platform/kernel/linux-rpi.git
17 months agomlxsw: spectrum_switchdev: Pass extack to mlxsw_sp_br_ban_rif_pvid_change()
Petr Machata [Thu, 13 Jul 2023 16:15:24 +0000 (18:15 +0200)]
mlxsw: spectrum_switchdev: Pass extack to mlxsw_sp_br_ban_rif_pvid_change()

Currently the reason for rejection of PVID manipulation is dumped to
syslog, and a generic -EBUSY is returned to the userspace. But
switchdev_handle_port_obj_add(), through which we get to
mlxsw_sp_port_vlans_add(), handles extack just fine, and we can pass the
message this way.

This improves visibility into reasons why the request to change PVID
was rejected. Before the change:

 # bridge vlan add dev br vid 2 self pvid untagged
 RTNETLINK answers: Device or resource busy
 (plus a syslog line)

After the change:

 # bridge vlan add dev br vid 2 self pvid untagged
 Error: mlxsw_spectrum: Can't change PVID, it's used by router interface.

Note that this particular error message is going away in the following
patches. However the ability to pass error messages through extack will be
useful more broadly for communicating in particular reasons why a RIF
failed to be created.

Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Danielle Ratson <danieller@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoMerge branch 'macsec-selftests'
David S. Miller [Fri, 14 Jul 2023 08:16:53 +0000 (09:16 +0100)]
Merge branch 'macsec-selftests'

Sabrina Dubroca says:

====================
net: add MACsec offload selftests

Patch 1 adds MACsec offload to netdevsim (unchanged from v2).

Patch 2 adds a corresponding selftest to the rtnetlink testsuite.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoselftests: rtnetlink: add MACsec offload tests
Sabrina Dubroca [Thu, 13 Jul 2023 13:20:24 +0000 (15:20 +0200)]
selftests: rtnetlink: add MACsec offload tests

Like the IPsec offload test, this requires netdevsim.

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonetdevsim: add dummy macsec offload
Sabrina Dubroca [Thu, 13 Jul 2023 13:20:23 +0000 (15:20 +0200)]
netdevsim: add dummy macsec offload

When the kernel is compiled with MACsec support, add the
NETIF_F_HW_MACSEC feature to netdevsim devices and implement
macsec_ops.

To allow easy testing of failure from the device, support is limited
to 3 SecY's per netdevsim device, and 1 RXSC per SecY.

v2:
 - nsim_macsec_add_secy, return -ENOSPC if secy_count isn't full but
   we can't find an empty slot (Simon Horman)
 - add sci_to_cpu to make sparse happy (Simon Horman)
 - remove set but not used secy variable (kernel test robot and
   Simon Horman)

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agodevlink: remove reload failed checks in params get/set callbacks
Jiri Pirko [Thu, 13 Jul 2023 09:44:19 +0000 (11:44 +0200)]
devlink: remove reload failed checks in params get/set callbacks

The checks in question were introduced by:
commit 6b4db2e528f6 ("devlink: Fix use-after-free after a failed reload").
That fixed an issue of reload with mlxsw driver.

Back then, that was a valid fix, because there was a limitation
in place that prevented drivers from registering/unregistering params
when devlink instance was registered.

It was possible to do the fix differently by changing drivers to
register/unregister params in appropriate places making sure the ops
operate only on memory which is allocated and initialized. But that,
as a dependency, would require to remove the limitation mentioned above.

Eventually, this limitation was lifted by:
commit 1d18bb1a4ddd ("devlink: allow registering parameters after the instance")

Also, the alternative fix (which also fixed another issue) was done by:
commit 74cbc3c03c82 ("mlxsw: spectrum_acl_tcam: Move devlink param to TCAM code").

Therefore, the checks are no longer relevant. Each driver should make
sure to have the params registered only when the memory the ops
are working with is allocated and initialized.

So remove the checks.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoMerge branch 'mv88e6xxx-phylink_pcs'
David S. Miller [Fri, 14 Jul 2023 07:51:49 +0000 (08:51 +0100)]
Merge branch 'mv88e6xxx-phylink_pcs'

Russell King says:

====================
Convert mv88e6xxx to phylink_pcs

This series (previously posted with further patches on the 26 June as
RFC) converts mv88e6xxx to phylink_pcs, and thus moves it from being
a pre-March 2020 legacy driver.

The first four patches lay the ground-work for the conversion by
adding four new methods to the phylink_pcs operations structure:

  pcs_enable() - called when the PCS is going to start to be used
  pcs_disable() - called when the PCS is no longer being used

  pcs_pre_config() - called before the MAC configuration method
  pcs_post_config() - called after the MAC configuration method
      Both of these are necessary for some of the mv88e639x
      workarounds.

We also add the ability to inform phylink of a change to the PCS
state without involving the MAC later, by providing
phylink_pcs_change() which takes a phylink_pcs structure rather than
a phylink structure. phylink maintains which instance the PCS is
conencted to, so internally it can do the right thing when the PCS
is in-use.

Then we provide some additional mdiobus and mdiodev accessors that
we will be using in the new PCS drivers.

The changes for mv88e6xxx follow, and the first one needs to be
explicitly pointed out - we (Andrew and myself) have both decided that
all possible approaches to maintaining backwards compatibility with DT
have been exhaused - everyone has some objection to everything that
has been proposed. So, after many years of trying, we have decided
that this is just an impossibility, and with this patch, we are now
intentionally and knowingly breaking any DT that does not specify the
CPU and DSA port fixed-link parameters. Hence why Andrew has recently
been submitting DT update patches. It is regrettable that it has come
to this.

Following this, we start preparing 88e6xxx for phylink_pcs conversion
by padding the mac_select_pcs() DSA method, and the internal hooks to
create and tear-down PCS instances. Rather than bloat the already very
large mv88e6xxx_ops structure, I decided that it would be better that
the new internal chip specific PCS methods are all grouped within their
own structure - and this structure can be declared in the PCS drivers
themselves.

Then we have the actual conversion patches, one for each family of PCS.

Lastly, we clean up the driver after conversion, removing all the now
redundant code.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: cleanup after phylink_pcs conversion
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:59 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: cleanup after phylink_pcs conversion

Now that mv88e6xxx is completely converted to using phylink_pcs
support, we have no need for the serdes methods. Remove all this
infrastructure. Also remove the __maybe_unused from
mv88e6xxx_pcs_select().

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: convert 88e639x to phylink_pcs
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:53 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: convert 88e639x to phylink_pcs

Convert the 88E6390, 88E6390X, and 88E6393X family of switches to use
the phylink_pcs infrastructure.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: convert 88e6352 to phylink_pcs
Russell King [Thu, 13 Jul 2023 08:42:48 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: convert 88e6352 to phylink_pcs

Convert the 88E6352 SERDES code to use the phylink_pcs infrastructure.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: convert 88e6185 to phylink_pcs
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:43 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: convert 88e6185 to phylink_pcs

Convert the 88E6185 SERDES code to use the phylink_pcs infrastructure.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: export mv88e6xxx_pcs_decode_state()
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:38 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: export mv88e6xxx_pcs_decode_state()

Rename and export the PCS state decoding function so our PCS can
make use of the functionality provided by this.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: add infrastructure for phylink_pcs
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:33 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: add infrastructure for phylink_pcs

Add infrastructure for phylink_pcs to the mv88e6xxx driver. This
involves adding a mac_select_pcs() hook so we can pass the PCS to
phylink at the appropriate time, and a PCS initialisation function.

As the various chip implementations are converted to use phylink_pcs,
they are no longer reliant on the legacy phylink behaviour. We detect
this by the use of this infrastructure, or the lack of any serdes.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: mv88e6xxx: remove handling for DSA and CPU ports
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:28 +0000 (09:42 +0100)]
net: dsa: mv88e6xxx: remove handling for DSA and CPU ports

As we now always use a fixed-link for DSA and CPU ports, we no longer
need the hack in the Marvell code to make this work. Remove it.

This is especially important with the conversion of DSA drivers to
phylink_pcs, as the PCS code only gets called if we are using
phylink for the port.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: mdio: add unlocked mdiobus and mdiodev bus accessors
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:22 +0000 (09:42 +0100)]
net: mdio: add unlocked mdiobus and mdiodev bus accessors

Add the following unlocked accessors to complete the set:
__mdiobus_modify()
__mdiodev_read()
__mdiodev_write()
__mdiodev_modify()
__mdiodev_modify_changed()
which we will need for Marvell DSA PCS conversion.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: phylink: add support for PCS link change notifications
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:17 +0000 (09:42 +0100)]
net: phylink: add support for PCS link change notifications

Add a function, phylink_pcs_change() which can be used by PCs drivers
to notify phylink about changes to the PCS link state.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: phylink: add pcs_pre_config()/pcs_post_config() methods
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:12 +0000 (09:42 +0100)]
net: phylink: add pcs_pre_config()/pcs_post_config() methods

Add hooks that are called before and after the mac_config() call,
which will be needed to deal with errata workarounds for the
Marvell 88e639x DSA switches.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: phylink: add pcs_enable()/pcs_disable() methods
Russell King (Oracle) [Thu, 13 Jul 2023 08:42:07 +0000 (09:42 +0100)]
net: phylink: add pcs_enable()/pcs_disable() methods

Add phylink PCS enable/disable callbacks that will allow us to place
IEEE 802.3 register compliant PCS in power-down mode while not being
used.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: ngbe: add Wake on Lan support
Mengyuan Lou [Thu, 13 Jul 2023 06:09:11 +0000 (14:09 +0800)]
net: ngbe: add Wake on Lan support

Implement ethtool_ops get_wol and set_wol.
Implement Wake-on-LAN support.

Wol requires hardware board support which use sub id
to identify.
Magic packets are checked by fw, for now just support
WAKE_MAGIC.

Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agonet: dsa: ar9331: Use maple tree register cache
Mark Brown [Wed, 12 Jul 2023 23:45:58 +0000 (00:45 +0100)]
net: dsa: ar9331: Use maple tree register cache

We now have a regmap cache which uses a maple tree to store the register
state, this is a more modern data structure and the regmap level code
using it makes a number of assumptions better tuned for modern hardware
than those made by the rbtree cache type that the at9331 driver uses.
Switch the ar9331 driver to use the more modern data structure.

This should have minimal practical impact, it's mainly code
modernisation.

Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoMerge branch 'sk-const'
David S. Miller [Fri, 14 Jul 2023 07:27:33 +0000 (08:27 +0100)]
Merge branch 'sk-const'

Guillaume Nault says:

====================
net: Mark the sk parameter of routing functions as 'const'.

The sk_getsecid security hook prevents the use of a const sk pointer in
several routing functions. Since this hook should only read sk data,
make its sk argument const (patch 1), then constify the sk parameter of
various routing functions (patches 2-4).

Build-tested with make allmodconfig.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agopptp: Constify the po parameter of pptp_route_output().
Guillaume Nault [Tue, 11 Jul 2023 13:06:26 +0000 (15:06 +0200)]
pptp: Constify the po parameter of pptp_route_output().

Make it explicit that this function doesn't modify the socket passed as
parameter.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoipv6: Constify the sk parameter of several helper functions.
Guillaume Nault [Tue, 11 Jul 2023 13:06:21 +0000 (15:06 +0200)]
ipv6: Constify the sk parameter of several helper functions.

icmpv6_flow_init(), ip6_datagram_flow_key_init() and ip6_mc_hdr() don't
need to modify their sk argument. Make that explicit using const.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoipv4: Constify the sk parameter of ip_route_output_*().
Guillaume Nault [Tue, 11 Jul 2023 13:06:14 +0000 (15:06 +0200)]
ipv4: Constify the sk parameter of ip_route_output_*().

These functions don't need to modify the socket, so let's allow the
callers to pass a const struct sock *.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agosecurity: Constify sk in the sk_getsecid hook.
Guillaume Nault [Tue, 11 Jul 2023 13:06:08 +0000 (15:06 +0200)]
security: Constify sk in the sk_getsecid hook.

The sk_getsecid hook shouldn't need to modify its socket argument.
Make it const so that callers of security_sk_classify_flow() can use a
const struct sock *.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 months agoMerge branch 'net-stmmac-replace-boolean-fields-in-plat_stmmacenet_data-with-flags'
Jakub Kicinski [Fri, 14 Jul 2023 03:57:55 +0000 (20:57 -0700)]
Merge branch 'net-stmmac-replace-boolean-fields-in-plat_stmmacenet_data-with-flags'

Bartosz Golaszewski says:

====================
net: stmmac: replace boolean fields in plat_stmmacenet_data with flags

As suggested by Jose Abreu: let's drop all 12 boolean fields in
plat_stmmacenet_data and replace them with a common bitfield.
====================

Link: https://lore.kernel.org/r/20230710090001.303225-1-brgl@bgdev.pl
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the en_tx_lpi_clockgating field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 09:00:01 +0000 (11:00 +0200)]
net: stmmac: replace the en_tx_lpi_clockgating field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-13-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the rx_clk_runs_in_lpi field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 09:00:00 +0000 (11:00 +0200)]
net: stmmac: replace the rx_clk_runs_in_lpi field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-12-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the int_snapshot_en field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:59 +0000 (10:59 +0200)]
net: stmmac: replace the int_snapshot_en field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-11-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the ext_snapshot_en field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:58 +0000 (10:59 +0200)]
net: stmmac: replace the ext_snapshot_en field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-10-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the multi_msi_en field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:57 +0000 (10:59 +0200)]
net: stmmac: replace the multi_msi_en field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-9-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the vlan_fail_q_en field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:56 +0000 (10:59 +0200)]
net: stmmac: replace the vlan_fail_q_en field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-8-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the serdes_up_after_phy_linkup field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:55 +0000 (10:59 +0200)]
net: stmmac: replace the serdes_up_after_phy_linkup field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-7-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the tso_en field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:54 +0000 (10:59 +0200)]
net: stmmac: replace the tso_en field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-6-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the has_sun8i field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:53 +0000 (10:59 +0200)]
net: stmmac: replace the has_sun8i field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-5-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the use_phy_wol field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:52 +0000 (10:59 +0200)]
net: stmmac: replace the use_phy_wol field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-4-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the sph_disable field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:51 +0000 (10:59 +0200)]
net: stmmac: replace the sph_disable field with a flag

Drop the boolean field of the plat_stmmacenet_data structure in favor of a
simple bitfield flag.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-3-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: replace the has_integrated_pcs field with a flag
Bartosz Golaszewski [Mon, 10 Jul 2023 08:59:50 +0000 (10:59 +0200)]
net: stmmac: replace the has_integrated_pcs field with a flag

struct plat_stmmacenet_data contains several boolean fields that could be
easily replaced with a common integer 'flags' bitfield and bit defines.

Start the process with the has_integrated_pcs field.

Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Andrew Halaney <ahalaney@redhat.com>
Link: https://lore.kernel.org/r/20230710090001.303225-2-brgl@bgdev.pl
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonfp: prevent dropped counter increment during probe
Ziyang Chen [Wed, 12 Jul 2023 12:35:51 +0000 (14:35 +0200)]
nfp: prevent dropped counter increment during probe

The dev_rx_discards counter will increment by one when an interface is
toggled up and down. The main reason is that the driver first sends a
`NFP_NET_CFG_CTRL_ENABLE` configuration packet to the NIC to perform port
initialisation when an interface is set up. But there is a race between
physical link up and free list queue initialization which may lead to the
configuration packet being discarded.

To address this problem a new bit NFP_NET_CFG_CTRL_FREELIST_EN is added to
perform free list initialisation on the NIC. The FREELIST_EN should be sent
in advance to initialize free list queue. When a port is set to down,
FREELIST_EN should be sent after CTRL_ENABLE to avoid packet discards.

Signed-off-by: Ziyang Chen <ziyang.chen@corigine.com>
Acked-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Link: https://lore.kernel.org/r/20230712123551.13858-1-louis.peens@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge branch 'selftests-mptcp-join-pass-args-in-new-env-vars'
Jakub Kicinski [Fri, 14 Jul 2023 03:34:54 +0000 (20:34 -0700)]
Merge branch 'selftests-mptcp-join-pass-args-in-new-env-vars'

Matthieu Baerts says:

====================
selftests: mptcp: join: pass args in new env vars

In this series, Geliang did some refactoring in the mptcp_join.sh file.

Patch 1 reduces the scope of some global env vars, only used by some
tests: easier to deal with.

Patch 2 uses a dedicated env var for fastclose case instead of re-using
addr_nr_ns2 with embedded info, clearer.

Patch 3 is similar but for the fullmesh case.

Patch 4 moves a positional but optional argument of run_tests() to an
env var like it has already been done with the other args, cleaner.
====================

Link: https://lore.kernel.org/r/20230712-upstream-net-next-20230712-selftests-mptcp-use-local-env-v1-0-f1c8b62fbf95@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mptcp: add speed env var
Geliang Tang [Wed, 12 Jul 2023 16:03:19 +0000 (18:03 +0200)]
selftests: mptcp: add speed env var

Use a new env var speed instead of using 'speed' argument of run_tests()
and do_transfer(). It can be set with 'fast', 'slow', or the actual speed
number:

  run_tests $ns1 $ns2 10.0.1.1 slow

  ->

  speed=slow \
          run_tests $ns1 $ns2 10.0.1.1.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/r/20230712-upstream-net-next-20230712-selftests-mptcp-use-local-env-v1-4-f1c8b62fbf95@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mptcp: add fullmesh env var
Geliang Tang [Wed, 12 Jul 2023 16:03:18 +0000 (18:03 +0200)]
selftests: mptcp: add fullmesh env var

Use a new env var fullmesh instead of passing 'fullmesh_*' to addr_nr_ns2.
It can be set with the actual value of addr_nr_ns2 now:

  addr_nr_ns2=fullmesh_1 \
          run_tests $ns1 $ns2 10.0.1.1

  ->

  fullmesh=1 \
          run_tests $ns1 $ns2 10.0.1.1.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/r/20230712-upstream-net-next-20230712-selftests-mptcp-use-local-env-v1-3-f1c8b62fbf95@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mptcp: add fastclose env var
Geliang Tang [Wed, 12 Jul 2023 16:03:17 +0000 (18:03 +0200)]
selftests: mptcp: add fastclose env var

Use a new env var fastclose instead of passing fastclose to addr_nr_ns2.
It can be set with 'server' or 'client':

  addr_nr_ns2=fastclose_client \
          run_tests $ns1 $ns2 10.0.1.1

  ->

  fastclose=client \
          run_tests $ns1 $ns2 10.0.1.1.

With this change, the fullmesh flag setting code can be moved into
pm_nl_set_endpoint() from do_transfer().

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/r/20230712-upstream-net-next-20230712-selftests-mptcp-use-local-env-v1-2-f1c8b62fbf95@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mptcp: set all env vars as local ones
Geliang Tang [Wed, 12 Jul 2023 16:03:16 +0000 (18:03 +0200)]
selftests: mptcp: set all env vars as local ones

It would be better to move the declaration of all the env variables to
do_transfer(), run_tests(), or pm_nl_set_endpoint() as local variables,
instead of exporting them globally at the beginning of the file.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Link: https://lore.kernel.org/r/20230712-upstream-net-next-20230712-selftests-mptcp-use-local-env-v1-1-f1c8b62fbf95@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf...
Jakub Kicinski [Fri, 14 Jul 2023 02:13:24 +0000 (19:13 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf-next

Alexei Starovoitov says:

====================
pull-request: bpf-next 2023-07-13

We've added 67 non-merge commits during the last 15 day(s) which contain
a total of 106 files changed, 4444 insertions(+), 619 deletions(-).

The main changes are:

1) Fix bpftool build in presence of stale vmlinux.h,
   from Alexander Lobakin.

2) Introduce bpf_me_mcache_free_rcu() and fix OOM under stress,
   from Alexei Starovoitov.

3) Teach verifier actual bounds of bpf_get_smp_processor_id()
   and fix perf+libbpf issue related to custom section handling,
   from Andrii Nakryiko.

4) Introduce bpf map element count, from Anton Protopopov.

5) Check skb ownership against full socket, from Kui-Feng Lee.

6) Support for up to 12 arguments in BPF trampoline, from Menglong Dong.

7) Export rcu_request_urgent_qs_task, from Paul E. McKenney.

8) Fix BTF walking of unions, from Yafang Shao.

9) Extend link_info for kprobe_multi and perf_event links,
   from Yafang Shao.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (67 commits)
  selftests/bpf: Add selftest for PTR_UNTRUSTED
  bpf: Fix an error in verifying a field in a union
  selftests/bpf: Add selftests for nested_trust
  bpf: Fix an error around PTR_UNTRUSTED
  selftests/bpf: add testcase for TRACING with 6+ arguments
  bpf, x86: allow function arguments up to 12 for TRACING
  bpf, x86: save/restore regs with BPF_DW size
  bpftool: Use "fallthrough;" keyword instead of comments
  bpf: Add object leak check.
  bpf: Convert bpf_cpumask to bpf_mem_cache_free_rcu.
  bpf: Introduce bpf_mem_free_rcu() similar to kfree_rcu().
  selftests/bpf: Improve test coverage of bpf_mem_alloc.
  rcu: Export rcu_request_urgent_qs_task()
  bpf: Allow reuse from waiting_for_gp_ttrace list.
  bpf: Add a hint to allocated objects.
  bpf: Change bpf_mem_cache draining process.
  bpf: Further refactor alloc_bulk().
  bpf: Factor out inc/dec of active flag into helpers.
  bpf: Refactor alloc_bulk().
  bpf: Let free_all() return the number of freed elements.
  ...
====================

Link: https://lore.kernel.org/r/20230714020910.80794-1-alexei.starovoitov@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Fri, 14 Jul 2023 00:33:19 +0000 (17:33 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

No conflicts or adjacent changes.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests/bpf: Add selftest for PTR_UNTRUSTED
Yafang Shao [Thu, 13 Jul 2023 02:56:42 +0000 (02:56 +0000)]
selftests/bpf: Add selftest for PTR_UNTRUSTED

Add a new selftest to check the PTR_UNTRUSTED condition. Below is the
result,

 #160     ptr_untrusted:OK

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/r/20230713025642.27477-5-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agobpf: Fix an error in verifying a field in a union
Yafang Shao [Thu, 13 Jul 2023 02:56:41 +0000 (02:56 +0000)]
bpf: Fix an error in verifying a field in a union

We are utilizing BPF LSM to monitor BPF operations within our container
environment. When we add support for raw_tracepoint, it hits below
error.

; (const void *)attr->raw_tracepoint.name);
27: (79) r3 = *(u64 *)(r2 +0)
access beyond the end of member map_type (mend:4) in struct (anon) with off 0 size 8

It can be reproduced with below BPF prog.

SEC("lsm/bpf")
int BPF_PROG(bpf_audit, int cmd, union bpf_attr *attr, unsigned int size)
{
switch (cmd) {
case BPF_RAW_TRACEPOINT_OPEN:
bpf_printk("raw_tracepoint is %s", attr->raw_tracepoint.name);
break;
default:
break;
}
return 0;
}

The reason is that when accessing a field in a union, such as bpf_attr,
if the field is located within a nested struct that is not the first
member of the union, it can result in incorrect field verification.

  union bpf_attr {
      struct {
          __u32 map_type; <<<< Actually it will find that field.
          __u32 key_size;
          __u32 value_size;
         ...
      };
      ...
      struct {
          __u64 name;    <<<< We want to verify this field.
          __u32 prog_fd;
      } raw_tracepoint;
  };

Considering the potential deep nesting levels, finding a perfect
solution to address this issue has proven challenging. Therefore, I
propose a solution where we simply skip the verification process if the
field in question is located within a union.

Fixes: 7e3617a72df3 ("bpf: Add array support to btf_struct_access")
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/r/20230713025642.27477-4-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agoselftests/bpf: Add selftests for nested_trust
Yafang Shao [Thu, 13 Jul 2023 02:56:40 +0000 (02:56 +0000)]
selftests/bpf: Add selftests for nested_trust

Add selftests for nested_strust to check whehter PTR_UNTRUSTED is cleared
as expected, the result as follows:

 #141/1   nested_trust/test_read_cpumask:OK
 #141/2   nested_trust/test_skb_field:OK                    <<<<
 #141/3   nested_trust/test_invalid_nested_user_cpus:OK
 #141/4   nested_trust/test_invalid_nested_offset:OK
 #141/5   nested_trust/test_invalid_skb_field:OK            <<<<
 #141     nested_trust:OK

The #141/2 and #141/5 are newly added.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/r/20230713025642.27477-3-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agobpf: Fix an error around PTR_UNTRUSTED
Yafang Shao [Thu, 13 Jul 2023 02:56:39 +0000 (02:56 +0000)]
bpf: Fix an error around PTR_UNTRUSTED

Per discussion with Alexei, the PTR_UNTRUSTED flag should not been
cleared when we start to walk a new struct, because the struct in
question may be a struct nested in a union. We should also check and set
this flag before we walk its each member, in case itself is a union.
We will clear this flag if the field is BTF_TYPE_SAFE_RCU_OR_NULL.

Fixes: 6fcd486b3a0a ("bpf: Refactor RCU enforcement in the verifier.")
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/r/20230713025642.27477-2-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agoMerge branch 'bpf-x86-allow-function-arguments-up-to-12-for-tracing'
Alexei Starovoitov [Thu, 13 Jul 2023 23:04:57 +0000 (16:04 -0700)]
Merge branch 'bpf-x86-allow-function-arguments-up-to-12-for-tracing'

Menglong Dong says:

====================
bpf, x86: allow function arguments up to 12 for TRACING

From: Menglong Dong <imagedong@tencent.com>

For now, the BPF program of type BPF_PROG_TYPE_TRACING can only be used
on the kernel functions whose arguments count less than or equal to 6, if
not considering '> 8 bytes' struct argument. This is not friendly at all,
as too many functions have arguments count more than 6. According to the
current kernel version, below is a statistics of the function arguments
count:

argument count | function count
7              | 704
8              | 270
9              | 84
10             | 47
11             | 47
12             | 27
13             | 22
14             | 5
15             | 0
16             | 1

Therefore, let's enhance it by increasing the function arguments count
allowed in arch_prepare_bpf_trampoline(), for now, only x86_64.

In the 1st patch, we save/restore regs with BPF_DW size to make the code
in save_regs()/restore_regs() simpler.

In the 2nd patch, we make arch_prepare_bpf_trampoline() support to copy
function arguments in stack for x86 arch. Therefore, the maximum
arguments can be up to MAX_BPF_FUNC_ARGS for FENTRY, FEXIT and
MODIFY_RETURN. Meanwhile, we clean the potential garbage value when we
copy the arguments on-stack.

And the 3rd patch is for the testcases of the this series.

Changes since v9:
- fix the failed test cases of trampoline_count and get_func_args_test
  in the 3rd patch

Changes since v8:
- change the way to test fmod_ret in the 3rd patch

Changes since v7:
- split the testcases, and add fentry_many_args/fexit_many_args to
  DENYLIST.aarch64 in 3rd patch

Changes since v6:
- somit nits from commit message and comment in the 1st patch
- remove the inline in get_nr_regs() in the 1st patch
- rename some function and various in the 1st patch

Changes since v5:
- adjust the commit log of the 1st patch, avoiding confusing people that
  bugs exist in current code
- introduce get_nr_regs() to get the space that used to pass args on
  stack correct in the 2nd patch
- add testcases to tracing_struct.c instead of fentry_test.c and
  fexit_test.c

Changes since v4:
- consider the case of the struct in arguments can't be hold by regs
- add comment for some code
- add testcases for MODIFY_RETURN
- rebase to the latest

Changes since v3:
- try make the stack pointer 16-byte aligned. Not sure if I'm right :)
- introduce clean_garbage() to clean the grabage when argument count is 7
- use different data type in bpf_testmod_fentry_test{7,12}
- add testcase for grabage values in ctx

Changes since v2:
- keep MAX_BPF_FUNC_ARGS still
- clean garbage value in upper bytes in the 2nd patch
- move bpf_fentry_test{7,12} to bpf_testmod.c and rename them to
  bpf_testmod_fentry_test{7,12} meanwhile in the 3rd patch

Changes since v1:
- change the maximun function arguments to 14 from 12
- add testcases (Jiri Olsa)
- instead EMIT4 with EMIT3_off32 for "lea" to prevent overflow
====================

Link: https://lore.kernel.org/r/20230713040738.1789742-1-imagedong@tencent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agoselftests/bpf: add testcase for TRACING with 6+ arguments
Menglong Dong [Thu, 13 Jul 2023 04:07:38 +0000 (12:07 +0800)]
selftests/bpf: add testcase for TRACING with 6+ arguments

Add fentry_many_args.c and fexit_many_args.c to test the fentry/fexit
with 7/11 arguments. As this feature is not supported by arm64 yet, we
disable these testcases for arm64 in DENYLIST.aarch64. We can combine
them with fentry_test.c/fexit_test.c when arm64 is supported too.

Correspondingly, add bpf_testmod_fentry_test7() and
bpf_testmod_fentry_test11() to bpf_testmod.c

Meanwhile, add bpf_modify_return_test2() to test_run.c to test the
MODIFY_RETURN with 7 arguments.

Add bpf_testmod_test_struct_arg_7/bpf_testmod_test_struct_arg_7 in
bpf_testmod.c to test the struct in the arguments.

And the testcases passed on x86_64:

./test_progs -t fexit
Summary: 5/14 PASSED, 0 SKIPPED, 0 FAILED

./test_progs -t fentry
Summary: 3/2 PASSED, 0 SKIPPED, 0 FAILED

./test_progs -t modify_return
Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED

./test_progs -t tracing_struct
Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/20230713040738.1789742-4-imagedong@tencent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agobpf, x86: allow function arguments up to 12 for TRACING
Menglong Dong [Thu, 13 Jul 2023 04:07:37 +0000 (12:07 +0800)]
bpf, x86: allow function arguments up to 12 for TRACING

For now, the BPF program of type BPF_PROG_TYPE_TRACING can only be used
on the kernel functions whose arguments count less than or equal to 6, if
not considering '> 8 bytes' struct argument. This is not friendly at all,
as too many functions have arguments count more than 6.

According to the current kernel version, below is a statistics of the
function arguments count:

argument count | function count
7              | 704
8              | 270
9              | 84
10             | 47
11             | 47
12             | 27
13             | 22
14             | 5
15             | 0
16             | 1

Therefore, let's enhance it by increasing the function arguments count
allowed in arch_prepare_bpf_trampoline(), for now, only x86_64.

For the case that we don't need to call origin function, which means
without BPF_TRAMP_F_CALL_ORIG, we need only copy the function arguments
that stored in the frame of the caller to current frame. The 7th and later
arguments are stored in "$rbp + 0x18", and they will be copied to the
stack area following where register values are saved.

For the case with BPF_TRAMP_F_CALL_ORIG, we need prepare the arguments
in stack before call origin function, which means we need alloc extra
"8 * (arg_count - 6)" memory in the top of the stack. Note, there should
not be any data be pushed to the stack before calling the origin function.
So 'rbx' value will be stored on a stack position higher than where stack
arguments are stored for BPF_TRAMP_F_CALL_ORIG.

According to the research of Yonghong, struct members should be all in
register or all on the stack. Meanwhile, the compiler will pass the
argument on regs if the remaining regs can hold the argument. Therefore,
we need save the arguments in order. Otherwise, disorder of the args can
happen. For example:

  struct foo_struct {
      long a;
      int b;
  };
  int foo(char, char, char, char, char, struct foo_struct,
          char);

the arg1-5,arg7 will be passed by regs, and arg6 will by stack. Therefore,
we should save/restore the arguments in the same order with the
declaration of foo(). And the args used as ctx in stack will be like this:

  reg_arg6   -- copy from regs
  stack_arg2 -- copy from stack
  stack_arg1
  reg_arg5   -- copy from regs
  reg_arg4
  reg_arg3
  reg_arg2
  reg_arg1

We use EMIT3_off32() or EMIT4() for "lea" and "sub". The range of the
imm in "lea" and "sub" is [-128, 127] if EMIT4() is used. Therefore,
we use EMIT3_off32() instead if the imm out of the range.

It works well for the FENTRY/FEXIT/MODIFY_RETURN.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/20230713040738.1789742-3-imagedong@tencent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agobpf, x86: save/restore regs with BPF_DW size
Menglong Dong [Thu, 13 Jul 2023 04:07:36 +0000 (12:07 +0800)]
bpf, x86: save/restore regs with BPF_DW size

As we already reserve 8 byte in the stack for each reg, it is ok to
store/restore the regs in BPF_DW size. This will make the code in
save_regs()/restore_regs() simpler.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/20230713040738.1789742-2-imagedong@tencent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
17 months agoMerge tag 'net-6.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 13 Jul 2023 21:21:22 +0000 (14:21 -0700)]
Merge tag 'net-6.5-rc2' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Including fixes from netfilter, wireless and ebpf.

  Current release - regressions:

   - netfilter: conntrack: gre: don't set assured flag for clash entries

   - wifi: iwlwifi: remove 'use_tfh' config to fix crash

  Previous releases - regressions:

   - ipv6: fix a potential refcount underflow for idev

   - icmp6: ifix null-ptr-deref of ip6_null_entry->rt6i_idev in
     icmp6_dev()

   - bpf: fix max stack depth check for async callbacks

   - eth: mlx5e:
      - check for NOT_READY flag state after locking
      - fix page_pool page fragment tracking for XDP

   - eth: igc:
      - fix tx hang issue when QBV gate is closed
      - fix corner cases for TSN offload

   - eth: octeontx2-af: Move validation of ptp pointer before its usage

   - eth: ena: fix shift-out-of-bounds in exponential backoff

  Previous releases - always broken:

   - core: prevent skb corruption on frag list segmentation

   - sched:
      - cls_fw: fix improper refcount update leads to use-after-free
      - sch_qfq: account for stab overhead in qfq_enqueue

   - netfilter:
      - report use refcount overflow
      - prevent OOB access in nft_byteorder_eval

   - wifi: mt7921e: fix init command fail with enabled device

   - eth: ocelot: fix oversize frame dropping for preemptible TCs

   - eth: fec: recycle pages for transmitted XDP frames"

* tag 'net-6.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (79 commits)
  selftests: tc-testing: add test for qfq with stab overhead
  net/sched: sch_qfq: account for stab overhead in qfq_enqueue
  selftests: tc-testing: add tests for qfq mtu sanity check
  net/sched: sch_qfq: reintroduce lmax bound check for MTU
  wifi: cfg80211: fix receiving mesh packets without RFC1042 header
  wifi: rtw89: debug: fix error code in rtw89_debug_priv_send_h2c_set()
  net: txgbe: fix eeprom calculation error
  net/sched: make psched_mtu() RTNL-less safe
  net: ena: fix shift-out-of-bounds in exponential backoff
  netdevsim: fix uninitialized data in nsim_dev_trap_fa_cookie_write()
  net/sched: flower: Ensure both minimum and maximum ports are specified
  MAINTAINERS: Add another mailing list for QUALCOMM ETHQOS ETHERNET DRIVER
  docs: netdev: update the URL of the status page
  wifi: iwlwifi: remove 'use_tfh' config to fix crash
  xdp: use trusted arguments in XDP hints kfuncs
  bpf: cpumap: Fix memory leak in cpu_map_update_elem
  wifi: airo: avoid uninitialized warning in airo_get_rate()
  octeontx2-pf: Add additional check for MCAM rules
  net: dsa: Removed unneeded of_node_put in felix_parse_ports_node
  net: fec: use netdev_err_once() instead of netdev_err()
  ...

17 months agoMerge tag 'trace-v6.5-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/trace...
Linus Torvalds [Thu, 13 Jul 2023 20:44:28 +0000 (13:44 -0700)]
Merge tag 'trace-v6.5-rc1-3' of git://git./linux/kernel/git/trace/linux-trace

Pull tracing fixes from Steven Rostedt:

 - Fix some missing-prototype warnings

 - Fix user events struct args (did not include size of struct)

   When creating a user event, the "struct" keyword is to denote that
   the size of the field will be passed in. But the parsing failed to
   handle this case.

 - Add selftest to struct sizes for user events

 - Fix sample code for direct trampolines.

   The sample code for direct trampolines attached to handle_mm_fault().
   But the prototype changed and the direct trampoline sample code was
   not updated. Direct trampolines needs to have the arguments correct
   otherwise it can fail or crash the system.

 - Remove unused ftrace_regs_caller_ret() prototype.

 - Quiet false positive of FORTIFY_SOURCE

   Due to backward compatibility, the structure used to save stack
   traces in the kernel had a fixed size of 8. This structure is
   exported to user space via the tracing format file. A change was made
   to allow more than 8 functions to be recorded, and user space now
   uses the size field to know how many functions are actually in the
   stack.

   But the structure still has size of 8 (even though it points into the
   ring buffer that has the required amount allocated to hold a full
   stack.

   This was fine until the fortifier noticed that the
   memcpy(&entry->caller, stack, size) was greater than the 8 functions
   and would complain at runtime about it.

   Hide this by using a pointer to the stack location on the ring buffer
   instead of using the address of the entry structure caller field.

 - Fix a deadloop in reading trace_pipe that was caused by a mismatch
   between ring_buffer_empty() returning false which then asked to read
   the data, but the read code uses rb_num_of_entries() that returned
   zero, and causing a infinite "retry".

 - Fix a warning caused by not using all pages allocated to store ftrace
   functions, where this can happen if the linker inserts a bunch of
   "NULL" entries, causing the accounting of how many pages needed to be
   off.

 - Fix histogram synthetic event crashing when the start event is
   removed and the end event is still using a variable from it

 - Fix memory leak in freeing iter->temp in tracing_release_pipe()

* tag 'trace-v6.5-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  tracing: Fix memory leak of iter->temp when reading trace_pipe
  tracing/histograms: Add histograms to hist_vars if they have referenced variables
  tracing: Stop FORTIFY_SOURCE complaining about stack trace caller
  ftrace: Fix possible warning on checking all pages used in ftrace_process_locs()
  ring-buffer: Fix deadloop issue on reading trace_pipe
  tracing: arm64: Avoid missing-prototype warnings
  selftests/user_events: Test struct size match cases
  tracing/user_events: Fix struct arg size match check
  x86/ftrace: Remove unsued extern declaration ftrace_regs_caller_ret()
  arm64: ftrace: Add direct call trampoline samples support
  samples: ftrace: Save required argument registers in sample trampolines

17 months agoMerge tag 'for-linus-6.5-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 13 Jul 2023 20:39:36 +0000 (13:39 -0700)]
Merge tag 'for-linus-6.5-rc2-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fixes from Juergen Gross:

 - a cleanup of the Xen related ELF-notes

 - a fix for virtio handling in Xen dom0 when running Xen in a VM

* tag 'for-linus-6.5-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/virtio: Fix NULL deref when a bridge of PCI root bus has no parent
  x86/Xen: tidy xen-head.S

17 months agoMerge tag 'sh-for-v6.5-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/glaubit...
Linus Torvalds [Thu, 13 Jul 2023 20:34:00 +0000 (13:34 -0700)]
Merge tag 'sh-for-v6.5-tag2' of git://git./linux/kernel/git/glaubitz/sh-linux

Pull sh fixes from John Paul Adrian Glaubitz:
 "The sh updates introduced multiple regressions.

  In particular, the change a8ac2961148e ("sh: Avoid using IRQ0 on SH3
  and SH4") causes several boards to hang during boot due to incorrect
  IRQ numbers.

  Geert Uytterhoeven has contributed patches that handle the virq offset
  in the IRQ code for the dreamcast, highlander and r2d boards while
  Artur Rojek has contributed a patch which handles the virq offset for
  the hd64461 companion chip"

* tag 'sh-for-v6.5-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/glaubitz/sh-linux:
  sh: hd64461: Handle virq offset for offchip IRQ base and HD64461 IRQ
  sh: mach-dreamcast: Handle virq offset in cascaded IRQ demux
  sh: mach-highlander: Handle virq offset in cascaded IRL demux
  sh: mach-r2d: Handle virq offset in cascaded IRL demux

17 months agotracing: Fix memory leak of iter->temp when reading trace_pipe
Zheng Yejian [Thu, 13 Jul 2023 14:14:35 +0000 (22:14 +0800)]
tracing: Fix memory leak of iter->temp when reading trace_pipe

kmemleak reports:
  unreferenced object 0xffff88814d14e200 (size 256):
    comm "cat", pid 336, jiffies 4294871818 (age 779.490s)
    hex dump (first 32 bytes):
      04 00 01 03 00 00 00 00 08 00 00 00 00 00 00 00  ................
      0c d8 c8 9b ff ff ff ff 04 5a ca 9b ff ff ff ff  .........Z......
    backtrace:
      [<ffffffff9bdff18f>] __kmalloc+0x4f/0x140
      [<ffffffff9bc9238b>] trace_find_next_entry+0xbb/0x1d0
      [<ffffffff9bc9caef>] trace_print_lat_context+0xaf/0x4e0
      [<ffffffff9bc94490>] print_trace_line+0x3e0/0x950
      [<ffffffff9bc95499>] tracing_read_pipe+0x2d9/0x5a0
      [<ffffffff9bf03a43>] vfs_read+0x143/0x520
      [<ffffffff9bf04c2d>] ksys_read+0xbd/0x160
      [<ffffffff9d0f0edf>] do_syscall_64+0x3f/0x90
      [<ffffffff9d2000aa>] entry_SYSCALL_64_after_hwframe+0x6e/0xd8

when reading file 'trace_pipe', 'iter->temp' is allocated or relocated
in trace_find_next_entry() but not freed before 'trace_pipe' is closed.

To fix it, free 'iter->temp' in tracing_release_pipe().

Link: https://lore.kernel.org/linux-trace-kernel/20230713141435.1133021-1-zhengyejian1@huawei.com
Cc: stable@vger.kernel.org
Fixes: ff895103a84ab ("tracing: Save off entry when peeking at next entry")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
17 months agoMerge branch 'net-sched-fixes-for-sch_qfq'
Paolo Abeni [Thu, 13 Jul 2023 09:12:01 +0000 (11:12 +0200)]
Merge branch 'net-sched-fixes-for-sch_qfq'

Pedro Tammela says:

====================
net/sched: fixes for sch_qfq

Patch 1 fixes a regression introduced in 6.4 where the MTU size could be
bigger than 'lmax'.

Patch 3 fixes an issue where the code doesn't account for qdisc_pkt_len()
returning a size bigger then 'lmax'.

Patches 2 and 4 are selftests for the issues above.
====================

Link: https://lore.kernel.org/r/20230711210103.597831-1-pctammela@mojatatu.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
17 months agoselftests: tc-testing: add test for qfq with stab overhead
Pedro Tammela [Tue, 11 Jul 2023 21:01:03 +0000 (18:01 -0300)]
selftests: tc-testing: add test for qfq with stab overhead

A packet with stab overhead greater than QFQ_MAX_LMAX should be dropped
by the QFQ qdisc as it can't handle such lengths.

Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Tested-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
17 months agonet/sched: sch_qfq: account for stab overhead in qfq_enqueue
Pedro Tammela [Tue, 11 Jul 2023 21:01:02 +0000 (18:01 -0300)]
net/sched: sch_qfq: account for stab overhead in qfq_enqueue

Lion says:
-------
In the QFQ scheduler a similar issue to CVE-2023-31436
persists.

Consider the following code in net/sched/sch_qfq.c:

static int qfq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
                struct sk_buff **to_free)
{
     unsigned int len = qdisc_pkt_len(skb), gso_segs;

    // ...

     if (unlikely(cl->agg->lmax < len)) {
         pr_debug("qfq: increasing maxpkt from %u to %u for class %u",
              cl->agg->lmax, len, cl->common.classid);
         err = qfq_change_agg(sch, cl, cl->agg->class_weight, len);
         if (err) {
             cl->qstats.drops++;
             return qdisc_drop(skb, sch, to_free);
         }

    // ...

     }

Similarly to CVE-2023-31436, "lmax" is increased without any bounds
checks according to the packet length "len". Usually this would not
impose a problem because packet sizes are naturally limited.

This is however not the actual packet length, rather the
"qdisc_pkt_len(skb)" which might apply size transformations according to
"struct qdisc_size_table" as created by "qdisc_get_stab()" in
net/sched/sch_api.c if the TCA_STAB option was set when modifying the qdisc.

A user may choose virtually any size using such a table.

As a result the same issue as in CVE-2023-31436 can occur, allowing heap
out-of-bounds read / writes in the kmalloc-8192 cache.
-------

We can create the issue with the following commands:

tc qdisc add dev $DEV root handle 1: stab mtu 2048 tsize 512 mpu 0 \
overhead 999999999 linklayer ethernet qfq
tc class add dev $DEV parent 1: classid 1:1 htb rate 6mbit burst 15k
tc filter add dev $DEV parent 1: matchall classid 1:1
ping -I $DEV 1.1.1.2

This is caused by incorrectly assuming that qdisc_pkt_len() returns a
length within the QFQ_MIN_LMAX < len < QFQ_MAX_LMAX.

Fixes: 462dbc9101ac ("pkt_sched: QFQ Plus: fair-queueing service at DRR cost")
Reported-by: Lion <nnamrec@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
17 months agoselftests: tc-testing: add tests for qfq mtu sanity check
Pedro Tammela [Tue, 11 Jul 2023 21:01:01 +0000 (18:01 -0300)]
selftests: tc-testing: add tests for qfq mtu sanity check

QFQ only supports a certain bound of MTU size so make sure
we check for this requirement in the tests.

Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Tested-by: Zhengchao Shao <shaozhengchao@huawei.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
17 months agonet/sched: sch_qfq: reintroduce lmax bound check for MTU
Pedro Tammela [Tue, 11 Jul 2023 21:01:00 +0000 (18:01 -0300)]
net/sched: sch_qfq: reintroduce lmax bound check for MTU

25369891fcef deletes a check for the case where no 'lmax' is
specified which 3037933448f6 previously fixed as 'lmax'
could be set to the device's MTU without any bound checking
for QFQ_LMAX_MIN and QFQ_LMAX_MAX. Therefore, reintroduce the check.

Fixes: 25369891fcef ("net/sched: sch_qfq: refactor parsing of netlink parameters")
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
17 months agosh: hd64461: Handle virq offset for offchip IRQ base and HD64461 IRQ
Artur Rojek [Mon, 10 Jul 2023 23:31:32 +0000 (01:31 +0200)]
sh: hd64461: Handle virq offset for offchip IRQ base and HD64461 IRQ

A recent change to start counting SuperH IRQ #s from 16 breaks support
for the Hitachi HD64461 companion chip.

Move the offchip IRQ base and HD64461 IRQ # by 16 in order to
accommodate for the new virq numbering rules.

Fixes: a8ac2961148e ("sh: Avoid using IRQ0 on SH3 and SH4")
Signed-off-by: Artur Rojek <contact@artur-rojek.eu>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Link: https://lore.kernel.org/r/20230710233132.69734-1-contact@artur-rojek.eu
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
17 months agosh: mach-dreamcast: Handle virq offset in cascaded IRQ demux
Geert Uytterhoeven [Sun, 9 Jul 2023 13:10:43 +0000 (15:10 +0200)]
sh: mach-dreamcast: Handle virq offset in cascaded IRQ demux

Take into account the virq offset when translating cascaded interrupts.

Fixes: a8ac2961148e8c72 ("sh: Avoid using IRQ0 on SH3 and SH4")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Link: https://lore.kernel.org/r/7d0cb246c9f1cd24bb1f637ec5cb67e799a4c3b8.1688908227.git.geert+renesas@glider.be
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
17 months agosh: mach-highlander: Handle virq offset in cascaded IRL demux
Geert Uytterhoeven [Sun, 9 Jul 2023 13:10:23 +0000 (15:10 +0200)]
sh: mach-highlander: Handle virq offset in cascaded IRL demux

Take into account the virq offset when translating cascaded IRL
interrupts.

Fixes: a8ac2961148e8c72 ("sh: Avoid using IRQ0 on SH3 and SH4")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Link: https://lore.kernel.org/r/4fcb0d08a2b372431c41e04312742dc9e41e1be4.1688908186.git.geert+renesas@glider.be
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
17 months agosh: mach-r2d: Handle virq offset in cascaded IRL demux
Geert Uytterhoeven [Sun, 9 Jul 2023 11:15:49 +0000 (13:15 +0200)]
sh: mach-r2d: Handle virq offset in cascaded IRL demux

When booting rts7751r2dplus_defconfig on QEMU, the system hangs due to
an interrupt storm on IRQ 20.  IRQ 20 aka event 0x280 is a cascaded IRL
interrupt, which maps to IRQ_VOYAGER, the interrupt used by the Silicon
Motion SM501 multimedia companion chip.  As rts7751r2d_irq_demux() does
not take into account the new virq offset, the interrupt is no longer
translated, leading to an unhandled interrupt.

Fix this by taking into account the virq offset when translating
cascaded IRL interrupts.

Fixes: a8ac2961148e8c72 ("sh: Avoid using IRQ0 on SH3 and SH4")
Reported-by: Guenter Roeck <linux@roeck-us.net>
Closes: https://lore.kernel.org/r/fbfea3ad-d327-4ad5-ac9c-648c7ca3fe1f@roeck-us.net
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/2c99d5df41c40691f6c407b7b6a040d406bc81ac.1688901306.git.geert+renesas@glider.be
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
17 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 13 Jul 2023 01:13:57 +0000 (18:13 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Alexei Starovoitov says:

====================
pull-request: bpf 2023-07-12

We've added 5 non-merge commits during the last 7 day(s) which contain
a total of 7 files changed, 93 insertions(+), 28 deletions(-).

The main changes are:

1) Fix max stack depth check for async callbacks, from Kumar.

2) Fix inconsistent JIT image generation, from Björn.

3) Use trusted arguments in XDP hints kfuncs, from Larysa.

4) Fix memory leak in cpu_map_update_elem, from Pu.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  xdp: use trusted arguments in XDP hints kfuncs
  bpf: cpumap: Fix memory leak in cpu_map_update_elem
  riscv, bpf: Fix inconsistent JIT image generation
  selftests/bpf: Add selftest for check_stack_max_depth bug
  bpf: Fix max stack depth check for async callbacks
====================

Link: https://lore.kernel.org/r/20230712223045.40182-1-alexei.starovoitov@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agowifi: cfg80211: fix receiving mesh packets without RFC1042 header
Felix Fietkau [Tue, 11 Jul 2023 11:50:52 +0000 (13:50 +0200)]
wifi: cfg80211: fix receiving mesh packets without RFC1042 header

Fix ethernet header length field after stripping the mesh header

Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/all/CT5GNZSK28AI.2K6M69OXM9RW5@syracuse/
Fixes: 986e43b19ae9 ("wifi: mac80211: fix receiving A-MSDU frames on mesh interfaces")
Reported-and-tested-by: Nicolas Escande <nico.escande@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230711115052.68430-1-nbd@nbd.name
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agowifi: rtw89: debug: fix error code in rtw89_debug_priv_send_h2c_set()
Zhang Shurong [Thu, 6 Jul 2023 02:45:00 +0000 (10:45 +0800)]
wifi: rtw89: debug: fix error code in rtw89_debug_priv_send_h2c_set()

If there is a failure during rtw89_fw_h2c_raw() rtw89_debug_priv_send_h2c
should return negative error code instead of a positive value count.
Fix this bug by returning correct error code.

Fixes: e3ec7017f6a2 ("rtw89: add Realtek 802.11ax driver")
Signed-off-by: Zhang Shurong <zhang_shurong@foxmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://lore.kernel.org/r/tencent_AD09A61BC4DA92AD1EB0790F5C850E544D07@qq.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: wwan: t7xx: Add AP CLDMA
Jose Ignacio Tornos Martinez [Tue, 11 Jul 2023 06:28:13 +0000 (08:28 +0200)]
net: wwan: t7xx: Add AP CLDMA

At this moment with the current status, t7xx is not functional due to
problems like this after connection, if there is no activity:
[   57.370534] mtk_t7xx 0000:72:00.0: [PM] SAP suspend error: -110
[   57.370581] mtk_t7xx 0000:72:00.0: can't suspend
    (t7xx_pci_pm_runtime_suspend [mtk_t7xx] returned -110)
because after this, the traffic no longer works.

The complete series 'net: wwan: t7xx: fw flashing & coredump support'
was reverted because of issues with the pci implementation.
In order to have at least the modem working, it would be enough if just
the first commit of the series is re-applied:
d20ef656f994 net: wwan: t7xx: Add AP CLDMA
With that, the Application Processor would be controlled, correctly
suspended and the commented problems would be fixed (I am testing here
like this with no related issue).

This commit is independent of the others and not related to the
commented pci implementation for the new features: fw flashing and
coredump collection.

Use v2 patch version of d20ef656f994 as JinJian Song suggests
(https://patchwork.kernel.org/project/netdevbpf/patch/20230105154215.198828-1-m.chetan.kumar@linux.intel.com/).

Original text from the commit that would be re-applied:

    d20ef656f994 net: wwan: t7xx: Add AP CLDMA
    Author: Haijun Liu <haijun.liu@mediatek.com>
    Date:   Tue Aug 16 09:53:28 2022 +0530

    The t7xx device contains two Cross Layer DMA (CLDMA) interfaces to
    communicate with AP and Modem processors respectively. So far only
    MD-CLDMA was being used, this patch enables AP-CLDMA.

    Rename small Application Processor (sAP) to AP.

Signed-off-by: Haijun Liu <haijun.liu@mediatek.com>
Co-developed-by: Madhusmita Sahu <madhusmita.sahu@intel.com>
Signed-off-by: Madhusmita Sahu <madhusmita.sahu@intel.com>
Signed-off-by: Moises Veleta <moises.veleta@linux.intel.com>
Signed-off-by: Devegowda Chandrashekar <chandrashekar.devegowda@intel.com>
Signed-off-by: M Chetan Kumar <m.chetan.kumar@linux.intel.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Reviewed-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230711062817.6108-1-jtornosm@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoipv6: rpl: Remove redundant skb_dst_drop().
Kuniyuki Iwashima [Mon, 10 Jul 2023 21:35:11 +0000 (14:35 -0700)]
ipv6: rpl: Remove redundant skb_dst_drop().

RPL code has a pattern where skb_dst_drop() is called before
ip6_route_input().

However, ip6_route_input() calls skb_dst_drop() internally,
so we need not call skb_dst_drop() before ip6_route_input().

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20230710213511.5364-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge branch 'mlxsw-add-port-range-matching-support'
Jakub Kicinski [Wed, 12 Jul 2023 23:57:23 +0000 (16:57 -0700)]
Merge branch 'mlxsw-add-port-range-matching-support'

Petr Machata says:

====================
mlxsw: Add port range matching support

Ido Schimmel writes:

Add port range matching support in mlxsw as part of tc-flower offload.

Patches #1-#7 gradually add port range matching support in mlxsw. See
patch #3 to understand how port range matching is implemented in the
device.

Patches #8-#10 add selftests.
====================

Link: https://lore.kernel.org/r/cover.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: forwarding: Add test cases for flower port range matching
Ido Schimmel [Tue, 11 Jul 2023 16:44:03 +0000 (18:44 +0200)]
selftests: forwarding: Add test cases for flower port range matching

Add test cases to verify that flower port range matching works
correctly. Test both source and destination port ranges, with different
combinations of IPv4/IPv6 and TCP/UDP, on both ingress and egress.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/9d47c9cd4522b2d335b13ce8f6c9b33199298cee.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mlxsw: Test port range registers' occupancy
Ido Schimmel [Tue, 11 Jul 2023 16:44:02 +0000 (18:44 +0200)]
selftests: mlxsw: Test port range registers' occupancy

Test that filters that match on the same port range, but with different
combination of IPv4/IPv6 and TCP/UDP all use the same port range
register by observing port range registers' occupancy via
devlink-resource.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/0a2eb63b234fb062ff011e80231868cc80000c81.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoselftests: mlxsw: Add scale test for port ranges
Ido Schimmel [Tue, 11 Jul 2023 16:44:01 +0000 (18:44 +0200)]
selftests: mlxsw: Add scale test for port ranges

Query the maximum number of supported port range registers using
devlink-resource and test that this number can be reached by configuring
tc filters with different port ranges. Test that an error is returned in
case the maximum number is exceeded.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/48eee181270d9f291e09d1858c7b26a3f7fcc164.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: spectrum_flower: Add ability to match on port ranges
Ido Schimmel [Tue, 11 Jul 2023 16:44:00 +0000 (18:44 +0200)]
mlxsw: spectrum_flower: Add ability to match on port ranges

Add the ability to match on port ranges by utilizing the previously
added port range registers and the port range key element. Up to two
port range registers can be used for each filter, one for source port
and another for destination port.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/df4385a9592917e9a22ebff339e0463e4a8dfa82.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: spectrum_acl: Pass main driver structure to mlxsw_sp_acl_rulei_destroy()
Ido Schimmel [Tue, 11 Jul 2023 16:43:59 +0000 (18:43 +0200)]
mlxsw: spectrum_acl: Pass main driver structure to mlxsw_sp_acl_rulei_destroy()

The main driver structure will be needed in this function by a
subsequent patch, so pass it. No functional changes intended.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/24d96a4e21310e5de2951ace58263db35e44a0df.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: spectrum_acl: Add port range key element
Ido Schimmel [Tue, 11 Jul 2023 16:43:58 +0000 (18:43 +0200)]
mlxsw: spectrum_acl: Add port range key element

Add the port range key element to supported key blocks so that it could
be used to match on the output of the port range registers. Each bit in
the element can be used to match on the output of the port range
register with the corresponding index.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/f0423f6ee9e36c6b0a426bc9995f42223c48f2db.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: spectrum_port_range: Add devlink resource support
Ido Schimmel [Tue, 11 Jul 2023 16:43:57 +0000 (18:43 +0200)]
mlxsw: spectrum_port_range: Add devlink resource support

Expose via devlink-resource the maximum number of port range registers
and their current occupancy. Besides the observability benefits, this
resource will be used by subsequent patches for scale and occupancy
tests.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/7945e0c715dc5efb1617f45f7560c1f1bd0bcf8a.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: spectrum_port_range: Add port range core
Ido Schimmel [Tue, 11 Jul 2023 16:43:56 +0000 (18:43 +0200)]
mlxsw: spectrum_port_range: Add port range core

The Spectrum ASICs have a fixed number of port range registers, each of
which maintains the following parameters:

* Minimum and maximum port.
* Apply port range for source port, destination port or both.
* Apply port range for TCP, UDP or both.
* Apply port range for IPv4, IPv6 or both.

Implement a port range core which takes care of the allocation and
configuration of these registers and exposes an API that allows
in-driver consumers (e.g., the ACL code) to request matching on a range
of either source or destination port.

These registers are going to be used for port range matching in the
flower classifier that already matches on EtherType being IPv4 / IPv6 and
IP protocol being TCP / UDP. As such, there is no need to limit these
registers to a specific EtherType or IP protocol, which will increase
the likelihood of a register being shared by multiple flower filters.

It is unlikely that a filter will match on the same range of both source
and destination ports, which is why each register is only configured to
match on either source or destination port. If a filter requires
matching on a range of both source and destination ports, it will
utilize two port range registers and match on the output of both.

For efficient lookup and traversal, use XArray to store the allocated
port range registers. The XArray uses RCU and an internal spinlock to
synchronise access, so there is no need for a dedicate lock.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/674f00539a0072d455847663b5feb504db51a259.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: resource: Add resource identifier for port range registers
Ido Schimmel [Tue, 11 Jul 2023 16:43:55 +0000 (18:43 +0200)]
mlxsw: resource: Add resource identifier for port range registers

Add a resource identifier for maximum number of layer 4 port range
register so that it could be later used to query the information from
firmware.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/59a8fec353d5ad9fbfb7612e4a7ff61eaedad445.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agomlxsw: reg: Add Policy-Engine Port Range Register
Ido Schimmel [Tue, 11 Jul 2023 16:43:54 +0000 (18:43 +0200)]
mlxsw: reg: Add Policy-Engine Port Range Register

Add the Policy-Engine Port Range Register that is used for configuring
port range identification.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/d1a1f53d758f7452cf5abfe006b23496076ec3e6.1689092769.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: txgbe: fix eeprom calculation error
Jiawen Wu [Tue, 11 Jul 2023 06:34:14 +0000 (14:34 +0800)]
net: txgbe: fix eeprom calculation error

For some device types like TXGBE_ID_XAUI, *checksum computed in
txgbe_calc_eeprom_checksum() is larger than TXGBE_EEPROM_SUM. Remove the
limit on the size of *checksum.

Fixes: 049fe5365324 ("net: txgbe: Add operations to interact with firmware")
Fixes: 5e2ea7801fac ("net: txgbe: Fix unsigned comparison to zero in txgbe_calc_eeprom_checksum()")
Signed-off-by: Jiawen Wu <jiawenwu@trustnetic.com>
Link: https://lore.kernel.org/r/20230711063414.3311-1-jiawenwu@trustnetic.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge tag 'for-linus' of https://github.com/openrisc/linux
Linus Torvalds [Wed, 12 Jul 2023 23:28:53 +0000 (16:28 -0700)]
Merge tag 'for-linus' of https://github.com/openrisc/linux

Pull OpenRISC fix from Stafford Horne:

 - During the 6.4 cycle my fpu support work broke ABI compatibility in
   the sigcontext struct. This was noticed by musl libc developers after
   the release. This fix restores the ABI.

* tag 'for-linus' of https://github.com/openrisc/linux:
  openrisc: Union fpcsr and oldmask in sigcontext to unbreak userspace ABI

17 months agotracing/histograms: Add histograms to hist_vars if they have referenced variables
Mohamed Khalfella [Wed, 12 Jul 2023 22:30:21 +0000 (22:30 +0000)]
tracing/histograms: Add histograms to hist_vars if they have referenced variables

Hist triggers can have referenced variables without having direct
variables fields. This can be the case if referenced variables are added
for trigger actions. In this case the newly added references will not
have field variables. Not taking such referenced variables into
consideration can result in a bug where it would be possible to remove
hist trigger with variables being refenced. This will result in a bug
that is easily reproducable like so

$ cd /sys/kernel/tracing
$ echo 'synthetic_sys_enter char[] comm; long id' >> synthetic_events
$ echo 'hist:keys=common_pid.execname,id.syscall:vals=hitcount:comm=common_pid.execname' >> events/raw_syscalls/sys_enter/trigger
$ echo 'hist:keys=common_pid.execname,id.syscall:onmatch(raw_syscalls.sys_enter).synthetic_sys_enter($comm, id)' >> events/raw_syscalls/sys_enter/trigger
$ echo '!hist:keys=common_pid.execname,id.syscall:vals=hitcount:comm=common_pid.execname' >> events/raw_syscalls/sys_enter/trigger

[  100.263533] ==================================================================
[  100.264634] BUG: KASAN: slab-use-after-free in resolve_var_refs+0xc7/0x180
[  100.265520] Read of size 8 at addr ffff88810375d0f0 by task bash/439
[  100.266320]
[  100.266533] CPU: 2 PID: 439 Comm: bash Not tainted 6.5.0-rc1 #4
[  100.267277] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-20220807_005459-localhost 04/01/2014
[  100.268561] Call Trace:
[  100.268902]  <TASK>
[  100.269189]  dump_stack_lvl+0x4c/0x70
[  100.269680]  print_report+0xc5/0x600
[  100.270165]  ? resolve_var_refs+0xc7/0x180
[  100.270697]  ? kasan_complete_mode_report_info+0x80/0x1f0
[  100.271389]  ? resolve_var_refs+0xc7/0x180
[  100.271913]  kasan_report+0xbd/0x100
[  100.272380]  ? resolve_var_refs+0xc7/0x180
[  100.272920]  __asan_load8+0x71/0xa0
[  100.273377]  resolve_var_refs+0xc7/0x180
[  100.273888]  event_hist_trigger+0x749/0x860
[  100.274505]  ? kasan_save_stack+0x2a/0x50
[  100.275024]  ? kasan_set_track+0x29/0x40
[  100.275536]  ? __pfx_event_hist_trigger+0x10/0x10
[  100.276138]  ? ksys_write+0xd1/0x170
[  100.276607]  ? do_syscall_64+0x3c/0x90
[  100.277099]  ? entry_SYSCALL_64_after_hwframe+0x6e/0xd8
[  100.277771]  ? destroy_hist_data+0x446/0x470
[  100.278324]  ? event_hist_trigger_parse+0xa6c/0x3860
[  100.278962]  ? __pfx_event_hist_trigger_parse+0x10/0x10
[  100.279627]  ? __kasan_check_write+0x18/0x20
[  100.280177]  ? mutex_unlock+0x85/0xd0
[  100.280660]  ? __pfx_mutex_unlock+0x10/0x10
[  100.281200]  ? kfree+0x7b/0x120
[  100.281619]  ? ____kasan_slab_free+0x15d/0x1d0
[  100.282197]  ? event_trigger_write+0xac/0x100
[  100.282764]  ? __kasan_slab_free+0x16/0x20
[  100.283293]  ? __kmem_cache_free+0x153/0x2f0
[  100.283844]  ? sched_mm_cid_remote_clear+0xb1/0x250
[  100.284550]  ? __pfx_sched_mm_cid_remote_clear+0x10/0x10
[  100.285221]  ? event_trigger_write+0xbc/0x100
[  100.285781]  ? __kasan_check_read+0x15/0x20
[  100.286321]  ? __bitmap_weight+0x66/0xa0
[  100.286833]  ? _find_next_bit+0x46/0xe0
[  100.287334]  ? task_mm_cid_work+0x37f/0x450
[  100.287872]  event_triggers_call+0x84/0x150
[  100.288408]  trace_event_buffer_commit+0x339/0x430
[  100.289073]  ? ring_buffer_event_data+0x3f/0x60
[  100.292189]  trace_event_raw_event_sys_enter+0x8b/0xe0
[  100.295434]  syscall_trace_enter.constprop.0+0x18f/0x1b0
[  100.298653]  syscall_enter_from_user_mode+0x32/0x40
[  100.301808]  do_syscall_64+0x1a/0x90
[  100.304748]  entry_SYSCALL_64_after_hwframe+0x6e/0xd8
[  100.307775] RIP: 0033:0x7f686c75c1cb
[  100.310617] Code: 73 01 c3 48 8b 0d 65 3c 10 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa b8 21 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 35 3c 10 00 f7 d8 64 89 01 48
[  100.317847] RSP: 002b:00007ffc60137a38 EFLAGS: 00000246 ORIG_RAX: 0000000000000021
[  100.321200] RAX: ffffffffffffffda RBX: 000055f566469ea0 RCX: 00007f686c75c1cb
[  100.324631] RDX: 0000000000000001 RSI: 0000000000000001 RDI: 000000000000000a
[  100.328104] RBP: 00007ffc60137ac0 R08: 00007f686c818460 R09: 000000000000000a
[  100.331509] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000009
[  100.334992] R13: 0000000000000007 R14: 000000000000000a R15: 0000000000000007
[  100.338381]  </TASK>

We hit the bug because when second hist trigger has was created
has_hist_vars() returned false because hist trigger did not have
variables. As a result of that save_hist_vars() was not called to add
the trigger to trace_array->hist_vars. Later on when we attempted to
remove the first histogram find_any_var_ref() failed to detect it is
being used because it did not find the second trigger in hist_vars list.

With this change we wait until trigger actions are created so we can take
into consideration if hist trigger has variable references. Also, now we
check the return value of save_hist_vars() and fail trigger creation if
save_hist_vars() fails.

Link: https://lore.kernel.org/linux-trace-kernel/20230712223021.636335-1-mkhalfella@purestorage.com
Cc: stable@vger.kernel.org
Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers")
Signed-off-by: Mohamed Khalfella <mkhalfella@purestorage.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
17 months agonet/sched: make psched_mtu() RTNL-less safe
Pedro Tammela [Tue, 11 Jul 2023 02:16:34 +0000 (23:16 -0300)]
net/sched: make psched_mtu() RTNL-less safe

Eric Dumazet says[1]:
-------
Speaking of psched_mtu(), I see that net/sched/sch_pie.c is using it
without holding RTNL, so dev->mtu can be changed underneath.
KCSAN could issue a warning.
-------

Annotate dev->mtu with READ_ONCE() so KCSAN don't issue a warning.

[1] https://lore.kernel.org/all/CANn89iJoJO5VtaJ-2=_d2aOQhb0Xw8iBT_Cxqp2HyuS-zj6azw@mail.gmail.com/

v1 -> v2: Fix commit message

Fixes: d4b36210c2e6 ("net: pkt_sched: PIE AQM scheme")
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230711021634.561598-1-pctammela@mojatatu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: ena: fix shift-out-of-bounds in exponential backoff
Krister Johansen [Tue, 11 Jul 2023 01:36:21 +0000 (18:36 -0700)]
net: ena: fix shift-out-of-bounds in exponential backoff

The ENA adapters on our instances occasionally reset.  Once recently
logged a UBSAN failure to console in the process:

  UBSAN: shift-out-of-bounds in build/linux/drivers/net/ethernet/amazon/ena/ena_com.c:540:13
  shift exponent 32 is too large for 32-bit type 'unsigned int'
  CPU: 28 PID: 70012 Comm: kworker/u72:2 Kdump: loaded not tainted 5.15.117
  Hardware name: Amazon EC2 c5d.9xlarge/, BIOS 1.0 10/16/2017
  Workqueue: ena ena_fw_reset_device [ena]
  Call Trace:
  <TASK>
  dump_stack_lvl+0x4a/0x63
  dump_stack+0x10/0x16
  ubsan_epilogue+0x9/0x36
  __ubsan_handle_shift_out_of_bounds.cold+0x61/0x10e
  ? __const_udelay+0x43/0x50
  ena_delay_exponential_backoff_us.cold+0x16/0x1e [ena]
  wait_for_reset_state+0x54/0xa0 [ena]
  ena_com_dev_reset+0xc8/0x110 [ena]
  ena_down+0x3fe/0x480 [ena]
  ena_destroy_device+0xeb/0xf0 [ena]
  ena_fw_reset_device+0x30/0x50 [ena]
  process_one_work+0x22b/0x3d0
  worker_thread+0x4d/0x3f0
  ? process_one_work+0x3d0/0x3d0
  kthread+0x12a/0x150
  ? set_kthread_struct+0x50/0x50
  ret_from_fork+0x22/0x30
  </TASK>

Apparently, the reset delays are getting so large they can trigger a
UBSAN panic.

Looking at the code, the current timeout is capped at 5000us.  Using a
base value of 100us, the current code will overflow after (1<<29).  Even
at values before 32, this function wraps around, perhaps
unintentionally.

Cap the value of the exponent used for this backoff at (1<<16) which is
larger than currently necessary, but large enough to support bigger
values in the future.

Cc: stable@vger.kernel.org
Fixes: 4bb7f4cf60e3 ("net: ena: reduce driver load time")
Signed-off-by: Krister Johansen <kjlx@templeofstupid.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Shay Agroskin <shayagr@amazon.com>
Link: https://lore.kernel.org/r/20230711013621.GE1926@templeofstupid.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agobpftool: Use "fallthrough;" keyword instead of comments
Quentin Monnet [Wed, 12 Jul 2023 15:23:22 +0000 (16:23 +0100)]
bpftool: Use "fallthrough;" keyword instead of comments

After using "__fallthrough;" in a switch/case block in bpftool's
btf_dumper.c [0], and then turning it into a comment [1] to prevent a
merge conflict in linux-next when the keyword was changed into just
"fallthrough;" [2], we can now drop the comment and use the new keyword,
no underscores.

Also update the other occurrence of "/* fallthrough */" in bpftool.

[0] commit 9fd496848b1c ("bpftool: Support inline annotations when dumping the CFG of a program")
[1] commit 4b7ef71ac977 ("bpftool: Replace "__fallthrough" by a comment to address merge conflict")
[2] commit f7a858bffcdd ("tools: Rename __fallthrough to fallthrough")

Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230712152322.81758-1-quentin@isovalent.com
17 months agotracing: Stop FORTIFY_SOURCE complaining about stack trace caller
Steven Rostedt (Google) [Wed, 12 Jul 2023 14:52:35 +0000 (10:52 -0400)]
tracing: Stop FORTIFY_SOURCE complaining about stack trace caller

The stack_trace event is an event created by the tracing subsystem to
store stack traces. It originally just contained a hard coded array of 8
words to hold the stack, and a "size" to know how many entries are there.
This is exported to user space as:

name: kernel_stack
ID: 4
format:
field:unsigned short common_type; offset:0; size:2; signed:0;
field:unsigned char common_flags; offset:2; size:1; signed:0;
field:unsigned char common_preempt_count; offset:3; size:1; signed:0;
field:int common_pid; offset:4; size:4; signed:1;

field:int size; offset:8; size:4; signed:1;
field:unsigned long caller[8]; offset:16; size:64; signed:0;

print fmt: "\t=> %ps\n\t=> %ps\n\t=> %ps\n" "\t=> %ps\n\t=> %ps\n\t=> %ps\n" "\t=> %ps\n\t=> %ps\n",i
 (void *)REC->caller[0], (void *)REC->caller[1], (void *)REC->caller[2],
 (void *)REC->caller[3], (void *)REC->caller[4], (void *)REC->caller[5],
 (void *)REC->caller[6], (void *)REC->caller[7]

Where the user space tracers could parse the stack. The library was
updated for this specific event to only look at the size, and not the
array. But some older users still look at the array (note, the older code
still checks to make sure the array fits inside the event that it read.
That is, if only 4 words were saved, the parser would not read the fifth
word because it will see that it was outside of the event size).

This event was changed a while ago to be more dynamic, and would save a
full stack even if it was greater than 8 words. It does this by simply
allocating more ring buffer to hold the extra words. Then it copies in the
stack via:

memcpy(&entry->caller, fstack->calls, size);

As the entry is struct stack_entry, that is created by a macro to both
create the structure and export this to user space, it still had the caller
field of entry defined as: unsigned long caller[8].

When the stack is greater than 8, the FORTIFY_SOURCE code notices that the
amount being copied is greater than the source array and complains about
it. It has no idea that the source is pointing to the ring buffer with the
required allocation.

To hide this from the FORTIFY_SOURCE logic, pointer arithmetic is used:

ptr = ring_buffer_event_data(event);
entry = ptr;
ptr += offsetof(typeof(*entry), caller);
memcpy(ptr, fstack->calls, size);

Link: https://lore.kernel.org/all/20230612160748.4082850-1-svens@linux.ibm.com/
Link: https://lore.kernel.org/linux-trace-kernel/20230712105235.5fc441aa@gandalf.local.home
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Reported-by: Sven Schnelle <svens@linux.ibm.com>
Tested-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
17 months agoftrace: Fix possible warning on checking all pages used in ftrace_process_locs()
Zheng Yejian [Wed, 12 Jul 2023 06:04:52 +0000 (14:04 +0800)]
ftrace: Fix possible warning on checking all pages used in ftrace_process_locs()

As comments in ftrace_process_locs(), there may be NULL pointers in
mcount_loc section:
 > Some architecture linkers will pad between
 > the different mcount_loc sections of different
 > object files to satisfy alignments.
 > Skip any NULL pointers.

After commit 20e5227e9f55 ("ftrace: allow NULL pointers in mcount_loc"),
NULL pointers will be accounted when allocating ftrace pages but skipped
before adding into ftrace pages, this may result in some pages not being
used. Then after commit 706c81f87f84 ("ftrace: Remove extra helper
functions"), warning may occur at:
  WARN_ON(pg->next);

To fix it, only warn for case that no pointers skipped but pages not used
up, then free those unused pages after releasing ftrace_lock.

Link: https://lore.kernel.org/linux-trace-kernel/20230712060452.3175675-1-zhengyejian1@huawei.com
Cc: stable@vger.kernel.org
Fixes: 706c81f87f84 ("ftrace: Remove extra helper functions")
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
17 months agotcp: add a scheduling point in established_get_first()
Jian Wen [Tue, 11 Jul 2023 03:24:05 +0000 (11:24 +0800)]
tcp: add a scheduling point in established_get_first()

Kubernetes[1] is going to stick with /proc/net/tcp for a while.

This commit reduces the scheduling latency introduced by
established_get_first(), similar to commit acffb584cda7 ("net: diag:
add a scheduling point in inet_diag_dump_icsk()").

In our environment, the scheduling latency affects the performance of
latency-sensitive services like Redis.

Changes in V2 :
 - call cond_resched() before checking if a bucket is empty as
   suggested by Eric Dumazet
 - removed the delay of synchronize_net() from the commit message

[1] https://github.com/google/cadvisor/blob/v0.47.2/container/libcontainer/handler.go#L130

Signed-off-by: Jian Wen <wenjian1@xiaomi.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230711032405.3253025-1-wenjian1@xiaomi.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agoMerge branch 'bpf-mem-cache-free-rcu'
Daniel Borkmann [Wed, 12 Jul 2023 21:45:23 +0000 (23:45 +0200)]
Merge branch 'bpf-mem-cache-free-rcu'

Alexei Starovoitov says:

====================
v3->v4:
- extra patch 14 from Hou to check for object leaks.
- fixed the race/leak in free_by_rcu_ttrace. Extra hunk in patch 8.
- added Acks and fixed typos.

v2->v3:
- dropped _tail optimization for free_by_rcu_ttrace
- new patch 5 to refactor inc/dec of c->active
- change 'draining' logic in patch 7
- add rcu_barrier in patch 12
- __llist_add-> llist_add(waiting_for_gp_ttrace) in patch 9 to fix race
- David's Ack in patch 13 and explanation that migrate_disable cannot be removed just yet.

v1->v2:
- Fixed race condition spotted by Hou. Patch 7.

v1:

Introduce bpf_mem_cache_free_rcu() that is similar to kfree_rcu except
the objects will go through an additional RCU tasks trace grace period
before being freed into slab.

Patches 1-9 - a bunch of prep work
Patch 10 - a patch from Paul that exports rcu_request_urgent_qs_task().
Patch 12 - the main bpf_mem_cache_free_rcu patch.
Patch 13 - use it in bpf_cpumask.

bpf_local_storage, bpf_obj_drop, qp-trie will be other users eventually.

With additional hack patch to htab that replaces bpf_mem_cache_free with bpf_mem_cache_free_rcu
the following are benchmark results:
- map_perf_test 4 8 16348 1000000
drops from 800k to 600k. Waiting for RCU GP makes objects cache cold.

- bench htab-mem -a -p 8
20% drop in performance and big increase in memory. From 3 Mbyte to 50 Mbyte. As expected.

- bench htab-mem -a -p 16 --use-case add_del_on_diff_cpu
Same performance and better memory consumption.
Before these patches this bench would OOM (with or without 'reuse after GP')
Patch 8 addresses the issue.

At the end the performance drop and additional memory consumption due to _rcu()
were expected and came out to be within reasonable margin.
Without Paul's patch 10 the memory consumption in 'bench htab-mem' is in Gbytes
which wouldn't be acceptable.

Patch 8 is a heuristic to address 'alloc on one cpu, free on another' issue.
It works well in practice. One can probably construct an artificial benchmark
to make heuristic ineffective, but we have to trade off performance, code complexity,
and memory consumption.

The life cycle of objects:
alloc: dequeue free_llist
free: enqeueu free_llist
free_llist above high watermark -> free_by_rcu_ttrace
free_rcu: enqueue free_by_rcu -> waiting_for_gp
after RCU GP waiting_for_gp -> free_by_rcu_ttrace
free_by_rcu_ttrace -> waiting_for_gp_ttrace -> slab
====================

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
17 months agobpf: Add object leak check.
Hou Tao [Thu, 6 Jul 2023 03:34:47 +0000 (20:34 -0700)]
bpf: Add object leak check.

The object leak check is cheap. Do it unconditionally to spot difficult races
in bpf_mem_alloc.

Signed-off-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230706033447.54696-15-alexei.starovoitov@gmail.com
17 months agobpf: Convert bpf_cpumask to bpf_mem_cache_free_rcu.
Alexei Starovoitov [Thu, 6 Jul 2023 03:34:46 +0000 (20:34 -0700)]
bpf: Convert bpf_cpumask to bpf_mem_cache_free_rcu.

Convert bpf_cpumask to bpf_mem_cache_free_rcu.
Note that migrate_disable() in bpf_cpumask_release() is still necessary, since
bpf_cpumask_release() is a dtor. bpf_obj_free_fields() can be converted to do
migrate_disable() there in a follow up.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/bpf/20230706033447.54696-14-alexei.starovoitov@gmail.com
17 months agobpf: Introduce bpf_mem_free_rcu() similar to kfree_rcu().
Alexei Starovoitov [Thu, 6 Jul 2023 03:34:45 +0000 (20:34 -0700)]
bpf: Introduce bpf_mem_free_rcu() similar to kfree_rcu().

Introduce bpf_mem_[cache_]free_rcu() similar to kfree_rcu().
Unlike bpf_mem_[cache_]free() that links objects for immediate reuse into
per-cpu free list the _rcu() flavor waits for RCU grace period and then moves
objects into free_by_rcu_ttrace list where they are waiting for RCU
task trace grace period to be freed into slab.

The life cycle of objects:
alloc: dequeue free_llist
free: enqeueu free_llist
free_rcu: enqueue free_by_rcu -> waiting_for_gp
free_llist above high watermark -> free_by_rcu_ttrace
after RCU GP waiting_for_gp -> free_by_rcu_ttrace
free_by_rcu_ttrace -> waiting_for_gp_ttrace -> slab

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/bpf/20230706033447.54696-13-alexei.starovoitov@gmail.com
17 months agoselftests/bpf: Improve test coverage of bpf_mem_alloc.
Alexei Starovoitov [Thu, 6 Jul 2023 03:34:44 +0000 (20:34 -0700)]
selftests/bpf: Improve test coverage of bpf_mem_alloc.

bpf_obj_new() calls bpf_mem_alloc(), but doing alloc/free of 8 elements
is not triggering watermark conditions in bpf_mem_alloc.
Increase to 200 elements to make sure alloc_bulk/free_bulk is exercised.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/bpf/20230706033447.54696-12-alexei.starovoitov@gmail.com
17 months agorcu: Export rcu_request_urgent_qs_task()
Paul E. McKenney [Thu, 6 Jul 2023 03:34:43 +0000 (20:34 -0700)]
rcu: Export rcu_request_urgent_qs_task()

If a CPU is executing a long series of non-sleeping system calls,
RCU grace periods can be delayed for on the order of a couple hundred
milliseconds.  This is normally not a problem, but if each system call
does a call_rcu(), those callbacks can stack up.  RCU will eventually
notice this callback storm, but use of rcu_request_urgent_qs_task()
allows the code invoking call_rcu() to give RCU a heads up.

This function is not for general use, not yet, anyway.

Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230706033447.54696-11-alexei.starovoitov@gmail.com
17 months agobpf: Allow reuse from waiting_for_gp_ttrace list.
Alexei Starovoitov [Thu, 6 Jul 2023 03:34:42 +0000 (20:34 -0700)]
bpf: Allow reuse from waiting_for_gp_ttrace list.

alloc_bulk() can reuse elements from free_by_rcu_ttrace.
Let it reuse from waiting_for_gp_ttrace as well to avoid unnecessary kmalloc().

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230706033447.54696-10-alexei.starovoitov@gmail.com
17 months agobpf: Add a hint to allocated objects.
Alexei Starovoitov [Thu, 6 Jul 2023 03:34:41 +0000 (20:34 -0700)]
bpf: Add a hint to allocated objects.

To address OOM issue when one cpu is allocating and another cpu is freeing add
a target bpf_mem_cache hint to allocated objects and when local cpu free_llist
overflows free to that bpf_mem_cache. The hint addresses the OOM while
maintaining the same performance for common case when alloc/free are done on the
same cpu.

Note that do_call_rcu_ttrace() now has to check 'draining' flag in one more case,
since do_call_rcu_ttrace() is called not only for current cpu.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/bpf/20230706033447.54696-9-alexei.starovoitov@gmail.com