OSDN Git Service

uclinux-h8/linux.git
2 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf
Jakub Kicinski [Thu, 9 Dec 2021 01:02:33 +0000 (17:02 -0800)]
Merge git://git./pub/scm/linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

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

1) Fix bogus compilter warning in nfnetlink_queue, from Florian Westphal.

2) Don't run conntrack on vrf with !dflt qdisc, from Nicolas Dichtel.

3) Fix nft_pipapo bucket load in AVX2 lookup routine for six 8-bit
   groups, from Stefano Brivio.

4) Break rule evaluation on malformed TCP options.

5) Use socat instead of nc in selftests/netfilter/nft_zones_many.sh,
   also from Florian

6) Fix KCSAN data-race in conntrack timeout updates, from Eric Dumazet.

* git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf:
  netfilter: conntrack: annotate data-races around ct->timeout
  selftests: netfilter: switch zone stress to socat
  netfilter: nft_exthdr: break evaluation if setting TCP option fails
  selftests: netfilter: Add correctness test for mac,net set type
  nft_set_pipapo: Fix bucket load in AVX2 lookup routine for six 8-bit groups
  vrf: don't run conntrack on vrf with !dflt qdisc
  netfilter: nfnetlink_queue: silence bogus compiler warning
====================

Link: https://lore.kernel.org/r/20211209000847.102598-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net...
Jakub Kicinski [Thu, 9 Dec 2021 00:36:12 +0000 (16:36 -0800)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2021-12-08

Yahui adds re-initialization of Flow Director for VF reset.

Paul restores interrupts when enabling VFs.

Dave re-adds bandwidth check for DCBNL and moves DSCP mode check
earlier in the function.

Jesse prevents reporting of dropped packets that occur during
initialization and fixes reporting of statistics which could occur with
frequent reads.

Michal corrects setting of protocol type for UDP header and fixes lack
of differentiation when adding filters for tunnels.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  ice: safer stats processing
  ice: fix adding different tunnels
  ice: fix choosing UDP header type
  ice: ignore dropped packets during init
  ice: Fix problems with DSCP QoS implementation
  ice: rearm other interrupt cause register after enabling VFs
  ice: fix FDIR init missing when reset VF
====================

Link: https://lore.kernel.org/r/20211208211144.2629867-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 9 Dec 2021 00:06:44 +0000 (16:06 -0800)]
Merge https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
bpf 2021-12-08

We've added 12 non-merge commits during the last 22 day(s) which contain
a total of 29 files changed, 659 insertions(+), 80 deletions(-).

The main changes are:

1) Fix an off-by-two error in packet range markings and also add a batch of
   new tests for coverage of these corner cases, from Maxim Mikityanskiy.

2) Fix a compilation issue on MIPS JIT for R10000 CPUs, from Johan Almbladh.

3) Fix two functional regressions and a build warning related to BTF kfunc
   for modules, from Kumar Kartikeya Dwivedi.

4) Fix outdated code and docs regarding BPF's migrate_disable() use on non-
   PREEMPT_RT kernels, from Sebastian Andrzej Siewior.

5) Add missing includes in order to be able to detangle cgroup vs bpf header
   dependencies, from Jakub Kicinski.

6) Fix regression in BPF sockmap tests caused by missing detachment of progs
   from sockets when they are removed from the map, from John Fastabend.

7) Fix a missing "no previous prototype" warning in x86 JIT caused by BPF
   dispatcher, from Björn Töpel.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpf: Add selftests to cover packet access corner cases
  bpf: Fix the off-by-two error in range markings
  treewide: Add missing includes masked by cgroup -> bpf dependency
  tools/resolve_btfids: Skip unresolved symbol warning for empty BTF sets
  bpf: Fix bpf_check_mod_kfunc_call for built-in modules
  bpf: Make CONFIG_DEBUG_INFO_BTF depend upon CONFIG_BPF_SYSCALL
  mips, bpf: Fix reference to non-existing Kconfig symbol
  bpf: Make sure bpf_disable_instrumentation() is safe vs preemption.
  Documentation/locking/locktypes: Update migrate_disable() bits.
  bpf, sockmap: Re-evaluate proto ops when psock is removed from sockmap
  bpf, sockmap: Attach map progs to psock early for feature probes
  bpf, x86: Fix "no previous prototype" warning
====================

Link: https://lore.kernel.org/r/20211208155125.11826-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: dsa: mv88e6xxx: fix "don't use PHY_DETECT on internal PHY's"
Russell King (Oracle) [Tue, 7 Dec 2021 10:32:43 +0000 (10:32 +0000)]
net: dsa: mv88e6xxx: fix "don't use PHY_DETECT on internal PHY's"

This commit fixes a misunderstanding in commit 4a3e0aeddf09 ("net: dsa:
mv88e6xxx: don't use PHY_DETECT on internal PHY's").

For Marvell DSA switches with the PHY_DETECT bit (for non-6250 family
devices), controls whether the PPU polls the PHY to retrieve the link,
speed, duplex and pause status to update the port configuration. This
applies for both internal and external PHYs.

For some switches such as 88E6352 and 88E6390X, PHY_DETECT has an
additional function of enabling auto-media mode between the internal
PHY and SERDES blocks depending on which first gains link.

The original intention of commit 5d5b231da7ac (net: dsa: mv88e6xxx: use
PHY_DETECT in mac_link_up/mac_link_down) was to allow this bit to be
used to detect when this propagation is enabled, and allow software to
update the port configuration. This has found to be necessary for some
switches which do not automatically propagate status from the SERDES to
the port, which includes the 88E6390. However, commit 4a3e0aeddf09
("net: dsa: mv88e6xxx: don't use PHY_DETECT on internal PHY's") breaks
this assumption.

Maarten Zanders has confirmed that the issue he was addressing was for
an 88E6250 switch, which does not have a PHY_DETECT bit in bit 12, but
instead a link status bit. Therefore, mv88e6xxx_port_ppu_updates() does
not report correctly.

This patch resolves the above issues by reverting Maarten's change and
instead making mv88e6xxx_port_ppu_updates() indicate whether the port
is internal for the 88E6250 family of switches.

  Yes, you're right, I'm targeting the 6250 family. And yes, your
  suggestion would solve my case and is a better implementation for
  the other devices (as far as I can see).

Fixes: 4a3e0aeddf09 ("net: dsa: mv88e6xxx: don't use PHY_DETECT on internal PHY's")
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Tested-by: Maarten Zanders <maarten.zanders@mind.be>
Link: https://lore.kernel.org/r/E1muXm7-00EwJB-7n@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoice: safer stats processing
Jesse Brandeburg [Sat, 13 Nov 2021 01:06:02 +0000 (17:06 -0800)]
ice: safer stats processing

The driver was zeroing live stats that could be fetched by
ndo_get_stats64 at any time. This could result in inconsistent
statistics, and the telltale sign was when reading stats frequently from
/proc/net/dev, the stats would go backwards.

Fix by collecting stats into a local, and delaying when we write to the
structure so it's not incremental.

Fixes: fcea6f3da546 ("ice: Add stats and ethtool support")
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agobpf: Add selftests to cover packet access corner cases
Maxim Mikityanskiy [Tue, 7 Dec 2021 08:15:21 +0000 (10:15 +0200)]
bpf: Add selftests to cover packet access corner cases

This commit adds BPF verifier selftests that cover all corner cases by
packet boundary checks. Specifically, 8-byte packet reads are tested at
the beginning of data and at the beginning of data_meta, using all kinds
of boundary checks (all comparison operators: <, >, <=, >=; both
permutations of operands: data + length compared to end, end compared to
data + length). For each case there are three tests:

1. Length is just enough for an 8-byte read. Length is either 7 or 8,
   depending on the comparison.

2. Length is increased by 1 - should still pass the verifier. These
   cases are useful, because they failed before commit 2fa7d94afc1a
   ("bpf: Fix the off-by-two error in range markings").

3. Length is decreased by 1 - should be rejected by the verifier.

Some existing tests are just renamed to avoid duplication.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211207081521.41923-1-maximmi@nvidia.com
2 years agonet: fec: only clear interrupt of handling queue in fec_enet_rx_queue()
Joakim Zhang [Mon, 6 Dec 2021 13:54:57 +0000 (21:54 +0800)]
net: fec: only clear interrupt of handling queue in fec_enet_rx_queue()

Background:
We have a customer is running a Profinet stack on the 8MM which receives and
responds PNIO packets every 4ms and PNIO-CM packets every 40ms. However, from
time to time the received PNIO-CM package is "stock" and is only handled when
receiving a new PNIO-CM or DCERPC-Ping packet (tcpdump shows the PNIO-CM and
the DCERPC-Ping packet at the same time but the PNIO-CM HW timestamp is from
the expected 40 ms and not the 2s delay of the DCERPC-Ping).

After debugging, we noticed PNIO, PNIO-CM and DCERPC-Ping packets would
be handled by different RX queues.

The root cause should be driver ack all queues' interrupt when handle a
specific queue in fec_enet_rx_queue(). The blamed patch is introduced to
receive as much packets as possible once to avoid interrupt flooding.
But it's unreasonable to clear other queues'interrupt when handling one
queue, this patch tries to fix it.

Fixes: ed63f1dcd578 (net: fec: clear receive interrupts before processing a packet)
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Reported-by: Nicolas Diaz <nicolas.diaz@nxp.com>
Signed-off-by: Joakim Zhang <qiangqing.zhang@nxp.com>
Link: https://lore.kernel.org/r/20211206135457.15946-1-qiangqing.zhang@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Jakub Kicinski [Wed, 8 Dec 2021 05:33:11 +0000 (21:33 -0800)]
Merge branch '40GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2021-12-06

This series contains updates to iavf and i40e drivers.

Mitch adds restoration of MSI state during reset for iavf.

Michal fixes checking and reporting of descriptor count changes to
communicate changes and/or issues for iavf.

Karen resolves an issue with failed handling of VF requests while a VF
reset is occurring for i40e.

Mateusz removes clearing of VF requested queue count when configuring
VF ADQ for i40e.

Norbert fixes a NULL pointer dereference that can occur when getting VSI
descriptors for i40e.
====================

Link: https://lore.kernel.org/r/20211206183519.2733180-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch 'net-phy-fix-doc-build-warning'
Jakub Kicinski [Wed, 8 Dec 2021 05:22:01 +0000 (21:22 -0800)]
Merge branch 'net-phy-fix-doc-build-warning'

Yanteng Si says:

====================
net: phy: Fix doc build warnings
====================

Link: https://lore.kernel.org/r/cover.1638776933.git.siyanteng@loongson.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: phy: Add the missing blank line in the phylink_suspend comment
Yanteng Si [Mon, 6 Dec 2021 08:12:28 +0000 (16:12 +0800)]
net: phy: Add the missing blank line in the phylink_suspend comment

Fix warning as:

Documentation/networking/kapi:147: ./drivers/net/phy/phylink.c:1657: WARNING: Unexpected indentation.
Documentation/networking/kapi:147: ./drivers/net/phy/phylink.c:1658: WARNING: Block quote ends without a blank line; unexpected unindent.

Signed-off-by: Yanteng Si <siyanteng@loongson.cn>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: phy: Remove unnecessary indentation in the comments of phy_device
Yanteng Si [Mon, 6 Dec 2021 08:12:27 +0000 (16:12 +0800)]
net: phy: Remove unnecessary indentation in the comments of phy_device

Fix warning as:

linux-next/Documentation/networking/kapi:122: ./include/linux/phy.h:543: WARNING: Unexpected indentation.
linux-next/Documentation/networking/kapi:122: ./include/linux/phy.h:544: WARNING: Block quote ends without a blank line; unexpected unindent.
linux-next/Documentation/networking/kapi:122: ./include/linux/phy.h:546: WARNING: Unexpected indentation.

Suggested-by: Akira Yokosawa <akiyks@gmail.com>
Signed-off-by: Yanteng Si <siyanteng@loongson.cn>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agogve: fix for null pointer dereference.
Ameer Hamza [Sun, 5 Dec 2021 18:38:10 +0000 (23:38 +0500)]
gve: fix for null pointer dereference.

Avoid passing NULL skb to __skb_put() function call if
napi_alloc_skb() returns NULL.

Fixes: 37149e9374bf ("gve: Implement packet continuation for RX.")
Signed-off-by: Ameer Hamza <amhamza.mgc@gmail.com>
Link: https://lore.kernel.org/r/20211205183810.8299-1-amhamza.mgc@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMAINTAINERS: net: mlxsw: Remove Jiri as a maintainer, add myself
Petr Machata [Mon, 6 Dec 2021 16:17:23 +0000 (17:17 +0100)]
MAINTAINERS: net: mlxsw: Remove Jiri as a maintainer, add myself

Jiri has moved on and will not carry out the mlxsw maintainership duty any
longer. Add myself as a co-maintainer instead.

Signed-off-by: Petr Machata <petrm@nvidia.com>
Acked-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://lore.kernel.org/r/45b54312cdebaf65c5d110b15a5dd2df795bf2be.1638807297.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch 'net-tls-cover-all-ciphers-with-tests'
Jakub Kicinski [Wed, 8 Dec 2021 04:18:09 +0000 (20:18 -0800)]
Merge branch 'net-tls-cover-all-ciphers-with-tests'

Vadim Fedorenko says:

====================
net: tls: cover all ciphers with tests

Recent patches to Kernel TLS showed that some ciphers are not covered
with tests. Let's cover missed.
====================

Link: https://lore.kernel.org/r/20211206213932.7508-1-vfedorenko@novek.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: tls: add missing AES256-GCM cipher
Vadim Fedorenko [Mon, 6 Dec 2021 21:39:32 +0000 (00:39 +0300)]
selftests: tls: add missing AES256-GCM cipher

Add tests for TLSv1.2 and TLSv1.3 with AES256-GCM cipher

Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: tls: add missing AES-CCM cipher tests
Vadim Fedorenko [Mon, 6 Dec 2021 21:39:31 +0000 (00:39 +0300)]
selftests: tls: add missing AES-CCM cipher tests

Add tests for TLSv1.2 and TLSv1.3 with AES-CCM cipher.

Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonetfilter: conntrack: annotate data-races around ct->timeout
Eric Dumazet [Tue, 7 Dec 2021 18:03:23 +0000 (10:03 -0800)]
netfilter: conntrack: annotate data-races around ct->timeout

(struct nf_conn)->timeout can be read/written locklessly,
add READ_ONCE()/WRITE_ONCE() to prevent load/store tearing.

BUG: KCSAN: data-race in __nf_conntrack_alloc / __nf_conntrack_find_get

write to 0xffff888132e78c08 of 4 bytes by task 6029 on cpu 0:
 __nf_conntrack_alloc+0x158/0x280 net/netfilter/nf_conntrack_core.c:1563
 init_conntrack+0x1da/0xb30 net/netfilter/nf_conntrack_core.c:1635
 resolve_normal_ct+0x502/0x610 net/netfilter/nf_conntrack_core.c:1746
 nf_conntrack_in+0x1c5/0x88f net/netfilter/nf_conntrack_core.c:1901
 ipv6_conntrack_local+0x19/0x20 net/netfilter/nf_conntrack_proto.c:414
 nf_hook_entry_hookfn include/linux/netfilter.h:142 [inline]
 nf_hook_slow+0x72/0x170 net/netfilter/core.c:619
 nf_hook include/linux/netfilter.h:262 [inline]
 NF_HOOK include/linux/netfilter.h:305 [inline]
 ip6_xmit+0xa3a/0xa60 net/ipv6/ip6_output.c:324
 inet6_csk_xmit+0x1a2/0x1e0 net/ipv6/inet6_connection_sock.c:135
 __tcp_transmit_skb+0x132a/0x1840 net/ipv4/tcp_output.c:1402
 tcp_transmit_skb net/ipv4/tcp_output.c:1420 [inline]
 tcp_write_xmit+0x1450/0x4460 net/ipv4/tcp_output.c:2680
 __tcp_push_pending_frames+0x68/0x1c0 net/ipv4/tcp_output.c:2864
 tcp_push_pending_frames include/net/tcp.h:1897 [inline]
 tcp_data_snd_check+0x62/0x2e0 net/ipv4/tcp_input.c:5452
 tcp_rcv_established+0x880/0x10e0 net/ipv4/tcp_input.c:5947
 tcp_v6_do_rcv+0x36e/0xa50 net/ipv6/tcp_ipv6.c:1521
 sk_backlog_rcv include/net/sock.h:1030 [inline]
 __release_sock+0xf2/0x270 net/core/sock.c:2768
 release_sock+0x40/0x110 net/core/sock.c:3300
 sk_stream_wait_memory+0x435/0x700 net/core/stream.c:145
 tcp_sendmsg_locked+0xb85/0x25a0 net/ipv4/tcp.c:1402
 tcp_sendmsg+0x2c/0x40 net/ipv4/tcp.c:1440
 inet6_sendmsg+0x5f/0x80 net/ipv6/af_inet6.c:644
 sock_sendmsg_nosec net/socket.c:704 [inline]
 sock_sendmsg net/socket.c:724 [inline]
 __sys_sendto+0x21e/0x2c0 net/socket.c:2036
 __do_sys_sendto net/socket.c:2048 [inline]
 __se_sys_sendto net/socket.c:2044 [inline]
 __x64_sys_sendto+0x74/0x90 net/socket.c:2044
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x44/0xd0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

read to 0xffff888132e78c08 of 4 bytes by task 17446 on cpu 1:
 nf_ct_is_expired include/net/netfilter/nf_conntrack.h:286 [inline]
 ____nf_conntrack_find net/netfilter/nf_conntrack_core.c:776 [inline]
 __nf_conntrack_find_get+0x1c7/0xac0 net/netfilter/nf_conntrack_core.c:807
 resolve_normal_ct+0x273/0x610 net/netfilter/nf_conntrack_core.c:1734
 nf_conntrack_in+0x1c5/0x88f net/netfilter/nf_conntrack_core.c:1901
 ipv6_conntrack_local+0x19/0x20 net/netfilter/nf_conntrack_proto.c:414
 nf_hook_entry_hookfn include/linux/netfilter.h:142 [inline]
 nf_hook_slow+0x72/0x170 net/netfilter/core.c:619
 nf_hook include/linux/netfilter.h:262 [inline]
 NF_HOOK include/linux/netfilter.h:305 [inline]
 ip6_xmit+0xa3a/0xa60 net/ipv6/ip6_output.c:324
 inet6_csk_xmit+0x1a2/0x1e0 net/ipv6/inet6_connection_sock.c:135
 __tcp_transmit_skb+0x132a/0x1840 net/ipv4/tcp_output.c:1402
 __tcp_send_ack+0x1fd/0x300 net/ipv4/tcp_output.c:3956
 tcp_send_ack+0x23/0x30 net/ipv4/tcp_output.c:3962
 __tcp_ack_snd_check+0x2d8/0x510 net/ipv4/tcp_input.c:5478
 tcp_ack_snd_check net/ipv4/tcp_input.c:5523 [inline]
 tcp_rcv_established+0x8c2/0x10e0 net/ipv4/tcp_input.c:5948
 tcp_v6_do_rcv+0x36e/0xa50 net/ipv6/tcp_ipv6.c:1521
 sk_backlog_rcv include/net/sock.h:1030 [inline]
 __release_sock+0xf2/0x270 net/core/sock.c:2768
 release_sock+0x40/0x110 net/core/sock.c:3300
 tcp_sendpage+0x94/0xb0 net/ipv4/tcp.c:1114
 inet_sendpage+0x7f/0xc0 net/ipv4/af_inet.c:833
 rds_tcp_xmit+0x376/0x5f0 net/rds/tcp_send.c:118
 rds_send_xmit+0xbed/0x1500 net/rds/send.c:367
 rds_send_worker+0x43/0x200 net/rds/threads.c:200
 process_one_work+0x3fc/0x980 kernel/workqueue.c:2298
 worker_thread+0x616/0xa70 kernel/workqueue.c:2445
 kthread+0x2c7/0x2e0 kernel/kthread.c:327
 ret_from_fork+0x1f/0x30

value changed: 0x00027cc2 -> 0x00000000

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 17446 Comm: kworker/u4:5 Tainted: G        W         5.16.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: krdsd rds_send_worker

Note: I chose an arbitrary commit for the Fixes: tag,
because I do not think we need to backport this fix to very old kernels.

Fixes: e37542ba111f ("netfilter: conntrack: avoid possible false sharing")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agoselftests: netfilter: switch zone stress to socat
Florian Westphal [Fri, 3 Dec 2021 14:33:23 +0000 (15:33 +0100)]
selftests: netfilter: switch zone stress to socat

centos9 has nmap-ncat which doesn't like the '-q' option, use socat.
While at it, mark test skipped if needed tools are missing.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agonetfilter: nft_exthdr: break evaluation if setting TCP option fails
Pablo Neira Ayuso [Tue, 30 Nov 2021 10:34:04 +0000 (11:34 +0100)]
netfilter: nft_exthdr: break evaluation if setting TCP option fails

Break rule evaluation on malformed TCP options.

Fixes: 99d1712bc41c ("netfilter: exthdr: tcp option set support")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agoselftests: netfilter: Add correctness test for mac,net set type
Stefano Brivio [Sat, 27 Nov 2021 10:33:38 +0000 (11:33 +0100)]
selftests: netfilter: Add correctness test for mac,net set type

The existing net,mac test didn't cover the issue recently reported
by Nikita Yushchenko, where MAC addresses wouldn't match if given
as first field of a concatenated set with AVX2 and 8-bit groups,
because there's a different code path covering the lookup of six
8-bit groups (MAC addresses) if that's the first field.

Add a similar mac,net test, with MAC address and IPv4 address
swapped in the set specification.

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agonft_set_pipapo: Fix bucket load in AVX2 lookup routine for six 8-bit groups
Stefano Brivio [Sat, 27 Nov 2021 10:33:37 +0000 (11:33 +0100)]
nft_set_pipapo: Fix bucket load in AVX2 lookup routine for six 8-bit groups

The sixth byte of packet data has to be looked up in the sixth group,
not in the seventh one, even if we load the bucket data into ymm6
(and not ymm5, for convenience of tracking stalls).

Without this fix, matching on a MAC address as first field of a set,
if 8-bit groups are selected (due to a small set size) would fail,
that is, the given MAC address would never match.

Reported-by: Nikita Yushchenko <nikita.yushchenko@virtuozzo.com>
Cc: <stable@vger.kernel.org> # 5.6.x
Fixes: 7400b063969b ("nft_set_pipapo: Introduce AVX2-based lookup implementation")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Tested-By: Nikita Yushchenko <nikita.yushchenko@virtuozzo.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agovrf: don't run conntrack on vrf with !dflt qdisc
Nicolas Dichtel [Fri, 26 Nov 2021 14:36:12 +0000 (15:36 +0100)]
vrf: don't run conntrack on vrf with !dflt qdisc

After the below patch, the conntrack attached to skb is set to "notrack" in
the context of vrf device, for locally generated packets.
But this is true only when the default qdisc is set to the vrf device. When
changing the qdisc, notrack is not set anymore.
In fact, there is a shortcut in the vrf driver, when the default qdisc is
set, see commit dcdd43c41e60 ("net: vrf: performance improvements for
IPv4") for more details.

This patch ensures that the behavior is always the same, whatever the qdisc
is.

To demonstrate the difference, a new test is added in conntrack_vrf.sh.

Fixes: 8c9c296adfae ("vrf: run conntrack only in context of lower/physdev for locally generated packets")
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Florian Westphal <fw@strlen.de>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agoice: fix adding different tunnels
Michal Swiatkowski [Mon, 22 Nov 2021 15:39:25 +0000 (16:39 +0100)]
ice: fix adding different tunnels

Adding filters with the same values inside for VXLAN and Geneve causes HW
error, because it looks exactly the same. To choose between different
type of tunnels new recipe is needed. Add storing tunnel types in
creating recipes function and start checking it in finding function.

Change getting open tunnels function to return port on correct tunnel
type. This is needed to copy correct port to dummy packet.

Block user from adding enc_dst_port via tc flower, because VXLAN and
Geneve filters can be created only with destination port which was
previously opened.

Fixes: 8b032a55c1bd5 ("ice: low level support for tunnels")
Signed-off-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoice: fix choosing UDP header type
Michal Swiatkowski [Tue, 16 Nov 2021 10:24:26 +0000 (11:24 +0100)]
ice: fix choosing UDP header type

In tunnels packet there can be two UDP headers:
- outer which for hw should be mark as ICE_UDP_OF
- inner which for hw should be mark as ICE_UDP_ILOS or as ICE_TCP_IL if
  inner header is of TCP type

In none tunnels packet header can be:
- UDP, which for hw should be mark as ICE_UDP_ILOS
- TCP, which for hw should be mark as ICE_TCP_IL

Change incorrect ICE_UDP_OF for none tunnel packets to ICE_UDP_ILOS.
ICE_UDP_OF is incorrect for none tunnel packets and setting it leads to
error from hw while adding this kind of recipe.

In summary, for tunnel outer port type should always be set to
ICE_UDP_OF, for none tunnel outer and tunnel inner it should always be
set to ICE_UDP_ILOS.

Fixes: 9e300987d4a8 ("ice: VXLAN and Geneve TC support")
Signed-off-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoice: ignore dropped packets during init
Jesse Brandeburg [Sat, 23 Oct 2021 00:28:17 +0000 (17:28 -0700)]
ice: ignore dropped packets during init

If the hardware is constantly receiving unicast or broadcast packets
during driver load, the device previously counted many GLV_RDPC (VSI
dropped packets) events during init. This causes confusing dropped
packet statistics during driver load. The dropped packets counter
incrementing does stop once the driver finishes loading.

Avoid this problem by baselining our statistics at the end of driver
open instead of the end of probe.

Fixes: cdedef59deb0 ("ice: Configure VSIs for Tx/Rx")
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoice: Fix problems with DSCP QoS implementation
Dave Ertman [Tue, 12 Oct 2021 20:31:21 +0000 (13:31 -0700)]
ice: Fix problems with DSCP QoS implementation

The patch that implemented DSCP QoS implementation removed a
bandwidth check that was used to check for a specific condition
caused by some corner cases.  This check should not of been
removed.

The same patch also added a check for when the DCBx state could
be changed in relation to DSCP, but the check was erroneously
added nested in a check for CEE mode, which made the check useless.

Fix these problems by re-adding the bandwidth check and relocating
the DSCP mode check earlier in the function that changes DCBx state
in the driver.

Fixes: 2a87bd73e50d ("ice: Add DSCP support")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoice: rearm other interrupt cause register after enabling VFs
Paul Greenwalt [Mon, 12 Jul 2021 11:54:25 +0000 (07:54 -0400)]
ice: rearm other interrupt cause register after enabling VFs

The other interrupt cause register (OICR), global interrupt 0, is
disabled when enabling VFs to prevent handling VFLR. If the OICR is
not rearmed then the VF cannot communicate with the PF.

Rearm the OICR after enabling VFs.

Fixes: 916c7fdf5e93 ("ice: Separate VF VSI initialization/creation from reset flow")
Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Tony Brelinski <tony.brelinski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoice: fix FDIR init missing when reset VF
Yahui Cao [Wed, 5 May 2021 21:18:00 +0000 (14:18 -0700)]
ice: fix FDIR init missing when reset VF

When VF is being reset, ice_reset_vf() will be called and FDIR
resource should be released and initialized again.

Fixes: 1f7ea1cd6a37 ("ice: Enable FDIR Configure for AVF")
Signed-off-by: Yahui Cao <yahui.cao@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agonet/qla3xxx: fix an error code in ql_adapter_up()
Dan Carpenter [Tue, 7 Dec 2021 08:24:16 +0000 (11:24 +0300)]
net/qla3xxx: fix an error code in ql_adapter_up()

The ql_wait_for_drvr_lock() fails and returns false, then this
function should return an error code instead of returning success.

The other problem is that the success path prints an error message
netdev_err(ndev, "Releasing driver lock\n");  Delete that and
re-order the code a little to make it more clear.

Fixes: 5a4faa873782 ("[PATCH] qla3xxx NIC driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20211207082416.GA16110@kili
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'linux-can-fixes-for-5.16-20211207' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Tue, 7 Dec 2021 18:32:04 +0000 (10:32 -0800)]
Merge tag 'linux-can-fixes-for-5.16-20211207' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
can 2021-12-07

The 1st patch is by Vincent Mailhol and fixes a use after free in the
pch_can driver.

Dan Carpenter fixes a use after free in the ems_pcmcia sja1000 driver.

The remaining 7 patches target the m_can driver. Brian Silverman
contributes a patch to disable and ignore the ELO interrupt, which is
currently not handled in the driver and may lead to an interrupt
storm. Vincent Mailhol's patch fixes a memory leak in the error path
of the m_can_read_fifo() function. The remaining patches are
contributed by Matthias Schiffer, first a iomap_read_fifo() and
iomap_write_fifo() functions are fixed in the PCI glue driver, then
the clock rate for the Intel Ekhart Lake platform is fixed, the last 3
patches add support for the custom bit timings on the Elkhart Lake
platform.

* tag 'linux-can-fixes-for-5.16-20211207' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  can: m_can: pci: use custom bit timings for Elkhart Lake
  can: m_can: make custom bittiming fields const
  Revert "can: m_can: remove support for custom bit timing"
  can: m_can: pci: fix incorrect reference clock rate
  can: m_can: pci: fix iomap_read_fifo() and iomap_write_fifo()
  can: m_can: m_can_read_fifo: fix memory leak in error branch
  can: m_can: Disable and ignore ELO interrupt
  can: sja1000: fix use after free in ems_pcmcia_add_card()
  can: pch_can: pch_can_rx_normal: fix use after free
====================

Link: https://lore.kernel.org/r/20211207102420.120131-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agocan: m_can: pci: use custom bit timings for Elkhart Lake
Matthias Schiffer [Mon, 15 Nov 2021 09:18:52 +0000 (10:18 +0100)]
can: m_can: pci: use custom bit timings for Elkhart Lake

The relevant datasheet [1] specifies nonstandard limits for the bit timing
parameters. While it is unclear what the exact effect of violating these
limits is, it seems like a good idea to adhere to the documentation.

[1] Intel Atom® x6000E Series, and Intel® Pentium® and Celeron® N and J
    Series Processors for IoT Applications Datasheet,
    Volume 2 (Book 3 of 3), July 2021, Revision 001

Fixes: cab7ffc0324f ("can: m_can: add PCI glue driver for Intel Elkhart Lake")
Link: https://lore.kernel.org/all/9eba5d7c05a48ead4024ffa6e5926f191d8c6b38.1636967198.git.matthias.schiffer@ew.tq-group.com
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: m_can: make custom bittiming fields const
Matthias Schiffer [Mon, 15 Nov 2021 09:18:51 +0000 (10:18 +0100)]
can: m_can: make custom bittiming fields const

The assigned timing structs will be defined a const anyway, so we can
avoid a few casts by declaring the struct fields as const as well.

Link: https://lore.kernel.org/all/4508fa4e639164b2584c49a065d90c78a91fa568.1636967198.git.matthias.schiffer@ew.tq-group.com
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agoRevert "can: m_can: remove support for custom bit timing"
Matthias Schiffer [Mon, 15 Nov 2021 09:18:50 +0000 (10:18 +0100)]
Revert "can: m_can: remove support for custom bit timing"

The timing limits specified by the Elkhart Lake CPU datasheets do not
match the defaults. Let's reintroduce the support for custom bit timings.

This reverts commit 0ddd83fbebbc5537f9d180d31f659db3564be708.

Link: https://lore.kernel.org/all/00c9e2596b1a548906921a574d4ef7a03c0dace0.1636967198.git.matthias.schiffer@ew.tq-group.com
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: m_can: pci: fix incorrect reference clock rate
Matthias Schiffer [Mon, 15 Nov 2021 09:18:49 +0000 (10:18 +0100)]
can: m_can: pci: fix incorrect reference clock rate

When testing the CAN controller on our Ekhart Lake hardware, we
determined that all communication was running with twice the configured
bitrate. Changing the reference clock rate from 100MHz to 200MHz fixed
this. Intel's support has confirmed to us that 200MHz is indeed the
correct clock rate.

Fixes: cab7ffc0324f ("can: m_can: add PCI glue driver for Intel Elkhart Lake")
Link: https://lore.kernel.org/all/c9cf3995f45c363e432b3ae8eb1275e54f009fc8.1636967198.git.matthias.schiffer@ew.tq-group.com
Cc: stable@vger.kernel.org
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Reviewed-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: m_can: pci: fix iomap_read_fifo() and iomap_write_fifo()
Matthias Schiffer [Thu, 18 Nov 2021 14:40:11 +0000 (15:40 +0100)]
can: m_can: pci: fix iomap_read_fifo() and iomap_write_fifo()

The same fix that was previously done in m_can_platform in commit
99d173fbe894 ("can: m_can: fix iomap_read_fifo() and iomap_write_fifo()")
is required in m_can_pci as well to make iomap_read_fifo() and
iomap_write_fifo() work for val_count > 1.

Fixes: 812270e5445b ("can: m_can: Batch FIFO writes during CAN transmit")
Fixes: 1aa6772f64b4 ("can: m_can: Batch FIFO reads during CAN receive")
Link: https://lore.kernel.org/all/20211118144011.10921-1-matthias.schiffer@ew.tq-group.com
Cc: stable@vger.kernel.org
Cc: Matt Kline <matt@bitbashing.io>
Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>
Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: m_can: m_can_read_fifo: fix memory leak in error branch
Vincent Mailhol [Sun, 7 Nov 2021 05:07:55 +0000 (14:07 +0900)]
can: m_can: m_can_read_fifo: fix memory leak in error branch

In m_can_read_fifo(), if the second call to m_can_fifo_read() fails,
the function jump to the out_fail label and returns without calling
m_can_receive_skb(). This means that the skb previously allocated by
alloc_can_skb() is not freed. In other terms, this is a memory leak.

This patch adds a goto label to destroy the skb if an error occurs.

Issue was found with GCC -fanalyzer, please follow the link below for
details.

Fixes: e39381770ec9 ("can: m_can: Disable IRQs on FIFO bus errors")
Link: https://lore.kernel.org/all/20211107050755.70655-1-mailhol.vincent@wanadoo.fr
Cc: stable@vger.kernel.org
Cc: Matt Kline <matt@bitbashing.io>
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: m_can: Disable and ignore ELO interrupt
Brian Silverman [Mon, 29 Nov 2021 22:26:28 +0000 (14:26 -0800)]
can: m_can: Disable and ignore ELO interrupt

With the design of this driver, this condition is often triggered.
However, the counter that this interrupt indicates an overflow is never
read either, so overflowing is harmless.

On my system, when a CAN bus starts flapping up and down, this locks up
the whole system with lots of interrupts and printks.

Specifically, this interrupt indicates the CEL field of ECR has
overflowed. All reads of ECR mask out CEL.

Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support")
Link: https://lore.kernel.org/all/20211129222628.7490-1-brian.silverman@bluerivertech.com
Cc: stable@vger.kernel.org
Signed-off-by: Brian Silverman <brian.silverman@bluerivertech.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: sja1000: fix use after free in ems_pcmcia_add_card()
Dan Carpenter [Wed, 24 Nov 2021 14:50:41 +0000 (17:50 +0300)]
can: sja1000: fix use after free in ems_pcmcia_add_card()

If the last channel is not available then "dev" is freed.  Fortunately,
we can just use "pdev->irq" instead.

Also we should check if at least one channel was set up.

Fixes: fd734c6f25ae ("can/sja1000: add driver for EMS PCMCIA card")
Link: https://lore.kernel.org/all/20211124145041.GB13656@kili
Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Oliver Hartkopp <socketcan@hartkopp.net>
Tested-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: pch_can: pch_can_rx_normal: fix use after free
Vincent Mailhol [Tue, 23 Nov 2021 11:16:54 +0000 (20:16 +0900)]
can: pch_can: pch_can_rx_normal: fix use after free

After calling netif_receive_skb(skb), dereferencing skb is unsafe.
Especially, the can_frame cf which aliases skb memory is dereferenced
just after the call netif_receive_skb(skb).

Reordering the lines solves the issue.

Fixes: b21d18b51b31 ("can: Topcliff: Add PCH_CAN driver.")
Link: https://lore.kernel.org/all/20211123111654.621610-1-mailhol.vincent@wanadoo.fr
Cc: stable@vger.kernel.org
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agodevlink: fix netns refcount leak in devlink_nl_cmd_reload()
Eric Dumazet [Sun, 5 Dec 2021 19:28:22 +0000 (11:28 -0800)]
devlink: fix netns refcount leak in devlink_nl_cmd_reload()

While preparing my patch series adding netns refcount tracking,
I spotted bugs in devlink_nl_cmd_reload()

Some error paths forgot to release a refcount on a netns.

To fix this, we can reduce the scope of get_net()/put_net()
section around the call to devlink_reload().

Fixes: ccdf07219da6 ("devlink: Add reload action option to devlink reload command")
Fixes: dc64cc7c6310 ("devlink: Add devlink reload limit option")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Moshe Shemesh <moshe@mellanox.com>
Cc: Jacob Keller <jacob.e.keller@intel.com>
Cc: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20211205192822.1741045-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoethtool: do not perform operations on net devices being unregistered
Antoine Tenart [Fri, 3 Dec 2021 10:13:18 +0000 (11:13 +0100)]
ethtool: do not perform operations on net devices being unregistered

There is a short period between a net device starts to be unregistered
and when it is actually gone. In that time frame ethtool operations
could still be performed, which might end up in unwanted or undefined
behaviours[1].

Do not allow ethtool operations after a net device starts its
unregistration. This patch targets the netlink part as the ioctl one
isn't affected: the reference to the net device is taken and the
operation is executed within an rtnl lock section and the net device
won't be found after unregister.

[1] For example adding Tx queues after unregister ends up in NULL
    pointer exceptions and UaFs, such as:

      BUG: KASAN: use-after-free in kobject_get+0x14/0x90
      Read of size 1 at addr ffff88801961248c by task ethtool/755

      CPU: 0 PID: 755 Comm: ethtool Not tainted 5.15.0-rc6+ #778
      Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-4.fc34 04/014
      Call Trace:
       dump_stack_lvl+0x57/0x72
       print_address_description.constprop.0+0x1f/0x140
       kasan_report.cold+0x7f/0x11b
       kobject_get+0x14/0x90
       kobject_add_internal+0x3d1/0x450
       kobject_init_and_add+0xba/0xf0
       netdev_queue_update_kobjects+0xcf/0x200
       netif_set_real_num_tx_queues+0xb4/0x310
       veth_set_channels+0x1c3/0x550
       ethnl_set_channels+0x524/0x610

Fixes: 041b1c5d4a53 ("ethtool: helper functions for netlink interface")
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Link: https://lore.kernel.org/r/20211203101318.435618-1-atenart@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoi40e: Fix NULL pointer dereference in i40e_dbg_dump_desc
Norbert Zulinski [Mon, 22 Nov 2021 11:29:05 +0000 (12:29 +0100)]
i40e: Fix NULL pointer dereference in i40e_dbg_dump_desc

When trying to dump VFs VSI RX/TX descriptors
using debugfs there was a crash
due to NULL pointer dereference in i40e_dbg_dump_desc.
Added a check to i40e_dbg_dump_desc that checks if
VSI type is correct for dumping RX/TX descriptors.

Fixes: 02e9c290814c ("i40e: debugfs interface")
Signed-off-by: Sylwester Dziedziuch <sylwesterx.dziedziuch@intel.com>
Signed-off-by: Norbert Zulinski <norbertx.zulinski@intel.com>
Signed-off-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Tested-by: Gurucharan G <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoi40e: Fix pre-set max number of queues for VF
Mateusz Palczewski [Fri, 16 Jul 2021 09:33:56 +0000 (11:33 +0200)]
i40e: Fix pre-set max number of queues for VF

After setting pre-set combined to 16 queues and reserving 16 queues by
tc qdisc, pre-set maximum combined queues returned to default value
after VF reset being 4 and this generated errors during removing tc.
Fixed by removing clear num_req_queues before reset VF.

Fixes: e284fc280473 (i40e: Add and delete cloud filter)
Signed-off-by: Grzegorz Szczurek <grzegorzx.szczurek@intel.com>
Signed-off-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Tested-by: Bindushree P <Bindushree.p@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoi40e: Fix failed opcode appearing if handling messages from VF
Karen Sornek [Fri, 14 May 2021 09:43:13 +0000 (11:43 +0200)]
i40e: Fix failed opcode appearing if handling messages from VF

Fix failed operation code appearing if handling messages from VF.
Implemented by waiting for VF appropriate state if request starts
handle while VF reset.
Without this patch the message handling request while VF is in
a reset state ends with error -5 (I40E_ERR_PARAM).

Fixes: 5c3c48ac6bf5 ("i40e: implement virtual device interface")
Signed-off-by: Grzegorz Szczurek <grzegorzx.szczurek@intel.com>
Signed-off-by: Karen Sornek <karen.sornek@intel.com>
Tested-by: Tony Brelinski <tony.brelinski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoiavf: Fix reporting when setting descriptor count
Michal Maloszewski [Tue, 26 Oct 2021 12:59:09 +0000 (12:59 +0000)]
iavf: Fix reporting when setting descriptor count

iavf_set_ringparams doesn't communicate to the user that

1. The user requested descriptor count is out of range. Instead it
   just quietly sets descriptors to the "clamped" value and calls it
   done. This makes it look an invalid value was successfully set as
   the descriptor count when this isn't actually true.

2. The user provided descriptor count needs to be inflated for alignment
   reasons.

This behavior is confusing. The ice driver has already addressed this
by rejecting invalid values for descriptor count and
messaging for alignment adjustments.
Do the same thing here by adding the error and info messages.

Fixes: fbb7ddfef253 ("i40evf: core ethtool functionality")
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Signed-off-by: Michal Maloszewski <michal.maloszewski@intel.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agonet: cdc_ncm: Allow for dwNtbOutMaxSize to be unset or zero
Lee Jones [Thu, 2 Dec 2021 14:34:37 +0000 (14:34 +0000)]
net: cdc_ncm: Allow for dwNtbOutMaxSize to be unset or zero

Currently, due to the sequential use of min_t() and clamp_t() macros,
in cdc_ncm_check_tx_max(), if dwNtbOutMaxSize is not set, the logic
sets tx_max to 0.  This is then used to allocate the data area of the
SKB requested later in cdc_ncm_fill_tx_frame().

This does not cause an issue presently because when memory is
allocated during initialisation phase of SKB creation, more memory
(512b) is allocated than is required for the SKB headers alone (320b),
leaving some space (512b - 320b = 192b) for CDC data (172b).

However, if more elements (for example 3 x u64 = [24b]) were added to
one of the SKB header structs, say 'struct skb_shared_info',
increasing its original size (320b [320b aligned]) to something larger
(344b [384b aligned]), then suddenly the CDC data (172b) no longer
fits in the spare SKB data area (512b - 384b = 128b).

Consequently the SKB bounds checking semantics fails and panics:

  skbuff: skb_over_panic: text:ffffffff830a5b5f len:184 put:172   \
     head:ffff888119227c00 data:ffff888119227c00 tail:0xb8 end:0x80 dev:<NULL>

  ------------[ cut here ]------------
  kernel BUG at net/core/skbuff.c:110!
  RIP: 0010:skb_panic+0x14f/0x160 net/core/skbuff.c:106
  <snip>
  Call Trace:
   <IRQ>
   skb_over_panic+0x2c/0x30 net/core/skbuff.c:115
   skb_put+0x205/0x210 net/core/skbuff.c:1877
   skb_put_zero include/linux/skbuff.h:2270 [inline]
   cdc_ncm_ndp16 drivers/net/usb/cdc_ncm.c:1116 [inline]
   cdc_ncm_fill_tx_frame+0x127f/0x3d50 drivers/net/usb/cdc_ncm.c:1293
   cdc_ncm_tx_fixup+0x98/0xf0 drivers/net/usb/cdc_ncm.c:1514

By overriding the max value with the default CDC_NCM_NTB_MAX_SIZE_TX
when not offered through the system provided params, we ensure enough
data space is allocated to handle the CDC data, meaning no crash will
occur.

Cc: Oliver Neukum <oliver@neukum.org>
Fixes: 289507d3364f9 ("net: cdc_ncm: use sysfs for rx/tx aggregation tuning")
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Link: https://lore.kernel.org/r/20211202143437.1411410-1-lee.jones@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoqede: validate non LSO skb length
Manish Chopra [Fri, 3 Dec 2021 17:44:13 +0000 (09:44 -0800)]
qede: validate non LSO skb length

Although it is unlikely that stack could transmit a non LSO
skb with length > MTU, however in some cases or environment such
occurrences actually resulted into firmware asserts due to packet
length being greater than the max supported by the device (~9700B).

This patch adds the safeguard for such odd cases to avoid firmware
asserts.

v2: Added "Fixes" tag with one of the initial driver commit
    which enabled the TX traffic actually (as this was probably
    day1 issue which was discovered recently by some customer
    environment)

Fixes: a2ec6172d29c ("qede: Add support for link")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Alok Prasad <palok@marvell.com>
Signed-off-by: Prabhakar Kushwaha <pkushwaha@marvell.com>
Signed-off-by: Ariel Elior <aelior@marvell.com>
Link: https://lore.kernel.org/r/20211203174413.13090-1-manishc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agobpf: Fix the off-by-two error in range markings
Maxim Mikityanskiy [Tue, 30 Nov 2021 18:16:07 +0000 (20:16 +0200)]
bpf: Fix the off-by-two error in range markings

The first commit cited below attempts to fix the off-by-one error that
appeared in some comparisons with an open range. Due to this error,
arithmetically equivalent pieces of code could get different verdicts
from the verifier, for example (pseudocode):

  // 1. Passes the verifier:
  if (data + 8 > data_end)
      return early
  read *(u64 *)data, i.e. [data; data+7]

  // 2. Rejected by the verifier (should still pass):
  if (data + 7 >= data_end)
      return early
  read *(u64 *)data, i.e. [data; data+7]

The attempted fix, however, shifts the range by one in a wrong
direction, so the bug not only remains, but also such piece of code
starts failing in the verifier:

  // 3. Rejected by the verifier, but the check is stricter than in #1.
  if (data + 8 >= data_end)
      return early
  read *(u64 *)data, i.e. [data; data+7]

The change performed by that fix converted an off-by-one bug into
off-by-two. The second commit cited below added the BPF selftests
written to ensure than code chunks like #3 are rejected, however,
they should be accepted.

This commit fixes the off-by-two error by adjusting new_range in the
right direction and fixes the tests by changing the range into the
one that should actually fail.

Fixes: fb2a311a31d3 ("bpf: fix off by one for range markings with L{T, E} patterns")
Fixes: b37242c773b2 ("bpf: add test cases to bpf selftests to cover all access tests")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211130181607.593149-1-maximmi@nvidia.com
2 years agotreewide: Add missing includes masked by cgroup -> bpf dependency
Jakub Kicinski [Thu, 2 Dec 2021 20:34:00 +0000 (12:34 -0800)]
treewide: Add missing includes masked by cgroup -> bpf dependency

cgroup.h (therefore swap.h, therefore half of the universe)
includes bpf.h which in turn includes module.h and slab.h.
Since we're about to get rid of that dependency we need
to clean things up.

v2: drop the cpu.h include from cacheinfo.h, it's not necessary
and it makes riscv sensitive to ordering of include files.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Krzysztof Wilczyński <kw@linux.com>
Acked-by: Peter Chen <peter.chen@kernel.org>
Acked-by: SeongJae Park <sj@kernel.org>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/all/20211120035253.72074-1-kuba@kernel.org/
Link: https://lore.kernel.org/all/20211120165528.197359-1-kuba@kernel.org/
Link: https://lore.kernel.org/bpf/20211202203400.1208663-1-kuba@kernel.org
2 years agonet: altera: set a couple error code in probe()
Dan Carpenter [Fri, 3 Dec 2021 10:11:28 +0000 (13:11 +0300)]
net: altera: set a couple error code in probe()

There are two error paths which accidentally return success instead of
a negative error code.

Fixes: bbd2190ce96d ("Altera TSE: Add main and header file for Altera Ethernet Driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: bcm4908: Handle dma_set_coherent_mask error codes
Jiasheng Jiang [Fri, 3 Dec 2021 03:31:06 +0000 (11:31 +0800)]
net: bcm4908: Handle dma_set_coherent_mask error codes

The return value of dma_set_coherent_mask() is not always 0.
To catch the exception in case that dma is not support the mask.

Fixes: 9d61d138ab30 ("net: broadcom: rename BCM4908 driver & update DT binding")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests: net/fcnal-test.sh: add exit code
Li Zhijian [Fri, 3 Dec 2021 02:32:13 +0000 (10:32 +0800)]
selftests: net/fcnal-test.sh: add exit code

Previously, the selftest framework always treats it as *ok* even though
some of them are failed actually. That's because the script always
returns 0.

It supports PASS/FAIL/SKIP exit code now.

CC: Philip Li <philip.li@intel.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobonding: make tx_rebalance_counter an atomic
Eric Dumazet [Fri, 3 Dec 2021 02:27:18 +0000 (18:27 -0800)]
bonding: make tx_rebalance_counter an atomic

KCSAN reported a data-race [1] around tx_rebalance_counter
which can be accessed from different contexts, without
the protection of a lock/mutex.

[1]
BUG: KCSAN: data-race in bond_alb_init_slave / bond_alb_monitor

write to 0xffff888157e8ca24 of 4 bytes by task 7075 on cpu 0:
 bond_alb_init_slave+0x713/0x860 drivers/net/bonding/bond_alb.c:1613
 bond_enslave+0xd94/0x3010 drivers/net/bonding/bond_main.c:1949
 do_set_master net/core/rtnetlink.c:2521 [inline]
 __rtnl_newlink net/core/rtnetlink.c:3475 [inline]
 rtnl_newlink+0x1298/0x13b0 net/core/rtnetlink.c:3506
 rtnetlink_rcv_msg+0x745/0x7e0 net/core/rtnetlink.c:5571
 netlink_rcv_skb+0x14e/0x250 net/netlink/af_netlink.c:2491
 rtnetlink_rcv+0x18/0x20 net/core/rtnetlink.c:5589
 netlink_unicast_kernel net/netlink/af_netlink.c:1319 [inline]
 netlink_unicast+0x5fc/0x6c0 net/netlink/af_netlink.c:1345
 netlink_sendmsg+0x6e1/0x7d0 net/netlink/af_netlink.c:1916
 sock_sendmsg_nosec net/socket.c:704 [inline]
 sock_sendmsg net/socket.c:724 [inline]
 ____sys_sendmsg+0x39a/0x510 net/socket.c:2409
 ___sys_sendmsg net/socket.c:2463 [inline]
 __sys_sendmsg+0x195/0x230 net/socket.c:2492
 __do_sys_sendmsg net/socket.c:2501 [inline]
 __se_sys_sendmsg net/socket.c:2499 [inline]
 __x64_sys_sendmsg+0x42/0x50 net/socket.c:2499
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x44/0xd0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

read to 0xffff888157e8ca24 of 4 bytes by task 1082 on cpu 1:
 bond_alb_monitor+0x8f/0xc00 drivers/net/bonding/bond_alb.c:1511
 process_one_work+0x3fc/0x980 kernel/workqueue.c:2298
 worker_thread+0x616/0xa70 kernel/workqueue.c:2445
 kthread+0x2c7/0x2e0 kernel/kthread.c:327
 ret_from_fork+0x1f/0x30

value changed: 0x00000001 -> 0x00000064

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 1082 Comm: kworker/u4:3 Not tainted 5.16.0-rc3-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: bond1 bond_alb_monitor

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agotcp: fix another uninit-value (sk_rx_queue_mapping)
Eric Dumazet [Thu, 2 Dec 2021 23:37:24 +0000 (15:37 -0800)]
tcp: fix another uninit-value (sk_rx_queue_mapping)

KMSAN is still not happy [1].

I missed that passive connections do not inherit their
sk_rx_queue_mapping values from the request socket,
but instead tcp_child_process() is calling
sk_mark_napi_id(child, skb)

We have many sk_mark_napi_id() callers, so I am providing
a new helper, forcing the setting sk_rx_queue_mapping
and sk_napi_id.

Note that we had no KMSAN report for sk_napi_id because
passive connections got a copy of this field from the listener.
sk_rx_queue_mapping in the other hand is inside the
sk_dontcopy_begin/sk_dontcopy_end so sk_clone_lock()
leaves this field uninitialized.

We might remove dead code populating req->sk_rx_queue_mapping
in the future.

[1]

BUG: KMSAN: uninit-value in __sk_rx_queue_set include/net/sock.h:1924 [inline]
BUG: KMSAN: uninit-value in sk_rx_queue_update include/net/sock.h:1938 [inline]
BUG: KMSAN: uninit-value in sk_mark_napi_id include/net/busy_poll.h:136 [inline]
BUG: KMSAN: uninit-value in tcp_child_process+0xb42/0x1050 net/ipv4/tcp_minisocks.c:833
 __sk_rx_queue_set include/net/sock.h:1924 [inline]
 sk_rx_queue_update include/net/sock.h:1938 [inline]
 sk_mark_napi_id include/net/busy_poll.h:136 [inline]
 tcp_child_process+0xb42/0x1050 net/ipv4/tcp_minisocks.c:833
 tcp_v4_rcv+0x3d83/0x4ed0 net/ipv4/tcp_ipv4.c:2066
 ip_protocol_deliver_rcu+0x760/0x10b0 net/ipv4/ip_input.c:204
 ip_local_deliver_finish net/ipv4/ip_input.c:231 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ip_local_deliver+0x584/0x8c0 net/ipv4/ip_input.c:252
 dst_input include/net/dst.h:460 [inline]
 ip_sublist_rcv_finish net/ipv4/ip_input.c:551 [inline]
 ip_list_rcv_finish net/ipv4/ip_input.c:601 [inline]
 ip_sublist_rcv+0x11fd/0x1520 net/ipv4/ip_input.c:609
 ip_list_rcv+0x95f/0x9a0 net/ipv4/ip_input.c:644
 __netif_receive_skb_list_ptype net/core/dev.c:5505 [inline]
 __netif_receive_skb_list_core+0xe34/0x1240 net/core/dev.c:5553
 __netif_receive_skb_list+0x7fc/0x960 net/core/dev.c:5605
 netif_receive_skb_list_internal+0x868/0xde0 net/core/dev.c:5696
 gro_normal_list net/core/dev.c:5850 [inline]
 napi_complete_done+0x579/0xdd0 net/core/dev.c:6587
 virtqueue_napi_complete drivers/net/virtio_net.c:339 [inline]
 virtnet_poll+0x17b6/0x2350 drivers/net/virtio_net.c:1557
 __napi_poll+0x14e/0xbc0 net/core/dev.c:7020
 napi_poll net/core/dev.c:7087 [inline]
 net_rx_action+0x824/0x1880 net/core/dev.c:7174
 __do_softirq+0x1fe/0x7eb kernel/softirq.c:558
 run_ksoftirqd+0x33/0x50 kernel/softirq.c:920
 smpboot_thread_fn+0x616/0xbf0 kernel/smpboot.c:164
 kthread+0x721/0x850 kernel/kthread.c:327
 ret_from_fork+0x1f/0x30

Uninit was created at:
 __alloc_pages+0xbc7/0x10a0 mm/page_alloc.c:5409
 alloc_pages+0x8a5/0xb80
 alloc_slab_page mm/slub.c:1810 [inline]
 allocate_slab+0x287/0x1c20 mm/slub.c:1947
 new_slab mm/slub.c:2010 [inline]
 ___slab_alloc+0xbdf/0x1e90 mm/slub.c:3039
 __slab_alloc mm/slub.c:3126 [inline]
 slab_alloc_node mm/slub.c:3217 [inline]
 slab_alloc mm/slub.c:3259 [inline]
 kmem_cache_alloc+0xbb3/0x11c0 mm/slub.c:3264
 sk_prot_alloc+0xeb/0x570 net/core/sock.c:1914
 sk_clone_lock+0xd6/0x1940 net/core/sock.c:2118
 inet_csk_clone_lock+0x8d/0x6a0 net/ipv4/inet_connection_sock.c:956
 tcp_create_openreq_child+0xb1/0x1ef0 net/ipv4/tcp_minisocks.c:453
 tcp_v4_syn_recv_sock+0x268/0x2710 net/ipv4/tcp_ipv4.c:1563
 tcp_check_req+0x207c/0x2a30 net/ipv4/tcp_minisocks.c:765
 tcp_v4_rcv+0x36f5/0x4ed0 net/ipv4/tcp_ipv4.c:2047
 ip_protocol_deliver_rcu+0x760/0x10b0 net/ipv4/ip_input.c:204
 ip_local_deliver_finish net/ipv4/ip_input.c:231 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ip_local_deliver+0x584/0x8c0 net/ipv4/ip_input.c:252
 dst_input include/net/dst.h:460 [inline]
 ip_sublist_rcv_finish net/ipv4/ip_input.c:551 [inline]
 ip_list_rcv_finish net/ipv4/ip_input.c:601 [inline]
 ip_sublist_rcv+0x11fd/0x1520 net/ipv4/ip_input.c:609
 ip_list_rcv+0x95f/0x9a0 net/ipv4/ip_input.c:644
 __netif_receive_skb_list_ptype net/core/dev.c:5505 [inline]
 __netif_receive_skb_list_core+0xe34/0x1240 net/core/dev.c:5553
 __netif_receive_skb_list+0x7fc/0x960 net/core/dev.c:5605
 netif_receive_skb_list_internal+0x868/0xde0 net/core/dev.c:5696
 gro_normal_list net/core/dev.c:5850 [inline]
 napi_complete_done+0x579/0xdd0 net/core/dev.c:6587
 virtqueue_napi_complete drivers/net/virtio_net.c:339 [inline]
 virtnet_poll+0x17b6/0x2350 drivers/net/virtio_net.c:1557
 __napi_poll+0x14e/0xbc0 net/core/dev.c:7020
 napi_poll net/core/dev.c:7087 [inline]
 net_rx_action+0x824/0x1880 net/core/dev.c:7174
 __do_softirq+0x1fe/0x7eb kernel/softirq.c:558

Fixes: 342159ee394d ("net: avoid dirtying sk->sk_rx_queue_mapping")
Fixes: a37a0ee4d25c ("net: avoid uninit-value from tcp_conn_request")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Tested-by: Alexander Potapenko <glider@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoinet: use #ifdef CONFIG_SOCK_RX_QUEUE_MAPPING consistently
Eric Dumazet [Thu, 2 Dec 2021 22:42:18 +0000 (14:42 -0800)]
inet: use #ifdef CONFIG_SOCK_RX_QUEUE_MAPPING consistently

Since commit 4e1beecc3b58 ("net/sock: Add kernel config
SOCK_RX_QUEUE_MAPPING"),
sk_rx_queue_mapping access is guarded by CONFIG_SOCK_RX_QUEUE_MAPPING.

Fixes: 54b92e841937 ("tcp: Migrate TCP_ESTABLISHED/TCP_SYN_RECV sockets in accept queues.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Tariq Toukan <tariqt@nvidia.com>
Acked-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests/tc-testing: Fix cannot create /sys/bus/netdevsim/new_device: Directory...
Li Zhijian [Fri, 3 Dec 2021 02:53:23 +0000 (10:53 +0800)]
selftests/tc-testing: Fix cannot create /sys/bus/netdevsim/new_device: Directory nonexistent

Install netdevsim to provide /sys/bus/netdevsim/new_device interface.

It helps to fix:
 # ok 97 9a7d - Change ETS strict band without quantum # skipped - skipped - previous setup failed 11 ce7d
 #
 #
 # -----> prepare stage *** Could not execute: "echo "1 1 4" > /sys/bus/netdevsim/new_device"
 #
 # -----> prepare stage *** Error message: "/bin/sh: 1: cannot create /sys/bus/netdevsim/new_device: Directory nonexistent
 # "
 #
 # -----> prepare stage *** Aborting test run.
 #
 #
 # <_io.BufferedReader name=5> *** stdout ***
 #

Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests/tc-testing: add missing config
Li Zhijian [Fri, 3 Dec 2021 02:53:22 +0000 (10:53 +0800)]
selftests/tc-testing: add missing config

qdiscs/fq_pie requires CONFIG_NET_SCH_FQ_PIE, otherwise tc will fail
to create a fq_pie qdisc.

It fixes following issue:
 # not ok 57 83be - Create FQ-PIE with invalid number of flows
 #       Command exited with 2, expected 0
 # Error: Specified qdisc not found.

Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests/tc-testing: add exit code
Li Zhijian [Fri, 3 Dec 2021 02:53:21 +0000 (10:53 +0800)]
selftests/tc-testing: add exit code

Mark the summary result as FAIL to prevent from confusing the selftest
framework if some of them are failed.

Previously, the selftest framework always treats it as *ok* even though
some of them are failed actually. That's because the script tdc.sh always
return 0.

 # All test results:
 #
 # 1..97
 # ok 1 83be - Create FQ-PIE with invalid number of flows
 # ok 2 8b6e - Create RED with no flags
[...snip]
 # ok 6 5f15 - Create RED with flags ECN, harddrop
 # ok 7 53e8 - Create RED with flags ECN, nodrop
 # ok 8 d091 - Fail to create RED with only nodrop flag
 # ok 9 af8e - Create RED with flags ECN, nodrop, harddrop
 # not ok 10 ce7d - Add mq Qdisc to multi-queue device (4 queues)
 #       Could not match regex pattern. Verify command output:
 # qdisc mq 1: root
 # qdisc fq_codel 0: parent 1:4 limit 10240p flows 1024 quantum 1514 target 5ms interval 100ms memory_limit 32Mb ecn drop_batch 64
 # qdisc fq_codel 0: parent 1:3 limit 10240p flows 1024 quantum 1514 target 5ms interval 100ms memory_limit 32Mb ecn drop_batch 64
[...snip]
 # ok 96 6979 - Change quantum of a strict ETS band
 # ok 97 9a7d - Change ETS strict band without quantum
 #
 #
 #
 #
 ok 1 selftests: tc-testing: tdc.sh <<< summary result

CC: Philip Li <philip.li@intel.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Acked-by: Davide Caratti <dcaratti@redhat.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests/fib_tests: Rework fib_rp_filter_test()
Peilin Ye [Wed, 1 Dec 2021 00:47:20 +0000 (16:47 -0800)]
selftests/fib_tests: Rework fib_rp_filter_test()

Currently rp_filter tests in fib_tests.sh:fib_rp_filter_test() are
failing.  ping sockets are bound to dummy1 using the "-I" option
(SO_BINDTODEVICE), but socket lookup is failing when receiving ping
replies, since the routing table thinks they belong to dummy0.

For example, suppose ping is using a SOCK_RAW socket for ICMP messages.
When receiving ping replies, in __raw_v4_lookup(), sk->sk_bound_dev_if
is 3 (dummy1), but dif (skb_rtable(skb)->rt_iif) says 2 (dummy0), so the
raw_sk_bound_dev_eq() check fails.  Similar things happen in
ping_lookup() for SOCK_DGRAM sockets.

These tests used to pass due to a bug [1] in iputils, where "ping -I"
actually did not bind ICMP message sockets to device.  The bug has been
fixed by iputils commit f455fee41c07 ("ping: also bind the ICMP socket
to the specific device") in 2016, which is why our rp_filter tests
started to fail.  See [2] .

Fixing the tests while keeping everything in one netns turns out to be
nontrivial.  Rework the tests and build the following topology:

 ┌─────────────────────────────┐    ┌─────────────────────────────┐
 │  network namespace 1 (ns1)  │    │  network namespace 2 (ns2)  │
 │                             │    │                             │
 │  ┌────┐     ┌─────┐         │    │  ┌─────┐            ┌────┐  │
 │  │ lo │<───>│veth1│<────────┼────┼─>│veth2│<──────────>│ lo │  │
 │  └────┘     ├─────┴──────┐  │    │  ├─────┴──────┐     └────┘  │
 │             │192.0.2.1/24│  │    │  │192.0.2.1/24│             │
 │             └────────────┘  │    │  └────────────┘             │
 └─────────────────────────────┘    └─────────────────────────────┘

Consider sending an ICMP_ECHO packet A in ns2.  Both source and
destination IP addresses are 192.0.2.1, and we use strict mode rp_filter
in both ns1 and ns2:

  1. A is routed to lo since its destination IP address is one of ns2's
     local addresses (veth2);
  2. A is redirected from lo's egress to veth2's egress using mirred;
  3. A arrives at veth1's ingress in ns1;
  4. A is redirected from veth1's ingress to lo's ingress, again, using
     mirred;
  5. In __fib_validate_source(), fib_info_nh_uses_dev() returns false,
     since A was received on lo, but reverse path lookup says veth1;
  6. However A is not dropped since we have relaxed this check for lo in
     commit 66f8209547cc ("fib: relax source validation check for loopback
     packets");

Making sure A is not dropped here in this corner case is the whole point
of having this test.

  7. As A reaches the ICMP layer, an ICMP_ECHOREPLY packet, B, is
     generated;
  8. Similarly, B is redirected from lo's egress to veth1's egress (in
     ns1), then redirected once again from veth2's ingress to lo's
     ingress (in ns2), using mirred.

Also test "ping 127.0.0.1" from ns2.  It does not trigger the relaxed
check in __fib_validate_source(), but just to make sure the topology
works with loopback addresses.

Tested with ping from iputils 20210722-41-gf9fb573:

$ ./fib_tests.sh -t rp_filter

IPv4 rp_filter tests
    TEST: rp_filter passes local packets [ OK ]
    TEST: rp_filter passes loopback packets [ OK ]

[1] https://github.com/iputils/iputils/issues/55
[2] https://github.com/iputils/iputils/commit/f455fee41c077d4b700a473b2f5b3487b8febc1d

Reported-by: Hangbin Liu <liuhangbin@gmail.com>
Fixes: adb701d6cfa4 ("selftests: add a test case for rp_filter")
Reviewed-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Peilin Ye <peilin.ye@bytedance.com>
Acked-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20211201004720.6357-1-yepeilin.cs@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch 'Fixes for kfunc-mod regressions and warnings'
Andrii Nakryiko [Thu, 2 Dec 2021 21:39:47 +0000 (13:39 -0800)]
Merge branch 'Fixes for kfunc-mod regressions and warnings'

Kumar Kartikeya says:

====================

This set includes fixes for two regressions and one build warning introduced by
the kfunc for modules series.

Changelog:
----------

v1 -> v2:
v1: https://lore.kernel.org/bpf/20211115191840.496263-1-memxor@gmail.com

 * Instead of demoting resolve_btfids warning to debug, only skip in case of
   set->cnt == 0.
====================

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
2 years agotools/resolve_btfids: Skip unresolved symbol warning for empty BTF sets
Kumar Kartikeya Dwivedi [Mon, 22 Nov 2021 14:47:42 +0000 (20:17 +0530)]
tools/resolve_btfids: Skip unresolved symbol warning for empty BTF sets

resolve_btfids prints a warning when it finds an unresolved symbol,
(id == 0) in id_patch. This can be the case for BTF sets that are empty
(due to disabled config options), hence printing warnings for certain
builds, most recently seen in [0].

The reason behind this is because id->cnt aliases id->id in btf_id
struct, leading to empty set showing up as ID 0 when we get to id_patch,
which triggers the warning. Since sets are an exception here, accomodate
by reusing hole in btf_id for bool is_set member, setting it to true for
BTF set when setting id->cnt, and use that to skip extraneous warning.

  [0]: https://lore.kernel.org/all/1b99ae14-abb4-d18f-cc6a-d7e523b25542@gmail.com

Before:

; ./tools/bpf/resolve_btfids/resolve_btfids -v -b vmlinux net/ipv4/tcp_cubic.ko
adding symbol tcp_cubic_kfunc_ids
WARN: resolve_btfids: unresolved symbol tcp_cubic_kfunc_ids
patching addr     0: ID       0 [tcp_cubic_kfunc_ids]
sorting  addr     4: cnt      0 [tcp_cubic_kfunc_ids]
update ok for net/ipv4/tcp_cubic.ko

After:

; ./tools/bpf/resolve_btfids/resolve_btfids -v -b vmlinux net/ipv4/tcp_cubic.ko
adding symbol tcp_cubic_kfunc_ids
patching addr     0: ID       0 [tcp_cubic_kfunc_ids]
sorting  addr     4: cnt      0 [tcp_cubic_kfunc_ids]
update ok for net/ipv4/tcp_cubic.ko

Fixes: 0e32dfc80bae ("bpf: Enable TCP congestion control kfunc from modules")
Reported-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20211122144742.477787-4-memxor@gmail.com
2 years agobpf: Fix bpf_check_mod_kfunc_call for built-in modules
Kumar Kartikeya Dwivedi [Mon, 22 Nov 2021 14:47:41 +0000 (20:17 +0530)]
bpf: Fix bpf_check_mod_kfunc_call for built-in modules

When module registering its set is built-in, THIS_MODULE will be NULL,
hence we cannot return early in case owner is NULL.

Fixes: 14f267d95fe4 ("bpf: btf: Introduce helpers for dynamic BTF set registration")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20211122144742.477787-3-memxor@gmail.com
2 years agobpf: Make CONFIG_DEBUG_INFO_BTF depend upon CONFIG_BPF_SYSCALL
Kumar Kartikeya Dwivedi [Mon, 22 Nov 2021 14:47:40 +0000 (20:17 +0530)]
bpf: Make CONFIG_DEBUG_INFO_BTF depend upon CONFIG_BPF_SYSCALL

Vinicius Costa Gomes reported [0] that build fails when
CONFIG_DEBUG_INFO_BTF is enabled and CONFIG_BPF_SYSCALL is disabled.
This leads to btf.c not being compiled, and then no symbol being present
in vmlinux for the declarations in btf.h. Since BTF is not useful
without enabling BPF subsystem, disallow this combination.

However, theoretically disabling both now could still fail, as the
symbol for kfunc_btf_id_list variables is not available. This isn't a
problem as the compiler usually optimizes the whole register/unregister
call, but at lower optimization levels it can fail the build in linking
stage.

Fix that by adding dummy variables so that modules taking address of
them still work, but the whole thing is a noop.

  [0]: https://lore.kernel.org/bpf/20211110205418.332403-1-vinicius.gomes@intel.com

Fixes: 14f267d95fe4 ("bpf: btf: Introduce helpers for dynamic BTF set registration")
Reported-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20211122144742.477787-2-memxor@gmail.com
2 years agoMerge tag 'net-5.16-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 2 Dec 2021 19:22:06 +0000 (11:22 -0800)]
Merge tag 'net-5.16-rc4' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from wireless, and wireguard.

  Mostly scattered driver changes this week, with one big clump in
  mv88e6xxx. Nothing of note, really.

  Current release - regressions:

   - smc: keep smc_close_final()'s error code during active close

  Current release - new code bugs:

   - iwlwifi: various static checker fixes (int overflow, leaks, missing
     error codes)

   - rtw89: fix size of firmware header before transfer, avoid crash

   - mt76: fix timestamp check in tx_status; fix pktid leak;

   - mscc: ocelot: fix missing unlock on error in ocelot_hwstamp_set()

  Previous releases - regressions:

   - smc: fix list corruption in smc_lgr_cleanup_early

   - ipv4: convert fib_num_tclassid_users to atomic_t

  Previous releases - always broken:

   - tls: fix authentication failure in CCM mode

   - vrf: reset IPCB/IP6CB when processing outbound pkts, prevent
     incorrect processing

   - dsa: mv88e6xxx: fixes for various device errata

   - rds: correct socket tunable error in rds_tcp_tune()

   - ipv6: fix memory leak in fib6_rule_suppress

   - wireguard: reset peer src endpoint when netns exits

   - wireguard: improve resilience to DoS around incoming handshakes

   - tcp: fix page frag corruption on page fault which involves TCP

   - mpls: fix missing attributes in delete notifications

   - mt7915: fix NULL pointer dereference with ad-hoc mode

  Misc:

   - rt2x00: be more lenient about EPROTO errors during start

   - mlx4_en: update reported link modes for 1/10G"

* tag 'net-5.16-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (85 commits)
  net: dsa: b53: Add SPI ID table
  gro: Fix inconsistent indenting
  selftests: net: Correct case name
  net/rds: correct socket tunable error in rds_tcp_tune()
  mctp: Don't let RTM_DELROUTE delete local routes
  net/smc: Keep smc_close_final rc during active close
  ibmvnic: drop bad optimization in reuse_tx_pools()
  ibmvnic: drop bad optimization in reuse_rx_pools()
  net/smc: fix wrong list_del in smc_lgr_cleanup_early
  Fix Comment of ETH_P_802_3_MIN
  ethernet: aquantia: Try MAC address from device tree
  ipv4: convert fib_num_tclassid_users to atomic_t
  net: avoid uninit-value from tcp_conn_request
  net: annotate data-races on txq->xmit_lock_owner
  octeontx2-af: Fix a memleak bug in rvu_mbox_init()
  net/mlx4_en: Fix an use-after-free bug in mlx4_en_try_alloc_resources()
  vrf: Reset IPCB/IP6CB when processing outbound pkts in vrf dev xmit
  net: qlogic: qlcnic: Fix a NULL pointer dereference in qlcnic_83xx_add_rings()
  net: dsa: mv88e6xxx: Link in pcs_get_state() if AN is bypassed
  net: dsa: mv88e6xxx: Fix inband AN for 2500base-x on 88E6393X family
  ...

2 years agoMerge tag 'trace-v5.16-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt...
Linus Torvalds [Thu, 2 Dec 2021 19:07:41 +0000 (11:07 -0800)]
Merge tag 'trace-v5.16-rc3' of git://git./linux/kernel/git/rostedt/linux-trace

Pull tracing fixes from Steven Rostedt:
 "Three tracing fixes:

   - Allow compares of strings when using signed and unsigned characters

   - Fix kmemleak false positive for histogram entries

   - Handle negative numbers for user defined kretprobe data sizes"

* tag 'trace-v5.16-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  kprobes: Limit max data_size of the kretprobe instances
  tracing: Fix a kmemleak false positive in tracing_map
  tracing/histograms: String compares should not care about signed values

2 years agoMerge tag 'for-linus-5.16-2' of git://github.com/cminyard/linux-ipmi
Linus Torvalds [Thu, 2 Dec 2021 18:56:16 +0000 (10:56 -0800)]
Merge tag 'for-linus-5.16-2' of git://github.com/cminyard/linux-ipmi

Pull IPMI fixes from Corey Minyard:
 "Some changes that went in 5.16 had issues. When working on the design
  a piece was redesigned and things got missed. And the message type was
  not being initialized when it was allocated, resulting in crashes.

  In addition, the IPMI driver has had a shutdown issue where it could
  still have an item in a system workqueue after it had been shutdown.
  Move to a private workqueue to avoid that problem"

* tag 'for-linus-5.16-2' of git://github.com/cminyard/linux-ipmi:
  ipmi:ipmb: Fix unknown command response
  ipmi: fix IPMI_SMI_MSG_TYPE_IPMB_DIRECT response length checking
  ipmi: fix oob access due to uninit smi_msg type
  ipmi: msghandler: Make symbol 'remove_work_wq' static
  ipmi: Move remove_work to dedicated workqueue

2 years agonet: dsa: b53: Add SPI ID table
Florian Fainelli [Thu, 2 Dec 2021 04:17:20 +0000 (20:17 -0800)]
net: dsa: b53: Add SPI ID table

Currently autoloading for SPI devices does not use the DT ID table, it
uses SPI modalises. Supporting OF modalises is going to be difficult if
not impractical, an attempt was made but has been reverted, so ensure
that module autoloading works for this driver by adding an id_table
listing the SPI IDs for everything.

Fixes: 96c8395e2166 ("spi: Revert modalias changes")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agogro: Fix inconsistent indenting
Jiapeng Chong [Thu, 2 Dec 2021 08:30:42 +0000 (16:30 +0800)]
gro: Fix inconsistent indenting

Eliminate the follow smatch warning:

net/ipv6/ip6_offload.c:249 ipv6_gro_receive() warn: inconsistent
indenting.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests: net: Correct case name
Li Zhijian [Thu, 2 Dec 2021 02:28:41 +0000 (10:28 +0800)]
selftests: net: Correct case name

ipv6_addr_bind/ipv4_addr_bind are function names. Previously, bind test
would not be run by default due to the wrong case names

Fixes: 34d0302ab861 ("selftests: Add ipv6 address bind tests to fcnal-test")
Fixes: 75b2b2b3db4c ("selftests: Add ipv4 address bind tests to fcnal-test")
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/rds: correct socket tunable error in rds_tcp_tune()
William Kucharski [Wed, 1 Dec 2021 14:45:22 +0000 (07:45 -0700)]
net/rds: correct socket tunable error in rds_tcp_tune()

Correct an error where setting /proc/sys/net/rds/tcp/rds_tcp_rcvbuf would
instead modify the socket's sk_sndbuf and would leave sk_rcvbuf untouched.

Fixes: c6a58ffed536 ("RDS: TCP: Add sysctl tunables for sndbuf/rcvbuf on rds-tcp socket")
Signed-off-by: William Kucharski <william.kucharski@oracle.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agomctp: Don't let RTM_DELROUTE delete local routes
Matt Johnston [Wed, 1 Dec 2021 08:07:42 +0000 (16:07 +0800)]
mctp: Don't let RTM_DELROUTE delete local routes

We need to test against the existing route type, not
the rtm_type in the netlink request.

Fixes: 83f0a0b7285b ("mctp: Specify route types, require rtm_type in RTM_*ROUTE messages")
Signed-off-by: Matt Johnston <matt@codeconstruct.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/smc: Keep smc_close_final rc during active close
Tony Lu [Wed, 1 Dec 2021 06:42:16 +0000 (14:42 +0800)]
net/smc: Keep smc_close_final rc during active close

When smc_close_final() returns error, the return code overwrites by
kernel_sock_shutdown() in smc_close_active(). The return code of
smc_close_final() is more important than kernel_sock_shutdown(), and it
will pass to userspace directly.

Fix it by keeping both return codes, if smc_close_final() raises an
error, return it or kernel_sock_shutdown()'s.

Link: https://lore.kernel.org/linux-s390/1f67548e-cbf6-0dce-82b5-10288a4583bd@linux.ibm.com/
Fixes: 606a63c9783a ("net/smc: Ensure the active closing peer first closes clcsock")
Suggested-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: Tony Lu <tonylu@linux.alibaba.com>
Reviewed-by: Wen Gu <guwen@linux.alibaba.com>
Acked-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoibmvnic: drop bad optimization in reuse_tx_pools()
Sukadev Bhattiprolu [Wed, 1 Dec 2021 05:48:36 +0000 (21:48 -0800)]
ibmvnic: drop bad optimization in reuse_tx_pools()

When trying to decide whether or not reuse existing rx/tx pools
we tried to allow a range of values for the pool parameters rather
than exact matches. This was intended to reuse the resources for
instance when switching between two VIO servers with different
default parameters.

But this optimization is incomplete and breaks when we try to
change the number of queues for instance. The optimization needs
to be updated, so drop it for now and simplify the code.

Fixes: bbd809305bc7 ("ibmvnic: Reuse tx pools when possible")
Reported-by: Dany Madden <drt@linux.ibm.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com>
Reviewed-by: Dany Madden <drt@linux.ibm.com>
Reviewed-by: Rick Lindsley <ricklind@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoibmvnic: drop bad optimization in reuse_rx_pools()
Sukadev Bhattiprolu [Wed, 1 Dec 2021 05:48:35 +0000 (21:48 -0800)]
ibmvnic: drop bad optimization in reuse_rx_pools()

When trying to decide whether or not reuse existing rx/tx pools
we tried to allow a range of values for the pool parameters rather
than exact matches. This was intended to reuse the resources for
instance when switching between two VIO servers with different
default parameters.

But this optimization is incomplete and breaks when we try to
change the number of queues for instance. The optimization needs
to be updated, so drop it for now and simplify the code.

Fixes: 489de956e7a2 ("ibmvnic: Reuse rx pools when possible")
Reported-by: Dany Madden <drt@linux.ibm.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com>
Reviewed-by: Dany Madden <drt@linux.ibm.com>
Reviewed-by: Rick Lindsley <ricklind@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/smc: fix wrong list_del in smc_lgr_cleanup_early
Dust Li [Wed, 1 Dec 2021 03:02:30 +0000 (11:02 +0800)]
net/smc: fix wrong list_del in smc_lgr_cleanup_early

smc_lgr_cleanup_early() meant to delete the link
group from the link group list, but it deleted
the list head by mistake.

This may cause memory corruption since we didn't
remove the real link group from the list and later
memseted the link group structure.
We got a list corruption panic when testing:

[  231.277259] list_del corruption. prev->next should be ffff8881398a8000, but was 0000000000000000
[  231.278222] ------------[ cut here ]------------
[  231.278726] kernel BUG at lib/list_debug.c:53!
[  231.279326] invalid opcode: 0000 [#1] SMP NOPTI
[  231.279803] CPU: 0 PID: 5 Comm: kworker/0:0 Not tainted 5.10.46+ #435
[  231.280466] Hardware name: Alibaba Cloud ECS, BIOS 8c24b4c 04/01/2014
[  231.281248] Workqueue: events smc_link_down_work
[  231.281732] RIP: 0010:__list_del_entry_valid+0x70/0x90
[  231.282258] Code: 4c 60 82 e8 7d cc 6a 00 0f 0b 48 89 fe 48 c7 c7 88 4c
60 82 e8 6c cc 6a 00 0f 0b 48 89 fe 48 c7 c7 c0 4c 60 82 e8 5b cc 6a 00 <0f>
0b 48 89 fe 48 c7 c7 00 4d 60 82 e8 4a cc 6a 00 0f 0b cc cc cc
[  231.284146] RSP: 0018:ffffc90000033d58 EFLAGS: 00010292
[  231.284685] RAX: 0000000000000054 RBX: ffff8881398a8000 RCX: 0000000000000000
[  231.285415] RDX: 0000000000000001 RSI: ffff88813bc18040 RDI: ffff88813bc18040
[  231.286141] RBP: ffffffff8305ad40 R08: 0000000000000003 R09: 0000000000000001
[  231.286873] R10: ffffffff82803da0 R11: ffffc90000033b90 R12: 0000000000000001
[  231.287606] R13: 0000000000000000 R14: ffff8881398a8000 R15: 0000000000000003
[  231.288337] FS:  0000000000000000(0000) GS:ffff88813bc00000(0000) knlGS:0000000000000000
[  231.289160] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  231.289754] CR2: 0000000000e72058 CR3: 000000010fa96006 CR4: 00000000003706f0
[  231.290485] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  231.291211] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  231.291940] Call Trace:
[  231.292211]  smc_lgr_terminate_sched+0x53/0xa0
[  231.292677]  smc_switch_conns+0x75/0x6b0
[  231.293085]  ? update_load_avg+0x1a6/0x590
[  231.293517]  ? ttwu_do_wakeup+0x17/0x150
[  231.293907]  ? update_load_avg+0x1a6/0x590
[  231.294317]  ? newidle_balance+0xca/0x3d0
[  231.294716]  smcr_link_down+0x50/0x1a0
[  231.295090]  ? __wake_up_common_lock+0x77/0x90
[  231.295534]  smc_link_down_work+0x46/0x60
[  231.295933]  process_one_work+0x18b/0x350

Fixes: a0a62ee15a829 ("net/smc: separate locks for SMCD and SMCR link group lists")
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Acked-by: Karsten Graul <kgraul@linux.ibm.com>
Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoFix Comment of ETH_P_802_3_MIN
Xiayu Zhang [Wed, 1 Dec 2021 02:57:13 +0000 (10:57 +0800)]
Fix Comment of ETH_P_802_3_MIN

The description of ETH_P_802_3_MIN is misleading.
The value of EthernetType in Ethernet II frame is more than 0x0600,
the value of Length in 802.3 frame is less than 0x0600.

Signed-off-by: Xiayu Zhang <Xiayu.Zhang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoethernet: aquantia: Try MAC address from device tree
Tianhao Chai [Wed, 1 Dec 2021 02:57:06 +0000 (20:57 -0600)]
ethernet: aquantia: Try MAC address from device tree

Apple M1 Mac minis (2020) with 10GE NICs do not have MAC address in the
card, but instead need to obtain MAC addresses from the device tree. In
this case the hardware will report an invalid MAC.

Currently atlantic driver does not query the DT for MAC address and will
randomly assign a MAC if the NIC doesn't have a permanent MAC burnt in.
This patch causes the driver to perfer a valid MAC address from OF (if
present) over HW self-reported MAC and only fall back to a random MAC
address when neither of them is valid.

Signed-off-by: Tianhao Chai <cth451@gmail.com>
Reviewed-by: Igor Russkikh <irusskikh@marvell.com>
Reviewed-by: Hector Martin <marcan@marcan.st>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoipv4: convert fib_num_tclassid_users to atomic_t
Eric Dumazet [Thu, 2 Dec 2021 02:26:35 +0000 (18:26 -0800)]
ipv4: convert fib_num_tclassid_users to atomic_t

Before commit faa041a40b9f ("ipv4: Create cleanup helper for fib_nh")
changes to net->ipv4.fib_num_tclassid_users were protected by RTNL.

After the change, this is no longer the case, as free_fib_info_rcu()
runs after rcu grace period, without rtnl being held.

Fixes: faa041a40b9f ("ipv4: Create cleanup helper for fib_nh")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: David Ahern <dsahern@kernel.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: avoid uninit-value from tcp_conn_request
Eric Dumazet [Tue, 30 Nov 2021 18:29:39 +0000 (10:29 -0800)]
net: avoid uninit-value from tcp_conn_request

A recent change triggers a KMSAN warning, because request
sockets do not initialize @sk_rx_queue_mapping field.

Add sk_rx_queue_update() helper to make our intent clear.

BUG: KMSAN: uninit-value in sk_rx_queue_set include/net/sock.h:1922 [inline]
BUG: KMSAN: uninit-value in tcp_conn_request+0x3bcc/0x4dc0 net/ipv4/tcp_input.c:6922
 sk_rx_queue_set include/net/sock.h:1922 [inline]
 tcp_conn_request+0x3bcc/0x4dc0 net/ipv4/tcp_input.c:6922
 tcp_v4_conn_request+0x218/0x2a0 net/ipv4/tcp_ipv4.c:1528
 tcp_rcv_state_process+0x2c5/0x3290 net/ipv4/tcp_input.c:6406
 tcp_v4_do_rcv+0xb4e/0x1330 net/ipv4/tcp_ipv4.c:1738
 tcp_v4_rcv+0x468d/0x4ed0 net/ipv4/tcp_ipv4.c:2100
 ip_protocol_deliver_rcu+0x760/0x10b0 net/ipv4/ip_input.c:204
 ip_local_deliver_finish net/ipv4/ip_input.c:231 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ip_local_deliver+0x584/0x8c0 net/ipv4/ip_input.c:252
 dst_input include/net/dst.h:460 [inline]
 ip_sublist_rcv_finish net/ipv4/ip_input.c:551 [inline]
 ip_list_rcv_finish net/ipv4/ip_input.c:601 [inline]
 ip_sublist_rcv+0x11fd/0x1520 net/ipv4/ip_input.c:609
 ip_list_rcv+0x95f/0x9a0 net/ipv4/ip_input.c:644
 __netif_receive_skb_list_ptype net/core/dev.c:5505 [inline]
 __netif_receive_skb_list_core+0xe34/0x1240 net/core/dev.c:5553
 __netif_receive_skb_list+0x7fc/0x960 net/core/dev.c:5605
 netif_receive_skb_list_internal+0x868/0xde0 net/core/dev.c:5696
 gro_normal_list net/core/dev.c:5850 [inline]
 napi_complete_done+0x579/0xdd0 net/core/dev.c:6587
 virtqueue_napi_complete drivers/net/virtio_net.c:339 [inline]
 virtnet_poll+0x17b6/0x2350 drivers/net/virtio_net.c:1557
 __napi_poll+0x14e/0xbc0 net/core/dev.c:7020
 napi_poll net/core/dev.c:7087 [inline]
 net_rx_action+0x824/0x1880 net/core/dev.c:7174
 __do_softirq+0x1fe/0x7eb kernel/softirq.c:558
 invoke_softirq+0xa4/0x130 kernel/softirq.c:432
 __irq_exit_rcu kernel/softirq.c:636 [inline]
 irq_exit_rcu+0x76/0x130 kernel/softirq.c:648
 common_interrupt+0xb6/0xd0 arch/x86/kernel/irq.c:240
 asm_common_interrupt+0x1e/0x40
 smap_restore arch/x86/include/asm/smap.h:67 [inline]
 get_shadow_origin_ptr mm/kmsan/instrumentation.c:31 [inline]
 __msan_metadata_ptr_for_load_1+0x28/0x30 mm/kmsan/instrumentation.c:63
 tomoyo_check_acl+0x1b0/0x630 security/tomoyo/domain.c:173
 tomoyo_path_permission security/tomoyo/file.c:586 [inline]
 tomoyo_check_open_permission+0x61f/0xe10 security/tomoyo/file.c:777
 tomoyo_file_open+0x24f/0x2d0 security/tomoyo/tomoyo.c:311
 security_file_open+0xb1/0x1f0 security/security.c:1635
 do_dentry_open+0x4e4/0x1bf0 fs/open.c:809
 vfs_open+0xaf/0xe0 fs/open.c:957
 do_open fs/namei.c:3426 [inline]
 path_openat+0x52f1/0x5dd0 fs/namei.c:3559
 do_filp_open+0x306/0x760 fs/namei.c:3586
 do_sys_openat2+0x263/0x8f0 fs/open.c:1212
 do_sys_open fs/open.c:1228 [inline]
 __do_sys_open fs/open.c:1236 [inline]
 __se_sys_open fs/open.c:1232 [inline]
 __x64_sys_open+0x314/0x380 fs/open.c:1232
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_64+0x54/0xd0 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Uninit was created at:
 __alloc_pages+0xbc7/0x10a0 mm/page_alloc.c:5409
 alloc_pages+0x8a5/0xb80
 alloc_slab_page mm/slub.c:1810 [inline]
 allocate_slab+0x287/0x1c20 mm/slub.c:1947
 new_slab mm/slub.c:2010 [inline]
 ___slab_alloc+0xbdf/0x1e90 mm/slub.c:3039
 __slab_alloc mm/slub.c:3126 [inline]
 slab_alloc_node mm/slub.c:3217 [inline]
 slab_alloc mm/slub.c:3259 [inline]
 kmem_cache_alloc+0xbb3/0x11c0 mm/slub.c:3264
 reqsk_alloc include/net/request_sock.h:91 [inline]
 inet_reqsk_alloc+0xaf/0x8b0 net/ipv4/tcp_input.c:6712
 tcp_conn_request+0x910/0x4dc0 net/ipv4/tcp_input.c:6852
 tcp_v4_conn_request+0x218/0x2a0 net/ipv4/tcp_ipv4.c:1528
 tcp_rcv_state_process+0x2c5/0x3290 net/ipv4/tcp_input.c:6406
 tcp_v4_do_rcv+0xb4e/0x1330 net/ipv4/tcp_ipv4.c:1738
 tcp_v4_rcv+0x468d/0x4ed0 net/ipv4/tcp_ipv4.c:2100
 ip_protocol_deliver_rcu+0x760/0x10b0 net/ipv4/ip_input.c:204
 ip_local_deliver_finish net/ipv4/ip_input.c:231 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ip_local_deliver+0x584/0x8c0 net/ipv4/ip_input.c:252
 dst_input include/net/dst.h:460 [inline]
 ip_sublist_rcv_finish net/ipv4/ip_input.c:551 [inline]
 ip_list_rcv_finish net/ipv4/ip_input.c:601 [inline]
 ip_sublist_rcv+0x11fd/0x1520 net/ipv4/ip_input.c:609
 ip_list_rcv+0x95f/0x9a0 net/ipv4/ip_input.c:644
 __netif_receive_skb_list_ptype net/core/dev.c:5505 [inline]
 __netif_receive_skb_list_core+0xe34/0x1240 net/core/dev.c:5553
 __netif_receive_skb_list+0x7fc/0x960 net/core/dev.c:5605
 netif_receive_skb_list_internal+0x868/0xde0 net/core/dev.c:5696
 gro_normal_list net/core/dev.c:5850 [inline]
 napi_complete_done+0x579/0xdd0 net/core/dev.c:6587
 virtqueue_napi_complete drivers/net/virtio_net.c:339 [inline]
 virtnet_poll+0x17b6/0x2350 drivers/net/virtio_net.c:1557
 __napi_poll+0x14e/0xbc0 net/core/dev.c:7020
 napi_poll net/core/dev.c:7087 [inline]
 net_rx_action+0x824/0x1880 net/core/dev.c:7174
 __do_softirq+0x1fe/0x7eb kernel/softirq.c:558

Fixes: 342159ee394d ("net: avoid dirtying sk->sk_rx_queue_mapping")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Link: https://lore.kernel.org/r/20211130182939.2584764-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: annotate data-races on txq->xmit_lock_owner
Eric Dumazet [Tue, 30 Nov 2021 17:01:55 +0000 (09:01 -0800)]
net: annotate data-races on txq->xmit_lock_owner

syzbot found that __dev_queue_xmit() is reading txq->xmit_lock_owner
without annotations.

No serious issue there, let's document what is happening there.

BUG: KCSAN: data-race in __dev_queue_xmit / __dev_queue_xmit

write to 0xffff888139d09484 of 4 bytes by interrupt on cpu 0:
 __netif_tx_unlock include/linux/netdevice.h:4437 [inline]
 __dev_queue_xmit+0x948/0xf70 net/core/dev.c:4229
 dev_queue_xmit_accel+0x19/0x20 net/core/dev.c:4265
 macvlan_queue_xmit drivers/net/macvlan.c:543 [inline]
 macvlan_start_xmit+0x2b3/0x3d0 drivers/net/macvlan.c:567
 __netdev_start_xmit include/linux/netdevice.h:4987 [inline]
 netdev_start_xmit include/linux/netdevice.h:5001 [inline]
 xmit_one+0x105/0x2f0 net/core/dev.c:3590
 dev_hard_start_xmit+0x72/0x120 net/core/dev.c:3606
 sch_direct_xmit+0x1b2/0x7c0 net/sched/sch_generic.c:342
 __dev_xmit_skb+0x83d/0x1370 net/core/dev.c:3817
 __dev_queue_xmit+0x590/0xf70 net/core/dev.c:4194
 dev_queue_xmit+0x13/0x20 net/core/dev.c:4259
 neigh_hh_output include/net/neighbour.h:511 [inline]
 neigh_output include/net/neighbour.h:525 [inline]
 ip6_finish_output2+0x995/0xbb0 net/ipv6/ip6_output.c:126
 __ip6_finish_output net/ipv6/ip6_output.c:191 [inline]
 ip6_finish_output+0x444/0x4c0 net/ipv6/ip6_output.c:201
 NF_HOOK_COND include/linux/netfilter.h:296 [inline]
 ip6_output+0x10e/0x210 net/ipv6/ip6_output.c:224
 dst_output include/net/dst.h:450 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ndisc_send_skb+0x486/0x610 net/ipv6/ndisc.c:508
 ndisc_send_rs+0x3b0/0x3e0 net/ipv6/ndisc.c:702
 addrconf_rs_timer+0x370/0x540 net/ipv6/addrconf.c:3898
 call_timer_fn+0x2e/0x240 kernel/time/timer.c:1421
 expire_timers+0x116/0x240 kernel/time/timer.c:1466
 __run_timers+0x368/0x410 kernel/time/timer.c:1734
 run_timer_softirq+0x2e/0x60 kernel/time/timer.c:1747
 __do_softirq+0x158/0x2de kernel/softirq.c:558
 __irq_exit_rcu kernel/softirq.c:636 [inline]
 irq_exit_rcu+0x37/0x70 kernel/softirq.c:648
 sysvec_apic_timer_interrupt+0x3e/0xb0 arch/x86/kernel/apic/apic.c:1097
 asm_sysvec_apic_timer_interrupt+0x12/0x20

read to 0xffff888139d09484 of 4 bytes by interrupt on cpu 1:
 __dev_queue_xmit+0x5e3/0xf70 net/core/dev.c:4213
 dev_queue_xmit_accel+0x19/0x20 net/core/dev.c:4265
 macvlan_queue_xmit drivers/net/macvlan.c:543 [inline]
 macvlan_start_xmit+0x2b3/0x3d0 drivers/net/macvlan.c:567
 __netdev_start_xmit include/linux/netdevice.h:4987 [inline]
 netdev_start_xmit include/linux/netdevice.h:5001 [inline]
 xmit_one+0x105/0x2f0 net/core/dev.c:3590
 dev_hard_start_xmit+0x72/0x120 net/core/dev.c:3606
 sch_direct_xmit+0x1b2/0x7c0 net/sched/sch_generic.c:342
 __dev_xmit_skb+0x83d/0x1370 net/core/dev.c:3817
 __dev_queue_xmit+0x590/0xf70 net/core/dev.c:4194
 dev_queue_xmit+0x13/0x20 net/core/dev.c:4259
 neigh_resolve_output+0x3db/0x410 net/core/neighbour.c:1523
 neigh_output include/net/neighbour.h:527 [inline]
 ip6_finish_output2+0x9be/0xbb0 net/ipv6/ip6_output.c:126
 __ip6_finish_output net/ipv6/ip6_output.c:191 [inline]
 ip6_finish_output+0x444/0x4c0 net/ipv6/ip6_output.c:201
 NF_HOOK_COND include/linux/netfilter.h:296 [inline]
 ip6_output+0x10e/0x210 net/ipv6/ip6_output.c:224
 dst_output include/net/dst.h:450 [inline]
 NF_HOOK include/linux/netfilter.h:307 [inline]
 ndisc_send_skb+0x486/0x610 net/ipv6/ndisc.c:508
 ndisc_send_rs+0x3b0/0x3e0 net/ipv6/ndisc.c:702
 addrconf_rs_timer+0x370/0x540 net/ipv6/addrconf.c:3898
 call_timer_fn+0x2e/0x240 kernel/time/timer.c:1421
 expire_timers+0x116/0x240 kernel/time/timer.c:1466
 __run_timers+0x368/0x410 kernel/time/timer.c:1734
 run_timer_softirq+0x2e/0x60 kernel/time/timer.c:1747
 __do_softirq+0x158/0x2de kernel/softirq.c:558
 __irq_exit_rcu kernel/softirq.c:636 [inline]
 irq_exit_rcu+0x37/0x70 kernel/softirq.c:648
 sysvec_apic_timer_interrupt+0x8d/0xb0 arch/x86/kernel/apic/apic.c:1097
 asm_sysvec_apic_timer_interrupt+0x12/0x20
 kcsan_setup_watchpoint+0x94/0x420 kernel/kcsan/core.c:443
 folio_test_anon include/linux/page-flags.h:581 [inline]
 PageAnon include/linux/page-flags.h:586 [inline]
 zap_pte_range+0x5ac/0x10e0 mm/memory.c:1347
 zap_pmd_range mm/memory.c:1467 [inline]
 zap_pud_range mm/memory.c:1496 [inline]
 zap_p4d_range mm/memory.c:1517 [inline]
 unmap_page_range+0x2dc/0x3d0 mm/memory.c:1538
 unmap_single_vma+0x157/0x210 mm/memory.c:1583
 unmap_vmas+0xd0/0x180 mm/memory.c:1615
 exit_mmap+0x23d/0x470 mm/mmap.c:3170
 __mmput+0x27/0x1b0 kernel/fork.c:1113
 mmput+0x3d/0x50 kernel/fork.c:1134
 exit_mm+0xdb/0x170 kernel/exit.c:507
 do_exit+0x608/0x17a0 kernel/exit.c:819
 do_group_exit+0xce/0x180 kernel/exit.c:929
 get_signal+0xfc3/0x1550 kernel/signal.c:2852
 arch_do_signal_or_restart+0x8c/0x2e0 arch/x86/kernel/signal.c:868
 handle_signal_work kernel/entry/common.c:148 [inline]
 exit_to_user_mode_loop kernel/entry/common.c:172 [inline]
 exit_to_user_mode_prepare+0x113/0x190 kernel/entry/common.c:207
 __syscall_exit_to_user_mode_work kernel/entry/common.c:289 [inline]
 syscall_exit_to_user_mode+0x20/0x40 kernel/entry/common.c:300
 do_syscall_64+0x50/0xd0 arch/x86/entry/common.c:86
 entry_SYSCALL_64_after_hwframe+0x44/0xae

value changed: 0x00000000 -> 0xffffffff

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 28712 Comm: syz-executor.0 Tainted: G        W         5.16.0-rc1-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Link: https://lore.kernel.org/r/20211130170155.2331929-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoocteontx2-af: Fix a memleak bug in rvu_mbox_init()
Zhou Qingyang [Tue, 30 Nov 2021 16:50:39 +0000 (00:50 +0800)]
octeontx2-af: Fix a memleak bug in rvu_mbox_init()

In rvu_mbox_init(), mbox_regions is not freed or passed out
under the switch-default region, which could lead to a memory leak.

Fix this bug by changing 'return err' to 'goto free_regions'.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_OCTEONTX2_AF=y show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: 98c561116360 (“octeontx2-af: cn10k: Add mbox support for CN10K platform”)
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Link: https://lore.kernel.org/r/20211130165039.192426-1-zhou1615@umn.edu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet/mlx4_en: Fix an use-after-free bug in mlx4_en_try_alloc_resources()
Zhou Qingyang [Tue, 30 Nov 2021 16:44:38 +0000 (00:44 +0800)]
net/mlx4_en: Fix an use-after-free bug in mlx4_en_try_alloc_resources()

In mlx4_en_try_alloc_resources(), mlx4_en_copy_priv() is called and
tmp->tx_cq will be freed on the error path of mlx4_en_copy_priv().
After that mlx4_en_alloc_resources() is called and there is a dereference
of &tmp->tx_cq[t][i] in mlx4_en_alloc_resources(), which could lead to
a use after free problem on failure of mlx4_en_copy_priv().

Fix this bug by adding a check of mlx4_en_copy_priv()

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_MLX4_EN=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: ec25bc04ed8e ("net/mlx4_en: Add resilience in low memory systems")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20211130164438.190591-1-zhou1615@umn.edu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agovrf: Reset IPCB/IP6CB when processing outbound pkts in vrf dev xmit
Stephen Suryaputra [Tue, 30 Nov 2021 16:26:37 +0000 (11:26 -0500)]
vrf: Reset IPCB/IP6CB when processing outbound pkts in vrf dev xmit

IPCB/IP6CB need to be initialized when processing outbound v4 or v6 pkts
in the codepath of vrf device xmit function so that leftover garbage
doesn't cause futher code that uses the CB to incorrectly process the
pkt.

One occasion of the issue might occur when MPLS route uses the vrf
device as the outgoing device such as when the route is added using "ip
-f mpls route add <label> dev <vrf>" command.

The problems seems to exist since day one. Hence I put the day one
commits on the Fixes tags.

Fixes: 193125dbd8eb ("net: Introduce VRF device driver")
Fixes: 35402e313663 ("net: Add IPv6 support to VRF device")
Cc: stable@vger.kernel.org
Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20211130162637.3249-1-ssuryaextr@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: qlogic: qlcnic: Fix a NULL pointer dereference in qlcnic_83xx_add_rings()
Zhou Qingyang [Tue, 30 Nov 2021 11:08:48 +0000 (19:08 +0800)]
net: qlogic: qlcnic: Fix a NULL pointer dereference in qlcnic_83xx_add_rings()

In qlcnic_83xx_add_rings(), the indirect function of
ahw->hw_ops->alloc_mbx_args will be called to allocate memory for
cmd.req.arg, and there is a dereference of it in qlcnic_83xx_add_rings(),
which could lead to a NULL pointer dereference on failure of the
indirect function like qlcnic_83xx_alloc_mbx_args().

Fix this bug by adding a check of alloc_mbx_args(), this patch
imitates the logic of mbx_cmd()'s failure handling.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_QLCNIC=m show no new warnings, and our
static analyzer no longer warns about this code.

Fixes: 7f9664525f9c ("qlcnic: 83xx memory map and HW access routine")
Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
Link: https://lore.kernel.org/r/20211130110848.109026-1-zhou1615@umn.edu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agokprobes: Limit max data_size of the kretprobe instances
Masami Hiramatsu [Wed, 1 Dec 2021 14:45:50 +0000 (23:45 +0900)]
kprobes: Limit max data_size of the kretprobe instances

The 'kprobe::data_size' is unsigned, thus it can not be negative.  But if
user sets it enough big number (e.g. (size_t)-8), the result of 'data_size
+ sizeof(struct kretprobe_instance)' becomes smaller than sizeof(struct
kretprobe_instance) or zero. In result, the kretprobe_instance are
allocated without enough memory, and kretprobe accesses outside of
allocated memory.

To avoid this issue, introduce a max limitation of the
kretprobe::data_size. 4KB per instance should be OK.

Link: https://lkml.kernel.org/r/163836995040.432120.10322772773821182925.stgit@devnote2
Cc: stable@vger.kernel.org
Fixes: f47cd9b553aa ("kprobes: kretprobe user entry-handler")
Reported-by: zhangyue <zhangyue1@kylinos.cn>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2 years agotracing: Fix a kmemleak false positive in tracing_map
Chen Jun [Wed, 24 Nov 2021 14:08:01 +0000 (14:08 +0000)]
tracing: Fix a kmemleak false positive in tracing_map

Doing the command:
  echo 'hist:key=common_pid.execname,common_timestamp' > /sys/kernel/debug/tracing/events/xxx/trigger

Triggers many kmemleak reports:

unreferenced object 0xffff0000c7ea4980 (size 128):
  comm "bash", pid 338, jiffies 4294912626 (age 9339.324s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<00000000f3469921>] kmem_cache_alloc_trace+0x4c0/0x6f0
    [<0000000054ca40c3>] hist_trigger_elt_data_alloc+0x140/0x178
    [<00000000633bd154>] tracing_map_init+0x1f8/0x268
    [<000000007e814ab9>] event_hist_trigger_func+0xca0/0x1ad0
    [<00000000bf8520ed>] trigger_process_regex+0xd4/0x128
    [<00000000f549355a>] event_trigger_write+0x7c/0x120
    [<00000000b80f898d>] vfs_write+0xc4/0x380
    [<00000000823e1055>] ksys_write+0x74/0xf8
    [<000000008a9374aa>] __arm64_sys_write+0x24/0x30
    [<0000000087124017>] do_el0_svc+0x88/0x1c0
    [<00000000efd0dcd1>] el0_svc+0x1c/0x28
    [<00000000dbfba9b3>] el0_sync_handler+0x88/0xc0
    [<00000000e7399680>] el0_sync+0x148/0x180
unreferenced object 0xffff0000c7ea4980 (size 128):
  comm "bash", pid 338, jiffies 4294912626 (age 9339.324s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<00000000f3469921>] kmem_cache_alloc_trace+0x4c0/0x6f0
    [<0000000054ca40c3>] hist_trigger_elt_data_alloc+0x140/0x178
    [<00000000633bd154>] tracing_map_init+0x1f8/0x268
    [<000000007e814ab9>] event_hist_trigger_func+0xca0/0x1ad0
    [<00000000bf8520ed>] trigger_process_regex+0xd4/0x128
    [<00000000f549355a>] event_trigger_write+0x7c/0x120
    [<00000000b80f898d>] vfs_write+0xc4/0x380
    [<00000000823e1055>] ksys_write+0x74/0xf8
    [<000000008a9374aa>] __arm64_sys_write+0x24/0x30
    [<0000000087124017>] do_el0_svc+0x88/0x1c0
    [<00000000efd0dcd1>] el0_svc+0x1c/0x28
    [<00000000dbfba9b3>] el0_sync_handler+0x88/0xc0
    [<00000000e7399680>] el0_sync+0x148/0x180

The reason is elts->pages[i] is alloced by get_zeroed_page.
and kmemleak will not scan the area alloced by get_zeroed_page.
The address stored in elts->pages will be regarded as leaked.

That is, the elts->pages[i] will have pointers loaded onto it as well, and
without telling kmemleak about it, those pointers will look like memory
without a reference.

To fix this, call kmemleak_alloc to tell kmemleak to scan elts->pages[i]

Link: https://lkml.kernel.org/r/20211124140801.87121-1-chenjun102@huawei.com
Signed-off-by: Chen Jun <chenjun102@huawei.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2 years agotracing/histograms: String compares should not care about signed values
Steven Rostedt (VMware) [Tue, 30 Nov 2021 17:31:23 +0000 (12:31 -0500)]
tracing/histograms: String compares should not care about signed values

When comparing two strings for the "onmatch" histogram trigger, fields
that are strings use string comparisons, which do not care about being
signed or not.

Do not fail to match two string fields if one is unsigned char array and
the other is a signed char array.

Link: https://lore.kernel.org/all/20211129123043.5cfd687a@gandalf.local.home/
Cc: stable@vgerk.kernel.org
Cc: Tom Zanussi <zanussi@kernel.org>
Cc: Yafang Shao <laoar.shao@gmail.com>
Fixes: b05e89ae7cf3b ("tracing: Accept different type for synthetic event fields")
Reviewed-by: Masami Hiramatsu <mhiramatsu@kernel.org>
Reported-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2 years agoiavf: restore MSI state on reset
Mitch Williams [Fri, 4 Jun 2021 16:53:28 +0000 (09:53 -0700)]
iavf: restore MSI state on reset

If the PF experiences an FLR, the VF's MSI and MSI-X configuration will
be conveniently and silently removed in the process. When this happens,
reset recovery will appear to complete normally but no traffic will
pass. The netdev watchdog will helpfully notify everyone of this issue.

To prevent such public embarrassment, restore MSI configuration at every
reset. For normal resets, this will do no harm, but for VF resets
resulting from a PF FLR, this will keep the VF working.

Fixes: 5eae00c57f5e ("i40evf: main driver core")
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: George Kuruvinakunnel <george.kuruvinakunnel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
2 years agoMerge tag 'sound-5.16-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Wed, 1 Dec 2021 18:07:39 +0000 (10:07 -0800)]
Merge tag 'sound-5.16-rc4' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of small fixes. A large series is found for ASoC tegra
  drivers to correct the control element handlings, while others are
  mostly for device-specific quirks and fix-ups"

* tag 'sound-5.16-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (25 commits)
  ALSA: hda/hdmi: fix HDA codec entry table order for ADL-P
  ALSA: hda: Add Intel DG2 PCI ID and HDMI codec vid
  ALSA: hda/cs8409: Set PMSG_ON earlier inside cs8409 driver
  ASoC: SOF: hda: reset DAI widget before reconfiguring it
  ASoC: cs35l41: Set the max SPI speed for the whole device
  ALSA: intel-dsp-config: add quirk for CML devices based on ES8336 codec
  ASoC: Intel: soc-acpi: add entry for ESSX8336 on CML
  ASoC: rk817: Add module alias for rk817-codec
  ASoC: soc-acpi: Set mach->id field on comp_ids matches
  ASoC: tegra: Fix kcontrol put callback in Mixer
  ASoC: tegra: Fix kcontrol put callback in ADX
  ASoC: tegra: Fix kcontrol put callback in AMX
  ASoC: tegra: Fix kcontrol put callback in SFC
  ASoC: tegra: Fix kcontrol put callback in MVC
  ASoC: tegra: Fix kcontrol put callback in AHUB
  ASoC: tegra: Fix kcontrol put callback in DSPK
  ASoC: tegra: Fix kcontrol put callback in DMIC
  ASoC: tegra: Fix kcontrol put callback in I2S
  ASoC: tegra: Fix kcontrol put callback in ADMAIF
  ASoC: tegra: Fix wrong value type in MVC
  ...

2 years agoMerge tag 'wireless-drivers-2021-12-01' of git://git.kernel.org/pub/scm/linux/kernel...
David S. Miller [Wed, 1 Dec 2021 15:40:43 +0000 (15:40 +0000)]
Merge tag 'wireless-drivers-2021-12-01' of git://git./linux/kernel/git/kvalo/wireless-drivers

Kalle Valo says:

====================
wireless-drivers fixes for v5.16

First set of fixes for v5.16. Mostly crash and driver initialisation
fixes, the fix for rtw89 being most important.

iwlwifi

* compiler, lockdep and smatch warning fixes

* fix for a rare driver initialisation failure

* fix a memory leak

rtw89

* fix const buffer modification causing a kernel crash

mt76

* fix null pointer access

* fix idr leak

rt2x00

* fix driver initialisation errors, a regression since v5.2-rc1
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge tag 'mlx5-fixes-2021-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Wed, 1 Dec 2021 14:50:13 +0000 (14:50 +0000)]
Merge tag 'mlx5-fixes-2021-11-30' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2021-11-30

This series provides bug fixes to mlx5 driver.
Please pull and let me know if there is any problem.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'mv88e6xxx-fixes'
David S. Miller [Wed, 1 Dec 2021 14:39:10 +0000 (14:39 +0000)]
Merge branch 'mv88e6xxx-fixes'

Marek Behún says:

====================
mv88e6xxx fixes (mainly 88E6393X family)

sending v2 of these fixes.

Original cover letter:

So I managed to discovered how to fix inband AN for 2500base-x mode on
88E6393x (Amethyst) family.

This series fixes application of erratum 4.8, adds fix for erratum 5.2,
adds support for completely disablign SerDes receiver / transmitter,
fixes inband AN for 2500base-x mode by using 1000base-x mode and simply
changing frequeny to 3.125 GHz, all this for 88E6393X.

The last commit fixes linking when link partner has AN disabled and the
device invokes the AN bypass feature. Currently we fail to link in this
case.

Changes since v1:
- fixed wrong operator in patch 3 (thanks Russell)
- added more comments about why BMCR_ANENABLE is used in patch 6 (thanks
  Russell)
- updated some return statements from
     if (something)
       return func();
     return 0;
  to
     if (something)
       err = func();
     return err;
  (err is set to 0 before the condition)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Link in pcs_get_state() if AN is bypassed
Marek Behún [Tue, 30 Nov 2021 17:01:51 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Link in pcs_get_state() if AN is bypassed

Function mv88e6xxx_serdes_pcs_get_state() currently does not report link
up if AN is enabled, Link bit is set, but Speed and Duplex Resolved bit
is not set, which testing shows is the case for when auto-negotiation
was bypassed (we have AN enabled but link partner does not).

An example of such link partner is Marvell 88X3310 PHY, when put into
the mode where host interface changes between 10gbase-r, 5gbase-r,
2500base-x and sgmii according to copper speed. The 88X3310 does not
enable AN in 2500base-x, and so SerDes on mv88e6xxx currently does not
link with it.

Fix this.

Fixes: a5a6858b793f ("net: dsa: mv88e6xxx: extend phylink to Serdes PHYs")
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Fix inband AN for 2500base-x on 88E6393X family
Marek Behún [Tue, 30 Nov 2021 17:01:50 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Fix inband AN for 2500base-x on 88E6393X family

Inband AN is broken on Amethyst in 2500base-x mode when set by standard
mechanism (via cmode).

(There probably is some weird setting done by default in the switch for
 this mode that make it cycle in some state or something, because when
 the peer is the mvneta controller, it receives link change interrupts
 every ~0.3ms, but the link is always down.)

Get around this by configuring the PCS mode to 1000base-x (where inband
AN works), and then changing the SerDes frequency while SerDes
transmitter and receiver are disabled, before enabling SerDes PHY. After
disabling SerDes PHY, change the PCS mode back to 2500base-x, to avoid
confusing the device (if we leave it at 1000base-x PCS mode but with
different frequency, and then change cmode to sgmii, the device won't
change the frequency because it thinks it already has the correct one).

The register which changes the frequency is undocumented. I discovered
it by going through all registers in the ranges 4.f000-4.f100 and
1e.8000-1e.8200 for all SerDes cmodes (sgmii, 1000base-x, 2500base-x,
5gbase-r, 10gbase-r, usxgmii) and filtering out registers that didn't
make sense (the value was the same for modes which have different
frequency). The result of this was:

    reg   sgmii 1000base-x 2500base-x 5gbase-r 10gbase-r usxgmii
  04.f002  005b       0058       0059     005c      005d    005f
  04.f076  3000       0000       1000     4000      5000    7000
  04.f07c  0950       0950       1850     0550      0150    0150
  1e.8000  0059       0059       0058     0055      0051    0051
  1e.8140  0e20       0e20       0e28     0e21      0e42    0e42

Register 04.f002 is the documented Port Operational Confiuration
register, it's last 3 bits select PCS type, so changing this register
also changes the frequency to the appropriate value.

Registers 04.f076 and 04.f07c are not writable.

Undocumented register 1e.8000 was the one: changing bits 3:0 from 9 to 8
changed SerDes frequency to 3.125 GHz, while leaving the value of PCS
mode in register 04.f002.2:0 at 1000base-x. Inband autonegotiation
started working correctly.

(I didn't try anything with register 1e.8140 since 1e.8000 solved the
 problem.)

Since I don't have documentation for this register 1e.8000.3:0, I am
using the constants without names, but my hypothesis is that this
register selects PHY frequency. If in the future I have access to an
oscilloscope able to handle these frequencies, I will try to test this
hypothesis.

Fixes: de776d0d316f ("net: dsa: mv88e6xxx: add support for mv88e6393x family")
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Add fix for erratum 5.2 of 88E6393X family
Marek Behún [Tue, 30 Nov 2021 17:01:49 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Add fix for erratum 5.2 of 88E6393X family

Add fix for erratum 5.2 of the 88E6393X (Amethyst) family: for 10gbase-r
mode, some undocumented registers need to be written some special
values.

Fixes: de776d0d316f ("net: dsa: mv88e6xxx: add support for mv88e6393x family")
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Save power by disabling SerDes trasmitter and receiver
Marek Behún [Tue, 30 Nov 2021 17:01:48 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Save power by disabling SerDes trasmitter and receiver

Save power on 88E6393X by disabling SerDes receiver and transmitter
after SerDes is SerDes is disabled.

Signed-off-by: Marek Behún <kabel@kernel.org>
Cc: stable@vger.kernel.org # de776d0d316f ("net: dsa: mv88e6xxx: add support for mv88e6393x family")
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Drop unnecessary check in mv88e6393x_serdes_erratum_4_6()
Marek Behún [Tue, 30 Nov 2021 17:01:47 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Drop unnecessary check in mv88e6393x_serdes_erratum_4_6()

The check for lane is unnecessary, since the function is called only
with allowed lane argument.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: mv88e6xxx: Fix application of erratum 4.8 for 88E6393X
Marek Behún [Tue, 30 Nov 2021 17:01:46 +0000 (18:01 +0100)]
net: dsa: mv88e6xxx: Fix application of erratum 4.8 for 88E6393X

According to SERDES scripts for 88E6393X, erratum 4.8 has to be applied
every time before SerDes is powered on.

Split the code for erratum 4.8 into separate function and call it in
mv88e6393x_serdes_power().

Fixes: de776d0d316f ("net: dsa: mv88e6xxx: add support for mv88e6393x family")
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/mlx5e: SHAMPO, Fix constant expression result
Ben Ben-Ishay [Sun, 31 Oct 2021 16:31:02 +0000 (18:31 +0200)]
net/mlx5e: SHAMPO, Fix constant expression result

mlx5e_build_shampo_hd_umr uses counters i and index incorrectly
as unsigned, thus the err state err_unmap could stuck in endless loop.
Change i to int to solve the first issue.
Reduce index check to solve the second issue, the caller function
validates that index could not rotate.

Fixes: 64509b052525 ("net/mlx5e: Add data path for SHAMPO feature")
Signed-off-by: Ben Ben-Ishay <benishay@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
2 years agonet/mlx5: Fix access to a non-supported register
Aya Levin [Sun, 7 Nov 2021 15:21:45 +0000 (17:21 +0200)]
net/mlx5: Fix access to a non-supported register

Validate MRTC register is supported before triggering a delayed work
which accesses it.

Fixes: 5a1023deeed0 ("net/mlx5: Add periodic update of host time to firmware")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>