platform/kernel/linux-starfive.git
12 months agoselftests: openvswitch: Catch cases where the tests are killed
Aaron Conole [Wed, 11 Oct 2023 19:49:37 +0000 (15:49 -0400)]
selftests: openvswitch: Catch cases where the tests are killed

In case of fatal signal, or early abort at least cleanup the current
test case.

Fixes: 25f16c873fb1 ("selftests: add openvswitch selftest suite")
Signed-off-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 months agoselftests: openvswitch: Add version check for pyroute2
Aaron Conole [Wed, 11 Oct 2023 19:49:36 +0000 (15:49 -0400)]
selftests: openvswitch: Add version check for pyroute2

Paolo Abeni reports that on some systems the pyroute2 version isn't
new enough to run the test suite.  Ensure that we support a minimum
version of 0.6 for all cases (which does include the existing ones).
The 0.6.1 version was released in May of 2021, so should be
propagated to most installations at this point.

The alternative that Paolo proposed was to only skip when the
add-flow is being run.  This would be okay for most cases, except
if a future test case is added that needs to do flow dump without
an associated add (just guessing).  In that case, it could also be
broken and we would need additional skip logic anyway.  Just draw
a line in the sand now.

Fixes: 25f16c873fb1 ("selftests: add openvswitch selftest suite")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Closes: https://lore.kernel.org/lkml/8470c431e0930d2ea204a9363a60937289b7fdbe.camel@redhat.com/
Signed-off-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 months agonet: more strict VIRTIO_NET_HDR_GSO_UDP_L4 validation
Willem de Bruijn [Wed, 11 Oct 2023 14:01:14 +0000 (10:01 -0400)]
net: more strict VIRTIO_NET_HDR_GSO_UDP_L4 validation

Syzbot reported two new paths to hit an internal WARNING using the
new virtio gso type VIRTIO_NET_HDR_GSO_UDP_L4.

    RIP: 0010:skb_checksum_help+0x4a2/0x600 net/core/dev.c:3260
    skb len=64521 gso_size=344
and

    RIP: 0010:skb_warn_bad_offload+0x118/0x240 net/core/dev.c:3262

Older virtio types have historically had loose restrictions, leading
to many entirely impractical fuzzer generated packets causing
problems deep in the kernel stack. Ideally, we would have had strict
validation for all types from the start.

New virtio types can have tighter validation. Limit UDP GSO packets
inserted via virtio to the same limits imposed by the UDP_SEGMENT
socket interface:

1. must use checksum offload
2. checksum offload matches UDP header
3. no more segments than UDP_MAX_SEGMENTS
4. UDP GSO does not take modifier flags, notably SKB_GSO_TCP_ECN

Fixes: 860b7f27b8f7 ("linux/virtio_net.h: Support USO offload in vnet header.")
Reported-by: syzbot+01cdbc31e9c0ae9b33ac@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/0000000000005039270605eb0b7f@google.com/
Reported-by: syzbot+c99d835ff081ca30f986@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/0000000000005426680605eb0b9f@google.com/
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 months agoqed: fix LL2 RX buffer allocation
Manish Chopra [Fri, 13 Oct 2023 13:18:12 +0000 (18:48 +0530)]
qed: fix LL2 RX buffer allocation

Driver allocates the LL2 rx buffers from kmalloc()
area to construct the skb using slab_build_skb()

The required size allocation seems to have overlooked
for accounting both skb_shared_info size and device
placement padding bytes which results into the below
panic when doing skb_put() for a standard MTU sized frame.

skbuff: skb_over_panic: text:ffffffffc0b0225f len:1514 put:1514
head:ff3dabceaf39c000 data:ff3dabceaf39c042 tail:0x62c end:0x566
dev:<NULL>

skb_panic+0x48/0x4a
skb_put.cold+0x10/0x10
qed_ll2b_complete_rx_packet+0x14f/0x260 [qed]
qed_ll2_rxq_handle_completion.constprop.0+0x169/0x200 [qed]
qed_ll2_rxq_completion+0xba/0x320 [qed]
qed_int_sp_dpc+0x1a7/0x1e0 [qed]

This patch fixes this by accouting skb_shared_info and device
placement padding size bytes when allocating the buffers.

Cc: David S. Miller <davem@davemloft.net>
Fixes: 0a7fb11c23c0 ("qed: Add Light L2 support")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 months agoMerge tag 'mlx5-fixes-2023-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Sat, 14 Oct 2023 01:00:27 +0000 (18:00 -0700)]
Merge tag 'mlx5-fixes-2023-10-12' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2023-10-12

This series provides bug fixes to mlx5 driver.

* tag 'mlx5-fixes-2023-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Fix VF representors reporting zero counters to "ip -s" command
  net/mlx5e: Don't offload internal port if filter device is out device
  net/mlx5e: Take RTNL lock before triggering netdev notifiers
  net/mlx5e: XDP, Fix XDP_REDIRECT mpwqe page fragment leaks on shutdown
  net/mlx5e: RX, Fix page_pool allocation failure recovery for legacy rq
  net/mlx5e: RX, Fix page_pool allocation failure recovery for striding rq
  net/mlx5: Handle fw tracer change ownership event based on MTRC
  net/mlx5: Bridge, fix peer entry ageing in LAG mode
  net/mlx5: E-switch, register event handler before arming the event
  net/mlx5: Perform DMA operations in the right locations
====================

Link: https://lore.kernel.org/r/20231012195127.129585-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoMerge branch 'intel-wired-lan-driver-updates-2023-10-11-i40e-ice'
Jakub Kicinski [Sat, 14 Oct 2023 00:57:07 +0000 (17:57 -0700)]
Merge branch 'intel-wired-lan-driver-updates-2023-10-11-i40e-ice'

Jacob Keller says:

====================
Intel Wired LAN Driver Updates 2023-10-11 (i40e, ice)

This series contains fixes for the i40e and ice drivers.

Jesse adds handling to the ice driver which resetis the device when loading
on a crash kernel, preventing stale transactions from causing machine check
exceptions which could prevent capturing crash data.

Mateusz fixes a bug in the ice driver 'Safe mode' logic for handling the
device when the DDP is missing.

Michal fixes a crash when probing the i40e driver in the event that HW
registers are reporting invalid/unexpected values.

The following are changes since commit a950a5921db450c74212327f69950ff03419483a:
  net/smc: Fix pos miscalculation in statistics

I'm covering for Tony Nguyen while he's out, and don't have access to create
a pull request branch on his net-queue, so these are sent via mail only.
====================

Link: https://lore.kernel.org/r/20231011233334.336092-1-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoice: Fix safe mode when DDP is missing
Mateusz Pacuszka [Wed, 11 Oct 2023 23:33:34 +0000 (16:33 -0700)]
ice: Fix safe mode when DDP is missing

One thing is broken in the safe mode, that is
ice_deinit_features() is being executed even
that ice_init_features() was not causing stack
trace during pci_unregister_driver().

Add check on the top of the function.

Fixes: 5b246e533d01 ("ice: split probe into smaller functions")
Signed-off-by: Mateusz Pacuszka <mateuszx.pacuszka@intel.com>
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Link: https://lore.kernel.org/r/20231011233334.336092-4-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoice: reset first in crash dump kernels
Jesse Brandeburg [Wed, 11 Oct 2023 23:33:33 +0000 (16:33 -0700)]
ice: reset first in crash dump kernels

When the system boots into the crash dump kernel after a panic, the ice
networking device may still have pending transactions that can cause errors
or machine checks when the device is re-enabled. This can prevent the crash
dump kernel from loading the driver or collecting the crash data.

To avoid this issue, perform a function level reset (FLR) on the ice device
via PCIe config space before enabling it on the crash kernel. This will
clear any outstanding transactions and stop all queues and interrupts.
Restore the config space after the FLR, otherwise it was found in testing
that the driver wouldn't load successfully.

The following sequence causes the original issue:
- Load the ice driver with modprobe ice
- Enable SR-IOV with 2 VFs: echo 2 > /sys/class/net/eth0/device/sriov_num_vfs
- Trigger a crash with echo c > /proc/sysrq-trigger
- Load the ice driver again (or let it load automatically) with modprobe ice
- The system crashes again during pcim_enable_device()

Fixes: 837f08fdecbe ("ice: Add basic driver framework for Intel(R) E800 Series")
Reported-by: Vishal Agrawal <vagrawal@redhat.com>
Reviewed-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Link: https://lore.kernel.org/r/20231011233334.336092-3-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoi40e: prevent crash on probe if hw registers have invalid values
Michal Schmidt [Wed, 11 Oct 2023 23:33:32 +0000 (16:33 -0700)]
i40e: prevent crash on probe if hw registers have invalid values

The hardware provides the indexes of the first and the last available
queue and VF. From the indexes, the driver calculates the numbers of
queues and VFs. In theory, a faulty device might say the last index is
smaller than the first index. In that case, the driver's calculation
would underflow, it would attempt to write to non-existent registers
outside of the ioremapped range and crash.

I ran into this not by having a faulty device, but by an operator error.
I accidentally ran a QE test meant for i40e devices on an ice device.
The test used 'echo i40e > /sys/...ice PCI device.../driver_override',
bound the driver to the device and crashed in one of the wr32 calls in
i40e_clear_hw.

Add checks to prevent underflows in the calculations of num_queues and
num_vfs. With this fix, the wrong device probing reports errors and
returns a failure without crashing.

Fixes: 838d41d92a90 ("i40e: clear all queues and interrupts")
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Link: https://lore.kernel.org/r/20231011233334.336092-2-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoMerge tag 'nf-23-10-12' of https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Jakub Kicinski [Sat, 14 Oct 2023 00:50:58 +0000 (17:50 -0700)]
Merge tag 'nf-23-10-12' of https://git./linux/kernel/git/netfilter/nf

Florian Westphal says:

====================
netfilter updates for net

Patch 1, from Pablo Neira Ayuso, fixes a performance regression
(since 6.4) when a large pending set update has to be canceled towards
the end of the transaction.

Patch 2 from myself, silences an incorrect compiler warning reported
with a few (older) compiler toolchains.

Patch 3, from Kees Cook, adds __counted_by annotation to
nft_pipapo set backend type.  I took this for net instead of -next
given infra is already in place and no actual code change is made.

Patch 4, from Pablo Neira Ayso, disables timeout resets on
stateful element reset.  The rest should only affect internal object
state, e.g. reset a quota or counter, but not affect a pending timeout.

Patches 5 and 6 fix NULL dereferences in 'inner header' match,
control plane doesn't test for netlink attribute presence before
accessing them. Broken since feature was added in 6.2, fixes from
Xingyuan Mo.

Last patch, from myself, fixes a bogus rule match when skb has
a 0-length mac header, in this case we'd fetch data from network
header instead of canceling rule evaluation.  This is a day 0 bug,
present since nftables was merged in 3.13.

* tag 'nf-23-10-12' of https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nft_payload: fix wrong mac header matching
  nf_tables: fix NULL pointer dereference in nft_expr_inner_parse()
  nf_tables: fix NULL pointer dereference in nft_inner_init()
  netfilter: nf_tables: do not refresh timeout when resetting element
  netfilter: nf_tables: Annotate struct nft_pipapo_match with __counted_by
  netfilter: nfnetlink_log: silence bogus compiler warning
  netfilter: nf_tables: do not remove elements if set backend implements .abort
====================

Link: https://lore.kernel.org/r/20231012085724.15155-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agonet: ti: icssg-prueth: Fix tx_total_bytes count
MD Danish Anwar [Thu, 12 Oct 2023 06:46:26 +0000 (12:16 +0530)]
net: ti: icssg-prueth: Fix tx_total_bytes count

ICSSG HW stats on TX side considers 8 preamble bytes as data bytes. Due
to this the tx_bytes of ICSSG interface doesn't match the rx_bytes of the
link partner. There is no public errata available yet.

As a workaround to fix this, decrease tx_bytes by 8 bytes for every tx
frame.

Fixes: c1e10d5dc7a1 ("net: ti: icssg-prueth: Add ICSSG Stats")
Signed-off-by: MD Danish Anwar <danishanwar@ti.com>
Link: https://lore.kernel.org/r/20231012064626.977466-1-danishanwar@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agodocs: fix info about representor identification
Mateusz Polchlopek [Thu, 12 Oct 2023 12:31:44 +0000 (08:31 -0400)]
docs: fix info about representor identification

Update the "How are representors identified?" documentation
subchapter. For newer kernels driver should use
SET_NETDEV_DEVLINK_PORT instead of ndo_get_devlink_port()
callback.

Fixes: 7712b3e966ea ("Merge branch 'net-fix-netdev-to-devlink_port-linkage-and-expose-to-user'")
Signed-off-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/20231012123144.15768-1-mateusz.polchlopek@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agonetlink: specs: devlink: fix reply command values
Jiri Pirko [Thu, 12 Oct 2023 11:58:11 +0000 (13:58 +0200)]
netlink: specs: devlink: fix reply command values

Make sure that the command values used for replies are correct. This is
only affecting generated userspace helpers, no change on kernel code.

Fixes: 7199c86247e9 ("netlink: specs: devlink: add commands that do per-instance dump")
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231012115811.298129-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agonet/smc: fix smc clc failed issue when netdevice not in init_net
Albert Huang [Wed, 11 Oct 2023 07:48:51 +0000 (15:48 +0800)]
net/smc: fix smc clc failed issue when netdevice not in init_net

If the netdevice is within a container and communicates externally
through network technologies such as VxLAN, we won't be able to find
routing information in the init_net namespace. To address this issue,
we need to add a struct net parameter to the smc_ib_find_route function.
This allow us to locate the routing information within the corresponding
net namespace, ensuring the correct completion of the SMC CLC interaction.

Fixes: e5c4744cfb59 ("net/smc: add SMC-Rv2 connection establishment")
Signed-off-by: Albert Huang <huangjie.albert@bytedance.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Link: https://lore.kernel.org/r/20231011074851.95280-1-huangjie.albert@bytedance.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agotcp: allow again tcp_disconnect() when threads are waiting
Paolo Abeni [Wed, 11 Oct 2023 07:20:55 +0000 (09:20 +0200)]
tcp: allow again tcp_disconnect() when threads are waiting

As reported by Tom, .NET and applications build on top of it rely
on connect(AF_UNSPEC) to async cancel pending I/O operations on TCP
socket.

The blamed commit below caused a regression, as such cancellation
can now fail.

As suggested by Eric, this change addresses the problem explicitly
causing blocking I/O operation to terminate immediately (with an error)
when a concurrent disconnect() is executed.

Instead of tracking the number of threads blocked on a given socket,
track the number of disconnect() issued on such socket. If such counter
changes after a blocking operation releasing and re-acquiring the socket
lock, error out the current operation.

Fixes: 4faeee0cf8a5 ("tcp: deny tcp_disconnect() when threads are waiting")
Reported-by: Tom Deseyn <tdeseyn@redhat.com>
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1886305
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/f3b95e47e3dbed840960548aebaa8d954372db41.1697008693.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoice: fix over-shifted variable
Jesse Brandeburg [Tue, 10 Oct 2023 20:30:59 +0000 (13:30 -0700)]
ice: fix over-shifted variable

Since the introduction of the ice driver the code has been
double-shifting the RSS enabling field, because the define already has
shifts in it and can't have the regular pattern of "a << shiftval &
mask" applied.

Most places in the code got it right, but one line was still wrong. Fix
this one location for easy backports to stable. An in-progress patch
fixes the defines to "standard" and will be applied as part of the
regular -next process sometime after this one.

Fixes: d76a60ba7afb ("ice: Add support for VLANs and offloads")
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
CC: stable@vger.kernel.org
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231010203101.406248-1-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agonet: dsa: bcm_sf2: Fix possible memory leak in bcm_sf2_mdio_register()
Jinjie Ruan [Wed, 11 Oct 2023 03:24:19 +0000 (11:24 +0800)]
net: dsa: bcm_sf2: Fix possible memory leak in bcm_sf2_mdio_register()

In bcm_sf2_mdio_register(), the class_find_device() will call get_device()
to increment reference count for priv->master_mii_bus->dev if
of_mdio_find_bus() succeeds. If mdiobus_alloc() or mdiobus_register()
fails, it will call get_device() twice without decrement reference count
for the device. And it is the same if bcm_sf2_mdio_register() succeeds but
fails in bcm_sf2_sw_probe(), or if bcm_sf2_sw_probe() succeeds. If the
reference count has not decremented to zero, the dev related resource will
not be freed.

So remove the get_device() in bcm_sf2_mdio_register(), and call
put_device() if mdiobus_alloc() or mdiobus_register() fails and in
bcm_sf2_mdio_unregister() to solve the issue.

And as Simon suggested, unwind from errors for bcm_sf2_mdio_register() and
just return 0 if it succeeds to make it cleaner.

Fixes: 461cd1b03e32 ("net: dsa: bcm_sf2: Register our slave MDIO bus")
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Suggested-by: Simon Horman <horms@kernel.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/r/20231011032419.2423290-1-ruanjinjie@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoMerge branch 'selftests-fib_tests-fixes-for-multipath-list-receive-tests'
Jakub Kicinski [Fri, 13 Oct 2023 21:28:24 +0000 (14:28 -0700)]
Merge branch 'selftests-fib_tests-fixes-for-multipath-list-receive-tests'

Ido Schimmel says:

====================
selftests: fib_tests: Fixes for multipath list receive tests

Fix two issues in recently added FIB multipath list receive tests.
====================

Link: https://lore.kernel.org/r/20231010132113.3014691-1-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoselftests: fib_tests: Count all trace point invocations
Ido Schimmel [Tue, 10 Oct 2023 13:21:13 +0000 (16:21 +0300)]
selftests: fib_tests: Count all trace point invocations

The tests rely on the IPv{4,6} FIB trace points being triggered once for
each forwarded packet. If receive processing is deferred to the
ksoftirqd task these invocations will not be counted and the tests will
fail. Fix by specifying the '-a' flag to avoid perf from filtering on
the mausezahn task.

Before:

 # ./fib_tests.sh -t ipv4_mpath_list

 IPv4 multipath list receive tests
     TEST: Multipath route hit ratio (.68)                               [FAIL]

 # ./fib_tests.sh -t ipv6_mpath_list

 IPv6 multipath list receive tests
     TEST: Multipath route hit ratio (.27)                               [FAIL]

After:

 # ./fib_tests.sh -t ipv4_mpath_list

 IPv4 multipath list receive tests
     TEST: Multipath route hit ratio (1.00)                              [ OK ]

 # ./fib_tests.sh -t ipv6_mpath_list

 IPv6 multipath list receive tests
     TEST: Multipath route hit ratio (.99)                               [ OK ]

Fixes: 8ae9efb859c0 ("selftests: fib_tests: Add multipath list receive tests")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/netdev/202309191658.c00d8b8-oliver.sang@intel.com/
Tested-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Tested-by: Sriram Yagnaraman <sriram.yagnaraman@est.tech>
Link: https://lore.kernel.org/r/20231010132113.3014691-3-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agoselftests: fib_tests: Disable RP filter in multipath list receive test
Ido Schimmel [Tue, 10 Oct 2023 13:21:12 +0000 (16:21 +0300)]
selftests: fib_tests: Disable RP filter in multipath list receive test

The test relies on the fib:fib_table_lookup trace point being triggered
once for each forwarded packet. If RP filter is not disabled, the trace
point will be triggered twice for each packet (for source validation and
forwarding), potentially masking actual bugs. Fix by explicitly
disabling RP filter.

Before:

 # ./fib_tests.sh -t ipv4_mpath_list

 IPv4 multipath list receive tests
     TEST: Multipath route hit ratio (1.99)                              [ OK ]

After:

 # ./fib_tests.sh -t ipv4_mpath_list

 IPv4 multipath list receive tests
     TEST: Multipath route hit ratio (.99)                               [ OK ]

Fixes: 8ae9efb859c0 ("selftests: fib_tests: Add multipath list receive tests")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/netdev/202309191658.c00d8b8-oliver.sang@intel.com/
Tested-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Tested-by: Sriram Yagnaraman <sriram.yagnaraman@est.tech>
Link: https://lore.kernel.org/r/20231010132113.3014691-2-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agotcp: Fix listen() warning with v4-mapped-v6 address.
Kuniyuki Iwashima [Tue, 10 Oct 2023 01:38:14 +0000 (18:38 -0700)]
tcp: Fix listen() warning with v4-mapped-v6 address.

syzbot reported a warning [0] introduced by commit c48ef9c4aed3 ("tcp: Fix
bind() regression for v4-mapped-v6 non-wildcard address.").

After the cited commit, a v4 socket's address matches the corresponding
v4-mapped-v6 tb2 in inet_bind2_bucket_match_addr(), not vice versa.

During X.X.X.X -> ::ffff:X.X.X.X order bind()s, the second bind() uses
bhash and conflicts properly without checking bhash2 so that we need not
check if a v4-mapped-v6 sk matches the corresponding v4 address tb2 in
inet_bind2_bucket_match_addr().  However, the repro shows that we need
to check that in a no-conflict case.

The repro bind()s two sockets to the 2-tuples using SO_REUSEPORT and calls
listen() for the first socket:

  from socket import *

  s1 = socket()
  s1.setsockopt(SOL_SOCKET, SO_REUSEPORT, 1)
  s1.bind(('127.0.0.1', 0))

  s2 = socket(AF_INET6)
  s2.setsockopt(SOL_SOCKET, SO_REUSEPORT, 1)
  s2.bind(('::ffff:127.0.0.1', s1.getsockname()[1]))

  s1.listen()

The second socket should belong to the first socket's tb2, but the second
bind() creates another tb2 bucket because inet_bind2_bucket_find() returns
NULL in inet_csk_get_port() as the v4-mapped-v6 sk does not match the
corresponding v4 address tb2.

  bhash2[] -> tb2(::ffff:X.X.X.X) -> tb2(X.X.X.X)

Then, listen() for the first socket calls inet_csk_get_port(), where the
v4 address matches the v4-mapped-v6 tb2 and WARN_ON() is triggered.

To avoid that, we need to check if v4-mapped-v6 sk address matches with
the corresponding v4 address tb2 in inet_bind2_bucket_match().

The same checks are needed in inet_bind2_bucket_addr_match() too, so we
can move all checks there and call it from inet_bind2_bucket_match().

Note that now tb->family is just an address family of tb->(v6_)?rcv_saddr
and not of sockets in the bucket.  This could be refactored later by
defining tb->rcv_saddr as tb->v6_rcv_saddr.s6_addr32[3] and prepending
::ffff: when creating v4 tb2.

[0]:
WARNING: CPU: 0 PID: 5049 at net/ipv4/inet_connection_sock.c:587 inet_csk_get_port+0xf96/0x2350 net/ipv4/inet_connection_sock.c:587
Modules linked in:
CPU: 0 PID: 5049 Comm: syz-executor288 Not tainted 6.6.0-rc2-syzkaller-00018-g2cf0f7156238 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/04/2023
RIP: 0010:inet_csk_get_port+0xf96/0x2350 net/ipv4/inet_connection_sock.c:587
Code: 7c 24 08 e8 4c b6 8a 01 31 d2 be 88 01 00 00 48 c7 c7 e0 94 ae 8b e8 59 2e a3 f8 2e 2e 2e 31 c0 e9 04 fe ff ff e8 ca 88 d0 f8 <0f> 0b e9 0f f9 ff ff e8 be 88 d0 f8 49 8d 7e 48 e8 65 ca 5a 00 31
RSP: 0018:ffffc90003abfbf0 EFLAGS: 00010293
RAX: 0000000000000000 RBX: ffff888026429100 RCX: 0000000000000000
RDX: ffff88807edcbb80 RSI: ffffffff88b73d66 RDI: ffff888026c49f38
RBP: ffff888026c49f30 R08: 0000000000000005 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000000000 R12: ffffffff9260f200
R13: ffff888026c49880 R14: 0000000000000000 R15: ffff888026429100
FS:  00005555557d5380(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000045ad50 CR3: 0000000025754000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 <TASK>
 inet_csk_listen_start+0x155/0x360 net/ipv4/inet_connection_sock.c:1256
 __inet_listen_sk+0x1b8/0x5c0 net/ipv4/af_inet.c:217
 inet_listen+0x93/0xd0 net/ipv4/af_inet.c:239
 __sys_listen+0x194/0x270 net/socket.c:1866
 __do_sys_listen net/socket.c:1875 [inline]
 __se_sys_listen net/socket.c:1873 [inline]
 __x64_sys_listen+0x53/0x80 net/socket.c:1873
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f3a5bce3af9
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 c1 17 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffc1a1c79e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000032
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f3a5bce3af9
RDX: 00007f3a5bce3af9 RSI: 0000000000000000 RDI: 0000000000000003
RBP: 00007f3a5bd565f0 R08: 0000000000000006 R09: 0000000000000006
R10: 0000000000000006 R11: 0000000000000246 R12: 0000000000000001
R13: 431bde82d7b634db R14: 0000000000000001 R15: 0000000000000001
 </TASK>

Fixes: c48ef9c4aed3 ("tcp: Fix bind() regression for v4-mapped-v6 non-wildcard address.")
Reported-by: syzbot+71e724675ba3958edb31@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=71e724675ba3958edb31
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20231010013814.70571-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
12 months agobonding: Return pointer to data after pull on skb
Jiri Wiesner [Tue, 10 Oct 2023 16:39:33 +0000 (18:39 +0200)]
bonding: Return pointer to data after pull on skb

Since 429e3d123d9a ("bonding: Fix extraction of ports from the packet
headers"), header offsets used to compute a hash in bond_xmit_hash() are
relative to skb->data and not skb->head. If the tail of the header buffer
of an skb really needs to be advanced and the operation is successful, the
pointer to the data must be returned (and not a pointer to the head of the
buffer).

Fixes: 429e3d123d9a ("bonding: Fix extraction of ports from the packet headers")
Signed-off-by: Jiri Wiesner <jwiesner@suse.de>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 months agoMerge tag 'net-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 12 Oct 2023 20:07:00 +0000 (13:07 -0700)]
Merge tag 'net-6.6-rc6' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Including fixes from CAN and BPF.

  We have a regression in TC currently under investigation, otherwise
  the things that stand off most are probably the TCP and AF_PACKET
  fixes, with both issues coming from 6.5.

  Previous releases - regressions:

   - af_packet: fix fortified memcpy() without flex array.

   - tcp: fix crashes trying to free half-baked MTU probes

   - xdp: fix zero-size allocation warning in xskq_create()

   - can: sja1000: always restart the tx queue after an overrun

   - eth: mlx5e: again mutually exclude RX-FCS and RX-port-timestamp

   - eth: nfp: avoid rmmod nfp crash issues

   - eth: octeontx2-pf: fix page pool frag allocation warning

  Previous releases - always broken:

   - mctp: perform route lookups under a RCU read-side lock

   - bpf: s390: fix clobbering the caller's backchain in the trampoline

   - phy: lynx-28g: cancel the CDR check work item on the remove path

   - dsa: qca8k: fix qca8k driver for Turris 1.x

   - eth: ravb: fix use-after-free issue in ravb_tx_timeout_work()

   - eth: ixgbe: fix crash with empty VF macvlan list"

* tag 'net-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (54 commits)
  rswitch: Fix imbalance phy_power_off() calling
  rswitch: Fix renesas_eth_sw_remove() implementation
  octeontx2-pf: Fix page pool frag allocation warning
  nfc: nci: assert requested protocol is valid
  af_packet: Fix fortified memcpy() without flex array.
  net: tcp: fix crashes trying to free half-baked MTU probes
  net/smc: Fix pos miscalculation in statistics
  nfp: flower: avoid rmmod nfp crash issues
  net: usb: dm9601: fix uninitialized variable use in dm9601_mdio_read
  ethtool: Fix mod state of verbose no_mask bitset
  net: nfc: fix races in nfc_llcp_sock_get() and nfc_llcp_sock_get_sn()
  mctp: perform route lookups under a RCU read-side lock
  net: skbuff: fix kernel-doc typos
  s390/bpf: Fix unwinding past the trampoline
  s390/bpf: Fix clobbering the caller's backchain in the trampoline
  net/mlx5e: Again mutually exclude RX-FCS and RX-port-timestamp
  net/smc: Fix dependency of SMC on ISM
  ixgbe: fix crash with empty VF macvlan list
  net/mlx5e: macsec: use update_pn flag instead of PN comparation
  net: phy: mscc: macsec: reject PN update requests
  ...

13 months agoMerge tag 'soc-fixes-6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Thu, 12 Oct 2023 18:52:23 +0000 (11:52 -0700)]
Merge tag 'soc-fixes-6.6-2' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "AngeloGioacchino Del Regno is stepping in as co-maintainer for the
  MediaTek SoC platform and starts by sending some dts fixes for the
  mt8195 platform that had been pending for a while.

  On the ixp4xx platform, Krzysztof Halasa steps down as co-maintainer,
  reflecting that Linus Walleij has been handling this on his own for
  the past few years.

  Generic RISC-V kernels are now marked as incompatible with the RZ/Five
  platform that requires custom hacks both for managing its DMA bounce
  buffers and for addressing low virtual memory.

 Finally, there is one bugfix for the AMDTEE firmware driver to prevent
 a use-after-free bug"

* tag 'soc-fixes-6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
  IXP4xx MAINTAINERS entries
  arm64: dts: mediatek: mt8195: Set DSU PMU status to fail
  arm64: dts: mediatek: fix t-phy unit name
  arm64: dts: mediatek: mt8195-demo: update and reorder reserved memory regions
  arm64: dts: mediatek: mt8195-demo: fix the memory size to 8GB
  MAINTAINERS: Add Angelo as MediaTek SoC co-maintainer
  soc: renesas: Make ARCH_R9A07G043 (riscv version) depend on NONPORTABLE
  tee: amdtee: fix use-after-free vulnerability in amdtee_close_session

13 months agoMerge tag 'pmdomain-v6.6-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 12 Oct 2023 18:34:23 +0000 (11:34 -0700)]
Merge tag 'pmdomain-v6.6-rc1-2' of git://git./linux/kernel/git/ulfh/linux-pm

Pull pmdomain fix from Ulf Hansson:

 - imx: scu-pd: Correct the DMA2 channel

* tag 'pmdomain-v6.6-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm:
  pmdomain: imx: scu-pd: correct DMA2 channel

13 months agonet/mlx5e: Fix VF representors reporting zero counters to "ip -s" command
Amir Tzin [Mon, 4 Sep 2023 15:26:47 +0000 (18:26 +0300)]
net/mlx5e: Fix VF representors reporting zero counters to "ip -s" command

Although vf_vport entry of struct mlx5e_stats is never updated, its
values are mistakenly copied to the caller structure in the VF
representor .ndo_get_stat_64 callback mlx5e_rep_get_stats(). Remove
redundant entry and use the updated one, rep_stats, instead.

Fixes: 64b68e369649 ("net/mlx5: Refactor and expand rep vport stat group")
Reviewed-by: Patrisious Haddad <phaddad@nvidia.com>
Signed-off-by: Amir Tzin <amirtz@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5e: Don't offload internal port if filter device is out device
Jianbo Liu [Tue, 12 Sep 2023 02:28:47 +0000 (02:28 +0000)]
net/mlx5e: Don't offload internal port if filter device is out device

In the cited commit, if the routing device is ovs internal port, the
out device is set to uplink, and packets go out after encapsulation.

If filter device is uplink, it can trigger the following syndrome:
mlx5_core 0000:08:00.0: mlx5_cmd_out_err:803:(pid 3966): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0xcdb051), err(-22)

Fix this issue by not offloading internal port if filter device is out
device. In this case, packets are not forwarded to the root table to
be processed, the termination table is used instead to forward them
from uplink to uplink.

Fixes: 100ad4e2d758 ("net/mlx5e: Offload internal port as encap route device")
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5e: Take RTNL lock before triggering netdev notifiers
Lama Kayal [Tue, 12 Sep 2023 07:06:24 +0000 (10:06 +0300)]
net/mlx5e: Take RTNL lock before triggering netdev notifiers

Hold RTNL lock when calling xdp_set_features() with a registered netdev,
as the call triggers the netdev notifiers. This could happen when
switching from nic profile to uplink representor for example.

Similar logic which fixed a similar scenario was previously introduced in
the following commit:
commit 72cc65497065 net/mlx5e: Take RTNL lock when needed before calling
xdp_set_features().

This fixes the following assertion and warning call trace:

RTNL: assertion failed at net/core/dev.c (1961)
WARNING: CPU: 13 PID: 2529 at net/core/dev.c:1961
call_netdevice_notifiers_info+0x7c/0x80
Modules linked in: rpcrdma rdma_ucm ib_iser libiscsi
scsi_transport_iscsi ib_umad rdma_cm ib_ipoib iw_cm ib_cm mlx5_ib
ib_uverbs ib_core xt_conntrack xt_MASQUERADE nf_conntrack_netlink
nfnetlink xt_addrtype iptable_nat nf_nat br_netfilter rpcsec_gss_krb5
auth_rpcgss oid_registry overlay mlx5_core zram zsmalloc fuse
CPU: 13 PID: 2529 Comm: devlink Not tainted
6.5.0_for_upstream_min_debug_2023_09_07_20_04 #1
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS
rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
RIP: 0010:call_netdevice_notifiers_info+0x7c/0x80
Code: 8f ff 80 3d 77 0d 16 01 00 75 c5 ba a9 07 00 00 48
c7 c6 c4 bb 0d 82 48 c7 c7 18 c8 06 82 c6 05 5b 0d 16 01 01 e8 44 f6 8c
ff <0f> 0b eb a2 0f 1f 44 00 00 55 48 89 e5 41 54 48 83 e4 f0 48 83 ec
RSP: 0018:ffff88819930f7f0 EFLAGS: 00010282
RAX: 0000000000000000 RBX: ffffffff8309f740 RCX: 0000000000000027
RDX: ffff88885fb5b5c8 RSI: 0000000000000001 RDI: ffff88885fb5b5c0
RBP: 0000000000000028 R08: ffff88887ffabaa8 R09: 0000000000000003
R10: ffff88887fecbac0 R11: ffff88887ff7bac0 R12: ffff88819930f810
R13: ffff88810b7fea40 R14: ffff8881154e8fd8 R15: ffff888107e881a0
FS:  00007f3ad248f800(0000) GS:ffff88885fb40000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000563b85f164e0 CR3: 0000000113b5c006 CR4: 0000000000370ea0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 <TASK>
 ? __warn+0x79/0x120
 ? call_netdevice_notifiers_info+0x7c/0x80
 ? report_bug+0x17c/0x190
 ? handle_bug+0x3c/0x60
 ? exc_invalid_op+0x14/0x70
 ? asm_exc_invalid_op+0x16/0x20
 ? call_netdevice_notifiers_info+0x7c/0x80
 call_netdevice_notifiers+0x2e/0x50
 mlx5e_set_xdp_feature+0x21/0x50 [mlx5_core]
 mlx5e_build_rep_params+0x97/0x130 [mlx5_core]
 mlx5e_init_ul_rep+0x9f/0x100 [mlx5_core]
 mlx5e_netdev_init_profile+0x76/0x110 [mlx5_core]
 mlx5e_netdev_attach_profile+0x1f/0x90 [mlx5_core]
 mlx5e_netdev_change_profile+0x92/0x160 [mlx5_core]
 mlx5e_vport_rep_load+0x329/0x4a0 [mlx5_core]
 mlx5_esw_offloads_rep_load+0x9e/0xf0 [mlx5_core]
 esw_offloads_enable+0x4bc/0xe90 [mlx5_core]
 mlx5_eswitch_enable_locked+0x3c8/0x570 [mlx5_core]
 ? kmalloc_trace+0x25/0x80
 mlx5_devlink_eswitch_mode_set+0x224/0x680 [mlx5_core]
 ? devlink_get_from_attrs_lock+0x9e/0x110
 devlink_nl_cmd_eswitch_set_doit+0x60/0xe0
 genl_family_rcv_msg_doit+0xd0/0x120
 genl_rcv_msg+0x180/0x2b0
 ? devlink_get_from_attrs_lock+0x110/0x110
 ? devlink_nl_cmd_eswitch_get_doit+0x290/0x290
 ? devlink_pernet_pre_exit+0xf0/0xf0
 ? genl_family_rcv_msg_dumpit+0xf0/0xf0
 netlink_rcv_skb+0x54/0x100
 genl_rcv+0x24/0x40
 netlink_unicast+0x1fc/0x2c0
 netlink_sendmsg+0x232/0x4a0
 sock_sendmsg+0x38/0x60
 ? _copy_from_user+0x2a/0x60
 __sys_sendto+0x110/0x160
 ? handle_mm_fault+0x161/0x260
 ? do_user_addr_fault+0x276/0x620
 __x64_sys_sendto+0x20/0x30
 do_syscall_64+0x3d/0x90
 entry_SYSCALL_64_after_hwframe+0x46/0xb0
RIP: 0033:0x7f3ad231340a
Code: d8 64 89 02 48 c7 c0 ff ff ff ff eb b8 0f 1f 00 f3
0f 1e fa 41 89 ca 64 8b 04 25 18 00 00 00 85 c0 75 15 b8 2c 00 00 00 0f
05 <48> 3d 00 f0 ff ff 77 7e c3 0f 1f 44 00 00 41 54 48 83 ec 30 44 89
RSP: 002b:00007ffd70aad4b8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c
RAX: ffffffffffffffda RBX: 0000000000c36b00 RCX:00007f3ad231340a
RDX: 0000000000000038 RSI: 0000000000c36b00 RDI: 0000000000000003
RBP: 0000000000c36910 R08: 00007f3ad2625200 R09: 000000000000000c
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001
 </TASK>
---[ end trace 0000000000000000 ]---
------------[ cut here ]------------

Fixes: 4d5ab0ad964d ("net/mlx5e: take into account device reconfiguration for xdp_features flag")
Signed-off-by: Lama Kayal <lkayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5e: XDP, Fix XDP_REDIRECT mpwqe page fragment leaks on shutdown
Dragos Tatulea [Mon, 25 Sep 2023 14:50:18 +0000 (17:50 +0300)]
net/mlx5e: XDP, Fix XDP_REDIRECT mpwqe page fragment leaks on shutdown

When mlx5e_xdp_xmit is called without the XDP_XMIT_FLUSH set it is
possible that it leaves a mpwqe session open. That is ok during runtime:
the session will be closed on the next call to mlx5e_xdp_xmit. But
having a mpwqe session still open at XDP sq close time is problematic:
the pc counter is not updated before flushing the contents of the
xdpi_fifo. This results in leaking page fragments.

The fix is to always close the mpwqe session at the end of
mlx5e_xdp_xmit, regardless of the XDP_XMIT_FLUSH flag being set or not.

Fixes: 5e0d2eef771e ("net/mlx5e: XDP, Support Enhanced Multi-Packet TX WQE")
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5e: RX, Fix page_pool allocation failure recovery for legacy rq
Dragos Tatulea [Fri, 29 Sep 2023 14:31:49 +0000 (17:31 +0300)]
net/mlx5e: RX, Fix page_pool allocation failure recovery for legacy rq

When a page allocation fails during refill in mlx5e_refill_rx_wqes, the
page will be released again on the next refill call. This triggers the
page_pool negative page fragment count warning below:

 [  338.326070] WARNING: CPU: 4 PID: 0 at include/net/page_pool/helpers.h:130 mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
  ...
 [  338.328993] RIP: 0010:mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [  338.329094] Call Trace:
 [  338.329097]  <IRQ>
 [  338.329100]  ? __warn+0x7d/0x120
 [  338.329105]  ? mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [  338.329173]  ? report_bug+0x155/0x180
 [  338.329179]  ? handle_bug+0x3c/0x60
 [  338.329183]  ? exc_invalid_op+0x13/0x60
 [  338.329187]  ? asm_exc_invalid_op+0x16/0x20
 [  338.329192]  ? mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [  338.329259]  mlx5e_post_rx_wqes+0x210/0x5a0 [mlx5_core]
 [  338.329327]  ? mlx5e_poll_rx_cq+0x88/0x6f0 [mlx5_core]
 [  338.329394]  mlx5e_napi_poll+0x127/0x6b0 [mlx5_core]
 [  338.329461]  __napi_poll+0x25/0x1a0
 [  338.329465]  net_rx_action+0x28a/0x300
 [  338.329468]  __do_softirq+0xcd/0x279
 [  338.329473]  irq_exit_rcu+0x6a/0x90
 [  338.329477]  common_interrupt+0x82/0xa0
 [  338.329482]  </IRQ>

This patch fixes the legacy rq case by releasing all allocated fragments
and then setting the skip flag on all released fragments. It is
important to note that the number of released fragments will be higher
than the number of allocated fragments when an allocation error occurs.

Fixes: 3f93f82988bc ("net/mlx5e: RX, Defer page release in legacy rq for better recycling")
Tested-by: Chris Mason <clm@fb.com>
Reported-by: Chris Mason <clm@fb.com>
Closes: https://lore.kernel.org/netdev/117FF31A-7BE0-4050-B2BB-E41F224FF72F@meta.com
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5e: RX, Fix page_pool allocation failure recovery for striding rq
Dragos Tatulea [Mon, 2 Oct 2023 11:05:29 +0000 (14:05 +0300)]
net/mlx5e: RX, Fix page_pool allocation failure recovery for striding rq

When a page allocation fails during refill in mlx5e_post_rx_mpwqes, the
page will be released again on the next refill call. This triggers the
page_pool negative page fragment count warning below:

 [ 2436.447717] WARNING: CPU: 1 PID: 2419 at include/net/page_pool/helpers.h:130 mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 ...
 [ 2436.447895] RIP: 0010:mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [ 2436.447991] Call Trace:
 [ 2436.447975]  mlx5e_post_rx_mpwqes+0x1d5/0xcf0 [mlx5_core]
 [ 2436.447994]  <IRQ>
 [ 2436.447996]  ? __warn+0x7d/0x120
 [ 2436.448009]  ? mlx5e_handle_rx_cqe_mpwrq+0x109/0x1d0 [mlx5_core]
 [ 2436.448002]  ? mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [ 2436.448044]  ? mlx5e_poll_rx_cq+0x87/0x6e0 [mlx5_core]
 [ 2436.448061]  ? report_bug+0x155/0x180
 [ 2436.448065]  ? handle_bug+0x36/0x70
 [ 2436.448067]  ? exc_invalid_op+0x13/0x60
 [ 2436.448070]  ? asm_exc_invalid_op+0x16/0x20
 [ 2436.448079]  mlx5e_napi_poll+0x122/0x6b0 [mlx5_core]
 [ 2436.448077]  ? mlx5e_page_release_fragmented.isra.0+0x42/0x50 [mlx5_core]
 [ 2436.448113]  ? generic_exec_single+0x35/0x100
 [ 2436.448117]  __napi_poll+0x25/0x1a0
 [ 2436.448120]  net_rx_action+0x28a/0x300
 [ 2436.448122]  __do_softirq+0xcd/0x279
 [ 2436.448126]  irq_exit_rcu+0x6a/0x90
 [ 2436.448128]  sysvec_apic_timer_interrupt+0x6e/0x90
 [ 2436.448130]  </IRQ>

This patch fixes the striding rq case by setting the skip flag on all
the wqe pages that were expected to have new pages allocated.

Fixes: 4c2a13236807 ("net/mlx5e: RX, Defer page release in striding rq for better recycling")
Tested-by: Chris Mason <clm@fb.com>
Reported-by: Chris Mason <clm@fb.com>
Closes: https://lore.kernel.org/netdev/117FF31A-7BE0-4050-B2BB-E41F224FF72F@meta.com
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5: Handle fw tracer change ownership event based on MTRC
Maher Sanalla [Wed, 6 Sep 2023 18:48:30 +0000 (21:48 +0300)]
net/mlx5: Handle fw tracer change ownership event based on MTRC

Currently, whenever fw issues a change ownership event, the PF that owns
the fw tracer drops its ownership directly and the other PFs try to pick
up the ownership via what MTRC register suggests.

In some cases, driver releases the ownership of the tracer and reacquires
it later on. Whenever the driver releases ownership of the tracer, fw
issues a change ownership event. This event can be delayed and come after
driver has reacquired ownership of the tracer. Thus the late event will
trigger the tracer owner PF to release the ownership again and lead to a
scenario where no PF is owning the tracer.

To prevent the scenario described above, when handling a change
ownership event, do not drop ownership of the tracer directly, instead
read the fw MTRC register to retrieve the up-to-date owner of the tracer
and set it accordingly in driver level.

Fixes: f53aaa31cce7 ("net/mlx5: FW tracer, implement tracer logic")
Signed-off-by: Maher Sanalla <msanalla@nvidia.com>
Reviewed-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5: Bridge, fix peer entry ageing in LAG mode
Vlad Buslov [Wed, 9 Aug 2023 09:10:57 +0000 (11:10 +0200)]
net/mlx5: Bridge, fix peer entry ageing in LAG mode

With current implementation in single FDB LAG mode all packets are
processed by eswitch 0 rules. As such, 'peer' FDB entries receive the
packets for rules of other eswitches and are responsible for updating the
main entry by sending SWITCHDEV_FDB_ADD_TO_BRIDGE notification from their
background update wq task. However, this introduces a race condition when
non-zero eswitch instance decides to delete a FDB entry, sends
SWITCHDEV_FDB_DEL_TO_BRIDGE notification, but another eswitch's update task
refreshes the same entry concurrently while its async delete work is still
pending on the workque. In such case another SWITCHDEV_FDB_ADD_TO_BRIDGE
event may be generated and entry will remain stuck in FDB marked as
'offloaded' since no more SWITCHDEV_FDB_DEL_TO_BRIDGE notifications are
sent for deleting the peer entries.

Fix the issue by synchronously marking deleted entries with
MLX5_ESW_BRIDGE_FLAG_DELETED flag and skipping them in background update
job.

Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5: E-switch, register event handler before arming the event
Shay Drory [Sun, 27 Aug 2023 10:31:53 +0000 (13:31 +0300)]
net/mlx5: E-switch, register event handler before arming the event

Currently, mlx5 is registering event handler for vport context change
event some time after arming the event. this can lead to missing an
event, which will result in wrong rules in the FDB.
Hence, register the event handler before arming the event.

This solution is valid since FW is sending vport context change event
only on vports which SW armed, and SW arming the vport when enabling
it, which is done after the FDB has been created.

Fixes: 6933a9379559 ("net/mlx5: E-Switch, Use async events chain")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agonet/mlx5: Perform DMA operations in the right locations
Shay Drory [Tue, 8 Aug 2023 11:21:30 +0000 (14:21 +0300)]
net/mlx5: Perform DMA operations in the right locations

The cited patch change mlx5 driver so that during probe DMA
operations were performed before pci_enable_device(), and during
teardown DMA operations were performed after pci_disable_device().
DMA operations require PCI to be enabled. Hence, The above leads to
the following oops in PPC systems[1].

On s390x systems, as reported by Niklas Schnelle, this is a problem
because mlx5_pci_init() is where the DMA and coherent mask is set but
mlx5_cmd_init() already does a dma_alloc_coherent(). Thus a DMA
allocation is done during probe before the correct mask is set. This
causes probe to fail initialization of the cmdif SW structs on s390x
after that is converted to the common dma-iommu code. This is because on
s390x DMA addresses below 4 GiB are reserved on current machines and
unlike the old s390x specific DMA API implementation common code
enforces DMA masks.

Fix it by performing the DMA operations during probe after
pci_enable_device() and after the dma mask is set,
and during teardown before pci_disable_device().

[1]
Oops: Kernel access of bad area, sig: 11 [#1]
LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
Modules linked in: xt_MASQUERADE nf_conntrack_netlink
nfnetlink xfrm_user iptable_nat xt_addrtype xt_conntrack nf_nat
nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 netconsole rpcsec_gss_krb5
auth_rpcgss oid_registry overlay rpcrdma rdma_ucm ib_iser ib_umad
rdma_cm ib_ipoib iw_cm libiscsi scsi_transport_iscsi ib_cm ib_uverbs
ib_core mlx5_core(-) ptp pps_core fuse vmx_crypto crc32c_vpmsum [last
unloaded: mlx5_ib]
CPU: 1 PID: 8937 Comm: modprobe Not tainted 6.5.0-rc3_for_upstream_min_debug_2023_07_31_16_02 #1
Hardware name: IBM pSeries (emulated by qemu) POWER9 (raw) 0x4e1202 0xf000005 of:SLOF,HEAD hv:linux,kvm pSeries
NIP:  c000000000423388 LR: c0000000001e733c CTR: c0000000001e4720
REGS: c0000000055636d0 TRAP: 0380   Not tainted (6.5.0-rc3_for_upstream_min_debug_2023_07_31_16_02)
MSR:  8000000000009033  CR: 24008884  XER: 20040000
CFAR: c0000000001e7338 IRQMASK: 0
NIP [c000000000423388] __free_pages+0x28/0x160
LR [c0000000001e733c] dma_direct_free+0xac/0x190
Call Trace:
[c000000005563970] [5deadbeef0000100] 0x5deadbeef0000100 (unreliable)
[c0000000055639b0] [c0000000003d46cc] kfree+0x7c/0x150
[c000000005563a40] [c0000000001e47c8] dma_free_attrs+0xa8/0x1a0
[c000000005563aa0] [c008000000d0064c] mlx5_cmd_cleanup+0xa4/0x100 [mlx5_core]
[c000000005563ad0] [c008000000cf629c] mlx5_mdev_uninit+0xf4/0x140 [mlx5_core]
[c000000005563b00] [c008000000cf6448] remove_one+0x160/0x1d0 [mlx5_core]
[c000000005563b40] [c000000000958540] pci_device_remove+0x60/0x110
[c000000005563b80] [c000000000a35e80] device_remove+0x70/0xd0
[c000000005563bb0] [c000000000a37a38] device_release_driver_internal+0x2a8/0x330
[c000000005563c00] [c000000000a37b8c] driver_detach+0x8c/0x160
[c000000005563c40] [c000000000a35350] bus_remove_driver+0x90/0x110
[c000000005563c80] [c000000000a38948] driver_unregister+0x48/0x90
[c000000005563cf0] [c000000000957e38] pci_unregister_driver+0x38/0x150
[c000000005563d40] [c008000000eb6140] mlx5_cleanup+0x38/0x90 [mlx5_core]

Fixes: 06cd555f73ca ("net/mlx5: split mlx5_cmd_init() to probe and reload routines")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
Tested-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
13 months agoMerge tag 'pinctrl-v6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Thu, 12 Oct 2023 17:48:19 +0000 (10:48 -0700)]
Merge tag 'pinctrl-v6.6-2' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pin control fixes from Linus Walleij:
 "Some pin control fixes for v6.6 which have been stacking up in my
  tree.

  Dmitry's fix to some locking in the core is the most substantial, that
  was a really neat fix.

  The rest is the usual assorted spray of minor driver fixes.

   - Drop some minor code causing warnings in the Lantiq driver

   - Fix out of bounds write in the Nuvoton driver

   - Fix lost IRQs with CONFIG_PM in the Starfive driver

   - Fix a locking issue in find_pinctrl()

   - Revert a regressive Tegra debug patch

   - Fix the Renesas RZN1 pin muxing"

* tag 'pinctrl-v6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: renesas: rzn1: Enable missing PINMUX
  Revert "pinctrl: tegra: Add support to display pin function"
  pinctrl: avoid unsafe code pattern in find_pinctrl()
  pinctrl: starfive: jh7110: Add system pm ops to save and restore context
  pinctrl: starfive: jh7110: Fix failure to set irq after CONFIG_PM is enabled
  pinctrl: nuvoton: wpcm450: fix out of bounds write
  pinctrl: lantiq: Remove unsued declaration ltq_pinctrl_unregister()

13 months agoIXP4xx MAINTAINERS entries
Krzysztof Hałasa [Wed, 11 Oct 2023 05:06:29 +0000 (07:06 +0200)]
IXP4xx MAINTAINERS entries

Update MAINTAINERS entries for Intel IXP4xx SoCs.

Linus has been handling all IXP4xx stuff since 2019 or so.

Signed-off-by: Krzysztof Hałasa <khalasa@piap.pl>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Deepak Saxena <dsaxena@plexity.net>
Link: https://lore.kernel.org/r/m3ttqxu4ru.fsf@t19.piap.pl
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
13 months agoMerge branch 'rswitch-fix-issues-on-specific-conditions'
Paolo Abeni [Thu, 12 Oct 2023 09:22:24 +0000 (11:22 +0200)]
Merge branch 'rswitch-fix-issues-on-specific-conditions'

Yoshihiro Shimoda says:

====================
rswitch: Fix issues on specific conditions

This patch series fix some issues of rswitch driver on specific
condtions.
====================

Link: https://lore.kernel.org/r/20231010124858.183891-1-yoshihiro.shimoda.uh@renesas.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agorswitch: Fix imbalance phy_power_off() calling
Yoshihiro Shimoda [Tue, 10 Oct 2023 12:48:58 +0000 (21:48 +0900)]
rswitch: Fix imbalance phy_power_off() calling

The phy_power_off() should not be called if phy_power_on() failed.
So, add a condition .power_count before calls phy_power_off().

Fixes: 5cb630925b49 ("net: renesas: rswitch: Add phy_power_{on,off}() calling")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agorswitch: Fix renesas_eth_sw_remove() implementation
Yoshihiro Shimoda [Tue, 10 Oct 2023 12:48:57 +0000 (21:48 +0900)]
rswitch: Fix renesas_eth_sw_remove() implementation

Fix functions calling order and a condition in renesas_eth_sw_remove().
Otherwise, kernel NULL pointer dereference happens from phy_stop() if
a net device opens.

Fixes: 3590918b5d07 ("net: ethernet: renesas: Add support for "Ethernet Switch"")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonetfilter: nft_payload: fix wrong mac header matching
Florian Westphal [Sun, 8 Oct 2023 17:36:53 +0000 (19:36 +0200)]
netfilter: nft_payload: fix wrong mac header matching

mcast packets get looped back to the local machine.
Such packets have a 0-length mac header, we should treat
this like "mac header not set" and abort rule evaluation.

As-is, we just copy data from the network header instead.

Fixes: 96518518cc41 ("netfilter: add nftables")
Reported-by: Blažej Krajňák <krajnak@levonet.sk>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonf_tables: fix NULL pointer dereference in nft_expr_inner_parse()
Xingyuan Mo [Mon, 9 Oct 2023 10:36:15 +0000 (18:36 +0800)]
nf_tables: fix NULL pointer dereference in nft_expr_inner_parse()

We should check whether the NFTA_EXPR_NAME netlink attribute is present
before accessing it, otherwise a null pointer deference error will occur.

Call Trace:
 <TASK>
 dump_stack_lvl+0x4f/0x90
 print_report+0x3f0/0x620
 kasan_report+0xcd/0x110
 __asan_load2+0x7d/0xa0
 nla_strcmp+0x2f/0x90
 __nft_expr_type_get+0x41/0xb0
 nft_expr_inner_parse+0xe3/0x200
 nft_inner_init+0x1be/0x2e0
 nf_tables_newrule+0x813/0x1230
 nfnetlink_rcv_batch+0xec3/0x1170
 nfnetlink_rcv+0x1e4/0x220
 netlink_unicast+0x34e/0x4b0
 netlink_sendmsg+0x45c/0x7e0
 __sys_sendto+0x355/0x370
 __x64_sys_sendto+0x84/0xa0
 do_syscall_64+0x3f/0x90
 entry_SYSCALL_64_after_hwframe+0x6e/0xd8

Fixes: 3a07327d10a0 ("netfilter: nft_inner: support for inner tunnel header matching")
Signed-off-by: Xingyuan Mo <hdthky0@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonf_tables: fix NULL pointer dereference in nft_inner_init()
Xingyuan Mo [Mon, 9 Oct 2023 10:36:14 +0000 (18:36 +0800)]
nf_tables: fix NULL pointer dereference in nft_inner_init()

We should check whether the NFTA_INNER_NUM netlink attribute is present
before accessing it, otherwise a null pointer deference error will occur.

Call Trace:
 dump_stack_lvl+0x4f/0x90
 print_report+0x3f0/0x620
 kasan_report+0xcd/0x110
 __asan_load4+0x84/0xa0
 nft_inner_init+0x128/0x2e0
 nf_tables_newrule+0x813/0x1230
 nfnetlink_rcv_batch+0xec3/0x1170
 nfnetlink_rcv+0x1e4/0x220
 netlink_unicast+0x34e/0x4b0
 netlink_sendmsg+0x45c/0x7e0
 __sys_sendto+0x355/0x370
 __x64_sys_sendto+0x84/0xa0
 do_syscall_64+0x3f/0x90
 entry_SYSCALL_64_after_hwframe+0x6e/0xd8

Fixes: 3a07327d10a0 ("netfilter: nft_inner: support for inner tunnel header matching")
Signed-off-by: Xingyuan Mo <hdthky0@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonetfilter: nf_tables: do not refresh timeout when resetting element
Pablo Neira Ayuso [Mon, 2 Oct 2023 09:57:42 +0000 (11:57 +0200)]
netfilter: nf_tables: do not refresh timeout when resetting element

The dump and reset command should not refresh the timeout, this command
is intended to allow users to list existing stateful objects and reset
them, element expiration should be refresh via transaction instead with
a specific command to achieve this, otherwise this is entering combo
semantics that will be hard to be undone later (eg. a user asking to
retrieve counters but _not_ requiring to refresh expiration).

Fixes: 079cd633219d ("netfilter: nf_tables: Introduce NFT_MSG_GETSETELEM_RESET")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonetfilter: nf_tables: Annotate struct nft_pipapo_match with __counted_by
Kees Cook [Tue, 3 Oct 2023 23:17:51 +0000 (16:17 -0700)]
netfilter: nf_tables: Annotate struct nft_pipapo_match with __counted_by

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

As found with Coccinelle[1], add __counted_by for struct nft_pipapo_match.

Cc: Pablo Neira Ayuso <pablo@netfilter.org>
Cc: Jozsef Kadlecsik <kadlec@netfilter.org>
Cc: Florian Westphal <fw@strlen.de>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: netfilter-devel@vger.kernel.org
Cc: coreteam@netfilter.org
Cc: netdev@vger.kernel.org
Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonetfilter: nfnetlink_log: silence bogus compiler warning
Florian Westphal [Thu, 5 Oct 2023 08:53:08 +0000 (10:53 +0200)]
netfilter: nfnetlink_log: silence bogus compiler warning

net/netfilter/nfnetlink_log.c:800:18: warning: variable 'ctinfo' is uninitialized

The warning is bogus, the variable is only used if ct is non-NULL and
always initialised in that case.  Init to 0 too to silence this.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202309100514.ndBFebXN-lkp@intel.com/
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agonetfilter: nf_tables: do not remove elements if set backend implements .abort
Pablo Neira Ayuso [Wed, 4 Oct 2023 11:12:58 +0000 (13:12 +0200)]
netfilter: nf_tables: do not remove elements if set backend implements .abort

pipapo set backend maintains two copies of the datastructure, removing
the elements from the copy that is going to be discarded slows down
the abort path significantly, from several minutes to few seconds after
this patch.

Fixes: 212ed75dc5fb ("netfilter: nf_tables: integrate pipapo into commit protocol")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
13 months agoocteontx2-pf: Fix page pool frag allocation warning
Ratheesh Kannoth [Tue, 10 Oct 2023 03:48:42 +0000 (09:18 +0530)]
octeontx2-pf: Fix page pool frag allocation warning

Since page pool param's "order" is set to 0, will result
in below warn message if interface is configured with higher
rx buffer size.

Steps to reproduce the issue.
1. devlink dev param set pci/0002:04:00.0 name receive_buffer_size \
   value 8196 cmode runtime
2. ifconfig eth0 up

[   19.901356] ------------[ cut here ]------------
[   19.901361] WARNING: CPU: 11 PID: 12331 at net/core/page_pool.c:567 page_pool_alloc_frag+0x3c/0x230
[   19.901449] pstate: 82401009 (Nzcv daif +PAN -UAO +TCO -DIT +SSBS BTYPE=--)
[   19.901451] pc : page_pool_alloc_frag+0x3c/0x230
[   19.901453] lr : __otx2_alloc_rbuf+0x60/0xbc [rvu_nicpf]
[   19.901460] sp : ffff80000f66b970
[   19.901461] x29: ffff80000f66b970 x28: 0000000000000000 x27: 0000000000000000
[   19.901464] x26: ffff800000d15b68 x25: ffff000195b5c080 x24: ffff0002a5a32dc0
[   19.901467] x23: ffff0001063c0878 x22: 0000000000000100 x21: 0000000000000000
[   19.901469] x20: 0000000000000000 x19: ffff00016f781000 x18: 0000000000000000
[   19.901472] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[   19.901474] x14: 0000000000000000 x13: ffff0005ffdc9c80 x12: 0000000000000000
[   19.901477] x11: ffff800009119a38 x10: 4c6ef2e3ba300519 x9 : ffff800000d13844
[   19.901479] x8 : ffff0002a5a33cc8 x7 : 0000000000000030 x6 : 0000000000000030
[   19.901482] x5 : 0000000000000005 x4 : 0000000000000000 x3 : 0000000000000a20
[   19.901484] x2 : 0000000000001080 x1 : ffff80000f66b9d4 x0 : 0000000000001000
[   19.901487] Call trace:
[   19.901488]  page_pool_alloc_frag+0x3c/0x230
[   19.901490]  __otx2_alloc_rbuf+0x60/0xbc [rvu_nicpf]
[   19.901494]  otx2_rq_aura_pool_init+0x1c4/0x240 [rvu_nicpf]
[   19.901498]  otx2_open+0x228/0xa70 [rvu_nicpf]
[   19.901501]  otx2vf_open+0x20/0xd0 [rvu_nicvf]
[   19.901504]  __dev_open+0x114/0x1d0
[   19.901507]  __dev_change_flags+0x194/0x210
[   19.901510]  dev_change_flags+0x2c/0x70
[   19.901512]  devinet_ioctl+0x3a4/0x6c4
[   19.901515]  inet_ioctl+0x228/0x240
[   19.901518]  sock_ioctl+0x2ac/0x480
[   19.901522]  __arm64_sys_ioctl+0x564/0xe50
[   19.901525]  invoke_syscall.constprop.0+0x58/0xf0
[   19.901529]  do_el0_svc+0x58/0x150
[   19.901531]  el0_svc+0x30/0x140
[   19.901533]  el0t_64_sync_handler+0xe8/0x114
[   19.901535]  el0t_64_sync+0x1a0/0x1a4
[   19.901537] ---[ end trace 678c0bf660ad8116 ]---

Fixes: b2e3406a38f0 ("octeontx2-pf: Add support for page pool")
Signed-off-by: Ratheesh Kannoth <rkannoth@marvell.com>
Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com>
Link: https://lore.kernel.org/r/20231010034842.3807816-1-rkannoth@marvell.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonfc: nci: assert requested protocol is valid
Jeremy Cline [Mon, 9 Oct 2023 20:00:54 +0000 (16:00 -0400)]
nfc: nci: assert requested protocol is valid

The protocol is used in a bit mask to determine if the protocol is
supported. Assert the provided protocol is less than the maximum
defined so it doesn't potentially perform a shift-out-of-bounds and
provide a clearer error for undefined protocols vs unsupported ones.

Fixes: 6a2968aaf50c ("NFC: basic NCI protocol implementation")
Reported-and-tested-by: syzbot+0839b78e119aae1fec78@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=0839b78e119aae1fec78
Signed-off-by: Jeremy Cline <jeremy@jcline.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20231009200054.82557-1-jeremy@jcline.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agoaf_packet: Fix fortified memcpy() without flex array.
Kuniyuki Iwashima [Mon, 9 Oct 2023 15:31:52 +0000 (08:31 -0700)]
af_packet: Fix fortified memcpy() without flex array.

Sergei Trofimovich reported a regression [0] caused by commit a0ade8404c3b
("af_packet: Fix warning of fortified memcpy() in packet_getname().").

It introduced a flex array sll_addr_flex in struct sockaddr_ll as a
union-ed member with sll_addr to work around the fortified memcpy() check.

However, a userspace program uses a struct that has struct sockaddr_ll in
the middle, where a flex array is illegal to exist.

  include/linux/if_packet.h:24:17: error: flexible array member 'sockaddr_ll::<unnamed union>::<unnamed struct>::sll_addr_flex' not at end of 'struct packet_info_t'
     24 |                 __DECLARE_FLEX_ARRAY(unsigned char, sll_addr_flex);
        |                 ^~~~~~~~~~~~~~~~~~~~

To fix the regression, let's go back to the first attempt [1] telling
memcpy() the actual size of the array.

Reported-by: Sergei Trofimovich <slyich@gmail.com>
Closes: https://github.com/NixOS/nixpkgs/pull/252587#issuecomment-1741733002 [0]
Link: https://lore.kernel.org/netdev/20230720004410.87588-3-kuniyu@amazon.com/
Fixes: a0ade8404c3b ("af_packet: Fix warning of fortified memcpy() in packet_getname().")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20231009153151.75688-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agopinctrl: renesas: rzn1: Enable missing PINMUX
Ralph Siemsen [Wed, 4 Oct 2023 20:00:08 +0000 (16:00 -0400)]
pinctrl: renesas: rzn1: Enable missing PINMUX

Enable pin muxing (eg. programmable function), so that the RZ/N1 GPIO
pins will be configured as specified by the pinmux in the DTS.

This used to be enabled implicitly via CONFIG_GENERIC_PINMUX_FUNCTIONS,
however that was removed, since the RZ/N1 driver does not call any of
the generic pinmux functions.

Fixes: 1308fb4e4eae14e6 ("pinctrl: rzn1: Do not select GENERIC_PIN{CTRL_GROUPS,MUX_FUNCTIONS}")
Signed-off-by: Ralph Siemsen <ralph.siemsen@linaro.org>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20231004200008.1306798-1-ralph.siemsen@linaro.org
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
13 months agonet: tcp: fix crashes trying to free half-baked MTU probes
Jakub Kicinski [Tue, 10 Oct 2023 17:36:51 +0000 (10:36 -0700)]
net: tcp: fix crashes trying to free half-baked MTU probes

tcp_stream_alloc_skb() initializes the skb to use tcp_tsorted_anchor
which is a union with the destructor. We need to clean that
TCP-iness up before freeing.

Fixes: 736013292e3c ("tcp: let tcp_mtu_probe() build headless packets")
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20231010173651.3990234-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoMerge tag 'ieee802154-for-net-2023-10-10' of git://git.kernel.org/pub/scm/linux/kerne...
Jakub Kicinski [Thu, 12 Oct 2023 00:22:15 +0000 (17:22 -0700)]
Merge tag 'ieee802154-for-net-2023-10-10' of git://git./linux/kernel/git/wpan/wpan

Stefan Schmidt says:

====================
pull-request: ieee802154 for net 2023-10-10

Just one small fix this time around.

Dinghao Liu fixed a potential use-after-free in the ca8210 driver probe
function.

* tag 'ieee802154-for-net-2023-10-10' of git://git.kernel.org/pub/scm/linux/kernel/git/wpan/wpan:
  ieee802154: ca8210: Fix a potential UAF in ca8210_probe
====================

Link: https://lore.kernel.org/r/20231010200943.82225-1-stefan@datenfreihafen.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoMerge tag 'fs_for_v6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/jack...
Linus Torvalds [Wed, 11 Oct 2023 21:21:15 +0000 (14:21 -0700)]
Merge tag 'fs_for_v6.6-rc6' of git://git./linux/kernel/git/jack/linux-fs

Pull quota regression fix from Jan Kara.

* tag 'fs_for_v6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  quota: Fix slow quotaoff

13 months agoMerge tag 'for-6.6-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Wed, 11 Oct 2023 20:58:32 +0000 (13:58 -0700)]
Merge tag 'for-6.6-rc5-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A revert of recent mount option parsing fix, this breaks mounts with
  security options.

  The second patch is a flexible array annotation"

* tag 'for-6.6-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: add __counted_by for struct btrfs_delayed_item and use struct_size()
  Revert "btrfs: reject unknown mount options early"

13 months agoMerge tag 'ata-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal...
Linus Torvalds [Wed, 11 Oct 2023 20:46:56 +0000 (13:46 -0700)]
Merge tag 'ata-6.6-rc6' of git://git./linux/kernel/git/dlemoal/libata

Pull ata fixes from Damien Le Moal:

 - Three fixes for the pata_parport driver to address a typo in the
   code, a missing operation implementation and port reset handling in
   the presence of slave devices (Ondrej)

 - Fix handling of ATAPI devices reset with the fit3 protocol driver of
   the pata_parport driver (Ondrej)

 - A follow up fix for the recent suspend/resume corrections to avoid
   attempting rescanning on resume the scsi device associated with an
   ata disk when the request queue of the scsi device is still suspended
   (in addition to not doing the rescan if the scsi device itself is
   still suspended) (me)

* tag 'ata-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata:
  scsi: Do not rescan devices with a suspended queue
  ata: pata_parport: fit3: implement IDE command set registers
  ata: pata_parport: add custom version of wait_after_reset
  ata: pata_parport: implement set_devctl
  ata: pata_parport: fix pata_parport_devchk

13 months agoMerge tag 'for-linus-2023101101' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 11 Oct 2023 20:27:44 +0000 (13:27 -0700)]
Merge tag 'for-linus-2023101101' of git://git./linux/kernel/git/hid/hid

Pull HID fixes from Benjamin Tissoires:

 - regression fix for i2c-hid when used on DT platforms (Johan Hovold)

 - kernel crash fix on removal of the Logitech USB receiver (Hans de
   Goede)

* tag 'for-linus-2023101101' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
  HID: logitech-hidpp: Fix kernel crash on receiver USB disconnect
  HID: i2c-hid: fix handling of unpopulated devices

13 months agoMerge tag 'printk-for-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/print...
Linus Torvalds [Wed, 11 Oct 2023 20:15:16 +0000 (13:15 -0700)]
Merge tag 'printk-for-6.6-rc6' of git://git./linux/kernel/git/printk/linux

Pull printk regression fix from Petr Mladek:

 - Avoid unnecessary wait and try to flush messages before checking
   pending ones

* tag 'printk-for-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux:
  printk: flush consoles before checking progress

13 months agoMerge branch 'rework/misc-cleanups' into for-linus
Petr Mladek [Wed, 11 Oct 2023 10:58:14 +0000 (12:58 +0200)]
Merge branch 'rework/misc-cleanups' into for-linus

13 months agobtrfs: add __counted_by for struct btrfs_delayed_item and use struct_size()
Gustavo A. R. Silva [Mon, 9 Oct 2023 20:44:54 +0000 (14:44 -0600)]
btrfs: add __counted_by for struct btrfs_delayed_item and use struct_size()

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

While there, use struct_size() helper, instead of the open-coded
version, to calculate the size for the allocation of the whole
flexible structure, including of course, the flexible-array member.

This code was found with the help of Coccinelle, and audited and
fixed manually.

Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
13 months agonet/smc: Fix pos miscalculation in statistics
Nils Hoppmann [Mon, 9 Oct 2023 14:40:48 +0000 (16:40 +0200)]
net/smc: Fix pos miscalculation in statistics

SMC_STAT_PAYLOAD_SUB(_smc_stats, _tech, key, _len, _rc) will calculate
wrong bucket positions for payloads of exactly 4096 bytes and
(1 << (m + 12)) bytes, with m == SMC_BUF_MAX - 1.

Intended bucket distribution:
Assume l == size of payload, m == SMC_BUF_MAX - 1.

Bucket 0                : 0 < l <= 2^13
Bucket n, 1 <= n <= m-1 : 2^(n+12) < l <= 2^(n+13)
Bucket m                : l > 2^(m+12)

Current solution:
_pos = fls64((l) >> 13)
[...]
_pos = (_pos < m) ? ((l == 1 << (_pos + 12)) ? _pos - 1 : _pos) : m

For l == 4096, _pos == -1, but should be _pos == 0.
For l == (1 << (m + 12)), _pos == m, but should be _pos == m - 1.

In order to avoid special treatment of these corner cases, the
calculation is adjusted. The new solution first subtracts the length by
one, and then calculates the correct bucket by shifting accordingly,
i.e. _pos = fls64((l - 1) >> 13), l > 0.
This not only fixes the issues named above, but also makes the whole
bucket assignment easier to follow.

Same is done for SMC_STAT_RMB_SIZE_SUB(_smc_stats, _tech, k, _len),
where the calculation of the bucket position is similar to the one
named above.

Fixes: e0e4b8fa5338 ("net/smc: Add SMC statistics support")
Suggested-by: Halil Pasic <pasic@linux.ibm.com>
Signed-off-by: Nils Hoppmann <niho@linux.ibm.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 months agonfp: flower: avoid rmmod nfp crash issues
Yanguo Li [Mon, 9 Oct 2023 11:21:55 +0000 (13:21 +0200)]
nfp: flower: avoid rmmod nfp crash issues

When there are CT table entries, and you rmmod nfp, the following
events can happen:

task1:
    nfp_net_pci_remove
          ↓
    nfp_flower_stop->(asynchronous)tcf_ct_flow_table_cleanup_work(3)
          ↓
    nfp_zone_table_entry_destroy(1)

task2:
    nfp_fl_ct_handle_nft_flow(2)

When the execution order is (1)->(2)->(3), it will crash. Therefore, in
the function nfp_fl_ct_del_flow, nf_flow_table_offload_del_cb needs to
be executed synchronously.

At the same time, in order to solve the deadlock problem and the problem
of rtnl_lock sometimes failing, replace rtnl_lock with the private
nfp_fl_lock.

Fixes: 7cc93d888df7 ("nfp: flower-ct: remove callback delete deadlock")
Cc: stable@vger.kernel.org
Signed-off-by: Yanguo Li <yanguo.li@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 months agonet: usb: dm9601: fix uninitialized variable use in dm9601_mdio_read
Javier Carrasco [Mon, 9 Oct 2023 22:26:14 +0000 (00:26 +0200)]
net: usb: dm9601: fix uninitialized variable use in dm9601_mdio_read

syzbot has found an uninit-value bug triggered by the dm9601 driver [1].

This error happens because the variable res is not updated if the call
to dm_read_shared_word returns an error. In this particular case -EPROTO
was returned and res stayed uninitialized.

This can be avoided by checking the return value of dm_read_shared_word
and propagating the error if the read operation failed.

[1] https://syzkaller.appspot.com/bug?extid=1f53a30781af65d2c955

Cc: stable@vger.kernel.org
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Reported-and-tested-by: syzbot+1f53a30781af65d2c955@syzkaller.appspotmail.com
Acked-by: Peter Korsgaard <peter@korsgaard.com>
Fixes: d0374f4f9c35cdfbee0 ("USB: Davicom DM9601 usbnet driver")
Link: https://lore.kernel.org/r/20231009-topic-dm9601_uninit_mdio_read-v2-1-f2fe39739b6c@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Wed, 11 Oct 2023 02:59:49 +0000 (19:59 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2023-10-11

We've added 14 non-merge commits during the last 5 day(s) which contain
a total of 12 files changed, 398 insertions(+), 104 deletions(-).

The main changes are:

1) Fix s390 JIT backchain issues in the trampoline code generation which
   previously clobbered the caller's backchain, from Ilya Leoshkevich.

2) Fix zero-size allocation warning in xsk sockets when the configured
   ring size was close to SIZE_MAX, from Andrew Kanner.

3) Fixes for bpf_mprog API that were found when implementing support
   in the ebpf-go library along with selftests, from Daniel Borkmann
   and Lorenz Bauer.

4) Fix riscv JIT to properly sign-extend the return register in programs.
   This fixes various test_progs selftests on riscv, from Björn Töpel.

5) Fix verifier log for async callback return values where the allowed
   range was displayed incorrectly, from David Vernet.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  s390/bpf: Fix unwinding past the trampoline
  s390/bpf: Fix clobbering the caller's backchain in the trampoline
  selftests/bpf: Add testcase for async callback return value failure
  bpf: Fix verifier log for async callback return values
  xdp: Fix zero-size allocation warning in xskq_create()
  riscv, bpf: Track both a0 (RISC-V ABI) and a5 (BPF) return values
  riscv, bpf: Sign-extend return values
  selftests/bpf: Make seen_tc* variable tests more robust
  selftests/bpf: Test query on empty mprog and pass revision into attach
  selftests/bpf: Adapt assert_mprog_count to always expect 0 count
  selftests/bpf: Test bpf_mprog query API via libbpf and raw syscall
  bpf: Refuse unused attributes in bpf_prog_{attach,detach}
  bpf: Handle bpf_mprog_query with NULL entry
  bpf: Fix BPF_PROG_QUERY last field check
====================

Link: https://lore.kernel.org/r/20231010223610.3984-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoethtool: Fix mod state of verbose no_mask bitset
Kory Maincent [Mon, 9 Oct 2023 13:36:45 +0000 (15:36 +0200)]
ethtool: Fix mod state of verbose no_mask bitset

A bitset without mask in a _SET request means we want exactly the bits in
the bitset to be set. This works correctly for compact format but when
verbose format is parsed, ethnl_update_bitset32_verbose() only sets the
bits present in the request bitset but does not clear the rest. The commit
6699170376ab fixes this issue by clearing the whole target bitmap before we
start iterating. The solution proposed brought an issue with the behavior
of the mod variable. As the bitset is always cleared the old val will
always differ to the new val.

Fix it by adding a new temporary variable which save the state of the old
bitmap.

Fixes: 6699170376ab ("ethtool: fix application of verbose no_mask bitset")
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20231009133645.44503-1-kory.maincent@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoMerge tag 'linux-can-fixes-for-6.6-20231009' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Wed, 11 Oct 2023 02:45:59 +0000 (19:45 -0700)]
Merge tag 'linux-can-fixes-for-6.6-20231009' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2023-10-09

Lukas Magel's patch for the CAN ISO-TP protocol fixes the TX state
detection and wait behavior.

John Watts contributes a patch to only show the sun4i_can Kconfig
option on ARCH_SUNXI.

A patch by Miquel Raynal fixes the soft-reset workaround for Renesas
SoCs in the sja1000 driver.

Markus Schneider-Pargmann's patch for the tcan4x5x m_can glue driver
fixes the id2 register for the tcan4553.

2 patches by Haibo Chen fix the flexcan stop mode for the imx93 SoC.

* tag 'linux-can-fixes-for-6.6-20231009' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  can: tcan4x5x: Fix id2_register for tcan4553
  can: flexcan: remove the auto stop mode for IMX93
  can: sja1000: Always restart the Tx queue after an overrun
  arm64: dts: imx93: add the Flex-CAN stop mode by GPR
  can: sun4i_can: Only show Kconfig if ARCH_SUNXI is set
  can: isotp: isotp_sendmsg(): fix TX state detection and wait behavior
====================

Link: https://lore.kernel.org/r/20231009085256.693378-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agonet: nfc: fix races in nfc_llcp_sock_get() and nfc_llcp_sock_get_sn()
Eric Dumazet [Mon, 9 Oct 2023 12:31:10 +0000 (12:31 +0000)]
net: nfc: fix races in nfc_llcp_sock_get() and nfc_llcp_sock_get_sn()

Sili Luo reported a race in nfc_llcp_sock_get(), leading to UAF.

Getting a reference on the socket found in a lookup while
holding a lock should happen before releasing the lock.

nfc_llcp_sock_get_sn() has a similar problem.

Finally nfc_llcp_recv_snl() needs to make sure the socket
found by nfc_llcp_sock_from_sn() does not disappear.

Fixes: 8f50020ed9b8 ("NFC: LLCP late binding")
Reported-by: Sili Luo <rootlab@huawei.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willy Tarreau <w@1wt.eu>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231009123110.3735515-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agomctp: perform route lookups under a RCU read-side lock
Jeremy Kerr [Mon, 9 Oct 2023 07:56:45 +0000 (15:56 +0800)]
mctp: perform route lookups under a RCU read-side lock

Our current route lookups (mctp_route_lookup and mctp_route_lookup_null)
traverse the net's route list without the RCU read lock held. This means
the route lookup is subject to preemption, resulting in an potential
grace period expiry, and so an eventual kfree() while we still have the
route pointer.

Add the proper read-side critical section locks around the route
lookups, preventing premption and a possible parallel kfree.

The remaining net->mctp.routes accesses are already under a
rcu_read_lock, or protected by the RTNL for updates.

Based on an analysis from Sili Luo <rootlab@huawei.com>, where
introducing a delay in the route lookup could cause a UAF on
simultaneous sendmsg() and route deletion.

Reported-by: Sili Luo <rootlab@huawei.com>
Fixes: 889b7da23abf ("mctp: Add initial routing framework")
Cc: stable@vger.kernel.org
Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/29c4b0e67dc1bf3571df3982de87df90cae9b631.1696837310.git.jk@codeconstruct.com.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agonet: skbuff: fix kernel-doc typos
Randy Dunlap [Sun, 8 Oct 2023 21:41:21 +0000 (14:41 -0700)]
net: skbuff: fix kernel-doc typos

Correct punctuation and drop an extraneous word.

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20231008214121.25940-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agos390/bpf: Fix unwinding past the trampoline
Ilya Leoshkevich [Tue, 10 Oct 2023 20:20:10 +0000 (22:20 +0200)]
s390/bpf: Fix unwinding past the trampoline

When functions called by the trampoline panic, the backtrace that is
printed stops at the trampoline, because the trampoline does not store
its caller's frame address (backchain) on stack; it also stores the
return address at a wrong location.

Store both the same way as is already done for the regular eBPF programs.

Fixes: 528eb2cb87bc ("s390/bpf: Implement arch_prepare_bpf_trampoline()")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20231010203512.385819-3-iii@linux.ibm.com
13 months agos390/bpf: Fix clobbering the caller's backchain in the trampoline
Ilya Leoshkevich [Tue, 10 Oct 2023 20:20:09 +0000 (22:20 +0200)]
s390/bpf: Fix clobbering the caller's backchain in the trampoline

One of the first things that s390x kernel functions do is storing the
the caller's frame address (backchain) on stack. This makes unwinding
possible. The backchain is always stored at frame offset 152, which is
inside the 160-byte stack area, that the functions allocate for their
callees. The callees must preserve the backchain; the remaining 152
bytes they may use as they please.

Currently the trampoline uses all 160 bytes, clobbering the backchain.
This causes kernel panics when using __builtin_return_address() in
functions called by the trampoline.

Fix by reducing the usage of the caller-reserved stack area by 8 bytes
in the trampoline.

Fixes: 528eb2cb87bc ("s390/bpf: Implement arch_prepare_bpf_trampoline()")
Reported-by: Song Liu <song@kernel.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20231010203512.385819-2-iii@linux.ibm.com
13 months agoMerge tag 'xsa441-6.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Linus Torvalds [Tue, 10 Oct 2023 18:31:42 +0000 (11:31 -0700)]
Merge tag 'xsa441-6.6-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fix from Juergen Gross:
 "A fix for the xen events driver:

  Closing of an event channel in the Linux kernel can result in a
  deadlock. This happens when the close is being performed in parallel
  to an unrelated Xen console action and the handling of a Xen console
  interrupt in an unprivileged guest.

  The closing of an event channel is e.g. triggered by removal of a
  paravirtual device on the other side. As this action will cause
  console messages to be issued on the other side quite often, the
  chance of triggering the deadlock is not negligible"

* tag 'xsa441-6.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/events: replace evtchn_rwlock with RCU

13 months agoKEYS: trusted: Remove redundant static calls usage
Sumit Garg [Fri, 6 Oct 2023 05:18:01 +0000 (10:48 +0530)]
KEYS: trusted: Remove redundant static calls usage

Static calls invocations aren't well supported from module __init and
__exit functions. Especially the static call from cleanup_trusted() led
to a crash on x86 kernel with CONFIG_DEBUG_VIRTUAL=y.

However, the usage of static call invocations for trusted_key_init()
and trusted_key_exit() don't add any value from either a performance or
security perspective. Hence switch to use indirect function calls instead.

Note here that although it will fix the current crash report, ultimately
the static call infrastructure should be fixed to either support its
future usage from module __init and __exit functions or not.

Reported-and-tested-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Link: https://lore.kernel.org/lkml/ZRhKq6e5nF%2F4ZIV1@fedora/#t
Fixes: 5d0682be3189 ("KEYS: trusted: Add generic trusted keys framework")
Signed-off-by: Sumit Garg <sumit.garg@linaro.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 months agoMerge tag 'irq-urgent-2023-10-10-v2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 10 Oct 2023 18:14:07 +0000 (11:14 -0700)]
Merge tag 'irq-urgent-2023-10-10-v2' of git://git./linux/kernel/git/tip/tip

Pull irq fixes from Thomas Gleixner:
 "A set of updates for interrupt chip drivers:

   - Fix the fail of the Qualcomm PDC driver on v3.2 hardware which is
     caused by a control bit being moved to a different location

   - Update the SM8150 device tree PDC resource so the version register
     can be read

   - Make the Renesas RZG2L driver correct for interrupts which are
     outside of the LSB in the TSSR register by using the proper macro
     for calculating the mask

   - Document the Renesas RZ2GL device tree binding correctly and update
     them for a few devices which faul to boot otherwise

   - Use the proper accessor in the RZ2GL driver instead of blindly
     dereferencing an unchecked pointer

   - Make GICv3 handle the dma-non-coherent attribute correctly

   - Ensure that all interrupt controller nodes on RISCV are marked as
     initialized correctly

  Maintainer changes:

   - Add a new entry for GIC interrupt controllers and assign Marc
     Zyngier as the maintainer

   - Remove Marc Zyngier from the core and driver maintainer entries as
     he is burried in work and short of time to handle that.

  Thanks to Marc for all the great work he has done in the past couple
  of years!

  Also note that commit 5873d380f4c0 ("irqchip/qcom-pdc: Add support for
  v3.2 HW") has a incorrect SOB chain.

  The real author is Neil. His patch was posted by Dmitry once and Neil
  picked it up from the list and reposted it with the bogus SOB chain.

  Not a big deal, but worth to mention. I wanted to fix that up, but
  then got distracted and Marc piled more changes on top. So I decided
  to leave it as is instead of rebasing world"

* tag 'irq-urgent-2023-10-10-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  MAINTAINERS: Remove myself from the general IRQ subsystem maintenance
  MAINTAINERS: Add myself as the ARM GIC maintainer
  irqchip/renesas-rzg2l: Convert to irq_data_get_irq_chip_data()
  irqchip/stm32-exti: add missing DT IRQ flag translation
  irqchip/riscv-intc: Mark all INTC nodes as initialized
  irqchip/gic-v3: Enable non-coherent redistributors/ITSes DT probing
  irqchip/gic-v3-its: Split allocation from initialisation of its_node
  dt-bindings: interrupt-controller: arm,gic-v3: Add dma-noncoherent property
  dt-bindings: interrupt-controller: renesas,irqc: Add r8a779f0 support
  dt-bindings: interrupt-controller: renesas,rzg2l-irqc: Document RZ/G2UL SoC
  irqchip: renesas-rzg2l: Fix logic to clear TINT interrupt source
  dt-bindings: interrupt-controller: renesas,rzg2l-irqc: Update description for '#interrupt-cells' property
  arm64: dts: qcom: sm8150: extend the size of the PDC resource
  irqchip/qcom-pdc: Add support for v3.2 HW

13 months agoMerge tag 'hyperv-fixes-signed-20231009' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 10 Oct 2023 18:01:21 +0000 (11:01 -0700)]
Merge tag 'hyperv-fixes-signed-20231009' of git://git./linux/kernel/git/hyperv/linux

Pull hyperv fixes from Wei Liu:

 - fixes for Hyper-V VTL code (Saurabh Sengar and Olaf Hering)

 - fix hv_kvp_daemon to support keyfile based connection profile
   (Shradha Gupta)

* tag 'hyperv-fixes-signed-20231009' of git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux:
  hv/hv_kvp_daemon:Support for keyfile based connection profile
  hyperv: reduce size of ms_hyperv_info
  x86/hyperv: Add common print prefix "Hyper-V" in hv_init
  x86/hyperv: Remove hv_vtl_early_init initcall
  x86/hyperv: Restrict get_vtl to only VTL platforms

13 months agoMerge tag 'v6.6-p4' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Tue, 10 Oct 2023 17:47:35 +0000 (10:47 -0700)]
Merge tag 'v6.6-p4' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fix from Herbert Xu:
 "Fix a regression in dm-crypt"

* tag 'v6.6-p4' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  dm crypt: Fix reqsize in crypt_iv_eboiv_gen

13 months agoMerge tag 'sound-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Tue, 10 Oct 2023 17:33:21 +0000 (10:33 -0700)]
Merge tag 'sound-6.6-rc6' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of pending fixes since a couple of weeks ago, which
  became slightly bigger than usual due to my vacation.

  Most of changes are about ASoC device-specific fixes while USB- and
  HD-audio received quirks as usual. All fixes, including two ASoC core
  changes, are reasonably small and safe to apply"

* tag 'sound-6.6-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (23 commits)
  ALSA: usb-audio: Fix microphone sound on Nexigo webcam.
  ALSA: hda/realtek: Change model for Intel RVP board
  ALSA: usb-audio: Fix microphone sound on Opencomm2 Headset
  ALSA: hda: cs35l41: Cleanup and fix double free in firmware request
  ASoC: dt-bindings: fsl,micfil: Document #sound-dai-cells
  ASoC: amd: yc: Fix non-functional mic on Lenovo 82YM
  ASoC: tlv320adc3xxx: BUG: Correct micbias setting
  ASoC: rt5682: Fix regulator enable/disable sequence
  ASoC: hdmi-codec: Fix broken channel map reporting
  ASoC: core: Do not call link_exit() on uninitialized rtd objects
  ASoC: core: Print component name when printing log
  ASoC: SOF: amd: fix for firmware reload failure after playback
  ASoC: fsl-asoc-card: use integer type for fll_id and pll_id
  ASoC: fsl_sai: Don't disable bitclock for i.MX8MP
  dt-bindings: ASoC: rockchip: Add compatible for RK3128 spdif
  ASoC: soc-generic-dmaengine-pcm: Fix function name in comment
  ALSA: hda/realtek - ALC287 merge RTK codec with CS CS35L41 AMP
  ASoC: simple-card: fixup asoc_simple_probe() error handling
  ASoC: simple-card-utils: fixup simple_util_startup() error handling
  ASoC: Intel: sof_sdw: add support for SKU 0B14
  ...

13 months agoRevert "btrfs: reject unknown mount options early"
David Sterba [Tue, 10 Oct 2023 13:27:56 +0000 (15:27 +0200)]
Revert "btrfs: reject unknown mount options early"

This reverts commit 5f521494cc73520ffac18ede0758883b9aedd018.

The patch breaks mounts with security mount options like

  $ mount -o context=system_u:object_r:root_t:s0 /dev/sdX /mn
  mount: /mnt: wrong fs type, bad option, bad superblock on /dev/sdX, missing codepage or helper program, ...

We cannot reject all unknown options in btrfs_parse_subvol_options() as
intended, the security options can be present at this point and it's not
possible to enumerate them in a future proof way. This means unknown
mount options are silently accepted like before when the filesystem is
mounted with either -o subvol=/path or as followup mounts of the same
device.

Reported-by: Shinichiro Kawasaki <shinichiro.kawasaki@wdc.com
Signed-off-by: David Sterba <dsterba@suse.com>
13 months agonet/mlx5e: Again mutually exclude RX-FCS and RX-port-timestamp
Will Mortensen [Fri, 6 Oct 2023 05:37:06 +0000 (22:37 -0700)]
net/mlx5e: Again mutually exclude RX-FCS and RX-port-timestamp

Commit 1e66220948df8 ("net/mlx5e: Update rx ring hw mtu upon each rx-fcs
flag change") seems to have accidentally inverted the logic added in
commit 0bc73ad46a76 ("net/mlx5e: Mutually exclude RX-FCS and
RX-port-timestamp").

The impact of this is a little unclear since it seems the FCS scattered
with RX-FCS is (usually?) correct regardless.

Fixes: 1e66220948df8 ("net/mlx5e: Update rx ring hw mtu upon each rx-fcs flag change")
Tested-by: Charlotte Tan <charlotte@extrahop.com>
Reviewed-by: Charlotte Tan <charlotte@extrahop.com>
Cc: Adham Faris <afaris@nvidia.com>
Cc: Aya Levin <ayal@nvidia.com>
Cc: Tariq Toukan <tariqt@nvidia.com>
Cc: Moshe Shemesh <moshe@nvidia.com>
Cc: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Will Mortensen <will@extrahop.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://lore.kernel.org/r/20231006053706.514618-1-will@extrahop.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonet/smc: Fix dependency of SMC on ISM
Gerd Bayer [Fri, 6 Oct 2023 12:58:47 +0000 (14:58 +0200)]
net/smc: Fix dependency of SMC on ISM

When the SMC protocol is built into the kernel proper while ISM is
configured to be built as module, linking the kernel fails due to
unresolved dependencies out of net/smc/smc_ism.o to
ism_get_smcd_ops, ism_register_client, and ism_unregister_client
as reported via the linux-next test automation (see link).
This however is a bug introduced a while ago.

Correct the dependency list in ISM's and SMC's Kconfig to reflect the
dependencies that are actually inverted. With this you cannot build a
kernel with CONFIG_SMC=y and CONFIG_ISM=m. Either ISM needs to be 'y',
too - or a 'n'. That way, SMC can still be configured on non-s390
architectures that do not have (nor need) an ISM driver.

Fixes: 89e7d2ba61b7 ("net/ism: Add new API for client registration")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Closes: https://lore.kernel.org/linux-next/d53b5b50-d894-4df8-8969-fd39e63440ae@infradead.org/
Co-developed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Gerd Bayer <gbayer@linux.ibm.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Simon Horman <horms@kernel.org> # build-tested
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Link: https://lore.kernel.org/r/20231006125847.1517840-1-gbayer@linux.ibm.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agoixgbe: fix crash with empty VF macvlan list
Dan Carpenter [Fri, 6 Oct 2023 12:53:09 +0000 (15:53 +0300)]
ixgbe: fix crash with empty VF macvlan list

The adapter->vf_mvs.l list needs to be initialized even if the list is
empty.  Otherwise it will lead to crashes.

Fixes: a1cbb15c1397 ("ixgbe: Add macvlan support for VF")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Link: https://lore.kernel.org/r/ZSADNdIw8zFx1xw2@kadam
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agoMerge branch 'add-update_pn-flag'
Paolo Abeni [Tue, 10 Oct 2023 08:30:33 +0000 (10:30 +0200)]
Merge branch 'add-update_pn-flag'

Radu Pirea says:

====================
Add update_pn flag

Patches extracted from
https://lore.kernel.org/all/20230928084430.1882670-1-radu-nicolae.pirea@oss.nxp.com/
Update_pn flag will let the offloaded MACsec implementations to know when
the PN is updated.
====================

Link: https://lore.kernel.org/r/20231005180636.672791-1-radu-nicolae.pirea@oss.nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonet/mlx5e: macsec: use update_pn flag instead of PN comparation
Radu Pirea (NXP OSS) [Thu, 5 Oct 2023 18:06:36 +0000 (21:06 +0300)]
net/mlx5e: macsec: use update_pn flag instead of PN comparation

When updating the SA, use the new update_pn flags instead of comparing the
new PN with the initial one.

Comparing the initial PN value with the new value will allow the user
to update the SA using the initial PN value as a parameter like this:
$ ip macsec add macsec0 tx sa 0 pn 1 on key 00 \
ead3664f508eb06c40ac7104cdae4ce5
$ ip macsec set macsec0 tx sa 0 pn 1 off

Fixes: 8ff0ac5be144 ("net/mlx5: Add MACsec offload Tx command support")
Fixes: aae3454e4d4c ("net/mlx5e: Add MACsec offload Rx command support")
Signed-off-by: Radu Pirea (NXP OSS) <radu-nicolae.pirea@oss.nxp.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonet: phy: mscc: macsec: reject PN update requests
Radu Pirea (NXP OSS) [Thu, 5 Oct 2023 18:06:35 +0000 (21:06 +0300)]
net: phy: mscc: macsec: reject PN update requests

Updating the PN is not supported.
Return -EINVAL if update_pn is true.

The following command succeeded, but it should fail because the driver
does not update the PN:
ip macsec set macsec0 tx sa 0 pn 232 on

Fixes: 28c5107aa904 ("net: phy: mscc: macsec support")
Signed-off-by: Radu Pirea (NXP OSS) <radu-nicolae.pirea@oss.nxp.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agoocteontx2-pf: mcs: update PN only when update_pn is true
Radu Pirea (NXP OSS) [Thu, 5 Oct 2023 18:06:34 +0000 (21:06 +0300)]
octeontx2-pf: mcs: update PN only when update_pn is true

When updating SA, update the PN only when the update_pn flag is true.
Otherwise, the PN will be reset to its previous value using the
following command and this should not happen:
$ ip macsec set macsec0 tx sa 0 on

Fixes: c54ffc73601c ("octeontx2-pf: mcs: Introduce MACSEC hardware offloading")
Signed-off-by: Radu Pirea (NXP OSS) <radu-nicolae.pirea@oss.nxp.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agonet: macsec: indicate next pn update when offloading
Radu Pirea (NXP OSS) [Thu, 5 Oct 2023 18:06:33 +0000 (21:06 +0300)]
net: macsec: indicate next pn update when offloading

Indicate next PN update using update_pn flag in macsec_context.
Offloaded MACsec implementations does not know whether or not the
MACSEC_SA_ATTR_PN attribute was passed for an SA update and assume
that next PN should always updated, but this is not always true.

The PN can be reset to its initial value using the following command:
$ ip macsec set macsec0 tx sa 0 off #octeontx2-pf case

Or, the update PN command will succeed even if the driver does not support
PN updates.
$ ip macsec set macsec0 tx sa 0 pn 1 on #mscc phy driver case

Comparing the initial PN with the new PN value is not a solution. When
the user updates the PN using its initial value the command will
succeed, even if the driver does not support it. Like this:
$ ip macsec add macsec0 tx sa 0 pn 1 on key 00 \
ead3664f508eb06c40ac7104cdae4ce5
$ ip macsec set macsec0 tx sa 0 pn 1 on #mlx5 case

Signed-off-by: Radu Pirea (NXP OSS) <radu-nicolae.pirea@oss.nxp.com>
Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
13 months agoscsi: Do not rescan devices with a suspended queue
Damien Le Moal [Wed, 4 Oct 2023 08:50:49 +0000 (17:50 +0900)]
scsi: Do not rescan devices with a suspended queue

Commit ff48b37802e5 ("scsi: Do not attempt to rescan suspended devices")
modified scsi_rescan_device() to avoid attempting rescanning a suspended
device. However, the modification added a check to verify that a SCSI
device is in the running state without checking if the device request
queue (in the case of block device) is also running, thus allowing the
exectuion of internal requests. Without checking the device request
queue, commit ff48b37802e5 fix is incomplete and deadlocks on resume can
still happen. Use blk_queue_pm_only() to check if the device request
queue allows executing commands in addition to checking the SCSI device
state.

Reported-by: Petr Tesarik <petr@tesarici.cz>
Fixes: ff48b37802e5 ("scsi: Do not attempt to rescan suspended devices")
Cc: stable@vger.kernel.org
Tested-by: Petr Tesarik <petr@tesarici.cz>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
13 months agoata: pata_parport: fit3: implement IDE command set registers
Ondrej Zary [Thu, 5 Oct 2023 20:55:59 +0000 (22:55 +0200)]
ata: pata_parport: fit3: implement IDE command set registers

fit3 protocol driver does not support accessing IDE control registers
(device control/altstatus). The DOS driver does not use these registers
either (as observed from DOSEMU trace). But the HW seems to be capable
of accessing these registers - I simply tried bit 3 and it works!

The control register is required to properly reset ATAPI devices or
they will be detected only once (after a power cycle).

Tested with EXP Computer CD-865 with MC-1285B EPP cable and
TransDisk 3000.

Signed-off-by: Ondrej Zary <linux@zary.sk>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
13 months agoata: pata_parport: add custom version of wait_after_reset
Ondrej Zary [Thu, 5 Oct 2023 20:55:58 +0000 (22:55 +0200)]
ata: pata_parport: add custom version of wait_after_reset

Some parallel adapters (e.g. EXP Computer MC-1285B EPP Cable) return
bogus values when there's no master device present. This can cause
reset to fail, preventing the lone slave device (such as EXP Computer
CD-865) from working.

Add custom version of wait_after_reset that ignores master failure when
a slave device is present. The custom version is also needed because
the generic ata_sff_wait_after_reset uses direct port I/O for slave
device detection.

Signed-off-by: Ondrej Zary <linux@zary.sk>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
13 months agoata: pata_parport: implement set_devctl
Ondrej Zary [Thu, 5 Oct 2023 20:55:57 +0000 (22:55 +0200)]
ata: pata_parport: implement set_devctl

Add missing ops->sff_set_devctl implementation.

Fixes: 246a1c4c6b7f ("ata: pata_parport: add driver (PARIDE replacement)")
Cc: stable@vger.kernel.org
Signed-off-by: Ondrej Zary <linux@zary.sk>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
13 months agoata: pata_parport: fix pata_parport_devchk
Ondrej Zary [Thu, 5 Oct 2023 20:55:56 +0000 (22:55 +0200)]
ata: pata_parport: fix pata_parport_devchk

There's a 'x' missing in 0x55 in pata_parport_devchk(), causing the
detection to always fail. Fix it.

Fixes: 246a1c4c6b7f ("ata: pata_parport: add driver (PARIDE replacement)")
Cc: stable@vger.kernel.org
Signed-off-by: Ondrej Zary <linux@zary.sk>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
13 months agohv/hv_kvp_daemon:Support for keyfile based connection profile
Shradha Gupta [Mon, 9 Oct 2023 10:38:40 +0000 (03:38 -0700)]
hv/hv_kvp_daemon:Support for keyfile based connection profile

Ifcfg config file support in NetworkManger is deprecated. This patch
provides support for the new keyfile config format for connection
profiles in NetworkManager. The patch modifies the hv_kvp_daemon code
to generate the new network configuration in keyfile
format(.ini-style format) along with a ifcfg format configuration.
The ifcfg format configuration is also retained to support easy
backward compatibility for distro vendors. These configurations are
stored in temp files which are further translated using the
hv_set_ifconfig.sh script. This script is implemented by individual
distros based on the network management commands supported.
For example, RHEL's implementation could be found here:
https://gitlab.com/redhat/centos-stream/src/hyperv-daemons/-/blob/c9s/hv_set_ifconfig.sh
Debian's implementation could be found here:
https://github.com/endlessm/linux/blob/master/debian/cloud-tools/hv_set_ifconfig

The next part of this support is to let the Distro vendors consume
these modified implementations to the new configuration format.

Tested-on: Rhel9(Hyper-V, Azure)(nm and ifcfg files verified)
Signed-off-by: Shradha Gupta <shradhagupta@linux.microsoft.com>
Reviewed-by: Saurabh Sengar <ssengar@linux.microsoft.com>
Reviewed-by: Ani Sinha <anisinha@redhat.com>
Signed-off-by: Wei Liu <wei.liu@kernel.org>
Link: https://lore.kernel.org/r/1696847920-31125-1-git-send-email-shradhagupta@linux.microsoft.com
13 months agonet: refine debug info in skb_checksum_help()
Eric Dumazet [Fri, 6 Oct 2023 17:33:54 +0000 (17:33 +0000)]
net: refine debug info in skb_checksum_help()

syzbot uses panic_on_warn.

This means that the skb_dump() I added in the blamed commit are
not even called.

Rewrite this so that we get the needed skb dump before syzbot crashes.

Fixes: eeee4b77dc52 ("net: add more debug info in skb_checksum_help()")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20231006173355.2254983-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
13 months agoselftests/bpf: Add testcase for async callback return value failure
David Vernet [Mon, 9 Oct 2023 16:14:14 +0000 (11:14 -0500)]
selftests/bpf: Add testcase for async callback return value failure

A previous commit updated the verifier to print an accurate failure
message for when someone specifies a nonzero return value from an async
callback. This adds a testcase for validating that the verifier emits
the correct message in such a case.

Signed-off-by: David Vernet <void@manifault.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20231009161414.235829-2-void@manifault.com
13 months agobpf: Fix verifier log for async callback return values
David Vernet [Mon, 9 Oct 2023 16:14:13 +0000 (11:14 -0500)]
bpf: Fix verifier log for async callback return values

The verifier, as part of check_return_code(), verifies that async
callbacks such as from e.g. timers, will return 0. It does this by
correctly checking that R0->var_off is in tnum_const(0), which
effectively checks that it's in a range of 0. If this condition fails,
however, it prints an error message which says that the value should
have been in (0x0; 0x1). This results in possibly confusing output such
as the following in which an async callback returns 1:

  At async callback the register R0 has value (0x1; 0x0) should have been in (0x0; 0x1)

The fix is easy -- we should just pass the tnum_const(0) as the correct
range to verbose_invalid_scalar(), which will then print the following:

  At async callback the register R0 has value (0x1; 0x0) should have been in (0x0; 0x0)

Fixes: bfc6bb74e4f1 ("bpf: Implement verifier support for validation of async callbacks.")
Signed-off-by: David Vernet <void@manifault.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20231009161414.235829-1-void@manifault.com
13 months agoxdp: Fix zero-size allocation warning in xskq_create()
Andrew Kanner [Sat, 7 Oct 2023 07:51:49 +0000 (10:51 +0300)]
xdp: Fix zero-size allocation warning in xskq_create()

Syzkaller reported the following issue:

  ------------[ cut here ]------------
  WARNING: CPU: 0 PID: 2807 at mm/vmalloc.c:3247 __vmalloc_node_range (mm/vmalloc.c:3361)
  Modules linked in:
  CPU: 0 PID: 2807 Comm: repro Not tainted 6.6.0-rc2+ #12
  Hardware name: Generic DT based system
  unwind_backtrace from show_stack (arch/arm/kernel/traps.c:258)
  show_stack from dump_stack_lvl (lib/dump_stack.c:107 (discriminator 1))
  dump_stack_lvl from __warn (kernel/panic.c:633 kernel/panic.c:680)
  __warn from warn_slowpath_fmt (./include/linux/context_tracking.h:153 kernel/panic.c:700)
  warn_slowpath_fmt from __vmalloc_node_range (mm/vmalloc.c:3361 (discriminator 3))
  __vmalloc_node_range from vmalloc_user (mm/vmalloc.c:3478)
  vmalloc_user from xskq_create (net/xdp/xsk_queue.c:40)
  xskq_create from xsk_setsockopt (net/xdp/xsk.c:953 net/xdp/xsk.c:1286)
  xsk_setsockopt from __sys_setsockopt (net/socket.c:2308)
  __sys_setsockopt from ret_fast_syscall (arch/arm/kernel/entry-common.S:68)

xskq_get_ring_size() uses struct_size() macro to safely calculate the
size of struct xsk_queue and q->nentries of desc members. But the
syzkaller repro was able to set q->nentries with the value initially
taken from copy_from_sockptr() high enough to return SIZE_MAX by
struct_size(). The next PAGE_ALIGN(size) is such case will overflow
the size_t value and set it to 0. This will trigger WARN_ON_ONCE in
vmalloc_user() -> __vmalloc_node_range().

The issue is reproducible on 32-bit arm kernel.

Fixes: 9f78bf330a66 ("xsk: support use vaddr as ring")
Reported-by: syzbot+fae676d3cf469331fc89@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/000000000000c84b4705fb31741e@google.com/T/
Reported-by: syzbot+b132693e925cbbd89e26@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/000000000000e20df20606ebab4f@google.com/T/
Signed-off-by: Andrew Kanner <andrew.kanner@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: syzbot+fae676d3cf469331fc89@syzkaller.appspotmail.com
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://syzkaller.appspot.com/bug?extid=fae676d3cf469331fc89
Link: https://lore.kernel.org/bpf/20231007075148.1759-1-andrew.kanner@gmail.com
13 months agoMerge tag 'irqchip-fixes-6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Thomas Gleixner [Mon, 9 Oct 2023 14:01:55 +0000 (16:01 +0200)]
Merge tag 'irqchip-fixes-6.6-2' of git://git./linux/kernel/git/maz/arm-platforms into irq/urgent

Pull irqchip fixes from Marc Zyngier:

  - DT binding updates for Renesas r8a779f0 and rzg2l

  - Let GICv3 honor the "dma-non-coherent" attribute for systems that
    rely on SW guessing what the HW supports

  - Fix the RISC-V INTC probing by marking all devices as initialised
    at once

  - Properly translate interrupt numbers from DT on stm32-exti

  - Use irq_data_get_irq_chip_data() in the rzg2l driver instead of
    blindly dereferencing the irq_data structure

  - Add a MAINTAINERS entry for the various ARM GIC irqchip drivers

  - Remove myself as the top-level irqchip/irqdomain maintainer

Link: https://lore.kernel.org/all/20231007121933.3840357-1-maz@kernel.org
13 months agoriscv, bpf: Track both a0 (RISC-V ABI) and a5 (BPF) return values
Björn Töpel [Wed, 4 Oct 2023 12:07:06 +0000 (14:07 +0200)]
riscv, bpf: Track both a0 (RISC-V ABI) and a5 (BPF) return values

The RISC-V BPF uses a5 for BPF return values, which are zero-extended,
whereas the RISC-V ABI uses a0 which is sign-extended. In other words,
a5 and a0 can differ, and are used in different context.

The BPF trampoline are used for both BPF programs, and regular kernel
functions.

Make sure that the RISC-V BPF trampoline saves, and restores both a0
and a5.

Fixes: 49b5e77ae3e2 ("riscv, bpf: Add bpf trampoline support for RV64")
Signed-off-by: Björn Töpel <bjorn@rivosinc.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20231004120706.52848-3-bjorn@kernel.org
13 months agoriscv, bpf: Sign-extend return values
Björn Töpel [Wed, 4 Oct 2023 12:07:05 +0000 (14:07 +0200)]
riscv, bpf: Sign-extend return values

The RISC-V architecture does not expose sub-registers, and hold all
32-bit values in a sign-extended format [1] [2]:

  | The compiler and calling convention maintain an invariant that all
  | 32-bit values are held in a sign-extended format in 64-bit
  | registers. Even 32-bit unsigned integers extend bit 31 into bits
  | 63 through 32. Consequently, conversion between unsigned and
  | signed 32-bit integers is a no-op, as is conversion from a signed
  | 32-bit integer to a signed 64-bit integer.

While BPF, on the other hand, exposes sub-registers, and use
zero-extension (similar to arm64/x86).

This has led to some subtle bugs, where a BPF JITted program has not
sign-extended the a0 register (return value in RISC-V land), passed
the return value up the kernel, e.g.:

  | int from_bpf(void);
  |
  | long foo(void)
  | {
  |    return from_bpf();
  | }

Here, a0 would be 0xffff_ffff, instead of the expected
0xffff_ffff_ffff_ffff.

Internally, the RISC-V JIT uses a5 as a dedicated register for BPF
return values.

Keep a5 zero-extended, but explicitly sign-extend a0 (which is used
outside BPF land). Now that a0 (RISC-V ABI) and a5 (BPF ABI) differs,
a0 is only moved to a5 for non-BPF native calls (BPF_PSEUDO_CALL).

Fixes: 2353ecc6f91f ("bpf, riscv: add BPF JIT for RV64G")
Signed-off-by: Björn Töpel <bjorn@rivosinc.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://github.com/riscv/riscv-isa-manual/releases/download/riscv-isa-release-056b6ff-2023-10-02/unpriv-isa-asciidoc.pdf
Link: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/releases/download/draft-20230929-e5c800e661a53efe3c2678d71a306323b60eb13b/riscv-abi.pdf
Link: https://lore.kernel.org/bpf/20231004120706.52848-2-bjorn@kernel.org
13 months agoprintk: flush consoles before checking progress
John Ogness [Fri, 6 Oct 2023 08:21:50 +0000 (10:21 +0200)]
printk: flush consoles before checking progress

Commit 9e70a5e109a4 ("printk: Add per-console suspended state")
removed console lock usage during resume and replaced it with
the clearly defined console_list_lock and srcu mechanisms.

However, the console lock usage had an important side-effect
of flushing the consoles. After its removal, consoles were no
longer flushed before checking their progress.

Add the console_lock/console_unlock dance to the beginning
of __pr_flush() to actually flush the consoles before checking
their progress. Also add comments to clarify this additional
usage of the console lock.

Note that console_unlock() does not guarantee flushing all messages
since the commit dbdda842fe96f89 ("printk: Add console owner and waiter
logic to load balance console writes").

Reported-by: Todd Brandt <todd.e.brandt@intel.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217955
Fixes: 9e70a5e109a4 ("printk: Add per-console suspended state")
Co-developed-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: John Ogness <john.ogness@linutronix.de>
Link: https://lore.kernel.org/r/20231006082151.6969-2-pmladek@suse.com