linux-2.6-microblaze.git
17 months agoMerge branch 'net-pcs-add-helpers-to-xpcs-and-lynx-to-manage-mdiodev'
Jakub Kicinski [Tue, 30 May 2023 04:46:55 +0000 (21:46 -0700)]
Merge branch 'net-pcs-add-helpers-to-xpcs-and-lynx-to-manage-mdiodev'

Russell King says:

====================
net: pcs: add helpers to xpcs and lynx to manage mdiodev

This morning, we have had two instances where the destruction of the
MDIO device associated with XPCS and Lynx has been wrong. Rather than
allowing this pattern of errors to continue, let's make it easier for
driver authors to get this right by adding a helper.

The changes are essentially:

1. Add two new mdio device helpers to manage the underlying struct
   device reference count. Note that the existing mdio_device_free()
   doesn't actually free anything, it merely puts the reference count.

2. Make the existing _create() and _destroy() PCS driver methods
   increment and decrement this refcount using these helpers. This
   results in no overall change, although drivers may hang on to
   the mdio device for a few cycles longer.

3. Add _create_mdiodev() which creates the mdio device before calling
   the existing _create() method. Once the _create() method has
   returned, we put the reference count on the mdio device.

   If _create() was successful, then the reference count taken there
   will "hold" the mdio device for the lifetime of the PCS (in other
   words, until _destroy() is called.) However, if _create() failed,
   then dropping the refcount at this point will free the mdio device.

   This is the exact behaviour we desire.

4. Convert users that create a mdio device and then call the PCS's
   _create() method over to the new _create_mdiodev() method, and
   simplify the cleanup.

We also have DPAA2 and fmem_memac that look up their PCS rather than
creating it. These could also drop their reference count on the MDIO
device immediately after calling lynx_pcs_create(), which would then
mean we wouldn't need lynx_get_mdio_device() and the associated
complexity to put the device in dpaa2_pcs_destroy() and pcs_put().
Note that DPAA2 bypasses the mdio device's abstractions by calling
put_device() directly.
====================

Link: https://lore.kernel.org/r/ZHCGZ8IgAAwr8bla@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: enetc: use lynx_pcs_create_mdiodev()
Russell King (Oracle) [Fri, 26 May 2023 10:14:50 +0000 (11:14 +0100)]
net: enetc: use lynx_pcs_create_mdiodev()

Use the newly introduced lynx_pcs_create_mdiodev() which simplifies the
creation and destruction of the lynx PCS.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: dsa: ocelot: use lynx_pcs_create_mdiodev()
Russell King (Oracle) [Fri, 26 May 2023 10:14:44 +0000 (11:14 +0100)]
net: dsa: ocelot: use lynx_pcs_create_mdiodev()

Use the newly introduced lynx_pcs_create_mdiodev() which simplifies the
creation and destruction of the lynx PCS.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: pcs: lynx: add lynx_pcs_create_mdiodev()
Russell King (Oracle) [Fri, 26 May 2023 10:14:39 +0000 (11:14 +0100)]
net: pcs: lynx: add lynx_pcs_create_mdiodev()

Add lynx_pcs_create_mdiodev() to simplify the creation of the mdio
device associated with lynx PCS. In order to allow lynx_pcs_destroy()
to clean this up, we need to arrange for lynx_pcs_create() to take a
refcount on the mdiodev, and lynx_pcs_destroy() to put it.

Adding the refcounting to lynx_pcs_create()..lynx_pcs_destroy() will
be transparent to existing users of these interfaces.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: stmmac: use xpcs_create_mdiodev()
Russell King (Oracle) [Fri, 26 May 2023 10:14:34 +0000 (11:14 +0100)]
net: stmmac: use xpcs_create_mdiodev()

Use the new xpcs_create_mdiodev() creator, which simplifies the
creation and destruction of the mdio device associated with xpcs.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: pcs: xpcs: add xpcs_create_mdiodev()
Russell King (Oracle) [Fri, 26 May 2023 10:14:29 +0000 (11:14 +0100)]
net: pcs: xpcs: add xpcs_create_mdiodev()

Add xpcs_create_mdiodev() to simplify the creation of the mdio device
associated with the XPCS. In order to allow xpcs_destroy() to clean
this up, we need to arrange for xpcs_create() to take a refcount on
the mdiodev, and xpcs_destroy() to put it.

Adding the refcounting to xpcs_create()..xpcs_destroy() will be
transparent to existing users of these interfaces.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
17 months agonet: mdio: add mdio_device_get() and mdio_device_put()
Russell King (Oracle) [Fri, 26 May 2023 10:14:24 +0000 (11:14 +0100)]
net: mdio: add mdio_device_get() and mdio_device_put()

Add two new operations for a mdio device to manage the refcount on the
underlying struct device. This will be used by mdio PCS drivers to
simplify the creation and destruction handling, making it easier for
users to get it correct.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf...
Jakub Kicinski [Sat, 27 May 2023 00:26:00 +0000 (17:26 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2023-05-26

We've added 54 non-merge commits during the last 10 day(s) which contain
a total of 76 files changed, 2729 insertions(+), 1003 deletions(-).

The main changes are:

1) Add the capability to destroy sockets in BPF through a new kfunc,
   from Aditi Ghag.

2) Support O_PATH fds in BPF_OBJ_PIN and BPF_OBJ_GET commands,
   from Andrii Nakryiko.

3) Add capability for libbpf to resize datasec maps when backed via mmap,
   from JP Kobryn.

4) Move all the test kfuncs for CI out of the kernel and into bpf_testmod,
   from Jiri Olsa.

5) Big batch of xsk selftest improvements to prep for multi-buffer testing,
   from Magnus Karlsson.

6) Show the target_{obj,btf}_id in tracing link's fdinfo and dump it
   via bpftool, from Yafang Shao.

7) Various misc BPF selftest improvements to work with upcoming LLVM 17,
   from Yonghong Song.

8) Extend bpftool to specify netdevice for resolving XDP hints,
   from Larysa Zaremba.

9) Document masking in shift operations for the insn set document,
   from Dave Thaler.

10) Extend BPF selftests to check xdp_feature support for bond driver,
    from Lorenzo Bianconi.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (54 commits)
  bpf: Fix bad unlock balance on freeze_mutex
  libbpf: Ensure FD >= 3 during bpf_map__reuse_fd()
  libbpf: Ensure libbpf always opens files with O_CLOEXEC
  selftests/bpf: Check whether to run selftest
  libbpf: Change var type in datasec resize func
  bpf: drop unnecessary bpf_capable() check in BPF_MAP_FREEZE command
  libbpf: Selftests for resizing datasec maps
  libbpf: Add capability for resizing datasec maps
  selftests/bpf: Add path_fd-based BPF_OBJ_PIN and BPF_OBJ_GET tests
  libbpf: Add opts-based bpf_obj_pin() API and add support for path_fd
  bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
  libbpf: Start v1.3 development cycle
  bpf: Validate BPF object in BPF_OBJ_PIN before calling LSM
  bpftool: Specify XDP Hints ifname when loading program
  selftests/bpf: Add xdp_feature selftest for bond device
  selftests/bpf: Test bpf_sock_destroy
  selftests/bpf: Add helper to get port using getsockname
  bpf: Add bpf_sock_destroy kfunc
  bpf: Add kfunc filter function to 'struct btf_kfunc_id_set'
  bpf: udp: Implement batching for sockets iterator
  ...
====================

Link: https://lore.kernel.org/r/20230526222747.17775-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: broadcom: Register dummy IRQ handler
Florian Fainelli [Thu, 25 May 2023 17:59:15 +0000 (10:59 -0700)]
net: phy: broadcom: Register dummy IRQ handler

In order to have our interrupt descriptor fully setup and in particular
the action, ensure that we register a full fledged interrupt handler.
This also allow us to set the interrupt polarity and flow through the
same call.

This is specifically necessary for kernel/irq/pm.c::suspend_device_irq
to set the interrupt descriptor to the IRQD_WAKEUP_ARMED state and
enable the interrupt for wake-up since it was still in a disabled state.

Without an interrupt descriptor we would have ran into cases where the
wake-up interrupt is not capable of waking up the system, specifically
if we resumed the system ACPI S5 using the Ethernet PHY. In that case
the Ethernet PHY interrupt would be pending by the time the kernel
booted, which it would acknowledge but then we could never use it as
a wake-up source again.

Fixes: 8baddaa9d4ba ("net: phy: broadcom: Add support for Wake-on-LAN")
Suggested-by: Doug Berger <doug.berger@broadcom.com>
Debugged-by: Doug Berger <doug.berger@broadcom.com>
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agotcp: remove unused TCP_SYNQ_INTERVAL definition
Neal Cardwell [Thu, 25 May 2023 14:57:36 +0000 (10:57 -0400)]
tcp: remove unused TCP_SYNQ_INTERVAL definition

Currently TCP_SYNQ_INTERVAL is defined but never used.

According to "git log -S TCP_SYNQ_INTERVAL net-next/main" it seems
the last references to TCP_SYNQ_INTERVAL were removed by 2015
commit fa76ce7328b2 ("inet: get rid of central tcp/dccp listener timer")

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agobpf: Fix bad unlock balance on freeze_mutex
Daniel Borkmann [Fri, 26 May 2023 10:13:56 +0000 (12:13 +0200)]
bpf: Fix bad unlock balance on freeze_mutex

Commit c4c84f6fb2c4 ("bpf: drop unnecessary bpf_capable() check in
BPF_MAP_FREEZE command") moved the permissions check outside of the
freeze_mutex in the map_freeze() handler. The error paths still jumps
to the err_put which tries to unlock the freeze_mutex even though it
was not locked in the first place. Fix it.

Fixes: c4c84f6fb2c4 ("bpf: drop unnecessary bpf_capable() check in BPF_MAP_FREEZE command")
Reported-by: syzbot+8982e75c2878b9ffeac5@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
18 months agolibbpf: Ensure FD >= 3 during bpf_map__reuse_fd()
Andrii Nakryiko [Thu, 25 May 2023 22:13:11 +0000 (15:13 -0700)]
libbpf: Ensure FD >= 3 during bpf_map__reuse_fd()

Improve bpf_map__reuse_fd() logic and ensure that dup'ed map FD is
"good" (>= 3) and has O_CLOEXEC flags. Use fcntl(F_DUPFD_CLOEXEC) for
that, similarly to ensure_good_fd() helper we already use in low-level
APIs that work with bpf() syscall.

Suggested-by: Lennart Poettering <lennart@poettering.net>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230525221311.2136408-2-andrii@kernel.org
18 months agolibbpf: Ensure libbpf always opens files with O_CLOEXEC
Andrii Nakryiko [Thu, 25 May 2023 22:13:10 +0000 (15:13 -0700)]
libbpf: Ensure libbpf always opens files with O_CLOEXEC

Make sure that libbpf code always gets FD with O_CLOEXEC flag set,
regardless if file is open through open() or fopen(). For the latter
this means to add "e" to mode string, which is supported since pretty
ancient glibc v2.7.

Also drop the outdated TODO comment in usdt.c, which was already completed.

Suggested-by: Lennart Poettering <lennart@poettering.net>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230525221311.2136408-1-andrii@kernel.org
18 months agoMerge branch 'mv88e6xxx-phylink-prepare'
David S. Miller [Fri, 26 May 2023 09:39:41 +0000 (10:39 +0100)]
Merge branch 'mv88e6xxx-phylink-prepare'

Russell King says:

====================
net: dsa: mv88e6xxx: prepare for phylink_pcs conversion

These two patches provide some preparation for converting the mv88e6xxx
DSA driver to use phylink_pcs rather than bolting the serdes bits into
the MAC calls.

In order to correctly drive mv88e6xxx hardware when the PCS code is
split, we need to force the link down while changing the configuration
of a port. This is provided for via the mac_prepare() and mac_finish()
methods, but DSA does not forward these on to DSA drivers.

Patch 1 adds support to the DSA core to forward these two methods to
DSA drivers, and patch 2 moves the code from mv88e6xxx_mac_config()
into the respective methods.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: mv88e6xxx: move link forcing to mac_prepare/mac_finish
Russell King (Oracle) [Thu, 25 May 2023 10:38:50 +0000 (11:38 +0100)]
net: dsa: mv88e6xxx: move link forcing to mac_prepare/mac_finish

Move the link forcing out of mac_config() and into the mac_prepare()
and mac_finish() methods. This results in no change to the order in
which these operations are performed, but does mean when we convert
mv88e6xxx to phylink_pcs support, we will continue to preserve this
ordering.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: add support for mac_prepare() and mac_finish() calls
Russell King (Oracle) [Thu, 25 May 2023 10:38:44 +0000 (11:38 +0100)]
net: dsa: add support for mac_prepare() and mac_finish() calls

Add DSA support for the phylink mac_prepare() and mac_finish() calls.
These were introduced as part of the PCS support to allow MACs to
perform preparatory steps prior to configuration, and finalisation
steps after the MAC and PCS has been configured.

Introducing phylink_pcs support to the mv88e6xxx DSA driver needs some
code moved out of its mac_config() stage into the mac_prepare() and
mac_finish() stages, and this commit facilitates such code in DSA
drivers.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: ynl: prefix uAPI header include with uapi/
Jakub Kicinski [Wed, 24 May 2023 17:09:01 +0000 (10:09 -0700)]
net: ynl: prefix uAPI header include with uapi/

To keep things simple we used to include the uAPI header
in the kernel in the #include <linux/$family.h> format.
This works well enough, most of the genl families should
have headers in include/net/ so linux/$family.h ends up
referring to the uAPI header, anyway. And if it doesn't
no big deal, we'll just include more info than we need.

Unless that is there is a naming conflict. Someone recently
created include/linux/psp.h which will be a problem when
supporting the PSP protocol. (I'm talking about
work-in-progress patches, but it's just a proof that assuming
lack of name conflicts was overly optimistic.)

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agosfc: handle VI shortage on ef100 by readjusting the channels
Pieter Jansen van Vuuren [Wed, 24 May 2023 09:36:38 +0000 (10:36 +0100)]
sfc: handle VI shortage on ef100 by readjusting the channels

When fewer VIs are allocated than what is allowed we can readjust
the channels by calling efx_mcdi_alloc_vis() again.

Signed-off-by: Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/core: Enable socket busy polling on -RT
Kurt Kanzenbach [Tue, 23 May 2023 11:15:18 +0000 (13:15 +0200)]
net/core: Enable socket busy polling on -RT

Busy polling is currently not allowed on PREEMPT_RT, because it disables
preemption while invoking the NAPI callback. It is not possible to acquire
sleeping locks with disabled preemption. For details see commit
20ab39d13e2e ("net/core: disable NET_RX_BUSY_POLL on PREEMPT_RT").

However, strict cyclic and/or low latency network applications may prefer busy
polling e.g., using AF_XDP instead of interrupt driven communication.

The preempt_disable() is used in order to prevent the poll_owner and NAPI owner
to be preempted while owning the resource to ensure progress. Netpoll performs
busy polling in order to acquire the lock. NAPI is locked by setting the
NAPIF_STATE_SCHED flag. There is no busy polling if the flag is set and the
"owner" is preempted. Worst case is that the task owning NAPI gets preempted and
NAPI processing stalls.  This is can be prevented by properly prioritising the
tasks within the system.

Allow RX_BUSY_POLL on PREEMPT_RT if NETPOLL is disabled. Don't disable
preemption on PREEMPT_RT within the busy poll loop.

Tested on x86 hardware with v6.1-RT and v6.3-RT on Intel i225 (igc) with
AF_XDP/ZC sockets configured to run in busy polling mode.

Suggested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Fri, 26 May 2023 03:56:19 +0000 (20:56 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge tag 'ib-leds-netdev-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Fri, 26 May 2023 03:37:28 +0000 (20:37 -0700)]
Merge tag 'ib-leds-netdev-v6.5' of git://git./linux/kernel/git/lee/leds

Lee Jones says:

====================
Immutable branch between LEDs and netdev due for the v6.5 merge window

Andrew Lunn says:

  Christian Marangi and I will be continuing the work of offloading LED
  blinking to Ethernet MAC and PHY LED controllers. The next set of
  patches is again cross subsystem, LEDs and netdev. It also requires
  some patches you have in for-leds-next:

  a286befc24e8 leds: trigger: netdev: Use mutex instead of spinlocks
  509412749002 leds: trigger: netdev: Convert device attr to macro
  0fd93ac85826 leds: trigger: netdev: Rename add namespace to netdev trigger enum modes
  eb31ca4531a0 leds: trigger: netdev: Drop NETDEV_LED_MODE_LINKUP from mode
  3fc498cf54b4 leds: trigger: netdev: Recheck NETDEV_LED_MODE_LINKUP on dev rename

  I'm assuming the new series will get nerged via netdev, with your
  Acked-by. Could you create a stable branch with these patches which
  can be pulled into netdev?

* tag 'ib-leds-netdev-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/leds:
  leds: trigger: netdev: Use mutex instead of spinlocks
  leds: trigger: netdev: Convert device attr to macro
  leds: trigger: netdev: Rename add namespace to netdev trigger enum modes
  leds: trigger: netdev: Drop NETDEV_LED_MODE_LINKUP from mode
  leds: trigger: netdev: Recheck NETDEV_LED_MODE_LINKUP on dev rename
====================

Link: https://lore.kernel.org/r/20230525111521.GA411262@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Fri, 26 May 2023 02:56:10 +0000 (19:56 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

Conflicts:

net/ipv4/raw.c
  3632679d9e4f ("ipv{4,6}/raw: fix output xfrm lookup wrt protocol")
  c85be08fc4fa ("raw: Stop using RTO_ONLINK.")
https://lore.kernel.org/all/20230525110037.2b532b83@canb.auug.org.au/

Adjacent changes:

drivers/net/ethernet/freescale/fec_main.c
  9025944fddfe ("net: fec: add dma_wmb to ensure correct descriptor values")
  144470c88c5d ("net: fec: using the standard return codes when xdp xmit errors")

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoselftests/bpf: Check whether to run selftest
Daniel Müller [Thu, 25 May 2023 23:22:48 +0000 (23:22 +0000)]
selftests/bpf: Check whether to run selftest

The sockopt test invokes test__start_subtest and then unconditionally
asserts the success. That means that even if deny-listed, any test will
still run and potentially fail.
Evaluate the return value of test__start_subtest() to achieve the
desired behavior, as other tests do.

Signed-off-by: Daniel Müller <deso@posteo.net>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20230525232248.640465-1-deso@posteo.net
18 months agoMerge tag 'net-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 25 May 2023 17:55:26 +0000 (10:55 -0700)]
Merge tag 'net-6.4-rc4' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Including fixes from bluetooth and bpf.

  Current release - regressions:

   - net: fix skb leak in __skb_tstamp_tx()

   - eth: mtk_eth_soc: fix QoS on DSA MAC on non MTK_NETSYS_V2 SoCs

  Current release - new code bugs:

   - handshake:
      - fix sock->file allocation
      - fix handshake_dup() ref counting

   - bluetooth:
      - fix potential double free caused by hci_conn_unlink
      - fix UAF in hci_conn_hash_flush

  Previous releases - regressions:

   - core: fix stack overflow when LRO is disabled for virtual
     interfaces

   - tls: fix strparser rx issues

   - bpf:
      - fix many sockmap/TCP related issues
      - fix a memory leak in the LRU and LRU_PERCPU hash maps
      - init the offload table earlier

   - eth: mlx5e:
      - do as little as possible in napi poll when budget is 0
      - fix using eswitch mapping in nic mode
      - fix deadlock in tc route query code

  Previous releases - always broken:

   - udplite: fix NULL pointer dereference in __sk_mem_raise_allocated()

   - raw: fix output xfrm lookup wrt protocol

   - smc: reset connection when trying to use SMCRv2 fails

   - phy: mscc: enable VSC8501/2 RGMII RX clock

   - eth: octeontx2-pf: fix TSOv6 offload

   - eth: cdc_ncm: deal with too low values of dwNtbOutMaxSize"

* tag 'net-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (79 commits)
  udplite: Fix NULL pointer dereference in __sk_mem_raise_allocated().
  net: phy: mscc: enable VSC8501/2 RGMII RX clock
  net: phy: mscc: remove unnecessary phydev locking
  net: phy: mscc: add support for VSC8501
  net: phy: mscc: add VSC8502 to MODULE_DEVICE_TABLE
  net/handshake: Enable the SNI extension to work properly
  net/handshake: Unpin sock->file if a handshake is cancelled
  net/handshake: handshake_genl_notify() shouldn't ignore @flags
  net/handshake: Fix uninitialized local variable
  net/handshake: Fix handshake_dup() ref counting
  net/handshake: Remove unneeded check from handshake_dup()
  ipv6: Fix out-of-bounds access in ipv6_find_tlv()
  net: ethernet: mtk_eth_soc: fix QoS on DSA MAC on non MTK_NETSYS_V2 SoCs
  docs: netdev: document the existence of the mail bot
  net: fix skb leak in __skb_tstamp_tx()
  r8169: Use a raw_spinlock_t for the register locks.
  page_pool: fix inconsistency for page_pool_ring_[un]lock()
  bpf, sockmap: Test progs verifier error with latest clang
  bpf, sockmap: Test FIONREAD returns correct bytes in rx buffer with drops
  bpf, sockmap: Test FIONREAD returns correct bytes in rx buffer
  ...

18 months agolibbpf: Change var type in datasec resize func
JP Kobryn [Thu, 25 May 2023 00:13:23 +0000 (17:13 -0700)]
libbpf: Change var type in datasec resize func

This changes a local variable type that stores a new array id to match
the return type of btf__add_array().

Signed-off-by: JP Kobryn <inwardvessel@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20230525001323.8554-1-inwardvessel@gmail.com
18 months agoMerge tag 'for-v6.4-rc' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux...
Linus Torvalds [Thu, 25 May 2023 17:26:36 +0000 (10:26 -0700)]
Merge tag 'for-v6.4-rc' of git://git./linux/kernel/git/sre/linux-power-supply

Pull power supply fixes from Sebastian Reichel:

 - Fix power_supply_get_battery_info for devices without parent devices
   resulting in NULL pointer dereference

 - Fix desktop systems reporting to run on battery once a power-supply
   device with device scope appears (e.g. a HID keyboard with a battery)

 - Ratelimit debug print about driver not providing data

 - Fix race condition related to external_power_changed in multiple
   drivers (ab8500, axp288, bq25890, sc27xx, bq27xxx)

 - Fix LED trigger switching from blinking to solid-on when charging
   finishes

 - Fix multiple races in bq27xxx battery driver

 - mt6360: handle potential ENOMEM from devm_work_autocancel

 - sbs-charger: Fix SBS_CHARGER_STATUS_CHARGE_INHIBITED bit

 - rt9467: avoid passing 0 to dev_err_probe

* tag 'for-v6.4-rc' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-power-supply: (21 commits)
  power: supply: Fix logic checking if system is running from battery
  power: supply: mt6360: add a check of devm_work_autocancel in mt6360_charger_probe
  power: supply: sbs-charger: Fix INHIBITED bit for Status reg
  power: supply: rt9467: Fix passing zero to 'dev_err_probe'
  power: supply: Ratelimit no data debug output
  power: supply: Fix power_supply_get_battery_info() if parent is NULL
  power: supply: bq24190: Call power_supply_changed() after updating input current
  power: supply: bq25890: Call power_supply_changed() after updating input current or voltage
  power: supply: bq27xxx: Use mod_delayed_work() instead of cancel() + schedule()
  power: supply: bq27xxx: After charger plug in/out wait 0.5s for things to stabilize
  power: supply: bq27xxx: Ensure power_supply_changed() is called on current sign changes
  power: supply: bq27xxx: Move bq27xxx_battery_update() down
  power: supply: bq27xxx: Add cache parameter to bq27xxx_battery_current_and_status()
  power: supply: bq27xxx: Fix poll_interval handling and races on remove
  power: supply: bq27xxx: Fix I2C IRQ race on remove
  power: supply: bq27xxx: Fix bq27xxx_battery_update() race condition
  power: supply: leds: Fix blink to LED on transition
  power: supply: sc27xx: Fix external_power_changed race
  power: supply: bq25890: Fix external_power_changed race
  power: supply: axp288_fuel_gauge: Fix external_power_changed race
  ...

18 months agobpf: drop unnecessary bpf_capable() check in BPF_MAP_FREEZE command
Andrii Nakryiko [Wed, 24 May 2023 22:54:19 +0000 (15:54 -0700)]
bpf: drop unnecessary bpf_capable() check in BPF_MAP_FREEZE command

Seems like that extra bpf_capable() check in BPF_MAP_FREEZE handler was
unintentionally left when we switched to a model that all BPF map
operations should be allowed regardless of CAP_BPF (or any other
capabilities), as long as process got BPF map FD somehow.

This patch replaces bpf_capable() check in BPF_MAP_FREEZE handler with
writeable access check, given conceptually freezing the map is modifying
it: map becomes unmodifiable for subsequent updates.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230524225421.1587859-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoMerge tag 'sound-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Thu, 25 May 2023 16:48:23 +0000 (09:48 -0700)]
Merge tag 'sound-6.4-rc4' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of small fixes:

   - HD-audio runtime PM bug fix

   - A couple of HD-audio quirks

   - Fix series of ASoC Intel AVS drivers

   - ASoC DPCM fix for a bug found on new Intel systems

   - A few other ASoC device-specific small fixes"

* tag 'sound-6.4-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda/realtek: Enable headset onLenovo M70/M90
  ASoC: dwc: move DMA init to snd_soc_dai_driver probe()
  ASoC: cs35l41: Fix default regmap values for some registers
  ALSA: hda: Fix unhandled register update during auto-suspend period
  ASoC: dt-bindings: tlv320aic32x4: Fix supply names
  ASoC: Intel: avs: Add missing checks on FE startup
  ASoC: Intel: avs: Fix avs_path_module::instance_id size
  ASoC: Intel: avs: Account for UID of ACPI device
  ASoC: Intel: avs: Fix declaration of enum avs_channel_config
  ASoC: Intel: Skylake: Fix declaration of enum skl_ch_cfg
  ASoC: Intel: avs: Access path components under lock
  ASoC: Intel: avs: Fix module lookup
  ALSA: hda/ca0132: add quirk for EVGA X299 DARK
  ASoC: soc-pcm: test if a BE can be prepared
  ASoC: rt5682: Disable jack detection interrupt during suspend
  ASoC: lpass: Fix for KASAN use_after_free out of bounds

18 months agoMerge tag 'platform-drivers-x86-v6.4-3' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 25 May 2023 16:42:15 +0000 (09:42 -0700)]
Merge tag 'platform-drivers-x86-v6.4-3' of git://git./linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fixes from Hans de Goede:
 "Nothing special to report just a few small fixes"

* tag 'platform-drivers-x86-v6.4-3' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/x86/intel/ifs: Annotate work queue on stack so object debug does not complain
  platform/x86: ISST: Remove 8 socket limit
  platform/mellanox: mlxbf-pmc: fix sscanf() error checking
  platform/x86/amd/pmf: Fix CnQF and auto-mode after resume
  platform/x86: asus-wmi: Ignore WMI events with codes 0x7B, 0xC0

18 months agoMerge tag 'm68k-for-v6.4-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/geert...
Linus Torvalds [Thu, 25 May 2023 16:33:03 +0000 (09:33 -0700)]
Merge tag 'm68k-for-v6.4-tag2' of git://git./linux/kernel/git/geert/linux-m68k

Pull m68k fix from Geert Uytterhoeven:

 - Fix signal frame issue causing user-space crashes on 68020/68030

* tag 'm68k-for-v6.4-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: Move signal frame following exception on 68020/030

18 months agos390/ism: Set DMA coherent mask
Niklas Schnelle [Wed, 24 May 2023 07:54:10 +0000 (09:54 +0200)]
s390/ism: Set DMA coherent mask

A future change will convert the DMA API implementation from the
architecture specific arch/s390/pci/pci_dma.c to using the common code
drivers/iommu/dma-iommu.c which the utilizes the same IOMMU hardware
through the s390-iommu driver. Unlike the s390 specific DMA API this
requires devices to correctly set the coherent mask to be allowed to use
IOVAs >2^32 in dma_alloc_coherent(). This was however not done for ISM
devices. ISM requires such addresses since currently the DMA aperture
for PCI devices starts at 2^32 and all calls to dma_alloc_coherent()
would thus fail.

Link: https://lore.kernel.org/all/20230310-dma_iommu-v9-1-65bb8edd2beb@linux.ibm.com/
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Reviewed-by: Pierre Morel <pmorel@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Link: https://lore.kernel.org/r/20230524075411.3734141-1-schnelle@linux.ibm.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agonet/mlx4: Use bitmap_weight_and()
Christophe JAILLET [Tue, 23 May 2023 18:17:52 +0000 (20:17 +0200)]
net/mlx4: Use bitmap_weight_and()

Use bitmap_weight_and() instead of hand writing it.

This saves a few LoC and is slightly faster, should it mater.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/a29c2348a062408bec45cee2601b2417310e5ea7.1684865809.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agoMerge branch 'net-tcp-make-txhash-use-consistent-for-ipv4'
Paolo Abeni [Thu, 25 May 2023 11:20:47 +0000 (13:20 +0200)]
Merge branch 'net-tcp-make-txhash-use-consistent-for-ipv4'

Antoine Tenart says:

====================
net: tcp: make txhash use consistent for IPv4

Series is divided in two parts. First two commits make the txhash (used
for the skb hash in TCP) to be consistent for all IPv4/TCP packets (IPv6
doesn't have the same issue). Last commit improve a hash-related doc.

One example is when using OvS with dp_hash, which uses skb->hash, to
select a path. We'd like packets from the same flow to be consistent, as
well as the hash being stable over time when using net.core.txrehash=0.
Same applies for kernel ECMP which also can use skb->hash.
====================

Link: https://lore.kernel.org/r/20230523161453.196094-1-atenart@kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agoDocumentation: net: net.core.txrehash is not specific to listening sockets
Antoine Tenart [Tue, 23 May 2023 16:14:53 +0000 (18:14 +0200)]
Documentation: net: net.core.txrehash is not specific to listening sockets

The net.core.txrehash documentation mentions this knob is for listening
sockets only, while sk_rethink_txhash can be called on SYN and RTO
retransmits on all TCP sockets.

Remove the listening socket part.

Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agonet: ipv4: use consistent txhash in TIME_WAIT and SYN_RECV
Antoine Tenart [Tue, 23 May 2023 16:14:52 +0000 (18:14 +0200)]
net: ipv4: use consistent txhash in TIME_WAIT and SYN_RECV

When using IPv4/TCP, skb->hash comes from sk->sk_txhash except in
TIME_WAIT and SYN_RECV where it's not set in the reply skb from
ip_send_unicast_reply. Those packets will have a mismatched hash with
others from the same flow as their hashes will be 0. IPv6 does not have
the same issue as the hash is set from the socket txhash in those cases.

This commits sets the hash in the reply skb from ip_send_unicast_reply,
which makes the IPv4 code behaving like IPv6.

Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agonet: tcp: make the txhash available in TIME_WAIT sockets for IPv4 too
Antoine Tenart [Tue, 23 May 2023 16:14:51 +0000 (18:14 +0200)]
net: tcp: make the txhash available in TIME_WAIT sockets for IPv4 too

Commit c67b85558ff2 ("ipv6: tcp: send consistent autoflowlabel in
TIME_WAIT state") made the socket txhash also available in TIME_WAIT
sockets but for IPv6 only. Make it available for IPv4 too as we'll use
it in later commits.

Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agoleds: trigger: netdev: Use mutex instead of spinlocks
Christian Marangi [Wed, 19 Apr 2023 21:07:43 +0000 (23:07 +0200)]
leds: trigger: netdev: Use mutex instead of spinlocks

Some LEDs may require to sleep while doing some operation like setting
brightness and other cleanup.

For this reason, using a spinlock will cause a sleep under spinlock
warning.

It should be safe to convert this to a sleepable lock since:
- sysfs read/write can sleep
- netdev_trig_work is a work queue and can sleep
- netdev _trig_notify can sleep

The spinlock was used when brightness didn't support sleeping, but this
changed and now it supported with brightness_set_blocking().

Convert to mutex lock to permit sleeping using brightness_set_blocking().

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230419210743.3594-6-ansuelsmth@gmail.com
18 months agoleds: trigger: netdev: Convert device attr to macro
Christian Marangi [Wed, 19 Apr 2023 21:07:42 +0000 (23:07 +0200)]
leds: trigger: netdev: Convert device attr to macro

Convert link tx and rx device attr to a common macro to reduce common
code and in preparation for additional attr.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230419210743.3594-5-ansuelsmth@gmail.com
18 months agoleds: trigger: netdev: Rename add namespace to netdev trigger enum modes
Christian Marangi [Wed, 19 Apr 2023 21:07:41 +0000 (23:07 +0200)]
leds: trigger: netdev: Rename add namespace to netdev trigger enum modes

Rename NETDEV trigger enum modes to a more symbolic name and add a
namespace to them.

Also add __TRIGGER_NETDEV_MAX to identify the max modes of the netdev
trigger.

This is a cleanup to drop the define and no behaviour change are
intended.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230419210743.3594-4-ansuelsmth@gmail.com
18 months agoleds: trigger: netdev: Drop NETDEV_LED_MODE_LINKUP from mode
Christian Marangi [Wed, 19 Apr 2023 21:07:40 +0000 (23:07 +0200)]
leds: trigger: netdev: Drop NETDEV_LED_MODE_LINKUP from mode

Putting NETDEV_LED_MODE_LINKUP in the same list of the netdev trigger
modes is wrong as it's used to set the link state of the device and not
to set a blink mode as it's done by NETDEV_LED_LINK, NETDEV_LED_TX and
NETDEV_LED_RX. It's also wrong to put this state in the same bitmap of the
netdev trigger mode and should be external to it.

Drop NETDEV_LED_MODE_LINKUP from mode list and convert to a simple bool
that will be true or false based on the carrier link. No functional
change intended.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230419210743.3594-3-ansuelsmth@gmail.com
18 months agoleds: trigger: netdev: Recheck NETDEV_LED_MODE_LINKUP on dev rename
Christian Marangi [Wed, 19 Apr 2023 21:07:39 +0000 (23:07 +0200)]
leds: trigger: netdev: Recheck NETDEV_LED_MODE_LINKUP on dev rename

Dev can be renamed also while up for supported device. We currently
wrongly clear the NETDEV_LED_MODE_LINKUP flag on NETDEV_CHANGENAME
event.

Fix this by rechecking if the carrier is ok on NETDEV_CHANGENAME and
correctly set the NETDEV_LED_MODE_LINKUP bit.

Fixes: 5f820ed52371 ("leds: trigger: netdev: fix handling on interface rename")
Cc: stable@vger.kernel.org # v5.5+
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Lee Jones <lee@kernel.org>
Link: https://lore.kernel.org/r/20230419210743.3594-2-ansuelsmth@gmail.com
18 months agoudplite: Fix NULL pointer dereference in __sk_mem_raise_allocated().
Kuniyuki Iwashima [Tue, 23 May 2023 16:33:05 +0000 (09:33 -0700)]
udplite: Fix NULL pointer dereference in __sk_mem_raise_allocated().

syzbot reported [0] a null-ptr-deref in sk_get_rmem0() while using
IPPROTO_UDPLITE (0x88):

  14:25:52 executing program 1:
  r0 = socket$inet6(0xa, 0x80002, 0x88)

We had a similar report [1] for probably sk_memory_allocated_add()
in __sk_mem_raise_allocated(), and commit c915fe13cbaa ("udplite: fix
NULL pointer dereference") fixed it by setting .memory_allocated for
udplite_prot and udplitev6_prot.

To fix the variant, we need to set either .sysctl_wmem_offset or
.sysctl_rmem.

Now UDP and UDPLITE share the same value for .memory_allocated, so we
use the same .sysctl_wmem_offset for UDP and UDPLITE.

[0]:
general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
CPU: 0 PID: 6829 Comm: syz-executor.1 Not tainted 6.4.0-rc2-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/28/2023
RIP: 0010:sk_get_rmem0 include/net/sock.h:2907 [inline]
RIP: 0010:__sk_mem_raise_allocated+0x806/0x17a0 net/core/sock.c:3006
Code: c1 ea 03 80 3c 02 00 0f 85 23 0f 00 00 48 8b 44 24 08 48 8b 98 38 01 00 00 48 b8 00 00 00 00 00 fc ff df 48 89 da 48 c1 ea 03 <0f> b6 14 02 48 89 d8 83 e0 07 83 c0 03 38 d0 0f 8d 6f 0a 00 00 8b
RSP: 0018:ffffc90005d7f450 EFLAGS: 00010246
RAX: dffffc0000000000 RBX: 0000000000000000 RCX: ffffc90004d92000
RDX: 0000000000000000 RSI: ffffffff88066482 RDI: ffffffff8e2ccbb8
RBP: ffff8880173f7000 R08: 0000000000000005 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000030000
R13: 0000000000000001 R14: 0000000000000340 R15: 0000000000000001
FS:  0000000000000000(0000) GS:ffff8880b9800000(0063) knlGS:00000000f7f1cb40
CS:  0010 DS: 002b ES: 002b CR0: 0000000080050033
CR2: 000000002e82f000 CR3: 0000000034ff0000 CR4: 00000000003506f0
Call Trace:
 <TASK>
 __sk_mem_schedule+0x6c/0xe0 net/core/sock.c:3077
 udp_rmem_schedule net/ipv4/udp.c:1539 [inline]
 __udp_enqueue_schedule_skb+0x776/0xb30 net/ipv4/udp.c:1581
 __udpv6_queue_rcv_skb net/ipv6/udp.c:666 [inline]
 udpv6_queue_rcv_one_skb+0xc39/0x16c0 net/ipv6/udp.c:775
 udpv6_queue_rcv_skb+0x194/0xa10 net/ipv6/udp.c:793
 __udp6_lib_mcast_deliver net/ipv6/udp.c:906 [inline]
 __udp6_lib_rcv+0x1bda/0x2bd0 net/ipv6/udp.c:1013
 ip6_protocol_deliver_rcu+0x2e7/0x1250 net/ipv6/ip6_input.c:437
 ip6_input_finish+0x150/0x2f0 net/ipv6/ip6_input.c:482
 NF_HOOK include/linux/netfilter.h:303 [inline]
 NF_HOOK include/linux/netfilter.h:297 [inline]
 ip6_input+0xa0/0xd0 net/ipv6/ip6_input.c:491
 ip6_mc_input+0x40b/0xf50 net/ipv6/ip6_input.c:585
 dst_input include/net/dst.h:468 [inline]
 ip6_rcv_finish net/ipv6/ip6_input.c:79 [inline]
 NF_HOOK include/linux/netfilter.h:303 [inline]
 NF_HOOK include/linux/netfilter.h:297 [inline]
 ipv6_rcv+0x250/0x380 net/ipv6/ip6_input.c:309
 __netif_receive_skb_one_core+0x114/0x180 net/core/dev.c:5491
 __netif_receive_skb+0x1f/0x1c0 net/core/dev.c:5605
 netif_receive_skb_internal net/core/dev.c:5691 [inline]
 netif_receive_skb+0x133/0x7a0 net/core/dev.c:5750
 tun_rx_batched+0x4b3/0x7a0 drivers/net/tun.c:1553
 tun_get_user+0x2452/0x39c0 drivers/net/tun.c:1989
 tun_chr_write_iter+0xdf/0x200 drivers/net/tun.c:2035
 call_write_iter include/linux/fs.h:1868 [inline]
 new_sync_write fs/read_write.c:491 [inline]
 vfs_write+0x945/0xd50 fs/read_write.c:584
 ksys_write+0x12b/0x250 fs/read_write.c:637
 do_syscall_32_irqs_on arch/x86/entry/common.c:112 [inline]
 __do_fast_syscall_32+0x65/0xf0 arch/x86/entry/common.c:178
 do_fast_syscall_32+0x33/0x70 arch/x86/entry/common.c:203
 entry_SYSENTER_compat_after_hwframe+0x70/0x82
RIP: 0023:0xf7f21579
Code: b8 01 10 06 03 74 b4 01 10 07 03 74 b0 01 10 08 03 74 d8 01 00 00 00 00 00 00 00 00 00 00 00 00 00 51 52 55 89 e5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d b4 26 00 00 00 00 8d b4 26 00 00 00 00
RSP: 002b:00000000f7f1c590 EFLAGS: 00000282 ORIG_RAX: 0000000000000004
RAX: ffffffffffffffda RBX: 00000000000000c8 RCX: 0000000020000040
RDX: 0000000000000083 RSI: 00000000f734e000 RDI: 0000000000000000
RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000296 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
 </TASK>
Modules linked in:

Link: https://lore.kernel.org/netdev/CANaxB-yCk8hhP68L4Q2nFOJht8sqgXGGQO2AftpHs0u1xyGG5A@mail.gmail.com/
Fixes: 850cbaddb52d ("udp: use it's own memory accounting schema")
Reported-by: syzbot+444ca0907e96f7c5e48b@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=444ca0907e96f7c5e48b
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230523163305.66466-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
18 months agonet: phylink: provide phylink_pcs_config() and phylink_pcs_link_up()
Russell King (Oracle) [Tue, 23 May 2023 15:31:50 +0000 (16:31 +0100)]
net: phylink: provide phylink_pcs_config() and phylink_pcs_link_up()

Add two helper functions for calling PCS methods. phylink_pcs_config()
allows us to handle PCS configuration specifics in one location, rather
than the two call sites. phylink_pcs_link_up() gives us consistency.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/E1q1TzK-007Exd-Rs@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge branch 'net-phy-mscc-support-vsc8501'
Jakub Kicinski [Thu, 25 May 2023 05:14:29 +0000 (22:14 -0700)]
Merge branch 'net-phy-mscc-support-vsc8501'

David Epping says:

====================
net: phy: mscc: support VSC8501

this updated series of patches adds support for the VSC8501 Ethernet
PHY and fixes support for the VSC8502 PHY in cases where no other
software (like U-Boot) has initialized the PHY after power up.

The first patch simply adds the VSC8502 to the MODULE_DEVICE_TABLE,
where I guess it was unintentionally missing. I have no hardware to
test my change.

The second patch adds the VSC8501 PHY with exactly the same driver
implementation as the existing VSC8502.

The (new) third patch removes phydev locking from
vsc85xx_rgmii_set_skews(), as discussed for v2 of the patch set.

The (now) fourth patch fixes the initialization for VSC8501 and VSC8502.
I have tested this patch with VSC8501 on hardware in RGMII mode only.
https://ww1.microchip.com/downloads/aemDocuments/documents/UNG/ProductDocuments/DataSheets/VSC8501-03_Datasheet_60001741A.PDF
https://ww1.microchip.com/downloads/aemDocuments/documents/UNG/ProductDocuments/DataSheets/VSC8502-03_Datasheet_60001742B.pdf
Table 4-42 "RGMII CONTROL, ADDRESS 20E2 (0X14)" Bit 11 for each of
them.
By default the RX_CLK is disabled for these PHYs. In cases where no
other software, like U-Boot, enabled the clock, this results in no
received packets being handed to the MAC.
The patch enables this clock output.
According to Microchip support (case number 01268776) this applies
to all modes (RGMII, GMII, and MII).

Other PHYs sharing the same register map and code, like
VSC8530/31/40/41 have the clock enabled and the relevant bit 11 is
reserved and read-only for them. As per previous discussion the
patch still clears the bit on these PHYs, too, possibly more easily
supporting other future PHYs implementing this functionality.

For the VSC8572 family of PHYs, having a different register map,
no such changes are applied.
====================

Link: https://lore.kernel.org/r/20230523153108.18548-1-david.epping@missinglinkelectronics.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: mscc: enable VSC8501/2 RGMII RX clock
David Epping [Tue, 23 May 2023 15:31:08 +0000 (17:31 +0200)]
net: phy: mscc: enable VSC8501/2 RGMII RX clock

By default the VSC8501 and VSC8502 RGMII/GMII/MII RX_CLK output is
disabled. To allow packet forwarding towards the MAC it needs to be
enabled.

For other PHYs supported by this driver the clock output is enabled
by default.

Fixes: d3169863310d ("net: phy: mscc: add support for VSC8502")
Signed-off-by: David Epping <david.epping@missinglinkelectronics.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: mscc: remove unnecessary phydev locking
David Epping [Tue, 23 May 2023 15:31:07 +0000 (17:31 +0200)]
net: phy: mscc: remove unnecessary phydev locking

Holding the struct phy_device (phydev) lock is unnecessary when
accessing phydev->interface in the PHY driver .config_init method,
which is the only place that vsc85xx_rgmii_set_skews() is called from.

The phy_modify_paged() function implements required MDIO bus level
locking, which can not be achieved by a phydev lock.

Signed-off-by: David Epping <david.epping@missinglinkelectronics.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: mscc: add support for VSC8501
David Epping [Tue, 23 May 2023 15:31:06 +0000 (17:31 +0200)]
net: phy: mscc: add support for VSC8501

The VSC8501 PHY can use the same driver implementation as the VSC8502.
Adding the PHY ID and copying the handler functions of VSC8502 is
sufficient to operate it.

Signed-off-by: David Epping <david.epping@missinglinkelectronics.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: mscc: add VSC8502 to MODULE_DEVICE_TABLE
David Epping [Tue, 23 May 2023 15:31:05 +0000 (17:31 +0200)]
net: phy: mscc: add VSC8502 to MODULE_DEVICE_TABLE

The mscc driver implements support for VSC8502, so its ID should be in
the MODULE_DEVICE_TABLE for automatic loading.

Signed-off-by: David Epping <david.epping@missinglinkelectronics.com>
Fixes: d3169863310d ("net: phy: mscc: add support for VSC8502")
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge branch 'bug-fixes-for-net-handshake'
Jakub Kicinski [Thu, 25 May 2023 05:05:25 +0000 (22:05 -0700)]
Merge branch 'bug-fixes-for-net-handshake'

Chuck Lever says:

====================
Bug fixes for net/handshake

Paolo observed that there is a possible leak of sock->file. I
haven't looked into that yet, but it seems to be separate from
the fixes in this series, so no need to hold these up.
====================

The submissions mentions net-next but it means netdev (perhaps
merge window left over when trees are converged). In any case,
it should have gone into net, but was instead applied to net-next
as commit deb2e484baf9 ("Merge branch 'net-handshake-fixes'").
These are fixes tho, and Chuck needs them to make progress with
the client so double-merging them into net... it is what it is :(

Link: https://lore.kernel.org/r/168381978252.84244.1933636428135211300.stgit@91.116.238.104.host.secureserver.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Enable the SNI extension to work properly
Chuck Lever [Thu, 11 May 2023 15:49:50 +0000 (11:49 -0400)]
net/handshake: Enable the SNI extension to work properly

Enable the upper layer protocol to specify the SNI peername. This
avoids the need for tlshd to use a DNS lookup, which can return a
hostname that doesn't match the incoming certificate's SubjectName.

Fixes: 2fd5532044a8 ("net/handshake: Add a kernel API for requesting a TLSv1.3 handshake")
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Unpin sock->file if a handshake is cancelled
Chuck Lever [Thu, 11 May 2023 15:49:17 +0000 (11:49 -0400)]
net/handshake: Unpin sock->file if a handshake is cancelled

If user space never calls DONE, sock->file's reference count remains
elevated. Enable sock->file to be freed eventually in this case.

Reported-by: Jakub Kacinski <kuba@kernel.org>
Fixes: 3b3009ea8abb ("net/handshake: Create a NETLINK service for handling handshake requests")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: handshake_genl_notify() shouldn't ignore @flags
Chuck Lever [Thu, 11 May 2023 15:48:45 +0000 (11:48 -0400)]
net/handshake: handshake_genl_notify() shouldn't ignore @flags

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Fixes: 3b3009ea8abb ("net/handshake: Create a NETLINK service for handling handshake requests")
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Fix uninitialized local variable
Chuck Lever [Thu, 11 May 2023 15:48:13 +0000 (11:48 -0400)]
net/handshake: Fix uninitialized local variable

trace_handshake_cmd_done_err() simply records the pointer in @req,
so initializing it to NULL is sufficient and safe.

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Fixes: 3b3009ea8abb ("net/handshake: Create a NETLINK service for handling handshake requests")
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Fix handshake_dup() ref counting
Chuck Lever [Thu, 11 May 2023 15:47:40 +0000 (11:47 -0400)]
net/handshake: Fix handshake_dup() ref counting

If get_unused_fd_flags() fails, we ended up calling fput(sock->file)
twice.

Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Fixes: 3b3009ea8abb ("net/handshake: Create a NETLINK service for handling handshake requests")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Remove unneeded check from handshake_dup()
Chuck Lever [Thu, 11 May 2023 15:47:09 +0000 (11:47 -0400)]
net/handshake: Remove unneeded check from handshake_dup()

handshake_req_submit() now verifies that the socket has a file.

Fixes: 3b3009ea8abb ("net/handshake: Create a NETLINK service for handling handshake requests")
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 25 May 2023 04:57:57 +0000 (21:57 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2023-05-24

We've added 19 non-merge commits during the last 10 day(s) which contain
a total of 20 files changed, 738 insertions(+), 448 deletions(-).

The main changes are:

1) Batch of BPF sockmap fixes found when running against NGINX TCP tests,
   from John Fastabend.

2) Fix a memleak in the LRU{,_PERCPU} hash map when bucket locking fails,
   from Anton Protopopov.

3) Init the BPF offload table earlier than just late_initcall,
   from Jakub Kicinski.

4) Fix ctx access mask generation for 32-bit narrow loads of 64-bit fields,
   from Will Deacon.

5) Remove a now unsupported __fallthrough in BPF samples,
   from Andrii Nakryiko.

6) Fix a typo in pkg-config call for building sign-file,
   from Jeremy Sowden.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpf, sockmap: Test progs verifier error with latest clang
  bpf, sockmap: Test FIONREAD returns correct bytes in rx buffer with drops
  bpf, sockmap: Test FIONREAD returns correct bytes in rx buffer
  bpf, sockmap: Test shutdown() correctly exits epoll and recv()=0
  bpf, sockmap: Build helper to create connected socket pair
  bpf, sockmap: Pull socket helpers out of listen test for general use
  bpf, sockmap: Incorrectly handling copied_seq
  bpf, sockmap: Wake up polling after data copy
  bpf, sockmap: TCP data stall on recv before accept
  bpf, sockmap: Handle fin correctly
  bpf, sockmap: Improved check for empty queue
  bpf, sockmap: Reschedule is now done through backlog
  bpf, sockmap: Convert schedule_work into delayed_work
  bpf, sockmap: Pass skb ownership through read_skb
  bpf: fix a memory leak in the LRU and LRU_PERCPU hash maps
  bpf: Fix mask generation for 32-bit narrow loads of 64-bit fields
  samples/bpf: Drop unnecessary fallthrough
  bpf: netdev: init the offload table earlier
  selftests/bpf: Fix pkg-config call building sign-file
====================

Link: https://lore.kernel.org/r/20230524170839.13905-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge branch 'libbpf: capability for resizing datasec maps'
Andrii Nakryiko [Wed, 24 May 2023 18:11:34 +0000 (11:11 -0700)]
Merge branch 'libbpf: capability for resizing datasec maps'

JP Kobryn says:

====================
Due to the way the datasec maps like bss, data, rodata are memory
mapped, they cannot be resized with bpf_map__set_value_size() like
non-datasec maps can. This series offers a way to allow the resizing of
datasec maps, by having the mapped regions resized as needed and also
adjusting associated BTF info if possible.

The thought behind this is to allow for use cases where a given datasec
needs to scale to for example the number of CPU's present. A bpf program
can have a global array in a data section with an initial length and
before loading the bpf program, the array length could be extended to
match the CPU count. The selftests included in this series perform this
scaling to an arbitrary value to demonstrate how it can work.
====================

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
18 months agolibbpf: Selftests for resizing datasec maps
JP Kobryn [Wed, 24 May 2023 00:45:37 +0000 (17:45 -0700)]
libbpf: Selftests for resizing datasec maps

This patch adds test coverage for resizing datasec maps. The first two
subtests resize the bss and custom data sections. In both cases, an
initial array (of length one) has its element set to one. After resizing
the rest of the array is filled with ones as well. A BPF program is then
run to sum the respective arrays and back on the userspace side the sum
is checked to be equal to the number of elements.
The third subtest attempts to perform resizing under conditions that
will result in either the resize failing or the BTF info being cleared.

Signed-off-by: JP Kobryn <inwardvessel@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/bpf/20230524004537.18614-3-inwardvessel@gmail.com
18 months agolibbpf: Add capability for resizing datasec maps
JP Kobryn [Wed, 24 May 2023 00:45:36 +0000 (17:45 -0700)]
libbpf: Add capability for resizing datasec maps

This patch updates bpf_map__set_value_size() so that if the given map is
memory mapped, it will attempt to resize the mapped region. Initial
contents of the mapped region are preserved. BTF is not required, but
after the mapping is resized an attempt is made to adjust the associated
BTF information if the following criteria is met:
 - BTF info is present
 - the map is a datasec
 - the final variable in the datasec is an array

... the resulting BTF info will be updated so that the final array
variable is associated with a new BTF array type sized to cover the
requested size.

Note that the initial resizing of the memory mapped region can succeed
while the subsequent BTF adjustment can fail. In this case, BTF info is
dropped from the map by clearing the key and value type.

Signed-off-by: JP Kobryn <inwardvessel@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/bpf/20230524004537.18614-2-inwardvessel@gmail.com
18 months agoMerge tag 'spi-fix-v6.4-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Wed, 24 May 2023 18:43:08 +0000 (11:43 -0700)]
Merge tag 'spi-fix-v6.4-rc3' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A collection of fixes that came in since the merge window, plus an
  update to MAINTAINERS.

  The Cadence fixes are coming from the addition of device mode support,
  they required a couple of incremental updates in order to get
  something that works robustly for both device and controller modes"

* tag 'spi-fix-v6.4-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: spi-cadence: Interleave write of TX and read of RX FIFO
  spi: dw: Replace spi->chip_select references with function calls
  spi: MAINTAINERS: drop Krzysztof Kozlowski from Samsung SPI
  spi: spi-cadence: Only overlap FIFO transactions in slave mode
  spi: spi-cadence: Avoid read of RX FIFO before its ready
  spi: spi-geni-qcom: Select FIFO mode for chip select

18 months agoMerge tag 'regulator-fix-v6.4-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 24 May 2023 18:37:03 +0000 (11:37 -0700)]
Merge tag 'regulator-fix-v6.4-rc3' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "Some fixes that came in since the merge window, nothing terribly
  exciting - a couple of driver specific fixes and a fix for the error
  handling when setting up the debugfs for the devices"

* tag 'regulator-fix-v6.4-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: mt6359: add read check for PMIC MT6359
  regulator: Fix error checking for debugfs_create_dir
  regulator: pca9450: Fix BUCK2 enable_mask

18 months agoMerge tag 'mmc-v6.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Wed, 24 May 2023 18:31:37 +0000 (11:31 -0700)]
Merge tag 'mmc-v6.4-rc1' of git://git./linux/kernel/git/ulfh/mmc

Pull MMC fixes from Ulf Hansson:
 "MMC core:

   - Fix error propagation for the non-block-device I/O paths

  MMC host:

   - sdhci-cadence: Fix an error path during probe

   - sdhci-esdhc-imx: Fix support for the 'no-mmc-hs400' DT property"

* tag 'mmc-v6.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: sdhci-esdhc-imx: make "no-mmc-hs400" works
  mmc: sdhci-cadence: Fix an error handling path in sdhci_cdns_probe()
  mmc: block: ensure error propagation for non-blk

18 months agoMerge branch 'net-pcs-xpcs-cleanups-for-clause-73-support'
Jakub Kicinski [Wed, 24 May 2023 16:13:30 +0000 (09:13 -0700)]
Merge branch 'net-pcs-xpcs-cleanups-for-clause-73-support'

Russell King says:

====================
net: pcs: xpcs: cleanups for clause 73 support

This series cleans up xpcs code, moving much of the clause 73 code
out of the driver into places where others can make use of it.

Specifically, we add a helper to convert a clause 73 advertisement
to ethtool link modes to mdio.h, and a helper to resolve the clause
73 negotiation state to phylink, which includes the pause modes.

In doing this cleanup, several issues were identified with the
original xpcs implementation:

1) it masks the link partner advertisement with its own advertisement
   so userspace can't see what the full link partner advertisement
   was.
2) it was always setting pause modes irrespective of the advertisements
   on either end of the link.
3) it was reading the STAT1 registers multiple times. Reading STAT1
   has the side effect of unlatching the link-down status, so
   multiple reads should be avoided.

This patch series addresses the first two first by addressing the
issues, and then by moving over to the new helpers. The third issue
is solved by restructuring the xpcs code.
====================

Link: https://lore.kernel.org/r/ZGyR/jDyYTYzRklg@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: avoid reading STAT1 more than once
Russell King (Oracle) [Tue, 23 May 2023 10:16:29 +0000 (11:16 +0100)]
net: pcs: xpcs: avoid reading STAT1 more than once

Avoid reading the STAT1 registers more than once while getting the PCS
state, as this register contains latching-low bits that are lost after
the first read.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: use phylink_resolve_c73() helper
Russell King (Oracle) [Tue, 23 May 2023 10:16:24 +0000 (11:16 +0100)]
net: pcs: xpcs: use phylink_resolve_c73() helper

Use phylink_resolve_c73() to resolve the clause 73 autonegotiation
result.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: correct pause resolution
Russell King (Oracle) [Tue, 23 May 2023 10:16:18 +0000 (11:16 +0100)]
net: pcs: xpcs: correct pause resolution

xpcs was indicating symmetric pause should be enabled regardless of
the advertisements by either party. Fix this to use
linkmode_resolve_pause() now that we're no longer obliterating the
link partner's advertisement by logically anding it with our own.

This is transitional, the function will be entirely replaced with
phylink_resolve_c73() in the following patch.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: correct lp_advertising contents
Russell King (Oracle) [Tue, 23 May 2023 10:16:13 +0000 (11:16 +0100)]
net: pcs: xpcs: correct lp_advertising contents

lp_advertising is supposed to reflect the link partner's advertisement
unmodified by the local advertisement, but xpcs bitwise ands it with
the local advertisement prior to calculating the resolution of the
negotiation.

Fix this by moving the bitwise and to xpcs_resolve_lpa_c73() so it can
place the results in a temporary bitmap before passing that to
ixpcs_get_max_usxgmii_speed().

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: use mii_c73_to_linkmode() helper
Russell King (Oracle) [Tue, 23 May 2023 10:16:08 +0000 (11:16 +0100)]
net: pcs: xpcs: use mii_c73_to_linkmode() helper

Convert xpcs clause 73 reading to use the newly introduced
mii_c73_to_linkmode() helper to translate the link partner
advertisement to an ethtool bitmap.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: pcs: xpcs: clean up reading clause 73 link partner advertisement
Russell King (Oracle) [Tue, 23 May 2023 10:16:03 +0000 (11:16 +0100)]
net: pcs: xpcs: clean up reading clause 73 link partner advertisement

Read the clause 73 link partner advertisement in a loop and then
translate to the ethtool modes.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phylink: add function to resolve clause 73 negotiation
Russell King (Oracle) [Tue, 23 May 2023 10:15:58 +0000 (11:15 +0100)]
net: phylink: add function to resolve clause 73 negotiation

Add a function to resolve clause 73 negotiation according to the
priority resolution function described in clause 73.3.6.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phylink: remove duplicated linkmode pause resolution
Russell King (Oracle) [Tue, 23 May 2023 10:15:53 +0000 (11:15 +0100)]
net: phylink: remove duplicated linkmode pause resolution

Phylink had two chunks of code virtually the same for resolving the
negotiated pause modes. Factor this down to one function.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: mdio: add clause 73 to ethtool conversion helper
Russell King (Oracle) [Tue, 23 May 2023 10:15:48 +0000 (11:15 +0100)]
net: mdio: add clause 73 to ethtool conversion helper

Add a helper to convert a clause 73 advertisement to an ethtool bitmap.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoALSA: hda/realtek: Enable headset onLenovo M70/M90
Bin Li [Wed, 24 May 2023 11:37:55 +0000 (19:37 +0800)]
ALSA: hda/realtek: Enable headset onLenovo M70/M90

Lenovo M70/M90 Gen4 are equipped with ALC897, and they need
ALC897_FIXUP_HEADSET_MIC_PIN quirk to make its headset mic work.
The previous quirk for M70/M90 is for Gen3.

Signed-off-by: Bin Li <bin.li@canonical.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20230524113755.1346928-1-bin.li@canonical.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
18 months agoMerge tag 'asoc-fix-v6.4-rc3' of https://git.kernel.org/pub/scm/linux/kernel/git...
Takashi Iwai [Wed, 24 May 2023 10:14:24 +0000 (12:14 +0200)]
Merge tag 'asoc-fix-v6.4-rc3' of https://git./linux/kernel/git/broonie/sound into for-linus

ASoC: Fixes for v6.4

A collection of fixes for v6.4, mostly driver specific but there's also
one fix for DPCM to avoid incorrectly repeated calls to prepare() which
can trigger issues on some systems.

18 months agoMerge branch 'devlink-port_del-new-cleanup'
David S. Miller [Wed, 24 May 2023 09:34:26 +0000 (10:34 +0100)]
Merge branch 'devlink-port_del-new-cleanup'

Jiri Pirko says:

====================
devlink: small port_new/del() cleanup

This patchset cleans up couple of leftovers after recent devlink locking
changes. Previously, both port_new/dev() commands were called without
holding instance lock. Currently all devlink commands are called with
instance lock held.

The first patch just removes redundant port notification.
The second one removes couple of outdated comments.
The last patch changes port_dev() to have devlink_port pointer as an arg
instead of port_index, which makes it similar to the rest of port
related ops.
====================

Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agodevlink: pass devlink_port pointer to ops->port_del() instead of index
Jiri Pirko [Tue, 23 May 2023 12:38:01 +0000 (14:38 +0200)]
devlink: pass devlink_port pointer to ops->port_del() instead of index

Historically there was a reason why port_dev() along with for example
port_split() did get port_index instead of the devlink_port pointer.
With the locking changes that were done which ensured devlink instance
mutex is hold for every command, the port ops could get devlink_port
pointer directly. Change the forgotten port_dev() op to be as others
and pass devlink_port pointer instead of port_index.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agodevlink: remove no longer true locking comment from port_new/del()
Jiri Pirko [Tue, 23 May 2023 12:38:00 +0000 (14:38 +0200)]
devlink: remove no longer true locking comment from port_new/del()

All commands are called holding instance lock. Remove the outdated
comment that says otherwise.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agodevlink: remove duplicate port notification
Jiri Pirko [Tue, 23 May 2023 12:37:59 +0000 (14:37 +0200)]
devlink: remove duplicate port notification

The notification about created port is send from devl_port_register()
function called from ops->port_new(). No need to send it again here,
so remove the call and the helper function.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge branch 'tools-ynl-byteorder'
David S. Miller [Wed, 24 May 2023 07:46:54 +0000 (08:46 +0100)]
Merge branch 'tools-ynl-byteorder'

Donald Hunter says:

====================
tools: ynl: Add byte-order support for struct members

This patchset adds support to ynl for handling byte-order in struct
members. The first patch is a refactor to use predefined Struct() objects
instead of generating byte-order specific formats on the fly. The second
patch adds byte-order handling for struct members.
====================

Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agotools: ynl: Handle byte-order in struct members
Donald Hunter [Tue, 23 May 2023 09:37:48 +0000 (10:37 +0100)]
tools: ynl: Handle byte-order in struct members

Add support for byte-order in struct members in the genetlink-legacy
spec.

Signed-off-by: Donald Hunter <donald.hunter@gmail.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agotools: ynl: Use dict of predefined Structs to decode scalar types
Donald Hunter [Tue, 23 May 2023 09:37:47 +0000 (10:37 +0100)]
tools: ynl: Use dict of predefined Structs to decode scalar types

Use a dict of predefined Struct() objects to decode scalar types in native,
big or little endian format. This removes the repetitive code for the
scalar variants and ensures all the signed variants are supported.

Signed-off-by: Donald Hunter <donald.hunter@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoipv6: Fix out-of-bounds access in ipv6_find_tlv()
Gavrilov Ilia [Tue, 23 May 2023 08:29:44 +0000 (08:29 +0000)]
ipv6: Fix out-of-bounds access in ipv6_find_tlv()

optlen is fetched without checking whether there is more than one byte to parse.
It can lead to out-of-bounds access.

Found by InfoTeCS on behalf of Linux Verification Center
(linuxtesting.org) with SVACE.

Fixes: c61a40432509 ("[IPV6]: Find option offset by type.")
Signed-off-by: Gavrilov Ilia <Ilia.Gavrilov@infotecs.ru>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge tag 'mlx5-fixes-2023-05-22' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Wed, 24 May 2023 07:40:14 +0000 (08:40 +0100)]
Merge tag 'mlx5-fixes-2023-05-22' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-fixes-2023-05-22

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

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: phy: avoid kernel warning dump when stopping an errored PHY
Russell King (Oracle) [Mon, 22 May 2023 15:58:08 +0000 (16:58 +0100)]
net: phy: avoid kernel warning dump when stopping an errored PHY

When taking a network interface down (or removing a SFP module) after
the PHY has encountered an error, phy_stop() complains incorrectly
that it was called from HALTED state.

The reason this is incorrect is that the network driver will have
called phy_start() when the interface was brought up, and the fact
that the PHY has a problem bears no relationship to the administrative
state of the interface. Taking the interface administratively down
(which calls phy_stop()) is always the right thing to do after a
successful phy_start() call, whether or not the PHY has encountered
an error.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge branch 'RTO_ONLINK'
David S. Miller [Wed, 24 May 2023 07:22:06 +0000 (08:22 +0100)]
Merge branch 'RTO_ONLINK'

Guillaume Nault says:

====================
ipv4: Remove RTO_ONLINK from udp, ping and raw sockets.

udp_sendmsg(), ping_v4_sendmsg() and raw_sendmsg() use similar patterns
for restricting their route lookup to on-link hosts. Although they use
slightly different code, they all use RTO_ONLINK to override the least
significant bit of their tos value.

RTO_ONLINK is used to restrict the route scope even when the scope is
set to RT_SCOPE_UNIVERSE. Therefore it isn't necessary: we can properly
set the scope to RT_SCOPE_LINK instead.

Removing RTO_ONLINK will allow to convert .flowi4_tos to dscp_t in the
future, thus allowing to properly separate the DSCP from the ECN bits
in the networking stack.

This patch series defines a common helper to figure out what's the
scope of the route lookup. This unifies the way udp, ping and raw
sockets get their routing scope and removes their dependency on
RTO_ONLINK.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoudp: Stop using RTO_ONLINK.
Guillaume Nault [Mon, 22 May 2023 14:38:07 +0000 (16:38 +0200)]
udp: Stop using RTO_ONLINK.

Use ip_sendmsg_scope() to properly initialise the scope in
flowi4_init_output(), instead of overriding tos with the RTO_ONLINK
flag. The objective is to eventually remove RTO_ONLINK, which will
allow converting .flowi4_tos to dscp_t.

Now that the scope is determined by ip_sendmsg_scope(), we need to
check its result to set the 'connected' variable.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoraw: Stop using RTO_ONLINK.
Guillaume Nault [Mon, 22 May 2023 14:38:02 +0000 (16:38 +0200)]
raw: Stop using RTO_ONLINK.

Use ip_sendmsg_scope() to properly initialise the scope in
flowi4_init_output(), instead of overriding tos with the RTO_ONLINK
flag. The objective is to eventually remove RTO_ONLINK, which will
allow converting .flowi4_tos to dscp_t.

The MSG_DONTROUTE and SOCK_LOCALROUTE cases were already handled by
raw_sendmsg() (SOCK_LOCALROUTE was handled by the RT_CONN_FLAGS*()
macros called by get_rtconn_flags()). However, opt.is_strictroute
wasn't taken into account. Therefore, a side effect of this patch is to
now honour opt.is_strictroute, and thus align raw_sendmsg() with
ping_v4_sendmsg() and udp_sendmsg().

Since raw_sendmsg() was the only user of get_rtconn_flags(), we can now
remove this function.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoping: Stop using RTO_ONLINK.
Guillaume Nault [Mon, 22 May 2023 14:37:57 +0000 (16:37 +0200)]
ping: Stop using RTO_ONLINK.

Define a new helper to figure out the correct route scope to use on TX,
depending on socket configuration, ancillary data and send flags.

Use this new helper to properly initialise the scope in
flowi4_init_output(), instead of overriding tos with the RTO_ONLINK
flag.

The objective is to eventually remove RTO_ONLINK, which will allow
converting .flowi4_tos to dscp_t.

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: ethernet: mtk_eth_soc: fix QoS on DSA MAC on non MTK_NETSYS_V2 SoCs
Arınç ÜNAL [Mon, 22 May 2023 10:57:43 +0000 (13:57 +0300)]
net: ethernet: mtk_eth_soc: fix QoS on DSA MAC on non MTK_NETSYS_V2 SoCs

The commit c6d96df9fa2c ("net: ethernet: mtk_eth_soc: drop generic vlan rx
offload, only use DSA untagging") makes VLAN RX offloading to be only used
on the SoCs without the MTK_NETSYS_V2 ability (which are not just MT7621
and MT7622). The commit disables the proper handling of special tagged
(DSA) frames, added with commit 87e3df4961f4 ("net-next: ethernet:
mediatek: add CDM able to recognize the tag for DSA"), for non
MTK_NETSYS_V2 SoCs when it finds a MAC that does not use DSA. So if the
other MAC uses DSA, the CDMQ component transmits DSA tagged frames to the
CPU improperly. This issue can be observed on frames with TCP, for example,
a TCP speed test using iperf3 won't work.

The commit disables the proper handling of special tagged (DSA) frames
because it assumes that these SoCs don't use more than one MAC, which is
wrong. Although I made Frank address this false assumption on the patch log
when they sent the patch on behalf of Felix, the code still made changes
with this assumption.

Therefore, the proper handling of special tagged (DSA) frames must be kept
enabled in all circumstances as it doesn't affect non DSA tagged frames.

Hardware DSA untagging, introduced with the commit 2d7605a72906 ("net:
ethernet: mtk_eth_soc: enable hardware DSA untagging"), and VLAN RX
offloading are operations on the two CDM components of the frame engine,
CDMP and CDMQ, which connect to Packet DMA (PDMA) and QoS DMA (QDMA) and
are between the MACs and the CPU. These operations apply to all MACs of the
SoC so if one MAC uses DSA and the other doesn't, the hardware DSA
untagging operation will cause the CDMP component to transmit non DSA
tagged frames to the CPU improperly.

Since the VLAN RX offloading feature configuration was dropped, VLAN RX
offloading can only be used along with hardware DSA untagging. So, for the
case above, we need to disable both features and leave it to the CPU,
therefore software, to untag the DSA and VLAN tags.

So the correct way to handle this is:

For all SoCs:

Enable the proper handling of special tagged (DSA) frames
(MTK_CDMQ_IG_CTRL).

For non MTK_NETSYS_V2 SoCs:

Enable hardware DSA untagging (MTK_CDMP_IG_CTRL).
Enable VLAN RX offloading (MTK_CDMP_EG_CTRL).

When a non MTK_NETSYS_V2 SoC MAC does not use DSA:

Disable hardware DSA untagging (MTK_CDMP_IG_CTRL).
Disable VLAN RX offloading (MTK_CDMP_EG_CTRL).

Fixes: c6d96df9fa2c ("net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging")
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agodocs: netdev: document the existence of the mail bot
Jakub Kicinski [Mon, 22 May 2023 23:09:03 +0000 (16:09 -0700)]
docs: netdev: document the existence of the mail bot

We had a good run, but after 4 weeks of use we heard someone
asking about pw-bot commands. Let's explain its existence
in the docs. It's not a complete documentation but hopefully
it's enough for the casual contributor. The project and scope
are in flux so the details would likely become out of date,
if we were to document more in depth.

Link: https://lore.kernel.org/all/20230522140057.GB18381@nucnuc.mle/
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230522230903.1853151-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agogve: Support IPv6 Big TCP on DQ
Coco Li [Mon, 22 May 2023 20:15:52 +0000 (13:15 -0700)]
gve: Support IPv6 Big TCP on DQ

Add support for using IPv6 Big TCP on DQ which can handle large TSO/GRO
packets. See https://lwn.net/Articles/895398/. This can improve the
throughput and CPU usage.

Perf test result:
ip -d link show $DEV
gso_max_size 185000 gso_max_segs 65535 tso_max_size 262143 tso_max_segs 65535 gro_max_size 185000

For performance, tested with neper using 9k MTU on hardware that supports 200Gb/s line rate.

In single streams when line rate is not saturated, we expect throughput improvements.
When the networking is performing at line rate, we expect cpu usage improvements.

Tcp_stream (unidirectional stream test, T=thread, F=flow):
skb=180kb, T=1, F=1, no zerocopy: throughput average=64576.88 Mb/s, sender stime=8.3, receiver stime=10.68
skb=64kb,  T=1, F=1, no zerocopy: throughput average=64862.54 Mb/s, sender stime=9.96, receiver stime=12.67
skb=180kb, T=1, F=1, yes zerocopy:  throughput average=146604.97 Mb/s, sender stime=10.61, receiver stime=5.52
skb=64kb,  T=1, F=1, yes zerocopy:  throughput average=131357.78 Mb/s, sender stime=12.11, receiver stime=12.25

skb=180kb, T=20, F=100, no zerocopy:  throughput average=182411.37 Mb/s, sender stime=41.62, receiver stime=79.4
skb=64kb,  T=20, F=100, no zerocopy:  throughput average=182892.02 Mb/s, sender stime=57.39, receiver stime=72.69
skb=180kb, T=20, F=100, yes zerocopy: throughput average=182337.65 Mb/s, sender stime=27.94, receiver stime=39.7
skb=64kb,  T=20, F=100, yes zerocopy: throughput average=182144.20 Mb/s, sender stime=47.06, receiver stime=39.01

Signed-off-by: Ziwei Xiao <ziweixiao@google.com>
Signed-off-by: Coco Li <lixiaoyan@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230522201552.3585421-1-ziweixiao@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: fix skb leak in __skb_tstamp_tx()
Pratyush Yadav [Mon, 22 May 2023 15:30:20 +0000 (17:30 +0200)]
net: fix skb leak in __skb_tstamp_tx()

Commit 50749f2dd685 ("tcp/udp: Fix memleaks of sk and zerocopy skbs with
TX timestamp.") added a call to skb_orphan_frags_rx() to fix leaks with
zerocopy skbs. But it ended up adding a leak of its own. When
skb_orphan_frags_rx() fails, the function just returns, leaking the skb
it just cloned. Free it before returning.

This bug was discovered and resolved using Coverity Static Analysis
Security Testing (SAST) by Synopsys, Inc.

Fixes: 50749f2dd685 ("tcp/udp: Fix memleaks of sk and zerocopy skbs with TX timestamp.")
Signed-off-by: Pratyush Yadav <ptyadav@amazon.de>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20230522153020.32422-1-ptyadav@amazon.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge branch 'splice-net-replace-sendpage-with-sendmsg-msg_splice_pages-part-1'
Jakub Kicinski [Wed, 24 May 2023 03:48:53 +0000 (20:48 -0700)]
Merge branch 'splice-net-replace-sendpage-with-sendmsg-msg_splice_pages-part-1'

David Howells says:

====================
splice, net: Replace sendpage with sendmsg(MSG_SPLICE_PAGES), part 1

Here's the first tranche of patches towards providing a MSG_SPLICE_PAGES
internal sendmsg flag that is intended to replace the ->sendpage() op with
calls to sendmsg().  MSG_SPLICE_PAGES is a hint that tells the protocol
that it should splice the pages supplied if it can and copy them if not.

This will allow splice to pass multiple pages in a single call and allow
certain parts of higher protocols (e.g. sunrpc, iwarp) to pass an entire
message in one go rather than having to send them piecemeal.  This should
also make it easier to handle the splicing of multipage folios.

A helper, skb_splice_from_iter() is provided to do the work of splicing or
copying data from an iterator.  If a page is determined to be unspliceable
(such as being in the slab), then the helper will give an error.

Note that this facility is not made available to userspace and does not
provide any sort of callback.

This set consists of the following parts:

 (1) Define the MSG_SPLICE_PAGES flag and prevent sys_sendmsg() from being
     able to set it.

 (2) Add an extra argument to skb_append_pagefrags() so that something
     other than MAX_SKB_FRAGS can be used (sysctl_max_skb_frags for
     example).

 (3) Add the skb_splice_from_iter() helper to handle splicing pages into
     skbuffs for MSG_SPLICE_PAGES that can be shared by TCP, IP/UDP and
     AF_UNIX.

 (4) Implement MSG_SPLICE_PAGES support in TCP.

 (5) Make do_tcp_sendpages() just wrap sendmsg() and then fold it in to its
     various callers.

 (6) Implement MSG_SPLICE_PAGES support in IP and make udp_sendpage() just
     a wrapper around sendmsg().

 (7) Implement MSG_SPLICE_PAGES support in IP6/UDP6.

 (8) Implement MSG_SPLICE_PAGES support in AF_UNIX.

 (9) Make AF_UNIX copy unspliceable pages.

Link: https://lore.kernel.org/r/20230316152618.711970-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230329141354.516864-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230331160914.1608208-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230405165339.3468808-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230406094245.3633290-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230411160902.4134381-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230515093345.396978-1-dhowells@redhat.com/
Link: https://lore.kernel.org/r/20230518113453.1350757-1-dhowells@redhat.com/
====================

Link: https://lore.kernel.org/r/20230522121125.2595254-1-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agounix: Convert unix_stream_sendpage() to use MSG_SPLICE_PAGES
David Howells [Mon, 22 May 2023 12:11:25 +0000 (13:11 +0100)]
unix: Convert unix_stream_sendpage() to use MSG_SPLICE_PAGES

Convert unix_stream_sendpage() to use sendmsg() with MSG_SPLICE_PAGES
rather than directly splicing in the pages itself.

This allows ->sendpage() to be replaced by something that can handle
multiple multipage folios in a single transaction.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Kuniyuki Iwashima <kuniyu@amazon.com>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoaf_unix: Support MSG_SPLICE_PAGES
David Howells [Mon, 22 May 2023 12:11:24 +0000 (13:11 +0100)]
af_unix: Support MSG_SPLICE_PAGES

Make AF_UNIX sendmsg() support MSG_SPLICE_PAGES, splicing in pages from the
source iterator if possible and copying the data in otherwise.

This allows ->sendpage() to be replaced by something that can handle
multiple multipage folios in a single transaction.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Kuniyuki Iwashima <kuniyu@amazon.com>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoip: Remove ip_append_page()
David Howells [Mon, 22 May 2023 12:11:23 +0000 (13:11 +0100)]
ip: Remove ip_append_page()

ip_append_page() is no longer used with the removal of udp_sendpage(), so
remove it.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoudp: Convert udp_sendpage() to use MSG_SPLICE_PAGES
David Howells [Mon, 22 May 2023 12:11:22 +0000 (13:11 +0100)]
udp: Convert udp_sendpage() to use MSG_SPLICE_PAGES

Convert udp_sendpage() to use sendmsg() with MSG_SPLICE_PAGES rather than
directly splicing in the pages itself.

This allows ->sendpage() to be replaced by something that can handle
multiple multipage folios in a single transaction.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoip6, udp6: Support MSG_SPLICE_PAGES
David Howells [Mon, 22 May 2023 12:11:21 +0000 (13:11 +0100)]
ip6, udp6: Support MSG_SPLICE_PAGES

Make IP6/UDP6 sendmsg() support MSG_SPLICE_PAGES.  This causes pages to be
spliced from the source iterator if possible, copying the data if not.

This allows ->sendpage() to be replaced by something that can handle
multiple multipage folios in a single transaction.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoip, udp: Support MSG_SPLICE_PAGES
David Howells [Mon, 22 May 2023 12:11:20 +0000 (13:11 +0100)]
ip, udp: Support MSG_SPLICE_PAGES

Make IP/UDP sendmsg() support MSG_SPLICE_PAGES.  This causes pages to be
spliced from the source iterator.

This allows ->sendpage() to be replaced by something that can handle
multiple multipage folios in a single transaction.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agotcp: Fold do_tcp_sendpages() into tcp_sendpage_locked()
David Howells [Mon, 22 May 2023 12:11:19 +0000 (13:11 +0100)]
tcp: Fold do_tcp_sendpages() into tcp_sendpage_locked()

Fold do_tcp_sendpages() into its last remaining caller,
tcp_sendpage_locked().

Signed-off-by: David Howells <dhowells@redhat.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>