larger than
u16. As recently discussed[1], this silences the following warning in
GCC 12+:

net/wireless/nl80211.c: In function 'nl80211_set_cqm_rssi.isra':
net/wireless/nl80211.c:12892:17: warning: 'memcpy' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overflow=]
12892 |                 memcpy(cqm_config->rssi_thresholds, thresholds,
      |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12893 |                        flex_array_size(cqm_config, rssi_thresholds,
      |                        ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12894 |                                        n_thresholds));
      |                                        ~~~~~~~~~~~~~~

A future change would be to clamp the subtraction to make sure it never
wraps around if nla_len is somehow less than NLA_HDRLEN, which would
have the additional benefit of being defensive in the face of nlattr
corruption or logic errors.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202311090752.hWcJWAHL-lkp@intel.com/ [1]
Cc: Johannes Berg <johannes@sipsolutions.net>
Cc: Jeff Johnson <quic_jjohnson@quicinc.com>
Cc: Michael Walle <mwalle@kernel.org>
Cc: Max Schulze <max.schulze@online.de>
Link: https://lore.kernel.org/r/20231202202539.it.704-kees@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20231206205904.make.018-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: microchip: use DSA_TAG_PROTO without _VALUE define

Correct the use of define DSA_TAG_PROTO_LAN937X_VALUE to
DSA_TAG_PROTO_LAN937X to improve readability.

Signed-off-by: Sean Nyekjaer <sean@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Link: https://lore.kernel.org/r/20231206160124.1935451-1-sean@geanix.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Use READ/WRITE_ONCE() for IP local_port_range.

Commit 227b60f5102cd added a seqlock to ensure that the low and high
port numbers were always updated together.
This is overkill because the two 16bit port numbers can be held in
a u32 and read/written in a single instruction.

More recently 91d0b78c5177f added support for finer per-socket limits.
The user-supplied value is 'high << 16 | low' but they are held
separately and the socket options protected by the socket lock.

Use a u32 containing 'high << 16 | low' for both the 'net' and 'sk'
fields and use READ_ONCE()/WRITE_ONCE() to ensure both values are
always updated together.

Change (the now trival) inet_get_local_port_range() to a static inline
to optimise the calling code.
(In particular avoiding returning integers by reference.)

Signed-off-by: David Laight <david.laight@aculab.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Acked-by: Mat Martineau <martineau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/4e505d4198e946a8be03fb1b4c3072b0@AcuMS.aculab.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'net-selftests-unique-namespace'

Hangbin Liu says:

====================
Convert net selftests to run in unique namespace (Part 2)

Here is the 2nd part of converting net selftests to run in unique namespace.
This part converts all bridge, vxlan, vrf tests.

Here is the part 1 link:
https://lore.kernel.org/netdev/20231202020110.362433-1-liuhangbin@gmail.com
====================

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

selftests/net: convert vrf-xfrm-tests.sh to run it in unique namespace

Here is the test result after conversion.

]# ./vrf-xfrm-tests.sh

No qdisc on VRF device
TEST: IPv4 no xfrm policy                                           [ OK ]
TEST: IPv6 no xfrm policy                                           [ OK ]
TEST: IPv4 xfrm policy based on address                             [ OK ]
TEST: IPv6 xfrm policy based on address                             [ OK ]
TEST: IPv6 xfrm policy with VRF in selector                         [ OK ]
TEST: IPv4 xfrm policy with xfrm device                             [ OK ]
TEST: IPv6 xfrm policy with xfrm device                             [ OK ]

netem qdisc on VRF device
TEST: IPv4 no xfrm policy                                           [ OK ]
TEST: IPv6 no xfrm policy                                           [ OK ]
TEST: IPv4 xfrm policy based on address                             [ OK ]
TEST: IPv6 xfrm policy based on address                             [ OK ]
TEST: IPv6 xfrm policy with VRF in selector                         [ OK ]
TEST: IPv4 xfrm policy with xfrm device                             [ OK ]
TEST: IPv6 xfrm policy with xfrm device                             [ OK ]

Tests passed:  14
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert vrf_strict_mode_test.sh to run it in unique namespace

Here is the test result after conversion.

 ]# ./vrf_strict_mode_test.sh

 ################################################################################
 TEST SECTION: VRF strict_mode test on init network namespace
 ################################################################################

     TEST: init: net.vrf.strict_mode is available                        [ OK ]

     TEST: init: strict_mode=0 by default, 0 vrfs                        [ OK ]

 ...

     TEST: init: check strict_mode=1                                     [ OK ]

     TEST: testns-HvoZkB: check strict_mode=0                            [ OK ]

 Tests passed:  37
 Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert vrf_route_leaking.sh to run it in unique namespace

Here is the test result after conversion.

 ]# ./vrf_route_leaking.sh

 ###########################################################################
 IPv4 (sym route): VRF ICMP ttl error route lookup ping
 ###########################################################################

 TEST: Basic IPv4 connectivity                                       [ OK ]
 TEST: Ping received ICMP ttl exceeded                               [ OK ]

 ...

 TEST: Basic IPv6 connectivity                                       [ OK ]
 TEST: Traceroute6 reports a hop on r1                               [ OK ]

 Tests passed:  18
 Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_vxlan_vnifiltering.sh to run it in unique namespace

Here is the test result after conversion.

]# ./test_vxlan_vnifiltering.sh
    TEST: Create traditional vxlan device                               [ OK ]
    TEST: Cannot create vnifilter device without external flag          [ OK ]
    TEST: Creating external vxlan device with vnifilter flag            [ OK ]
...
    TEST: VM connectivity over traditional vxlan (ipv6 default rdst)    [ OK ]
    TEST: VM connectivity over metadata nonfiltering vxlan (ipv4 default rdst)  [ OK ]

Tests passed:  27
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_vxlan_under_vrf.sh to run it in unique namespace

Here is the test result after conversion.

]# ./test_vxlan_under_vrf.sh
Checking HV connectivity                                           [ OK ]
Check VM connectivity through VXLAN (underlay in the default VRF)  [ OK ]
Check VM connectivity through VXLAN (underlay in a VRF)            [ OK ]

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_vxlan_nolocalbypass.sh to run it in unique namespace

Here is the test result after conversion.

]# ./test_vxlan_nolocalbypass.sh
TEST: localbypass enabled                                           [ OK ]
TEST: Packet received by local VXLAN device - localbypass           [ OK ]
TEST: localbypass disabled                                          [ OK ]
TEST: Packet not received by local VXLAN device - nolocalbypass     [ OK ]
TEST: localbypass enabled                                           [ OK ]
TEST: Packet received by local VXLAN device - localbypass           [ OK ]

Tests passed:   6
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_vxlan_mdb.sh to run it in unique namespace

Here is the test result after conversion.

]# ./test_vxlan_mdb.sh

Control path: Basic (*, G) operations - IPv4 overlay / IPv4 underlay
--------------------------------------------------------------------
TEST: MDB entry addition                                            [ OK ]

...

Data path: MDB torture test - IPv6 overlay / IPv6 underlay
----------------------------------------------------------
TEST: Torture test                                                  [ OK ]

Tests passed: 620
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_bridge_neigh_suppress.sh to run it in unique namespace

Here is the test result after conversion.

]# ./test_bridge_neigh_suppress.sh

Per-port ARP suppression - VLAN 10
----------------------------------
TEST: arping                                                        [ OK ]
TEST: ARP suppression                                               [ OK ]

...

TEST: NS suppression (VLAN 20)                                      [ OK ]

Tests passed: 148
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

selftests/net: convert test_bridge_backup_port.sh to run it in unique namespace

There is no h1 h2 actually. Remove it. Here is the test result after
conversion.

]# ./test_bridge_backup_port.sh

Backup port
-----------
TEST: Forwarding out of swp1                                        [ OK ]
TEST: No forwarding out of vx0                                      [ OK ]
TEST: swp1 carrier off                                              [ OK ]
TEST: No forwarding out of swp1                                     [ OK ]
...
Backup nexthop ID - ping
------------------------
TEST: Ping with backup nexthop ID                                   [ OK ]
TEST: Ping after disabling backup nexthop ID                        [ OK ]

Backup nexthop ID - torture test
--------------------------------
TEST: Torture test                                                  [ OK ]

Tests passed:  83
Tests failed:   0

Acked-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'ethtool_puts'

Justin Stitt says:

====================
ethtool: Add ethtool_puts()

This series aims to implement ethtool_puts() and send out a wave 1 of
conversions from ethtool_sprintf(). There's also a checkpatch patch
included to check for the cases listed below.

This was sparked from recent discussion here [1]

The conversions are used in cases where ethtool_sprintf() was being used
with just two arguments:
|       ethtool_sprintf(&data, buffer[i].name);
or when it's used with format string: "%s"
|       ethtool_sprintf(&data, "%s", buffer[i].name);
which both now become:
|       ethtool_puts(&data, buffer[i].name);

The first case commonly triggers a -Wformat-security warning with Clang
due to potential problems with format flags present in the strings [3].

The second is just a bit weird with a plain-ol' "%s".

Changes found with Cocci [4] and grep [5].

[1]: https://lore.kernel.org/all/202310141935.B326C9E@keescook/
[2]: https://lore.kernel.org/all/?q=dfb%3Aethtool_sprintf+AND+f%3Ajustinstitt
[3]: https://lore.kernel.org/all/202310101528.9496539BE@keescook/
[4]: (script authored by Kees w/ modifications from Joe)
@replace_2_args@
expression BUF;
expression VAR;
@@

-       ethtool_sprintf(BUF, VAR)
+       ethtool_puts(BUF, VAR)

@replace_3_args@
expression BUF;
expression VAR;
@@

-       ethtool_sprintf(BUF, "%s", VAR)
+       ethtool_puts(BUF, VAR)

-       ethtool_sprintf(&BUF, "%s", VAR)
+       ethtool_puts(&BUF, VAR)

[5]: $ rg "ethtool_sprintf\(\s*[^,)]+\s*,\s*[^,)]+\s*\)"

Signed-off-by: Justin Stitt <justinstitt@google.com>
---
Changes in v5:
- updated documentation to include info about the lack of a trailing newline
  (Thanks Russell)
- rebased onto mainline
- Link to v4: https://lore.kernel.org/r/20231102-ethtool_puts_impl-v4-0-14e1e9278496@google.com

Changes in v4:
- update documentation to match:
  https://lore.kernel.org/all/20231028192511.100001-1-andrew@lunn.ch/

- Link to v3: https://lore.kernel.org/r/20231027-ethtool_puts_impl-v3-0-3466ac679304@google.com

Changes in v3:
- fix force_speed_maps merge conflict + formatting (thanks Vladimir)
- rebase onto net-next (thanks Andrew, Vladimir)
- change subject (thanks Vladimir)
- fix checkpatch formatting + implementation (thanks Joe)
- Link to v2: https://lore.kernel.org/r/20231026-ethtool_puts_impl-v2-0-0d67cbdd0538@google.com

Changes in v2:
- wrap lines better in replacement (thanks Joe, Kees)
- add --fix to checkpatch (thanks Joe)
- clean up checkpatch formatting (thanks Joe, et al.)
- rebase against next
- Link to v1: https://lore.kernel.org/r/20231025-ethtool_puts_impl-v1-0-6a53a93d3b72@google.com
====================

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

net: Convert some ethtool_sprintf() to ethtool_puts()

This patch converts some basic cases of ethtool_sprintf() to
ethtool_puts().

The conversions are used in cases where ethtool_sprintf() was being used
with just two arguments:
|       ethtool_sprintf(&data, buffer[i].name);
or when it's used with format string: "%s"
|       ethtool_sprintf(&data, "%s", buffer[i].name);
which both now become:
|       ethtool_puts(&data, buffer[i].name);

Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

checkpatch: add ethtool_sprintf rules

Add some warnings for using ethtool_sprintf() where a simple
ethtool_puts() would suffice.

The two cases are:

1) Use ethtool_sprintf() with just two arguments:
|       ethtool_sprintf(&data, driver[i].name);
or
2) Use ethtool_sprintf() with a standalone "%s" fmt string:
|       ethtool_sprintf(&data, "%s", driver[i].name);

The former may cause -Wformat-security warnings while the latter is just
not preferred. Both are safely in the category of warnings, not errors.

Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ethtool: Implement ethtool_puts()

Use strscpy() to implement ethtool_puts().

Functionally the same as ethtool_sprintf() when it's used with two
arguments or with just "%s" format specifier.

Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Madhuri Sripada <madhuri.sripada@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'nfp-add-ext_ack-messages-to-supported-callbacks'

Louis Peens says:

====================
nfp: add ext_ack messages to supported callbacks

This is a mostly cosmetic series to add error messages to devlink and
ethtool callbacks which supports them but did not have them added in the
nfp driver yet.
====================

Link: https://lore.kernel.org/r/20231206151209.20296-1-louis.peens@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfp: devlink: add extended ack report messages

Add descriptive error messages to common devlink failures to
be more user friendly.

Signed-off-by: Ryno Swart <ryno.swart@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Link: https://lore.kernel.org/r/20231206151209.20296-3-louis.peens@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

nfp: ethtool: add extended ack report messages

Add descriptive error messages to common ethtool failures to be more
user friendly.

Update `nfp_net_coalesce_para_check` to only check one argument, which
facilitates unique error messages.

Additionally, three error codes are updated to `EOPNOTSUPP` to reflect
that these operations are not supported.

Signed-off-by: Ryno Swart <ryno.swart@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Link: https://lore.kernel.org/r/20231206151209.20296-2-louis.peens@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge git://git./linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

Conflicts:

drivers/net/ethernet/stmicro/stmmac/dwmac5.c
drivers/net/ethernet/stmicro/stmmac/dwmac5.h
drivers/net/ethernet/stmicro/stmmac/dwxgmac2_core.c
drivers/net/ethernet/stmicro/stmmac/hwif.h
  37e4b8df27bc ("net: stmmac: fix FPE events losing")
  c3f3b97238f6 ("net: stmmac: Refactor EST implementation")
https://lore.kernel.org/all/20231206110306.01e91114@canb.auug.org.au/

Adjacent changes:

net/ipv4/tcp_ao.c
  9396c4ee93f9 ("net/tcp: Don't store TCP-AO maclen on reqsk")
  7b0f570f879a ("tcp: Move TCP-AO bits from cookie_v[46]_check() to tcp_ao_syncookie().")

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'net-6.7-rc5' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bpf and netfilter.

  Current release - regressions:

   - veth: fix packet segmentation in veth_convert_skb_to_xdp_buff

  Current release - new code bugs:

   - tcp: assorted fixes to the new Auth Option support

  Older releases - regressions:

   - tcp: fix mid stream window clamp

   - tls: fix incorrect splice handling

   - ipv4: ip_gre: handle skb_pull() failure in ipgre_xmit()

   - dsa: mv88e6xxx: restore USXGMII support for 6393X

   - arcnet: restore support for multiple Sohard Arcnet cards

  Older releases - always broken:

   - tcp: do not accept ACK of bytes we never sent

   - require admin privileges to receive packet traces via netlink

   - packet: move reference count in packet_sock to atomic_long_t

   - bpf:
      - fix incorrect branch offset comparison with cpu=v4
      - fix prog_array_map_poke_run map poke update

   - netfilter:
      - three fixes for crashes on bad admin commands
      - xt_owner: fix race accessing sk->sk_socket, TOCTOU null-deref
      - nf_tables: fix 'exist' matching on bigendian arches

   - leds: netdev: fix RTNL handling to prevent potential deadlock

   - eth: tg3: prevent races in error/reset handling

   - eth: r8169: fix rtl8125b PAUSE storm when suspended

   - eth: r8152: improve reset and surprise removal handling

   - eth: hns: fix race between changing features and sending

   - eth: nfp: fix sleep in atomic for bonding offload"

* tag 'net-6.7-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (62 commits)
  vsock/virtio: fix "comparison of distinct pointer types lacks a cast" warning
  net/smc: fix missing byte order conversion in CLC handshake
  net: dsa: microchip: provide a list of valid protocols for xmit handler
  drop_monitor: Require 'CAP_SYS_ADMIN' when joining "events" group
  psample: Require 'CAP_NET_ADMIN' when joining "packets" group
  bpf: sockmap, updating the sg structure should also update curr
  net: tls, update curr on splice as well
  nfp: flower: fix for take a mutex lock in soft irq context and rcu lock
  net: dsa: mv88e6xxx: Restore USXGMII support for 6393X
  tcp: do not accept ACK of bytes we never sent
  selftests/bpf: Add test for early update in prog_array_map_poke_run
  bpf: Fix prog_array_map_poke_run map poke update
  netfilter: xt_owner: Fix for unsafe access of sk->sk_socket
  netfilter: nf_tables: validate family when identifying table via handle
  netfilter: nf_tables: bail out on mismatching dynset and set expressions
  netfilter: nf_tables: fix 'exist' matching on bigendian arches
  netfilter: nft_set_pipapo: skip inactive elements during set walk
  netfilter: bpf: fix bad registration on nf_defrag
  leds: trigger: netdev: fix RTNL handling to prevent potential deadlock
  octeontx2-af: Update Tx link register range
  ...

Merge tag 'cgroup-for-6.7-rc4-fixes' of git://git./linux/kernel/git/tj/cgroup

Pull cgroup fix from Tejun Heo:
 "Just one fix.

  Commit f5d39b020809 ("freezer,sched: Rewrite core freezer logic")
  changed how freezing state is recorded which made cgroup_freezing()
  disagree with the actual state of the task while thawing triggering a
  warning. Fix it by updating cgroup_freezing()"

* tag 'cgroup-for-6.7-rc4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup_freezer: cgroup_freezing: Check if not frozen

Merge tag 'wq-for-6.7-rc4-fixes' of git://git./linux/kernel/git/tj/wq

Pull workqueue fix from Tejun Heo:
 "Just one patch to fix a bug which can crash the kernel if the
  housekeeping and wq_unbound_cpu cpumask configuration combination
  leaves the latter empty"

* tag 'wq-for-6.7-rc4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: Make sure that wq_unbound_cpumask is never empty

Merge tag 'regmap-fix-v6.7-rc4' of git://git./linux/kernel/git/broonie/regmap

Pull regmap fix from Mark Brown:
 "An incremental fix for the fix introduced during the merge window for
  caching of the selector for windowed register ranges. We were
  incorrectly leaking an error code in the case where the last selector
  accessed was for some reason not cached"

* tag 'regmap-fix-v6.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: fix bogus error on regcache_sync success

Merge tag 'devicetree-fixes-for-6.7-1' of git://git./linux/kernel/git/robh/linux

Pull devicetree fixes from Rob Herring:

 - Fix dt-extract-compatibles for builds with in tree build directory

 - Drop Xinlei Lee <xinlei.lee@mediatek.com> bouncing email

 - Fix the of_reconfig_get_state_change() return value documentation

 - Add missing #power-domain-cells property to QCom MPM

 - Fix warnings in i.MX LCDIF and adi,adv7533

* tag 'devicetree-fixes-for-6.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: display: adi,adv75xx: Document #sound-dai-cells
  dt-bindings: lcdif: Properly describe the i.MX23 interrupts
  dt-bindings: interrupt-controller: Allow #power-domain-cells
  of: dynamic: Fix of_reconfig_get_state_change() return value documentation
  dt-bindings: display: mediatek: dsi: remove Xinlei's mail
  dt: dt-extract-compatibles: Don't follow symlinks when walking tree

Merge tag 'platform-drivers-x86-v6.7-3' of git://git./linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fixes from Ilpo Järvinen:

 - Fix i8042 filter resource handling, input, and suspend issues in
   asus-wmi

 - Skip zero instance WMI blocks to avoid issues with some laptops

 - Differentiate dev/production keys in mlxbf-bootctl

 - Correct surface serdev related return value to avoid leaking errno
   into userspace

 - Error checking fixes

* tag 'platform-drivers-x86-v6.7-3' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/mellanox: Check devm_hwmon_device_register_with_groups() return value
  platform/mellanox: Add null pointer checks for devm_kasprintf()
  mlxbf-bootctl: correctly identify secure boot with development keys
  platform/x86: wmi: Skip blocks with zero instances
  platform/surface: aggregator: fix recv_buf() return value
  platform/x86: asus-wmi: disable USB0 hub on ROG Ally before suspend
  platform/x86: asus-wmi: Filter Volume key presses if also reported via atkbd
  platform/x86: asus-wmi: Change q500a_i8042_filter() into a generic i8042-filter
  platform/x86: asus-wmi: Move i8042 filter install to shared asus-wmi code

Merge tag 'x86-int80-20231207' of git://git./linux/kernel/git/tip/tip

Pull x86 int80 fixes from Dave Hansen:
 "Avoid VMM misuse of 'int 0x80' handling in TDX and SEV guests.

  It also has the very nice side effect of getting rid of a bunch of
  assembly entry code"

* tag 'x86-int80-20231207' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/tdx: Allow 32-bit emulation by default
  x86/entry: Do not allow external 0x80 interrupts
  x86/entry: Convert INT 0x80 emulation to IDTENTRY
  x86/coco: Disable 32-bit emulation by default on TDX and SEV

vsock/virtio: fix "comparison of distinct pointer types lacks a cast" warning

After backporting commit 581512a6dc93 ("vsock/virtio: MSG_ZEROCOPY
flag support") in CentOS Stream 9, CI reported the following error:

    In file included from ./include/linux/kernel.h:17,
                     from ./include/linux/list.h:9,
                     from ./include/linux/preempt.h:11,
                     from ./include/linux/spinlock.h:56,
                     from net/vmw_vsock/virtio_transport_common.c:9:
    net/vmw_vsock/virtio_transport_common.c: In function ‘virtio_transport_can_zcopy‘:
    ./include/linux/minmax.h:20:35: error: comparison of distinct pointer types lacks a cast [-Werror]
       20 |         (!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
          |                                   ^~
    ./include/linux/minmax.h:26:18: note: in expansion of macro ‘__typecheck‘
       26 |                 (__typecheck(x, y) && __no_side_effects(x, y))
          |                  ^~~~~~~~~~~
    ./include/linux/minmax.h:36:31: note: in expansion of macro ‘__safe_cmp‘
       36 |         __builtin_choose_expr(__safe_cmp(x, y), \
          |                               ^~~~~~~~~~
    ./include/linux/minmax.h:45:25: note: in expansion of macro ‘__careful_cmp‘
       45 | #define min(x, y)       __careful_cmp(x, y, <)
          |                         ^~~~~~~~~~~~~
    net/vmw_vsock/virtio_transport_common.c:63:37: note: in expansion of macro ‘min‘
       63 |                 int pages_to_send = min(pages_in_iov, MAX_SKB_FRAGS);

We could solve it by using min_t(), but this operation seems entirely
unnecessary, because we also pass MAX_SKB_FRAGS to iov_iter_npages(),
which performs almost the same check, returning at most MAX_SKB_FRAGS
elements. So, let's eliminate this unnecessary comparison.

Fixes: 581512a6dc93 ("vsock/virtio: MSG_ZEROCOPY flag support")
Cc: avkrasnov@salutedevices.com
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Arseniy Krasnov <avkrasnov@salutedevices.com>
Link: https://lore.kernel.org/r/20231206164143.281107-1-sgarzare@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/smc: fix missing byte order conversion in CLC handshake

The byte order conversions of ISM GID and DMB token are missing in
process of CLC accept and confirm. So fix it.

Fixes: 3d9725a6a133 ("net/smc: common routine for CLC accept and confirm")
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Link: https://lore.kernel.org/r/1701882157-87956-1-git-send-email-guwen@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: wangxun: fix changing mac failed when running

in some bonding mode, service need to change mac when
netif is running. Wangxun netdev add IFF_LIVE_ADDR_CHANGE
priv_flag to support it.

Signed-off-by: duanqiangwen <duanqiangwen@net-swift.com>
Link: https://lore.kernel.org/r/20231206095044.17844-1-duanqiangwen@net-swift.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: microchip: provide a list of valid protocols for xmit handler

Provide a list of valid protocols for which the driver will provide
it's deferred xmit handler.

When using DSA_TAG_PROTO_KSZ8795 protocol, it does not provide a
"connect" method, therefor ksz_connect() is not allocating ksz_tagger_data.

This avoids the following null pointer dereference:
 ksz_connect_tag_protocol from dsa_register_switch+0x9ac/0xee0
 dsa_register_switch from ksz_switch_register+0x65c/0x828
 ksz_switch_register from ksz_spi_probe+0x11c/0x168
 ksz_spi_probe from spi_probe+0x84/0xa8
 spi_probe from really_probe+0xc8/0x2d8

Fixes: ab32f56a4100 ("net: dsa: microchip: ptp: add packet transmission timestamping")
Signed-off-by: Sean Nyekjaer <sean@geanix.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231206071655.1626479-1-sean@geanix.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'generic-netlink-multicast-fixes'

Ido Schimmel says:

====================
Generic netlink multicast fixes

Restrict two generic netlink multicast groups - in the "psample" and
"NET_DM" families - to be root-only with the appropriate capabilities.
See individual patches for more details.
====================

Link: https://lore.kernel.org/r/20231206213102.1824398-1-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

drop_monitor: Require 'CAP_SYS_ADMIN' when joining "events" group

The "NET_DM" generic netlink family notifies drop locations over the
"events" multicast group. This is problematic since by default generic
netlink allows non-root users to listen to these notifications.

Fix by adding a new field to the generic netlink multicast group
structure that when set prevents non-root users or root without the
'CAP_SYS_ADMIN' capability (in the user namespace owning the network
namespace) from joining the group. Set this field for the "events"
group. Use 'CAP_SYS_ADMIN' rather than 'CAP_NET_ADMIN' because of the
nature of the information that is shared over this group.

Note that the capability check in this case will always be performed
against the initial user namespace since the family is not netns aware
and only operates in the initial network namespace.

A new field is added to the structure rather than using the "flags"
field because the existing field uses uAPI flags and it is inappropriate
to add a new uAPI flag for an internal kernel check. In net-next we can
rework the "flags" field to use internal flags and fold the new field
into it. But for now, in order to reduce the amount of changes, add a
new field.

Since the information can only be consumed by root, mark the control
plane operations that start and stop the tracing as root-only using the
'GENL_ADMIN_PERM' flag.

Tested using [1].

Before:

 # capsh -- -c ./dm_repo
 # capsh --drop=cap_sys_admin -- -c ./dm_repo

After:

 # capsh -- -c ./dm_repo
 # capsh --drop=cap_sys_admin -- -c ./dm_repo
 Failed to join "events" multicast group

[1]
 $ cat dm.c
 #include <stdio.h>
 #include <netlink/genl/ctrl.h>
 #include <netlink/genl/genl.h>
 #include <netlink/socket.h>

 int main(int argc, char **argv)
 {
  struct nl_sock *sk;
  int grp, err;

  sk = nl_socket_alloc();
  if (!sk) {
  fprintf(stderr, "Failed to allocate socket\n");
  return -1;
  }

  err = genl_connect(sk);
  if (err) {
  fprintf(stderr, "Failed to connect socket\n");
  return err;
  }

  grp = genl_ctrl_resolve_grp(sk, "NET_DM", "events");
  if (grp < 0) {
  fprintf(stderr,
  "Failed to resolve \"events\" multicast group\n");
  return grp;
  }

  err = nl_socket_add_memberships(sk, grp, NFNLGRP_NONE);
  if (err) {
  fprintf(stderr, "Failed to join \"events\" multicast group\n");
  return err;
  }

  return 0;
 }
 $ gcc -I/usr/include/libnl3 -lnl-3 -lnl-genl-3 -o dm_repo dm.c

Fixes: 9a8afc8d3962 ("Network Drop Monitor: Adding drop monitor implementation & Netlink protocol")
Reported-by: "The UK's National Cyber Security Centre (NCSC)" <security@ncsc.gov.uk>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231206213102.1824398-3-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

psample: Require 'CAP_NET_ADMIN' when joining "packets" group

The "psample" generic netlink family notifies sampled packets over the
"packets" multicast group. This is problematic since by default generic
netlink allows non-root users to listen to these notifications.

Fix by marking the group with the 'GENL_UNS_ADMIN_PERM' flag. This will
prevent non-root users or root without the 'CAP_NET_ADMIN' capability
(in the user namespace owning the network namespace) from joining the
group.

Tested using [1].

Before:

 # capsh -- -c ./psample_repo
 # capsh --drop=cap_net_admin -- -c ./psample_repo

After:

 # capsh -- -c ./psample_repo
 # capsh --drop=cap_net_admin -- -c ./psample_repo
 Failed to join "packets" multicast group

[1]
 $ cat psample.c
 #include <stdio.h>
 #include <netlink/genl/ctrl.h>
 #include <netlink/genl/genl.h>
 #include <netlink/socket.h>

 int join_grp(struct nl_sock *sk, const char *grp_name)
 {
  int grp, err;

  grp = genl_ctrl_resolve_grp(sk, "psample", grp_name);
  if (grp < 0) {
  fprintf(stderr, "Failed to resolve \"%s\" multicast group\n",
  grp_name);
  return grp;
  }

  err = nl_socket_add_memberships(sk, grp, NFNLGRP_NONE);
  if (err) {
  fprintf(stderr, "Failed to join \"%s\" multicast group\n",
  grp_name);
  return err;
  }

  return 0;
 }

 int main(int argc, char **argv)
 {
  struct nl_sock *sk;
  int err;

  sk = nl_socket_alloc();
  if (!sk) {
  fprintf(stderr, "Failed to allocate socket\n");
  return -1;
  }

  err = genl_connect(sk);
  if (err) {
  fprintf(stderr, "Failed to connect socket\n");
  return err;
  }

  err = join_grp(sk, "config");
  if (err)
  return err;

  err = join_grp(sk, "packets");
  if (err)
  return err;

  return 0;
 }
 $ gcc -I/usr/include/libnl3 -lnl-3 -lnl-genl-3 -o psample_repo psample.c

Fixes: 6ae0a6286171 ("net: Introduce psample, a new genetlink channel for packet sampling")
Reported-by: "The UK's National Cyber Security Centre (NCSC)" <security@ncsc.gov.uk>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231206213102.1824398-2-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge branch 'fixes-for-ktls'

John Fastabend says:

====================
Couple fixes for TLS and BPF interactions.
====================

Link: https://lore.kernel.org/r/20231206232706.374377-1-john.fastabend@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

bpf: sockmap, updating the sg structure should also update curr

Curr pointer should be updated when the sg structure is shifted.

Fixes: 7246d8ed4dcce ("bpf: helper to pop data from messages")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20231206232706.374377-3-john.fastabend@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: tls, update curr on splice as well

The curr pointer must also be updated on the splice similar to how
we do this for other copy types.

Fixes: d829e9c4112b ("tls: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Reported-by: Jann Horn <jannh@google.com>
Link: https://lore.kernel.org/r/20231206232706.374377-2-john.fastabend@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

x86/tdx: Allow 32-bit emulation by default

32-bit emulation was disabled on TDX to prevent a possible attack by
a VMM injecting an interrupt on vector 0x80.

Now that int80_emulation() has a check for external interrupts the
limitation can be lifted.

To distinguish software interrupts from external ones, int80_emulation()
checks the APIC ISR bit relevant to the 0x80 vector. For
software interrupts, this bit will be 0.

On TDX, the VAPIC state (including ISR) is protected and cannot be
manipulated by the VMM. The ISR bit is set by the microcode flow during
the handling of posted interrupts.

[ dhansen: more changelog tweaks ]

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@vger.kernel.org> # v6.0+

x86/entry: Do not allow external 0x80 interrupts

The INT 0x80 instruction is used for 32-bit x86 Linux syscalls. The
kernel expects to receive a software interrupt as a result of the INT
0x80 instruction. However, an external interrupt on the same vector
also triggers the same codepath.

An external interrupt on vector 0x80 will currently be interpreted as a
32-bit system call, and assuming that it was a user context.

Panic on external interrupts on the vector.

To distinguish software interrupts from external ones, the kernel checks
the APIC ISR bit relevant to the 0x80 vector. For software interrupts,
this bit will be 0.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@vger.kernel.org> # v6.0+

x86/entry: Convert INT 0x80 emulation to IDTENTRY

There is no real reason to have a separate ASM entry point implementation
for the legacy INT 0x80 syscall emulation on 64-bit.

IDTENTRY provides all the functionality needed with the only difference
that it does not:

  - save the syscall number (AX) into pt_regs::orig_ax
  - set pt_regs::ax to -ENOSYS

Both can be done safely in the C code of an IDTENTRY before invoking any of
the syscall related functions which depend on this convention.

Aside of ASM code reduction this prepares for detecting and handling a
local APIC injected vector 0x80.

[ kirill.shutemov: More verbose comments ]
Suggested-by: Linus Torvalds <torvalds@linuxfoundation.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@vger.kernel.org> # v6.0+

x86/coco: Disable 32-bit emulation by default on TDX and SEV

The INT 0x80 instruction is used for 32-bit x86 Linux syscalls. The
kernel expects to receive a software interrupt as a result of the INT
0x80 instruction. However, an external interrupt on the same vector
triggers the same handler.

The kernel interprets an external interrupt on vector 0x80 as a 32-bit
system call that came from userspace.

A VMM can inject external interrupts on any arbitrary vector at any
time.  This remains true even for TDX and SEV guests where the VMM is
untrusted.

Put together, this allows an untrusted VMM to trigger int80 syscall
handling at any given point. The content of the guest register file at
that moment defines what syscall is triggered and its arguments. It
opens the guest OS to manipulation from the VMM side.

Disable 32-bit emulation by default for TDX and SEV. User can override
it with the ia32_emulation=y command line option.

[ dhansen: reword the changelog ]

Reported-by: Supraja Sridhara <supraja.sridhara@inf.ethz.ch>
Reported-by: Benedict Schlüter <benedict.schlueter@inf.ethz.ch>
Reported-by: Mark Kuhne <mark.kuhne@inf.ethz.ch>
Reported-by: Andrin Bertschi <andrin.bertschi@inf.ethz.ch>
Reported-by: Shweta Shinde <shweta.shinde@inf.ethz.ch>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@vger.kernel.org> # v6.0+: 1da5c9b x86: Introduce ia32_enabled()
Cc: <stable@vger.kernel.org> # v6.0+

Merge tag 'nf-23-12-06' of git://git./linux/kernel/git/netfilter/nf

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

1) Incorrect nf_defrag registration for bpf link infra, from D. Wythe.

2) Skip inactive elements in pipapo set backend walk to avoid double
   deactivation, from Florian Westphal.

3) Fix NFT_*_F_PRESENT check with big endian arch, also from Florian.

4) Bail out if number of expressions in NFTA_DYNSET_EXPRESSIONS mismatch
   stateful expressions in set declaration.

5) Honor family in table lookup by handle. Broken since 4.16.

6) Use sk_callback_lock to protect access to sk->sk_socket in xt_owner.
   sock_orphan() might zap this pointer, from Phil Sutter.

All of these fixes address broken stuff for several releases.

* tag 'nf-23-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: xt_owner: Fix for unsafe access of sk->sk_socket
  netfilter: nf_tables: validate family when identifying table via handle
  netfilter: nf_tables: bail out on mismatching dynset and set expressions
  netfilter: nf_tables: fix 'exist' matching on bigendian arches
  netfilter: nft_set_pipapo: skip inactive elements during set walk
  netfilter: bpf: fix bad registration on nf_defrag
====================

Link: https://lore.kernel.org/r/20231206180357.959930-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2023-12-06

We've added 4 non-merge commits during the last 6 day(s) which contain
a total of 7 files changed, 185 insertions(+), 55 deletions(-).

The main changes are:

1) Fix race found by syzkaller on prog_array_map_poke_run when
   a BPF program's kallsym symbols were still missing, from Jiri Olsa.

2) Fix BPF verifier's branch offset comparison for BPF_JMP32 | BPF_JA,
   from Yonghong Song.

3) Fix xsk's poll handling to only set mask on bound xsk sockets,
   from Yewon Choi.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  selftests/bpf: Add test for early update in prog_array_map_poke_run
  bpf: Fix prog_array_map_poke_run map poke update
  xsk: Skip polling event check for unbound socket
  bpf: Fix a verifier bug due to incorrect branch offset comparison with cpu=v4
====================

Link: https://lore.kernel.org/r/20231206220528.12093-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: microchip: move ksz_chip_id enum to platform include

With the ksz_chip_id enums moved to the platform include file for ksz
switches, platform code that instantiates a device can now use these to
set ksz_platform_data::chip_id.

Signed-off-by: Daniel Danzberger <dd@embedd.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: dsa: microchip: properly support platform_data probing

The ksz driver has bits and pieces of platform_data probing support, but
it doesn't work.

The conventional thing to do is to have an encapsulating structure for
struct dsa_chip_data that gets put into dev->platform_data. This driver
expects a struct ksz_platform_data, but that doesn't contain a struct
dsa_chip_data as first element, which will obviously not work with
dsa_switch_probe() -> dsa_switch_parse().

Pointing dev->platform_data to a struct dsa_chip_data directly is in
principle possible, but that doesn't work either. The driver has
ksz_switch_detect() to read the device ID from hardware, followed by
ksz_check_device_id() to compare it against a predetermined expected
value. This protects against early errors in the SPI/I2C communication.
With platform_data, the mechanism in ksz_check_device_id() doesn't work
and even leads to NULL pointer dereferences, since of_device_get_match_data()
doesn't work in that probe path.

So obviously, the platform_data support is actually missing, and the
existing handling of struct ksz_platform_data is bogus. Complete the
support by adding a struct dsa_chip_data as first element, and fixing up
ksz_check_device_id() to pick up the platform_data instead of the
unavailable of_device_get_match_data().

The early dev->chip_id assignment from ksz_switch_register() is also
bogus, because ksz_switch_detect() sets it to an initial value. So
remove it.

Also, ksz_platform_data :: enabled_ports isn't used anywhere, delete it.

Link: https://lore.kernel.org/netdev/20231204154315.3906267-1-dd@embedd.com/
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Daniel Danzberger <dd@embedd.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'dsa-microchip-rmii-reference'

Ante Knezic says:

====================
net: dsa: microchip: enable setting rmii reference

KSZ88X3 devices can select between internal and external RMII reference clock.
This patch series introduces new device tree property for setting reference
clock to internal.

---
V5:
  - move rmii-clk-internal to be a port device tree property.
V4:
  - remove rmii_clk_internal from ksz_device, as its not needed any more
  - move rmii clk config as well as ksz8795_cpu_interface_select to
    ksz8_config_cpu_port
V3:
  - move ksz_cfg from global switch config to port config as suggested by Vladimir
    Oltean
  - reverse patch order as suggested by Vladimir Oltean
  - adapt dt schema as suggested by Conor Dooley
V2:
  - don't rely on default register settings - enforce set/clear property as
    suggested by Andrew Lunn
  - enforce dt schema as suggested by Conor Dooley
====================

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

net: dsa: microchip: add property to select internal RMII reference clock

Microchip KSZ8863/KSZ8873 have the ability to select between internal
and external RMII reference clock. By default, reference clock
needs to be provided via REFCLKI_3 pin. If required, device can be
setup to provide RMII clock internally so that REFCLKI_3 pin can be
left unconnected.
Add a new "microchip,rmii-clk-internal" property which will set
RMII clock reference to internal. If property is not set, reference
clock needs to be provided externally.

While at it, move the ksz8795_cpu_interface_select() to
ksz8_config_cpu_port() to get a cleaner call path for cpu port.

Signed-off-by: Ante Knezic <ante.knezic@helmholz.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

dt-bindings: net: microchip,ksz: document microchip,rmii-clk-internal

Add documentation for selecting reference rmii clock on KSZ88X3 devices

Signed-off-by: Ante Knezic <ante.knezic@helmholz.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower: fix for take a mutex lock in soft irq context and rcu lock

The neighbour event callback call the function nfp_tun_write_neigh,
this function will take a mutex lock and it is in soft irq context,
change the work queue to process the neighbour event.

Move the nfp_tun_write_neigh function out of range rcu_read_lock/unlock()
in function nfp_tunnel_request_route_v4 and nfp_tunnel_request_route_v6.

Fixes: abc210952af7 ("nfp: flower: tunnel neigh support bond offload")
CC: stable@vger.kernel.org # 6.2+
Signed-off-by: Hui Zhou <hui.zhou@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge tag 'parisc-for-6.7-rc5' of git://git./linux/kernel/git/deller/parisc-linux

Pull parisc fix from Helge Deller:
 "A single line patch for parisc which fixes the build in tinyconfig
  configurations:

   - Fix asm operand number out of range build error in bug table"

* tag 'parisc-for-6.7-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Fix asm operand number out of range build error in bug table

Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-12-05 (ice, i40e, iavf)

This series contains updates to ice, i40e and iavf drivers.

Michal fixes incorrect usage of VF MSIX value and index calculation for
ice.

Marcin restores disabling of Rx VLAN filtering which was inadvertently
removed for ice.

Ivan Vecera corrects improper messaging of MFS port for i40e.

Jake fixes incorrect checking of coalesce values on iavf.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  iavf: validate tx_coalesce_usecs even if rx_coalesce_usecs is zero
  i40e: Fix unexpected MFS warning message
  ice: Restore fix disabling RX VLAN filtering
  ice: change vfs.num_msix_per to vf->num_msix
====================

Link: https://lore.kernel.org/r/20231205211918.2123019-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: rtnetlink: remove local list in __linkwatch_run_queue()

Due to linkwatch_forget_dev() (and perhaps others?) checking for
list_empty(&dev->link_watch_list), we must have all manipulations
of even the local on-stack list 'wrk' here under spinlock, since
even that list can be reached otherwise via dev->link_watch_list.

This is already the case, but makes this a bit counter-intuitive,
often local lists are used to _not_ have to use locking for their
local use.

Remove the local list as it doesn't seem to serve any purpose.
While at it, move a variable declaration into the loop using it.

Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20231205170011.56576dcc1727.I698b72219d9f6ce789bd209b8f6dffd0ca32a8f2@changeid
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: dsa: mv88e6xxx: Restore USXGMII support for 6393X

In 4a56212774ac, USXGMII support was added for 6393X, but this was
lost in the PCS conversion (the blamed commit), most likely because
these efforts where more or less done in parallel.

Restore this feature by porting Michal's patch to fit the new
implementation.

Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Tested-by: Michal Smulski <michal.smulski@ooma.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Fixes: e5b732a275f5 ("net: dsa: mv88e6xxx: convert 88e639x to phylink_pcs")
Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
Link: https://lore.kernel.org/r/20231205221359.3926018-1-tobias@waldekranz.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tcp: do not accept ACK of bytes we never sent

This patch is based on a detailed report and ideas from Yepeng Pan
and Christian Rossow.

ACK seq validation is currently following RFC 5961 5.2 guidelines:

   The ACK value is considered acceptable only if
   it is in the range of ((SND.UNA - MAX.SND.WND) <= SEG.ACK <=
   SND.NXT).  All incoming segments whose ACK value doesn't satisfy the
   above condition MUST be discarded and an ACK sent back.  It needs to
   be noted that RFC 793 on page 72 (fifth check) says: "If the ACK is a
   duplicate (SEG.ACK < SND.UNA), it can be ignored.  If the ACK
   acknowledges something not yet sent (SEG.ACK > SND.NXT) then send an
   ACK, drop the segment, and return".  The "ignored" above implies that
   the processing of the incoming data segment continues, which means
   the ACK value is treated as acceptable.  This mitigation makes the
   ACK check more stringent since any ACK < SND.UNA wouldn't be
   accepted, instead only ACKs that are in the range ((SND.UNA -
   MAX.SND.WND) <= SEG.ACK <= SND.NXT) get through.

This can be refined for new (and possibly spoofed) flows,
by not accepting ACK for bytes that were never sent.

This greatly improves TCP security at a little cost.

I added a Fixes: tag to make sure this patch will reach stable trees,
even if the 'blamed' patch was adhering to the RFC.

tp->bytes_acked was added in linux-4.2

Following packetdrill test (courtesy of Yepeng Pan) shows
the issue at hand:

0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+0 bind(3, ..., ...) = 0
+0 listen(3, 1024) = 0

// ---------------- Handshake ------------------- //

// when window scale is set to 14 the window size can be extended to
// 65535 * (2^14) = 1073725440. Linux would accept an ACK packet
// with ack number in (Server_ISN+1-1073725440. Server_ISN+1)
// ,though this ack number acknowledges some data never
// sent by the server.

+0 < S 0:0(0) win 65535 <mss 1400,nop,wscale 14>
+0 > S. 0:0(0) ack 1 <...>
+0 < . 1:1(0) ack 1 win 65535
+0 accept(3, ..., ...) = 4

// For the established connection, we send an ACK packet,
// the ack packet uses ack number 1 - 1073725300 + 2^32,
// where 2^32 is used to wrap around.
// Note: we used 1073725300 instead of 1073725440 to avoid possible
// edge cases.
// 1 - 1073725300 + 2^32 = 3221241997

// Oops, old kernels happily accept this packet.
+0 < . 1:1001(1000) ack 3221241997 win 65535

// After the kernel fix the following will be replaced by a challenge ACK,
// and prior malicious frame would be dropped.
+0 > . 1:1(0) ack 1001

Fixes: 354e4aa391ed ("tcp: RFC 5961 5.2 Blind Data Injection Attack Mitigation")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Yepeng Pan <yepeng.pan@cispa.de>
Reported-by: Christian Rossow <rossow@cispa.de>
Acked-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20231205161841.2702925-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

ipv6: add debug checks in fib6_info_release()

Some elusive syzbot reports are hinting to fib6_info_release(),
with a potential dangling f6i->gc_link anchor.

Add debug checks so that syzbot can catch the issue earlier eventually.

BUG: KASAN: slab-use-after-free in __hlist_del include/linux/list.h:990 [inline]
BUG: KASAN: slab-use-after-free in hlist_del_init include/linux/list.h:1016 [inline]
BUG: KASAN: slab-use-after-free in fib6_clean_expires_locked include/net/ip6_fib.h:533 [inline]
BUG: KASAN: slab-use-after-free in fib6_purge_rt+0x986/0x9c0 net/ipv6/ip6_fib.c:1064
Write of size 8 at addr ffff88802805a840 by task syz-executor.1/10057

CPU: 1 PID: 10057 Comm: syz-executor.1 Not tainted 6.7.0-rc2-syzkaller-00029-g9b6de136b5f0 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/10/2023
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0xd9/0x1b0 lib/dump_stack.c:106
print_address_description mm/kasan/report.c:364 [inline]
print_report+0xc4/0x620 mm/kasan/report.c:475
kasan_report+0xda/0x110 mm/kasan/report.c:588
__hlist_del include/linux/list.h:990 [inline]
hlist_del_init include/linux/list.h:1016 [inline]
fib6_clean_expires_locked include/net/ip6_fib.h:533 [inline]
fib6_purge_rt+0x986/0x9c0 net/ipv6/ip6_fib.c:1064
fib6_del_route net/ipv6/ip6_fib.c:1993 [inline]
fib6_del+0xa7a/0x1750 net/ipv6/ip6_fib.c:2038
__ip6_del_rt net/ipv6/route.c:3866 [inline]
ip6_del_rt+0xf7/0x200 net/ipv6/route.c:3881
ndisc_router_discovery+0x295b/0x3560 net/ipv6/ndisc.c:1372
ndisc_rcv+0x3de/0x5f0 net/ipv6/ndisc.c:1856
icmpv6_rcv+0x1470/0x19c0 net/ipv6/icmp.c:979
ip6_protocol_deliver_rcu+0x170/0x13e0 net/ipv6/ip6_input.c:438
ip6_input_finish+0x14f/0x2f0 net/ipv6/ip6_input.c:483
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ip6_input+0xa1/0xc0 net/ipv6/ip6_input.c:492
ip6_mc_input+0x48b/0xf40 net/ipv6/ip6_input.c:586
dst_input include/net/dst.h:461 [inline]
ip6_rcv_finish net/ipv6/ip6_input.c:79 [inline]
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ipv6_rcv+0x24e/0x380 net/ipv6/ip6_input.c:310
__netif_receive_skb_one_core+0x115/0x180 net/core/dev.c:5529
__netif_receive_skb+0x1f/0x1b0 net/core/dev.c:5643
netif_receive_skb_internal net/core/dev.c:5729 [inline]
netif_receive_skb+0x133/0x700 net/core/dev.c:5788
tun_rx_batched+0x429/0x780 drivers/net/tun.c:1579
tun_get_user+0x29e3/0x3bc0 drivers/net/tun.c:2002
tun_chr_write_iter+0xe8/0x210 drivers/net/tun.c:2048
call_write_iter include/linux/fs.h:2020 [inline]
new_sync_write fs/read_write.c:491 [inline]
vfs_write+0x64f/0xdf0 fs/read_write.c:584
ksys_write+0x12f/0x250 fs/read_write.c:637
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x40/0x110 arch/x86/entry/common.c:82
entry_SYSCALL_64_after_hwframe+0x63/0x6b
RIP: 0033:0x7f38e387b82f
Code: 89 54 24 18 48 89 74 24 10 89 7c 24 08 e8 b9 80 02 00 48 8b 54 24 18 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 31 44 89 c7 48 89 44 24 08 e8 0c 81 02 00 48
RSP: 002b:00007f38e45c9090 EFLAGS: 00000293 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 00007f38e399bf80 RCX: 00007f38e387b82f
RDX: 00000000000003b6 RSI: 0000000020000680 RDI: 00000000000000c8
RBP: 00007f38e38c847a R08: 0000000000000000 R09: 0000000000000000
R10: 00000000000003b6 R11: 0000000000000293 R12: 0000000000000000
R13: 000000000000000b R14: 00007f38e399bf80 R15: 00007f38e3abfa48
</TASK>

Allocated by task 10044:
kasan_save_stack+0x33/0x50 mm/kasan/common.c:45
kasan_set_track+0x25/0x30 mm/kasan/common.c:52
____kasan_kmalloc mm/kasan/common.c:374 [inline]
__kasan_kmalloc+0xa2/0xb0 mm/kasan/common.c:383
kasan_kmalloc include/linux/kasan.h:198 [inline]
__do_kmalloc_node mm/slab_common.c:1007 [inline]
__kmalloc+0x59/0x90 mm/slab_common.c:1020
kmalloc include/linux/slab.h:604 [inline]
kzalloc include/linux/slab.h:721 [inline]
fib6_info_alloc+0x40/0x160 net/ipv6/ip6_fib.c:155
ip6_route_info_create+0x337/0x1e70 net/ipv6/route.c:3749
ip6_route_add+0x26/0x150 net/ipv6/route.c:3843
rt6_add_route_info+0x2e7/0x4b0 net/ipv6/route.c:4316
rt6_route_rcv+0x76c/0xbf0 net/ipv6/route.c:985
ndisc_router_discovery+0x138b/0x3560 net/ipv6/ndisc.c:1529
ndisc_rcv+0x3de/0x5f0 net/ipv6/ndisc.c:1856
icmpv6_rcv+0x1470/0x19c0 net/ipv6/icmp.c:979
ip6_protocol_deliver_rcu+0x170/0x13e0 net/ipv6/ip6_input.c:438
ip6_input_finish+0x14f/0x2f0 net/ipv6/ip6_input.c:483
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ip6_input+0xa1/0xc0 net/ipv6/ip6_input.c:492
ip6_mc_input+0x48b/0xf40 net/ipv6/ip6_input.c:586
dst_input include/net/dst.h:461 [inline]
ip6_rcv_finish net/ipv6/ip6_input.c:79 [inline]
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ipv6_rcv+0x24e/0x380 net/ipv6/ip6_input.c:310
__netif_receive_skb_one_core+0x115/0x180 net/core/dev.c:5529
__netif_receive_skb+0x1f/0x1b0 net/core/dev.c:5643
netif_receive_skb_internal net/core/dev.c:5729 [inline]
netif_receive_skb+0x133/0x700 net/core/dev.c:5788
tun_rx_batched+0x429/0x780 drivers/net/tun.c:1579
tun_get_user+0x29e3/0x3bc0 drivers/net/tun.c:2002
tun_chr_write_iter+0xe8/0x210 drivers/net/tun.c:2048
call_write_iter include/linux/fs.h:2020 [inline]
new_sync_write fs/read_write.c:491 [inline]
vfs_write+0x64f/0xdf0 fs/read_write.c:584
ksys_write+0x12f/0x250 fs/read_write.c:637
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x40/0x110 arch/x86/entry/common.c:82
entry_SYSCALL_64_after_hwframe+0x63/0x6b

Freed by task 5123:
kasan_save_stack+0x33/0x50 mm/kasan/common.c:45
kasan_set_track+0x25/0x30 mm/kasan/common.c:52
kasan_save_free_info+0x2b/0x40 mm/kasan/generic.c:522
____kasan_slab_free mm/kasan/common.c:236 [inline]
____kasan_slab_free+0x15b/0x1b0 mm/kasan/common.c:200
kasan_slab_free include/linux/kasan.h:164 [inline]
slab_free_hook mm/slub.c:1800 [inline]
slab_free_freelist_hook+0x114/0x1e0 mm/slub.c:1826
slab_free mm/slub.c:3809 [inline]
__kmem_cache_free+0xc0/0x180 mm/slub.c:3822
rcu_do_batch kernel/rcu/tree.c:2158 [inline]
rcu_core+0x819/0x1680 kernel/rcu/tree.c:2431
__do_softirq+0x21a/0x8de kernel/softirq.c:553

Last potentially related work creation:
kasan_save_stack+0x33/0x50 mm/kasan/common.c:45
__kasan_record_aux_stack+0xbc/0xd0 mm/kasan/generic.c:492
__call_rcu_common.constprop.0+0x9a/0x7a0 kernel/rcu/tree.c:2681
fib6_info_release include/net/ip6_fib.h:332 [inline]
fib6_info_release include/net/ip6_fib.h:329 [inline]
rt6_route_rcv+0xa4e/0xbf0 net/ipv6/route.c:997
ndisc_router_discovery+0x138b/0x3560 net/ipv6/ndisc.c:1529
ndisc_rcv+0x3de/0x5f0 net/ipv6/ndisc.c:1856
icmpv6_rcv+0x1470/0x19c0 net/ipv6/icmp.c:979
ip6_protocol_deliver_rcu+0x170/0x13e0 net/ipv6/ip6_input.c:438
ip6_input_finish+0x14f/0x2f0 net/ipv6/ip6_input.c:483
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ip6_input+0xa1/0xc0 net/ipv6/ip6_input.c:492
ip6_mc_input+0x48b/0xf40 net/ipv6/ip6_input.c:586
dst_input include/net/dst.h:461 [inline]
ip6_rcv_finish net/ipv6/ip6_input.c:79 [inline]
NF_HOOK include/linux/netfilter.h:314 [inline]
NF_HOOK include/linux/netfilter.h:308 [inline]
ipv6_rcv+0x24e/0x380 net/ipv6/ip6_input.c:310
__netif_receive_skb_one_core+0x115/0x180 net/core/dev.c:5529
__netif_receive_skb+0x1f/0x1b0 net/core/dev.c:5643
netif_receive_skb_internal net/core/dev.c:5729 [inline]
netif_receive_skb+0x133/0x700 net/core/dev.c:5788
tun_rx_batched+0x429/0x780 drivers/net/tun.c:1579
tun_get_user+0x29e3/0x3bc0 drivers/net/tun.c:2002
tun_chr_write_iter+0xe8/0x210 drivers/net/tun.c:2048
call_write_iter include/linux/fs.h:2020 [inline]
new_sync_write fs/read_write.c:491 [inline]
vfs_write+0x64f/0xdf0 fs/read_write.c:584
ksys_write+0x12f/0x250 fs/read_write.c:637
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x40/0x110 arch/x86/entry/common.c:82
entry_SYSCALL_64_after_hwframe+0x63/0x6b

Second to last potentially related work creation:
kasan_save_stack+0x33/0x50 mm/kasan/common.c:45
__kasan_record_aux_stack+0xbc/0xd0 mm/kasan/generic.c:492
insert_work+0x38/0x230 kernel/workqueue.c:1647
__queue_work+0xcdc/0x11f0 kernel/workqueue.c:1803
call_timer_fn+0x193/0x590 kernel/time/timer.c:1700
expire_timers kernel/time/timer.c:1746 [inline]
__run_timers+0x585/0xb20 kernel/time/timer.c:2022
run_timer_softirq+0x58/0xd0 kernel/time/timer.c:2035
__do_softirq+0x21a/0x8de kernel/softirq.c:553

The buggy address belongs to the object at ffff88802805a800
which belongs to the cache kmalloc-512 of size 512
The buggy address is located 64 bytes inside of
freed 512-byte region [ffff88802805a800, ffff88802805aa00)

The buggy address belongs to the physical page:
page:ffffea0000a01600 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x28058
head:ffffea0000a01600 order:2 entire_mapcount:0 nr_pages_mapped:0 pincount:0
flags: 0xfff00000000840(slab|head|node=0|zone=1|lastcpupid=0x7ff)
page_type: 0xffffffff()
raw: 00fff00000000840 ffff888013041c80 ffffea0001e02600 dead000000000002
raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
page_owner tracks the page as allocated
page last allocated via order 2, migratetype Unmovable, gfp_mask 0x1d20c0(__GFP_IO|__GFP_FS|__GFP_NOWARN|__GFP_NORETRY|__GFP_COMP|__GFP_NOMEMALLOC|__GFP_HARDWALL), pid 18706, tgid 18699 (syz-executor.2), ts 999991973280, free_ts 996884464281
set_page_owner include/linux/page_owner.h:31 [inline]
post_alloc_hook+0x2d0/0x350 mm/page_alloc.c:1537
prep_new_page mm/page_alloc.c:1544 [inline]
get_page_from_freelist+0xa25/0x36d0 mm/page_alloc.c:3312
__alloc_pages+0x22e/0x2420 mm/page_alloc.c:4568
alloc_pages_mpol+0x258/0x5f0 mm/mempolicy.c:2133
alloc_slab_page mm/slub.c:1870 [inline]
allocate_slab mm/slub.c:2017 [inline]
new_slab+0x283/0x3c0 mm/slub.c:2070
___slab_alloc+0x979/0x1500 mm/slub.c:3223
__slab_alloc.constprop.0+0x56/0xa0 mm/slub.c:3322
__slab_alloc_node mm/slub.c:3375 [inline]
slab_alloc_node mm/slub.c:3468 [inline]
__kmem_cache_alloc_node+0x131/0x310 mm/slub.c:3517
__do_kmalloc_node mm/slab_common.c:1006 [inline]
__kmalloc+0x49/0x90 mm/slab_common.c:1020
kmalloc include/linux/slab.h:604 [inline]
kzalloc include/linux/slab.h:721 [inline]
copy_splice_read+0x1ac/0x8f0 fs/splice.c:338
vfs_splice_read fs/splice.c:992 [inline]
vfs_splice_read+0x2ea/0x3b0 fs/splice.c:962
splice_direct_to_actor+0x2a5/0xa30 fs/splice.c:1069
do_splice_direct+0x1af/0x280 fs/splice.c:1194
do_sendfile+0xb3e/0x1310 fs/read_write.c:1254
__do_sys_sendfile64 fs/read_write.c:1322 [inline]
__se_sys_sendfile64 fs/read_write.c:1308 [inline]
__x64_sys_sendfile64+0x1d6/0x220 fs/read_write.c:1308
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x40/0x110 arch/x86/entry/common.c:82
page last free stack trace:
reset_page_owner include/linux/page_owner.h:24 [inline]
free_pages_prepare mm/page_alloc.c:1137 [inline]
free_unref_page_prepare+0x4fa/0xaa0 mm/page_alloc.c:2347
free_unref_page_list+0xe6/0xb40 mm/page_alloc.c:2533
release_pages+0x32a/0x14f0 mm/swap.c:1042
tlb_batch_pages_flush+0x9a/0x190 mm/mmu_gather.c:98
tlb_flush_mmu_free mm/mmu_gather.c:293 [inline]
tlb_flush_mmu mm/mmu_gather.c:300 [inline]
tlb_finish_mmu+0x14b/0x6f0 mm/mmu_gather.c:392
exit_mmap+0x38b/0xa70 mm/mmap.c:3321
__mmput+0x12a/0x4d0 kernel/fork.c:1349
mmput+0x62/0x70 kernel/fork.c:1371
exit_mm kernel/exit.c:567 [inline]
do_exit+0x9ad/0x2ae0 kernel/exit.c:858
do_group_exit+0xd4/0x2a0 kernel/exit.c:1021
get_signal+0x23be/0x2790 kernel/signal.c:2904
arch_do_signal_or_restart+0x90/0x7f0 arch/x86/kernel/signal.c:309
exit_to_user_mode_loop kernel/entry/common.c:168 [inline]
exit_to_user_mode_prepare+0x121/0x240 kernel/entry/common.c:204
irqentry_exit_to_user_mode+0xa/0x40 kernel/entry/common.c:309
asm_sysvec_apic_timer_interrupt+0x1a/0x20 arch/x86/include/asm/idtentry.h:645

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20231205173250.2982846-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: display: adi,adv75xx: Document #sound-dai-cells

When using audio from ADV7533 or ADV7535 and describing the audio
card via simple-audio-card, the '#sound-dai-cells' needs to be passed.

Document the '#sound-dai-cells' property to fix the following
dt-schema warning:

imx8mn-beacon-kit.dtb: hdmi@3d: '#sound-dai-cells' does not match any of the regexes: 'pinctrl-[0-9]+'
from schema $id: http://devicetree.org/schemas/display/bridge/adi,adv7533.yaml#

Signed-off-by: Fabio Estevam <festevam@denx.de>
Acked-by: Adam Ford <aford173@gmail.com>
Link: https://lore.kernel.org/r/20231206093643.2198562-1-festevam@gmail.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: lcdif: Properly describe the i.MX23 interrupts

i.MX23 has two LCDIF interrupts instead of a single one like other
i.MX devices.

Take this into account for properly describing the i.MX23 LCDIF
interrupts.

This fixes the following dt-schema warning:

imx23-olinuxino.dtb: lcdif@80030000: interrupts: [[46], [45]] is too long
        from schema $id: http://devicetree.org/schemas/display/fsl,lcdif.yaml#

Signed-off-by: Fabio Estevam <festevam@denx.de>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20231206112337.2234849-1-festevam@gmail.com
Signed-off-by: Rob Herring <robh@kernel.org>

selftests/bpf: Add test for early update in prog_array_map_poke_run

Adding test that tries to trigger the BUG_IN during early map update
in prog_array_map_poke_run function.

The idea is to share prog array map between thread that constantly
updates it and another one loading a program that uses that prog
array.

Eventually we will hit a place where the program is ok to be updated
(poke->tailcall_target_stable check) but the address is still not
registered in kallsyms, so the bpf_arch_text_poke returns -EINVAL
and cause imbalance for the next tail call update check, which will
fail with -EBUSY in bpf_arch_text_poke as described in previous fix.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/20231206083041.1306660-3-jolsa@kernel.org

bpf: Fix prog_array_map_poke_run map poke update

Lee pointed out issue found by syscaller [0] hitting BUG in prog array
map poke update in prog_array_map_poke_run function due to error value
returned from bpf_arch_text_poke function.

There's race window where bpf_arch_text_poke can fail due to missing
bpf program kallsym symbols, which is accounted for with check for
-EINVAL in that BUG_ON call.

The problem is that in such case we won't update the tail call jump
and cause imbalance for the next tail call update check which will
fail with -EBUSY in bpf_arch_text_poke.

I'm hitting following race during the program load:

  CPU 0                             CPU 1

  bpf_prog_load
    bpf_check
      do_misc_fixups
        prog_array_map_poke_track

                                    map_update_elem
                                      bpf_fd_array_map_update_elem
                                        prog_array_map_poke_run

                                          bpf_arch_text_poke returns -EINVAL

    bpf_prog_kallsyms_add

After bpf_arch_text_poke (CPU 1) fails to update the tail call jump, the next
poke update fails on expected jump instruction check in bpf_arch_text_poke
with -EBUSY and triggers the BUG_ON in prog_array_map_poke_run.

Similar race exists on the program unload.

Fixing this by moving the update to bpf_arch_poke_desc_update function which
makes sure we call __bpf_arch_text_poke that skips the bpf address check.

Each architecture has slightly different approach wrt looking up bpf address
in bpf_arch_text_poke, so instead of splitting the function or adding new
'checkip' argument in previous version, it seems best to move the whole
map_poke_run update as arch specific code.

  [0] https://syzkaller.appspot.com/bug?extid=97a4fe20470e9bc30810

Fixes: ebf7d1f508a7 ("bpf, x64: rework pro/epilogue and tailcall handling in JIT")
Reported-by: syzbot+97a4fe20470e9bc30810@syzkaller.appspotmail.com
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Cc: Lee Jones <lee@kernel.org>
Cc: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20231206083041.1306660-2-jolsa@kernel.org

netfilter: xt_owner: Fix for unsafe access of sk->sk_socket

A concurrently running sock_orphan() may NULL the sk_socket pointer in
between check and deref. Follow other users (like nft_meta.c for
instance) and acquire sk_callback_lock before dereferencing sk_socket.

Fixes: 0265ab44bacc ("[NETFILTER]: merge ipt_owner/ip6t_owner in xt_owner")
Reported-by: Jann Horn <jannh@google.com>
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

netfilter: nf_tables: validate family when identifying table via handle

Validate table family when looking up for it via NFTA_TABLE_HANDLE.

Fixes: 3ecbfd65f50e ("netfilter: nf_tables: allocate handle and delete objects via handle")
Reported-by: Xingyuan Mo <hdthky0@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

netfilter: nf_tables: bail out on mismatching dynset and set expressions

If dynset expressions provided by userspace is larger than the declared
set expressions, then bail out.

Fixes: 48b0ae046ee9 ("netfilter: nftables: netlink support for several set element expressions")
Reported-by: Xingyuan Mo <hdthky0@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

netfilter: nf_tables: fix 'exist' matching on bigendian arches

Maze reports "tcp option fastopen exists" fails to match on
OpenWrt 22.03.5, r20134-5f15225c1e (5.10.176) router.

"tcp option fastopen exists" translates to:
inet
  [ exthdr load tcpopt 1b @ 34 + 0 present => reg 1 ]
  [ cmp eq reg 1 0x00000001 ]

.. but existing nft userspace generates a 1-byte compare.

On LSB (x86), "*reg32 = 1" is identical to nft_reg_store8(reg32, 1), but
not on MSB, which will place the 1 last. IOW, on bigendian aches the cmp8
is awalys false.

Make sure we store this in a consistent fashion, so existing userspace
will also work on MSB (bigendian).

Regardless of this patch we can also change nft userspace to generate
'reg32 == 0' and 'reg32 != 0' instead of u8 == 0 // u8 == 1 when
adding 'option x missing/exists' expressions as well.

Fixes: 3c1fece8819e ("netfilter: nft_exthdr: Allow checking TCP option presence, too")
Fixes: b9f9a485fb0e ("netfilter: nft_exthdr: add boolean DCCP option matching")
Fixes: 055c4b34b94f ("netfilter: nft_fib: Support existence check")
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Closes: https://lore.kernel.org/netfilter-devel/CAHo-OozyEqHUjL2-ntATzeZOiuftLWZ_HU6TOM_js4qLfDEAJg@mail.gmail.com/
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

netfilter: nft_set_pipapo: skip inactive elements during set walk

Otherwise set elements can be deactivated twice which will cause a crash.

Reported-by: Xingyuan Mo <hdthky0@gmail.com>
Fixes: 3c4287f62044 ("nf_tables: Add set type for arbitrary concatenation of ranges")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

netfilter: bpf: fix bad registration on nf_defrag

We should pass a pointer to global_hook to the get_proto_defrag_hook()
instead of its value, since the passed value won't be updated even if
the request module was loaded successfully.

Log:

[   54.915713] nf_defrag_ipv4 has bad registration
[   54.915779] WARNING: CPU: 3 PID: 6323 at net/netfilter/nf_bpf_link.c:62 get_proto_defrag_hook+0x137/0x160
[   54.915835] CPU: 3 PID: 6323 Comm: fentry Kdump: loaded Tainted: G            E      6.7.0-rc2+ #35
[   54.915839] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014
[   54.915841] RIP: 0010:get_proto_defrag_hook+0x137/0x160
[   54.915844] Code: 4f 8c e8 2c cf 68 ff 80 3d db 83 9a 01 00 0f 85 74 ff ff ff 48 89 ee 48 c7 c7 8f 12 4f 8c c6 05 c4 83 9a 01 01 e8 09 ee 5f ff <0f> 0b e9 57 ff ff ff 49 8b 3c 24 4c 63 e5 e8 36 28 6c ff 4c 89 e0
[   54.915849] RSP: 0018:ffffb676003fbdb0 EFLAGS: 00010286
[   54.915852] RAX: 0000000000000023 RBX: ffff9596503d5600 RCX: ffff95996fce08c8
[   54.915854] RDX: 00000000ffffffd8 RSI: 0000000000000027 RDI: ffff95996fce08c0
[   54.915855] RBP: ffffffff8c4f12de R08: 0000000000000000 R09: 00000000fffeffff
[   54.915859] R10: ffffb676003fbc70 R11: ffffffff8d363ae8 R12: 0000000000000000
[   54.915861] R13: ffffffff8e1f75c0 R14: ffffb676003c9000 R15: 00007ffd15e78ef0
[   54.915864] FS:  00007fb6e9cab740(0000) GS:ffff95996fcc0000(0000) knlGS:0000000000000000
[   54.915867] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   54.915868] CR2: 00007ffd15e75c40 CR3: 0000000101e62006 CR4: 0000000000360ef0
[   54.915870] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   54.915871] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   54.915873] Call Trace:
[   54.915891]  <TASK>
[   54.915894]  ? __warn+0x84/0x140
[   54.915905]  ? get_proto_defrag_hook+0x137/0x160
[   54.915908]  ? __report_bug+0xea/0x100
[   54.915925]  ? report_bug+0x2b/0x80
[   54.915928]  ? handle_bug+0x3c/0x70
[   54.915939]  ? exc_invalid_op+0x18/0x70
[   54.915942]  ? asm_exc_invalid_op+0x1a/0x20
[   54.915948]  ? get_proto_defrag_hook+0x137/0x160
[   54.915950]  bpf_nf_link_attach+0x1eb/0x240
[   54.915953]  link_create+0x173/0x290
[   54.915969]  __sys_bpf+0x588/0x8f0
[   54.915974]  __x64_sys_bpf+0x20/0x30
[   54.915977]  do_syscall_64+0x45/0xf0
[   54.915989]  entry_SYSCALL_64_after_hwframe+0x6e/0x76
[   54.915998] RIP: 0033:0x7fb6e9daa51d
[   54.916001] Code: 00 c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 2b 89 0c 00 f7 d8 64 89 01 48
[   54.916003] RSP: 002b:00007ffd15e78ed8 EFLAGS: 00000246 ORIG_RAX: 0000000000000141
[   54.916006] RAX: ffffffffffffffda RBX: 00007ffd15e78fc0 RCX: 00007fb6e9daa51d
[   54.916007] RDX: 0000000000000040 RSI: 00007ffd15e78ef0 RDI: 000000000000001c
[   54.916009] RBP: 000000000000002d R08: 00007fb6e9e73a60 R09: 0000000000000001
[   54.916010] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000006
[   54.916012] R13: 0000000000000006 R14: 0000000000000000 R15: 0000000000000000
[   54.916014]  </TASK>
[   54.916015] ---[ end trace 0000000000000000 ]---

Fixes: 91721c2d02d3 ("netfilter: bpf: Support BPF_F_NETFILTER_IP_DEFRAG in netfilter link")
Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>
Acked-by: Daniel Xu <dxu@dxuuu.xyz>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

leds: trigger: netdev: fix RTNL handling to prevent potential deadlock

When working on LED support for r8169 I got the following lockdep
warning. Easiest way to prevent this scenario seems to be to take
the RTNL lock before the trigger_data lock in set_device_name().

======================================================
WARNING: possible circular locking dependency detected
6.7.0-rc2-next-20231124+ #2 Not tainted
------------------------------------------------------
bash/383 is trying to acquire lock:
ffff888103aa1c68 (&trigger_data->lock){+.+.}-{3:3}, at: netdev_trig_notify+0xec/0x190 [ledtrig_netdev]

but task is already holding lock:
ffffffff8cddf808 (rtnl_mutex){+.+.}-{3:3}, at: rtnl_lock+0x12/0x20

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (rtnl_mutex){+.+.}-{3:3}:
       __mutex_lock+0x9b/0xb50
       mutex_lock_nested+0x16/0x20
       rtnl_lock+0x12/0x20
       set_device_name+0xa9/0x120 [ledtrig_netdev]
       netdev_trig_activate+0x1a1/0x230 [ledtrig_netdev]
       led_trigger_set+0x172/0x2c0
       led_trigger_write+0xf1/0x140
       sysfs_kf_bin_write+0x5d/0x80
       kernfs_fop_write_iter+0x15d/0x210
       vfs_write+0x1f0/0x510
       ksys_write+0x6c/0xf0
       __x64_sys_write+0x14/0x20
       do_syscall_64+0x3f/0xf0
       entry_SYSCALL_64_after_hwframe+0x6c/0x74

-> #0 (&trigger_data->lock){+.+.}-{3:3}:
       __lock_acquire+0x1459/0x25a0
       lock_acquire+0xc8/0x2d0
       __mutex_lock+0x9b/0xb50
       mutex_lock_nested+0x16/0x20
       netdev_trig_notify+0xec/0x190 [ledtrig_netdev]
       call_netdevice_register_net_notifiers+0x5a/0x100
       register_netdevice_notifier+0x85/0x120
       netdev_trig_activate+0x1d4/0x230 [ledtrig_netdev]
       led_trigger_set+0x172/0x2c0
       led_trigger_write+0xf1/0x140
       sysfs_kf_bin_write+0x5d/0x80
       kernfs_fop_write_iter+0x15d/0x210
       vfs_write+0x1f0/0x510
       ksys_write+0x6c/0xf0
       __x64_sys_write+0x14/0x20
       do_syscall_64+0x3f/0xf0
       entry_SYSCALL_64_after_hwframe+0x6c/0x74

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(rtnl_mutex);
                               lock(&trigger_data->lock);
                               lock(rtnl_mutex);
  lock(&trigger_data->lock);

 *** DEADLOCK ***

8 locks held by bash/383:
 #0: ffff888103ff33f0 (sb_writers#3){.+.+}-{0:0}, at: ksys_write+0x6c/0xf0
 #1: ffff888103aa1e88 (&of->mutex){+.+.}-{3:3}, at: kernfs_fop_write_iter+0x114/0x210
 #2: ffff8881036f1890 (kn->active#82){.+.+}-{0:0}, at: kernfs_fop_write_iter+0x11d/0x210
 #3: ffff888108e2c358 (&led_cdev->led_access){+.+.}-{3:3}, at: led_trigger_write+0x30/0x140
 #4: ffffffff8cdd9e10 (triggers_list_lock){++++}-{3:3}, at: led_trigger_write+0x75/0x140
 #5: ffff888108e2c270 (&led_cdev->trigger_lock){++++}-{3:3}, at: led_trigger_write+0xe3/0x140
 #6: ffffffff8cdde3d0 (pernet_ops_rwsem){++++}-{3:3}, at: register_netdevice_notifier+0x1c/0x120
 #7: ffffffff8cddf808 (rtnl_mutex){+.+.}-{3:3}, at: rtnl_lock+0x12/0x20

stack backtrace:
CPU: 0 PID: 383 Comm: bash Not tainted 6.7.0-rc2-next-20231124+ #2
Hardware name: Default string Default string/Default string, BIOS ADLN.M6.SODIMM.ZB.CY.015 08/08/2023
Call Trace:
 <TASK>
 dump_stack_lvl+0x5c/0xd0
 dump_stack+0x10/0x20
 print_circular_bug+0x2dd/0x410
 check_noncircular+0x131/0x150
 __lock_acquire+0x1459/0x25a0
 lock_acquire+0xc8/0x2d0
 ? netdev_trig_notify+0xec/0x190 [ledtrig_netdev]
 __mutex_lock+0x9b/0xb50
 ? netdev_trig_notify+0xec/0x190 [ledtrig_netdev]
 ? __this_cpu_preempt_check+0x13/0x20
 ? netdev_trig_notify+0xec/0x190 [ledtrig_netdev]
 ? __cancel_work_timer+0x11c/0x1b0
 ? __mutex_lock+0x123/0xb50
 mutex_lock_nested+0x16/0x20
 ? mutex_lock_nested+0x16/0x20
 netdev_trig_notify+0xec/0x190 [ledtrig_netdev]
 call_netdevice_register_net_notifiers+0x5a/0x100
 register_netdevice_notifier+0x85/0x120
 netdev_trig_activate+0x1d4/0x230 [ledtrig_netdev]
 led_trigger_set+0x172/0x2c0
 ? preempt_count_add+0x49/0xc0
 led_trigger_write+0xf1/0x140
 sysfs_kf_bin_write+0x5d/0x80
 kernfs_fop_write_iter+0x15d/0x210
 vfs_write+0x1f0/0x510
 ksys_write+0x6c/0xf0
 __x64_sys_write+0x14/0x20
 do_syscall_64+0x3f/0xf0
 entry_SYSCALL_64_after_hwframe+0x6c/0x74
RIP: 0033:0x7f269055d034
Code: c7 00 16 00 00 00 b8 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 f3 0f 1e fa 80 3d 35 c3 0d 00 00 74 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 c3 0f 1f 00 48 83 ec 28 48 89 54 24 18 48
RSP: 002b:00007ffddb7ef748 EFLAGS: 00000202 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 0000000000000007 RCX: 00007f269055d034
RDX: 0000000000000007 RSI: 000055bf5f4af3c0 RDI: 0000000000000001
RBP: 000055bf5f4af3c0 R08: 0000000000000073 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000202 R12: 0000000000000007
R13: 00007f26906325c0 R14: 00007f269062ff20 R15: 0000000000000000
 </TASK>

Fixes: d5e01266e7f5 ("leds: trigger: netdev: add additional specific link speed mode")
Cc: stable@vger.kernel.org
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/fb5c8294-2a10-4bf5-8f10-3d2b77d2757e@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

dt-bindings: interrupt-controller: Allow #power-domain-cells

MPM provides a single genpd. Allow #power-domain-cells = <0>.

Fixes: 54fc9851c0e0 ("dt-bindings: interrupt-controller: Add Qualcomm MPM support")
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Link: https://lore.kernel.org/r/20231129-topic-mpmbindingspd-v2-1-acbe909ceee1@linaro.org
Signed-off-by: Rob Herring <robh@kernel.org>

Merge branch 'octeontx2-af-miscellaneous-fixes'

Geetha sowjanya says:

====================
octeontx2-af: miscellaneous fixes

The series of patches fixes various issues related to mcs
and NIX link registers.

v3-v4:
 Used FIELD_PREP macro and proper data types.

v2-v3:
 Fixed typo error in patch 4 commit message.

v1-v2:
 Fixed author name for patch 5.
 Added Reviewed-by.
====================

Link: https://lore.kernel.org/r/20231205080434.27604-1-gakula@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: Update Tx link register range

On new silicons the TX channels for transmit level has increased.
This patch fixes the respective register offset range to
configure the newly added channels.

Fixes: b279bbb3314e ("octeontx2-af: NIX Tx scheduler queue config support")
Signed-off-by: Rahul Bhansali <rbhansali@marvell.com>
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: Add missing mcs flr handler call

If mcs resources are attached to PF/VF. These resources need
to be freed on FLR. This patch add missing mcs flr call on PF FLR.

Fixes: bd69476e86fc ("octeontx2-af: cn10k: mcs: Install a default TCAM for normal traffic")
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: Fix mcs stats register address

This patch adds the miss mcs stats register
for mcs supported platforms.

Fixes: 9312150af8da ("octeontx2-af: cn10k: mcs: Support for stats collection")
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: Fix mcs sa cam entries size

On latest silicon versions SA cam entries increased to 256.
This patch fixes the datatype of sa_entries in mcs_hw_info
struct to u16 to hold 256 entries.

Fixes: 080bbd19c9dd ("octeontx2-af: cn10k: mcs: Add mailboxes for port related operations")
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: Adjust Tx credits when MCS external bypass is disabled

When MCS external bypass is disabled, MCS returns additional
2 credits(32B) for every packet Tx'ed on LMAC. To account for
these extra credits, NIX_AF_TX_LINKX_NORM_CREDIT.CC_MCS_CNT
needs to be configured as otherwise NIX Tx credits would overflow
and will never be returned to idle state credit count
causing issues with credit control and MTU change.

This patch fixes the same by configuring CC_MCS_CNT at probe
time for MCS enabled SoC's

Fixes: bd69476e86fc ("octeontx2-af: cn10k: mcs: Install a default TCAM for normal traffic")
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

octeontx2-af: cn10k: Increase outstanding LMTST transactions

Currently the number of outstanding store transactions issued by AP as
a part of LMTST operation is set to 1 i.e default value.
This patch set to max supported value to increase the performance.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Link: https://lore.kernel.org/r/20231205055241.26355-1-gakula@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'tcp-ao-fixes'

Dmitry Safonov says:

====================
TCP-AO fixes

Changes from v4:
- Dropped 2 patches on which there's no consensus. They will require
  more work TBD if they may made acceptable. Those are:
  o "net/tcp: Allow removing current/rnext TCP-AO keys on TCP_LISTEN sockets"
  o "net/tcp: Store SNEs + SEQs on ao_info"

Changes from v3:
- Don't restrict adding any keys on TCP-AO connection in VRF, but only
  the ones that don't match l3index (David)

Changes from v2:
- rwlocks are problematic in net code (Paolo)
  Changed the SNE code to avoid spin/rw locks on RX/TX fastpath by
  double-accounting SEQ numbers for TCP-AO enabled connections.

Changes from v1:
- Use tcp_can_repair_sock() helper to limit TCP_AO_REPAIR (Eric)
- Instead of hook to listen() syscall, allow removing current/rnext keys
  on TCP_LISTEN (addressing Eric's objection)
- Add sne_lock to protect snd_sne/rcv_sne
- Don't move used_tcp_ao in struct tcp_request_sock (Eric)

I've been working on TCP-AO key-rotation selftests and as a result
exercised some corner-cases that are not usually met in production.

Here are a bunch of semi-related fixes:
- Documentation typo (reported by Markus Elfring)
- Proper alignment for TCP-AO option in TCP header that has MAC length
  of non 4 bytes (now a selftest with randomized maclen/algorithm/etc
  passes)
- 3 uAPI restricting patches that disallow more things to userspace in
  order to prevent it shooting itself in any parts of the body
- SNEs READ_ONCE()/WRITE_ONCE() that went missing by my human factor
- Avoid storing MAC length from SYN header as SYN-ACK will use
  rnext_key.maclen (drops an extra check that fails on new selftests)
====================

Link: https://lore.kernel.org/r/
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/tcp: Don't store TCP-AO maclen on reqsk

This extra check doesn't work for a handshake when SYN segment has
(current_key.maclen != rnext_key.maclen). It could be amended to
preserve rnext_key.maclen instead of current_key.maclen, but that
requires a lookup on listen socket.

Originally, this extra maclen check was introduced just because it was
cheap. Drop it and convert tcp_request_sock::maclen into boolean
tcp_request_sock::used_tcp_ao.

Fixes: 06b22ef29591 ("net/tcp: Wire TCP-AO to request sockets")
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/tcp: Don't add key with non-matching VRF on connected sockets

If the connection was established, don't allow adding TCP-AO keys that
don't match the peer. Currently, there are checks for ip-address
matching, but L3 index check is missing. Add it to restrict userspace
shooting itself somewhere.

Yet, nothing restricts the CAP_NET_RAW user from trying to shoot
themselves by performing setsockopt(SO_BINDTODEVICE) or
setsockopt(SO_BINDTOIFINDEX) over an established TCP-AO connection.
So, this is just "minimum effort" to potentially save someone's
debugging time, rather than a full restriction on doing weird things.

Fixes: 248411b8cb89 ("net/tcp: Wire up l3index to TCP-AO")
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/tcp: Limit TCP_AO_REPAIR to non-listen sockets

Listen socket is not an established TCP connection, so
setsockopt(TCP_AO_REPAIR) doesn't have any impact.

Restrict this uAPI for listen sockets.

Fixes: faadfaba5e01 ("net/tcp: Add TCP_AO_REPAIR")
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net/tcp: Consistently align TCP-AO option in the header

Currently functions that pre-calculate TCP header options length use
unaligned TCP-AO header + MAC-length for skb reservation.
And the functions that actually write TCP-AO options into skb do align
the header. Nothing good can come out of this for ((maclen % 4) != 0).

Provide tcp_ao_len_aligned() helper and use it everywhere for TCP
header options space calculations.

Fixes: 1e03d32bea8e ("net/tcp: Add TCP-AO sign to outgoing packets")
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Documentation/tcp: Fix an obvious typo

Yep, my VIM spellchecker is not good enough for typos like this one.

Fixes: 7fe0e38bb669 ("Documentation/tcp: Add TCP-AO documentation")
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: linux-doc@vger.kernel.org
Reported-by: Markus Elfring <Markus.Elfring@web.de>
Closes: https://lore.kernel.org/all/2745ab4e-acac-40d4-83bf-37f2600d0c3d@web.de/
Signed-off-by: Dmitry Safonov <dima@arista.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

Merge branch 'there-are-some-bugfix-for-the-hns-ethernet-driver'

Jijie Shao says:

====================
There are some bugfix for the HNS ethernet driver

There are some bugfix for the HNS ethernet driver
---
changeLog:
v2 -> v3:
  - Refine the commit msg as Wojciech suggestions
  - Reconstruct the "hns_mac_link_anti_shake" function suggested by Wojciech
  v2: https://lore.kernel.org/all/20231204011051.4055031-1-shaojijie@huawei.com/
v1 -> v2:
  - Fixed the internal function is not decorated with static issue, suggested by Jakub
  v1: https://lore.kernel.org/all/20231201102703.4134592-1-shaojijie@huawei.com/
---
====================

Link: https://lore.kernel.org/r/20231204143232.3221542-1-shaojijie@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: hns: fix fake link up on xge port

If a xge port just connect with an optical module and no fiber,
it may have a fake link up because there may be interference on
the hardware. This patch adds an anti-shake to avoid the problem.
And the time of anti-shake is base on tests.

Fixes: b917078c1c10 ("net: hns: Add ACPI support to check SFP present")
Signed-off-by: Yonglong Liu <liuyonglong@huawei.com>
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: hns: fix wrong head when modify the tx feature when sending packets

Upon changing the tx feature, the hns driver will modify the
maybe_stop_tx() and fill_desc() functions, if the modify happens
during packet sending, will cause the hardware and software
pointers do not match, and the port can not work anymore.

This patch deletes the maybe_stop_tx() and fill_desc() functions
modification when setting tx feature, and use the skb_is_gro()
to determine which functions to use in the tx path.

Fixes: 38f616da1c28 ("net:hns: Add support of ethtool TSO set option for Hip06 in HNS")
Signed-off-by: Yonglong Liu <liuyonglong@huawei.com>
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: atlantic: Fix NULL dereference of skb pointer in

If is_ptp_ring == true in the loop of __aq_ring_xdp_clean function,
then a timestamp is stored from a packet in a field of skb object,
which is not allocated at the moment of the call (skb == NULL).

Generalize aq_ptp_extract_ts and other affected functions so they don't
work with struct sk_buff*, but with struct skb_shared_hwtstamps*.

Found by Linux Verification Center (linuxtesting.org) with SVACE

Fixes: 26efaef759a1 ("net: atlantic: Implement xdp data plane")
Signed-off-by: Daniil Maximov <daniil31415it@gmail.com>
Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Link: https://lore.kernel.org/r/20231204085810.1681386-1-daniil31415it@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>